Full text of "The American naturalist"

See other formats

This is a digital copy of a book that was preserved for generations on library shelves before it was carefully scanned by Google as part of a project
to make the world's books discoverable online.

It has survived long enough for the copyright to expire and the book to enter the public domain. A public domain book is one that was never subject
to copyright or whose legal copyright term has expired. Whether a book is in the public domain may vary country to country. Public domain books
are our gateways to the past, representing a wealth of history, culture and knowledge that's often difficult to discover.

Marks, notations and other marginalia present in the original volume will appear in this file - a reminder of this book's long journey from the
publisher to a library and finally to you.

Usage guidelines

Google is proud to partner with libraries to digitize public domain materials and make them widely accessible. Public domain books belong to the
public and we are merely their custodians. Nevertheless, this work is expensive, so in order to keep providing this resource, we have taken steps to
prevent abuse by commercial parties, including placing technical restrictions on automated querying.

We also ask that you:

+ Make non-commercial use of the files We designed Google Book Search for use by individuals, and we request that you use these files for
personal, non-commercial purposes.

+ Refrain from automated querying Do not send automated queries of any sort to Google's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a large amount of text is helpful, please contact us. We encourage the
use of public domain materials for these purposes and may be able to help.

+ Maintain attribution The Google "watermark" you see on each file is essential for informing people about this project and helping them find
additional materials through Google Book Search. Please do not remove it.

+ Keep it legal Whatever your use, remember that you are responsible for ensuring that what you are doing is legal. Do not assume that just
because we believe a book is in the public domain for users in the United States, that the work is also in the public domain for users in other
countries. Whether a book is still in copyright varies from country to country, and we can't offer guidance on whether any specific use of
any specific book is allowed. Please do not assume that a book's appearance in Google Book Search means it can be used in any manner
anywhere in the world. Copyright infringement liability can be quite severe.

About Google Book Search

Google's mission is to organize the world's information and to make it universally accessible and useful. Google Book Search helps readers
discover the world's books while helping authors and publishers reach new audiences. You can search through the full text of this book on the web

at http : //books . google . com/|

The American naturalist

Essex Institute, American

Society of Naturalists, JSTOR (Organization)

Digitized by

Google

2044 106 436 397

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

THE AMERICAN NATURAUST

Digitized by

Google

Digitized by

Google

THE

AMERICAN NATURALIST

A Monthly Journal

Devoted to the Advancement of the Biological Sciences

With Special Reference to the Factors of Evolution

VOLUME XLVIII

NEW YORK
THE SCIENCE PRESS

1914

Digitized by

Google

PRESS OF

THE NEW ERA PRINTINO COMPANY

LANCASTER PA.

Digitized by

Google

roi. zLTm, 10. MS jahiust, i>i4

THE

AMERICAN
NATURALIST

A M0HTEL7 JOXnUTAL

Jtaroted to fhe AdTaneement of the Biological Scienoot witt

Special Beforence to the Factors of Evolntioii

OOVTBVTS

Pa§€

I. A OeiMtte Analysli of th» Ohanf m produced by 8«loetloa la BxptrlsiMiU

wltli TobMOO. Proftnor £. M. Eabt and H. K. Hatbs .... 5

n. aynaadromorpboiu AnU, dMorlbod dnring tbe Dtcade, 1908-1918

ProfeMor William Morton Whbelkb 49

m. Sliorler ArttolM and Diiouitloa :

On tbe Besnlts of Inbreeding a Mendellan Population — ^A Correction and Ex*
tendon of Prerions Conclusions. Dr. Batmomd Pbarl— Isolati<m and Selec-
tion allied in Principle. Dr. John T. Gulick - 07

THE 80IEN0E FBE88

IiAHOASTEB, PA. OABBI80H. H. T.

■ZW 70BK: SVB-STAnOH M

Digitized by

Google

The American Naturalist

MSS Intended for poblioatfon and booke, etc., Intended for review ehoufd be
sent to the Editor of THE AMERICAN NATURALIST, Garrlson-on- Hudson, New York.

Short articles oontaining eommarlet of research work bearing on the
probleas of organic evelntien are especially welcome, and will bo given preference
in publication.

One hundreo reprints of contributions are supplied to authors firee of charge.
Further reprints will be supplied at cost.

Subscriptions and advertiseaents should be sent to the publishers. The
subscription price is four dollars a year. Foreign postage is fifty cents and
Canadian postage twenty-five cents additional. The charge fbr single copies is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Qarrison, N. Y.

NEW YORK : Sub-Station 84

Bntored m teoond-elaai matter, April 2. 1906, at tht Post Offiet at Lancaitar, Pa., imder tbt At% of

Congren of Mareh 8, 1879.

FOR SALE

ARCTIC. ICELAND and GREENLAND

BIRDS' SKINS,

W«II Prepared Low Prioss

Partloulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Ensland

TENTH EDITION.

THE MICROSCOPE,

sn introduetiioii to ICieroooopie Moihodi and to Hiitology,
bj SncovHnrBT Gaob, of Oomeil UniTendty. Orer
860 largo ootSTO paaM and aboTO 250 fiicarw in this new
andltaflyxoitediditlon. Frioe 12.00, pot^aid.

COMSTOCK PUBL18H1NQ CO., Ithaca, N. Y.

TO OOLOQISTS

and ENTOnOLOQISTS

W. Fe H. ROSENBERG

Importer of Exotio Zoologioal Spedmemi
57, Haverstock HiH, London, N. W., Enfflaad

B^gB to annoiinoe the pablioation of two new
Price Usta : Na 18, Ezotie L^idoptera (over 8000
ipedea), No. 10, Birda' Egga (oyer 1100 spedci).

Tbeee will be mailed free on applioation, as well
aa any of the following liata:

Birdakins (over 6000 apeetea) ; Mammals (oyer
900 speoles); Reptiles, Batraohians and Fiahea (over
400 8peoieB).

Laigest stock in the world of specimens in all
branches of Zoology.

ALL MUSEUMS SHOULD
WRITE FOR THESE LISTS

All specimens sent on approval.
Please state which lists are required and give
name of this periodical.

PINE LAND 5HELL5

I have in itock and for sale nearly ten thoasavd spedet
of land thellt, alto a very good itock of marine. All are ac-
curately labeled. I am prepared to make up collectloni 11-
luitrating every phase of the subject and from nearly all
countries and islands. A fine illustrated booklet on the
beautifnl shells of the Philippines, and other circulars, will
be mailed free. Strietly scientific collections purchased fCr
cash. Correspondence solicited with travelers and ezplorera.
Only the finest material handled. WALTER P. WEBB,
202 Westminister Road« Rochester, N. T., U. 8. A.

Natural History

Preparations, Collections
and Models

of acknowledged first-rate style at very low
prices for Museums and Oolleges

Outfit and Completion of Natural History

Cabinets. Own Preparation works

Catalogues free of charge

Dr. Schluter's Natural Science

Establishment
Halle on S GERMANY

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVni January, 19U No. 566

A GENETIC ANALYSIS OF THE CHANGES PRO-
DUCED BY SELECTION IN EXPERIMENTS
WITH TOBACCO*

PROFESSOR E. M. EAST and H. K. HAYES
BussET Institution of Habvard Univebsitt

The Problem

In 1903 Johannsen announced that continued selection
of the extreme values of certain quantitative characters
in successive self-fertilized generations of a number of
strains of beans had produced no changes in the mean
values of the characters. He concluded that these par-
ticular strains were homozygous for the gametic factors
whose interaction resulted in the characters investigated,
that these homozygous characters may be properly de-
scribed by one or more gametic factors nonvariable in
transmissible qualities and properties, and that the varia-
tions observed in the characters of any single fraternity
were due entirely to the action of environmental condi-
tions during ontogeny and were not inherited. Funda-
mentally, these conclusions were a recognition of the gen-
eral value of Mendelian description for all forms of in-
heritance through sexual reproduction, combined with an

1 These inyestigations were conducted with funds furnished by the Con-
necticut Agricultural Experiment Station from their Adams' appropria-
tions, bj the Bureau of Plant Industry of the United States Department of
Agriculture, and by the Bussey Institution of Harvard University, and the
writers desire to take this opportunity of expressing their sincere appre-
ciation of this hearty cooperation which made the work possible.

Digitized by

Google

6 THE AMERICAN NATURALIST [V0L.XLVIII

admission of disbelief in the inheritance of ordinary
adaptive changes. The latter conception was Weismann-
ian in that all inherited variations were held to be changes
in the germ cells. It was not necessary to suppose it im-
possible for the environment to produce such changes and
therefore to have been of no value during the course of
evolution, but merely to suppose that during the compara-
tively short period of experimental investigations no gam-
etic variations have occurred traceable to such a cause.
For his first conclusion to be justified, it was assumed that
the changes which every biologist knows do follow the
continuous selection of extremes under certain conditions
are to be interpreted entirely by the segregation and re-
combination of hypothetical gametic factors which are
constant in their reactions under identical conditions.

Numerous investigators working on ''pure lines" with
different material corroborated Johannsen's conclusions,
and, as it was seen to be possible to interpret in the same
manner changes made by selection in experiments where
self-fertilized lines were not used, such as those of the
Vilmorins and others on sugar beets and those of the
Illinois Agricultural Experiment Station on maize, many
biologists accepted them and considered them a great ad-
vance over former conceptions of the mechanism of
heredity. On the other hand, there were those who main-
tained a skeptical attitude, the chief criticism directed
against the conception being that all progress due to
selection must have a limit, which in many of these ex-
periments had already been reached, and that even if re-
sults were being obtained action might be too slow to be
detected.

The Material

These criticisms were reasonable when applied to cer-
tain specific cases, and in 1908 the experiments reported
in this paper were designed with the hope of testing their
validity, using the species ordinarily grown for commer-
cial tobacco, Nicotiana tabacum, as the material. This
plant satisfies the conditions which are requisite for

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 7

material used in pure line studies. It has characters that
can be estimated readily and accurately and which are
affected only slightly by external conditions. It is easily
grown, is naturally self -fertilized, reproduces prolifically,
and is known in many markedly different varieties. In
fact, it is an ideal subject for work of this kind.

The investigations were not patterned after the stand-
ard type set by Johannsen wherein the constancy of suc-
cessive generations of pure lines grown from selected
extremes were tested, since even if it were possible to
gather a quantity of data at all comparable to that col-
lected by Johannsen ( :09) and Jennings ( :08) in their
brilliant investigations, the criticisms mentioned above
might still be made. The plan chosen was that of cross-
ing two varieties of tobacco which differed in a character
complex easily and precisely determined, and of selecting
extremes from a number of families of the F2 generation.
If Johannsen 's views be incorrect, such continued selec-
tion should affect each family in the same degree. If his
conclusions be justified, selection should reach an end-
poiat in different generations in different families, and
there should be no relation between the number of genera-
tions required to reach this end-point and the progress
that is possible.

There should be no need of a historical summary of the
previous investigations that have been interpreted as cor-
roborating or refuting Johannsen 's conclusions. Such
summaries have been made in other papers. It should be
mentioned, however, that the classical researches of Pearl
(:11) on the inheritance of fecundity in the domestic
fowl have been so planned and executed that certain of
the criticisms directed against Johannsen mentioned above
are not justified, yet Pearl finds himself thoroughly in
accord with the Danish physiologist's position.

Several hundred varieties of Nicotiana tahacum exist
which differ from each other by definite botanical char-
acters, yet only two general characters suitable for our
purpose were found. We desired to confine our observa-
tions to quantitative characters that were influenced but

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVm

little by environment, and number of leaves and size of
corolla were the only ones that satisfied this requirement.
Such character differences as height of plant and size of
leaf, while undoubtedly transmissible, are influenced so
strongly in their development by nutrition that work with
them is exceedingly diflBcult. For example, if a certain
variety of Nicotiana tahacum is grown under the best of
field conditions, the longest leaves are about 28 inches and
the total height about 6 feet, but a portion of the same
seed fraternity may be grown to maturity in 4-inch pots
without reaching a height of over 16 inches or having
leaves longer than 4 inches. On the other hand, several
experiments conducted in the same manner have shown
no difference between the frequency curves of variation
in number of leaves or of size of corolla, whether starved
in small pots or grown under optimum conditions. The
character complex number of leaves was chosen for this
investigation rather than the size of corolla because vari-
eties that differ greatly in number of leaves are common.

TABLE I

Frequency Distribution op Number op Leaves per Plant when
Starved in Small Pots

(Compare with frequency distribution under normal field conditions at
Forest Hills, Massachusetts, in Tables VU and XI)

No. of Lmtm per Plant

Plant No.

(6-l)-l

3
6

10
8

8
16
8
0
6
10

7
12

1
5
14
0
2
8

(e-2)
(6-2)-2

(se-i)

(66-2)

15
2

14
3

8
12

3
17

3
16

8
4

Previous Work of the '* Havana'' X *' Sumatra'' Cross

Several crosses have been made between varieties of
tobacco that had a mean difference of seven or eight
leaves, but the majority of the data reported here were
collected from the descendants of a cross made by A. D.
Shamel between the types known in Connecticut as
' * Havana ' ' and * ' Sumatra. ' ' The ' * Havana ' ' parent was

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 9

from a variety that had been grown for a number of years
at Granby, Connecticut. It averages about 20 leaves per
plant although ranging from 16 to 25 leaves. The aver-
age height is about 1.4 m. and the average leaf area about
7 sq. dm. The *' Sumatra'' parent was a type specimen
of a variety that had been introduced into Connecticut to
be grown under cloth shade. It averages between 26 and
27 leaves per plant with a range of from 21 to 32 leaves.
The average height is nearly 2.0 m., but the average leaf
area is only about 3 sq. dm.

According to Shamel, the first hybrid generation of
this cross developed somewhat more vigorously than the
parent types and was uniform in its habit of growth.
The second generation, he thought, was hardly more vari-
able than the first. Several F3 families, the progeny of
inbred F2 individuals, were grown in 1906 and proved to
be a variable lot. One of these plants produced 26 small,
round-pointed leaves with short intemodes between them.
This plant was thought by Mr. E. Halladay, upon whose
farm the experiment was conducted, and Mr. J. B. Stewart,
of the U. S. Department of Agriculture, to be worth sav-
ing from its promise of producing a desirable commercial
type.

In 1907 the Department of Agriculture made an agree-
ment with Mr. Halladay to grow two acres of tobacco for
experimental purposes, and on his own initiative Mr.
Halladay grew a number of plants from inbred seed of
the one that bore 26 leaves. This selection, numbered 2
h-29 in accordance with the department nomenclature,
was comparatively uniform in appearance and several
plants were selfed. In Mr. Halladay 's absence, how-
ever, all of the plants were ** topped,'' except one that
happened to be rather late. This plant was selfed. It
had 26 medium-sized, round leaves and grew to about the
same height as the Connecticut Havana.

In view of Mr. Halladay 's high opinion of the type, the
seed of this plant and the remaining seed of its parent
were planted in 1908. The plants of this generation pre-
sented a uniform appearance and promised a high grade

Digitized by

Google

10 THE AMERICAN NATURALIST [VoL.XLVra

of wrapper tobacco, but the crop when cured lacked uni-
formity. Some leaves of exceptionally high quality were
produced, but the crop in general lacked that characteris-
tic known as ^* grain'' and had too large a proportion of
heavy leaves — the so-called *'tops."

From this 1908 generation 100 seed plants were selfed,
their leaves harvested, cured and fermented separately,
and data on quality recorded. The type was also grown
conunercially on a large scale. The commercial results,
however, have been reported in another paper. We are to
consider only the results gf the selection experiment that
began in 1908, through the cooperation between the U. S.
Department of Agriculture and the Connecticut Agricul-
tural Experiment Station, a joining of forces that in 1909
included the Bussey Institution of Harvard University.
Shamel ( :07) considered the strain produced by this cross
to be the result of a mutation. From a study of the
data from the previous work on the cross it seemed to the
writers that a different interpretation of the results might
be made. While it was not impossible that the many-
leaved type that had been isolated was the result of a
mutation, it appeared much more probable that it had
arisen through a recombination of Mendelian factors.
The type had the habit of growth and size of leaf of the
pure '* Havana" variety and the number of leaves of the
''Sumatra" variety, a combination that might reason-
ably be expected to be the result of the Mendelian law.

Results on the Reciprocal Cross, ''Sumatra"
X ^'Havana"
To test the hypothesis that the new tobacco was the
result of such recombination and could be reproduced
whenever desired, the reciprocal of the original cross was
made in 1910. The female parent, "Sumatra," was the
direct descendant of a sister of the plant used as the
male parent of the original cross by Shamel in 1903
through seven generations of selfed plants. The male
parent, "Havana," was from the commercial field of the
Windsor Tobacco Growers' Corporation at Bloomfield,

Digitized by

Google

No. 566] CHANGES PRODUCED BY SELECTION 1 1

Connecticut. It was a descendant in a collateral line of
the plant used by Shamel in 1903 as the female parent in
his cross.

Table n, giving the frequency distribution for the num-
ber of leaves of the two parents and the first and the
second hybrid generations, is a complete justification of
our prediction as to how the hybrid type produced by
Shamel originated. The ** Sumatra" and the F^ genera-
tion were grown at New Haven, Connecticut, in 1911, the
*' Havana" was grown at Bloomfield, Connecticut, in 1911
from conmaercial seed of the same variety as the plant
used for the male parent, while the Fg generation was
grown at New Haven, Connecticut, in 1912. The F, gen-
eration, producing an average of 23.3 ±: .14 leaves per
plant, is intermediate in leaf number, since the * * Havana ' '
variety shows an average leaf number per plant of 19.8
± .08 and the *^ Sumatra" variety 26.5 ±: .11. The varia-
tion as determined by the coeflBcient of variability is some-
what less for the F^ than for either parent. The value
for the ** Sumatra" variety is 6.64 per cent. ±.28 per
cent, for the ** Havana" variety 6.98 per cent. ±: .27 per
cent, and for the F^ generation 6.24 per cent. ± .41 per
cent. Taking into consideration the probable error in
each case, one may say that the variability of the three
populations is almost the same.

The variability of the F2 generation, however, is greatly
increased. This is shown by the high coefficient of vari-
ability, 10.29 ±: .23 per cent., although a glance at the fre-
quency distribution with its range of from 18 to 31 leaves
brings home the point without recourse to biometrical
calculation.

The appearance of the plants in the field corroborated
the data of Table II in other characters. The Fi genera-
tion was intermediate in the various leaf characters, such
as shape, size and texture, that distinguish '* Sumatra"
from * * Havana ' ' tobacco, and in these characters it seemed
as uniform as either of the parental varieties. On the other
hand, the Fg generation was extremely variable. Some
plants could not be distinguished from the pure *'Suma-

Digitized by

Google

12 THE AMERICAN NATURALIST [VoL-XLVm

tra,'' others resembled ** Havana/' although of course the
majority were intermediate in various degrees. Several
plants combined the leaf size and habit of growth of the
** Havana" parent with the leaf number of the ** Suma-
tra'' parent. In other words, plcmts were produced in
the F2 generation by the recombination of Mendelian fac-
tors that exactly repeated the type which Shamel had ob-
tained in the F^ generation of the reciprocal cross made
in 1903 and which he thought was due to a mutation.
This fulfilled adequately the prediction made by us in
1908.

Results of Selecting fob High Numbeb and Low Num-
ber OF Leaves in the ** Havana" X ** Sumatra"

Cross

Li describing the reproduction of Shamel 's hybrid with
numerous large leaves by a reciprocal cross, there has
been a chronological inversion. This was done simply to
show that the original hybrid known commercially as
**The Halladay" was actually a recombination of Men-
delian factors in which the ** Havana" and the ** Suma-
tra" varieties differed. We will now describe the effects
of selection on the original '* Halladay hybrid."

It wiU be recalled that the selection experiment which
is the principal subject of this paper began with the self-
ing of 100 seed plants of Shamel's Halladay hybrid in
1908. These plants were the F4 and F^ generations of the
cross ''Havana" X ''Sumatra." Plants numbered from
1 to 49 were the F4 generation ; those numbered from 50
to 100 were the F5 generation. They were apparently
breeding true for the short habit of growth and large-
sized leaf of the "Havana" parent and the goodly num-
ber of leaves of the "Sumatra" parent. The casual ob-
server either would have said with Shamel that here was
a mutation breeding as true as any tobacco variety, or
that a fixed hybrid, a hybrid homozygous in all of its
gametic factors, had been produced. Accurate data
taken on the progeny of those of the F4 and Fg seed plants
which it was possible for us to grow in our limited space,

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 13

however, show that such judgments would have been
superficial. The general type of the plant did appear to
be fixed, but the frequency distribution for number of
leaves of the F^ and F^ populations were not the same.
Strictly speaking, they were not fixed. What would be
the result of selecting (and selfing) extremes from these
ddflferent families for a number of years? A tentative
answer to this question is to be obtained by examining
the remainder of our tables.

The tables are arranged roughly in the order of the
effect that selection has had in changing the mean of the
various families that were the starting points of this part
of the experiment. The selections were grown near Bloom-
field, Connecticut, on the light sandy loam of that region,
soil typical of that which produces the famous Connecti-
cut River Valley wrapper tobacco. Duplicate experi-
ments with several of the original families were made at
New Haven, Connecticut, however, on an impoverished
soil not fitted to grow a good quality of tobacco even after
supplying large quantities of tobacco fertilizer, and in
the condition used not fitted to grow good crops of any
kind. Two families were also grown in triplicate, the
third selections being planted at Forest Hills, Massachu-
setts, on a very fine type of rich garden land which brought
out maximum luxuriance of growth, but which did not
produce good tobacco quality. These experiments were
not true repetitions of the experiments at Bloomfield,
Connecticut, since aliquot portions of the seed from the
selfed plant grown there were not sent to the other places
to be grown. But they were duplicates in that each
family came from the same F4 or Fg mother plant,
although, beginning with the Fg or Fe population, differ-
ent selfed seed plants furnished the starting point of selec-
tions carried on independently. In this way there were
afforded a greater number of chances to see what selec-
tion could do.

Table III shows the results obtained from family No.
77. This family arose from an F5 plant having 23 leaves,
one below the modal leaf number if we may juTJge from

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVTEI

C.V.

4 4 1 HI

CO CO CO c

1.38 *.05
1.76*. 08
1.45*. 10
2.51*. 06

19.8*. 08
26.5*. 11
23.3*. 14
24.4*. 08

E
I

■^ :c^

•^ :r-

U3 :oo

:2 ;a

;s-§

^a^'s

iSj-^K

-22S

•«28

0» 00

«s

e»

■$

r-t

]

\
X

fli

S
^

8.37*. 26
6.60*. 18
9.27*. 27
6.38*. 33
9.32*. 30
6.92*. 20
7.48*. 21

1.54*. 05
1.41*. 04
2.03*. 06
1.35*. 07
2.32*. 07
1.84*. 05
1.93*. 06

18.4*. 07
21.3*. 05
21.9*. 08
22.4 *.ll
24.9 *.ll
26.6*. 07
25.8 *.08

WOX

||8g^g8

• iH fH

fH fH i-l

d ^^

lo d CO

*S2

258

'^-ssss

«*<=g8S

-^^s^^ggs

-"2:825? S!5

^gfeSSSS

^S5;s:2<^»«

^sss::

ssss^

sss

l^ d UD

»«•

CO
fH

JOIM

sssss§g

TUkOlOJ^X

d ^ O O O ^ d

1-1 ^4 tH O ^^ ^^ ^^

aon«i»o»o

^££££££

w^ y^ fH fH

Digitized by

Google

No. 565]

CHANGES PRODUCED BY SELECTION

c/3

l»H>i

JO jaqmpfj

VLAOIOJWA

I -H -B -N -H -H «
r* -H i^ 00 iQ CO 1-1
^ CO 00 CO c^ C9 1>-
co lO 00 CO oi r^ t>I

HI HI 1 1 Hi 1 1

,H fH ^ 1-1 CJ f-4 C4

•H -I -H HI* V -l' -M*

0> r>; f^ -^ -^ lO r>;

8 e5 SS N S S S

S 5 »H S Q 5 S

^ ^ e< ^ fH ^

t fH CO ^ *-«

d CO »H

fH <o r^o CO N

r^o CO N

d ^ CO -<

SS5S2"

N .H O » O 1-1 w
^^ tH ^^ Q fH f-4 1-4

9 9 O^ 9 0> O) O)

aoDu»a»{)

^ ^ pEt (Z4 pE4 (Z4 pt,

III V S* V

f^ r^ r^ r^ i^ i^ i^

O Q

^U3

sas

fl fl -H ^l -H -H

SfeSSSS

toco CO CO CO CO

82§8Sg

Hi HI HI HI HI HI

(/)

B^!^^!

SSSfegfe

HI HI HI HI HI HI

0> M — ■»|l -< 0»

SiSSiSSS

O '-^ o> ^ o kO

*;HCO ooScob-

CO ?l <N <N

• ^^ fH

• «-l W

^tO^

-"SS

-*ss

OJ CO ^ o »0 OJ

•s

2»SSSSg

s»sgss

S'SSS£

S
a

««2§»''

-^ O Oi CO C< -4

^o^ ^ :

w^ w^ w^ . . .

ssssssss

21-5%

», a

»H o g o •-« c<

0>0» O) O) O) O)

« t« « t« <» *

PEiPM^M^fob

•-<

»-« i-i «-4

-^ ^ M CI C^

^^^j>ii

r^ t^ r^ r^ r^ r^

www|

Digitized by

Google

16 THE AMERICAN NATURALIST [VoL.XLVin

the F2 generation of the reciprocal cross where the mode
was at 24 to 25 leaves. The F^ fraternity that it pro-
duced was somewhat smaller than one would wish if
he were to be confident of the calculations made. The
mode is 22 leaves and the mean nearly the same, 22.4
± .11 leaves. From among these plants, a minus variant
having 20 leaves and a plus variant having 27 leaves were
selected to produce the F^ generation. The modes in this
generation are 21 and 25 leaves, respectively, a difference
of 4 leaves ; and the means are 21.9 ± .08 and 24.9 ± .11
leaves, respectively, a difference of 3 leaves. Progress in
both directions continued when a 20-leaved plant was
selected to carry on the minus strain, and a 30-leaved
plant was selected to carry on the plus strain. The modal
classes of the Fg generation are 21 leaves in the minus
selection and 26 leaves in the plus selection, while the
means are 21.3 ±: .05 leaves and 26.6 ± .07 leaves, respect-
ively. In the Fg generation the plus selection was lost,
but the minus selection grown from a 20-leaved plant had
the mode dropped to 18 leaves and the mean to 18.4 ± .08
leaves. In order not to lose the plus selection entirely,
however, more of the Fg generation seed was grown in
1912. The mode is the same as in 1911, but the mean
dropped slightly to 25.8 ±: .08 leaves.

Here one notices what is very common throughout the
experiment ; the extremes selected for mother plants were
not members of the most extreme classes. This means
simply that vigorous healthy specimens were always
selected as the mother plants, and often the most extreme
variants did not come up to the standard. It is hardly
just to criticize this procedure, however, for with the best
care that it was possible to give, the experiments with
several families were terminated on account of non-
germination of seed or for some similar reason, it being
impossible, on account of the pressure of other work, to
self many plants in each selection. Even where seed
from several mother plants was collected, it did not in-
sure the continuation of that selection. The necessary
space and care involved in growing so many seedlings in

Digitized by

Google

No. 565] CHANGES PBODUOED BY SELECTION

S5^8|

1 4 -a 1 4

ssass

t^fOfOt^tO

^.SM^.B.

HI 1 HI HI 4

feggSS

fH iH fH O fH

8SSSS

« HI 4 4 4

Ttm

sasss

»H .

'»H01

««

»<e«

1
I

I
1

rt e«

"s^as

"SSS?!

s^sias

ss^s:!!

S5!g22

ggg«e,

S2S»-

2««

»H

ssssss

ilt

1-4090 ^
«H vH S 1-4 1-1

0»0»90»0»

>'o

jfl

•>•«••>

Pe,|s,Pe,|s,Pe,

7 7

-»-\*-S ^^-'^

^1-4 04 04

J,,j,aJ,,j,

^ ^ vH ^ vH

'%•

s«

29255288

4 4 HI -R 1 -H 1

SS853SSS^

^' 00 id <d id <d «

4 -H HI -H 1 HI -H

gssgggg

^ CJ «H «H 1-4 fH ej

1 hi' hi' 4 4 1 HI*

coi-«oieo^«o

vH ^ ,^ *H «H vH

fH«

ot*

CO 04

•H

''^-S

^«

""SS

sasg

s^'-ssstsa

SSSSSSSJ*

gSSSIgSS"

"Sfeaa**-

eoggjg.^

,»io<o

t^e^^

fH fH »H

vH . •

5 "5

ssssssss

*H fH .-4 O fH *H fH

2222222

£p:i£g£££

»H

«H fH 04
fH «H *H ^

■i

Digitized by

Google

18 THE AMERICAN NATURALIST [Voi^XLVni

isolated seed pans filled with sterilized soil made it im-
possible to start more than two sets of plants for each
plus and each minus selection. Generally both sets grew
perfectly, but occasionally both failed, and in that case it
was usually too late in the season to start a third set even
if it were available.

The second part of Table III shows the results obtained
on the poor soil of New Haven, Connecticut, with the same
family. There was continuous progress in both direc-
tions. The minus selections during the three generations
show a constant reduction of mode, the figures being 23,
22 and 21; the plus selections show an even greater in-
crease in mode, the figures being 25, 27 and 28. The same
decrease and increase occur in the means until in the F^
generation there is a difference of nearly 9 leaves, the cal-
culated means being 20.9db.08 leaves and 29.7±.14 leaves,
respectively.

Figs. 1 and 2 show typical plants of the plus and minus
strains of this family as developed by 3 years of selection.
Fig. 3 illustrates an interesting change of phyllotaxy in
some plants of (77--2)-l-l as grownat New Haven in 1912.

Passing to the data on Family No. 76 (Table IV) there
is the same evidence of the effectiveness of selection, ex-
cluding the minus strain in 1910, of which only 31 plants
were healthy. This effect is markedly less than with the
other family. The mode of the minus selection remained
at 24 leaves and the mean was reduced only from 24.1
± .11 leaves to 23.9 ± .05 leaves, — hardly a significant
figure. The mode of the plus selection crept up to 26-27
and the mean to 26.9 ± .07 leaves, there being here one
more generation than in the case of the minus strain.

Table V gives the data on plus and minus selections of
Family No. 19 at Bloomfield for two generations. The
original family stock of the Fg generation has the mode at
27 leaves and the mean at about 26 leaves. A 24-leaved
plant of this generation became the parent of the minus
strain, giving in the Fe generation a population with the
same mode and a slightly higher mean (26.9 ± .08 leaves).
Continuation of the strain through a 24-leaved plant gave

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION

an F7 population with the mode one class lower and the
mean at 25.8 db .09 leaves. Whether this slight reduction
really means anything we are unable to say. At least, if
it yields at all to selection,
the progress is very slow.
On the other hand, a con-
siderable gain has been
made in the plus selec-
tions. The mode rose im-
mediately to 29 leaves
when the progeny of a 29-
leaved plant were grown,
and went up to 30 leaves
the next generation, the
modal condition being the
same as the number of
leaves of the parent plant.
The means are 26.3db.l0
leaves, 28.7 d= .10 leaves
and 29.2 it .08 leaves, the
amount of progress being
— as may be seen — 2.4
leaves and 0.5 leaf in the
two successive genera-
tions. This result appar-
ently indicates a slowing
down of the effect of selec-
tion.

The continuation of the
table gives the results ob-
tained at New Haven on
this same family. Here
there are data from three
generations, and these
data modify the conclu-
sions based on the results

obtained at Bloomfield. Both plus and minus strains
nearly parallel the Bloomfield results for two generations.

Fig. 1. Plant op Halladat Ha-
vana Tobacco (77-2) -1-1, which Av-
BBAOES 29.7 Leaves Psb Plant. It
IS THB Result of Three Ybabs of Se-
lection FOB High Leaf Numbeb in
Family 77, which Averaged 22.4
Leaves Per Plant in 1909. New
Haven, 1912.

Digitized by

Google

20 THE AMERICAN NATURALIST [VoL.XLVra

the F7 generation means being 28.3 ± .11 leaves and 25.1
± .15 leaves, respectively, but in the Fg generations they
differ. Selecting minus extremes for the first two genera-

no. 2. Plant or Halladat Havana Tobacco (77-1) -1-1, which Atbsaobs
20.9 LSAYBS PiB Plant. It is thb Rbsui^t of Thbbb Ybajis or Sblbction fob
Low Lbat Numbbb in Family 77. Nbw Havbn, 1912.

tions reduced the mean of that line from 26.3 ± .10 leaves
to 25.1 ± .15 leaves, but the third selected generation (Fg)
had a higher mean than the original family (27.3 ± .08
leaves) . The parent plant of this Fj generation produced

Digitized by

Google

No. 666] CHANGES PRODUCED BY SELECTION 21

24 leaves, and as the strain indicated that it was hetero-
zygous for a number of factors by showing a coeflScient of
variability of 8.29 d= .42 per cent-, it is possible that the
selected parent plant may have belonged gametically to a
higher class than was indicated somatically; nevertheless,
it can not be denied that three generations of selected
minus extremes have produced no results. This conclu-
sion is not valid for the plus strain. Starting with 26.3 ±.
.10 as the mean number of leaves (Fg) , the succeeding gen-
erations had means of 27.1 it .07 leaves, 28.3 dz .11 leaves
and 30.0 it .11 leaves. The differences are 0.8, 12 and 1.7
leaves, respectively. Progressive change has certainly f ol-

Fio. 8. Chanob of Phyllotaxy in Some Plants of (77-2) -1-1 Grown in New

Haven in 1912.

Digitized by

Google

THE AMEBIC AN NATURALIST [Vol. XLVin

lowed, and unless one considers that the results of 1912 are
somewhat too high (probably a valid assumption), the
change has increased instead of decreased. Naturally
there must be a decreased momentum in change of mean
time, but this decrease is not yet shown by the figures.

Pig. 4. Plant op Halladay Ha-
TANA Tobacco (19-2) -1-2, which Av-
BSAOBS 30 Lbaves Peb Plant. It
IS THB Rbsuivt of Thebb Ybabs of Sb-
lbction fob hloh leaf numbeb in
Family 19, which in 1909 Avebaqed
26.3 Leaves Peb Plant. New Haven,
1912.

Fig. 5. Plant op Halladat Ha-
vana Tobacco (19-1) -1-1, which Av-
BBAGES 27.3 Leaves Peb Plant. Thbbb
Ybabs of Selection fob Low Lbaf
Numbbb Havb Pbovbd Unsuccessful.
New Haven, 1912.

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 23

Bepresentative plants of the plus and minus strains of
family 19 as obtained by three years of selection at New
Haven are shown in Figs. 4 and 5.

Family No. 5 (Table VI) shows a decrease in mode
from 28 to 26 leaves, and a similar decrease in mean from
28.1 ±: .06 leaves to 26.6 ± .09 leaves as a result of the first
minus selection. A second minus selection, however, in-
dicates either that the future progress is to be very slow
or that the entire effect of selection was manifested in the
first selected generation.

With the three parts of Table VH we take up the re-
sults on Family No. 6 at all three stations. The minus
strain was carried on only two generations at Bloomfield,
but with this exception there are data upon three genera-
tions. At Bloomfield the two generations of selected
minus extremes resulted in 0.6 leaf decrease in the mean,
but at New Haven the results were negative, the means
advancing from 25.8 ±: .06 leaves to 27.9 zt .12 leaves in
three generations, while at Forest Hill the mean remained
practically the same. Surely selection was unprofitable
here.

The first year of selection from the other end of the
curve, however, resulted in marked progress. The mean
advanced nearly 5 leaves in each case. The original F5
mean is 25.8 ±: .06 leaves, but the three Fe means are 30.7
dz .09, 29.6 dt .08 and 30.8 zt .12 leaves. This is a remark-
able concurrence of results. The means in the two suc-
ceeding generations were about the same in the Bloomfield
and New Haven experiments, but there was another defi-
nite advance at Forest Hills. Such a result should not
be unexpected. If the Fe generation were almost but not
quite a homozygous lot, and if one assumes that selection
of extremes from homozygous population has no effect
in shifting the mean, it would frequently happen that
some individuals selected to continue the line would be
homozygous in all factors and some heterozygous in one
or more factors.

The cause of the peculiar distribution of the population
(high variability) of the Fg generation grown in Bloom-

Digitized by

Google

THE AMERICAN NATURALIST [VoL.XLYin

►

9
2

^ O O

CO "^ IC 00 K 00

»H C^ •-• 1-1 1-1 1-1

•H* I I 4l' -H* -H*
22SggS

lO b<^ lO iO CO CO

•H I -H -I -H I

•H -H -H -H -H -H
lo CO iH ci CO q

CO CO 00 O) 00 O)

c^ c^ c5 c^ c5 N

g 1-4 ^ <0 (

1-t iM ^ 1-1 lO

1-4 ,-1 M lO 00 00

'SS$S

I »0 CO CO
• lO ^ CO

!?SK?3i

•28

S SSScO CO

1-I O ^O i-< N

l-l f^ O 1^ •-• 1-4
Oi Oi Oi Oi Oi Oi

to kO to ub >o >o

: : : : :^

I ; ; I i c<

: : : : i*^

?5SSg8S

•H O 0> O 1-I CI

*H f-4 Q 1-^ 1-^ >H

QOf

pL, pL, pC(s, pE4 ;Z4

^ wt N cj A

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION

ssu

sssg

■1

► •

HI 1 HI 1 HI HI HI

t. ti <6 i6 >o <o ^

s&sssgs

HI 1 HI 1 HI 1 HI

Sgfe^SSS

e^ ,M 1-4 1-4 *-l fH iH

sssssss

HI HI HI 1 HI HI HI

a<4<Koo«h:e4

S;^SS^Sgi

PW^L

SSilsSS

^i4 •

o j

^i* fH

O « 1-4

•*•

CO CO o

1-4

M«

1-4 CV| «
^N 1-4

w-^'-'^g®S

t* 1-4 C^ fH d ,-4 1-4

^ lO WJ«

CM 1-4 1-4 1-4 Q «C lO

1-4 1-4 1-4 l5 CO c5

S*Sgg§§J:

52S3SS"

s§5r;-=-

sgssis--

;;(

SSSS"-

»o o o <o : :

1-4 1-4 1-4 . .

CO »0 11 ^ ;

. w4 • • • '

111

^s^n^^u

d f-i o 9 o 1-4 es

^ .H iH p l-» 1-4 f-l 1

o> A a» o o> o> o

• !.••«•.• 1

Pc Pc ^ Pn Ph {>, p^

7 7

J**^ ^J4

xx^ ^xx

f^ti><i>3.^<a

'>-

'^^

%^>^'>—

' 1

00 o

»c

00 •*

1-4 1-4 d iH W iH .H

►

•H -H -H -H -H 1 1

fcSSSSSfe

^ ^ »o <o »o ^ CO

ss^sssss

HI HI HI HI HI HI HI

^^^^^3^

S&g828S

-^'

HI HI HI HI 1 HI HI

O 00 00 00 00 *-4 00

^s^^s^s

mox

sss&S!;s

Is"

t*M

d lO ^

O b- 1-4

2SS

1-4

822

^jj^OOO

1-4

"^POW ^ W 1-I

ss^s"-

S«

ssas**

SS8!:

ssas

1-4 1-4 id

CO •-• w «0

1-4

'. '. '.'•^

III

ssssssss

s%^

d 1-* O 9 O 1-4 d

^ 1-4 1-4 <5 1-4 1-4 1-4

Oi 0> Oi w) Oi Oi o>

«• r- •«•••- w

^ ^ pC( pt4 ^ pC( ^

7 cji

77 ?T

JLJL^ ^JkJk

1-11-4 1-4 d d d

iiiS)iii

m
It

Digitized by

Google

THE AMERICAN NATURALIST [VoL-XLVHI

SSSI

4i -H -H -H -H

tA ,H fH ^ a»
« ^ ^ ^. '^I

lO CO CO t^ lO

ss&gg

1 d

•H -H -H 4 -H

1 00

t* COOCVI 00

•^ ^ ^ t* ^

feSSgg

•H 1 -H 1 -H

« ^ 0> »H O

SSSSISS;

8S88g

i-» C< i-» N i-»

1-4

«-

»c

•^

1-4

' a

1 «

■*§

§3«

52«&?l

SS5S52

o>t* K »o ^

-**< CO i-» lO

CO CI ^ o> i-<

^cocoio

CO t* •-• t* :

CO CV| ^ .

•H*

; i-» w 00 :

! N •-•

:^^»c :

• fH • • •

amber
Leaves
Parent

§ss;ss

^V.Vl

•^ o o

i-«OC5 O i-«

liH 1-4 O (-^ 1-H

OJOJ OJ OSO>

1-4 1-4 ^H !-• 1-4

»••«•••«

p^p^p^p^p^

N «

Js^ ^X

1-11-1 W CV|

jz;

^itsH

0» CI

CO o

^ I

»-Iei

W CI CI CI

•il -il

•H 4i -H 1

K CI

22SS

cot^

CO CO

CO 00 lo CO

SSS8

1 HI

1111

S2SS

1-4 fi

^ e^ »H ^'

ssss

•il -1

•H -H •« -H

1-« t*

lO 1-4 oo o

^sss

CO o CO CO lo 00 ^

OS •-< 1-4 »^ 00 CO
CI N i-« d i-« CI

-•

f^ •

CI CI

a »o

-^f »^

1-4 1-1 «o

C< 1-1

Tf O 1-1

1-4 1-1 T*«

CO
CO

»0 CI

CO -* CO '

»0 CI

w4 rft

o> a CO

oj CO a

►

»o -*

w4 r^ia

:§§

-^Sqjg

*sss ,

r^ a o O)

CI CO »o

t^ lO

S CO c3

1^ »-<

S§^

s
s^

b- CO C1 1-»

t^ lO

CO^

imber
Leaves
Parent

ssg^ss^ss

1 ^^^

Year
Grown

CI i-« O O O i-< CI

l-« 1-4 1-H O l-« "^ *H

05 ojo>2 2® 2

»> f • « «D !• »•

§:§

PC4 P^ pLi P^ pLi pt4 pC4

• ••

^ ^ fM fH

1-4 iH 1-4 d CI CI

c5|(A (A ci(A <A ci

,M iH i-H ^H 1^ 1-H «M

fy^y^

CICICICICICICI ^

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 27

field is not clear. It is possible that the plants having
from 18 to 23 leaves were diseased, but no such condition
could be recognized in the field. Again, it is possible
that a few Havana plants were mixed in by mistake,
although as the leaves of the selection are characteris-
tically diflferent from Havana and as the plants with low
leaf numbers resembled the remainder of the row, this
supposition is improbable. The most likely explanation
is that mutation occurred in a few gametes of the mother
plant, a condition that did arise, or that we assume to
have arisen, in Family 41 (see Table X). At any rate,
the change did not follow the path of selection.

In Figs. 6 and 7 are shown typical plants of Family No.
6 obtained by three years of selection in the effort to pro-
duce strains of high and low leaf number, respectively.

Family No. 34 (Table VllI) is peculiar — although this
is not the only time the phenomenon occurred — in that
the Fg population grown from a 24-leaved F4 plant seems
not to have given the true mean. Plants with a low num-
ber of leaves (22 and 20) were selfed to carry on the
minus strain, but both gave means higher than was shown
by the Fg generation. Perhaps further selection will
produce results, but the case is not a hopeful one. The
only evidence for such an assumption is the increased
mean of the F^ plus strain. If it is assumed that 24.0 is
nearer the true mean of the F5 population than the 22.9
actually calculated, then the jump to 27.0 =t .08 leaves in
the F7 generation gives us a basis for expecting results in
Fg in the minus strain.

Nothing can be said as yet about the minus strain of
Family No. 12 (Table IX), for it happened that the first
selection was a complete failure. Six plants were ob-
tained, but the lowest number of leaves was 29. One of
these plants was selfed and gave an F^ population having
a mean of 28.7 ±: .09 leaves. Unfortunately the selections
from this fraternity did not germinate and in 1912 we had
to fall back on the reserve seed from which the 1911 crop
came. The crops of 1911 and 1912 are therefore dupli-
cates. The plus strain made an advance from 24.5 ib .10

Digitized by

Google

28 THE AMERICAN NATURALIST [VouXLVHI

leaves to either 26.8 ±l .07 or 29.0 ± .08 leaves. The first
advance is 1.6, the second 0.7. We can give no explana-

FiG. 6. Plant op Halladay Havana Tobacco (6-2) -1-1, which AvERAass 80.2
Lbaves Pub Plant. It is the Result op Thbeb Years op Selection fob High

LlAF NUMBEB in FAMILT 6, WHICH AVERAGED 25.8 LEAVES PEB PLANT IN I909w

Nbw Haven, 1912.

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 29

tion of the failure of the results of 1911 and 1912 to dupli-
cate. This is the greatest deviation obtained in the course
of our experiments. The results of 1912 are probably
too high. It is yet too early to say whether or not this

FxQ. 7. PLANT ov Haklapat Hatana Tobacco (6-1) -1-1, which Atbbagu 27.9
Lbatbs Pbb Plakt. Thbbb Ybabs ov Sblbction to Dbcrbase the Leaf Numbbb

OP THIS TTFB HATB PBOTBD UNSDCCXSSFULb NBW HAYEN, 1912.

Digitized by

Google

30 THE AMERICAN NATURALIST [Vol. XLVIII

strain is decreasing in the average annual shift of the
mean.

Family No. 41 shown in Table X gave perhaps the most
peculiar results of any of the selections. It may be that
no great shifting of the mean toward the minus end of the
curve should have been expected, because the minus
mothers were each rather high in number of leaves. There
was one with 25 leaves and one with 24 leaves. This was
unfortunate, but was made necessary by the number of
late and diseased (mosaic) plants in the selection. Never-
theless, each of these plants was below the mean of the
previous generation and if a marked change would have
followed the selection of extreme individuals, some change
should have followed the selections of the individuals that
were the actual mothers. But in spite of this fact the
mean persistently rose from 23.9 ± .07 leaves to 26.3 it .08
leaves, then to 28.1 zb .07 leaves, although the duplicate of
this selection grown in 1912 went down slightly to 27.4
± .07 leaves. In the plus strain successive generations
of mothers having 28 and 30 leaves caused a small upward
shift of the mean ; it became first 25.7 ± .09 leaves then
25.6 it .14 leaves, although the 1912 duplicate of the last
population had a mean of 26.9 db .08 leaves.

The extraordinary phenomenon to which we wish to
call particular attention, however, is not this behavior of
the minus and plus strains in the regular selection ex-
periment, but rather the origin of a few-leaved strain
from a single individual that appeared in the Fe genera-
tion of the plus strain. Eeferring to the table, it will be
seen that in this generation a 12-leaved plant appeared.
This is really a peculiar phenomenon, for we had never
before observed a normal 12-leaved plant among the many
thousands that have come under our observation. They
do not occur. In this population the plant with the next
lowest numbers of leaves had 20 leaves, and in classes 20
and 21 there was only a single plant of each. This 12-
leaved plant was selfed and gave rise to a population
ranging from 8 leaves to 30 leaves, and having a vari-
ability of 23.50 per cent. ± .11 per cent. The mean of the

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 31

distribution was 19.8 ±: .28 leaves. A 10-leaved plant of
this lot was selfed and gave a progeny with a mean of
17.9 ±L .08 leaves and a variability of 11.24 per cent. ±: .33
per cent. What interpretation can be given these facts!

We believe a distinct mutation occurred, a mutation
different from those of DeVries. At least DeVries be-
lieves that the mutations that he has observed always
breed true. If the following hypothesis as to the origin
of the 12-leaved plant be true, it is unnecessary to sup-
pose with DeVries that mutations always breed true or
even that they often breed true. Of course DeVries be-
lieves that his (Enothera mutations obey laws different
from those of whose mechanism we know a little. He be-
lieves that species crosses always breed true; that they
do not Mendelize. This belief we hold to be unfounded.
Species crosses have never been shown to breed true.
There have been statements to the effect that crosses be-
tween Rubiis species breed true, but no good evidence has
been submitted in their support ; while the data of Tam-
mes ( :11) on Linum species crosses, Davis ( :21) on (Eno-
thera species crosses, and of East ( :13) on Nicotiana
species crosses, concur in showing that species as well as
varieties obey MendePs Law of segregation and recom-
bination. Furthermore, we think that Heribert-Nilsson's
(:12) beautiful experiments on DeVries *s own material
show that the latter did not collect suflSciently exact data
on his own crosses to find out whether they bred true or
not.

If one is to believe that a mutation in a hermaphroditic
plant breeds true he must suppose that constitutional
changes occur both in the male and the female gam-
etes, or that the change occurs after fertilization. But it
seems more probable that such a change will take place
either in the one or the other gamete and not in both. This
we believe to be the explanation of the appearance of the
12-leaved tobacco plant. A mutation occurred in either
an egg cell or a pollen cell. It does not matter in which
one it is assumed because there is no evidence favoring
either case to the exclusion of the other. This cell with

Digitized by

Google

THE AMERICAN NATURALIST [VoL.XLVin

9 g

22S8gJg:2S

cdid cdr^r^^ ^cot^

sssssssss

•4

ii^i^SSSg

; ;

I I I *^ I

§S

00 t*

:*^ :^ :

<oo

. CO 'lO •

SS"

^«-a :

§!SlS-2»-®S§ :

SSS«S2®S :

s«

S!;!SS5S''S-

SgSS5!5-S°

S-SSSS'S-

OOClOggCOOftgO

2gS"*2-

•^

M g« Ti<«2"«

« jo-ieooo -Hg

O^iOjjC* b-

N -COOO

.»H00

1-4

2!;

■*

;

*«4

;

III

sssssssss

aS2i23S2S

a>a>o>9a>0fta>a>a>

,.^^,HfHf1fHl-4f-4l-4

££££iS£^££

.:2

(

6.71*. 66
6.31*. 20
6.44*. 21
6.86 *.18
6.18*. 21
6.60*. 18

0(5

1.44 *.14
1.57 *.06
1.69 *.05
1.66*. 05
1.66 *.06
1.60 *.06

26.2 *.19
24.9 *.07

24.7 *.07
24.2*. 06
26.7*. 08

26.8 *.07

8822§8

; 1

M j

-«

00 CO

.C^CJJ;

^r^cDCgj^j

~8S258

«8S5Sfe

'^sisgja

«58Sg22

«S89^^

-^•ss

e«coco

i*^^

0»

• • 1-4

SSiSiSS^

1912
1911
1910
1909
1910
1911

(i^p^£p^£pE4

Digitized by

Google

No. 565] CHANGES PBODUCED BY SELECTION

HI HI hi HI HI 1 HI

O CI CO $ S CO ^

lo CO <d CO to r«^ r^

HI 1 HI 1 HI HI 1

^SSSSSSS;

f.^ ,M ^4 1-4 ^ iH iH

^•

HI HI HI HI HI HI HI
-^ O CI CI ^ -^ to

•*•

\
t

I
1

O »H

^ <I4(

•oma

S^'S

'"•'SJSS i

"*^s«gsg

<°"a2SS8S

S25?5aS?5

S85:?2Sa3

5§5i!g"g}'-

J5»g5„u,«

««.Oj-

S5»

^SSSSSSS

^'1

CI fH O ® O 1-I CI

iH iH 1.H O «^ iH ^4

A^^^^^^

J,
1

1-4
1

6.14*. 22
4.50 *.17
5.01*^
6.86*. 18
4.26*. 20
4.91 *.19
4.42*. 17

Q
oS

1.59*. 06
1.17 *.05
1.27 *.06
1.66 *.05
1.16 *.05
1.42*. 06
1.18*. 05

25.9*. 08
26.0*. 06
25.3*. 09
24.2*. 06
27.2*. 08
28.9*. 08
26.7*. 06

SSSS85S

»H !-• ^^ CO ^^ »H !-•

1
1

«-( •

<o :

2 :

"S-

to CO

Nt-ge

22''*^6gg

S;S22S2S

^gggs"*?;

■^ 1-I <p CI t* ^ t*
•* CO CI t* ^

fiSSS-^

lO.Ht-»

d -iCI ;-J
CO

•^

ill

sssisss^;;

d fHOCjOfHCI
1-4 iH lH Q l-< ^H ^^

2222222

a O

i^^^^^A

,3
,1

3
1

Digitized by

Google

34 THE AMERICAN NATURALIST [Vol. XLVTEI

a changed gametic constitution, — a loss of gametic fac-
tors,— was fertilized by an unchanged cell. The un-
changed cell may have had any of the gametic possibil-
ities open to the germ cells of the 28-leaved plant of the
Fg family in which the mutation arose, and we know that
certain factors in this plant were heterozygous, for pro-
gressive change followed the selection of a plus extreme
in the next generation. The 12-leaved plant was there-
fore a hybrid. It resulted from the union of a mutating
germ cell of the mother plant that furnished the F^ gen-
eration with an unchanged germ cell. We can even as-
sume that the mutating germ cell, if fertilized by another
of the same kind, would have produced a plant with less
than 12 leaves. The reasons for believing this are simple.
There is experimental evidence (Hayes, 1912) that the
Fj generation of a cross between varieties differing in
their number of leaves is intermediate in character. Our
12-leaved plant is the lone representative of such an Fj
generation. The Fg generation therefore should give
plants with less than 12 leaves, and in fact such plants
did occur. The distribution marked Fa in the table is
the F2 generation, and this accounts for its extreme vari-
ability. The distribution marked Fb is the Fg generation,
and its variability is less than half that of the preceding
generation.

Family No. 56 was the second family to be grown at all
three of the experimental statioiis (Table XI). It arose
from a 26-leaved plant of the F5 generation which pro-
duced an Fe progeny with a mean of 24.2 ± .06 leaves and
a mode at 24 leaves. The three generations of the minus
strain grown at Bloomfield remained practically the same.
The last generation did indeed show a mean 1.0 leaf
higher than the original population, but no dependence
can be placed in data from only 25 plants. The data on
the minus selections grown at New Haven are for this
reason a little more dependable. They show a fluctuat-
ing mean, but no progress due to selection, the F9 genera-
tion having a little higher mean than the F^ generations.
The three minus selections grown at Forest Hills also

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 35

resulted in higher means, those for F^, Fg and F^ being
25.3 It: .09, 26.0 ± .06 and 25.9 it .08 leaves, respectively.

This peculiar result implies only that the mean of the
original F^ population which was grown at Bloomfield
was lower than it would have been if grown on the Forest
Hills ^ soil. This is not a direct effect of environment on
the growing plant. It has been shown conclusively in
our pot experiments, as stated before, that starvation or
optimum feeding has scarcely any effect on the number of
leaves, although it has a marked effect on the develop-
ment of many other characters. On the other hand, en-
vironment does appear to have a marked effect on the
number of leaves that a plant is to develop, if it acts
during the development of the seed. It is well known by
plant physiologists that the environment produces many
of its effects very early in the life history of the indi-
vidual or in the development of the organ concerned. For
example, the so-called light leaves of the beech with two
layers of palisade cells are differentiated from the shade
leaves with only one row of palisade cells by the amount
of light that falls on a branch during the season preceding
the development of the leaves: that is, it is determined
during the laying down of the bud from which the next
season's growth of twig and leaves comes. This period
during which a particular change is possible is called the
critical i)eriod for that change by plant physiologists.
Thus a plant may have hundreds of critical periods in its
ontogeny, each marking an end-point of development be-
yond which a certain feature is irrevocably fixed. For
example, the critical period for that cell division that de-
termines leaf size in the beech is much later than that
which determines the number of layers of palisade cells.

Now the critical period for influencing the number of
leaves of the tobacco plant is practically at an end when
the embryo plant goes into the resting stage of the seed.
Before that time the number of leaves may be influenced
by the external and the internal influences that form the
total environment of the mother plant; after that time
environment has little influence on the number of leaves.

Digitized by

Google

36 THE AMERICAN NATURALIST [VoL-XLVHI

The rise in the mean of the population of the Fg genera-
tion of Family No. 56 is due partially to the effect of en-
vironment, therefore, in that the mother plant was grown
under better conditions, but is probably not to any great
extent due to the conditions under which the plants them-
selves were produced.

The better environment of the mother plants does not
account for all the rise in the means in populations Fg
and Fg, but it accounts for part of it. It will be noticed
that all of the populations grown at Forest Hills had
higher means than those grown at Bloomfield and New
Haven, although the Fe mother plants were grown at
Bloomfield and not at Forest Hills. The greatest shift
of the mean, however, comes in the Fg and Fg generations,
for the mother plants of both of these populations were
grown on the more fertile soil. There is a simple ex-
planation of these facts, an explanation that is of great
economic importance to practical tobacco growers. A
part of the rise in mean at Forest Hills was due to set-
ting the plants in the field there when they were in an
earlier stage of development than those at Bloomfield and
New Haven. They were not set earlier in the season (at
least, one year they were set early, one year they were set
at the average time and the third year they were set late),
but they were set as small plants. When small plants
(about 4 inches high) are set in the open the root system
is equal to the task of supporting the aerial parts and the
plants start right in to growing normally. There is no
period of passivity. The plants produce leaves spaced
with normal internodes and these leaves develop suffi-
ciently to have a commercial value. But when the plants
reach a height of 8 or 10 inches in the seed pans or seed
beds and are then set in the field, the normal metabolism
is likely to be upset for a time. The plant takes some
time to recover its equilibrium and start a normal growth.
During this period basal leaves begin to develop, but the
internodes are so close together that they do not obtain
their aliquot share of nutriment, hence they grow only to
one quarter or one third their normal size and soon wither

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION

and drop off. The leaf scars are left, but they are so
close together that it is difficult to make a correct count of
the number of leaves. But more important than this,

Fio. 8. Plant of Hallaoat Ha-
TANA Tobacco (56-2) -1-1, which Av-
■BAGBS 27.6 Leaves Peb Plant. It
IS THE Result of These Yeabs of Se-
lection FOB High Leaf Numbbb in
Family 56, which in 1900 Atebagbd
24.2 Leaves Peb Plant. New Haven,
1012.

Fig. 9. Plant of Halladat Ha-
vana Tobacco (56-1) -1-1, which Av-
EBAOES 24.4 Leaves Peb Plant. Thbee
Ybabs of Selection fob Low Leaf
NuMBEB Have Pboved Unsuccessful,
New Haven, 1912.

Digitized by

Google

THE AMERICAN NATURALIST [VoL-XLVHl

H
M

7.77*. 26
6.52*. 22
6.93*. 27
4.62*. 40
7.98*. 33
6.03*. 20
7.15*. 23

1.67 *.06
1.53*. 05
1.69*. 07
1.03*. 09
1.82*. 08
1.45*. 05
1.63*. 05

21.5*. 08
23.5*. 07
24.4*. 09
22.3 *.13
22.8*. 11
24.0*. 07
22.8*. 07

||$;S§3^

• • *H

•*»H C^

^^o

rH rH CO

W FH rH

oc^ o>

eiooo

^2§5

»H coco

'^SS

*ss

ssss^gs?

SSS^^SSS

5S2SSSS8

gs^^s^g

gco

c« CO o to

o»

W O »H

■ fH »H

• fH •

III

^%'Z

ei »H o® o «-« c^

fH *H fH Q vH fH fH

p^^p^pCp^p^p?

1 1

???

^ ^ m

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 39

the tobacco grower loses an average of from one to two
of his most valuable leaves.

The pins strain of Family No. 56, which we were dis-
cussing when we digressed to speak of the critical periods
of development, did show a considerable shifting of the
mean following the selection of high-leaved mother plants.
In the Bloomfield selections the mean went from 24.2
di .06 to 26.7 ±z .08 leaves, then to 26.8 ± .07 leaves ; in the
New Haven experiment the mean shifted to 27.4 ±.08
leaves, — a gain of 3.2 leaves, — and then dropped to 26.4
±.11 leaves, recovering again in the F^ generation to
27.5 ± .11 leaves ; in the Forest Hills experiment the suc-
cessive means were 27.2 ± .08, 28.9 ± .08 and 26.7 ± .06
leaves. Summing up the data from this experiment, it
may be assumed to be reasonably certain that no progress
resulted from the selection of minus extremes, but that
there was a slight effect gradually diminishing in quan-
tity when plus extremes were selected.

Bepresentative plants of Family 56 obtained by three
years of selection in the effort to produce strains of high
and low leaf number, respectively, are shown in Figs.
8 and 9.

Family No. K (Table XH) was grown on a farm near
the Bloomfield experiments, in 1910. The records of the
Fg generation consisted of the number of leaves of only
31 plants. From among these individuals two plants
were self ed to become the mothers of the Fe generation.
Since no dependence can be placed on the Fj distribution
by reason of the few plants and since it is not absolutely
certain that the mother plants of Fe had 20 leaves each,
the selection really began in 1911 with theF^ generation.
There is a difference between the minus strain and the
plus strain in 1911 and 1912, — 0.5 leaves the first year and
1.3 leaves the second year, — ^however, so that one may
assume the possibility of a slow shifting of the mean in
both directions.

The data on Family No. 73 are shown in Table XHI.
This family came from a 28-leaved plant, one of the
highest of the Fg generation. The Fe progeny of this

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVIH

>•

00 « e« -^ f-i 0& 00

1-1 »H C« »H W »H ,H

■H -11 41 41 -H -H -H

ssss^jss

CO lO t. lO o to o

sssssss

4 HI 4 HI HI HI I

1H »H »H 1H »H iH •-<

SSSSfeSS

* -H 41 41 41 HI 41

o> w »ooo« oo

SS8S8SS

Total.

22S8S5S5g

CO « C< CO C« W CO

CO iH "* .-1

s
^

C^ C^ (N 0>

2''ag"2"=

t^ (N 0> CO to t* "*
CO CS C^ -^ i-^ Cs| »H

£ S O O W CO 01

gf:S£:5SK

CO (N T-4 CO »C t* CO
CO CO CO CO O "^ t^

52g35gS§S

2'2«§5"?§

00 ^iO Oit^^ ^

o • »H i CO »-• lo

»H

gJSSSSSS

i^t

(N ^O CftO iH(N

.-1 ^^ »H Q »H »H »H

2 22 S2S

& o

pi;p?p?p?p;p?£

•TV 71"

S&SS&

QQ
O A

O ^

c5S

41 41 41 41

SS5SS

OiOl^iO

ssgs

HI HI HI HI

c/5

S^S^

^ ,H *H 1-H

oSSS

-i

41 41 41 41

t* CO CO CO

sss§

c^ cq c^ c«

^ : : :

CO 1-1 CO 1-1

CO tT CO .-1

•*»

O (N C^l CO

\;i

(M ^ C^ ^

CO b» o o

•^ CO CO CO

O <N CO .H

CO CO CO OS

SSS55S

.-1 »0 1-1 00

(N CO CO CO

CO iH »H lO

,-( ,H C^ .-H

CO C^ ''t' t^

»o

»H

gS^

III

CO CO C^ CO

^*^^

o o

C« rH O S

1-4 *H *H Q

OiOiOscS

fH *H fH *H

IS§

« • f «

• IS

fi<^P^pH

fH fH

C) C4 M

c^ cii cii e^

00 00 00 00

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 41

individual showed a mean of 26.9 ± .06 leaves, and from
among them plants having 25 and 29 leaves, respectively,
were selected to start the minus and the plus lines. These
two mother plants gave F7 populations alike as to mean,
but differing by one class as to mode. The minus line
had the higher mode. The extremes of this generation
used in carrying on the experiment differed by 8 leaves,
and the resulting progenies apparently followed the selec-
tion. The means are 25.6 it .07 and 28.2 it .09 leaves.
Whether these shifted means represent a permanent
change or not we are not prepared to say. The minus
mean is probably somewhere near the correct figure for
in the Fg generation it was practically the same, but in
the Fg generation of the plus strain the mean dropped
from 28.2 ±.09 leaves to 26.7 it .13 leaves. This is a
slightly lower point than that of the original Fg distribu-
tion, but it was calculated from only 76 individuals. A
conservative estimate of the significance of the results
would probably be as follows: the mean of the minus
strain has shifted slightly but permanently and is now
fixed, while the mean of the plus strain has not changed
but has shown evidence of some heterozygosis in one gen-
eration.

We come finally to consider Families No. 27 and No. 82,
the data on which are listed in Tables XIV and XV. Two
generations of both plus and minus . selection were re-
corded for Family No. 27, but only plus selections of
Family No. 82 were grown. There is no necessity for
considering either in detail because a simple inspection of
the tables shows that selection has accomplished nothing.

Conclusions

The cumbersome and no doubt dry details of the ex-
periments to the close of the year 1912 having been de-
scribed, let us give a brief resume of the conclusions that
we believe may reasonably be drawn from the data that
have been offered. There can be no doubt that the orig-
inal *'Halladay" type of tobacco, isolated and propa-

Digitized by

Google

42 THE AMERICAN NATURALIST [Voi^XLVin

gated by Mr. Shamel and Mr. Halladay from the cross
between ^^ Havana" and *' Sumatra" tobaccos, arose
through the segregation and recombination of the Men-
delian factorial differences of the two plants, and not as
a mutation. It is simply a union of the factors that stand
for leaf size and height of plant in the '^Havana" variety
with the factors that bring about leaf shape and high
number of leaves in the ^^ Sumatra" variety. It hap-
pened that the somatic characters of these varieties ac-
count for all the characters of the hybrid. At the same
time one must remember that strains were obtained by
selection that averaged higher in number of leaves than did
even the * * Sumatra * * parent. We can only conclude from
this fact that the difference between the ''Havana" and
the ''Sumatra" varieties in leaf number is greater fac-
torially than somatically. Besides certain factors com-
mon to the two varieties, the factors for leaf number in
"Havana" tobacco might be represented by the letters
AA, and those of "Sumatra" tobacco by the letters BB,
CC, DD, EE. By recombination, this would give plants
with a smaller number of leaves than the "Havana"
variety and plants with a greater number of leaves than
the "Sumatra" variety. Both combinations were ob-
tained; and further, the theory has been shown to be cor-
rect by the results of other crosses where both types ap-
peared (Hayes, *12). It is probably unwise to suggest too
concrete a factorial analysis of the cross, yet the factorial
difference assumed above will account for all of the facts
obtained, by simple recombination. We assume a factor
in the heterozygous condition to account for the produc-
tion of one leaf and a factor in the homozygous condition
to account for the production of two leaves. The mean
of the " Havana" variety is about 20 leaves and the mean
of the "Sumatra" variety about 26 leaves. Somatically
there is a difference of 6 leaves or three factorial pairs
for which to account. But in order to have the theory
coincide with the facts there must be at least one (pos-
sibly two or three) factorial difference that does not show
in the two varieties. The meaning of this statement can

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 43

be shown best by an illustration. The 20 leaves of the
" Havana*' variety and the first 20 leaves of the ** Suma-
tra'* variety are represented by 10 pairs of factors, of
which nine are the same and one different in the two
strains. The ** Havana'* variety is nine leaf factors plus
A A, the first 20 leaves of the '^Sumatra" variety are nine
leaf factors (the same as those in the ** Havana") plus
BB. The additional leaf factors of the ** Sumatra" are
CC, DD. EE. With these assumptions, the recombina^
tions of a tetra-hybrid will represent our facts fairly
accurately. But, as was stated above, it does not seem
wise to take this interpretation of the facts too literally.
That some such factorial combination will represent our
facts superficially there can be no doubt, but in reality if
one could grow hundreds of thousands of individuals and
follow the behavior of each he would likely find himself
constrained to represent his breeding facts by a much
more complex system. There would probably be gametic
couplings and factorial differences whose main effect
would be on some entirely different character or complex
of characters, but which would have some slight jurisdic-
tion over leaf determination. To become diagrammatical,
the unit characters of a house are its cornices, its win-
dows, its floors and what not, but a collection of these
components is not a house. We may even exchange
dormer windows with our neighbor, but we can exchange
them only if they fit. Again, we may put on a coat of
paint, a color unit, but this color unit affects the appear-
ance of many other parts that are just as truly units.

The essential part of our conception of the origin of
this hybrid type is that recombinations of characters
quantitative in their nature can be expected and predicted
in crosses in exactly the same manner as is done with
qualitative characters. On the other hand, it must be
borne in mind that here was a hybrid type that appeared
to be breeding true to the general characters that we have
described, in the F4 generation. That it was not breed-
ing true is clear from the results of the selection experi-
ments, yet out of the small number of Fg and Fe families

Digitized by

Google

44 THE AMERICAN NATURALIST [VoL-XLVHI

taken tinder observation at least two were found to be
breeding true for all practical purposes in the Fg and F^
generations. We were able to reproduce the ^* Havana '^
type by continued selection in Family 77 and were able
to produce strains breeding approximately true to 30
leaves or so by the selection of mother plants in several
families. But can we say that any of our families are
now fixed so that no progress can be made by selection?
We can not. But we can say that some of them are so
constant that it would be a loss of time for selection to be
continued for economic results. It is important to know
whether plant or animal populations can reach such a
state of constancy by inbreeding that no profitable results
can afterwards be obtained by the practical breeder. We
believe it demonstrated by even these few data that such
a state, a homozygous condition, occurs in a definite pro-
portion of Fa offspring, and can be propagated commer-
cially at once if a suflBcient number of families are grown
to be relatively certain of including the desired com-
bination.

As to the problem of theoretical importance, the ques-
tion of the true constancy of homozygotes generation
after generation, we believe it to be fair to conclude that
a state so constant is reached, that even for the theoret-
ical purposes of experimental genetics it may be assumed
as actually constant. Further experiment and larger
numbers may show that selection can always cause a shift
in the mean, but will necessarily be a shift so slight that
it can be detected only by a long-continued experiment
and enormous numbers. Assuming for the purpose of
argument that this is the case, the matter would affect
only the question of the trend of evolution. It may come
to be believed, from evidence now unknown, that evolu-
tion may progress slowly in this manner, but if it does,
its course can hardly be demonstrated experimentally be-
yond a reasonable doubt. The problems of experimental
genetics can be attacked, however, from the standpoint
that experimental evidence of the shifting of the mean of
a homozygous population by selection is negligible.

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 45

Mutations may occur. We have shown the origin of
one family by a very wide mutation. In this particular
case it was not diflScult to show that a constitutional
change took place in a single germ cell of the mother
plant. It was only by a lucky chance that this fact could
be demonstrated, for with smaller changes such proof
would be impossible ; but there is no reason to believe that
this phenomenon is imique or even rare. It is much more
reasonable to assume that mutations usually arise in
single gametes than that the same change occurs simul-
taneously in many germ cells. One should expect the
somatic result of a mutation in an hermaphroditic plant
— ^the sporting plant itself — ^not to breed true, therefore,
but to behave as an Fj hybrid between a mutating and an
unchanged germ cell. It is true that the mutations ob-
served by DeVries in (Enothera Lamarckiana are sup-
posed to have bred true, but this is sometimes question-
able even from DeVries 's own data. The Lamarckiana
'* mutants** that did breed true are much more reason-
ably explained as segregates from complex hybrids.
They can be interpreted by Mendelism with no essential
outstanding facts, but if they are to be interpreted as
mutations, several discrepancies between what actually
occurred and what should be expected on DeVries 's own
theory must be explained. It must be shown why the
changes took place in numerous germ cells, — in both the
male and the female gametes, — and why these germ cells
always fused at fertilization ; for the changed germ cells
must have fused with each other because majiy Lamarck-
iana plants were produced by the same mother plants that
produced the mutations, while the mutations are sup-
posed to have bred true. On the only other possible theory
of mutation, that the change occurred in the developing
zygote after fertilization, one would have to explain why
the mutants did not often appear as bud variations, in-
stead of these being much rarer than the supposed muta-
tions, as is actually the case.

We do not deny the theory of mutation as modified to

Digitized by

Google

46 THE AMERICAN NATURALIST [Vol. XL VIII

assume only that constitutional changes usually occur in
the germ cells, but on this belief the sporting plants must
often be F^ hybrids, and the plant breeder must resort to
selection to isolate his pure mutation. And by the same
reasoning one gametic change may produce many new
creations, for there is a chance to recombine it with all the
known gametic differences in the species.

No one can say how often mutations arise. It is likely
that changes other than the one observed took place in
our tobacco experiments, but it is not likely that they
are suflBciently numerous to base a system of selection
within a pure race on the possibility of their occurrence.
The fact that no changes ensued that could be detected in
several of our selected lines is an argument against it.
The comparatively large jumps are the ones likely to
have the greatest economic importance, and these are
easily detected without refined methods of procedure.
Small jumps can be economically important only if they
are numerous, and, as there are absolutely no data to
show either that they are numerous or that changes can
be produced rapidly within homozygous pure lines through
any other cause, it seems unwise to recommend that the
practical breeder expend time and money to bring about
results that either can not be expected at all or that are
so slow and so trifling that they can not be detected in
carefully planned and accurately executed genetic inves-
tigations. On the other hand, the results of the last de-
cade show that important economic results can be ob-
tained easily and surely by selection from artificial hy-
brids or from the natural hybrids that occur in cross-
fertilized species by the recombination of Mendelian
factors. We believe, therefore, that the isolation of ho-
mozygous strains from mixtures that are either mechan-
ical or physiological, that are either made artificially or
are found in nature, offers the only method of procedure
that the practical plant breeder will find financially
profitable.

Finally, we should like to call attention again to the

Digitized by

Google

No. 565] CHANGES PRODUCED BY SELECTION 47

practical importance of determining the duration of the
period in the course of which particular plant characters
are responsive to the action of environmental influences.
The character complex that has been the basis of this
study is a striking illustration of how results from such
investigations may be applicable to farm practise. One
may plant a portion of the seed from a self-pollinated
tobacco plant on poor soil or on good soil and the average
number of leaves per plant and the general variation of
the plants in number of leaves will remain nearly the
same in both cases.* But seed selected from mother
plants grown on the good soil will produce plants aver-
aging slightly higher in leaf number than the plants com-
ing from seed on mother plants whose environment is
poor. Consequently, it is better to select seed from well-
developed mother plants — smother plants whose environ-
ment has been good — ^than from mediocre mother plants.
There is no question here of the inheritance of an acquired
character or of continuing to raise the number of leaves
by cultural treatment. One simply takes advantage of
the fact that during seed formation there is a period of
mobility at which time the potential number of leaves of
the young plant are practically fixed. Pending the end
of this critical period, the number of leaves can be in-
fluenced by external conditions within the limit of fluctu-
ating variability.

In the same connection, the effect of time of planting
on the tobacco plant should again be mentioned, as this
also emanates from environmental change. The actual
number of leaves is, of course, practically fixed at the
time of setting the plants in the field, but this is not true
of the number of leaves that will have a commercial
value. For example, a seedling with 26 potential leaves
is planted. If it is planted when about four inches high,
the general physiological disturbance due to transplanta-
tion is negligible and the plant continues its normal cycle
of development without a pause, bringing to maturity

3 Gamer's (:12) results on Maryland Mammoth are an exception to this
statement because this variety is indeterminate in growth.

Digitized by

Google

48 THE AMERICAN NATURALIST [Vol-XLVTEI

about 22 leaves. If planting is delayed until the seedling
is eight or ten inches high, there is a different state of
affairs. Development is arrested, the plant pauses to ad-
just itself to the change. It soon recovers and continues
its normal ontogeny, but the period of reduced growth
has left an ineffaceable record. Several of the leaves —
among them the more valuable leaves — ^have been so
affected during this readjustment, that they develop to
only a fraction the size that they should attain because
the intemodes between them are so short, due to the con-
stricted development that normal metabolism does not
occur. Thus there is a loss of one or two leaves, which
on several acres of tobacco may make the difference be-
tween profit and loss. Hence, the grower should not de-
lay setting his plants in the field until they have become
overgrown in the seed bed.

March, 1913

LITERATURE CITED

Davis, B. M. G«netical Studies in CEnothera, III. Amsr. Nat., 46: 377-

427. 1912.
East, E. M. Inheritance of Flower Size in Crosses between Niootiana

Species. Bot, Gae,, 55: 177-188. 1913.
East, E. M., and Hayes, H. K. Inheritance in Maize. Conn. Agr. Exp. Sta.

BuU. 167: 1-142. 1911.
Garner, W. W. Some Observations on Tobacco Breeding. Ann. Rpt. Amer.

Breed. Assoc., 8: 458-468. 1912.
Hayes, H. K. Correlation and Inheritance in Nicotiana Tdbacum. Conn.

Agr. Exp. Sta. BuU. 171: 1-45. 1912.
Heribert-Nilsson, N. Die Variabilitat der (Enothera LamarcHana und das

Problem der Mutation. Ztschr. Abstam, u. Vereh,, 8: 89-231. 1912.
Jennings, H. S. Heredity, Variation and Evolution in Protozoa, I. Jour.

Exp. Zool., 5: 577-632. 1908.
. Heredity, Variation and Evolution in Protozoa, II. Proo. Amer,

Phil. Soc, 47: 393-546. 1908.

Assertive Mating, Variability and Inheritance of Size, 'in the Con-

jugation of Paramecvum. Jour, Exp. Zool., 11: 1-133. 1911.
Johannsen, W. Uber Erblichkeit in Populationen und in reinen Linien.

Jena, Gustav Fischer, pp. 1-615. 1903.
Pearl, Raymond. Inheritance of Fecundity in the Domestic Fowl. Ameb.

Nat., 45: 321-345. 1911.
Shamel, A. D. New Tobacco Varieties. Yearbook U. S. Dept. Agr., 1906:

387-404. 1907.
Tammes, Tine. Das Verhalten fluktuierend variierender Merkmale bei der

Bastardierung. Bee. Trav. Bot. N^erl., 8: 201-288. 1911.

Digitized by

Google

GYNANDEOMOEPHOUS ANTS DESCEIBED DUE-
ING THE DECADE 1903-1913

Pbofbssob WILLIAM MORTON WHEELER
BussiT Institutiok, Habvabd Ukivxbsitt

In 1903 I described six gynandromorphous ants and
reviewed the previously recorded cases, seventeen in
number. Although many thousand ants have since passed
through my hands, I have failed to find any additional
cases. Other observers, however, have been more for-
tunate and have described seven within the past decade.
As these are all very interesting, it seems advisable to
give a brief account of them as a sequel to my former
paper.

1. Lateral. Gynandbomobph of Cabdiocondyla batesi

FOBBL. VAB. NIGBA FoBEL. SaNTSCHI (1903, p. 324,

Fig. 5, i)

This specimen is female on the right and partly male
on the left side. The male portions are sharply marked
off from the black female portions by their testaceous red
color. The line of demarcation, very clear in front, starts
at the anterior clypeal border and divides the head into two
nearly equal parts, but leaves the median ocellus on the
male side. It then divides the pronotum down the middle
and the three anterior quarters of themesonotum. Thence
the line fades out on the right side so that the whole pos-
terior border of the mesonotum is male. Three quarters
of the prescutellum and the anterior half of the scutel-
lum are male. The epinotum and the abdomen are female
throughout, but the female genitalia are slightly asym-
metrical on the left side. The fore and middle legs on
this side and a portion of the mesosternum are male.
There are wings on both sides, but the anterior one on the
female side was lost after capture. Those on the left

Digitized by

Google

50 THE AMERICAN NATURALIST [VoL.XLVin

side are well-developed, with distinct venation and pale
pterostigma, and are inserted in a distinctly male area.
The specimen was not dissected.

Santschi found this ant in a nest with females at Kai-
rouan, Tunis, but without males, either of the winged or
of the ergatomorphic type, which is peculiar to this and
some of the other species of Cardiocondyla. His atten-
tion was attracted by the bizarre movements of the speci-
men, as it turned around rather quickly in circles about
10 cm. in diameter, with the male portion inside. In
other words, owing either to the asymmetry of its brain
and visual organs or to differences in the length of the
legs on the two sides of the body, it made circus move-
ments like a normal insect which has had one of its eyes
or optic ganglia injured.

2. Lateral Gynandromorph of Anergates atratulus
ScHENCK. — ^Adlbrz (1908, p. 3, Fig. 1, a, h, c, d and /)

An imperfect lateral gynandromorph, with the head
largely male on the left, female on the right side, the light
color of the male being sharply marked off from the dark
color of the female only anteriorly. Thorax in front
female, with wings equally developed on both sides (the
male Anergates is wingless and pupoid!), but with pale
(male) coloration on the left and dark (female) colora-
tion on the right side, the line of division between the
two neither sharp nor straight and the whole postscutel-
lum blackish brown. Abdomen with irregular arrange-
ment of color. Petiole black on the right, grayish yellow
on the left ; postpetiole mostly blackish brown, but with
a large grayish yellow spot on the left side of its anterior
surface. Third dorsal tergite blackish brown on the right,
grayish yellow on the left side. Remainder of gaster
grayish yellow, tinged here and there with pale brown.
Third tergite with a median longitudinal groove which
runs back on to the succeeding segment as in the virgin
female. The left side of the abdomen has seven com-
plete segments and well-developed genitalia; the right

Digitized by

Google

No. 565] QTNANDBOMOEPHOUS ANTS 51

side has only six complete segments and a membranous,
incomplete seventh^ The genitalia on the right side are
imperfect, the volsella being represented only by a piece
corresponding to its dorsal portion and the stipes is com-
pletely lacking. The legs are of the female type, except
the left fore leg, which is male, although the tibial spur
(strigil) is pectinate as in the female. This spur is known
to be nonpectinate in male Swedish, but pectinate in male
Swiss Anergates specimens.

On dissecting this specimen, which he took from a large
Anergaies-Tetramorium colony near Arkosund in Oster-
gotland, Sweden, Adlerz found on the left side a well-
developed vesicula seminalis, receiving a vas deferens
half as long. No traces of female reproductive organs
nor of the poison gland and vesicle could be detected.

Of particular interest was the behavior of this gynan-
dromorph, because, as Adlerz says, it evidently felt itself
to be a male but was treated by the normal males in the
colony as a female. Its movements were somewhat live-
lier than those of normal males, and it at first made feeble
attempts to copulate with the females and was treated
with indifference by the males. A few days later it be-
came more energetic and persistently attempted to copu-
late, especially with one particular female, although
always unsuccessfully while it was under observation.
It was evidently inflamed with the insatiable sexual appe-
tite so characteristic of the normal Anergates males,
which, being wingless, always mate with their sisters be-
fore they fly out of the parental nest. On the following
day, however, a normal male made the most persistent
efforts for several hours to mate with, this same gynan-
dromorphous individual. Adlerz concludes that

this indicates that the males regarded it as a female. Of course, we
may suppose that its wings made it seem like a female and attracted the
male, but from the fact that males attempt to mate even with female
pupae and therefore with a stage which has not yet developed wings, it
is more probable that the male was attracted to the gynandromorph by
some female odor. At any rate the double nature of the gynandromorph

Digitized by

Google

62 THE AMEBIC AN NATURALIST [VoL-XLVHI

is even more strongly indicated by the facts just recorded than by its
morphol6gical peculiarities.

3. Latbbal Gynandbomorph of Anbrgates atbatulus
ScHENCK. — Adlebz (1908, p. 5, Fig. 2, a, 6, c, d and e)

An imperfect lateral gynandromorph, male on the left,
female on the right side, resembling the preceding speci-
men, but with the dark female color more pronotmced on
the male side of the head There were well-developed
wings on both sides of the thorax, which was of the female
form though dark on the right and pale on the left side,
except the epinotum, which was grayish yellow through-
out. Abdomen in color and form almost typically male,
with the genitalia well-developed on both sides, but with
a feeble mid-dorsal impression recalling the condition in
the virgin female. Legs of the female type, except the
left fore one, which is somewhat shorter and thicker as in
the male and with the tibial spur (strigil) cleft but not
pectinated.

Dissection showed the reproductive organs to be in the
same condition as in the preceding specimen; i. e,, they
were present only on the left side and consisted of a
rather large vesicula seminalis with its vas deferens. No
traces of female reproductive organs, nor of a sting or
poison apparatus were to be found.

This specimen was taken from the same nest as the
preceding.

4. Lateral Gynandromorph (Ergatandromorph) of
Formica sanguinea Latreille. — Donisthorpe (1909,

p. 464, Fig. 1)

A nearly complete lateral ergatandromorph, with the
right antenna, mandible and eye, and right and median
ocellus male and the left antenna, mandible, eye and ocel-
lus of the worker type. Head black, except the left
mandible, left half of clypeus, left cheek and a small patch
in front of the eye, which are red. Thorax and petiole

Digitized by

Google

No.6e6] GTNANDB0M0BPH0U8 ANTS 58

male on the right, worker on the left, the line of division
running to the left of the median line so that the black of
the right side of the mesonotnm encroaches on the red
color of the left side. Petiole and gaster sharply divided
into black right and red left halves, the right half of the
latter also with male pilosity and sculpture. External
male genitalia and anal stemite on the right side. The
red and black coloration is sharply divided on the venter,
but the C0X8B are all black and red as on the male, and the
legs on both sides are somewhat infuscated. Tarsi longer
on the right (male) side. Wings well developed, on the
right side only, with pale veins and stigma and more like
those of the female. Length 7 mm.

This specimen was taken by Mr. Donisthrope July 20
or 21 from a large colony in Bewdley Forest, England.

5. Latebal Gtnakdbomobph of Fobmica sakouinba
Latbeille. — DoKiSTHOBPB (1909, p. 464, Fig. 2)

A nearly complete lateral gynandromorph, male on the
left, female on the right side. The head is of the female
type, rather small, with both of the antennae and the ocelli
female and the left eye a little larger than the right.
Head black, clypeus and right mandible red ; thorax evenly
divided into a black left and red right half, but only the
upper right comer of the epinotum red. A piece of the
scutellum and postscutellum red on the left side where
the wing is inserted. Petiole sharply divided into a red
right and left black half. Gaster black, the pilosity and
sculpture on the right half female, on the left half male,
the color being sharply defined on the venter. Legs and
coxaB female on the right, male on the left side. External
genitalia of the male type present on the left side. Both
pairs of wings fully developed, but the stigma and veins
darker as in the male. Length 9 mm.

This specimen was taken from the same colony as the
preceding.

Digitized by

Google

64 THE AMERICAN NATURALIST [VoL.XLVin

6. Fbontal Gynandbomobph of Solbnopsis fugax
Latreillb. — Santschi (1910, p. 649)

The head and thorax in this specimen are female, the
pedicel and gaster male. The head is somewhat smaller
than in normal females. The copulatory organs are those
of the normal male. Santschi remarks that it ** would be
interesting to observe the sexual behavior of such an indi-
vidual possessing a female brain and male genitalia.*'

7. Lateral Gynandbomobph (Ebgatandbomobph) of

MyBMICA SCABBINODIS NyLANDBB. — DONISTHOBPE

(1913, p. 44, PL I)

A nearly complete lateral ergatandromorph ; worker on
the right, male on the left side, the former being blackish,
the latter reddish yellow. Eight half of head larger than
the left, but with a smaller eye, striatorugose; right an-
tenna yellow, with a three- jointed club, its scape with the
usual strong lateral tooth at the basal flexure. Bight
half of thorax yellow, its epinotal half with a strong spine ;
right half of petiole and postpetiole yellow, rugose and
punctured ; right half of gaster pale fuscous yellow. Legs
on the right side of the worker type, yellow. Left side of
head blackish, ptmctate, not striatorugose, with a larger
eye and the median and left ocellus ; its antenna fuscous,
with four-jointed club. Left half of thorax blackish, its
epinotal portion unarmed ; left half of petiole and post-
petiole smooth, fuscous black. The greater part of the
left half of the gaster had been eaten away but the re-
mainder was darker fuscous th€in the right. Legs on left
side of the male type, fuscous ; wings on the left side only.

Donisthorpe remarks that this specimen, which was
picked up dead by Mr. Dollman at Ditchling, England, ap-
proaches the var. sabuleti Meinert in having the left
ant ennal scape longer than in the typical male scabrinodis
and the tooth on the right antenna large.

In conclusion I would call attention to a peculiar ant
described by Mayr (1868, p. 60) from the Baltic amber

Digitized by

Google

No. 565] GTNANDBOMOBPHOUS ANTS 55

and designated as a *'Zwitter** (gynandromorph) of
Hypoclinea constricta Mayr, or Iridomyrmex constrictits
as we must now call the species. Through the kindness
of Prof. A. Tornquist, of the University of Konigsberg,
I have been able to examine this specimen in connection
with many other amber FormicidsB. The general struc-
ture of the head, thorax and gaster is that of a worker,
though the thorax is not typical, as the base of the epino-
tum is less convex and less abruptly elevated, so that the
angle between it and the declivity is less pronounced in
profile. Mayr does not mention that the eyes are decid-
edly larger and more convex than in the normal worker
and therefore more like those of the male. There are a
few small white spots or bubbles on the vertex, which re-
semble small ocelli, but these organs seem to be actually
absent. The antennsB are 13-jointed and very long, as in
the male ; the scapes, however, are like those of the worker,
but extend well beyond the posterior borders of the head,
whereas joints 2-11 of the funiculi are cylindrical, sub-
equal and fully three times as long as broad, the terminal
joint being somewhat longer than these, the first shorter.
In the gaster, whioh is shaped as in the normal worker,
there are five distinctly visible segments, but the tip shows
•clearly the small, hairy, external genital valves (stipes)
of the male. The legs are also more slender than in the
normal worker and therefore more like those of the male.
At first sight thissingular insect seems to be a gynandro-
morph, as Mayr supposed, or more specifically, an erga-
tandromorph of the blended type, with worker characters
preponderating in the trunk and those of the male pre-
ponderating in the eyes, appendages and genitalia. It is
possible, however, to regard this specimen as an ergato-
morphic male, like those which occur normally in certain
species of Ponera, Cardiocondyla, Formicoxenus, Sym-
myrmica and Technomyrmex. Unfortunately we are not
in a position to decide between these alternatives, because
we are dealing with a single fossil specimen and are not
even sure that it belongs to the species to which Mayr

Digitized by

Google

56 THE AMERICAN NATURALIST [VoL.XLVni

assigned it. Still the case is interesting if only because
it suggests the further question as to whether the ergato-
morphic males in the genera above cited may be regarded
as originally frontal ergatandromorphs, with worker
head and thorax and male gaster, that have become the
only males of the species. If this is true, the ergato-
morphic males may have arisen by mutation from patho-
logical or teratological forms and have been preserved in
certain species in which peculiarities of habit rendered the
fecundation of the virgin females in the nest by wingless
males more advantageous than the type of mating ex-
hibited by the nuptial flight. A moment *s reflection shows
that the nuptial flight is a highly advantageous institu-
tion in common ants that form large colonies, but must be
as decidedly disadvantageous in the case of very small,
rare ants whose colonies are very sporadic and comprise
only a few individuals. This is actually the condition
seen in all the species with ergatomorphic males in the
genera Ponera, Cardiocondyla, Formicoxenus, Symmyr-
mica and Technomyrmex, and may be supposed, there-
fore, to accotmt for the substitution of the wingless, erga-
tomorphic for the normal winged males in these species.

LITERATURE

1908. Adlerzi G. Zwei Gynandroinorpheii von Anergates atratulus Schenck.

Arhiv. for ZooL, V, 1908, No. 2, 6 pp., 2 pis.

1909. Donisthorpe, H. S. J. K. Formica sanguinea, Ltr., at Bewdley, with

an Account of a Slave-raid, and Description of two Gynandro-

morphs, etc. Zoologist, 1909, pp. 463-466, 2 figs.
1913. Donisthorpe, H. S. J. K. Some Notes on tihe Genus Myrmica, Ent.

Becord, XXV, 1913, pp. 1-8, 42-51, 1 pi., 10 text figs.
1868. Mayr, G. Die Ameisen des batischen Bernsteins. Beiir, Naturk,

Preussens, 1. Konigsberg, 1868, pp. iv + 102, 5 pis.
1907. Santschi, F. Fourmis de Tunisie Captur^es en 1906. Bev. Suisse

Zool, XV, 2, 1907, pp. 305-334, 7 figs.

1910. Santschi, F. CJontributions k la Faune Entomologique de la

Roumanie. Formicides Captur^es par Mr. A. L. Montandon.
Bull, 8oc. 8ci. Bucharest, XIX, No. 4, 1910, pp. 648-651.
1903. Wheeler, W. M. Some New Gynandromorphous Ants, with a Review
of the Previously Recorded Cases. Bull, Amer, Mus, Nat, Hist^
XIX, 1903, pp. 653-683, 11 figs.

Digitized by

Google

SHORTER ARTICLES AND DISCUSSION

ON THE RESULTS OP INBREEDING A MENDELIAN

POPULATION: A CORRECTION AND EXTENSION

OP PREVIOUS CONCLUSIONS*

In a recent paper by the present writer on inbreeding,* the con-
clusion was reached {loc, cit., p. 608)

that no increase in the proportion of homozygotes in the population
follows inbreeding save under one or the other of two special condi-
tionsy viz.:

(a) Continued self-fertilization.

(b) Some form of gametic assortative mating which increases the
natural probability of like gametes uniting to form zygotes.

This conclusion is entirely correct as it stands, but also barren,
for it overlooks the very essential fact that any sort of inbreed-
ing involves in greater or less degree ''gametic assortative mat-
ing." The mathematical demonstration on page 608 of the paper
referred to is also entirely correct so far as it goes, but it stops
too soon. Up to the third generation of brother X sister mating
starting from a population of complete heterozygotes there is no
increase in the proportion of homozygotes beyond that prevailing
in a general Mendelian population. In the fourth and later gen-
erations there is, however. The blunder, kindly pointed out to
me by Professor E. M. East, which in retrospect seems altogether
too stupid even to be possible, was in the failure to recognize that
after the second generation the constitution of the family would
no longer be the same as that of the population. This is the point
which makes illegitimate the extension by induction of the results
up to the third generation to the generations beyond.

The genei|l conclusion of the former paper quoted above,
should then l^as follows : An increase in the proportion of homo-
zygotes in the population will follow inbreeding of any sort,
though at different rates for different types of inbreeding,
because any inbreeding involves homogamy (or assortative fnat-
ing) in some degree.

Having made clear the location and nature of the error I desire
now to show in some detail exactly what results follow from con-

1 Papers from the Biologieal Laboratory of the Maine Agricultural Ex-
periment Station^ No. 54.

1 < < A Contribution Towards an Analysis of the Problem of Inbreeding, * '
Ambucan Naturalist, Vol. XLVn, pp. 577-614, 1913.

Digitized by

Google

58 THE AMEBIC AN NATURALIST [VoL-XLVHI

tinued brother X sister mating in a Mendelian population. To
this we may now proceed.

The Distribution op a Mendeuan Population in Succbssivb

Generations with Continued Brother X Sister

Mating

Let us start with a population composed entirely of complete
heterozygotes. We shall consider a single character pair, A
denoting the dominant character, and a the recessive. The com-
plete heterozygote individual will then be Aa, and will produce
in equal numbers A and a gametes.

In making an analysis of the effect of inbreeding on the popu-
lation it will be necessary to deal not merely with the distribu-
tion of individuals in each generation, but also with the distribu-
tion of families of the several types. Each mating will produce
an array of families, as well as an array of individuals. The
standard family throughout this discussion is taken as including
32 individuals, of which 16 are males and 16 females. It is
further assumed that there is no sex-linkage of characters, and
that in any family there will be an equal number of brothers and
sisters of each zygotic constitution represented. One family of
16 pairs of brothers and sisters will make 16 matings and pro-
duce 16 families of 32 individuals each. This constant rate of
fertility is assumed throughout the discussion.

Every mating made is of a brother with his sister.

With so much by way of preliminary definition of the limita-
tions of this investigation, let us proceed to the actual analysis.

First Oeneration

Constitution of the Population. — ^By hyi)othesis all individuals
are Aa.

Proportion of Homozygotes in this Oeneration. — 0 per cent,
of the whole population.

Matings to Produce the Second Oeneration. — Start with one
brother X sister pair of individuals from this population. The
mating will be Aa X Aa. This will produce one family, 8AA
+ 8Aa + 8aA + 8aa.

Second Generation

Constitution of the Population. — 8-4.-4. + SAa -f- 8-4.a -f- 8aa.
Proportion of Homozygotes in this Oeneration. — 50 per cent,
of the whole population.

Digitized by

Google

No. 565] 8H0BTEB ARTICLES AND DISCUSSION

Matings io Produce the Third Oeneration. — ^The matmgs of
the one family of this generation will be as follows:

cfo"

d'd'

(1) AA

(9) oA

(3) oA

(2) AA

(5) AA

(10) o4

(7) oA

(3) AA

(9) AA

(11) aA

(11) a4

(4) AA

(13) AA

(12) oA

(15) oA

(5) Aa

(2) .ia

(13) aa

(4) «a

(6) Aa

(14) ao

(8) aa

(7) Aa

(10) 4a

(15) aa

(12) aa

(8) Aa

(14) 4a

(16) aa

(16) oa

TfctVd Oeneration Families Produced. — (Note: the numbers
in parenthesis are to identify matings and their consequent
families.)

(1)

33
16
16

(2)

16
16
32

(3)

(4)

(5)

Vft

8
8

16
8
8

(«)
(7)
(8)

(9)

8
8

16
8
8

(10)
(11)
(12)

(13)

32
16
16

(14)

(15)

(16)

Third Oeneration
Constitution of the Population. —

128AA + 1284a + 12804 + 128aa.

Proportion of Homozygotes in this Oeneration. — 50 per cent,
of the whole population.

Fourth Oeneration Families Produced. — The third generation
families, when mated, will produce families as follows:

Summarized this gives the following fourth generation families
produced :

(1)4 86 families like (1),
(2)4 24 families like (2),

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVIH

(3)4 4 famiUes
(4)4 24 families
(5)4 80 families
(6)4 24 families
(7)4 4 families
(8)4 24 families
(9)4 36 families

like (4),
like (5),
like (6),
like (8),
like (13),
like (14),
like (16),

Aa ' Aa

Family (l)t will produce . . 16 families of constitution

32
32
16
16

8
8

Families (2)t will produce 4 families of constitution

(3)i, (5)i will produce. . . . +4 families of constitution

"s"

and (9)i will produce +4 families of constitution

each will produce +4 families of constitution

and (13)i

8
8

"s"

each

Families (6)i will produce . 1 family of constitution

32
16

"ie *

1 1

(7)i, (10)t will produce. . . +2 families of constitution

and (ll)l will produce t , - +1 family of constitutinn

***8 "
16

+2 families of constitution
+4 families of constitution
+2 families of constitution
+1 family of constitution

16
8

**32 *
16

"s"

+2 families of constitution

+1 family of constitution

Families (8)1 will produce 4 families of constitution
(12)i, (14)i will produce. . +4 families of constitution
and (15)t will produce. . . +4 families of constitution
each will produce -{'4 families of constitution

8
16

8
'16 '

8
16
16

Fftmily (lft)t Will produce 16 fftmiliAif of constitution

Fourth Oeneraiion
Constitution of the Population. —

2560 AA + 1536Aa + 1536aA + 2560aa.

Proportion of Hotnozygotes in this Generation. — 62.5 per cent,
of the whole population.

Fifth Generation Families Produced. — The third generation
families, when mated, will produce families as follows :

(1).

will produce 36 X 16 =

576 families like

(1).

(2).

will prodDce 24 X 4 =

96 families like

(1).

+ 96

families like

(2).

+ 96

families like

(5).

+ 96

families like

(6).

(3).

will produce 4 X 16 =

64 families like

(6).

(4),

wiU produce 24 X 4 =

96 famiUes like

(1).

+ 96

families like

(2).

+ 96

families like

(5).

+ 96

families like

(6).

Digitized by

Google

Ko.665] 8H0BTER ARTICLES AND DISCUSSION 61

(5)4 will produee 80 families like (l)t

+ 80 X 2 = 160 familiee like (2).

+ 80 families like (4),

+ 160 families like (5).

+ 80 X 4 = 320 families like (6),

+ 160 families like (8).

+ 80 famines like (13),

+ 160 families like (14),

+ 80 families like (16),

(6)4 will produce 24 X 4 = 96 families like (6),

+ 96 families like (8),

+ 96 familiee like (14),

+ 96 families like (16),

(7)4 will produce 4 X 16 = 64 families like (6),

(8)4 will produce 24 X 4= 96 families like (6),

+ 96 families like (8),

+ 96 families like (14),

+ 96 families like (16),

(9)4 will produce 36 X 16 = 576 families like (16),

Summarized this gives the following fifth generation families
produced :

(1), 848 families like (1),
(2), 352 families like (2),
(3), 80 families like (4),
(4), 352 families like (5),
(5), 832 families like (6),
(6), 352 famUies like (8),
(7), 80 families like (13),
(8), 352 families like (14),
(9), 848 families like (16),

4096 (=16X256).

Fifth Oeneration

Constitution of the Population, —

44,73644 + 20,4804a + 20,48004 + 44,736aa.
Proportion of Homozygotes in This Oeneration. — 68.75 per
cent, of the whole population.
Sixth generation families produced:

17,216 families like (1),

4,480 famiUes like (2),

832 families like (4),

4,480 families like (5),

11,520 families like (6),

4,480 families like (8),

832 families like (13),

4,480 famities like (14),

Digitized by

Google

THE AMERICAN NATURALIST [VoL.XLVin
17,216 families like (16),

65,636 (=16X4,096).

Sixth Oeneration
Constitution of the Population, —

786,43244 + 262,1444a + 262,144a4 + 786,432aa.

Proportion of Homozygotes in This Oeneration. — ^75 per cent,
of the whole population.

From this point on it will not be necessary to carry out the
work in detail. The final results are given in Table I for four
more generations.

TABLE I

Showing the Constitution of the Population After 7 to 10
Generations of Brother x Sister Mating

Gener-
fttion

<Ul

Percentage
of Homozj-

gotesin
Whole Pop-
ulation

13,369,344

224.395,264

3.724,541,952

61,337.501,696

3,407,872

44,040,192

570,425,344

7,381,975.040

3,407,872

44.040,192

570,425,344

7,381,975.040

13,369,344

224,395,264

3.724,541,952

61,337.601,696

79.69
83.59
86.72
89.26

It is evident that the proportion of homozygotes is approaching
100 per cent, in the same manner as in the case of self-fertiliza-
tion, worked out by East, Jennings and others, but at a slower
rate.

In a later paper I hope to take up the problem of the general
formulae for finding the constitution of a Mendelian population
after n generations of inbreeding of the different types, and at
the same time discuss the relation of these results to the coeffi-
cients of inbreeding described in my former paper. It should be
specifically mentioned that, in the light of the data here set forth,
those criticisms of the conclusions of East and Hayes made in my
former paper^ which were based on the erroneous assumption of
a fundamental difference between self-fertilization and all other
forms of inbreeding in respect to homozygosis, have no validity
whatever. It scarcely needs to be said that the blunder on the
theoretical side here corrected in no wise affects thS usefulness of
inbreeding coefficients. Raymond Peabl

« Cf. Pearl, loc. cit., p. 606, 609 and 610.

Digitized by

y Google

ISOLATION AND SELECTION ALLIED IN PRINCIPLE

There are those who folly recognize the influence of natural
selection in transforming the hereditary characters of a species,
but are unable to see how isolation should have any effect of that
kind. They say that you may divide a species into two branches
between which all possibility of crossing is completely prevented,
but if the environment surrounding each branch is the same, the
natural selection to which each is subjected will be the same, and
no divei^nce of character will take place. They forget that the
separate branches, if prevented from crossing for many genera-
tions, are sure to develop different types of variation, and in due
time different methods of using the same environment, and are
therefore liable to subject themselves to different foims of selec-
tion. Again they forget that when the power of dispersal is
highly developed in a species it may be exposed to diverse en-
vironments and therefore to diversity of selecting influences, and
still remain one harmonious species, because free crossing is
maintained between all parts of the species. As long as there is
no isolation of different branches, that is, while free crossing con-
tinues, there is no permanent divergence resulting in diverse
races or species, even though the one species is exposed to differ-
ent forms of selection in different parts of its habitat.

Diversity of evolution, producing many divergent forms of
animals, could never have arisen without continuous isolation be-
tween the different forms.

Again there are those who maintain that selection unaided by
isolation can not produce transformation. It is true that diver-
gent groups can not be produced and intensified without isola-
tion; but a given race may be transformed by selection without
being divided into two groups by isolation.

Heredity with variation is the active cause of transformation ;
isolation and selection are the conditions that shape the f6rms of
heredity and variation.

It is a law of heredity, that, if those of a given stock that are
most alike in hereditary characters mate with each other, there
will be a tendency in their offspring to a stronger emphasis of
that character.

Digitized by

Google

64 THE AMERICAN NATURALIST [Vol.XLVHI

Another law of heredity is that as long as free crossing is main-
tained between the different forms of a species these forms can
not become widely divergent. The elephant and the mouse could
never have been developed from one original stock while free
crossing continued.

Now there are many ways by which the free crossing of one
variation with others of the same species may be prevented, but
they all come under two groups.

Under selection are classed all the influences enabling certain
variations to reproduce more successfully than other variations,
and 80 preventing free crossing between the successful and the
unsucessful. Under isolation are classed all the influences that
prevent living, and sexually reproducing creatures, from freely
crossing.

Under normal conditions there is no crossing between the ass
and the horse, though there is reason to believe that the early
ancestors of each were of one stock freely interbreeding and pro-
ducing fertile offspring. If isolation had not existed for ages be-
tween them, they could not have become the separate creatures
that they now are. Heredity can combine only compatible char-
acters. In some cases, incompatible characters arise between
creatures of the same race preventing any crossing between them,
as when a dextrally twisted moUusk produces a sinistrally twisted
one; but, in most cases, such incompatibility arises only after
isolation, through geographical separation, for many generations.

In view of these facts, is it not plain, that, in the case of a
variable and plastic organism, races more or less divergent will
be produced, if for many generations the organism is divided
into branches that are prevented from crossing! Is not such a
result just as sure as the gradual transformation of the race
under a slow change of climate, when the successful variations
are prevented from crossing with the unsuccessful variations!

John T. Guuck

Honolulu, T. H.

Digitized by

Google

)

\
\

The American Journal of Science

Established by Benjamin Silliman in 1818

The Leading Scientific Journal in the United States

Devoted to the Physical and Natural Soiences, with special reference to
Physics, and Chemistry on the one hand, and to Geology and Mineralogy
on the other.

Editor: EDWARD 8. DANA.
Aasoolftto Editors : ProfMSOr OEORaB L. OOODALE. JOHN TROWBRIDOE, W. Q. FARLOW and WM. M. DAVIS
of Caxnbridse ; ProfesBors A. E. VERRILL, HENRY 8. WILLIAMS and L. V. PIRSSON. of New
HaTon; Professor JOSEPH 8. AIMES, of Baltimore ; MR. J. 8. DILLER. of WasMngton.

Two volames annually, in monthly numbers of about 80 pages each.

This Joamal ended its first series of 50 volumes as a quarterly in 1845 ; its second
series of 50 volames as a two-monthly in 1870 ; its third series as a monthly ended Dec-
ember, 1895. A Fourth Series commenced in 1896.

Subscription price* $6 per year or 50 cents a number, postage prepaid in the United States
$6.25 to Canada ; $6,40 to Couotries in the Postal Union. Back numbers at reduced prices.

j^rTen- Volume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address

The American Journal of Science
New Haven, Conn.

University Control

By J. MoKbbn Cattell, Professor of Psycholc^ in Columbia University
Tofrther with a nrias of Two Hundred and Ninety-nino Unsigned Letters by Leading Men of Seienoe
[ Aeadeaaie Posttkms and Articles by Josbjph Jastbow, Qaonoa T. Ladd, Jokv J. Sramfaoa, J. E.
s, J. If oKsaii Cattbll, GaoBaa M. Stbaitoh, Sxbwabt Patos, Jokv Jat Chapmah, Jambs P.
Muasiis and Jaoob Goiru> Sobubman.

A ireat TaHeiy of qnestioiis oonceming general university administration are dealt with in an original and
jsetjilul wsyw — AToliirf •

Theee qnotations and examples are taken from Professor CatteU's informed and thorough discussion of the
aiib|«et of university oontrol, a subject upon which he has had much to say of late, finding occasion for oaustis
■ lliiisiii of existing Amen an couditionB, and standing as the champion of an academic democracy and a teach-
Ina professioD upon which a man may enter without forfeiting his self-respeot. — Th§ DidL

Senteoess and paragraphs that betoken the expert, highly-trained mind, the suggestions that oome to t^
ffffli a d tsU us that a new day is about to dawn in eduoatioual wilting. — ThB Botkm Bvemng TramacripL

SCIENCB AND EDUCATION

A seriea of volames for the promotion ofsoientiRo reae&roh and educational progress
TOLUME L The Foundations of Science. By H. PoxNOABi. Containing the authorised English trauAlation
hf Qwowam Bbuob Halstbo of ** Science and Hypothesis,'' "The Value of Science," and "Science
and Method.'' Price, $3.00 net.
TOLCTME n. Medical fiesearch and Bdncatlon. By Bzobabd M. Pbabob, Whjjam H. Wblgb, W. H.
HowBUi, FBAWKiiOr P. Malz*, Lbwblltb F. Babkbb, Chablbs S. Mznot, W. B. Cakmon, W. T.
CcfmnoMAM^ Thbobald Smith, G. N. Stbwabt, C. M. Jackson, E. F. Lton, Jambs B. Hbrbiok, Jobh
M. DoDSOMt C R. Babdbxr, W. OPHeLs, 8. J. Mbiasbb, Jambs Ewino, W. W. Ejbn, Hbhbt H. Dobald-
aoai; GBBwnAm A. Hbbtbb, and Hbbbt P. Bowditcb. Price, $3.00 net.
VOLUME m. Ifntr^rsity Control. Norn Bmtdy. Paget x+48^ Price, %3M nei.

OAKKisoN, N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The American Naturalist

A ilcMlhlr Joanial}e^bliahed iii'1867. D;Tolad to tli« Advancement of the BiologieiJ Sctaeot
witb Spoeial RoforoBCo to the Facton of Oryaiiie EvolotioB and Heredity

• > ^ 1 ^ • ... .

CONTENTS or THE JULY NUMBER . ^
Doetrlnts held ai VitaUna. ProfeMor H. & Jen^ngs.
Tht PrMenoe of the Barred Plumage Pattern in the

White Leghorn Breed of Fowli. Dr. Philip B.

Hadley. ' '
Shorter Artiolet and Diseuislon: The Geologic Work

of Termite* in the Belgian Congo. Donald BteeL
Noteiand LlteraChre^ Work in Geneiicl>roblemiin

Protozoa at Yale. Profeasor^A. R. Hiddleton.

Notes on lohthyology. President DaTid Starr

Jordan.

OONTENTB OF THe' AUGUST NUMBER

Oeneti(*al Studies on Oenothera. IV. Dr. Bradley

Hoore Davis. •
The Inflnenoe of Protracted and Intermittent Pasting

upbnQrowth. Dr Sergiuf Morgnlis.
Cambrian Holothurians. Austin H. Clark.
Shorter Abides and Discustlon: Viability and CouiT

ling in Drosophila. P. W. Whiting. TheResnltt

Obtaiu< d by crossing sea mail L. and Enchlaenfc

mezicana Schrmd, Mary Q. Lacy. ,

OONTINT8 OF THE •EPTEMBER NUMBER
Thft Natoral History of «h« Nine-banded Armadmo
ofTexM. Professor H.H. Newman.

Genatical Studies on Otnothera. IV. Dr. Bradley
M. DaTis.

NotMon a Dlflerential Mortality obserrwl betw««ii
TWMtmohsonrUandT.moUtor. Dr. Boss Aiken

OOftDtf.

CONTENTS OF THE OCTOBER NUMBER

A Contribution towards an Analysis of the Problem
of Inbreeding. Dr. Raymond Pearl.

The Inheritance of Coat Color in Hones. Professor
W. S. Anderson.

The Variations in the Numberof Vertebrae and Ven-
tral Scutes in Two Snakes of the Qenus Reglna.
Professor Alexander G. Ruthren and Crystal

Shorter Articles and Reports : The Simultaneous

fessor R. A. JSmerson. The Fourth International
Genetic Conference : Dr. Frank M. Surfaee.

OONTENTB OF THE NOVEMBER NUMBER

The Bffleet on tho OfTspring of IntozlcAting the Male
PMwit and tht Transmission of the Dsfttets to

Shortor Articles Mid Discnssion: Reciprocal Crosses
between Reeye's Pheesant and the Common
Rlngneek Pheesant prodncing Unlike Hybrids.
JohnCPhilUps.

CONTENTS OF THE DECEMBER NUMBER

The Fixation of Character in Organisms. By Edward
Sinnott.

Ramaley.
Supplementary Studies on the Differential Mortality

with Respect to Seed Weight in the Gemination

ef Garden Beans, n. Dr. J. Arthur Harris,
Shorter Articles and Discussion : A Croas iUTolTlag

FourPainofHendelianCaiaractersinMioe. C

C. Little. J. C.PhiUips.
Index to Volume XLVH.

Single Number 40 Cents Yeariy SubMrriptkm, $4.00
Th« NATUiUUJST wm be emt to fe.#w MibecribM lor foor montlM to

THE SCIENCE PRESS

Garruon. N. Y. Sab^tation 84 : NEW YORK LancMtcr. Pa.

Digitized by

Google i

VOL. ZL7m, HO. 666 FEBKVABT, 1914

THE

AMERICAN
NATURALIST

A MOHTHLT JOUBSAL

Seroted to the Advancement of the Biological Scieneea wtOk

Special Beference to the Factors of Evolution

COHTEHTS

Page
L Some Hew Varieties of Rats and Onlnea-plgs and their Relations to Prob-
lems of Color Innerltanoe. Professor W. E. Castlb .... 65

ZI. ** Dominant" and'* RecesBlye" Bpottlnf In lUoe. C. C. Little - - T4
m. On DUforentlal Mortality with respect to Seed Weight occurring In Fl«ld

Onltnres of Plsnm sativum. Dr. J. Artuub Hakbis - . - - 83
rv. Tne Inheritance of a Recurring Somatic Variation In Variegated Ears of

MaUe. Professor R. A. Embrson --------87

T. Restoration of Bdaphosaurus crudger Cope. Professor E. C. Case - - 117
VI. Shorter Articles and Discussion :

Humidiiy~ft Neglected Factor in EnyiroDmental Work. Dr. Frank E. Lutz 122

THE SOIENOE PRESS

LAH0A8TEB, PA. GABBI80N, H. T.

NEW YORK: SUB-STATION 84

Digitized by

Google

Two Important Biological Journals

The Journal of Qenetics

Edited bj
W. BATESON and K. C. PUNNETT
A periodical for the pablioation of reoords of original reeearoh in heredity, Tariation, and allied topioi.
The JoUBKAX also oontaina, from time to time, artioies snmmarizing the existing state of knowledge in the
Tarioos branches of genetics. The numbers are fully illustrated, and, when the subject-matter demands it,
colored plates are freely employed. The Journal is issued in parts as material accumulates, and a volume
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7.60 a Tolome ;
single copies, |2.50.

3iometrik:£i

Edited by KARL PEARSON
A journal for the statistical study of biological problems, appearing about four times a year. A Tolnme
containing about 500 pages, with plates and tables, is issued annually. Volume X b^ns in 1914. The
nine preceding Tolumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
craniometry, medicine, heredity, etc. Subscription price, $7.60 a volume ; single copies, 12.60.

Prices on back numbers of these journals vary, and postage from London is charged on back volamee
and single copies.

Published in America for the Cambridge University Press of England by

The University of OHicas:o Press

ChicasTO, Illinois

University Control

By J. MoKsBN Cattbll, Professor of Psychology in Golombift Uniyersity

Together with a series of Two Hundred and ^nnety-nine Unsigned Letters by Tieading Men of Sefattos
holding Academic Positions and Articles by Joseph Jabtbow, Qbobob T. Ladd, Johh J. Stbysmsoh, J. S.
GBnoBTON, J. lioKmma GATmJk, QaoBoa M. STBATfov, Stbwabt Paton, Jomr Jat Chapmav, Jaiom P.
MuNBoa and Jacob Gould Sghubman.

A great variety of questions ooaceming general nniyecaity administration are dealt with in an origiiial and
helpful way. — Nature.

These quotations and examples are taken from Professor Cattell's informed and thorough discussion of ihm
subject of university oontrol« a subjeot upon which he has had much to say of late, finding oooasion for caosti*
eritioism of existing Ameri -an oonditions, and standing as the champion of an academic demoeraoy and a tnaoh
log profession upon which a man may enter without forfeiting his sdf-respect. — Th§ DiaL

Sentences and paragraphs that betoken the expert, highly-trained mind, the suggestions that come to v^
fresh a d tell us that a new day is about to dawn in educational writing.— TAs BotUm Ewemno TratueripL

SCIENCE AIND EDUCATIOIN

A series of voltunea for the promotion of acienti£o reaearcb and educational programs
VOLUME I. The Fonndationa of Science. By H. PoufOABi. Containing the authorised English translatioa

by Gbobob Bbvcb Halbtbd of "Science and Hypothesis," "The Value of Sdenoe," and "Scieno^

and Method." Price, $3.00 net.
VOLUME IL Medical Research and Education. By Riohabd M. Pbabcb, WnuAM H. Wblob, W. H.

HowBLL, Fbankliv P. MAUi, Lbwblltb F. Babebb, Chablbs S. Minot, W. B. Camhov, W* T.

CouNonjiAK, Thbobald Bmitb, G. N. Stbwabt, C. M. Jaokson, E. P. Lton, Jambs B. Hbbbiok, Jobs

M. DoDBOM, C. R. BABnBBN, W. OphOlb, S. J. Mbltzbb, Jambs Ewzno, W. W. Ebb», Hbbbt H. Dokjju>-

Boir, Cbbistiah A. Hbbtbb, and Hbnbt P. Bowditob, Price, $3.00 net.
VOLUME III. Unirecsity Control. Now Beady. Pagm x-\-48^ Price, %8M asl.

GARRISON, N. Y. THE SCIENCE PRESS Lancaster, pa,

SUB-STATION 64, NEW YORK CITY

Digitized by

Google

Digitized by

Google

66 THE AMERICAN NATURALIST [Vol. XLVm

race and as a mutation or unit-character variation, retro-
gressive in nature (i. e., due to loss of some normal con-
stituent from the germplasm). Each is a simple Mende-
lian recessive character in crosses with wild race, and
with certain at least of the tame varieties. The two varia-
tions have not as yet been combined by intercrossing, but
this will be attempted soon and, I doubt not, with entire
success.

My first information about the new variations was
obtained from Fur cmd Feather, the oflScial organ of the
English fanciers, in which appeared advertisements of
*Hhe new variety*' of black-eyed yellow rat. Now as
long ago as 1903 Bateson had commented on the singular
absence of a ** yellow'* variety among rats, noteworthy
because nearly all mammals kept in captivity have such
varieties ; and I have since been bold enough to publish
some speculations as to why this variation had not made
its appearance. Consequently I was mudh excited to
learn that it actually had appeared. Miss M. Douglas,
one of the editors of Fur and Feather, and secretary of
the National Mouse and Eat Club (of England) very
kindly answered my inquiries about the new varieties and
put me in communication with the ** originators,'* who
have given so clear and full accounts of their procedure
in establishing the new varieties that even the genetic be-
havior of the variations is fairly certain, though I pur-
pose to confirm this fully with experiments which are
already in progress.

The pink-eyed variation made its appearance first, so
far as known, about 1910 or 1911, but it had probably
been in existence for some time and become rather widely
diflFused throughout the central part of England, for at
about the same time pink-eyed wild rats were caught at
or near Preston and at Chesterfield, cities some 65 miles
apart. I am informed that Mr. T. Eobinson at Preston
and Mr. W. E. Marriott at Chesterfield independently
established the ** pink-eyed fawn" variety, or what would
better be called the pink-eyed agouti variety, since appar-

Digitized by

Google

No. 566] NEW VARIETIES OF EATS AND GUINEA P1Q8 67

ently it differs from the wild gray (or agouti) variety by
the pink-eyed variation alone. It is not a true yellow
variety at all genetically, though (like the pink-eyed gray
mouse) it resembles one superficially because of the yel-
low ticking of the agouti fur.

It is also quite distinct genetically from the albino
variation seen in white rats, yet its * * dirty white' ' color is
enough like the appearance of the albino to permit mis-
taking one for the other. Possibly this is why the pink-
eyed variation may have been for some time overlooked.

Mr. Robinson has not answered my inquiries, but Mr.
Mariott writes in detail about his observations and ex-
periments.

Under date of October 11, 1913, he says :

The first rat with any semblance of fawn in it that I had was canght
in a trap on a provision merchant's premises in Chesterfield. Yon conld
searcely call it a fawn, bnt more of a cream or dirty white. I have also
had fonr others similar to this one, 2 caught at the same place and 2
caught at a malt-house in close proximity to the other premises, [in aU]
3 bucks and 2 does, but the only one that I was able to get to breed was
the first brought to me, which was a buck. When first caught it was
very wild, in fact it appeared to me to be more wild than an ordinary
wild rat. It was a source of trouble getting it to mate, killing no less
than 20 does before mating. I eventually got it mated to 2 does, one a
pure white for at least 10 generations, and one black-and-white hooded-
and-striped, or Japanese rat. The result of the pure white cross was 2
young, a buck and a doe, wMch were agoutis with no white at all,^ The
result from the Japanese cross was 7 young, 5 does and 2 bucks, which
were the color of Irish agoutis being agouti color with a white stripe
running underneath. These results naturaUy caused me great disap-
I>ointment as I was expecting a fawn colored young one. When the
young were old enough I mated father and daughter, result nil; mother
and son, result ntl; brother and sister. The brother and sister mating
from the pure white cross produced 2 fawn colored rats, a buck and a
doe, and 5 agoutis.^ The brother and sister mating from the Japanese
cross produced 2 f awn-and-white Japanese, 1 cream-and-white Japanese,

1 Italics mine. Note the reversion to full wild color. This shows the
pink-eyed variation to be entirely different in nature from the ordinary
albino variation.

sNote the return of "fawn" (pink-eyed &gouti) as a recessive character
in approximately 1 in 4 young.

Digitized by

Google

68 THE AMERICAN NATURALIST [Vol. XLVHI

1 black-and-white Japanese, and 4 agoutis.^ The fawns and fawn-and-
whites resulting from these crosses were much deeper in color than the
wild grandsire. Mated one with another they gave a proportion of
about 2 fawn colored or fawn-and-white in 7 young.* I may say in
conclusion that the original wild rat was in shape of body, skull, etc.,
as the ordinary brown or agouti rat that we have running wild in
this district.

Mr. Marriott sold a ** fawn-and-white" (pink-eyed
hooded agouti) bnck to Mr. E. F. Tilling, of Hessenford,
who also ** originated '^ the second variation, the ** black-
eyed yellow," or true yellow variation. His results from
the pink-eyed variation confirm those of Mr. Marriott.

Mr. Tilling writes under date of October 18, 1913 :

I see by Fur and Feather this week that you are interested in the
yellow and cream varieties of rats. I am also much interested in these
and have produced the latter variety within the last few months. We
have 2 kinds over here, the yellow-and-white hooded with pink eyes and
the self yellow (and cream) with black eyes. Both are quite distinct.
The first mentioned was produced some 2 or 3 years ago. Mr. Harriott,
of Chesterfield; bred the first I heard of from a wild caught fawn. He
bred a couple of yellow and white hooded bucks of which Miss Douglas
bought one and I the other. I mated miae to about 15 does of various
colors and definite strains. He was a splendid breeder and got some very
fine youngsters, but not one of his own color from the first cross^ I
subsequently mated him to some of his daughters and they produced a
good proportion of yellow-and-white young.' These are now fairly
plentiful over here and are in the hands of several fanciers.

Of the other kinds, black-eyed fawns and creams, the first one ex-
hibited and from which all mine are descended, was a very fine wild
caught, deep colored, fawn specimen. I got her partly tame and ex-
hibited her at the National Mouse and Rat Club's annual show at Bristol
on November 27 and 28, 1912, where she won first in the self class and

8 " Fawn- and white Japanese'* means (to me) pink-eyed agouti with the
"Japanese" color pattern (hooded). The formation of this class of young
shows the hooded pattern ("Japanese") to be independent in transmission
of the pink-eyed variation. * * Cream-and-white Japanese," I interpret as
pink-eyed hlack (non-agouti) hooded. "Black-and-white Japanese" is the
familiar black hooded. We should expect this mating to produce also self
pink-eyed agouti and self pink-eyed black which are not mentioned.

4 The Mendelian expectation is 2 in 8.

5 Italics mine. Note again the recessive nature of the variation.

*Not real yellow-and-white, as already explained^ but pink-eyed agouti-
and-white or black-and-white.

Digitized by

Google

No. 566] NEW VARIETIES OF RATS AND GUINEA PIGS 69

was well commented upon in the fanciers' papers. From this doe I have
built up my strain of black-eyed creams. I mated her to a self black
buck and she bred 8 youngsters all wild colored.^ This is the only
litter I had from her, as shortly afterward, during my illness, my man
while transferring her from one cage to another let her get away and
was unable to recapture her. However, I have bred from her young-
sters, mating brother and sister, and the litters have invariably con-
tained at least 1 fawn or cream^ each time. I have now just bred for
the first time from the 3 first does so produced, again mating them to
their brother and the result is litters of 7, 5 and 7, respectively, all self
creams.®

From the statements of Messrs. Marriott and Tilling,
it is evident that the two variations, which they, respect-
ively, have introduced into the rat fancy, are both reces-
sive in heredity, as are also the three previously known
Mendelizing color variations of rats, viz., (1) the albino
variation (with uncolored coat and eyes) ; (2) the black
variation (lacking the agouti ticking of the fur) ; and (3)
the piebald ** hooded^' pattern of white and colored fur.
Each of these is known to be an independent Mendelizing
tmit-character. If the new variations are as supposed in-
dependent of each other and of those previously known,
they will make possible the immediate four-fold increase
in number of the previously known color varieties of rats.
If for the present we adopt a simplified terminology (as I
have elsewhere suggested) for the different color varia-
tions, employing small letters for such as are recessive
in heredity, we may use the following set of symbols :

White (albino) =w,

Black = h,

Hooded = h,

Pink-eyed = p,

Yellow = tf,

7 This shows that the original yellow animal was potentially an agouti.
A pair of yellows which Mr. Tilling has sent me have light bellies and I
presume are also potentially agoutis.

8 "Cream" here probably means yellow not transmitting agouti. It
probably lacks the lighter belly as do yellow rabbits which do not transmit
agouti.

9 This shows that extracted yellows breed true to yellow. Hence the
Tariation is recessive, as in rabbits and guinea-pigs, not dominant as in mice.

Digitized by

Google

70 TRE AMERICAN NATURALIST [VOL-XLVHI

By various combinations of these variations, if each is
independent of all the others, 32 varieties become possi-
ble. Half of these varieties will be albinos, white and so
visibly indistinguishable. The other 16, we have reason
to suppose, will look different from each other. Pre-
viously we had but four of these, the first four in the fol-
lowing list of the theoretically possible 16.

1. Normal or wUd agouti,

2. h black,

3. h hooded,

4. hh black hooded,

5. p pink-eyed,

6. pb pink-eyed black,

7. ph pink-eyed hooded,

8. pbh pink-eyed black hooded,

9. y yellow,

10. yh yellow black (t. e,, non agouti yellow),

11. yh yellow hooded,

12. yp yellow pink-eyed,

13. yhh yellow black hooded,

14. yph yellow pink-eyed black,

15. yph yellow pink-eyed hooded,

16. ypbh yellow pink-eyed black hooded.

Varieties 1-4 have been known for some time; they
have constituted the fancier's entire repertoire up to the
present time. Varieties 5 and 9 have apparently arisen
as wild sports obtained by Marriott and Tilling, respect-
ively. By crosses these gentlemen have apparently ob-
tained varieties 6, 7, 8, and probably 10. Varieties 11-16
are as yet unknown, but will doubtless soon be produced.
Corresponding with each of the 16 colored varieties, an
uncolored one should be possible of production, which
would transmit in crosses with any colored variety the
characteristics indicated by its formula. Albinos cor-
responding to colored varieties 1-4 are positively known
to occur ; their symbols would be w, wb, wh and wbh, re-
spectively. Symbols for the remaining 12 expected varie-
ties may be formed in like fashion, by prefixing w to the
combinations already given.

All the five unit-character variations, which in different
combinations are responsible for the color varieties of

Digitized by

Google

No. 566] NEW VARIETIES OF BATS AND GUINEA PIGS 71

rats, have their parallels in other mammals. Albinism
and white-spotting (which in rats takes the form of the
hooded pattern) are among the commonest. They occur
in practically all mammals from mice to men. Albinism
appears to consist in such a modification of metabolism
that the process of pigment-formation can take place only
feebly or not at all. That particular process which seems
chiefly affected is the production of yellow pigment.
Albinos, so far as I know, never produce genuine yellow
pigment, though they may produce considerable quanti-
ties of black or brown pigment, as in the case of the
Himalayan rabbit. An undescribed variety of guinea-
pig, which I obtained about two years ago in Peru, may
bear as much black pigment in its coat as wild cavies do,
yet it forms no yellow pigment at all. Further this varia-
tion behaves as the allelomorph of ordinary albinism, in-
dicating that it is probably of the same genetic character.
For this reason we may provisionally consider the albin-
ism of mammals as due to a loss of the ability to form
yellow pigment. This usually, if not always, involves a
lessened capacity to form other pigments also, so that it
seems probable that the same chemical process, which
produces yellow pigment as an end-product, is ordinarily
involved also in producing the higher oxidation stages
seen in brown and black pigment. In albinos this process
would seem to be omitted, or to be accomplished by some
step which does not involve the production of yellow
pigment.

The yellow variation is extremely common in mammals.
Yellow varieties, which at opposite extremes of intensity
of pigmentation are known as cream and red, occur
among horses, cattle, hogs, cats, dogs, rabbits, guinea-
pigs, mice and human beings. In this variation pigment
oxidation stops at the yellow stage, usually throughout
the coat but not in the eye. Described in negative terms
a yellow variety is one in which black and brown are sup-
pressed or restricted. Black and brown, though usually
restricted to the eye in yellow varieties, may occur also in

Digitized by

Google

72 THE AMERICAN NATURALIST [Vol. XLVIH

small quantities in the fur. Examples are found among
horses (bay and dun varieties), cattle (the Jersey breed),
dogs (the common dirty yellow variety), rabbits (the
* 'tortoise-shell'' variety), mice and guinea-pigs, and
probably red-haired human beings also.

Black varieties of mammals arise in two genetically
distinct ways. One is a quantitative increase or exten-
sion of black, the reverse of what happens in yellow varie-
ties, so that black encroaches on regions normally yellow
or may even obliterate them altogether. Examples are
found in black squirrels, in which the agouti yellow tick-
ing of the fur is almost, but not quite, obliterated by black
pigment. But the ''black'' variation of rats, mice,
guinea-pigs and ordinary rabbits results from a total
loss, not a covering up, of the yellow ticking of the fur
seen in agouti varieties. Genetically it is quite distinct
from the other kind of black. It is a recessive variation
and so breeds true.

The pink-eyed variation is the rarest of all the five
enumerated as occuring in rats. It has been known here-
tofore only in mice, though I have recently obtained it
also in guinea-pigs from Peru, where it seems to be well
established.

In this variation the capacity to form yellow pigment
is unimpaired, but only traces of black or brown pigment
are produced. Consequently varieties which possess the
other genetic factors of normal yellow animals have fully
pigmented (yellow) fur, but with very faintly pigmented
(pink) eyes, when they possess this factor. If, however,
they possess the other genetic factors of black, brown, or
agouti varieties, along with this pink-eyed variation, then
both the fur and the eyes are very faintly pigmented.
From this results the seeming paradox that pink-eyed
blacks are less heavily pigmented than pink-eyed yellows,
so that in rats the fanciers have called the former
"creams," the latter ''fawns."

When pink-eyed animals are crossed with albinos, off-
spring fully colored (eyes and all) result, as was first

Digitized by

Google

No. 566] NEW VARIETIES OF BATS AND GUINEA PIGS 73

shown by Darbishire some ten years ago. This indicates
that the two variations are not only genetically distinct,
bnt are physiologically complementary. The albino has
defective metabolism for producing yellow (and in conse-
quence brown and black also) ; the pink-eyed animal has
the full mechanism for forming yellow, but its brown and
black producing mechanism is defective. Together they
possess the full mechanism of normal color production.
Hence the reversion on crossing.

White spotting is clearly due to neither of the above
modifications, but to a different change in the metabolism
so that no pigment at all is produced. For an albino rab-
bit or guinea-pig may, as already observed, bear consider-
able black or brown pigment, but a white spot either on
an albino, on a pink-eyed animal, or on a fully colored
animal is entirely devoid of pigment. The paradox of a
white spot on an albino is obtainable by crossing a white-
spotted colored race with an albino race, which develops
some pigment in the fur, as for example the Himalayan
race of rabbits and guinea-pigs. In this way English-
marked Himalayan rabbits and spotted albino guinea-
pigs have been produced in my laboratory.

Postscript: While this paper was in press, Mr. Tilling,
in reply to a further inquiry, wrote that his original black-
eyed yellow rat was caught on a ship at Liverpool. The
fact that the pink-eyed variety was found in the same gen-
eral region leads him to believe that both variations were
introduced on ships from some foreign country. It would
be of much interest to know from what country or coun-
tries. Any information on this point obtainable from
rat-catchers or others would be welcome.

Digitized by

Google

*^ DOMINANT" AND ^'EECESSIVE" SPOTTING IN

MICE

C. C. LITTLE,
BussEY Institution, Habvahd University

Intboductoby

The inheritance of spotting has long proved of interest
to animal geneticists. The nature of spotting is snch
as to afford an excellent chance to observe quantitative
fluctuation and variations of very minute size. Further-
more, the fact that spotted varieties are found in all the
rapidly breeding smaller domesticated mammals has led
to a widespread investigation of its phenomena of in-
heritance.

One of the most clean-cut and constant types of spot-
ting which has been studied is that of the ** hooded*' pat-
tern in rats. This character was studied independently
by Doncaster (1905) and by Castle and McCurdy (1907).
All these observers agree that this form of spotting is
due to a recessive Mendelizing unit which gives a 1:3
ratio in crosses with self-colored races.

In mice there has been no such well-localized pattern
recorded and a series of spotted forms has been described
which vary from black-eyed whites on one end of the
series to heavily colored animals having only a few
white hairs on the forehead or on the belly at the other
extreme.

Cuenot, who did considerable work on the inheritance
of spotting in mice, came to the conclusion that spotting
is due to a group of recessive spotting factors which he
describes as pl,,p2, p3, p4, etc. His figures, however,
show a single unit character difference as 3 : 1 and 1 : 1
ratios prove.

Up to 1908 all the spotting in mice was classed as re-
cessive to solid-colored coat. At that time, however,

Digitized by

Google

No. 566] SPOTTING IN MICE ^ 76

Miss Dnrham described the appearance of dominant spot-
ting in addition to the recessive form which she also had
experimented with* Such a dominant form of spotting
is supposed, by Bateson, to be due to the addition of some
factor for restriction of pigment formation in certain
areas. This produces a dominant form of spotting as
contrasted with the recessive type, which, he holds, is due
merely to the loss of the **self ^* factor.

Hagedoom (1912) gives data to show that the domi-
nant form of spotting occurs in mice and in addition con-
siders it as produced by a factor analogous to that which
produces the dominant ** English'* spotting in rabbits.

The object of this paper is to present certain evidence
concerning the nature of dominant and recessive spot-
ting in mice ; to discuss in its light the results of the above-
mentioned investigations ; and to criticize one additional
point in Hagedoom 's work with mice.

EXPEBIMENTAL

Materials. — ^Among several wild mice caught during
the spring of 1911 was one individual with a white spot
' or ** blaze** on the forehead between the eyes. This spot

3 or ** blaze** was about one quarter of an inch in length

and one eighth of an inch in width. This mouse, an adult
j^ male, was transferred to a breeding cage and a series of

/ experiments was started to* determine whether the

''blaze** character was inherited and, if so, in what way.
As at that time no adult wild females were available
from unrelated stock the wild ''blaze** male (SI) was
crossed with a female from a dilute brown race. In
many ways this dilute brown race was the best possible
material for such a cross. It was very closely inbred,
being descended from a single pair of animals, progeny of
which had been free from out-crossing for naore than a
year. Further, it had never given, nor has it ever given
in hundreds of young, an animal with the slightest trace
of a spot, even on the tail, where white bands are fre-
quently seen in wild mice. Besides this the race was vig-
orous and active and yet easy to handle.

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVHI

Results

As a result of mating SI ** blaze" with a female of this
dilute brown race, two litters, totalling eight young, were
produced. All these young were self-colored without a
trace of white, and, as expected, all resembled the male in
coat color.

The Fi generation self s were then crossed in two ways,
(1) inter se and (2) with animals of the dilute brown self
race to which their mother belonged. It is hoped that a
detailed account of all the matings made may be pub-
lished later, but for the present purposes certain of the
crosses under the first heading will suflSce.

When Fi was crossed inter se, two sorts of young were
produced, namely, those with white and those without.
While all of the latter tyi>e may be classed as self, the
former were of two general sorts: (1) those with a
*' blaze '^ as large or larger than that of 81, these we may
call ** blaze'' animals; and (2) those with only a few
white hairs on the forehead, which we may call few white-
haired (f.w.h.) animals.

The exact numbers in this cross were

0/ftpriDg

Ptrents

Self

F.W.H.

Blase

56 X 55— 58

519 X58

518 X 58

When the Fj few white-haired animals were bred to-
gether they produced three types of young: few white-
haired, blaze and self, as follows.

Oflbpring

Parents

Self

F.W.H.

Blaze

3,030 X 3,028

11
5

6
6

3,043 X 3.028

16 1 11

One further fact is also of interest. Various descendants
of F2 ** blaze" animals, which should breed as recessives,

Digitized by

Google

No. 566]

SPOTTING IN MICE

have given the following resnlts. The generation num-
bers may be disregarded as they refer to another method
of classification. It is to be remembered that the parents
in the tabulation given below, are all ** blaze'* in
character.

Yoaog Prodooed

Genention

BUM

BlaMMdVeo-i
tnl White j F.W.H.

ToUl

F4B

F»B

F.B

157

6
60
53

27
5
0

1
3
0
0

247

128

Total ' 269

125 1 36

434

If the ** blaze '* is a true Mendelian recessive we should
expect all 434 offspring to have some white on them.
The figures show that 430 of the 434 are of this type;
that is to say, approximately 1 per cent, are self.

It is possible to account for the occasional production
of selfs even if the ** blaze'* character is a true recessive,
if we supposed that there are supplementary factors
which may influence color development; and it is quite
conceivable that such is the case.

The chief point of interest in the crosses given above
is that while spotting behaves in F^ as a recessive, certain
of the F2 spotted individuals fulfil the requirements of
dominant spotting by producing self offspring.

The spotting came from a single individual and can
scarcely be considered to be of two distinct types.

We may now consider the bearing of these results on
the work of Miss Durham and Hagedoom.

Miss Dukham's Results
Miss Durham (1908) gives a detailed account of a re-
cessive type of spotting in mice. The numbers she ob-
tained are extensive, and the case seems well established,
coming as it does in corroboration of the work of Cuenot,
Darbishire and others. In the same papers she records
the occurrence of a dominant spotted type of mice. Bate-
son (1909), commenting on the case, compares it with the

Digitized by

Google

78 THE AMERICAN NATURALIST [Vol. XLVm

dominant ^^ English'' spotting in rabbits but also agrees
that, in the case of mice, there is no criterion to enable one
to distinguish somatically between the dominant and re-
cessive forms. This, of course, is not the case in rabbits
where the ** English'* pattern differs visibly from the
''Dutch" spotting, which Hurst (1905) found to be re-
cessive to self. Bateson also considers that the case of
dominant spotting in mice, reported by Miss Durham, is
the result of a different spotting factor from that pro-
ducing recessive spotting.

In terms of the presence and absence hypothesis this
means that the dominant form possesses a factor for re-
striction of pigmentation which self forms lack. This
fact becomes of interest when Miss Durham's experi-
mental results are closely examined.

In the race which gave rise to the dominant spotting
the following conditions are seen.

A sooty yellow spotted mouse of unknown origin was
crossed with a black-eyed white (spotted) animal (of
Atlee's strain). Among other progeny was obtained a
black-eyed white mouse with ''agouti ears." This
mouse. No. 21 (spotted), was crossed with an albino (car-
rying chocolate), No. 35, and gave among its progeny No.
69, a black self mouse. This black animal. No. 69 was
crossed with an albino (carrying chocolate), No. 34, and
from these two individuals came the dominant spotted
race.

Now inasmuch as No. 34 and No. 35, the albinos, were
not supposed to carry spotting, the dominant spotting
must be considered as probably coming from No. 69, a
black self animal. We know that this animal must carry
spotting as a recessive character since its parent. No. 21,
was spotted.

If, therefore, this animal was the progenitor of the
dominant spotted race, and if he carried a recessive spot-
ting, as it seems certain he did, we must suppose that one
of three things has happened to the recessive spotting
which he carried.

Digitized by

Google

No. 666] SPOTTING IN MICE 79

1. It may have been completely lost, failing to manifest
itself in his germ celk.

2. It may have continued to exist and to be inherited
together with the dominant type of spotting.

3. It may have been changed to a so-called *' dominant*'
type of spotting simply by the nature of modifying sup-
plem^itary factors which it encountered during ontogeny.

The first two cases necessitate the origin of the *' domi-
nant*' spotting by a mutation in no way connected with
the previous recessive spotting. In the first case, more-
over, we should have to suppose the disappearance of the
recessive spotting character in a manner entirely con-
trary to any principle of Mendelian heredity. In the sec-
ond case the occurrence of the two types of spotting side
by side in the same litters of young would so complicate
the experiments that analysis would be diflScult if not im-
possible, on Miss Durham's results.

There is good reason to believe that the third possible
explanation is the correct one. It accounts for the for-
merly *' recessive" type of spotting. It presupposes no
fundamentally different appearance of the two types
of spotting. Moreover, it is very probable that the al-
bino race brings in the modifying factors necessary to
give the apparent change in the type of spotting. The
addition of a factor as presupposed by the presence and
absence hypothesis is not proved by the results obtained
nor is it necessary to account for them.

That the presence and absence hypothesis does not
apply to all cases of spotting is seen in the case of the
** blaze" mice in my experiments. Here, if Fi animals
had been given me as a starting point for experimenta-
tion, I should conclude the spotting to be recessive, while
if Fj spotted animals were given as a starting point the
conclusion would be inevitable, that spotting should be
considered dominant. Yet it is one and the same spot-
ting in both cases. It is certain that ''self" and ''blaze"
are alternative conditions, but it is equally certain that
they differ from each other rather as two degrees of a

Digitized by

Google

80 THE AMERICAN NATURALIST [Vol. XLVIH

single process, one greater, the other less, than as the
presence and absence of one or more unit characters.

Hagedoobn's Wobk

Hagedoorn's work shows the danger of the modem
tendency to produce factors upon the slightest provoca-
tion. While adding, in experimental work, only a single
litter of young bearing on the problem, he gives a symbol
for a factor for dominant spotting in mice, and further
considers it as due to a factor similar to that producing
the dominant ** English" spotting of rabbits. He refers
to Morgan's work with black-eyed white and self mice
as being a study of this dominant factor in mice. Mor-
gan himself suggests that if black-eyed white mice repre-
sent the extremes of the spotted series the appearance
of spotted animals in crosses with selfs is due to a
strengthening of the spotting factor or to a change in
dominance. This is far different from supposing the
addition of an entirely new inhibiting factor comparable
to the English pattern in rabbits. Cuenot with mice and
Castle (1905) with guinea-pigs have shown that black-
eyed whites are the extreme of the recessive spotted
series and it is almost certain that Morgan's explanation
of the results, as due to a change in dominance, is the
correct one. It is, of course, obvious that the presence
and absence hypothesis fails to explain any change of
dominance of a single character.

To treat '^dominant" spotting in mice as due to the
presence of a definite unit-character is exceeding present
experimental facts, while to consider it similar in nature
to the ^* English" spotting of rabbits is still less justified.

One other point in Hagedoorn's work is of such a
nature as to require further experimentation before it
can be accepted.

This is the case (on page 126) of ^^mutual repulsion be-
tween two factors." In this case, Hagedoom mated to-
gether agouti animals heterozygous in factor A (for color
production) and in factor G (for the agouti pattern).

Digitized by

Google

No. 566] SPOTTING IN MICE 8 1

Such animals would ordinarily form gametes AG, Ag, aO
and ag in equal numbers. These by independent recom-
bination would form

1 AAGG

2 AaGG I

4 AaGg .

1 aaGG ^

2 aoGg I
1 aagg J

4 albino.

But Hagedoom gives figures which show that the pro-
portion which he obtains is nearer 2 agouti ; 1 black and
1 albino. This he supposes to be due to the fact that A
and G can never go into the same gamete.

Now let us see what happens if this is the case. The
original heterozygotes will form only two kinds of gam-
etes instead of four, these will be aG and Ag. Now in the
recombination of these gametes the following result will
be obtained.

1 oG aG=:l albino,

2 oG Ag = 2 agouti,
1 AgAg = l black.

So far, so good, but the trouble comes in testing the
albinos. Here I may quote from Hagedoom, p. 126 :

. . . thirteen of these albinos have been tested by mating with black.
Without exception they have given black or equal numbers of black
and albino young. . . . But never has one of those albinos produced
a single agouti young in a mating with black. Counting together the
colored young of such families I get 89 black young?-

This result is indeed remarkable, for on Hagedoom' s
own hypothesis the albinos should have produced in such
matings nothing hut agouti young, '^ since they are all,
by his hypothesis, homozygous for the agouti factor.
The evidence is incontestable ; no repulsion of A and G
can have occurred. Has there been any coupling of these
two factors! If such was the case only gametes AG and

Digitized by

Google

82 THE AMERICAN NATURALIST [Vol. XLVm

ag would have been formed and this would have given
only agoutis and albinos in a 3 :1 ratio, while Hagedoorn
reports **73 agouti, 37 blacks^ and 32 albinos/'

The case then is nothing so simple as ^* repulsion'* or
** coupling,'' it includes failure to segregate and com-
plete disappearance of a dominant Mendelian factor; G
the factor for agouti.

Since numerous investigators of color inheritance in
mice have never found the agouti factor anything but a
normal Mendelizing factor epistatic to black, and since
Hagedoorn himself seems to have become mixed in his in-
terpretation, it seems that the case proves or shows little
until a satisfactory answer can be found to the question
of what has become of the agouti factor.

Conclusions
The facts above given lead to the following conclu-
sions :

1. The so-called dominant type of spotting in mice does
not differ from *^self" color by the presence of a unit
character which ^^self " lacks. The presence and absence
hypothesis fails to account for the shifting dominance
seen in spotting in mice.

2. It is misleading to describe, under the same symbol,
the so-called ^^ dominant" spotting of mice and the Eng-
lish spotting in rabbits.

3. It seems probable that differences in *' dominance"
of spotting in mice are due to modifying supplementary
factors and such spotting might be termed *'unsup-
pressed" and *' suppressed" spotting rather than *' domi-
nant" and *' recessive" in the Mendelian sense.

4. Hagedoorn 's hypothesis of repulsion between the
color factor. A, and the agouti factor, G, is incorrect.

November 19, 1913.
1 Italics mine.

Digitized by

Google

ON DIFFERENTIAL MORTALITY WITH RESPECT

TO SEED WEIGHT OCCURRING IN FIELD

CULTURES OF PISUM SATIVUM

DR. J. ABTHUR HAERIS
Cabnbgib Institution or Washington

In two papers which have already appeared in these
pageSy^ I have shown that for the dwarf varieties of
Phaseolus vulgaris the mortality of apparently perfect
seeds (failure to germinate or to complete the life cyde)
is not random, but differential, or selective.

It seemed highly desirable to extend these studies to
other forms. Pisum sativum naturally occurred to me as
affording suitable experimental material — ^both because
of the wide range of seed characteristics and the conve-
nience with which it may be bred. I had no pedigreed seed
and consequently began work in the spring of 1913 with
commercial stock. About 1,000 seeds from each of ten
early (dwarf) varieties purchased from the Thorbum
seed company were weighed, individually labelled and
planted in short rows scattered over one of the fields of
the Station for Experimental Evolution. Conditions
were not the best, and the mortality was high.

Table I^ gives the weights in units of .025 gram range'

1 Harris, J. Artbnr, "On Differential Mortality with Bespeet to Seed
Weight Oeenrring in Field Colturee of Fhaseolus vulgaris," Amsb. Nat.,
46: 512-525, 1912; < ' Sapplementary Studies on the Differential Mortality
with Bespeet to Seed Weight in the Germination of Garden Beans/' Amsb.
Nat. [in press].

2 For conyenienee the series may be designated bj letters: A, Witham
Wonder; B, American Wonder; C, Premiam Qem; D, Little Gem; E, Nott's
Excelsior; F, Sutton's Excelsior; O, Laztonian; H, little Marvel; /, Peter
Pan ; /, English Wonder.

> Glass 1 = 0.00(K025 gram, . . . class 4 = .07&-.100, dass 5 = .100-
.125, and so on. Thus to obtain means or standard deviations of weights in
grams, deduct .5 from the values in the tables and multiplj bj .025.

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVHI

TABLE I
Weight of Seeds which Geeminated

Series 4

11 12 18 14 15 I 16 17 18 19 20 ToUls

A
B
C
D
E
F
O
H
I
J

65100
40107
63 117 173
36 106 191

76
134

4gi05
9 27

107,126

105
63
27

116
16

142

170

126

167

183

112

105

114

161

19
36
11
18
3
69
159
63
88
13

2 —

11 4
Bl 23

6 2
26101

2 —

468
604
549
606
367
391
631
603
633
665

TABLE n
Weight op Seeds which Failed to Gekminatb

Series

10 11 12 IS 14 16

16 17

19 I 20 ToUls

A
B
C
D
E
F
O
H
I
J

63 102
2

66
37
116
13
6
17
13
66

116

88127

210180

46 84

61 29

93 78

104;117

122 95
93

125

42
84
34
37
23
153

16
36
18
11
4
114

2 1 —
3—2

4l 1

66, 71|110

76 86
93! 29

26 35
125; 91

66162

2 —

7 -

22 3

656
400
461
399
637
613
376
404
374
439

of the seeds which germinated.^ Table II gives the same
distributions for the seeds which failed to germinate.
The physical constants*^ with their probable errors are
given in Tables III-IV.

Taking the differences, germinated less failed, in order
to have the positive sign if elimination tends to increase
mean weight or variability of weight and the negative
sign if it tends to decrease these constants in the popula-
tion of seeds which grow as compared with those which
fail, I find the differences shown in Table V,

4 When the plantlets were about three inches high the labels for seeds
which had failed to germinate were collected. The distributions for the seeds
which had germinated were then obtained bj subtraction from the weight
seriations prepared before planting. Some of the plants subsequently died.

B Sheppard 's correction was applied to the second moments.

Digitized by

Google

No. 566]

DIFFERENTIAL MORTALITY

TABLE in
Physical Constants tob Qxedb whiob Gbionatid

SUndard DotUUod and

Coeffloient of Variation

SariM

Mmui And Probable Error i

Frobabla Error

and Probabla Error

9.2541^.062

2.000*. 044

21.610*. 498

10.581^^.038

1.371*. 027

12.954*. 256

10.078*. 037

1.294*. 026

12.844*. 266

10.355^.033 ;

1.211*. 023

11.692 *.230

9.790*. 042 1

1.193*. 030

12.185*. 308

11.568 *.054 I

1.671 *.038

13.583*. 334

13.090*. 043 1

1.612 *.031

12.313*. 237

11.186 *.037 1

1.362*. 026

12.178*. 240

14.269*. 057

2.134*. 041

14.958*. 290

9.773*. 040

1.429*. 029

14.622*. 300

TABLE rV
Physioal Constants fob Skeds which Failed to Gbionatk

Seriw

Mean and Probable

Standard DoTlation and

Coefficient of Variation

Error

Probable Error

and Probable Error

8.993*. 057

2.003*. 041

22.286*. 472

10.898*. 041

1.236*. 030

11.346 *.274

9.913*. 045

1.439*. 032

14.512*. 829

10.048*. 041

1.229*. 029

12.234*. 296

9.488*. 030

1.122*. 021

11.826*. 227

11.726*. 045

1.653*. 032

14.097*. 277

12.816*. 062

1.787*. 044

13.945*. 350

10.869*. 049

1.447*. 034

13.317*. 322

13.225*. 089

2.552*. 063

19.298*. 493

10.009*. 044

1.376*. 031

13.749*. 311

TABLE V
CoicPABisoN OF Physical Constants fob Skkdb Gkbminatino with those

FOB SXEDS FaILINO TO QlBMINATB

Difference In Mean

Difference In Standard

Difference in Coefficient

Series

and Probable Error of

DoTiation and Probable

of Variation and Probable

Difference

Error of Difference

+ .261*. 085

-.003*. 060

-1.676*. 686

- .316*. 067

+.134*. 040

+1.608*. 375

+ .165*. 068

-.144*.041

-1.669*. 423

+ .307*. 063

-.019*.037

- .642*. 375

+ .302*. 051

+.071*. 036

+ .358*. 382

- .158*. 070

-.082*. 049

- .513*. 434

+ .274*. 075

-.175*. 054

-1.632 *.422

+ .317*. 062

-.085*. 044

-1.139 *.401

+1.044 *.105

-.418*. 074

-4.340*. 572

- .236*.060

+.053*. 042

+ .873*. 432

Consider first the differences in the mean weight.
Seven are positive and three are negative. All of the

Digitized by

Google

86 THE AMERICAN NATURALIST [VoL-XLVHI

seven positive differences are at least 2.5 times their prob-
able error; four of them are over five times their prob-
able error. The mortality is therefore almost certainly
selective, with a tendency to leave the surviving popula-
tion with seeds distinctly heavier on the average than
those which were planted. On the other hand, there are
the three cases in which the seeds which produced plant-
lets were on the average lighter than those which failed
to germinate. One of these differences is only 2.2 times
its probable error, and so perhaps not statistically trust-
worthy. Of the other two, one is over 5.5 times and the
other nearly 4 times its probable error. There can be
little doubt that in at least one of these cases there is a
tendency for the lighter seeds to show a viability greater
than that of the heavier. In garden beans, too, strong
evidences of differences between strains in this regard
have been pointed out.

The interpretation of the variabilities offers greater
diflSculties than does that of the means. More data and
more refined methods of analysis are necessary for a final
solution of the problem. It appears, however, that in
seven of the ten series the variability of the seeds which
survived is less than that of those which failed. This is
true whether absolute variability as measured by the
standard deviation or relative variability as expressed
by the coeflScient of variation be used in the comparison.

As far as these data go, therefore, they are in general
accord with those for Phaseolus. In both of these Legu-
minosae the mortality which occurs before germination is
not random but differential. But in both cases, and espe-
cially in Pisum where the seeds used are of commercial,
not pedigreed, stock and number as yet only about 10,000,
far wider series of experiments and much refinement of
methods of analysis are necessary to establish fully the
nature and the immediate (physical or chemical) cause of
this selective death rate.

Cold Spring Hasbor^ N, Y.,
July 28, 1913

Digitized by

Google

THE INHERITANCE OF A EECURRING SOMATIC

VARIATION IN VARIEGATED EARS

OFMAIZE^

PROFESSOR R. A. EMERSON

UNiyEBsrnr of Nsbraska

Intbodtjotion

The inheritance of variegation has special interest and
importance in genetics. It is with forms of variegation
that the only two certainly known cases of non-Mendelian
inheritance have had to do. I refer to Banr's experiments
with Pelargonium, in which crosses of green-leaved and
white-leaved forms exhibited somatic segregations in Fj
that bred tme in later generations, and to Correns^s work
with Mirabilis, which showed green and white leaf color,
to be inherited through the mother only. De Vries^s con-
ception of *' ever-sporting '* varieties was apparently
fomided largely upon the behavior of variegated flowers
in pedigree cultures, from which he reached the conclusion
that the variegated color pattern and the monochromatic
condition arising from it as sports are non-Mendelian in
inheritance. Correns, however, has shown that in Mira-
hilis jcUapa the inheritance of these sports is distinctly
lifendelian, and the results of East and Hayes indicate the
same for Zea mays. In this paper I shall present data
from maize and attempt to show how they can be inter-
preted in strictly Mendelian terms.

Variegation is distinguished from other color patterns
by its incorrigible irregularity. It is perhaps most often
seen in the coloration of flowers and leaves but also occurs
in fruits, seeds, stems, and even roots of various plants.
It is characteristic of the ears of certain varieties of maize
known, at least in the Middle West, as ^ ^ calico ' * com. In

iThe experimental results reported here were presented at the Cleveland
meeting of the American Societj" of Naturalists, January, 1913. Besearch
l)Qlletin No. 4 of the Nebraska Agricultural Experiment Station.

Digitized by

Google

88 THE AMERICAN NATURALIST [Vol. XLVni

these varieties the pericarp of most of the grains has f ew
to many narrow stripes of dark red, the remaining area
being colorless or showing a sort of washed-out red.
Often broad red stripes appear on some grains, a single
stripe covering from perhaps one tenth to nine tenths of
the grain. Not uncommonly there are entirely colorless
grains (so far as pericarp is concerned) and also solid red
grains scattered over the ear. Much more rarely there
is found a **freak*^ ear with a large patch of self -red or
nearly self-red grains. Or sometimes an ear is composed
largely of red or almost red grains with a small patch of
striped or nearly colorless grains. In such cases it is not
uncommon for the margin of the red area to cut across a
grain so that one side — always the side toward the red
patch — ^is red and the other side colorless or striped. Ears
that are colorless throughout, except for a single striped
grain, are not unknown and there are even known ears
that are red except for a single striped grain. Very rarely
a plant has one self-red ear and one variegated ear on the
same stalk. It is also conceivable that all the ears of a
plant might thus become red, but of course such a red-
eared plant rising as a bud-sport could not ordinarily be
distinguished from a red-eared plant arising as a seed-
sport.

Variegated ears generally have variegated cobs, the
amount of red in the cob ordinarily varying with the
amount of red on the grains. In some **freaks" a part
of the cob is solid red and the rest variegated. In a few
such cases the red part of the cob corresponds exactly in
position to the freak patch of grains. This is more fre-
quently true when the grains of the freak patch are dark
variegated than when they are self-red. In other ears
there is no change in the cob corresponding to the change
in the grains. The husks of variegated ears are also
rather commonly variegated. In a few freak ears the red
side of the ear is enclosed in reddish husks, the remainder
of the husks being light striped. Red-eared plants aris-
ing as seed-sports always have solid red cobs and usually
solid reddish husks.

Digitized by

Google

No. 566] INHERITANCE IN EAB8 OF MAIZE 89

The first account, so far as I am aware, of the inherit-
ance of the striking somatic variations so commonly found
in variegated plants was given by de Vries* in his dis-
cussion of ever-sporting varieties. The study was made
in the years from 1892 to 1896 with a variety of Antir-
rhinum with striped flowers. De Vries^s records are re-
produced diagrammatically in Fig. 1.

Pi Striped

plant

Striped Red

plants plants

90% 10%

-L

( I ( 1

Fs Striped Red Striped R6d

plants plants plants plants

98% 2% 24% 76%

I '" 1

Striped Red

branches branches

i-^-. I ' 1

Fi Striped Red Striped Red

plants plants plants plants

98% 2% 29% 71%

I ^ I ^

F4 Striped Red Striped Red

plants plants plants plants

96% 5% 16% 84%

FIO. 1. DUOBAM FBOK I« VBIBS*8 RSCOKDS SHOWING THB INHBEITANCII OF

Yabibgation AMD Sblf-bbd IN THE Flowbrs ov Antirrhinum.

Of these results de Vries says :

From these figures it is manifest that the red and striped types differ
from one another not only in their visible attributes, but also in the
degree of their heredity. The striped individuals repeat their peculiarity
in 90-98 per cent, of their progeny, 2-10 per cent, sporting into the uni-
form red color. On the other hand, the red individuals are constant in
71-84 per cent, of their offspring, while 16-29 per cent, go over to the
striped type. Or in one word : both types are inherited to a high degree,
but the striped type is more strictly inherited than the red one.

De Vries 's results were in some respects very similar
to those of Correns and it is probable that he would have
interpreted them in the same way had he then been famil-
iar with Mendelian phenomena.

aVries, Hugo de, "Species and Varieties," pp. 309-328 (1905).

Digitized by

Google

THE AMEBIC AN NATURALIST [Vol. XLVm

Correns^ has reported results of a careful study of the
inheritance of the self -green condition appearing as a
bud-sport on variegated-leaved plants of Mirabilis jalapa,
and also of a self-color appearing in striped-flowered
plants of the same species. His results for self-green
and variegation of the leaves are shown diagranunatically
in Fig. 2. The results are stated in approximate per-
centages. I have seen no report in which the detailed
records were given.

Variegated
plant

I —

Variegated
branch

Variegated
plants
100-a*

Vgtd.
100-a
, I

Green

plants

I —

Variegated

plants

Green
branch

>66
I

<33

Green

plants

Green
a

Vgtd.
25

Green Green Vgtd.
76 100 100-a

Green
a

Vgtd. Green >66 <33
branch branch

V G V G V G G

ranch I I I

-i-^

Vgtd.
25

Green Green
75 100

x\ xx\

66 33

VGVGGG VGVGG VGVGGG

Fi 100-a a 25 75 25 75 100 100-a a 25 75 100 100 100-a a 25 75 100 100-a a 25 75 100 100

FIO. 2.

COBBBNS'S DIAOBAK SHOWING THB INHBRITANCB OF VABIBQATION AND

Sblf-gbbbn in thb Lbaves of Mirabilis jalapa.

The diagram shows that a variegated branch of a varie-
gated plant produces in F^ mainly variegated plants, but
occasionally a wholly green plant, while a green branch
from the same plant produces in Fj 25 per cent, varie-
gated and 75 per cent, green plants. The Fj variegated
plants, however produced, behave in later generations
just like the original variegated parent plant. The Fi
green plants, whether produced from green or variegated
branches, are always of two sorts, namely, those that are
homozygous and therefore breed true green, and those

• Correns, C, Ber, Deutsch, Bot. Gesel, 28: 418-434, 1910. Der Uher-
gang aua dem homoeygoiischen in einen heteroeygoiischen Zustand im selben
Individuum hei hunihlditrigen wnd gestreifihlUhenden MirdbUia-Sippen,

* Numerals indicate approximate percentages; a =0-10 per cent.

Digitized by

Google

100

No. 566] INHERITANCE IN EAB8 OF MAIZE 91

that are heterozygous and therefore produce progenies of
green and variegated individuals in a ratio of approxi-
mately 3 to 1. Correns points out that a green branch of
a variegated plant behaves as though it belonged not to a
variegated plant at all, but to a hybrid between a varie-
gated plant and a green one, in which green is dominant,
and that half of the germ cells produced by the green
branch carry a factor for green and the other half a factor
for variegation. Similar results were secured from
branches with self-colored flowers on plants with striped
flowers, except that such branches produce few if any
more self-colored plants than are produced by branches
with striped flowers. Plants with self-colored flowers, no
matter how they arise, behave as they would if they had
occurred in an Fj progeny of a cross of striped by self-
colored plants.

Results of Experiments with Maize
Hartley* in 1902 gave an account of an experiment with
variegated maize. In a comparatively pure white strain,
which occasionally produced a red ear, there was found an
ear similar to some of the *' freak '^ ears noted earlier in
this paper. It is described as being red except for a spot
covering about one fifth of the surface, in which the grains
were white with fine red streaks. The excellent plate ac-
companying the account, however, shows that most of the
**red'* grains had white streaks at the crown and that the
cob was light-colored, not red. From the near-red grains
of this ear there was produced a crop of 84 red ears and
86 pure white ones^ while from the variegated grains of
the same ear there came 39 light variegated ears and 36
white ones. Hartley refers to the parent ear as a ' ' sport
or sudden variation from the type*' but does not indicate
whether the *Hype'' in mind was the white variety or the
red ears occasionally produced by it. Both the color of
the grains and cob and the production of about 50 per
cent, of white ears from both the red and the variegated
grains indicate very clearly that the parent ear was a

* Hartley, C. P., Yearbook, U. S. Bept. Agr., 1902: 543-544.

Digitized by

Google

92 THE AMERICAN NATURALIST [Vol. XLVIII

heterozygous, variegated one and that it probably came

from a white seed crossed by a stray grain of pollen from

a variegated-eared plant, just as the occasional red ears j

in the white variety were certainly produced by stray pol- '

len from red-eared plants.

More recently East and Hayes^ reported like behavior
of a similarly variegated ear. An ear having on one side
solid red grains and on the other white and very light j

variegated grains, similar to some of the ''freak'' ears
noted earlier in this paper furnished the material for the
test. The ear was produced from a white seed in a field
of otherwise pure white corn and was therefore doubtless
heterozygous for pericarp color and was probably pol-
linated in large part from plants without pericarp color,
so that 50 per cent, white-eared plants were to be expected
in its progeny. The white, the light variegated and the
solid red grains were planted separately. The white and
the variegated seeds alike produced light variegated and
white ears, 15 of the former and 15 of the latter. The red
seeds produced 22 white ears and 22 solid red ears. The
authors' interpretation of these results is that the white
seed which gave rise to the original colored ear had been
fertilized by pollen from a red-eared plant and that the
Fi plant, ''due to produce a red ear varied, somatically so
that one half of the ear was red and one half striped."
The authors further state :

This variation was transmitted by seeds, but at the same time the
hybrid character of its seeds was unchanged as shown by their segrega-
tion into reds and whites in the next generation and the normal segre-
gation of the hybrid dark reds in a further generation.

In the light of my own observations, it is equally pos-
sible and seems more likely that the white seed from which
the original red-and-variegated ear came was the result
of pollination from a plant with variegated ears, and that
the somatic variation was from variegated grains to solid
red grains rather than from red to variegated. But the
important fact is that a somatic variation was later in-
herited in a strictly Mendelian way.

5 East, E. M., and Hayes, H. K., Bui. Conn. Agr. Expt. Sta., 167: 106-107.
1911.

Digitized by

Google

No. 566]

INHERITANCE IN EARS OF MAIZE

In 1909 I obtained results somewhat similar to those re-
ported by East and Hayes. A few ''freak" ears were
secured, mainly from local and national com expositions.
Nothing was learned as to their parentage or pollination.
Obviously, however, the parentage of the red, the varie-
gated, and the white grains of any one ear was the same,
and it is reasonable to suppose that the different sorts of
grains of any one ear were pollinated with approximately
the same kind or the same mixture of pollen. The results,
as shown below, were essentially like those of Hartley and
of East and Hayes.

Number of Plants with

Seed! PlftDted

Red Ears

Variegated Eari

White Ears

Self-red

43
0

Variegated and white

The results from four other ears were somewhat differ-
ent, probably owing to differences in their pollination.
(See Fi^. 3.) They were as follows:

Flo. 3. A, •' freak " ear of maize ; B, progeny of striped seeds ; C, progeny of

self-red seeds.

Digitized by

Google

THE AMERICAN NATURALIST [Voi.. XLVHI

Number of Plants with

Seed! Planted

Red Ears

Variegated Ears

White Ears

Self-red

128
8

32
103

Variegated and white

Two other ears of similar history, while they gave quite
as striking results as those noted above, probably do not
belong here since none of their immediate progeny were
variegated and no variegated ears have occurred in later
generations. These two ears were made up of red grains
and white grains only. The results were as follows :

Seeds Planted
Red....

Number of Plants
Red Ears White Ears

rw-

White.

.77
. 0

85
122

The white ears bred true in later generations and the
red ears produced reds and whites in typical Mendelian
fashion. No such somatic variations as these have oc-
curred in my cultures of self-red or white maize, so that I
have been unable to study them further. Somatic varia-
tions in variegated corn, however, are not rare. Unfor-
tunately several of the most pronounced of those occur-
ring in my cultures were open-pollinated and therefore
of little or no use in a careful study. I have therefore
been obliged to make use in large part of the few solid
red and nearly solid red grains scattered over otherwise
more or less evenly variegated ears.

From twenty-three self-pollinated, variegated ears of
plants that were homozygous for pericarp color, grains
with various amounts of red were selected and planted.
The results are summarized as follows :

Namberof Plants with

Seeds Planted

Self-red Ears

Variegated Ears

Non-red Ears

Self-red

8
56

5
33

9
16
34
22
394
22

Nearly self-red

More than half red

Less than half red

0
0

Narrow red stripes

Non-red

Digitized by

Google

No. 566]

INHERITANCE IN EAB8 OF MAIZE

Besides these 23 ears, 20 other selfed ears from homo-
zygous plants contained only narrow-striped seeds from
which there were produced 16 plants with red ears, 280
with variegated ears, and none with white ears. Similarly
21 selfed ears with narrow-striped seeds only, from plants
that were heterozygous for pericarp color, produced 28
plants with red ears, 411 with variegated ears, and 208
with non-red* ears. Variously colored grains from 42
self-pollinated, heterozygous, variegated ears gave the
following results :

r Namber of Planti with

Seeds Planted

Self-red Ears

Variegated Ears

Non-red^ Ears

Sdf-red

15
17
46

8
67

767

Neariy self-red

More than one half red

Narrow red strioes

300

Non-red

In the progenies of these 63 self-pollinated ears that
were heterozygous for pericarp color, there were approxi-
mately 2.5 plants with pericarp color to one without it.
All the classes of grains from self -red to non-red yielded
both colored and non-colored ears, thus indicating, as
already shown by East and Hayes, that the somatic varia-
tion in the seeds does not change their hybrid character.
Considering only the plants with pericarp color, in the
progenies of both heterozygous and homozygous varie-
gated ears, 106 progenies in all, marked differences are
seen in the percentages of self-red ears from seeds of the
different color classes, as follows :

« Some of these ears had what I have termed ^' half -red" pericarp, i. e.,
periearp with a reddish color extending part way from the base to the
crown of the seeds. (See Ann. Bpt. Nebr. Agr. Expt. Sta., 24: 62. 1911.)
Half-red differs from self -red and variegated red not only in distribution
but also in almost never developing f uUy in the heterozygous condition. It
is hypostatic to self -red, but shows between the red stripes of variegated
seeds. Since its pr^ence does not mask either self -red or variegated-red
and since it is strictly allelomorphic to both of them, half -red is here in-
cluded with non-red. Variegated ears have never, in my observation, pro-
duced half -red grains aa somatic variations.

T Some of these were half -red. See footnote 6.)

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVm

Number of PUnU with

Per Cent Self-red
AmongColored

SeeaaFUnted

Self-rMl Ears

Variegated Ears

Self-red

23
73
55
13
134
1

10
24
85
56
1.852
32

69.7

Nearly self-red

75.3

More than one half red

Leas than one half red

Narrow red stripes

39.3

18.8

6.7

Non-red

3.0

In comparison with the cases reported by Hartley and
by East and Hayes and one of my first cultures from
open-pollinated ears, in all of which red grains produced
no variegated ears and striped grains no red ones, the
striking features of the results from these 106 self -pol-
linated ears are the facts that the wholly red grains
yielded some variegated as well as red ears and that the
striped grains and even the wholly non-red grains yielded
some red as well as variegated ears. The percentages noted
above indicate in a general way that for self -pollinated,
variegated ears, the more red there is in the seed planted
the larger the percentage of red ears in the progeny.
These records, however, do not give a wholly trustworthy
indication of the mode of inheritance of the somatic vari-
ations concerned here. If there is a modification of some
factor in the female gametes, associated with a visible
modification of somatic cells of the pericarp and even at
times of the cob and husks, modifications that do not be-
come visible until long after the gametes are formed, may
there not be a similar modification of the same factor in
the male gametes, though here not associated with any
visible change in somatic cells because of the fact that the
staminate inflorescence dies too soon after the pollen is
shed? If male gametes do carry such modified factors
and if the modification is as irregular in occurrence as the
somatic modifications seen in variegated ears, so that any
part of the tassel, from all to none, may produce gametes
with the modified factor while not showing any visible
somatic modification, it is obvious that the real nature of
the male gametes of any variegated-eared maize plant
can not be foretold. The mere fact that a variegated ear
is self-pollinated, therefore, does not insure that its seeds
are fertilized with pollen of known character.

Digitized by

Google

No. 566] INHERITANCE IN EARS OF MAIZE 97

That the male gametes of variegated-eared maize do
often carry factors for self-red is shown by crosses of
pure non-red strains with pollen from plants with varie-
gated ears. The plants that furnished the pollen for
these crosses were in some cases the same ones whose self-
pollinated ears were concerned in the records discussed
above. The results of these crosses are summarized here.
Eight non-red ears crossed by plants that were homozy-
gous for pericarp color yielded 17 red-eared, 116 varie-
gated-eared and 8 white-eared® plants. Similarly, 14 ears
of pure non-red strains crossed by pollen from plants
heterozygous for pericarp color yielded 26 red-eared, 192
variegated-eared and 229 white-eared plants. Consider-
ing merely the plants with colored ears, 22 crossed ears
produced 43 red-eared to 308 variegated-eared plants, or
a little over 12 per cent, self-red.

Since the male gametes of variegated-eared com have
now been shown occasionally to carry a factor for self-
red, it is obvious that only from crosses of variegated-
eared plants with pollen from pure non-colored strains,
can a definite idea of the inheritance of the somatic varia-
tions in pericarp color be gained.* Twelve ears from
homozygous, variegated plants cross-poUinated by non-
red strains might have afforded important evidence, but
for the fact that 7 of them contained only narrow-striped
grains and the other 5 no fully or even nearly self-red
grains. The results are summarized here :

Namberof Plants with

Seeds Planted

Self-red Ean

Variegated Ean

Non-red Ean.

More than one half red

Less than one half red

Narrow red s^pes

5
0
2
0

281

0
0
0

Non-red

< Some of the 8 white ears may have been extreme light tjpes of varie-
gation, for in some other cases very light variegated and wholly white ears
have been observed on the same plant. And of course some of them may
have been due to accidental pollination of the parent ear.

• Though the genetic factors for pigment patterns in maize seem to be
distinct from the factors for the pigment concerned in these patterns, no
non-colored maize that I have used has ever given any indication in crosses
of carrying pattern factors.

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVHI

The principal facts of interest here are the production
of only one red-eared plant to about 140 variegated-eared
ones from narrow-striped seeds, and of about one red-
eared to two variegdted-eared plants from seeds with
from one half to perhaps three fourths red.

Of 20 variegated ears, heterozygous for pericarp color,
that were crossed with pollen from pure non-colored
strains, 5 had only narrow-striped grains and 15 had
variously broad-striped grains and even some self-red
ones. The summaries of these crosses are as follows :

Number of Plmnts with

Seeds Planted

Self-red Ears

Variegated Ears

Non-red Ears

Self-red

9
5
4
3

7
0

0
0
2
5
265

Nearly self-red

More than one half red

Less than one half red

Narrow red stripes

301

Non-red

Here again, just as with homozygous, variegated ears,
the more red there is in the pericarp the more likely are
the female gametes to carry a factor for self -red. While
the number of individuals dealt with are too few to aflford
reliable evidence, it is suggestive to note that the ratio of
red-eared to variegated-eared plants, though not the ratio
of red-eared to total plants, is greater in case of parent
ears that are heterozygous than of those that are homozy-
gous for variegated pericarp.

So far nothing has been said of the results in genera-
tions later than the one grown from the selected seeds
(Fi). Let us now see what results follow when the varie-
gated ears and the red ears produced as explained above
become the parents of second generations (Fg) from the
selected seeds. The variegated ears so produced behave
like the original variegated ears from which seeds were
selected and their progenies have, therefore, been included
in the data already presented. There remains only to
present the records of the progenies of red ears.

Data are available from 7 Fi red ears obtained from
self -pollinated, homozygous, variegated plants. Five of

Digitized by

Google

No. 666]

INHERITANCE IN EARS OF MAIZE

these red ears were self -pollinated and two were crossed
with pure white-eared plants. The results in F, and Fg
were as follows :

KumberofFluitswith

Seeds PlAiiUd from

Self-red
Esn

Ysriecmted
Een

Kon-red
Ears

Fi redfl from selfed, homo., vgtd. Pi's
5 wiTfl jwlf^yl

119
46

9
16

26
40

37
45

2
• 0

0
0

2 ears X white

Fs reda from self ed Ft reds
1 ear i»lf«l

1 ear selfed

Fs rods from Fi rods X white
1 ear selfed

2 ears X white

The above is approximately what would have been ex-
pected, had the F^ red ears that arose from self-polli-
nated, homozygous, variegated-eared plants been pro-
duced by a cross between red-eared and variegated-eared
races.

Of the Fj reds arising from self-pollinated, heterozy-
gous, variegated-eared plants, nine were selfed and two
were crossed with whites. The results secured in Fj and
Fo follow:

Seeds Planted from

Self-red
Eurt

Variegsted
Ears

Non-red
Ears

Fi rods from selfed, hetero., vgtd. Pi's
3 eara selfed (a)

104

106

59
23

23
7
0
0

12
0

1 ear X white

6 earo selfed

1 ear X white

Fs rods from selfed Fi rods of (a)
4 eara selfed

1 ear selfed

From the above it api)ears that the F^ red ears, arising
from self-pollinated, heterozygous, variegated-eared
plants behave in some cases as if they were hybrids be-
tween red-eared and variegated-eared races and in other
cases as if they were hybrids between red-eared and
white-eared races.

Of the four possible sorts of red-eared ** sports*' from
variegated-eared plants, two remain to be treated. Be-

Digitized by

Google

100

TEE AMERICAN NATURALIST [Vol. XLVm

cause of their similar behavior they will be considered
together here. Of the Fj red ears arising from homozy-
gous, variegated-eared plants that had been crossed with
white-eared races, three were self-pollinated and two
crossed with whites. Of the Fi red ears arising from
heterozygous, variegated-eared plants that had been
crossed with white-eared races, four were selfed. The
results in Fg and F3 are :

Seeds Planted from

Self-red
Ears

Variegated
Ears

NoD-red
Ears

Fi reds from vgtd. Pi's X white
Pi's homozygous

3 ears selfed

54
34

102

32
43

0
0

2 ears X whites

Pi's heterozygous
4 ears selfed

Fi reds from selfed Ft reds

3 ears selfed

1 ear selfed

So far as these results go they indicate that F^ reds
arising from crosses between both homozygous and heter-
ozygous, variegated-eared plants and white-eared races
behave as if they were hybrids between red-eared and
white-eared races.

One homozygous, variegated-eared plant was cross-
pollinated by a homozygous red race. From the varie-
gated ear produced, self-red, nearly self-red, and narrow-
striped seeds were planted. All resulted, of course, in
red-eared Fj plants, 16 in all. A self-pollinated Fi red
ear from a narrow-striped seed gave in F2 24 red-eared
and 11 variegated-eared plants — somewhat fewer reds
than were to have been expected. An Fi red ear from a
nearly self-red grain, when cross-poUinated with non-red,
yielded 9 reds and 11 variegated in Fg. A third Fj red-
eared plant, this one from a self-red grain of the varie-
gated parent ear, bred true red in Fg. One ear of this Fj
plant was selfed and yielded 14 reds in Fg, and another
ear was cross-pollinated by non-red and yielded 29 reds.

There are various other somatic variations rather fre-
quently seen in maize, but they are apparently not in-

Digitized by

Google

No. 566] INHERITANCE IN EARS OF MAIZE 101

herited. There are sometimes found variegated ears
with a large patch of self-red cob but with little or no cor-
responding change in the color of the overlying grains.
I have as yet no evidence that this somatic variation in
cob color is inherited through the seeds of the self-red
part of the cob. Such seeds apparently always produce
ears with variegated grains and variegated cobs, just as
do other seeds of the same parent ear. Of course varie-
gated seeds from a self -red patch of cob occasionally give
rise to a self -red ear, as discussed in detail in this paper,
and such red ears always have self -red cobs, but this is
also true of all self -red ears, whether or not they are pro-
duced by red or by variegated seeds and without respect
to whether the part of the cob underlying these seeds is
self-red, finely variegated, or entirely white.

Another form of somatic variation seen in ears of maize
is the occurrence of patches of considerable size, the
grains of which, though variegated, are much darker in
color than the grains of the rest of the ear. Such patches
of grains are often quite as strikingly distinct in appear-
ance as patches of self-red grains, and are apparently
even more likely to correspond exactly in outline with an
underlying patch of self-red cob than are patches of self-
red grains. Moreover, such dark, variegated grains often
present a rather definite color pattern. The crowns are
often made to appear almost solid red by the widening
and convergence at the crown of narrow red stripes ex-
tending down toward the base of the grain particularly on
the side opposite the germ. Another type of dark, varie-
gated grains differs from the lighter, variegated grains
of the same ear principally in the greater development of
the somewhat washed-out red apparently underlying the
dark red stripes of the variegation pattern proper. I
have grown numerous progenies from dark and light
variegated grains of the same ears, but as yet have no
evidence that such somatic variations are inherited. Not-
withstanding this, I have strains of maize breeding true
to a very dark type of variegation, others to a medium

Digitized by

Google

102 THE AMERICAN NATURALIST [Vol. XLVm

sort of variegation, and still others to exceedingly light
types of variegation. There can be no doubt that some of
these different types of variegation are inherited, bnt the
mode of inheritance in crosses has not been fully worked
out.

One other form of grain coloration that might be called
an extremely dark type of variegation is to be noted. The
grains are self-red throughout except for a nearly color-
less crown formed by converging light stripes extending
some way down the side of the grain opposite the germ,
almost exactly the reverse of one of the types of dark
variegation described above. Variegations of this sort
behave in inheritance almost exactly like fully self-red
grains, giving a large percentage of red-eared prog^iy.
And these red ears are apparently always fully self-red,
never showing the pattern of converging light lines seen
in the parent seeds. Many such seeds have been included
in the results recorded earlier in this paper where they
were listed as ** nearly self-red.**

Intebpbbtation of Besults
Anyinterpretation of the data presented here must take
account of these facts: (1) that the more red there is in
the pericarp the more frequently do red ears occur in the
progeny, and (2) that such red ears behave just as if they
were Fj hybrids between red and variegated or red and
white races. The development of red in the pericarp is
evidently associated with and perhaps due to a modifica-
tion of some Mejidelian factor for pericarp color in the
somatic cells. The zygotic formula of a plant homozy-
gous for variegated i)ericarp may be designated as W,
and that of a plant heterozygous for variegated pericarp
as V — . If in any somatic cell VV, from unknown causes,
a V factor were transformed into a factor for self -color,
8, that cell would then have the formula VS. Any peri-
carp cells descended from it would without further modi-
fication be red. If all the pericarp cells of a seed were
thus descei^ded, the seed would be self -red, just as it would

Digitized by

Google

No. 666] INHERITANCE IN EARS OF MAIZE 108

if the plant bearing it were a hybrid between pure red and
vari^ated races. Moreover, one half of the gametes
arising from such somatic cells wonld carry V and one
half would carry 8, just as if the plant were a hybrid of
red and variegated types. Or, if both V factors were
changed, the grains would be self-red as before, but all
instead of half the gametes would carry S. If, however,
the modification from VV to V8 should occur very early
in the life of the plant, or even of the embryo, all the ears
of the plant might thereby become self -red, and one half
of all the gametes both male and female might then carry
8 and the other half F as in the ordinary hybrid. Or the
plant might then become a sectorial chimera with one
variegated ear and one red ear, the gametes from the one
side of the plant all carrying V. If the modification
occur much later, say soon after the ear begins to form,
there might then be merely a solid patch of red grains on
an otherwise variegated ear. In this case only those
gametes arising from these smaller masses of tissue would
carry half 8 and half F. If, however, the modification
occur after the grains begin to form, the latter might be
perhaps three fourths red, or one half red, or merely have
narrow stripes of red, depending upon the amount of peri-
carp directly descended from the modified cell. In this
case it seems reasonable to assume that the larger the
mass of modified tissue the greater the chance that the
gametes concerned should carry 8. Finally, if in certain
grains the change never occurs, they should show no red
and the gametes formed in connection* with them should
all carry F, none 8.

Similarly, it may be assumed that in any cell of a heter-
ozygous, variegated-eared plant, F — , the F factor may
as before become an 8 factor. The effect on pericarp
color would be exactly the same as in a homozygous, vari-
egated plant, and, of the gametes arising from the modi-
fied tissue, one half would carry iS^ as in the other case,
but the other half, instead of carrying F, would carry no
factor and would be represented by — .

Digitized by

Google

104 THE AMERICAN NATURALIST [Vol. XLVIH

If the interpretation suggested here is correct, it is to
be expected that the more red there is in the pericarp of
any seeds, i. e., the larger the mass of tissue descended
from the cell in which the change from V to S took place,
the greater the chance that the female gametes concerned
carried the factor S. With heterozygous, variegated-
eared plants, V — , however, never more than half of the
gametes concerned could carry S even in case of self-red
grains, the other half of the gametes carrying no factor,
— . Of the heterozygous, variegated ears the progenies
of which have been reported here, some were self ed, some
crossed with white, and some open-pollinated. From self-
pollinated ears, self-red and nearly self-red seeds yielded
32 red-eared, 9 variegated-eared, and 14 non-red-eared
plants, or practically 58 per cent, self-red. This excess of
self-red ears may be due, in part at least, to the presence
of the S factor in some of the male gametes concerned, but
the numbers are too small to give very reliable indica-
tions. From similar ears that instead of being selfed
were crossed with white, so that the results could not have
been influenced by factors present in the male gametes,
self-red and nearly self -red seeds produced 14 plants with
red ears and 13 with non-red ears, or about 52 per cent,
red. While these numbers are very small, the fact that
no variegated ears were produced, but that every ear with
any red color was self-red, is noteworthy. From the
open-pollinated, heterozygous ears included in my cul-
tures self-red seeds gave progenies consisting of 171 red-
eared, 32 variegated-eared, and 102 non-red-eared plants,
or about 56 per cent. red.

In case of homozygous, variegated-eared plants, VV, all
the gametes associated with seeds that later become self-
red could carry /S only if both V factors of the somatic cells
from which the gametes arise were changed to S factors.
Because of the rarity of changes from V to S, unless both
V factors are influenced alike by whatever causes the
change, so that both change simultaneously to S factors,
the chance is slight that more than one will ever change.

Digitized by

Google

No. 566] INHERITANCE IN EARS OF MAIZE 106

In the latter case only about 50 per cent, of the gametes
associated with self-red grains of homozygous, varie-
gated ears could be expected to carry S, just as in the
case of heterozygous ears. None of the open-pollinated
ears whose progenies I have grown were homozygous for
variegated pericarp, and none of the homozygous ears
that had been crossed with white contained any self red or
nearly self -red seeds. The only data, therefore, that bear
upon the point at issue are those obtained from self -pol-
linated, homozygous, variegated ears. The self -red and
nearly self-red seeds of such ears produced 64 red-eared
and only 25 variegated-eared plants, or about 72 per cent,
self-red. This may mean that in some cases both V
factors were changed to S factors, but the results may
just as likely be due to the presence of S in an unusually
large percentage of the male gametes concerned. The
production of the 25 variegated-eared plants, however, is
very good evidence that, in at least a very considerable
number of cases, not more than one of the two V factors
could have been changed to S.

If the change from F to iS^ should happen to occur at such
a time that the grain rudiments became sectorial chimeras
consisting of say one half modified cells and one half un-
modified ones, one half of the pericarp would be expected
to show red color and the other half no color. It would
be expected further that the chances of a particular
gamete's arising from a modified or from an unmodified
ceU would be equal. If then one half of the gametes asso-
ciated with these one-half -red grains arise from cells in
which only one of the V factors has been changed to S,
one fourth of the gametes should carry S and three
fourths should carry F, or one f ourth S', one fourth F, and
one half — , depending upon whether the ears concerned
are homozygous or heterozygous for variegated pericarp.
Such grains from homozygous ears should, therefore,
whether selfed or crossed by white, yield about one red
ear to three variegated ones. Similarly, from hetero-
zygous ears, grains with one half their pericarp red should

Digitized by

Google

106 THE AMERICAN NATURALIST [VOL.XLVm

yield about one red to two variegated to one white if self-
pollinated and one red to one variegated to two white if
crossed by white. (This is on the assumption that no S
factors are carried by the male gametes.) Let us assume
that by lumping together all the seeds listed in the fore-
going records as *^more than one half red*' and as *4ess
than one half red'* the whole lot would average about one
half red, and compare the results with the expectation as
noted above. From grains of these two classes from
homozygous ears both selfed and crossed by white, there
resulted 19 red-eared and 82 variegated-eared plants, or
a ratio of about 1:4.3 instead of 1:3. From heterozy-
gous ears self-pollinated grains of these two classes
yielded 54 red-eared, 85 variegated-eared, and 52 white-
eared plants, and similar grains crossed by white yielded
7 red-eared, 7 variegated-eared, and 20 white-eared plants,
or ratios of 1.04:1.63:1 and 1:1:2.86 instead of 1:2:1
and 1:1:2, respectively. The observed ratios are cer-
tainly suggestive but must not be given undue importance,
for there is no assurance that the seeds used really aver-
aged one half red and no assurance that some of the male
gametes in the case of the selfed seeds did not carry S.

We must now examine the results secured in genera-
tions later than F^, and note whether the hypothesis under
consideration applies equally well to them.

It will be recalled that Fj red-eared plants that arose
from homozygous, variegated ears which had been self-
pollinated (see page 99) yielded in Fa only red-eared and
variegated-eared progeny. On our assumption the for-
mula of the parent variegated ears was VV, but the red
grains of these ears were VS and the gametes associated
with them therefore either V or S or all S. Female
gametes carrying S would have produced red. ears in Fj
whether the male gametes carried S or V, and female
gametes with V could not have produced red ears except
when the male gametes uniting with them carried S. The
Fi red-eared plants must therefore have been VS or S8,
the former being expected much more frequently than the

Digitized by

Google

No. 566] INHERITANCE IN EARS OF MAIZE 107

latter, owing to the rarity of S' in male gametes. Only 7
such red ears were tested and all yielded red and varie-
gated ears in typical Mendelian ratios, showing that all
of them were V8 like any Fi hybrid between red and
variegated races. Of two Fj reds from self ed Fi *s, one
again yielded reds and variegates and one apparently
bred true red. Three Fj reds, from Fj reds crossed by
whites, yielded reds and whites only— typical Mendelian
results throughout.

When Fj red-eared plants arose from either homozy-
gous or heterozygous, variegated ears that had been cross-
pollinated by whites they yielded only red-eared and
white-eared, never variegated-eared, offspring (see page
100), just as if they were Fi ears of a cross of reds with
whites. By hypothesis the parent variegated-eared plants
were V — -and VV, and their red grains 8 — and 8V (or
possibly 88). The gametes associated with such grains
were therefore 8 and — , and 8 and V (or possibly all iS).
The male gametes from white races were all — . The Fi

plants were therefore 8 — , V — , and , only those with

8 — having red ears. The five red-eared Fi plants that
were tested produced in Fj red-eared and white-eared
plants in Mendelian ratios. Of the Fg red-eared plants
one bred true in Fj and three again segregated into reds
and whites.

When heterozygous, variegated, parent ears were self-
pollinated, the Fi red-eared plants behaved in some cases
like hybrids of red with variegated races and in other
cases like hybrids of red with white races (see page 99).
Our assumption is that the variegated-eared parent plants
were V — and their red grains 8 — . The gametes asso-
ciated with these red grains were of course 8 and — . The
male gametes of the same plants were doubtless largely
V and — y though a few were probably 8. The Fj plants

must therefore have been , V — , 8 — , 8V or 88. Reds

with 88 would be expected only rarely, and of the 11 Fj
reds tested none had that formula, else they would have
bred true in Fj. Seven of the 11 Fi reds evidently were

Digitized by

Google

108 TBE AMEBIC AN NATURALIST [VouXLVm

8 — , for they yielded Fg progenies consisting of reds and
whites only. Four of the 11 were obviously 8V, for they
yielded Fg's of reds and variegates only. Of the latter
Fg reds, one bred tme in F3 and four again segregated
into reds and variegates.

From a self -red seed of a homozygous, variegated ear
that had been cross-pollinated by a pure red race, an F^
red-eared plant was produced and this plant bred true red
in Fg. From a nearly self-red seed of the same varie-
gated, parent ear, an Fj red was produced but yielded
reds and variegates in Fg just as did a similar Fj ear
from a seed with narrow red stripes (see page 100). The
variegated parent ear was VV and the red and near-red
grains probably VS. The gametes associated with these
grains were V and S. The male gametes were all S.
Therefore the Fg reds were in part V8 and in part 8S.

By way of summary, it is recalled that, in all, 28 Fj red-
eared plants were tested by Fg progenies. Only one of
these bred true and that one came from a red grain of an
ear that had been cross-pollinated by a pure red race.
Disregarding the three Fg red-eared plants thus produced
and the 9 red ears produced from seeds of variegated ears
that had been cross-pollinated by white races and that
therefore could not have bred true, there remain 16 F^
reds, none of which bred true in Fg. Had these Fj red-
eared plants behaved as did the F^ green-leaved plants
produced by green branches of variegated-leaved parents
in Correns's experiments, approximately 5 of the 16
should have bred true. It will be recalled that Correns
found that such green branches always produced green-
leaved and variegated-leaved plants in the ratio of 3:1,
and that one of the three bred true and the other two
again segregated, just as must have happened if the green
branch had been a part of an Fj hybrid of green with
variegated instead of a part of a homozygous variegated
plant.

The difference between Zea and Mirabilis is, however,
not a fundamental one, but is due merely to the circum-

Digitized by

Google

No. 566] INHERITANCE IN EARS OF MAIZE 109

stance that MirabUis has perfect flowers while Zea is
monecious. In MirabUis both male and female gametes
of a green branch arise from somatic cells in which the V
factor has changed to a G factor. If a change in only one
V factor is responsible for the production of the green
branch, the somatic cells of such a branch must all be VG
and the results reported hj Correns are the only ones to
be expected. With Zea mays, however, all the grains of
one ear of a variegated-eared plant might arise from cells
having V8, so that half of the female gametes would carry
S, while little or no corresponding change might take
place in the staminate inflorescence and therefore no (or
very few) male gametes would carry 8. From such an
ear of maize only about one half, instead of three fourths,
of the Fj plants should have red ears and none (or very
few), instead of one third, of the Fj plants should breed
true.

The occasional green plants {^'a" per cent.) arising
from variegated branches in Correns 's experiments with
MirabUis are more nearly comparable to Fj red-eared
maize plants than are the green plants arising from green
branches. It is quite conceivable that on a variegated
branch the male gametes might arise from cells that are
VG, while the female gametes arise from cells that are
VV, or the reverse, though this difference between male
and female gametes would hardly be so common an occur-
rence as with maize where the staminate and pistillate in-
florescences are situated so far apart. It is worthy of
note in this connection that of the occasional green plants
produced by selfed seed of variegated plants in Correns *s
experiments with MirabUis (see diagram, Fig. 2), less
than one third bred true and more than two thirds segre-
gated into green and variegated. (Correns indicates this
merely by the signs < and > in connection with 33 per
cent, and 66 per cent, respectively, in his diagram, and
gives no indication of how much less than 33 per cent,
bred true or how much more than 66 per cent, segregated.)

De Vries's results with Antirrhinum yield readily to

Digitized by

Google

1 1 0 THE AMERICAN NATURALIST [Vol. XLVin

the same analysis used with Zea and Mirabilis. Selfed
seed from striped-flowered branches gave a small per
cent. — ^from 2 to 10 — of red-flowered plants. Only a few
of the red-flowered plants were tested and these were
found to yield 76 per cent, red to 24 per cent, striped.
Selfed seed from red-flowered branches of striped-flow-
ered plants yielded 71 per cent, red-flowered and 29 per
cent, striped-flowered plants, approximating the 75 per
cent, and 25 per cent, indicated by Correns^s results with
Mirabilis. None of these red-flowered plants bred true,
but only one test, and that of only a few plants, was made.
The results were 84 per cent, red-flowered and 16 per cent,
striped-flowered plants. It seems quite likely that had
de Vries tested more red-flowered plants he would have
found some of them to breed true.

Correns ^s results with striped and red flowers of Mirdb-
His differed in one impK)rtant respect from his results
with variegated and green plants of the same species, as
well as from the principal results with Zea reported here
and from de Vries 's results with striped-flowered and red-
flowered forms of Antirrhinum. When red-flowered
plants arose from striped-flowered varieties of Mirabilis,
they behaved just as did the green plants that arose from
variegated forms. But selfed seeds from wholly red-
flowered branches of otherwise striped-flowered plants
yielded little if any larger percentages of red-flowered
plants than did selfed seeds from striped-flowered
branches of the same plants. It would seem that in case
of Mirabilis flowers, when the self pattern arises as a
somatic variation from the variegated pattern there is no
corresponding change in the Mendelian factors for these
patterns. In case of seed-sports from variegated-flow-
ered to red-flowered plants, however, the factors for vari-
egation are affected just as in case of green plants arising
from variegated ones and of red-eared maize plants aris-
ing from variegated-eared ones. The apparently non-
inherited somatic variations of maize plants, noted briefly
earlier in this paper, are possibly of the same nature as

Digitized by

Google

No. 566] INHEBITANCE IN EARS OF MAIZE 111

the somatic variations in variegated flowers of Mirabilis.
Some of these variations in maize are self -red cob patches
on otherwise variegated cobs^ and dark, variegated grains
occurring in patches or scattered over light, variegated
ears.

Gbnbbal Considbrations

The experiments of de Vries, Correns, Hartley, and
East and Hayes, as well as the records reported in this
paper, all indicate that certain somatic variations are in-
herited in strictly Mendelian fashion. All these somatic
variations consist in the appearance of self-colors on
plants that are normally variegated in pattern. The fact
that variegated plants occasionally throw both bud-sports
and seed-sports with self -colors is not, in general, to be
taken as an indication that the variegated plants in ques-
tion are heterozygous. Such behavior seems to be insep-
arably associated with variegation. Coi'rens has pointed
out {loc. dt.) that variegated MirahUis plants can not be
considered mosaics of green and **chlorina** tyi)es due to
heterozygosis, since they do not segregate into chlorina
and green, but into variegated and green. The same rea-
soning applies to variegation in the color of maize ears.
Variegated-eared plants do not throw reds and whites, but
reds and variegates. The conclusion seems irresistible
that self-color occurring as a somatic variation is due to
the change of a Mendelian factor for variegation into a
factor for self -color. If this be granted, the behavior of
these variations in later generations is a mere matter of
simple Mendelian inheritance.

From the title of his paper and the tone of his discus-
sion, it is clear that Correns regards, as the most signifi-
cant feature of these inherited somatic variations, the
change from a homozygous to a heterozygous condition.
He even refers to them as cases of **vegetativen Bastar-
dierung'^ or *'autohybridization.** To me, however, the
essential feature is the change of one Mendelian factor
into another. The fact that this modification of genetic
factors results in a change from homozygosis to heterozy-

Digitized by

Google

112 THE AMERICAN NATURALIST [Vol. XLVIH

gosis seems wholly incidental. It follows from the circum-
stance that usually only one of the two V factors of so-
matic cells is modified. My own data do not in fact show
that the change always affects only one of the factors at a
time. While the results prove that this is true in a part
of the cases at least, the Fj ratios suggest the possibility
of both factors being modified in some cases.

It is of course utterly impossible at the present time to
conceive of the cause or even of the nature of this change
in factors from V to 8. We can only conjecture at pres-
ent as to whether the change may possibly be associated
with changing metabolic processes in the maturing plant,
or perhaps be connected in some way with changing ex-
ternal infiuences, or even be a quality inherent in the V
factor itself. It is perhaps significant that in maize, at
least, the change, whatever its cause, occurs very rarely
early in the life of the plant and apparently becomes in-
creasingly more frequent as the plant matures. Wholly
red ears in variegated-eared plants are extremely rare;
large patches of red grains are somewhat less rare ; indi-
vidual red grains occur on most variegated ears; red
stripes on the individual grains are very frequent, in fact
all but universal in some strains, though in other strains
— ^very light variegated ones — ^there may be only a few
striped grains on a whole ear, the others being wholly
colorless. As a matter of fact, even the presence of an
ear with red pericarp throughout on a variegated-eared
plant may not be good evidence that the change in factors
occurred before the ear began to form. If the change
took place before the ear was laid down, it would seem
that the cob should always be self -red, since the red-eared
progeny of such modified grains of the variegated parent
plant invariably have red cobs, and cob and pericarp
colors are coupled absolutely in later generations. But
red ears, or nearly red ears, with light variegated instead
of red cobs, have been found to occur as somatic variations
on variegated-eared plants. Such behavior suggests that
sometimes the factor change may occur almost simul-

Digitized by

Google

No. 566] INHEBITANCE IN EABS OF MAIZE 113

taneously in the rudiments of every grain so that the
grains become self-red while the cob remains variegated.
We might, of course, account for the appearance of self-
colored grains on a variegated cob on the basis of sepa-
rate factors for cob and pericarp color^® by the assump-
tion that one of these factors may be modified while the
other remains unchanged. But we should then have the
no less difficult problem of accounting for the universal
appearance of red cobs with Fj red ears without respect
to whether the parent grains stood on red or variegated
cobs."

Forced to its logical limit, our conception of the V f ac-,
tor is that of a sort of temporary inhibitor, an inhibitor
that sooner or later loses its power to inhibit color devel-
opment, a power that once lost is ordinarily never re-
gained. Of course it may be that there is present in varie-
gated maize merely a dominant factor for self -color, 8, that
is temporarily inactive, but that sooner or later becomes
permanently active. Even if this be true, S as an active
factor and /S as an inactive factor are certainly as distinct
in inheritance as they are in development and therefore
deserve to be designated separately. And since in one
case there results self -color and in the other variegation,
the factors may as well be called 8 and V as anything else.
It is of course also conceivable that the 8 factor may re-
peatedly arise de novo, though this seems very unlikely.

Whatever our conception of the nature of the factors
for variegation and for self-color in maize ears, these
factors are certainly as distinct in inheritance as any two
factors could well be. Moreover, there is abundant evi-
dence, which can not be given here, that they are strictly
allelomorphic, as indeed they must necessarily be if one
arises by modification of the other — this on the assump-
tion that the factors are definitely localized in certain

10 Evidence that there are distinct factors for cob and pericarp color was
presented in a previous paper on coupling and allelomorphism in maize.
Ann, Bpt. Nebr. Agr. Expt. Sta., 24: 59-90. 1911.

11 This problem is discussed in another paper on the simultaneous modifi-
cation of distinct Mendelian factors. Ameb. Nat., 47: 633-636. 1913.

Digitized by

Google

114 THE AMERICAN NATURALIST [Vol. XLVm

chromosomes. Furthermore, these factors are to be re-
garded as pattern factors. Though they must influence
the development of the pigment in order to produce a pat-
tern at all, they are now known to be distinct in inherit-
ance from the factors for pigment — a fact that I have
been able to show by use of a race of maize with a peculiar
brown pericarp in addition to races with red pericarp.

SUMMABY

A somatic variation in maize is shown to be inherited in
simple Mendelian fashion. The variation has to do with
the development of a dark red pigment (or in one stock
a brown pigment) in the pericarp of the grains, often
associated with the development of an apparently similar
pigment in the cob and husks.

Plants in which this pigment has a variegated pattern
may show any amount of red pericarp, including wholly
self-red ears, large or small patches of self-red grains,
scattered self-red grains, grains with a single stripe of
red covering from perhaps nine tenths to one tenth of the
surface, grains with several prominent stripes and those
with a single minute streak, ears with most of the grains
prominently striped and ears that are non-colored except
for a single partly colored grain, and probably also plants
with wholly self-red and others with wholly colorless ears.

It is shown that the amount of pigment developed in the
pericarp of variegated seeds bears a definite relation to
the development of color in the progeny of such seeds.
This relation is not such that seeds showing say nine
tenths, one half, or one tenth red will produce or even tend
to produce plants whose ears as a whole or whose indi-
vidual grains are, respectively, nine tenths, one half, or
one tenth red. Experimental results indicate rather that
the more color in the pericarp of the seeds planted the
more likely are they to produce plants with wholly self-
red ears, and, correspondingly, the less likely to yield
plants with variegated ears.

Self-red ears thus produced are shown to behave in in-

Digitized by

Google

No. 566] INHEBITANCE IN EABS OF MAIZE 115

heritance just as if they were hybrids between self-red
and variegated races or between self-red and non-red
races, the behavior in any given case depending upon
whether the parent variegated ears were homozygous or
heterozygous for variegated pericarp and whether they
were self-pollinated or crossed with white.

It is suggested that these results may be interpreted by
the assumption that a genetic factor for variegation, V,
is changed to a self -color factor, 8, in a somatic cell. All
pericarp cells directly descended from this modified cell
will, it is assumed, develop color, and of the gametes aris-
ing from such modified cells one half will carry the S
factor and one half the V factor if only one of the two V
factors of the somatic cells is changed, or all such gametes
will carry 8 if both V factors are changed.

The V factor is thought of as a sort of temporary, re-
cessive inhibitor that sooner or later permanently loses
its power to inhibit color development, becoming thereby
an S factor. Or it may be that the dominant factor, 8,
is temporarily inactive, but sooner or later becomes per-
manently active. Again, the 8 factor may repeatedly
arise de novo. The cause of any such change in factors
is beyond intelligent discussion at present.

Tie results of Correns with Mirabilis and of de Vries
with Antirrhinum are shown to be subject to the same
analysis as that used to interpret the results secured with
maize.

Digitized by

Google

^'^•; J

,'1^

Digitized by

Google

EESTOEATION OF EDAPHOSAUEUS CRUCIGEB

COPE

Professoe E. C. case
University of Michigan

In the year 1882 Cope described from the Permian beds
of Texas, an imperfect reptilian skull which he called
Edaphosaurus pogonias. Two years later he described
for the first time, the wonderful vertebra with elongate
spines bearing lateral projections on the sides. These
verteibraB he assigned to the same genus as the skull but
later they were removed to a separate genus as he con-
sidered that the two specimens represented different
forms of reptilian life. The vertebrae with long spines
and cross pieces were placed in the genus Naosaurus —
** Ship-lizard,*' a name suggested by the fancied resem-
blance of the spines with their lateral projections to the
masts and yard-arms of a full-rigged ship.

From the time of the original description until 1907 the
two genera were regarded as distinct but in that year
Case^ suggested that the two genera should be united and
that the skull described as Edaphosaurus by Cope be-
longed with the vertebral column and limb bones de-
scribed under the name Naosaurus. The similar condi-
tion of elongate spines, but without cross pieces, on the
vertebrae of the carnivorous genus Dimetrodon very nat-
urally led to the belief that the two forms Edaphosaurus
and Dimetrodon were similar in other parts of the body
and Naosaurus merely exhibited something of the extrav-
agance in spines, rugosities, tubercles, etc., which is such
a common feature in the most highly specialized members
of any group which is approaching the final stages of its
family or generic life. The close relationship of the two
genera was so probable that it was accepted by all paleon-

1 Publicatioii 55, Carnegie Institutioii of Washington.

117

Digitized by

Google

118 THE AMERICAN NATURALIST [Vol. XLVm

talogists and even Case was very reticent in his sugges-
tion that they were much farther apart than was usually
thought. Following the generally conceived idea of Nao-
saurus a composite mount was prepared in the American
Museum of Natural History in New York in which the
skull and limb bones of a Dimetrodon were associated
with the vertebral column of a Naosaurus. This restora-
tion was published by Dr. Osborn in the Bulletin of the
American Museum and a model of the creature in the
flesh was prepared under his direction by Mr. Chas.
Knight. Case in his ** Revision of the Pelycosauria of
North America" republished this restoration by Osborn
but at the same time published an alternative restoration
in which the skull described as Edaphosaurus was asso-
ciated with the vertebral column of Naosaurus and the
two genera were united under the former name, as it had
priority.

T^he composite restoration prepared at the American
Museum has gained wide circulation in the text books but
later discoveries have shown that it was unfortunate. In
the summer of 1911 Dr. B . v. Huene, of Tiibingen, while
a guest of the joint expedition from the universities of
Chicago and Michigan to the Permo-Carboniferous beds
of New Mexico, discovered the remains of a skeleton of
Edaphosaurus in which both the skull and a portion of
the vertebral column were preserved. As the vertebrae
bore the typical cross-pieces of the genus Naosaurus the
identity of the two genera was established but new evi-
dence was speedily coming; Case in the summer of 1912
discovered in the Permo-Carboniferous beds of Archer
County, Texas, the nearly perfect vertebral column of an
Edaphosaurus (Naosaurus) cruciger Cope with the limb
bones, and a crushed skull, identical with the skull origin-
ally described as Edaphosaurus.

From this skeleton, now preserved in the museum of
the University of Michigan, the author has prepared the
restoration shown in B ig. 1. The only conjectural parts
are the size of the feet and the length of the tail ; the re-

Digitized by

Google

No. 566] EDAPH08AUBUS CBUCIGEB 119

mainder is based upon careful measurements from a
single specimen. So far from being a carnivorous, rap-
torial animal similar to Dimetrodon, Edaphosaurus was
harmless, moUuscivorous or insectivorous with possibly
some ability to masticate vegetable matter. The edges
of the jaws were lined with sharp conical teeth and upon
the palate and the dentary bones were strong plates sup-
porting numerous blunt, conical teeth. The head in all
specimens recovered seems rather small for the size of
the body and in this is peculiar in the Permo-Carbonif-
erous reptilian fauna, in which the reverse is the rule.
The shape of the head in the restoration is taken from the
nearly perfect and undistorted skull in the museum of the
University of Chicago. The elevated dorsal spines begin
with the third vertebrae and speedily reach a considerable
height. The lateral projections are elongate at the base
of the spine but above the middle are reduced to mere
nodules irregularly arranged. The author is not in ac-
cord with the suggestion made by Jaekel and Abel that
the spines were separate, and can see no reason for the
suggestion made by the former that the spines were mov-
able. The strongly interlocking zygapophyses render such
an idea impossible to any one familiar with the skeleton.
Nor does the author believe that the spines were of any
use to the creature as offensive or defensive weapons;
rather, as he has frequently expressed himself, he believes
that they were in the nature of excessive growths which
may have had their inception and impetus in some useful
function, but grew beyond that use as the animal became
more specialized. The union of the spines into a thin
dorsal fin is far more probable and the idea is supported
by the presence of rugosities and the channels of small
nutrient vessels such as would lie beneath a thick dermal
covering. The anterior and posterior faces of the bases
of the spines have sharp, low ridges which give place to
shallow grooves farther up the spine ; only near the top
are the spines similar on all sides. Moreover in the liv-
ing genus Basiliscus, which has elevated dorsal spines.

Digitized by

Google

120 THE AMERICAN NATURALIST [VOL.XLVin

and in the genera of the chameleons in which the same
thing occurs, for example, Chameleo cristatus Stutch., the
spines are united into a thin dorsal crest by the integu-
ment and are further united by a thin membrane carry-
ing scattered muscle fibers. The outline of the dorsal
fin shown in the restoration is suggested by all the speci-
mens in which the spines have been preserved. The sharp
recurvature of the spines in the lumbar region is less
pronounced in the specimen from which the restoration
was drawn than in some other and it is possible that in
other species there was even more of an overhang of the
posterior end. The spines are abruptly shortened in the
pelvic region and rapidly decrease on the tail. The length
of the tail is not known but in all probability was elongate
rather than short and stumpy.

The limbs were short and heavy with the forearm and
foreleg shorter than the proximal segment of the limb, a
condition which is quite common in slow moving forms or
those of aquatic or palustrial habit, and just the reverse
of the condition found in the active, raptorial Dimetro-
don. The bones of the feet have not been found in posi-
tion, but in the great Brier Creek Bone-bed in Archer
County, Texas, excavated by an expedition from the Uni-
versity of Michigan in the summer of 1913, numerous
large foot bones of a character different from those of
Dimetrodon or the cotylosaur Diadectes were found as-
sociated with the spines of Edaphosaurus and with large
claws. It is believed that the foot of that animal was of
goodly size and armed with sharp claws well fitted for
digging in the soft earth or vegetation, tearing open rot-
ten logs and overturning rocks in search of food.

It has been noted by all collectors in the Texas beds that
isolated vertebrae of Edaphosaurus are among the most
common fossils found but that any portion of an asso-
ciated skeleton is extremely rare. This has led to the
suggestion that the remains of the animals were trans-
ported for some distance after death, probably by rivers
from a higher land.

Digitized by

Google

No. 566] EDAPE08AURUS CBUCIGEB 121

Edaphosaurus was a highly specialized creature, slug-
gish in movement and entirely harmless, living upon mol-
luscs, insects and perhaps vegetation. It probably lived
in the woods or near swamps at some distance from the
lowlands upon which were deposited the deltas which
make up the Wichita and Clear Fork formations.

In conclusion the author wishes to express his thanks to
Dr. Euthven, of the University of Michigan, for many
valuable suggestions in arranging the pose and propor-
tions of the restoration, and to Mr. Irwin Christman, of
the American Museum, for the painstaking care with
which his suggestions have been followed in making the
drawing.2

s A full aceonnt of the known tpedmens of Edaphotaurui and NaoiauruM
and a complete Bynonjmj of the two genera will be found in Publications
55 and 181 of the Carnegie Institution of Washington.

Digitized by

Google

SHORTER ARTICLES AND DISCUSSION

HTOIIDITY— A NEGLECTED 'FACTOR IN ENVIRON-
MENTAL WORK

An admittedly rough but probably fair estimate of the relative
interest which has been taken in the relation of the various
environmental factors to insects, at least, may be made from the
fact thart; Bachmetjew in his admirable compilation^ of the work
along these lines devotes, in round numbers, four hundred pages
to temperature, one hudred and fifty to food and chemicals,
seventy to light, forty-five to humidity, fifteen to electricity and
magnetism and thirty to mechanical and other factors. Why is
it that temperature is given about a third more attention than all
the other factors put together? Is it true that it is nearly ten
times as interesting or important as humidity ?

A partial answer to the first question undoubtedly is that tem-
perature is easily controlled as well as measured, whereas humid-
ity, for example, is not easily controlled and the means of
measuring humidity in small containers are untrustworthy and
expensive. Furthermore, work with temperature gives results.
The unfortunate part is that these results have usually been as-
cribed wholly to temperature.

In the course of some work at the Carnegie Station for Ex-
perimental Evolution I found that I could change to a surprising
extent the markings on the larvae of a moth {Isia isabella) by
varying the temperature at which they fed and moulted. How-
ever, such changes were much more definite when the tempera-
ture was kept constant and humidity varied. I did not have the
necessary apparatus for getting accurate control of either factor,
but I feel confident that temperature had little or no direct in-
fluence. It was acting through its influence upon humidity. .

It would seem unnecessary to urge upon experimenters such a
fundamental principle in the logic of cause and effect, but the
fact is that with only two or three exceptions none of the more
than a hundred papers having to do with the effect of tempera-
ture upon insects tell us anything about the effect of temperature

1 * * Experimentelle Entomologische Studien vom phjsikalisch-chemischen
Standpunkt aus. '* Zweiter Band. Sophia, 1907.

122

Digitized by

Google

No. 566] SHORTER ARTICLES AND DISCUSSION 1 23

per se. A few state that the atmosphere was ** moist" or *'dry,"
but even then how moist or how dry is not usually mentioned
unless it is believed to be saturated or absolutely free from moist-
ure. It is clearly incumbent upon the one who makes such a criti-
cism to show, either by his own work or in a review of that of
others, that humidity is a factor of such importance that the
criticism is worth the making — especially since the point is so
self-evident and has been made in the past. The following notes
are an attempt to justify the preceding.

The experiments of many wofkers show that when lepidop-
terous pupae are subjected to abnormal temperature part, at
least, of the adults which emerge differ from the normal. The
observations have usually been made on color changes, and
Fisher^ especially has shown that warm conditions (36*' to 41 '^ C.)
produce the same or similar effects as do cold conditions (O*' to
10^ C), also that hot conditions (42** to 46^ C.) produce effects
wliich are similar to those produced by freezing ( — 20"* to 0** C).
Fisher apparently had no means of successfully controlling the
humidity but Tower* claims to have had this in his ** Investiga-
tion of Evolution in Chrysomelid Beetles of the Genus Leptino-
tarsa" and he obtained similar results, stating them as follows:

The result produced by either a higher or a lower temperature is the
development of a greater amount of pigmentation and a consequent me«
lanic tendency in variations. This stimulus in both directions to increased
pigmentation reaches a maximum between 5® and 7^ C. deviation from
normal. Beyond these, as the temperature further deviates, there is
a rapid fall in melanism, first to the normal, and then to a condition
below normal, until a marked albinic tendency is found; and this de-
crease in pigmentation continues until the zero point is reached, be-
yond which no pigment whatever is produced. The zero point is
reached much sooner, however, in high-temperature experiments than
in low.

Tower then gives the results of experiments in which all the
environmental conditions, except humidity, are '* normal.''
Normal humidity for Leptinotarsa decemlineata is taken as rang-
ing from 43 per cent, to saturation with an average of 74 per cent.
The humidity in various experiments ranges from 10 per cent, to
saturation. The lowest natural humidity of which I have seen
a record is 5 per cent. It occurred in Death Valley, California,

2 866 Archiv fiir Bassen- und GeselUchafts-Biologie, 1907, IV, pp. 761-
793, for Fisher's statement concerning criticisms of bis conclusions.
8 Carnegie Institution of Washington, Publication No. 48, 1906.

Digitized by

Google

124

THE AMEBIC AN NATURALIST [Vol. XLVHI

where the monthly means for May to September inclusive varied
from 20 per cent, to 27 per cent. The annual mean at Cairo,
Egypt, is 56 per cent, and at Ghardaia (Algerian Sahara) is
50 per cent, at 7 a.m. and 26 per cent, at 1 p.m. The humidity
at Buitenzorg, Java, during the height of the rainy season fluc-
tuates between 70 per cent, and 97 per cent, during the day.
Naturally, when dew is being deposited the humidity is practi-
cally 100 per cent. It will be seen then that even Tower's ex-
treme averages (see below) are not beyond the range of
possibility in nature, although they are as great as it is possible
to use in experimental work, since at an average of 34 per cent,
humidity only 0.4 per cent, of the larvae reached the adult stage
and atmosphere can not be kept supersaturated.

The beetles were seriated according to an arbitrary scale in
which '^20 equals total melanism and 0 total albinism." It is
difScult to suggest a better method of measuring the extent of
melanism than this, although we could wish for diagrams to aid
us in grasping just what the scale means. I have tabulated the
experiments and interpolated the normal data.

ReUtlTe Humidity

Per Cent, of
Mortality

Melanism

Arerage

Bange

Mode

Range

100
95
84
74
66
60
50
34

100-100
82-100
55-100
43-100
33-100
30-100
25-83
10-55

90
30
15

80
92
99.6

4
7

12
9

11
5
3
2

2-9

3-11

7-16

6-lS

6-18

3-11

1-7

1-4

It will be seen that mortality increases rapidly as the humidity
departs from normal but this can not account for the change
in color since the range of melanism is doubled and in three of
the experiments even the mode falls below the normal range. As
stated by the author:

The results of experiments with deviations of humidity are almost
exactly the same as those which were obtained from experiments with
deviations of temperature. Such deviations from the normal either to-
ward an increase or a decrease, produce up to a maximum increased
pigmentation and a consequent melanic tendency, but beyond this the
effect is reversed, pigmentation is retarded, and the tendency toward
albinism becomes more and more pronounced as the deviation from the
normal becomes greater.

Digitized by

Google

No. 566] EDAPHOSAUBUS CBUCI6EB 126

The point which concerns the present discussion is that not
only does humidity have a definite regularly acting influence, but
that its results are similar to those of temperature and, as with
temperature, plus and minus variations of certain intensities
bring about similar effects. If, as has usually happened, the hu-
midity is not controlled in experimental work on the effect of
temperature, how can it be said that the observed results are the
effect of changes in temperature t

Tower made certain experiments in which both temperature
and humidity are abnormal, normal average temperature being
taken as 22.2** C. Unfortunately, proof reading or something of
the sort was faulty when it came to publication. Experiment
26 would be the most valuable for our present purpose, but the
table includes records of relative humidity 35 and 39 per cent,
above normal, ». e,, relative humidities of 109 and 113 per cent.,
respectively, if, as in the other experiments, 74 per cent, is
'* normal" humidity. These are clearly impossible. The text
figure illustrating this experiment does not help us since hu-
midities are not given and furthermore the temperatures in the
figure are rather consistently one degree different from those
given in the table. Since there are two errors in text-figure 15,
which illustrates the experiments with humidity as the only vari-
able, it is likely that the figure is the thing that is at fault here.
Several other similar discrepancies could be pointed out (as, for
example, the temperatures in experiment 24, which concerns the
combination effect of humidity and temperature) but it is prob-
able that the author's notebook records are correct and the tem-
perature discrepancies in the published report are so slight that
we may accept his conclusion. It is

that when temperature and moisture are the variables in a given en-
vironmental complex, the trend of general color modification is con-
trolled by moisture (relative humidity), excepting in conditions where
the temperature deviation is so excessive that the ordinary physiological
and developmental processes are greatly inhibited. In experiments
approximating natural environmental complexes, however, moisture is
the dominant factor in influencing coloration.

Even if there were no other reasons for urging the necessity
of taking humidity into account, I feel that Tower's work would
be ample justification. Before taking up those reasons let us
notice several cases where, on account of the striking results of
the experiments, we must regret our lack of information as to the
real cause or the relation of the several causes.

Digitized by

Google

126 THE AMERICAN NATURALIST [Vol. XLVm

This same work of Tower is one of them. The effects just noted
were merely ontogenetic. However, he made other experiments
in which the effect seemed to be passed on by heredity. The fac-
tors in the various experiments with L, decemlineata were 35^,
45 per cent, and low atmosphere pressure (p. 287) ; **hot, dry"
(p. 288) ; **hot, dry and low pressure'' (p. 288) ; and '*hot,
moist'' (p. 291), probably 31.2°, 94 per cent. Those with L.
mulitceniata were 30° and saturation (p. 292 and p. 293) ; and.
the one with L, undecemlineata was * * 10 C. above the average and
a relative humidity of 40 per cent.*' The work is of such im-
portance because of its pioneer character that it would be un-
gracious to complain too strongly, but the fact is that it is
impossible to tell from the data given whether the effects are
caused by humidity or by temperature or by a combination of
the two. Bateson's idea that there are no effects to be explained
need not concern us here.

There is a long series of interesting papers starting in 1895
by Fischer. As has already been mentioned, he finds that certain
high temperature grades produce effects which are similar to
those produced by certain low temperature grades. The con-
ditions of humidity are rarely mentioned, not to say considered.
However, he occasionally confesses that they are important, as
when he tells us* that it is necessary to have the warm air dry and
the cold air moist in order to get similar forms of Vanessa by the
application of moderate cold and moderate heat. I suspect hu-
midity largely enters into the other experiments also for in one
with high temperature,** which gave the same results as certain
low temperatures and presumably high humidity he says the hu-
midity was high.

Like Tower's experiments with beetles these concern color
alone. Pictet* and Federley,'' especially, have considered the
effect of environmental factors upon the form of lepidopterous
scales. Federley calls his work ''Temperatur-experimente" and
Pictet ** Influence de I'Humidite" but neither enables us to dif-
ferentiate the effects of the two factors, although both obtained
striking results. Kominsky® modified to a considerable extent

^Algemeine Zeiischrift fur Eniomologie, VIII, p. 274, 1903.

5 Illustrierte Zeiischrift fiir Eniomologie, IV, p. 134, 1899.

^MSmoires de la SocUiS de Physique et d'Risioire Naturelle de Gen^e,
XXXV, Fasc. 1, 1905.

7 Festschrift fiir Palmen, No. 16, Helsingfors, 1905.

8 Zool, JahrhUcher. Aht. fiir Allg, Zool. und Physiologic, pp. 321-338,
1911.

Digitized by

Google

No. 566] SHORTER ARTICLES AND DISCUSSION • 1 27

not only the color and form of scales but also the form of an-
tennae, legs and other body parts of Lepidoptera. He exposed
the pupae to 42.5** C, humidity not given; 38° to 39° C. and 42*
to 43° C, relative humidity 80; 8° C, "high humidity"; 0° C,
''very high humidity"; —7.5 to 5° C, relative humidity 80-90
and 50; and — 11** C, humidity not given. For the most part
the humidity was high and probably had much to do with the
results, but we can not be certain.

All the experiments just considered were made upon pupae.
It should be remembered that only about one fourth of the weight
of lepidopterous pupae consists of solids, and that the only way
they can replace fluids lost by evaporation is by chemical changes
in these solids. It is probable that they do so to some extent,
although this has not been accurately determined. It is known
that under normal conditions pupae lose in weight and the per-
centage of solids increases. Naturally, a change in the humidity
of the surrounding air would modify this physiological process
and it is difiScult to believe that it has not quite as much effect as
changes in temperature, the humidity remaining the same. It
is easy to see that, if the air is made more absorptive or less ab-
sorptive either by the temperature changes themselves or by other
means, and then the physiological activities are slowed or quick-
ened by temperature changes, the effects will be much greater
and might easily pass as due entirely to the temperature changes.

The species which have wet and dry season forms in regions
where the temperature is fairly constant throughout the year, as
well as the tendency for the animals of moist regions to be mel-
anic and of arid regions to be light colored, speak for the impor-
tant influence of humidity. But there is another point in
distribution to be considered. The study of distribution was
long, and still is, largely an effort to get the ranges of animals
and plants to fit isotherms. When yearly averages do not work,
winter minima or summer maxima or accumulated temperatures
are tried. The success which often attends these efforts shows
that man is very ingenious and also that temperature is really
one of the controlling factors, but it does not show that it is the
only factor or, in fact, that it has any direct influence.

The areas of grassland and forest in North America cut across
isotherms as though they were merely political boundaries but
Transeau* has shown that if we plot the ratio of temperature to
• Ameb. Nat., XXXIX, pp. 875-889, 1905.

Digitized by

Google

1 28 THE AMEBIC AN NATURALIST [Vol. XLVm

humidity we get a very close corresjwndenee between distribution
and climatic factors. Schimper^® has brought together a great
deal of evidence which indicates that, as far as plants are con-
cerned, even the major divisions of the world's surface into arctic,
temperate and tropical are fundamentally a question of the de-
mand for and supply of water.

Furthermore, if recent climatic changes have an effect upon
the origin of new characters and the distribution of the organisms
possessing certain characters, humidity is deserving of more
attention than temperature, since practically the only evidence
we have of such changes concerns humidity.

It should not be forgotten that even aquatic oi^anisms are
subject to what amounts to changes in humidity. Peat bog plants
take on many characteristics of a desert flora, although their
roots are covered with water. It is water, however, which is
not easily available, because of the chemicals which it carries.
It is water which is physiologically dry.

Finally, the great amount of work which has been done upon
artificial parthenogenesis and related subjects is, in a way, a
study of the influence of environmental factors. The obvious
factors concerned have usually been various chemicals but at
foundation humidity, in a broad sense, the addition or withdrawal
of water by osmosis seems to be a factor of prime importance.

Frank E. Lutz
American Museum op Natural History

10 "Plant Geography upon a PhTsiological Basis," translated by W. B.
Fischer. Oxford, 1903.

Digitized by

Google

]

The American Naturalist

MSS. inteBdtfl fbr psblicatioB aid books, eto., lotendod for review should be
sect to the Editor of THE AMERICAN NATURALIST, GarrieoB-on-HudeoD, New York.

Short artleles eootaiDlDg euBoiarlee of research work bearing on the
problems of organio evolution are especially welcouie, and will be given preference
la publication.

One hundred reprints of contributions are supplied to authors fk-ee of charge.
Farther reprints will be supplied at cost.

Subscriptions and advertisements should be sent to the publishers. The
sobscrtptlon price is four dollars a year. Foreign postage is fifty cents and
Canadian postage twenty-five cents addKlonai. The charge for single copies is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS
Lancaster, Pa. Qarrison, N. Y.

NEW YORK : Sub-Station 84

Entered ei seoond-elaM matter, April 2, 1908, at the Poit Office at Lancaiter, Pa., under the Aet of

Congreii of Mareh 8, 1870.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavllc, North Iceland, Via Leidle, Ensrland

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon.
7 or 8 volumes, please report, stating cash price, stat-
ing condition, binding and dates of volumes.

P. C. HARRIS,
Box 2244 Boston, Massachusetts

TO OOLOQISTS

and ENTOnOLOGISTS

W. F. H. ROSENBERG

Importer of Ezotio Zoological Specimens
57, Haverstock Hill, London, N* W., finglaad

Begs to announce the publication of two new
Price Lists : No. 18, Exotic Lepidoptera (over 8000
species), No. 19, Birds' £^ (over 1100 species).

These will be mailed free on application, as well
as any of the following lists:

Birdskins (over 5000 species) ; Mammals (over
aOOspedes); Beptiles, Batraobians and Fishes (over
400 species).

Lai^gest stock in the world of specimens in all
t»«nohes of Zool<^^.

ALL MUSEUMS SHOULD
WRITE FOR THESE LISTS

AJ] specimens sent on approval.
Please state which lists are reqnired and give
name of this periodical.

FINE LAND 5HBLL5

I have in stock and for sale nearly ten thousand spedet
of land shells, also a very good stock of marine. All are ac-
curately labeled. I am prepared to make up collections Il-
lustrating every phase of the subject and from nearly all
countries and Islands. A fine illustrated booklet on the
beautiful shells of the Philippines, and other circulars, will
be mailed free. Strictly scieDtiflc collections purchased for
cash. Correspondence solicited with travelers and explorers.
Only the finest material handled. WALTER P. WEBB,
202 Westminister Road, Rochester, N. T., U. S. A.

The University of Chicago

Offers instnictian durinf the Sum-
mer Quarter on the same basis as
during the other quartets of the
academic year.

The undosnuluate colleges, the
graduate schooU, and the profes-
sional schools provide courses in
I ArtB, Htmvainrm, Scimnem,
Commmrem and AaminiMtra'
Hon, Law, Mmdieinm, Educa*
I Hon» andDivinity. Instruction
I is given by regular members of the
I University staff which b augmented
b the summer by appointment of
professors and instructcn from
I other institutions.

Summer Quarter, 1914
Itt Term June 15-Jnly 22
2d Term July 23- Aug. 28

Detailed announcements wiU be
sent upon application.

The University of Chicago

Mitchell Tower Chicago, IlKnoia

Digitized by

Google

The American Naturalist

Ml ■■tiWhhiJ fai 1867, Dwrotod to iIm Ainmtmmm^ W lb* Pioloiicri
SpMiid iUlMMM to tU FMton W OrcMie ErofartlM ana Hmdlly

ooNTcmra op the Auourr number

Geiwtioftl Stodiet on 0«iioUi«m. IV. Br. Bndlty

HooreDATii,
The Infloftne* of Protneted and Intermittent Fiftlnf

upon Growth. Dr. 8en;iiii Morgnlli,
dmbrien HolothnrUuM. Anttln H. Clerk*
Shorter Artiolet end Diaeosiion: Vlebllity end Ooapr

linginSrosophile, P.W. Whiting. TheBemilte

Obteined'by erosiing wtm meii L. end Kochleene

wexleene Bohred, Kery G. Leoy.

OONTCMTB OF THC MPTCMBEII NUMBKR ^
VheNetnrel Hlftoryof the Nine-bended AxmediUe
of Texea. rrofimg H. H. Kewnum.

Genetieel Stodiee onOenothere.
H. DeTlB.

IV. Dc Bredley

Derwlniimlnl^Qreetrjr. Dr. Beiiheel Zoa.

Noteeone Dtfferentlel Mortelitj ohferred bciiweeB
TenebroebtenriiendT.siolitor. Dr.
Gartner.

CONTCNTS OP THC OCTOBCII NUMBCII

▲ Ck>ntrlbntlon toweidi en AnelTeie ef the Ffoblem
of Inbreeding. Dr. Beymond PeerL

The Inheritenee of Ooet Color in Hotiei. ProltMor
W.& AndeiKMU

The Verietiont In the Nnaberof VertebnB end Ven*
irel Sentte in Two Bneket of the Genoa Begine.
Profenor Alezender G. ButhTen end Czyitel
Thompaon.

Shorter Artiolet end Beporta: The SImiilleBeona
KodUicetion of Diatinet Mendellen Vecton: Pto-
feaaorB.A.£meraon. TheFonithlntemetlooel
Genetic Oonftoenee : Dr. Frenk M. Sorfeee.

OONTKNTB OP THC NOVCMBCll NUMBCK

The Blleet on the Oflipring of IntozleettBg the Mele
Perent end the Tzenamlaaion of the Defeeta to
Bobaegnent Genemtiona. Dr.CherleaB.8toekeEd.

Bnpplementery Stodlea on the DUEerentlel Moitalitgr
with raapect to Seed Weight in theGfreifnetton
of Gerden Beena, Dr. J. Arthur Herxla.

Shorter Artlolea end Diaciiaaion : Bedpiooel Croaeea
between Beeye'a Pheeaent end the CnButmw
Bingneek Pheeaent prodndng Unlike Hybclde.
JohaCPhUUpe.

CONTCNTS OP THC DCCCMBCII NUBIBCII

TheFittttonofCherectertnOrgeniama. ByBdweid
Sinnott.

Inheritenee of Left^hendedneaa. Ptef eiaor TnnoU

Bemeley.
Snpplementery Btndiea on the Diflerentiel MorteU^

with Beapect to Seed Weight in the Germlnetlon

•f Gerden Beena. H. Dr. J. Arthur Herria.
Shorter Artielea end Diacnaalon : A Croaa inyolTing

FonrPeiraofHendelienChereeterainlliee. C

C. Little, J. CPhOllpe.
Index to Volume XLVn.

OONTCNTB OP THC JANUARY NUMBER
A Genetie Anelyaia of the Chengee predneed by

Selection in Xzperimenta with Tobecco. Pro-

feMor B. M. Beat end H. K. Heyea.
Gynendromorphoiia Anta, deecribed daring the I>e>

cede. 1906-1918. Piofeaeor WlUIem Mortoa

Wheeler.

Shorter Artielea end DiaeoMlon: On the Beenlte of
Inbreeding e Mendellen Fopoletlon— A Cocree-
tion end Bxtenaion of PreTlooa Ooneluaioiia.
Dr. Beymond Feerl— laoletfon end Bdeetloii
elUed in Principle. Dr. John T. Gnliek.

Single Number 40 Centc
TIm NATURALIST win U

Yeerir SulMcr^tiom $4.00

gobecrfrMPe lor four moaitfie lor Ona Dolbr

THE SCIENCE PRESS

GurrisoHy N. Y.

SttlKSutioii84s NEW YORK

Lancaster, Pk.

Digitized by

Google

nj.. ZLvm,ao.'M7. ubch, mm

THE

AMERICAN
NATURALIST

A HOSTELS JOmUTAL

Itorotad to fht AdTaaoemeiit of the Biologioftl Seitaeei wltii

Special Beference to the Factors of Xrolution

CaVTSHTS

Fagt
U Tbe BAet of Bsctoat of DtitrlbutiOB on SpoclaUon. Asa C. Chanducr - ISO

n. Bloloffjr of llio Tliysaaoivtora. Dr. A. Fbankliii Sbull - - - Ifl

m. SlMrt^ Artldos sad Oorrospoadoaoo : The Endemic Mammals of the Britiih

lales. Pn>lbiaorT.D.A.CocKBBBLL 177

IV. X^tos and Lltorataro : Swingle on Variation in Fi Citroi Hybrids and the

Theory of Zygotaxis. Br. Orlakd £. Whitb - ISS

THE SCIENCE PRESS

JiAHOABTER, PA. OARKI80N, IT. T.

NEW 70BK : SVB-ST ATION 84

Digitized by

Google

Two Important Biological Journals

The Journal of Qeneticfi

Edited by
W. BATESON and K. C. PUNNETT
A periodical for the pablioation of reoorde of original reeearch in heredity, variation, and allied topior.
The JouBKAL also contains, from time to time, articles summarising the exieting state of knowledge in the
varions branches of genetics. The numbers are fully illustrated, and, when the suligeot-matter demands it,
oolored plates are freely employed. The Journal is issued in parts as material accumulates, and a volu&e
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7.60aTolame;
single copies, |2.50.

3iometrik;a

Edited by EARL PEAB80N
A journal for the statistical study of biol<^gioal problems, appearing about four times a year. A volnme
containing about 500 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
nine preceding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
craniometry, medicine, heredity, etc. Subscription price, $7.60 a volume ; single copies, ^3.50.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes
and single copies.

Published in America for the Cambridge University Press of England by

The University of Chicasro Press
ChicasfO^ Illinois

University Control

By J. MoKbbk Cattell, Profeesor of Psychology in Golombia University

Together with a series of Two Hundred and Ninety-nine Unsigned Letters by I«eading Mso of Bdtmm
holding Aeademio Positions and Articles by JoBKm Jastbow, OaoBon T. LAi>n, Jomr J. BramraoM. S. E.
Cbuobvoh, J. MoKiBN Cattbll, QaoBon M. Stbattoh, Stswabt Patoh, Jomr Jat Csapmah, jAiOi P.
Mfmxon and Jaoob Gould BcmnofAH.

A grsat Tariety of questions eonoeming general unlTersity administration are dealt with in an cricinal sad
helpful way**— ^oivfe.

These quotations and esamples are taken from Professor Cattell's informed and thorough diseussion of thi
suhjeet of uniTersity oontrol, a subjeot upon whieh he hue had much to say of late, finding cNNsaaicm for eanslii
criticism of eristing Ameri an oooditions, and standing as the ehampion of an academic demoeraoy and a tcaoh-
ing profession upon which a man may enter without forfeiting his self-reiveot. — Tk$ DiaL

Sentenoes and paragraphs that betoken the expert* highly-trmined mind* the suggestions that oome to »
fresh a d tell us that a new day is about to dawn in edueatioual writing.^rA# Boston Swemng TromBoripL

SCIENCE AIND EDUCATION

A BerieB of rolumes for the promotion of Boientifto roaearch und eduoaiionml progre$9
VOLUME L The Foundations of Science. By H. PonvoAai. Containing the authorised English transiatioB

by Qaoaan Bavca EULsran of "Scienoe and Hypothesis.'* "The Value of Sdenoe," and "Bcisoss

and Method." Price, $3.00 net.
VOLUME IL Medical Research and Sdocatioii. By RioBAsn M. Pbabcb, Wxulxak H. Wklgh, W. H.

HowxLL, FBAinajor P. Mall, Lbwillts F. Bahctb, CHARLaa 8. Muror, W. B. Canvoh, W. T.

ConroiuiAir. Tkbobald Smiih. O. N. Stswabt, C. M. Jackson. E. P. Lton, jAicaa B. H»«mti7g, Some

M. DoDsoH, C. IL BABomi, W. OPBth^, 8. J. Mbltsbx, Jambs Ewxnq, W. W. K»bn, Hbmbt H. Doxaw

§am, GHUsnAH A. Hmmn, and Hwcbt P. Bownirca. Price, $3.00 net.
VOLUME m. Unirenity Control. HimBmEd^. Paoe$»+48^ Priet. tSM mti.

QARR150N, N. V. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

Digitized by

Google

130 THE AMERICAN NATUBALIST [YouXLYUl

In the pasty mucli of the work that has been done on
zoogeography has dealt with a study of the facts of dis-
tribution, both present and past, as they sta9d, together
with a study of the factors influencing distribution and
speculations regarding the explanation of somo of the
interesting and apparently anomalous facts thus brought
to light. In all of this work, the distribution of animals
has been considered almost entirely as the effect of cer-
tain biological and geological causes. The present paper
is intended to show that the distribution of animals is not
only the effect of other causes, but is in itself the cause
of other effects, and that extent of distribution has a
direct influence on the modification and speciation of the
group concerned.

To find out how far-reaching and how potent is this
effect, much further study is necessary, not only of the
distribution of various groups, but of their classification
and systematic relationships as well.

In brief, the effect of extent of distribution on groups
of different systematic rank may be stated as foUows:
As the range of a group of animals, be it genus, family, or
order, is extended, the species increase out of proportion
to the genera, the genera out of proportion to the families,
and the families out of proportion to the orders. In
other words, if we assume that in a distributional area
of certain extent, there are three genera and six species,
in a distributional area of twice that size, there will not
be six genera and twelve species, but more probably only
four or five genera, and twelve species ; i. e., if in the first
case the index of modification (a term here used to indi-
cate the average number of species per genus) be two, in
the second case it will be greater than two.

As new distributional areas are added, other factors
remaining equal, there is a constant increase in number
of species and subspecies, going hand in hand with a
diminishing rate of increase in genera, the result being a
constantly larger index of modification as the area in-
habited by a group of animals is extended.

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 131

It should be remarked that a unit of area in this con-
nection should be considered a distributional unit, not a
geographical unit. In other words, while the addition of
one hundred square miles might or might not involve a
change in the life of a region, the addition of a new ^^life
zone,'* **fauna,'* or association" (see p. 155) would
inevitably involve a biotic change, and therefore the addi-
tion of one or several of any of these distributional areas
should be considered as an addition of a unit, comparable
to another unit of similar kind.

Two possible ways of testing this hypothesis present
themselves. "We may compare the faunas of distribu-
tional areas of dissimilar size, or we may compare the
specific and generic differentiation found within families
occupying areas of different extent. The former method
we should expect to work out with a fair degree of
accuracy, but the latter involves so many modifying cir-
cumstances that even if sufficient data were at hand, it
would be difficult to prove anything by it. In the first
place there is the difficulty of comparing, in a distribu-
tional sense, the areas occupied by different families,
since, as pointed out above, the geographic areas do not
necessarily coincide at all with distributional areas; in
the second place, while it is justifiable to compare the
speciation of a family in one region with the speciation of
the same family in another region, it is of doubtful value
to compare the speciation of one family with that of
another in the same or different regions, unless the other
factors controlling their speciation be comparable or
nearly so. In view of this there are few families which
could be advantageously compared with each other as to
speciation in relation to extent of distribution, yet in the
families which do seem to lend themselves to such a com-
parison, the evidence aU points towards the correctness
of the law here proposed.

The bats seem as favorable for such an interfamily
comparison as any group of mammals that could be
selected, and the table (Table I) of their distribution by

Digitized by

Google

132

TRE AMERICAN NATURALIST [Vol. XLVHl

TABLE I

Distribution and Spegiation of Families of Ghibopteba

Data Derived from Sdater and Sclater (1899)

Family

Distribution

Gen.

I Index
SP. I of Mod.

Cosmopolitan

Warm parts of both hemispheres .

VespertilionidaB .
Emballonuridffi .

Ptoropodidae Old World,

Rhinolophidae lOld World

Nycteridse Warm parts of Old World

Phyllostomidas ^ Neotropical

17
15
18
6
2
36

190
79

110
61
15
81

11.18
6.27
6.11

10.16
7.60
2.25

families is significant. One family, the VespertilionidaB,
is cosmopolitan, inhabiting every zoologic region and
every life zone, and it has 11.18 species per genus, the
highest of any family of bats. The Phyllostomidae, on the
other hand, has the narrowest range, occupying only the
warm zones of one zoologic region, namely, the neotropic,
and has in 36 genera only 81 species, giving 2.25 as the

TABLE n

DlSTBIBUnON AND SPECIATION OF FAMILIES OF INSEGTIYOBA

Data Derived from Sclater and Sclater (1899)

Family

Distribution

Gen.

8p.

126

21 .

2 !

Index
of Mod.

Soricidffi.

Erinaceidffi .

TalpidsB.
TupaiidsB.

. . . Palearctic, Ethiopian, Oriental and

Nearctic resions, all zones | 11

. . . Palearctic, Ethiopian, and Oriental

regions

. . . Palearctio and Nearctic regions, tem-
perate zones only

. . . Oriental region, warm zones

MaoroscelidaB .Ethiopian region, warm zones

Potamogalidffi I Central Africa and Madagascar,

I tropical zones

Galeopithecidffi {Malay only, forests, tropical zones. . . ,

Chrysochloridas .... !South Africa I

Centetidffi Madagascar ,

Solenodontidas Cuba and Hayti |

11.36

8.00

2.27
7.60
5.66

1.50
2.00
7.00
3.00
2.00

index of modification. The other figures in this table are
significant, but the indices of modification in the families
RhinolophidaB and Nycteridae are abnormally large, and
will probably be reduced by subsequent subdivision of
genera, or discovery of new forms.
Table 11 shows the generic and specific differentiation

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 133

of the various families of Insectivores, but as some of the
families have not been as intensively studied as others,
and as the conditions affecting their distribution and
speciation are so different in different families, we could
hardly expect accurate results, and yet the table clearly
shows a tendency for the families having wider ranges
to have a higher index of modification, the almost cosmo-
politan shrews, for instance', having 11.36 species per
genus, and the families with restricted range (Galeopi-
thecidae, Solenodontidae, Centetidae and Potamogalidae),
having only 1 to 3 species per genus. The Talpidae and
Chrysochloridae do not seem to conform in their speciation
to what should be expected.

When the specific and generic subdivisions of all the
families of mammals have been worked out more per-
fectly, and their ranges in a distributional sense, i. e.,
through life zones, faunas, and associations, are more
accurately known, some interesting facts concerning the
relation between their indices of modification, and the
extent of their ranges, might be brought out.

It is interesting to note that there is a considerable
number of conspicuous examples of wide-ranging genera
which are remarkably poor in species. Among carnivo-
rous mammals there are many such cases, these animals
seeming to be adaptable to an almost unlimited range of
environmental conditions without modification, or, in
other words, their germ plasm, is not stimulated to change
by altered conditions of climate or environment. The
tiger, for instance, is equally at home in the bleak frozen
stepi>es of Siberia, or in the hot humid jungles of India.
The genus Cynaelurus is widely distributed over the
Ethiopian and Oriental regions, and yet it contains but a
single species, with several geographic races. Among
birds there are a number of similar examples, the most
striking case, perhaps, being Pandion, a cosmopolitan
genus with but a single species. The same peculiar condi-
tion occurs among lower animals, as for instance in the
Dinoflagellate genus Diplopsalis, which is cosmopolitan

Digitized by

Google

134

THE AMERICAN NATUBALIST [Vol. XLVHI

TABLE inj

Speciation of Mammals in Various Distbibutional Areas in CSauvobnu

Data from GrinneU {191$ A), (1908), Orinnell and Swarth (191S)

Boreal aod Upper Trmnsltioo Zooee

Group

Sao Jae. Mts.
(SSOSq.M.)

San Bern. Mta.
(MOSq.M.)

Sierra Range
(80,000 8q.ll.)

OoD. ! 8p.
1 1

Gen. 8p.

Gen.

8p.

Ungulata

1 , 1

1
1

1
7
1
1
1
1
1
2

2
3

1
1

4
4

Bovids

1
3

2
2

1
3

3
3

Cervids

10
6

1
1

2(7)

(2)

1(2)

(1)

1
1

2
2

12
6

1
1

2(8)

(2)

1(2)

(1)

(2)

1
1

3
3

Antilocapridffi

Rodentia

Sdurids

Aplodontids

Murids

1
17

Geomyids

Heteromyidtt

24apodida

2
2

Ereihisontids

Oohotonida

Leporids

1
3
4

Camivora

FelidflB

Canida

Muatelidffi

Prooyonidffi

Uraids

1
1

Inaeotivora

SoriddflB

Talpids

Cheiroptera

Phylloetomids

Vespertilionids

Moloeaids

7
7

Total

17(22)

20(26)

45 ; 100

Indioes of modifioa-
tion

1.11

1.17 (1.81)

2.22

in warm and temperate seas^ and yet is composed of not
more than two species. No adequate explanation of these
exceptional cases has been offered, and it is probable that
their speciation, or lack of it, is due to conditions of their
existence or constitution which we do not understand, or
do not recognize.

Digitized by

Google

No. 667] EFFECT OF DISTBIBUTION ON 8PECIATI0N 135

To test the law by comparison of faunas of areas of
different extent, a series of tabular comparisons of the
faunas of various regions of different size and character
was made. In all of these tabulations, care has been taken
in the choice of areas for comparison to make them of un-
equal size from a distributional point of view, and to
make them reasonably comparable. An arctic and a
tropical region, for example, are not considered reason-
ably comparable as regards number of genera and
species, nor is a region on the outskirts of the range of a
group considered comparable with a region near its
center of distribution.

Table m shows a comparison of the mammals of vari-
ous parts of California. The regions compared are as
follows: (A) the boreal and transition zones of (a) the
San Jacinto Mountain range, (b) the San Bernardino
Mountain range, and (c) the entire Sierra range, includ-
ing the Warner and Shasta Mountains to the north, and
the San Bemardinos and San Jacintos to the south; (B)
a comparison of all the zones of (a) the San Jacinto
Mountains with the immediately adjoining country, (&)
the Sierra range as defined above, and including their
foothills, and (c) the entire state.

A careful study of Table III brings out a number of
interesting and significant facts, and bears out the law
here proposed with unexpected accuracy, barring one
seeming exception which, as we shall see later, can not
truly be considered as such.

Let us compare first the three areas in which only the
two uppermost life zones are involved, and from which
the species invading only the lower Transition zone have
also been excluded. First, a word as to the areas com-
pared. The Boreal and Transition zones of the Sierras
take in over one half of all the representation of these
zones within the whole state. These zones of the San
Bernardino and San Jacinto mountain masses are, as
compared with the entire range, very small indeed, and
comprise almost as small areas as could justifiably be

Digitized by

Google

t36

THE AMERICAN NATURALIST [Vol. XLVIH

TABLE niB

(Data as above)
(Data as in Table III A)

Group

Ungulata

BovidsB

Cervida

Antilooaprids

Rodentia

Sduridffi

Castorids.

Aplodontids

Morids

Geomyids

Heteromyids

Zspodids

Erethiiontida

Oohotonids

Leporida

Carnivora

FeGda

Canida

Must^da

Prooyonida

yrsida

InBeotivora

Soridda

Talpida

Cheiroptera

Ph3^o«tomida

VeepertUionida

MoloKida

Total

Indices of modifi'
eation.

All Zoom

San Jm. Mts.

(2,600 Sq.M.)

Gen.

16
6

Sp.

41
7

4
10

Sierra Range
(60,000 Sq.M.)

Geo.

1
10

1
4
1
1
1
2

2
3

7
2
1

Sp.

110

3
9
13
3
1

11
1

California
(158,000 Sq.M.)

Geo.

1
2

4
1
1
1
3

2
3
9
2

4
2

1
8
2

6
17

4
2

14
6

1
21

170

810

1.85

2.03

4.56

considered to be individual f annal nnits. The San Jacin-
tos are somewhat smaller than the San Bemardinos, bnt
the difference is almost inconsiderable when compared
with the Sierras. Before examining the table, let us see

Digitized by

Google

Na 567] EFFECT OF DISTRIBUTION ON SPECIATION 1 37

what conditions in number of genera and species wonld
be expected in these three areas. The San Bemardinos,
being almost as small a fannal unit as should be sepa-
rately considered, we should ezi)ect to approach a mini-
mum index of modification^ i. e., a minimum number of
species per genus, approaching one as a limit. On the
San Jacintos, these being smaller than the San Bemardi-
nos, we shotdd expect fewer types according to the law
suggested by Grinnell and Swarth (1913), that the num-
ber of persistent types in a disconnected area varies
directly with the size of the area. On the entire Sierra
range we should expect, due to the greatly increased
territory, a considerable increase in genera, but a very
much greater increase in species. Looking now at Table
m, we find that with the single exception of the car-
nivores on the San Bernardino Mountains, not one dis-
crepancy exists. The Ungulates, Insectivores and bats
are represented by the same numbers of genera and
species on both of the small areas, and all of them show
a marked increase in genera and species on the larger
area, in every case with an increase in the index of
modification.

The rodents, which show a larger degree of differentia-
tion than any of the other groups, show a very interesting
advance in the index of modification as the area is ex-
tended. The carnivores, as stated above, show a seeming
discrepancy, inasmuch as there are six genera and six
species existing on the San Jacintos, and only two genera
and two species on the San Bemardinos, whereas, if they
conformed with our laws of distribution, we should expect
at least six, and possibly seven or eight, species to be
found there. On page 35 of Grinnell 's ** Biota of the San
Bernardino Mountains'' (1908) we find reference to a
number of carnivores now rare or extinct on the San
Bemardinos, which undoubtedly have been exterminated
by man within the last fifty years. Counting these forms,
which it seems to me we are justified in doing, the table
bears out the law without a single exception, not only for

Digitized by

Google

138

THE AMERICAN NATURALIST [Vol. XLVIII

the total of mammalian forms, but the totals for each
order and for each family.

In comparing the three areas in which all the life zones
are involved, the truth of the effect of extended distribu-
tion on speciation is still more forcibly impressed upon
us. In this case we are comparing areas which are suc-
cessively larger in size, the San Jacintos, with their foot-
hills and low passes involving the fauna of an area of
about 2,500 square miles, the Sierras, about 60,000 square
miles, and the whole state of California about 158,000
square miles. The following table, derived from Table
III, is very significant in showing the diminishing in-

Genera

Species

Index of ModificaUon

Group

SanJac.

Sier.

CaL

8anJao.{ Sier.

Gal.

SanJao.

Sier.

Cal.

Unsulates

Rodents

Carnivores

Insectivores...
Cheiroptera. ..

4
31
17

6
11

2
41
10

110

10
203
51
20
26

1.00
2.56
1.11
1.00

yr5

1.85

2.33
3.93
1.93
2.40
1.71

2,60
6.45
3.00
3.33
2.36

Totals

170

310

2.93 1 4.56

crease of genera, and the constantly increasing addition
of species as the area is enlarged.

By comparing the upper zones of the San Jacintos
with the San Jacintos as a whole, and the upper zones of
the Sierras with the Sierras as a whole (see Table HI),
we find that increasing the life zones has in a lesser
degree the same effect as increasing the geographic area
regardless of zones; in other words, adding life zones
tends to have the same effect on speciation as adding
faunas and associations without life zones. The follow-
ing table (derived from Table HI) illustrates this:

Mammal!

San Jac.
(Upper Zonae)

San Jac.
(All Zonae)

Sierras
(Upper Zones)

Sierras
(All Zones)

Genera

18
20
1.11

34
63

1.85

45
100
2.22

Species

170

Index of mod

2.93

Another rough test of the hypothesis was made in a
comparison of the mammalian faunas of some of our

Digitized by

Google

No. 667] EFFECT OF DISTRIBUTION ON SPECIATION 139

large continental islands and zoologic regions^ the results
being shown in Table IV. The data used in this table are

TABLE IV
Spichation of Mammals in Yarious Continental Islands and Zoolooio

Begions
Daia from Solater and Sclater (1899)

Group

AfHca

AuttnUUn
Region

Austrftli*

N«w

Gnine*

MadaffMOAr
(228,000 Bq.M.)

8p.

Qen.

Sp.

Qen.

8p.

Gen.

8p.

Gen.

Sp.

Gen.

tJngulatfls. . .

Rodents

Caniivores...
InBoctivores..

Bats

Lemurs

Primates

Hyraces

Elephants. ..
Edentates...
Marsupials . .
Monotremes.

156

196

101

41
22

1
1
2

69
83

144
5

8
26

36
3

18
39

36
3

5
16

14
2

1
13

9
20
21
36

1
7
7
9
12
11

Totals

685

128

301

169

100

Index of mod-
ification. . .

5.35

4.12

2.86

2.59

by no means up to date, being taken from the summaries
in Sclater and Sclater (1899), but the subsequent additions
to the faunas of the places concerned, and the splitting
up of genera and species, have probably been approxi-
mately proportionate in each of the five areas, and there-
fore the figures used are sufficiently accurate to be signifi-
cant Comparing Africa, the Australian region, Australia,
New Guinea and Madagascar, which rank in size in the
order given, we find that the indices of modification of
their mammalian faunas are as follows: Africa 5.35,
Australian region 4.12, Australia 2.86, New Guinea
2.59, and Madagascar 2.13. Certainly these figures are
significant.

Comparing the mammalian faunas of the various
islands of the Philippine Archipelago (Table V), we find
that there is even here some corroboration of our law of

Digitized by

Google

140

THE AMERICAN NATURALIST [Vol. XLVIII

TABLE V

Speoation of Mammals in Islands of thb Philippink Abohipelago

Data from Holliaier {1912)

Island

Sq. Miles

Sp.

Gen.

Index of Mod.

Luson ....
Mindanao .
Samar. . . .
Negros. . . .
Panay ....
Palawan. .
Mindoro. .

Leyte

Cebu

Bohol

Masbate. .

40.969
36.292
5,031
4,881
4,611
4.027
3.851
2,722
1,762
1,441
1.236

1.80
1.90
1.23
1.07
1.26
1.16
1.54
1.12
1.14
1.00
1.25

speciation. Considering the large element of chance in
the animal population of a group of islands of such small
size as those of the Philippines, where the various islets
are at a varying distance from each other, and their
faunas have originated from diflFerent sources, the rela-
tion between their size and the differentiation of their
forms is remarkably regular. In Table V, where the main
islands have been listed in order of size, with their num-
bers of genera and species of mammals, the deer have
been excluded entirely, since their generic and specific
differentiation is in too chaotic a state to be used. The
most striking fact brought out by the table is the lead
which the two large islands, Luzon and Mindanao, show,
not only in total number of forms, but in index of modifi-
cation as well. With the possible exception of Mindoro
and Palawan, practically none of the smaller islands is
supporting as large a variety of mammalian forms as
could be expected of it, a fact which might be explained
in a number of ways.

In all of the tabulations given, the marine mammals
have been entirely excluded since the factors affecting
their distribution and speciation are so different from
those of terrestrial mammals. In the majority of cases
marine mammalian families have a paucity both of genera
and species, a circumstance brought about by a number
of factors. Generally speaking, large, wide-ranging

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 141

forms, or forms which are poor in nmnbers of individ-
nalsy are poor in genera and species, possibly due to the
comparative uniformity of their environment, which is
usually coincident. Most marine mammals are of these
kinds, and their paucity of types is emphasized by the
comparative uniformity of their environment, even in the
most widespread groups. From a distributional point of
view, i. e., taking into account life zones, faunas and
associations, a cosmopolitan, oceanic, surface group of
animals does not range through as great a variety of
ecologic niches and environmental and climatic condi-
tions as does a cosmopolitan terrestrial group.

In order to determine whether the principles of distri-
bution and differentiation here set forth would apply to
birds as well as to mammals, a number of series of com-
parisons was made as with mammals, and with exactly
comparable results.

TABLE VI

SpsaATioN OF Birds in Various California Arbas

Data from GrinneU (19133), {1908), WiUett (191g)

Group

8*11 Bernardino Mts.
(2,000 Sq.M.)

Southern Californi*
(«0,000,8q. M.)

California
(158.000 8q.M.)

Gen.

8p.

Gen.

Sp.

Gen.

Sp.

Pluseres

1
1
3
1
2
2
1

19
7

10
2
3
4
5
7
5

114

197

PiearuD

Striges

Aecqiitres

CohimbflB.

<Vlli

Grues

Waden

Anseres

other water birds...

Total

121

163

199

186

347

Index of mod

1.:

1.30

1.87

Table VI gives a comparison of genera and species
of resident birds of (a) the San Bernardino Mountain
region, (&) Southern California, and (c) California as
a whole. Almost without exception, in each individual
group of birds there is a reduction in the index of modi-

Digitized by

Google

142

THE AMERICAN NATURALIST [Vol. XLYIH

fication as the area is restricted from Califomia to the
Pacific Coast region of Southern California, and finally
to the San Bernardino region. The totals reflect the trend
in each gronp. While in the largest area the nnmber of
genera is considerably less than double what it is in the
smallest, the number of species is mpre nearly tripled.
The Southern Califomia area is intermediate.

TABLE vn

Speciation of Besidint Birds in Australll and Tasmania

Data from North {1901-1909)

Group

AostnOia
(2,947,000 Sq.M.)

Tasmania
(26.000 8q.M.)

Sp.

0«n.

Fam.

Sp.

0«]i.

Fam.

Passeres •

804

119

26
6
2
2
3

7
1

11
11

7
1
9
9

PfoaiuD

Striges

Acoipitres

Paittaoi

Total

426

170

Index of cenerio mod

4.36

2.83

Index of speoifio mod.

2.30

1.22

Table VII shows a comparison of the families, genera,
and species of resident birds of Australia and Tasmania,
from North (1901-1909), Here again, in addition to a
very marked diminution of the total number of types in
Tasmania as compared with Australia, each group shows
a considerable decrease in the ratio of genera to families,
namely, from 4.35 in Australia to 2.83 in Tasmania, and
of species, to genera going from 2.30 in Australia to 1.22
in Tasmania.

Table VlJJL is a similar comparison of (a) the resident
birds of Ireland, from Hartert (1912), (6) the resident
birds of all the British Isles, from Hartert (1912), (c) all
the species of the Palaearctic region, the great majority
of which are resident in one part or another, from Dresser
(1902), (d) all the species of Japan, many of which are
not resident, from Ogawa (1908), and (e) all the species
of Kamtschatka, where the majority are resident, from

Digitized by

Google

No. 667] EFFECT OF DISTRIBUTION ON SPECIATION 143

TABLE Vni

I^KSATioN OF Birds in Yabious Paleabgtio Begions

Data from Hartert (191g), Dresser (190S), Oganoa (1908) and

Stejneger (1885)

Group

Ireland

(82,688
8q/M.)

Britiih
IslM

Palearotio

Beffion
(19,160,000
Bq. m!)

Japan

Kam-
taohatka
(106,000
Sq.k)

Sp.

Oen.

Sp.

Gen.

Sp.

Oen.

8p.

Gen.

8p.

Gen.

Paflmran

4
2
7
4
4

10
4
1
8

4
2
4
2
4
9
4
1
8
14

6
12-

8
16

16
30

42
7
4
7
2
7

11
6
1

610
81
34
66
29
76
97
34
31
64

129

116
21
11
21
6
19
32
13
12
24
36

180
34
14
23
12
11
46
27
23
89
83

12
21
28

8
4

16
0
6

26
1
0

PicaruB

Strises

Aodpitres

Cdnmbte

OfJli

T.fmi<^1^ .....

Gmes

Waders

Animrm

Other water birds

Total

124

188

113

1,261

310

491

204

183

112

Index of mod. .........

1.42

l.AA

4.nn

2.41)

l.ftS

Stejneger (1885). The increase in index of modification
from Lreland to the British Isles^ and then to the entire
Palaearctic region, is almost exactly what should be ex-
pected. The greater number of both genera and species
in Japan as compared with Kamtschatka reflects the
greater variety of ecologic niches in a warm country as
compared with a cold one of comparable size. A com-
parison of the resident species of Japan with the resident
si)ecies of the British Isles would be of very great inter-
est, but such a list of Japanese birds is not available. The
very striking similarity between the speciation of birds in
Kamtschatka, and that in the British Isles, both in num-
ber of genera and of species, is very remarkable. The
interesting manner in which the balance of nature is pre-
served is shown by the large representation of raptorial
birds to parallel the abundance of shore birds and Anseres.
That reptiles and amphibians are influenced in their
speciation by tiieir (iistribution is indicated by Table IX,
which shows a comparison of the genera and species of
amphibians, lizards, and snakes, in three of the geo*
graphic areas defined by Cope (1898).

Digitized by

Google

144

TRE AMERICAN NATUBALIST [Vol. XLVUl

TABLE IX

Spegiation of Amphibia and Bsptilia in Nobth Amebicah Absas

Data from Cope (1889), (1898}

Group

Lower Californim District
(12,000 Sq.M.)

Wettern Sab-region
(600,000 Sq. M^)

Medioolnmblan Begiea
(4.600.000 Sq.lE)

8p.

Gen.

Index of
Mod.

8p.

Gen.

Index of
Mod.

Sp.

Gen.

Index of
Mod.

Amphibia

Lacertilis

Ophidia

4
17
16

8
13
12

1^
1.30
1.33

23
28
20

2.30
2.15
2.22

130
143
191

28
31
46

4.64
4.61
4.24

The ** Lower California district'^ consists of only the
tip of Lower California; the ** Western subregion^' em-
braces the Pacific slope of North America from Northern
Mexico, east of the Sierras, to Oregon, where it crosses
the Sierras to the Bocky Mountains, including northern
Idaho, eastern Montana, and most of British Columbia.
The **Medicolumbian region'^ includes northern and cen-
tral Mexico, and most of the United States and Canada
north to a line drawn diagonally from New England to
Alaska, interdigitating on its border with the **Holarctic
region. '*

The almost exactly parallel increase in the indices of
modification in the three groups of cold-blooded verte-
brates considered, as the area is extended, is quite remark-
able. All three groups average from 1.25 to 1.33 species
per genus in the smallest area, from 2.15 to 2.30 in the
intermediate area, and from 4.24 to 4.64 in the largest
area.

As suggested by Professor Kofoid, a factor influencing
speciation in such diverse vertebrates as mammals, birds,
reptiles, and amphibians, should be very widely appli-
cable to speciation in the entire animal kingdom.

A series of statistics relating to various orders of in-
sects and other invertebrates has been compiled to ascer-
tain whether in these groups as well as in vertebrates, the
number of species increases out of proportion to the
genera, as the size of the area, in a distributional sense,
is enlarged.

Digitized by

Google

No. 507] EFFECT OF DI8TBIBUTI0N ON 8PECIAT10N 146

TABLE X

Spsciation of "ELAmiDM IN Yabious Abbas of XJNBQUiii Sub

Data from Sohwan (1906)

RefloD

Sq.MUes

Sp.

6«D.

674

246

438

160

177

126

386

137

Index of Hod.

Africa.

India

Borneo. .
Sumatra.

Java

Oylon. . .

Australia

New Guinea. .
New Zealand.
Tasmania. . . .

11,770,000
228,000

1.760,000

296,700

184.000

60,000

26.338

2.947,000

312,000

104,760

26,000

10.43
6.80

8.26
3.76
4.31
3.37
3.42

9.19
3.06
5.70
1.86

Table X was compUed to show the number of genera
and species of beetles of the family ElateridsB in various
continents and islands, the regions chosen for comparison
being well defined areas of unequal size.

A careful inspection of this table shows that with only-
two exceptions the indices of modification are directly
proportional to the size of the areas. Borneo and New
Guinea, however, not only show a smaller index of modi-
fication than should be expected of them, but are poor in
total number of types. Nevertheless, when we reflect that
these two islands are not nearly so thoroughly known to
science as are the other areas considered in the table, no
great significance can be attached to their seeming paucity
of known types.

Table XI shows tiie number of genera and species of
Limnophilidse, a family of Trichoptera, in eastern North
America (east of the Bockies) as compared with North
America as a whole. It will be noticed that while in the

TABLE XI
SPldATION OF LnCNOPHILIDil (TBICHOPTBU) IN NOETH AlCESIOA

Data from Ulmer (1907)

Bagton

8q. MilM

8p.

0«n.

Index of Mod.

North America

8,000.000
6,000,000

98
46

27
20

3.63

Eastern North Ameriea

2.26

Digitized by

Google

146

TKE AMEBIC AN NATURALIST [Vol. XLVIII

larger area the number of species is more than double
what it is in the smaller area, the increase in genera is
only about one third, increasing the index of modification
from 2.25 to 3.63.

Table XTT shows practically the same thing in the case
of the hawk moths of the family Sphingidae.

TABLE XII

Speciation of Sphinqidji in Amebioan and African Abkas

Data from de Rothschild and Jordan (1907)

Area

Sq. MilM

Gen. Index of Mod.

West Indies

Mexico and Central America

South America

Mex., Cent. Am., and S. Am

Mex., Cent. Am.. S. Am., and W. I.

Bourbon

Madagascar

Africa

Africa and Mad

Africa, Mad., and Bourbon

76,000

976,200

7.000,000

7.975,200

8,051,200

965

228,000

11,772.000

12.000.000

12,000,965

122

197

237

262

166

195

197

3.05
3.58
5.62
5.92
6.39

1.40
1.95
3.45
3.67
3.71

In this case two series of tabulations were made, one
showing the number of genera and species in various
Neotropical areas, and combinations of these areas, the
other showing a similar tabulation for various Ethiopian
areas, with similar combinations. It will be observed that
the speciation in the West Indies is very large for the size
of the area involved, but when we consider the abundant
opportunity that has been given for isolation to operate,
this is not surprising. The index of modification is quite
low. Mexico and Central America have a larger specia-
tion, compared with South America, than would normally
be expected, the reason being that Central America is the
American center of distribution. The index of modifica-
tion, however, reflects the smaller size of the area, being
considerably lower than that for South America. The in-
crease in index of modification from 5.62 to 6.39, as areas
are successively added to South America, is significant.
Looking now at the Ethiopian regions, we find that there
is the same disproportionate increase of species over

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 147

genera in successively larger area6, the index of modifica-
tion increasing from 1.40 in the small island of Bourbon
to 1.95 in Madagascar, and 3.45 in Africa. Combining
Africa and Madagascar, this is increased to 3.67, and with
the island of Bourbon, to 3.71.
Table AIJJL is one of especial interest, since it deals

TABLE XIII

SPECIATION OF MABINE GAMMABIDIA (AMPHIPODA) IN VaBIOUS SBAS

Data from Stebhing (1906)

Aras

8p.

G«ii.

Index of Mod.

Medit4¥TTAneftn 8«a

147
311
476
65
588
645
735
1,333

67
140
176

44
101
207
214
313

2.19

Arotio Ocean ^ .

2.22

N, Atlfrntri« Oc«ftn

2.70

S. Atlantic Ocean

1.47

Arctic and N. Atlantic

3.07

Arctic, N. Atlantic, and S. Atlantic

3.11

Arctic, N. Atlantic S. Atiantio and Med. Sea

Whole family

3.43
4.22

with a marine instead of a terrestrial group. It embodies
the results of a compilation of the marine genera and
species of Amphipoda of the suborder Gammaridea in a
number of the oceans and seas of the world. Since it is
primarily a cold-loving group, the largest numbers are
found in the cold seas, the Arctic and North Atlantic being
the home of considerably over half of the known marine
si)ecies. It is very likely that when the Antarctic regions
have been studied as thoroughly as the northern regions,
the number of species from tiiat part of the world will be
very considerably increased. At the time of Stebbing's
work on Amphii)oda, our knowledge of Antarctic and
contiguous areas was very meager.

The steady increase of the index of modification from
the smaller to the larger seas is striking. The Mediter-
ranean Sea, although it is the most thoroughly known of
all, has the lowest index of modification, namely 2.19, the
Arctic Ocean comes next with 2.22, and then the North
Atlantic with 2.70. The small number of species from the
South Atlantic and Antarctic regions has already been

Digitized by

Google

1 48 TEE AMERICAN NATURALIST [Vol. XLVIII

mentioned, and its low index of modification may be at-
tributed to the same sort of imperfect knowledge as in the
case of Borneo and New Guinea in Table X. The con-
stant growth of the number of species per genus from
2.22 to 3.43 as the various seas and oceans are added to-
gether, exactly parallels the results obtained in a similar
way for a terrestrial group in Table XII. The comparison
of the speciation of the largest area for which it was
worked out, with the speciation of the entire group, many
species and genera of which inhabit fresh water, is inter-
esting, jumping as it does from 3.43 to 4.22. From the
facts brought to light by this table it can hardly be doubted
that practically the same influence is brought to bear on
the speciation of marine as on terrestrial organisms by
the extent of their distribution.

The theoretical explanation here proposed for this phe-
nomenon involves a number of complex problems relating
to evolution and speciation, including isolation, effect of
time, causes of specific and generic modification, etc., each
of which will be dealt with in the following pages as they
seem to influence the law here proposed.

Let us first consider the factor of isolation in relation
to the production of new forms. As excellently stated by
Cook (1909), isolation can not be considered as a cause or
factor in evolution, since changes in the characters of
species are not dependent upon the subdivision of species
to form additional species. To quote from him :

The separation of species into two or more parts allows the parts
to become different, but there is every reason to believe that evolutionary
changes of the same kind would take place if the species were not
divided. That the isolated groups become different, does not indicate
that isolation assists in the process of change. It gives the contrary
indication that changes are restricted by isolation. If isolation did not
confine the new characters to the group in which they arise, the groups
would remain alike, instead of becoming different. . . . Isolation is
the shears that splits the species, not the loom that weaves it.

Therefore, while isolation can not be considered a factor
in evolution, it is an important factor in speciation.
Species vary in many directions or orthogenetically pro-

Digitized by

Google

No. 567] EFFECT OF DISTBIBUTION ON SPECIATION 149

gress in a definite direction, but the trend of variation or
progression may be different in one locality, and tend
towards a different result, from that of another locality.
Whether the evolution, usually in more or less divergent
directions, of segregated groups of individuals be looked
upon (1) as the accumulation of numerous slight varia-
tions which have a different average character in any two
portions of a species, as originally explained by Darwin
(1859, Chap, 4) or (2) purely as the result of natural selec-
tion, as argued by Wallace (1858), or (3) as the result of
a change in the average character of two portions due to
the imeven occurrence of mutations in the two portions,
a conclusion reached by Dewar and Finn (1909, p. 380), or
(4) as the result of orthogenetic evolutionary tendencies
inherent in the species and influenced by the environment,
as Eimer suggested (1897, Chap. 1), does not concern us
here, — ^the general tendency appears to be that two iso-
lated portions of a species as a general rule trend in
different directions, and diverge farther and farther as
long as they are isolated.

It is assumed that the greater the length of time given
for the influence of isolation to be felt, the farther apart
are the two originally identical divisions likely to trend,
however the dissimilar evolution be interpreted. As
stated by Tower (1906), in speaking of the method of
evolution of the CShrysomelid genus Leptinotarsa,

We can interpret the conditions found by any of the current
hypotheses; but explaining a condition by an hypothesis is not the
same as that the conditions found are evidence in support of an
hypothesis, although it is often so used.

The existence of distinct variations, subspecies, and
ultimately species and genera, in isolated areas is a too
frequently observed phenomenon to be looked upon as
anything else than a self-evident truth, but that this should
necessarily be considered as supporting any particular
theory of evolution can not be argued.

The profound results of prolonged isolation may be
observed in the fauna of some of our long-separated con-

Digitized by

Google

150 THE AMERICAN NATURALIST [VouXLVIH

tinental islands, such as Madagascar, Australia and New
Zealand. Decreasing degrees of isolation may be observed
in our West Indian islands, where some generic differenti-
ation has occurred ; in the Santa Barbara islands, where
there has been a differentiation of species; and the de-
tached mountain ranges of Southern California, where the
upper life zones are at present in an isolated condition,
but have been so only long enough to develop a few new
subspecies, and to lose many of the types of the mother
range, in accordance with the law proposed by Grinnell
and Swarth (1913) that **the smaller the disconnected
area of a given zone, or distributional area of any other
rank, the fewer the types which are persistent therein."

From this it is apparent that the time element, in con-
junction with isolation, may have a very decided effect on
the number of genera and species in a family, but since,
from a geologic point of view, animals appear to have
reached a new equilibrium very quickly after a geographic
change, the time element may have little effect on the num-
bers of genera and species relative to each other in any
given area. In other words, as fast as new genera are
produced in a given area, the species within the genera
will tend to be produced in the same ratio, thus leaving
the index of modification unaffected.

As an example of the effect of time and isolation let us
take a hypothetical case. Let us assume that a certain
island became divided into two islands of unequal size,
and that after a short period of segregation, just long
enough for the fauna to readjust itself to the smaller
areas and reach a new equilibrium, we had say six species
in three genera on the larger island, and three of the same
species in two of the genera on the smaller one. After
a long period of isolation we should have approximately
the same number of genera and species on the two islands,
but they would have diverged to generic differentiation.
In other words, the effect of time in conjunction with iso-
lation is to increase the number of genera and species in
the family, while the index of modification undergoes little
change.

Digitized by

Google

No- 567] EFFECT OF DISTRIBUTION ON SPECIATION 151

This leads ns to a consideration of the factors involved
in the differentiation of genera as contrasted with the
differentiation of species. In general it may be said that
extrinsic modifications, i. e., those which are in some way
connected with changes in temperature, humidity, char-
acter of flora, food, and other environmental conditions,
and which usually affect such characters as color, size,
length of hair, etc., lead to differentiation of species and
subspecies primarily. On the other hand, intrinsic modi-
fications, i. e., those which are related directly or indi-
rectly to a change in the habits or mode of life of the
animal or the occupation of a new niche in nature, usually,
if not always, lead to generic or family differentiation,
since it is evident that changes fitting an animal to live
arboreally instead of terrestrially, for instance, are of
such a nature, that if they are perpetuated and carried to
perfection, will not stop at specific difference but will
become of generic importance.

It might be argued that there are no modifications which
might not, if carried far enough, ultimately lead to generic
differentiation. This is possible, but very improbable,
because the modifications here alluded to as ** extrinsic*^
are of such a nature that in the varying climatic condi-
tions there are likely to be intermediate forms which make
the division of the more widely separated ones into genera
impracticable. In the case of our *4ntrinsic*' modifica-
tions, intermediate forms are not so likely to exist when
once the incipient changes leading to an altered mode of
life have reached a fair degree of perfection.

As a concrete example of what is meant by extrinsic
and intrinsic modifications, let us take the squirrels of a
given region, say eastern North America. There are four
genera to be distinguished, — Sciurius, Tamias, Sciuro-
pterus and Arctomys. The genus Sciurus contains
strictly arboreal, mostly nut-eating, omnivorous forms.
Tamias includes forms which are terrestrial, diurnal,
dwelling in natural or artificial holes and crevices, and
with a device for carrying food in their cheeks. Sciuro-

Digitized by

Google

162 THE AMERICAN NATURALIST [Vol. XLVHI

pterus is an arboreal tjrpe which is nocturnal, and has de-
veloped characters which enable it more easily to travel
from tree to tree. Arctomys is the most highly modified
form, and has departed most widely in its habits; it is
entirely terrestrial, seeks shelter in artificial burrows,
eats, grass, and hibernates.

Were we to study the characters separating these gen-
era, we should find that they are all characters which
enable the animal best to occupy the ecologic niche it fills.
If now we select any one of these genera and examine its
species, we perceive that the differences we find are not
such as could clearly be related to differences in mode of
life or habits, but rather such differences as are induced
by the circumstances mentioned above, such differences
being size, color, length of feet and tail, texture of fur,
etc.-i. e., extrinsic variations.

An interesting example of both extrinsic and intrinsic
modii&cations in an incipient stage may be foxmd in the
song-sparrows of western United States. Let us compare
the form of the humid northwest coast belt, Melospiza
melodia morphna, with the form of the arid Arizona des-
erts, M. m. fallaaj. The differences to be observed in color
and size are very noticeable, and would undoubtedly lead
to their separation into two distinct species were it not for
the complete chain of intermediate forms. But even if the
chain of intermediate forms were not complete, and after
a period of segregation the numerous intergrading sub-
species became broken up into a few well-marked species,
nevertheless, unless a change in mode of life of the bird
were involved, however far the extremes of color and size
might tend, they could not be given generic distinction
because of the intermediate forms, inhabiting semi-arid or
semi-humid regions, which would be almost certain to
exist It hapens, however, that Melospiza melodia mor-
phna, and M. m. fallax, do differ considerably in mode of
life, the former being a beach comber, the latter a nomad
of the desert It would be expected, therefore, tiiat if
these two subspecies were isolated, the modifications re-

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 163

lated to their difference in mode of life, already shown
in an incipient manner, woxdd soon lead to their generic
differentiation.

It is not argaed that nnder a given set of ecologic con-
ditions, only one type could be produced, nor that accord-
ing to the idea of some zoologists, as set forth and refuted
by Grinnell and Swarth (1913), should individuals of one
geographic race be transplanted into the region of a dif-
ferent geographic race, the first race would assume within
a few generations all the characters of the second race.
Whether the changes due to the influence of the environ-
ment be looked upon as the results of natural selection
and adaptation, or merely as the results of a stimulus to
the germ plasm, the new type would not necessarily be
always the same, this, however, depending upon the num-
ber of potential responses in the type, and, as excellently
shown by Buthven (1909) in his study of evolution in the
genus Thamnophis, upon the modifications previously
undergone by the type we are dealing with.

It is very evident that there are many variations in
animals which seem to fall into neither the extrinsic nor
intrinsic category, but which are neutral and vary inde-
pendently of climate or habits, and may be inherited phy-
logenetic tendencies. It is very largely due to these
neutral variations, frequently to be ascribed to ortho-
genetic evolution, tending in different directions in dif-
ferent places, and giv^i an opportunity to diverge by iso-
lation, that different species may be produced to occupy
regions of similar climatic and environmental conditions,
and different genera may be found occup3dng the same
ecologic niches.

To choose an example in the same family quoted before,
we may cite the case of Tamias in eastern North America,
and Eutamias in western North America. In this case the
characters separating the genera are not clearly related
to their mode of life, the chief difference being the loss of
one small premolar in Tamias, and its retention in Euta-
mias. The extent of divergence of these neutral varia-

Digitized by

Google

154 THE AMERICAN NATURALIST [Vol. XLVIH

tions depends on the duration of geographic segregation,
and may therefore be of specific, generic, family, or ordi-
nal rank.

To sum up, specific modifications may be of three kinds :
(1) extrinsic modifications, induced by changes of climate
and environmental conditions; (2) neutral modifications,
due to a different trend of evolution in segregated regions ;
(3) incipient generic modifications. On the other hand,
generic modification may be either intrinsic modifications,
concomitant with changes in mode of life or habits of the
animal, or neutral modifications as above, given generic
value by a longer i>eriod of segregation.

Having dwelt for some length on these preliminary con-
siderations, let us now apply them to the case in hand and
see how they affect differentiation into species and genera
through extension of range.

It is a well-known biological fact that different types of
a group of animals, at least of higher animals, are found
associated with different ei^vironments ; nearly related
species do not, as a rule, live comfortably together in the
same environment, and nearly related genera do not
occupy the same ecologic niche in a given zoogeographical
area. This does not seem to hold true for animals of
lower organization, as conclusively shown by Kofoid
(1907). It is common for a group of animals, unless hin-
dered by an impassable barrier or unfavorable environ-
mental conditions, not only to continually extend its range
into new territory, but also to attempt to live in as many
different niches in nature as possible within a given area.
Such attempts to invade new ecologic niches are frequently
concomitant with heritable modifications better fitting
them to occupy their new situation, though it is difficult to
say whether these modifications are causes or results of
the change in mode of life. However this may be looked
upon, the tendency to occupy new niches in nature is fre-
quently accompanied by intrinsic modifications, and there-
fore by generic differentiation.

From this we may safely assume that in a given area

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 155

a family of animals, by adaptive evolution, will approach
a maximum of generic differentiation which can be sup-
ported in that area. In other words, every suitable eco-
logic niche which is represented in the region considered
will be invaded by the family, and even in a small area
there is likely to be a considerable generic differentiation,
especially if isolation has had any opportunity to operate
within the area, in breaking up the genera and species.

Let us assume that in one unit of area a certain family,
Sciuridae for example, was represented by three genera,
each with three species. Second, let us assume that this
family kept spreading into additional units of area. With
each new unit, the chance of new suitable ecologic niches
being represented would decrease, and therefore the
chance of new genera being represented would decrease,
since if a genus were fitted for its niche in nature imder
certain conditions of climate and environment, it would
in the majority of cases not be likely to undergo any
radical changes in the occupation of the same niche
under somewhat altered conditions of climate and environ-
ment; i. e., the stimulus for intrinsic modification would
be lacking.

On the other hand, with each additional unit of area, the
chances of the combined conditions of temperature, hu-
midity, and environment being different, would remain the
same. In other words, the chances of the three dimen-
sions influencing the life of a region, i. e., **life zone**
(controlled by temperature), ** fauna'* (controlled by hu-
midity), and ** association" (controlled by the effect of
the other two plus a number of other environmental con-
ditions), intersecting at the same point would be almost
equally improbable with each succeeding unit of area.
Since it is changes in *'life zone,*' ** fauna,** or ** associa-
tion** which produce extrinsic changes, and therefore lead
to differentiation of species and subspecies primarily, the
increment of species would average nearly the same for
each succeeding unit of area, other factors remaining
equal. It should also be taken into consideration that

Digitized by

Google

166 THE AMERICAN NATURALIST [Vol. XL VIII

with the invasion of new zoogeographic areas, contact
with allied forms is frequently experienced, and oppor-
tunity is thus afforded for cross breeding and hybridiza-
tion, the result of which upon the germ plasm appears to
be as influential in the production of new forms as is the
shock of new environmental conditions. The constant
increase in species and subspecies accompanying invasion
of new territory, going hand in hand with a diminishing
increase in genera, results in the constantly larger index
of modification as the area inhabited by a group is
extended.

SUMMABY

1. Extent of distribution has a direct influence on the
speciation of the group concerned in this way, that as the
range of a group of animals is extended, the species in-
crease out of proportion to the genera, the genera out of
proportion to the families, and the families out of pro-
portion to the orders.

2. Comparisohof different families having unequal geo-
graphic ranges is usually inaccurate due to the great dif-
ferences in the other factors controlling their speciation.
Those families which do lend themselves to such a com-
parison show decidedly the effect of extent of distribu-
tion, e. g., the bats and some of the insectivores, the fami-
lies of widest distribution having the largest indices of
modification. A number of exceptions exist in the form of
certain wide ranging genera which have a paucity of
species. We have no adequate explanation for this
phenomenon.

3. Comparison of the faunas of areas of different size
gives very accurate results. A number of tabulations show
as a whole an invariable increase in the index of modifica-
tion as the distributional area is extended by the addition
of either life zones, faunas, or associations. Such tabu-
lar comparisons were made for all the classes of ter-
restrial vertebrates, for several families of insects, and
for the marine Amphipoda of the suborder Gammaridea.
Allowing for explicable exceptions, the increase in number

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 167

of lower systematic groups out of proportion to the in-
crease of higher systematic groups as the area considered
is enlarged is a remarkably constant and wide-spread
phenomenon.

4. The theoretical explanation here proposed for this
phenomenon involves a number of complex problems
relating to evolution and speciation, including isolation^
the time element, and causes of specific and generic
modification.

5. Isolation is an important factor in speciation, since
the separation of species into two or more parts allows
the parts to become different The degree of divergence
of the s^regated parts is largely dependent upon the
duration of segregation.

6. Time, in conjunction with isolation and evolution,
tends to increase the number of genera and species in a
family, but the index of modification, t. e., the average
number of species per genus, remains approximately the
same in a given area.

7. Three types of modifications in animals may be
named:— first, ** extrinsic'* modifications, which are in-
duced by climate and other environmental conditions, and
which lead to differentiation of species and subspecies
primarily; second, ** intrinsic'' modifications, which are
concomitant with a change in habits or mode of life of the
animal, due to the occupation of a new ecologic niche, and
which usually lead to generic or family differentiation;
and third, neutral. modifications, which are merely the
result of the natural tendency of all animals to vary and
to be subject to more or less orthogenetic evolution,—
modifications which can not be correlated with environ-
mental conditions, nor with a change in mode of life of
the animal, but which may be influenced largely by in-
herited tendencies. Such modifications are responsible for
the production, through isolation, of different species to
live under the same climatic and environmental conditions,
and of different genera to occupy the same ecologic niche.

8. Specific modifications may be of three kinds: (1) ex-

Digitized by

Google

168 THE AMERICAN NATURALIST [V0L.XLVIII

trinsic modifications, (2) neutral variations in segregated
regions, ( 3 ) incipient generic modifications. Generic modi-
fications may be (1) intrinsic modifications, or (2) neutral
varations, given generic value by a longer i)eriod of
segregation.

9. Since different types of a group of animals are
usually found associated with different environmental con-
ditions or different ecologic niches, and since it is common
for animals, if unhindered, not only to extend their range
continually into new territory, but also to occupy new
ecologic niches, and since these tendencies lead to specific
and generic differentiations, respectively, any given area
will have a differentiation of species proportionate to its
variety of environmental conditions, and of genera pro-
portionate to its variety of suitable ecologic niches.

10. Since, as the area of distribution is extended, the
chance of new conditions of climate and environment being
represented remains approximately the same, the increase
in number of species is nearly proportional to the increase
in the area of distribution, but since the chance of new
ecologic niches being represented in most cases constantly
decreases, the increase in genera proceeds at an ever-
diminishing rate. This, going hand in hand with the
nearly constant increase in species or subspecies, results
in a constantly increasing index of modification.

LITERATURE CITED
Cook, O. F.

1908. Evolution without Isolation. Am. Nat., 4g, 727-731.
Cope, E. D.

1889. The Batrachia of North America. Smithsonian Inst. Nation. Mus.

Bull., S4, 1-525, 86 pis., 119 figs, in text.
1898. The Crocodilians, lazards and Snakes of North America. Smith-
sonian Inst. Nation. Mus. Rep., 18^, 155-1270, 36 pis., 347 figs,
in text.
Darwin, Ch.

1859. The Origin of Species, 2d ed., 1869. New York, D. Appleton & Co.,
xi, 440, 1 pi.
Dewar, D., and Finn, F.

1909. The Making of Species. London, John Lane Co., xix, 400, 15 pis.

Digitized by

Google

No. 567] EFFECT OF DISTRIBUTION ON SPECIATION 169

Dreflser^ H. E.

1902. A Mannal of Palearetie Birds. London^ Vols. 1 and 2, 922, 2 pis.
Eimer, Th. G. H.

1897. Die £^nt8tehllllg der Aiten anf Orund von Yererbung erworbung
Eigenschaften naeh den Gesetzen organisehen Waehsens, II.
Leipzig, W. Engelmann, xvi, 513, 235 figs, in text.
Orinnen, J.

1908. The Biota of the San Bernardino Mountains. Uniy. Calif. PabL

Zool., 5, 1-170, pis. 1-24.
191ZA. A Distribational List of the Mammals of California. Proc, Calif.

Acad, Set,, 4tli series, S, 265-390, pis. 15-16.
1913B. A Distribational List of the Birds ot California. Mss.
Grinnell, J.^ and Swarth, H. S.

1913. An Aeconni of the Birds and Mammals of the San Jacinto Area
of Southern California, with Bemarks upon the Behavior of
Geographic Baces on the Margins of their Habitats. Vjoy,
Calif. Publ. Zool., 10, 197-, pis. 6-10, figs. 1-3 (in press).
Hartert, E., Jourdain, F. C. B., Ticehurst, N. F., and Witherby, H. F.

1912. A Handlist of British Birds. London, Witherby & Co., XII, 237.
HolHster, N.

1912. A List of the Mammals of the Philippine Islands, EzclusiTe of the
Cetacea. Phihpp. J. Set., D, 7, 1-64.
Kofoid, C. A.

1901. The Limitations of Isolation in the Origin of Species. Science,
N. S., es, 500-606.
North, A, J.

1901-1909. Nests and Eggs of Birds Found Breeding in Australia and
Tasmania. Austr. Musi, Sydney, Sp. Catalogue I, Vol. 1, vii,
366, pis. A 1-8, B 1-7, text figs.; Vol. 2, vii, 380, pis. A 9-13,
B 8-13.
Ogawa, M.

1908. A Handlist of the Birds of Japan. Annot. Zool. Jap., Tokyo, (>,

337-420.
de Bothschild, W., and Jordan, K.
1907, Lepidoptera, F^m. Sphingids. Genera Ineectorum (ed. by Wyts-

man. P.), 57, 157, 8 col. pis.
Buthven, A. G.

1909. A Contribution to the Theory of Orthogenesis. Aic. Nat., 43,

401-409.
Bchwarz, O.

1906. Coleoptera, Fam. Elateridee. Genera Insectorum (ed. by Wytsman,
P.), 46, 370, 6 col. pis.
Sclater, W. L., and Sclater, P. L.
1899. The Geography of Mammals. London, Eegan Paul, Trench,
Trubner & Co., xviii, 335, 8 maps, 51 figs, in text.
Stebbing, T. B. B.
1906. Amphipoda. I. Gammaridea. Das Tierreich (ed. by F. E.
Schulze). Berlin, B. FriedlAnders Son, 21 xxzix, 806, numerous
figs, in text.

Digitized by

Google

160 THE AMERICAN NATURALIST [Vol. XLVni

Stpjneger, L.
1885. Results of Ornithological Explorations in the Commander Islands

and in Kamtschatka. Smithsonian Inst. Nation. Mus. Bull., t9,

382, 8 pis.
Tower, W. L.

1906. An Investigation of Erolution in Chrysomelid Beetles of the GenuB

Leptinotarsa, Washington, Cam. Inst. Publ., 4^, x, 321, 30 pis.,
31 figs, in text.
Ulmer, G.

1907. Trichoptera. Genera Insectorum (ed. by Wytsman, P.), 60, 259,

13 col. pis., 28 black pis.
Wallace, A. R.

1858. On the Tendency of Varieties to Depart Indefinitely from the
Original Type. J, Proc, Linn. 8oo,, London, S, 58-^2.
Willett, G.

1912. Birds of the Pacific Slope of Southern California. Pae. Coast
Avifauna, 7, 1-122.

Digitized by

Google

BIOLOGY OF THE THYSANOPTEBA^ ^
DR. A. FRANKLIN SHULL

XJNIVEBSITr OF MlOHIOAN

I. FACTORS GOVERNING LOCAL DISTRIBUTION

Intboduction

The Thysanoptera, commonly called thrips, are only
beginning to be known, in this conntry, by systematic
entomologists. The systematic knowledge is mostly con-
tained in the monograph of Hinds (1902), a more recent
synopsis by Monlton (1911), and a few other papers deal-
ing with new species and with relationships, prominent
among which is the work of Jones (1912). Biologically
the group is still less known. A considerable number of
papers have been issued from experiment stations, de-
scribing the life history {egg^ larval, pupal and adult
stages) and habits of thrips of economic importance. Be-
sides these the principal recent work of a biological
nature is a paper of my own (Shull, 1911), on the ecology,
method of locomotion, mode of reproduction, and dissemi-
nation. The life cycle of most species is still largely un-
known«

The first section of this paper is an attempt to carry
into further detail the study of the ecology of the Thy-
sanoptera. The first ecological scheme, so far as I am
aware, worked out for the Thysanoptera was that of Jor-
dan (1888), who divided thrips into three classes: first,
the flower-dwellers ; second, the leaf -dwellers ; and third,
all other thrips (for example, those living on fungi, under
wet leaves, under bark of trees, on roots, on lichens, etc.).
The inadequacy of this classification, and the difficulty of
applying schemes of ecology adapted to other groups of
insects, was pointed out in my earlier paper, where I pro-

1 GontributionB from the Zoological Laboratory ot the Uniyersity of
Miehigaiiy No. 142 (Biological Station Series, Zoological Publication,
No. 10).

161

Digitized by

Google

162 THE AMERICAN NATURALIST [VouXLVHl

posed a new scheme, based on my observations in the
field. In this scheme, Thysanoptera were divided into
two groups : (1) interstitial species, those living in closely
concealed situations, as among the florets of composite
flowers, or in clusters of young leaves; and (2) super-
ficial species, those living on exposed surfaces, for ex-
ample, the surface of leaves. The interstitial species
were further divided into an anthophilous division
(flower-dwellers) and a phlceophilous division (those
living under bark scales on trees). The superficial spe-
cies were either poephilous (on grass) or phyllophilous
(on leaves of plants other than grasses). The distinction
between poephilous and phyllophilous seemed warranted,
since grass-dwellers were found on many different
grasses, but rarely on other kinds of leaves.

Such a classification undoubtedly describes the facts,
but does not explain why the habitats named are the ones
chosen(f). The factors determining habitat were be-
lieved by me at that time to be character of food, and pro-
tection afforded. In some species one of these factors
predominated, in other species the other factor, while
others may have been influenced largely by both. In the
light of recent ecological studies, however, the explana-
tion of local distribution in terms of such general environ-
mental factors seems inadequate. Largely owing to the
work of Shelf ord (1911) upon the tiger-beetles, much
emphasis is now being placed upon the ecological impor-
tance of physiological factors. With a view to relating the
distribution of Thysanoptera to the physiology (more
specificially, behavior) of the various species, and thus
explaining that distribution in more definite terms, the
experiments and observations recorded in this paper
were made.

This work was done largely at the University of Michi-
gan Biological Station, at Douglas Lake, Michigan, sup-
plemented by observations at Ann Arbor, Michigan, in
Ohio and elsewhere.

Digitized by

Google

No. 567] BIOLOGY OF THE THTSANOPTEBA 163

Facts to be Explained

The following are some of the facts of habits and dis-
tribntion of the more abnndant species for which physio-
logical explanations were sought. Some of these facts
are stated in my former paper, some of them doubtless
the common property of all thysanopterists ; others, so
far as I know, have never been recorded.

Euthrips tritici is found almost exclusively in situa-
tions where it is concealed, as among the florets of com-
posite flowers, in clusters of young leaves, or in almost
any close crevice where the tissues are not too hard or
tough to be pierced. It appears to make little difference
what species of plant is inhabited, provided a concealed
situation is available. In the paper cited above (Shall,
1911) I gave a list of seventy species of plant on which
Euthrips tritici was taken, and I have since collected it on
a number of plants not included in that list. But with
rare exceptions, it has been found in crevices where it was
not readily visible. In related plants, it is always more
abundant in those affording concealed situations. Thus,
in white clover {Tri folium repens) and in red clover
(jT. pratense)y this thrips is usually abundant; while on
the related yellow, and white, sweet clovers (Melilotus
officinalis and M. alba, respectively), growing along with
the red and white clovers, Euthrips tritici is usually rare
or wanting. The flowers of Melilotus are widely sepa-
rated from one another on the stem, and do not afford
concealment (ShuU, 1911).

If, while Euthrips is in one of these crevices, it is dis-
turbed, as by gently rubbing or pressing the flower, it
quickly comes out of its retreat and crawls rapidly away,
or takes to flight. The larvae show the same behavior as
the adults in this regard, except, of course, that they do
not fly.

Anaphothrips striatus is found usually on grasses of
various kinds, rarely on leaves of other plants. The spe-
cies of grass seems to make little difference. Some indi-
viduals are found in perfectly exposed situations, as on

Digitized by

Google

164 THE AMEBIC AN NATURALIST [VoL.XLVin

the upper side of grass blades, others more or less con-
cealed in the rolled up young leaves (Shnll, 1911). I have
found, however, that among the adults, those in exposed
situations are almost exclusively females, while those in
the rolled young leaves are either males or females. (For
the first time on record, the males of this species, as will
be shown in the second part of this paper, have been found
in considerable numbers.) The larvsB, according to my
observations, may be either exposed or concealed; the ex-
posed ones are predominantly the older larvae.

In one of the grasses {Spartina michauxiana) on which
Anaphothrips was found in abundance at Douglas Lake,
Michigan, the leaves bear on the upper surface a set of
fine, but prominent, ridges running parallel to the axis of
the leaf. Adult females and larvae of Anaphothrips on
the exposed parts of these leaves were always lodged be-
tween the tops of these ridges, and almost invariably
unth their heads toward the base of the leaf. If disturbed,
they began to crawl along the crest of one of these ridges
toward the base of the leaf. It was possible to force them
to turn in the opposite direction, but if allowed to do so
they soon turned again toward the base of the leaf, often
continuing until they were among the rolled young leaves
in the center of the top of the plant.

Anthothrips verbasci is found exclusively on one spe-
cies of plant, the common mullein {Verbascum thapsus).
Furthermore, it is rare that a specimen of mullein, of con-
siderable size, is found free from the mullein thrips. Most
of the thrips are found among the florets or seed pods of
the spike. Less commonly they are to be seen on exposed
surfaces, as on the leaves or stem lower on the plant; but
these exposed individiials are mostly adults. The larvse
are usually hidden on the flower sp^e unless that situa-
tion is crowded by a large number of larvae ; and the larva
that are occasionally found exposed are mostly nearly
fully grown.

Anthothrips niger was not abundant enough during my
stay at Douglas Lake that many observations of its

Digitized by

Google

No. 567] BIOLOGY OF THE THTSANOPTEBA 165

habitat and behavior could be made. One f act, howeyer,
is of interest in connection with an experiment to be de-
scribed. While the adults live mostly on flowers, some-
times concealed, sometimes more or less exposed, the
larv» were always found concealed; moreover, it was
ivith difficulty that the larva could be driven from their
retreat by pressing the flowers. Frequently such vigor-
ous squeezing was necessary to dislodge them that the
larv® emerging were injured ; and a flower so treated was
often found later to contain numerous dead larvsB. In
this respect, the behavior of this species is in considerable
contrast to that, for example, of Euthrips tritici.

The habitats and behavior described above can be ^^ex-
plained'' in large measure if we say, as I at first proposed
(1911), that certain species seek protection, or that cer-
tain other species have specific food requirements. Thus,
it might be said that Euthrips tritici seeks safety in
crevices, and flees danger when disturbed ; that Anapho-
thrips striatus ** prefers'' grass for food, that it requires
as much protection as its commissarial activities permit,
and that its habitat and behavior are such as best fulfill
these requirements. Anthothrips verbasci might be said
to be limited to one article of diet, while protection is a
minor matter.

This explanation might be acceptable as far as it goes,
were it not that no species is immune to attack. I have
seen larvae of Anthothrips verbasci frequently captured
by various bugs. Heads of mullein where thrips are
found nearly always bear bugs of the family Capsidae,
and observations convince me that they prey almost
wholly on the larvae of the mullein thrips. The degree to
which they check the thrips was tested experimentally as
follows : Two mullein spikes of approximately equal size
and equally infected with thrips were selected. The
predatory bugs were removed from one of them, after
which the spike was enclosed in a thin muslin bag. Two
weeks later the bag was removed. The enclosed spike
bore a large number of full-grown larvae, a few had

Digitized by

Google

186 THE AMERICAN NATURALIST [Vol. XLVni

pupated, and many were crawling on the inside of the
bag. The spike that was exposed, on the other hand, bore
but little over half the number of larvae that were on the
protected one, none were quite full grown, and none had
pupated. Since nothing in the climatic conditions (heavy
rains, for example) could have caused this difference,
it is to be inferred that predatory bugs had devoured the
larger larvae in considerable numbers.

Yet Anthothrips verbasci, according to my earlier ex-
planation, ** chooses'' its habitat almost exclusively in re-
lation to food, protection being a minor consideration.

Can we not explain habitat and behavior in these in-
sects in some way not implying choice, especially choice
between conflicting preferences ? May we not assume that
certain elements of behavior are what they are without
reference to their usefulness? If we grant the possibility
of an aflSrmative answer to these questions, the experi-
ments about to be described will have significance.

Experiments on Behavioe
The following experiments were designed to show the
reaction of the commoner species of Thysanoptera to
what seemed to me the most probable external agents
affecting their distribution and behavior, namely, light,
contact and gravity. Inasmuch as I was not primarily
interested in how a given reaction was brought about, but
only in its end result, the experiments were rather crude.
Eefinements were unnecessary, and their omission en-
abled me to use much greater numbers of individuals than
would otherwise have been possible. From ten to forty
repetitions of each test were usually made. The experi-
ments are described by species, only representative ex-
periments being given.

Euthrips tritici
Light. Exp. 1. — ^Adults of this species were placed in
a glass tube about three feet long and one inch in diam-
eter, closed at the ends with corks. One end of the tube

Digitized by

Google

No. 567] BIOLOGY OF THE THTSANOPTERA 167

was turned toward a small window, while the room was
rather dimly lighted. All the thrips crawled rapidly
toward the window. When the position of the tube was
reversed, the thrips reversed their crawling, again going
toward the iraidow. The reaction was definite and in-
variable.

Exp. 3. — ^A close-fitting sleeve of black building paper
was slipped over one half of the glass tube used in experi-
ment 1. The thrips were collected at the exposed end by
turning that end for a few minutes toward the window.
The covered end of the tube was then turned toward the
window. The thrips crawled rapidly toward the light,
until they reached the shadow of the sleeve. Here they
crawled about, apparently aimlessly, for half an hour an
inch or two within the sleeve or just outside it.

Contact. Exp. 1. — ^When, in the light experiments, the
tube was reversed in position as soon as the thrips
reached one end, the insects immediately turned toward
the opposite end. But if the tube was allowed to rest for
some time, the thrips became settled quietly between the
glass and the sloping surface of the cork. The tube could
then be carefully reversed, and most of the thrips re-
mained lodged between cork and glass for many minutes,
some of them for hours. The positive reaction to contact
counteracted the positive reaction to light.

Exp. 25. — ^A larva of this species was placed on a glass
plate, upon which rested a microscope slide. When the
larva in its crawling reached the slide, it came to rest in
the angle formed by the glass plate and the edge of the
slide. It remained there many minutes until disturbed.

Gravity. Exp. 17. — ^An adult female was placed in a
glass tube which was enclosed in a black sleeve to exclude
Ught, and the tube placed in a vertical position. The posi-
tion of the thrips was marked with a wax pencil before
putting on the sleeve. The sleeve was then removed mo-
mentarily at frequent intervals, and the position and
direction of crawling of the insect noted. Most fre-
quently it was found lower than the previous position.

Digitized by

Google

168 TRE AMEBIC AN NATURALIST [Vol. XLVHI

and crawling downward. This was not always the case,
however.

Of other specimens tried, some showed positive geo-
tropism more definitely, some less definitely than the one
described. None showed a negative reaction in the ma-
jority of cases.

Anaphothrips striatus

Light. Exps. 5 and 7. — ^Adnlts of this species were
shaken ont on a sheet of white paper near a window, and
the course of their crawling was plotted as accurately as
possible in my notes. Some individuals were decidedly
negative to light, crawling directty away from the window
every time they were tried, regardless of the direction in
which they happened to be headed when they touched the
paper. Others were indifferent to light, crawling in vari-
ous directions. Most of the males used were decidedly
negative to light, females usually indifferent.

Exp. 10. — ^Females taken from the exposed portions of
leaves of Spartina michauxiana, and tested as above,
were found in nearly every case to be indifferent to light.
Females from the curled young leaves of the same plants
were as a rule negative to light.

Exp. 6. — ^Larvae were usually found indifferent to light,
regardless of whether they came from exposed or con-
cealed situations.

Exp. 15. — ^A single larva taken from the exposed part
of a leaf, when placed in a glass tube one end of which
was directed toward the window, crawled steadily toward
the window. When the position of the tube was reversed,
the larva at once reversed its direction. The tube was
then placed in a black sleeve to exclude the light, and kept
there for an hour. When it was removed, the larva
showed for some minutes a decidedly negative reaction to
light. Later, however, its behavior became indefinite,
and soon became markedly positive. Darkness had ap-
parently temporarily reversed its reaction.

Contact. Exp. 22. — ^A female of this species which was

Digitized by

Google

No. 567] BIOLOGY OF THE THJSANOPTEBA 169

negative to light was placed on a sheet of blotting paper.
A small square of glass was placed over her, and sup-
ported at one edge, so that in crawling away from the
window the thrips approached the edge of the glass which
was in contact with the paper. She soon became lightly
wedged between the glass and the blotter, and came to
rest. Blotter, thrips and glass were tiien carefully
turned through 180 degrees so that the negative reaction
to light would have led the thrips out of its crevice; but
she remained there for a long time. Positive reaction to
contact overcame the negative reaction to Ught.

Another female, indifferent to light, was placed under
a sinoilar glass. In her random crawling she became
wedged between the blotter and glass, and, notwithstand-
ing that the blotter was occasionally turned in the mean-
time, remained there several hours, until I lifted the glass.

Another female, not negative to light, was placed under
a similar glass square. She crawled from under it, but
happened to crawl against the edge of the microscope
slide that supported the glass cover. She settled quickly
into the right angle formed by the slide and the blotter,
and remained there a long time.

Gravity. Exp. 21. — ^A female which was indifferent to
light was placed in a glass tube, and the tube set in a
vertical position. The thrips immediately began to crawl
downward. The tube was reversed, and the thrips im-
mediately reversed its direction. A sleeve was placed
over the tube to exclude the light, and frequently removed
temporarily to observe the position of the thrips. In
every case she was found crawling downward.

"Wlien the tube was held in an oblique position, the re-
sult was the same ; the thrips crawled down the slope. If
she was already crawling down, a slope of 5 to 10 degrees
was found to be sufficient to keep her going in the same
direction. But to reverse the direction of crawling, it
was necessary to create a slope of about 45 degrees in the
opposite direction. The same positive geotropism was
shown when the thrips was placed on an inclined sheet of

Digitized by

Google

1 70 THE AMERICAN NATURALIST [Vol. XLVUI

paper; but being here at liberty to fly, she soon inter-
rupted the experiment.

Numerous other females were tried, and all showed
positive geotropism, some more promptly than others,
but all perfectly definitely. A single male tested showed
no definite reaction to gravity. A larva, nearly full
grown, subjected to the same tests, showed as definite a
positive reaction to gravity as did any of the females.

With the possible exception of the males, therefore,
Anaphothrips striatus is decidedly positive to gravity.

Anthothrips verbasci

Light. Exp. 4. — ^Adults of this species, shaken out on
a paper near a window, crawled in various directions.
None of them showed any definite reaction to light.

Numerous larvae, none of them over three fourths
grown, crawled directly away from the window in every
instance.

Exp. 12. — ^In this experiment adults from concealed
places in mullein spikes were compared with those from
exposed situations. They were shaken out on a sheet of
paper near a window, and the direction of crawling noted.
In every case, those from concealed situations showed a
fairly definite negative reaction to light. Of those from
exposed situations, two were plainly negative, the re-
maining ten indifferent to light.

Exp. 11. — ^Larvae taken from concealment in a mullein
spike were tested, on a sheet of paper, for their reaction
to light. Those of the smaller sizes crawled directly away
from the windoiv. Those nearly full grown, while on the
whole negative, crawled in a more or less devious path
away from the window;. One reddish larva, which from
its color and size must have been nearly ready to pupate,
was especially indefinite in its reaction to light.

Contact. Exp. 18. — ^Larvae of various sizes, which were
found to be negative to light, were placed on a blotter
under a square of glass supported at one edge, as de-
scribed for Anaphothrips striatus. When, in crawling

Digitized by

Google

No. 567] BIOLOGY OF THE THTSANOFTERA 171

away from the window, they became wedged lightly be-
tween glass and blotter, and came to rest, the blotter with
all on it was turned through 180 degrees. The larvae
turned their bodies so that their heads were directed away
from the window, but did not crawl away. The positive
reaction to contact overcame the negative response to
light.

An adult tested in the same manner as the larvae above
described did not come to rest under the glass square.
But happening to crawl against the microscope slide
which supported the glass, the thrips came to rest in the
right angle formed by the blotter and the edge of the
slide, and remained there a long time.

Gravity. Exps. 13 and 20. — ^Adults and larvae were
put, one at a time, into a glass tube, which was set in a
vertical position, and covered with a black sleeve to ex-
clude light. Some were examined at frequent intervals,
others were left half an hour without examination. In
every case the thrips were found almost precisely where
they were put at the beginning of the experiment. This
species is therefore indifferent to gravity.

Anthothrips niger
Light. Exp. 2. — The red larvae of this species were
shaken out on a paper near a window, as described in
other experiments. In every case the larva crawled away
from the window for a few seconds at firsts then slowly
turned toward the window, and continued indefinitely
toward the light. Once while the larva was crawling
toward the light, I tapped the paper vigorously with a
pencil, so that the thrips was lifted slightly from the
paper and let drop; it immediately reversed its direction,
crawling from the window, but in a few seconds turned
again toward the light. The paper was jarred frequently,
but always with the same result. To show whether the
jarring made the response to light negative, or merely
reversed whatever the larva was doing at the instant, the
tapping was repeated at intervals of one or two seconds.

Digitized by

Google

172 THE AMERICAN NATURALIST [VoL.XLVni

At the first tap, the larva, which had been crawling
toward the window, immediately tnmed away from the
light. Before it resumed its positive response to the light,
the paper was tapped again ; the negative response con-
tinued. In this way the larva conld be kept crawling away
from the light indefinitely. Disturbance makes the reac-
tion of the larva to light temporarily negative; otherwise
it is positive.

Summary op Expbeimbnts

Euthrips tritici, when disturbed, is positively photo-
tropic in both larval and adult stages. It is positively
stereotropic, and the stereotropism is stronger than pho-
totropism, at least under certain circumstances. Some
individuals appear to be on the whole positively geo-
tropic ; others are indiflPerent.

Anaphothrips striatus. — ^Adult males are usually nega-
tively phototropic. Females taken from exposed situa-
tions are usually indifferent to light, those from concealed
situations usually negative. The larvae are usually in-
different to light, regardless of the kind of place from
which they are taken ; a single larva that was positive was
made negative by keeping it in the dark. Adults are posi-
tively stereotropic. The females and larvae are positively
geotropic.

Anthothrips verbasci. — ^Adults taken from concealed
situations are usually negatively phototropic, those from
exposed places tend to be indifferent to light. The larvae
are all negatively phototropic, except the full-grown ones,
which may be indifferent. The larvae are plainly posi-
tively stereotropic, the adults less plainly so, or not at all.
Neither adult nor larva responds to gravity.

Inteepbetation op the Experiments in thbib Bblation
TO Distribution and Behavioe op Thrips in Nature
With the evidence from these experiments before us,
may we not interpret the observed distribution and be-
havior of the Thysanoptera in nature somewhat as fol-
lows ? Instead of explaining the fact that Euthrips tritici

Digitized by

Google

No. 667] BIOLOGY OF THE THTSANOPTEBA 173

always lives in concealed situations as due to a demand
for protection, we may assume that it is due to the strong
positive stereotropism of this species — aided in some
cases by positive geotropism, where the flower inhabited
is upright, but notwithstanding positive geotropism
where the flower is inverted. The rapid escape by crawl-
ing or flight when disturbed is not due to the fact that
this is the best way of avoiding danger, but to the posi-
tive reaction to light. Other species avoid danger by
going deeper into crevices, because they are negatively
res]>onsive to light.

Anaphothrips striatus lives on grasses doubtless be-
cause it can not live on any other food, or because the
reproductive processes are not stimulated by any other
host plant. But their distribution and behavior on the
grasses may be explained largely in terms of their reac-
tions to the three agents tested in the experiments. The
males usually live in concealed situations on the plants
(curled-up leaves) because they are mostly negatively
phototropic, and crawl down the leaves until they reach
these concealed situations. Females may live either in
exposed or in concealed places, for some of them are
negative to light, others indifferent. The larvae are either
exposed or concealed, because they are indifferent to
light. The eggs from which they hatch are probably laid
by negatively phototropic females in the young curled
leaves, and tiie leaves unfold as the larvsB develop; this
explains why the exposed larvsB are much larger, on the
average, than are those concealed in the young leaves.
Perhaps the relation of cause and effect as here stated is
reversed, at least for some cases. Concealment — caused
in one way or another — ^may lead to negative phototro-
pism, as in the larva which was made temporarily nega-
tively phototropic by being kept in the dark.

The adults are lodged between the ridges on the upper
side of the leaves of the grass Spartina, not for the sake
of protection, it seems to me, but because they are posi-
tively stereotropic. Doubtless between the ridges is the

Digitized by

Google

174 THE AMEBIC AN NATURALIST [Vol. XLVni

place where they can best suck the juices of the plant, but
there is no need to assume that they deliberately choose
this location in order to get their food most easily. Both
adults and larvae rest on these leaves with their heads
directed toward the base of the leaf, and crawl toward
the base of the leaf if disturbed, not because protection is
most quickly to be found among the curled leaves at the
center of the plant, but because the thrips are positively
geotropic.

Anthothrips verbasci. — The larvae of this species live
hidden among the flowers of the mullein spike, not be-
cause they must get their food there, for they can get it
from any part of the plant; nor do they hide there, it
seems to me, to secure protection. They remain in these
crevices because, excepting the largest larvae, they are
positively stereotropic and negatively phototropic. The
adults are sometimes exposed, sometimes concealed, prob-
ably because in the former case they are usually indiffer-
ent to light, in the latter case negatively phototropic. (Or
may they be made negative or indifferent according as
they live — ^f or one reason or another — concealed or ex*
posed?)

Thus, while Anthothrips verbasci is limited to one food
plant, and the food requirements are therefore probably
exceedingly important, yet the distribution and behavior
of the insects on this plant may be explained without ap-
pealing to anything like ** choice*' in other matters.

Regarding Anthothrips niger, I wish to call attention
to but one fact. The difficulty with which the larvae are
driven forth from a flower in which they live appears to
be due, not to a persistent attempt at concealment, but to
the fact that on being disturbed they are temporarily
negatively phototropic; if the disturbance is continued,
the negative response continues.

The only argument which, it appears to me, could be
advanced in favor of assuming that the Thysanoptera
choose their locations, instead of adopting simple re-
sponse to external stimuli as the correct explanation of

Digitized by

Google

No. 567] BIOLOGY OF THE THTSANOPTEBA 176

distribution, is the possibility that they have learned that
certain modes of behavior are best suited (for example)
to continued safety.

The reply to such an argument is first, that most of my
studies on behavior have been made in an unsettled re-
gion, where the enemies of thrips incident to civilization
are practically wanting, and where even the natural
enemies are not abundant. It could hardly be assumed
that every individual would learn to avoid its enemies in
the course of its short lifetime, yet certain species seem
to be invariable in their response to certain agents.
Furthermore, many of the larvae tested in the experi-
ments could have been but a few days old. It is incred-
ible that their reactions should have been, as in fact they
were, as definite and invariable as those of older larvae,
if these responses were dependent on experience.

It seems to me, therefore, that the only satisfactory ex-
planation of outdoor behavior and distribution of the
Thysanoptera lies in the assumption that they are in
large measure the result of responses to simple stimuli,
and do not imply any degree of choice.

Obigin and Adaptivenbss op Besponses to External

Stimuli

The origin of such responses in Thysanoptera as have
been described above is not, I believe, discoverable. Pur-
poseful they most probably are not, as I have shown, if
by purpose we mean conscious direction of actions to
some end. But adaptive they no doubt are in many cases.
Perhaps they are all adaptive, but I confess that my
powers of analysis are not keen enough to prove such a
view correct. That Euthrips tritici is positively photo-
tropic when disturbed is no doubt the cause of frequent
escapes from danger. One may even believe the negative
phototropism of larvae of Anthothrips verbasci to be
adaptive, because they are much more sluggish than is
Euthrips tritici, and could not escape quickly even if they
should emerge into the light. They are probably safest.

Digitized by

Google

176 THE AMERICAN NATURALIST [VoL.XLVm

theref ore, if , when disturbed, they retire into still deeper
crevices. But I am nnable to discover the adaptiveness
of the response of the larvsB of Anthothrips niger to light
— at first negative, on being distorbed, bnt soon becoming
positive. Nor can I understand why the males of Ana-
phothrips striatus are more definitely negative to light
than are the females or larvsB. These reactions seem to
me to be useless.

We need not demand that all of these responses be
adaptive, any more than that they be purposeful. Re-
sponses have arisen, no one knows how. They have been
preserved, and we can but speculate as to the method of
their preservation. Natural selection may be respon-
sible for the preservation of the useful, and it may have
eliminated res]>onses that were harmful. But other re-
sponses of no value whatever, but likewise harmless, may
have been allowed to persist, without help or hindrance
from selection.

(To he continued.)

Digitized by

Google

SHORTER ARTICLES AND CORRESPONDENCE

THE ENDEMIC MAMMALS OF THE BRITISH ISLANDS

When, in 1891, I was collecting information to be used by
Dr. A. E. Wallace in preparing the second edition of his ''Island
Life," I found much skepticism among naturalists concerning
the alleged endemic or precinctive elements of the British fauna.
Dr. Wallace was able to give lists of supposed precinctive species
and varieties belonging to several groups, but for the mammals
he was obliged to state, "it is the opinion of the best authorities
that we possess neither a distinct species nor distinguishable
variety." We little imagined that about twenty years later the
British Museum would issue a work describing ten species and
twenty subspecies of mammals peculiar to the British Islands;
twenty-one of these being actually undescribed at the time I
made my enquiries, and the rest then reposing quietly in the
synonymy. Still less did we imagine that such a revision, when
made, would be the work of an American, coming over from the
United States National Museum to show Europeans the neglected
wonders of their own fauna ! The Catalogue of the Mammals of
Western Europe, by Mr. G. S. Miller, published last year by the
British Museum, is certainly one of the most remarkable zoolog-
ical works ever produced, and is well worthy of the attention of
all naturalists, whether specially interested in the Mammalia
or not. While so many students of genetics are giving us the
results of their experiments in breeding mammals, it is worth
while to turn also to the results of nature's long-time breeding
exx>eriments, so clearly set forth by Mr. Miller in the volume cited.
What, after all, is the connection between the phenomena seen by
the breeder and the facts of mammalian evolution t Do species
and subspecies differ by "units,'' and do the variations observed
in captivity correspond in any way to the recorded specific and
snbspecific differences T

A complete analysis of Mr. Miller ^s volume can not be made at
the present time, but I have extracted the list, given below, of
the forms supposed to be confined to the British Islands, giving
their distribution and principal distinctive characters. I have
added to Mr. Miller's list three quite recently described animals.
On examining the list, it appears that a few of the species must
belong to the older fauna of the country, not wholly exterminated
by the glacial ice and periods of partial submergence. Such are

177

Digitized by

Google

178 THE AMERICAN NATURALIST [VouXLVm

Mustela hibemica of Ireland and Microtus orcadensis of the
Orkney Islands. It is at least suggestive, in this connection, that
so many of the Scottish islands yield animals differing from
those of the mainland. In the majority of cases, however, the
peculiar British mammals are closely related to those of the con-
tinent, and might well be of very recent origin. There is a
decided tendency to darker colors, such as has been noted also
among British moths. In spite of this tendency, however, some
forms are lighter than their relatives, the most conspicuous case
being the light-tailed British squirrel. In several cases the differ-
ence noted has in part to do with particular phases; thus the
squirrel has no dark phase, and the ermine does not turn so white
in winter. The British red grouse, it will be remembered, is
peculiar in lacking a white winter phase. Some of these differ-
ences may be due to the direct effect of the mild and moist
British climate, and would perhaps disappear in the descendants
of British animals taken elsewhere. The experiments on birds by
Beebe are very suggestive in this connection. In other cases, the
distinctions are such as might readily result from changes in one
or two ** units," such as are observed in experimental breeding.
"When we have a variable type, subject to losses and new combi-
nations of unit characters, it is perhaps to be expected that
different groups of individuals, isolated from one another, will
after a time produce different homozygous combinations. That
is to say, the result comes from a long series of ** accidents,"
which will probably not be duplicated in two different places. In
this way mere isolation may be an adequate cause of modification,
providing always that through variation degrees of hetero-
zygosity have arisen.

In the common house mouse, Mus musculus, Hagedoom^ has
isolated and figured a great number of color varieties, for nearly
all of which he bas constructed zygotic formulae. Little* has
also described and figured a similar series of varieties, appar-
ently in ignorance of Hagedporn 's paper, which he does not cite.
He gives zygotic formulae for thirty-two different varieties, but
not all of them are visibly different. Albino varieties, resulting
from the dropping out of a particular determiner, may be pro-
duced, corresponding in other respects to each of the thirty-two
colored forms, although they all look alike, and will only show
their true characters on crossing. Several of the varieties show

1 Zeii. f. ind, Ahst. Ver,, 1912.

'"Experimental Studies of the Inheritance of Color in Mice," 1913.

Digitized by

Google

No. 567] SHORTER ARTICLES AND CORRESPONDENCE 179

noteworthy fluctuating variability, due to differences in ex-
pression.

Muts musculus, then, is very conspicuously variable in color;
yet Miller's book records only one subspecies, that of the Medi-
terranean region and the Azores, which is less dusky and more
yeUowish, with the under parts buffy grayish. It possibly agrees
with Little's ** dilute black agouti" variety. On the other hand,
M. musculus has a recognized subspecies in Mexico, whete it
must have developed since the species was introduced by man.
The mice of St. Eilda and the Faroe Islands, although given as
distinct species, are derivatives of Mus musculus, differing in
other points than color. In connection with subspecific differ-
ences in size, Sumner's experiments with different temperatures
should be noted, since they prove that differences of temperature
might lead to readily measurable differences in dimensions,
wholly unconnected with losses of determiners or new zygotic
combinations. Whether or not diverse conditions of this sort
would ultimately affect the germ plasm, their effects would be
patent long before and quite independently of any such modifi-
cation. On the whole, the poverty of Mu^ muscuiu^ in subspecies
would suggest that the variations observed by breeders are not,
as a rule, the stuff that new subspecies are made of. Against this
argument may well be adduced the fa6t that M. musculus is an
urban animal, constantly traveling about, so that incipient races
do not remain isolated. Here the closely related rats, Epimys,
are worth considering. For Europe Miller can only recognize
the Norway, Black and Alexandrian rats, all widespread, prac-
tically cosmopolitan. Yet in the Malay Archipelago, where
Epimys is distributed over myriads of islands, large and small,
the species are innumerable. One can almost take a map and
indicate where new species of Epimys are to be found, namely, on
those islands still unexplored. Years ago, when the writer was
actively engaged in studying the British MoUusca and Lepi-
doptera, the question of endemic forms was constantly in mind ;
but in those days we failed to discriminate properly between the
different classes of "varieties. ' ' We made the mistake of looking
for well-marked ''sports" or aberrations, rather than for con-
stant but only slightly distinguished local races. There was a
practical reason for this, in the fact that by searching the litera-
tore we could ascertain whether a well-marked variation had
been reported from the continent ; whereas the determination of
subspecific types analogous to those described by Miller among

Digitized by

Google

130 THE AMEBIC AN NATURALIST [VouXLVm

mammals required long series from different parts of Eiirope,
and these we did not possess, and could not readily obtain. Miller,
following the custom of mammalogists, lays great stress on sub-
species, but almost ignores individual variations, except such as
are expressed by the statistical data regarding size. By reading
the synonymy, one can see that many such variations have re-
ceived names, and I can not doubt that the time will come when
iheOd names will be generally used. In this case, it will be
extremely desirable to use the same adjectival name for analogous
varieties of different species, and beyond the limits of subspecies
it ought not to be held that a name once used in a genus can not
be employed again. It may be true that most or all of the '* indi-
vidual" varieties can be expressed by zygotic formulae, but one
can not remember all these formulae, nor use them in speech with
any comfort. Moreover, they have to do with the germinal con-
stitution rather than the patent characters. Little provides all
his varieties with polynomial English appellations, but would not
Latin varietal names be better f Following his theory con-
cerning the pigments, some of the varieties receive names
which do not suggest the animals at all; thus ''brown-eyed
yellow," according to the apparently excellent colored plate,
is light orange-ferruginous, while "sooty-yellow" is dark gray
with yellowish under parts. Morgan* describes a wild variety of
M. musctdv^ from Colorado, which he calls ''mauve," but from
the detailed account it is rather "fauve," namely, fulvous op
yellowish brown. It must be similar to the Old World subspecies
azoricUs, or possibly that subspecies introduced t If we had
standard scientific names for the different forms, we should try
to compare our specimens with the types or descriptions of those
names, and it would not be left to authors to use such miscellane-
ous descriptive terms as might occur to them. For Mus musculus,
possibly Little's apparently excellent colored plates might be
made the standards for a series of names. Thus his Fig. 9
(pi. 8) is the animal named niger as long ago as 1801; Fig. 10,
the dilute black, would naturally take the name nigrescens.
Fig. 12 is probably albicans of Billberg, 1827.

Mammals Peculiab to tbs British Islands
In$ectivora
8orex aranem castaneus (Jenyns 1838). Great Britain. Not so dark as true
araneuB,

tAnn, N, T. Acad, 8ci., XXT, p. 106.

Digitized by

Google

No. 567] 8H0BTEB ARTICLES AND COBBESPONDENCE ISi

Borex granti (Barrett-HamUton and Hinton 1918). Inner Hebrides. Dif*
fera from araneus bj the eontraat between bright-colored flanks and
dusky npper parts; teeih also different.

Veamys fodiens hicohr (Shaw 1791). Qreat Britain. Under parts nsoally
washed witit wood-brown instead of bnffy whitish; skull smaller.

Chiroptera

Khwolaphus ferrum-equinum insulanuB Barrett-Hamilton 1910. Central and
8. England. Wing shorter.

MMnolophui hippasideros tninutiu (Montagu 1808). England and Ireland.
Wing shorter.

Camivora

Musiela erminea stabUis (Barrett-Hamilton 1904). Mainland of Great
Britain. Bather large, with large teeth; color somewhat different, a
little darker above. Change to white in winter less complete and regu-
lar than in continental forms.

Miuiela erminea riciruB (Miller 1907). Islands of Islay and Jura, Scot-
land. Smaller than etdbiUa; proportions of skull different.

Mustela hibemica (Thomas and Barrett-Hamilton 1895). Ireland and Isle
of Man. Quite distinct; recognized by combination of black-tipped,
heavily penciled tail with entirely dark ear and upper lip. Superficially
like certain North American forms.

Felia sylvestria grampia (Miller 1907). Scotland; formerly throughout
Great Britain. Darker, with more pronounced black markings.

Bodentia
Lepus europcBus occidenialis de Winton 1898. England, Scotland and Isle

of Man. Buffy tints rich and dark.
Lepus iimidus scotieus (Hilzheimer 1906). Highlands of Scotland.

Smaller; rarely becomes so white in winter as Alpine race.
Lepus T^emicus Bell 1837. Ireland. Distinguished by the strongly russet

color and partial or complete absence of white winter coat. Larger than

seoticus.
Bvotomys aistoni Barrett-fiamilton and Hinton 1913. Island of Mull,

Hebrides.
Evotomys glareolw hritannicus (Miller 1900). Great Britain. Smaller;

color less intense.
Evotomys skomerensis Barrett-Hamilton 190S. Skomer Island, off coast of

Wales. Color nbove unusually light and bright; skull peculiar.
Microtus agrestis esBsul Miller 1908. North and South Uist, Hebrides. Be-

sembles true agrestis of Scandinavia; teeth peculiar, a character usually

present which elsewhere in the species occurs as a rather rare anomaly.
Mierotus agrestis maegiUivraii Barrett-Hamflton and Hinton 1913. Island

of Islay, Hebrides.
Mierotus agrestis Iwrtus (Bellamy 1889). England and South Scotland.

Smaller than typical agrestis; upper parts noticeably tinged with rus-
set, and venter washed with wood-brown.
Mierotus agrestis negleotua (Jenyns 1841). Highlands of Scotland. Not

so small as Mrtus; upper parts darker.

Digitized by

Google

182 THE AMERICAN NATURALIST [Vol. XL VIII

Miorotus orcadenm Millais 1904. South Orkney Islands. Belated to If.
aamius of Guernsey and the Pleistocene M, corneri of South England.
Distinguished by its large size and dark color.

Miorotus sandayensis (Millais 1905). Sanday Island, N. Orkney group.
Allied to orcadensis, but skull differing; upper parts much lighter.

Miorotus sandayensis westrcB MiUer 1908. Westray Island, N. Orkney group.
Not 80 pale as in typical form; teeth differing a little. *

Arvioola amphilius (L. 1758). Typical subspecies. England and South
Scotland. Large; color moderately dark.

Arvioola amphibius ater (Macgillivray 1832) = reta Miller 1910. Scot-
land, except southward. Darker, melanism frequent. The name was
changed on account of EypudoBus terrestris var. ater Billberg 1827, but
the change is perhaps needless, as Billberg 's animal was not a sub-
species, and has not been treated as a species or subspecies under
Arvioola.

Apodemus hehridensis (de Win ton 1895). Lewis and Barra islands, Hebri-
des. Large, with small ears; color dark.

Apodemus hirtensis (Barrett-Hamilton 1899). Island of St. Kilda. Near
hehridensis, but skull larger and color darker.

Apodemus fridariensis (Kinnear 1906). Fair Isle, Shetland group. Large;
skull peculiar; colors also somewhat peculiar.

Apodemus fia/vioollis wintoni (Barrett-Hamilton 1900). England. Under
parts with duller color, pectoral spot more diffuse.

Mus muralis Barrett-Hamilton 1899. Island of St. Kilda. Like M, musoulus
but feet and tail less slender; skull peculiar.

Mus fceroensis (Clarke 1904). Faroe Islands. Larger than musoulus and
muralis; hind foot very robust; tail thickened.

Sciurus vulgaris leuoourus Kerr 1792. Great Britain and Ireland. Small;
tail drab, fading in summer to cream buff. No dark phase.

Vngulata

Cervus elaphus sootious Lonnberg 1906. Great Britain. Color darker and

less gray than in the related Norwegian form.
Capreolus oapreolus thotti Lonnberg 1910. Great Britain. Darker, face

darker than body.

I thought it of interest to compare the above British list with
a similar one for the Spanish peninsula (Spain and Portugal).
The latter area is continuous northward with France, but the
Pyrenees constitute a barrier. The Iberian peninsula differs so
much in its recent geological history from Britain, and is at the
same time so much more southern, that we should expect to find
the faunal elements very different. This expectation is realized,
yet the difference in numbers between the two lists is not very
great, and the number of Iberian forms treated as distinct
species is exactly the same (12) as that for the British Islands.
This suprising result is evidently due to the numerous small

Digitized by

Google

No. 567] SHOETEB ARTICLES AND CORRESPONDENCE 183

islands of the British group, sti<;h islands being wanting around
the coasts of Spain.

Mammals Peculiab to thx
Inseetivora
Talpa ocoidentaUs (Cabr.).
Oalemys pyrenaious rufulus

(GraeDs.).
8<fr€x araneus granariut Miller.
Neomys anomalus Cabr.
Crocidura mimuJa cantabra (Cabr.).
Crocidura russula dnira MiUer.
Brinaceus europcBus hispanioua B.-

Ham.

Chiroptera

(None.)

Camivora
Cants lupu8 signatus Cabr.
Cants luptu deiianvs Cabr.
Meles metes marianensis (Graells.).
Maries foina mediterranea

(B.-Ham.).
Mnstela nivalis iberica (B.-Haiii.).

(Also Balearic Is.).
Mustelaputoriusaureolus (B.-Ham.).
Mungos toiddringtonii (Gray).
Oenetta genetta (L.), typical nibsp.
Felts sylvestris tartessia (Miller).
Lynx pardeUus Miller.

Bodeniia
Lepus granatensis Bosenb.

(Also Balearic Is.).
Leptu granatensis galUseius Miller.

Univebsity of Colokado

Spanish (Iberian) Pininstjla
Eliomys lusitanieus (Beuvens).
Glis glis pyrenaious Cabr.
Microttu agrestis rosianus (Bocage).
Mieroius asturianus Miller.
Arvicola sapidtu Miller, typical

subsp.
Pitymys lusitanieus (Gerbe).
Pitymys maruB (Major).
Pitymys pelandoMus Miller.
Pitymys depressus Miller.
Pitymys iberious (Gerbe), typical

sabsp.
Pitymys iberious centralis Miller.
Pitymys iberious pasouus Miller.
Pitymys iberictu regulus Miller.
Mus spicilegus i kfanicus Miller.
Mus spicUegtu hispanicus Miller.
Soiurtu vtUgaris numantius MiUer.
Soiunu VtUgaris infusoatus (Oabr.).
Sdmus VtUgaris segura Miller.
Sciurus vulgaris bcsticus (Cabr.).

UngtUata
Bus sorofa castUianus Thomas^
8us sorofa bcetious Thomas.
Cervus elaphus Mspanieus Hilxh.
Capreolus oapreohu oantu Miller.
Capra pyrenaica lusitanica (Franca).
Capra pyrenaioa victoria Cabr.
Capra pyrenaica hispanioa (Schimp.).
Bupicapra parva (Cabr.).

T. D. A. COCEERELL

LITEBATUBE CITED
BatesoB, W. Mendel's Principles of Heredity Cambridge (England)

University Press. 1909. 396 pp.
Castle, W. E. Heredity of Coat Characters in Guinea-pigs and Babbits.

PubL Carnegie Inst of Wash. No. 23, 1905.
Cannot, L. La loi de Mendel et I'h^r^dit^ de la pigmentation chez les

souns. 4me note. Arch, Zool exp. et g4n. Notes et Bevue, 1905.
Darbishire, A. D. Notes on the Besults of Crossing Japanese Waltzing

Mice with European Albino Baees. Biometriica, Vol. 2, p. 101, 1902.

Digitized by

Google

184 THE AMERICAN NATURALIST [VouXLVHl

Poncsster, L. On the Inheritance of Coat Colour in Bate. Proo, Camb,
Phil. Soc, Vol. 12, pt 4, p. 215, 1905.

Durham, F. M. A Preliminary Account of the Inheritance of Coat Colors
in Mice. Bept. Evol. C't'ee. B07. Soc., IV, 1908.

Hagedoorn, A. L. The Genetic Factors in the Development of the House-
mouse which Influence the Coat Color, with Notes on Such Factors in
the Development of Other Bodents. Zeit fur indukt, Abst. u. Vererh,,
Bd. 6, pp. 97-136, 1912.

McCurdj, H., and Castle, W. E. Selection and Crossbreeding in Belation to
the Inheritance of Coat-pigments and Coat-patterns in Bats and Mice.
Publ. Carnegie Inst, of Wash., No. 70, 1907.

Morgan, T. H. Becent Experiments on the Inheritance of Coat Color in
Mice. Am. Nat., Vol. 43, pp. 494-510, 1909.

Digitized by

Google

NOTES AND LITERATURE

SWINGLE^ ON VARIATION IN F^ CITRUS HYBRIDS
AND THE THEORY OP ZYGOTAXIS

SwiKGiiB in two recent papers has published some very inter-
esting observations on Citrus species and their F^ hybrids. On
the basis of these observations, the somewhat startling statement
is tnade that current theories of heredity and variation give no
adequate explanation of variability in F^ hybrid generations
from ''pure bred'' parent strains. Swingle assumes this vari-
ability to be so great that qualitative differences in chromosomes
can not account for it. As the chromosomes in the F^ hybrid
remain unfused until synapsis, there is said to be no opportunity
for quantitative exchange of hereditary substance, so that this
variation can not be accounted for on this basis. Hence,

if proof can be given to show that in certain specific cases, pairs of
gametes of identical hereditary composition* give rise to very diverse
organisms, the way has been opened for a general reinvestigation of the
validity of our modem theories of heredity.

The term "pure bred" as used by Swingle implies that cer-
tain Citrus species reproduce themselves in a relatively faithful
manner from seed, there being no overlapping of distinguishing
specific characters and very little variation of these characters
intraspecifically. C. aurantium and C. trifoliata are examples
of such widely separated species. The former has been grown
from seed in Florida for two hundred years, and though varia-
tions have appeared, they are said to differ but little from the
general type of C. a/urantium, and in no way to approximate
that of C. trifoliata.

On the basis of evidence of this kind, Swingle believes the
various Citrus species (C aurantiuniy C. trifoliata, C. medica
limonum, etc.) breed true in nearly all their characters and
especially in those which differentiate them from one another.
Hence, for genetic studies, the germ cells of these species are

1 Swingle, W. T^ "VariAtion in First Oentration Hybrids (Imperfect
Domisanoe) : Its Possible Explanation through Zjgotaxis,'' IV® Conf. In-
temat. de Genetique^ Paris, 1911, pp. 381-394; <<Some New Citrus Fruits,"
Amer. Breed. Ma^,, 4: 83-95, 1913.

s The italics are mj own.

185

Digitized by

Google

186 THE AMERICAN NATURALIST [VouXLVni

assumed, in respect to these diflFerential characters, to be pure;
or, expressed in more technical language, each species is for the
characters under observation, genotypically homozygous. This
assumption is based on wholly inadequate evidence, as will be
shown later.

Citrus trifoliata crossed with other Citrus species (C auran-
tium, etc.) gave Fj hybrid families showing a large degree of
variability, even when the seeds from a single cross having
identical male and female parents were grown. This variability
expressed itself in foliage, habit of growth, and fruit, and was
especially noticeable in the latter, the fruits of the Fj individuals
showing differences in color, size, texture, shape, number of seeds,
and flavor. For example, from a single cross of C. trifoliata X C.
aurantium, the 11 resulting hybrid seeds gave rise to Fj plants
(citranges) differing in foliage, habit of growth, and very strik-
ingly in fruit. The fruit of one of these citranges, the "Morton,"
was smooth, round, very large, and orange-colored ; those of the
**Colman" were rather flattened, globose, pubescent, yellow, al-
most seedless, and lacked the disagreeable oil common to the
others; while those of still another type, the ** Willi ts," were
often monstrously fingered. The ** Phelps" was bitter, while the
''Saunders" almost lacked this quality. The "Rustic" often has
double fruits with many seeds, and a habit of growth more like
its aurantium parent.

When varieties of the lemon were crossed with C. trifoliata,
still greater differences in the F, generation (citremons) resulted.
These consisted largely of "abnormal" foliage developments.
Hypophylls, though absent in the common Citrus species are ex-
tremely characteristic of C. trifoliata. About 20 per cent, of the
lesnon-trifoliata hybridis developed an intensified form of this
character, and this proportion occurred in each case in crosses
involving three different varieties of lemon. The tangerine
orange X grape fruit (tangelo) in the F^ generation was almost
as variable as the citrange families. Fj hybrids between the
West Indian lime and the kumquat (limequat) were strikingly
different in such characters as aroma, flavor, acidity of pulp and
thickness of skin.

Although much stress has been laid on the differences in these
Fi hybrids, there were numerous similarities. For example, all
the Citrus hybrids involving C. trifoliata in their parentage have
compound, semi-evergreen leaves, increased hardiness and fruits

Digitized by

Google

No. 567] SHORTER ARTICLES AND CORRESPONDENCE 187

with abundant bitterish, acid juice. Two of the citranges (Col-
man and Cunningham) have the pubescent fruit character of
C. trifoliata, while the others are smooth-skinned.

The author's data led him to formulate in substance the follow-
ing conclusions, which I have grouped and stated in my own
language.

1. Citrus species are but slightly variable in the characters
which diflferentiate them, and, in the sense that no overlapping
takes place, may be said to breed true, their germ cells being
genetically pure for these differential characters.

2. Individual plants of the Fj hybrid generations between these
species are strikingly variable, although all are, in a given cross,
the zygotic product of pairs of gametes of ** identical hereditary
composition."

3. Modern theories of heredity can not account for this varia-
tion.

These are not the conclusions, however, in which all present-
day geneticists would! concur. In the first place, few ''modem"
geneticists would take Swingle's view concerning the **pure
breeding" ability of the various Citrus species, nor even of C.
aurantium. Webber, in the Encyclopedia of American Horti-
culture, notes that 70 varieties of the common sweet orange are
grown within our borders, and although a few varieties are
fairly constant, the majority of these do not breed true from seed.
Practically the same idea has been gained by certain prominent
taxonomists of the genus Citrus. De CandoUe specifically calls
attention to the remarkable variability of the whole group ; and
Professor Hume of Florida remarks on the same fact in certain
Experiment Station publications. As to the variability among
the individuals in the special strains used by Swingle in his breed-
ing work, no data are given, so that it can not be afSrmed that
inbred progeny from them would have been duplicates as far as
hereditary characters are concerned. Citrus plants naturally
cross fertilize, and from this cause alone no dependence can be
placed on their ability to produce progeny, which are exact dupli-
cates of themselves when inbred; in fact, the inference is that
they would not. Hence, as far as intraspecific constancy of
hereditary characters is concerned. Swingle's statement can not
be accepted until more exact information is produced.

Swingle says no interspecific gradations occur between these
various species, especially C. trifoliaia and C. aurantium. Qrant-

Digitized by

Google

188 THE AMERICAN NATURALIST [Vol. XLVUI

ing this, the two species have clearcut differences in leaves (ever-
green or deciduous, unifoliolate or compound), in resistance to
cold (difference in ability to withstand certain degrees of tem-
perature) and in numerous fruit characters (presence or absence
of pubescence, quality of juice, quantity of seed, size of fruit,
etc.).

From the standpoint of modem theories of heredity as regards
variation in Pi hybrid generations, it matters little whether so-
called species intergrade or whether their differences are clear-cut
and all variation is intraspecific. In either case, if crosses were
made, variation among the F^ individuals from a single family
might or might not occur. In either case, no violence to modem
theories of heredity would result and no new problems would
arise. But if two species that differ from each other in part or
all of their characters, but breed true intra-specifically (geno-
typically homozygous) are crossed, and Fi variation results, then
modem theories of heredity would be compelled to change front
and invoke the aid of new hypotheses. Swingle's data, assuming
that intraspecific variation in Citrus species occurs, does not
present a problem of this kind at all. C. aurantium and C. tri-
foliatu each possess distinctive characters, but convincing data are
not at hand to warrant any belief in the homozygosity of these
differential eharacters or of even those the two species may have
in common. The evidence directly, and one might almost say
conclusively, opposes such a conclusion. If these species are not
homozygous in all of their characters, then one can not affirm, in
the light of modem theories, that all the gametes produced by a
particular group of individuals called a species are identical in
hereditary composition, nor even that the gametes of one indi-
vidual of such a species are identical as to hereditary potenti-
alities. At the risk of wasting valuable space by repeating what
is extremely common knowledge to genetic students, let us assume,
for the purpose of argument, that C. (mrantium and C. irifolicUa
are homozygous in all their respective characters except one. In
the former, the character A is heterozygous and peculiar to this
species. Likewise, in C, trifoUata, B is heterozygous and differ-
ential. All the remaining^ characters of the two species may be
symbolized, respectively, by the formulaB XX and TT. When
XXAabb (C. aurantium) is crossed with YYaaBd (C. trifoltata)^
the resulting progeny would appear in the approximate propor-
tion of 1 XYAaBb : 1 XYAabl : 1 XYaaBb : IXYaabb, providing

Digitized by

Google

No. 567] NOTES AND LITEBATURE 189

A and B are single factor characters. In the majority of char-
acters, the Fi hybrids would be intermediate or possess those of
either one or the other parent, since all the F^ individuals would
be alike as far as any hereditary quality symbolized by XY is con-
cerned, providing the plants were all grown under the same en-
vironmental conditions. But these F^ individuals would not be
alike as regards the inheritance of the characters A and B. Ex-
perimental evidence from crosses of this kind show us that four
different Fj forms may result, the distinctions between them aris-
ing from the presence or absence, through inheritance, of the
characters A and B. Dominance is assumed to be absent in this
illustration.

Swingle's Citrus hybrids, though involving greater complexity
because a large number of parental characters instead of two are
probably heterozygous, are of the same general type as those of
the illustration and lend themselves to the same interpretation.
Owing to the absence of sufficient exact experimental data, one
can not speak of unit characters and factors in these hybrids, but
one may say without violence to modem theories of heredity that
one or both of the parents involved in the crosses which produced
the Colman and the Cunningham were heterozygous in the factors
or factor for pubescence, that various size factors were hetero-
zygous and that one parent was homozygous for absence and one
for presence of the factors for hardiness, compound leaves and
evergreen foliage.

Fi variation in Citrus hybrids then, in the light of the data at
hand, apparently results from differences in the gametic compo-
sition of the heterozygous parents.

Swingle calls attention to other cases of variation in F^ hy-
brids from two pure stocks which support his contention that this
phenomenon of Fi variation is very general, though usually
obscured through variation due to heterozygous parent stock.
CoUins and Kempton* crossed a race of com breeding true to
waxy endosperm with one constant for homy endosperm. Homy
endosperm was dominant in F^ and the Fj generation segregated
in the expected ratio of 1 waxy to 3 homy kemels. This ratio
represented the average proportion of each when the ears of all
the plants were lumped together. The F, progeny of each selfed

tColHos, G. N., and Kempton, J., n, 1912, ''Inheritance of Waxy Endo-
spenn in Hybrids of Ohinese Corn,'' IV** Conf. Internal, de Genetique, 1911,
p. 34T; alto Cir«. No. 120, Bur. of P. I., V. 8. Dept of Agr., 1913.

Digitized l5y

Google

190 THE AMERICAN NATURALIST [Vol. XLVHI

Fj plant when taken by itself gave some ears as low as 13.7 per
cent, waxy, while others exceeded the expected proportions and
gave ears as high as 33.3 per cent. waxy. The investigators point
out that this variation is not the result of the laws of chance as
the deviation is far greater in many cases than the probable error.
Therefore, says Swingle,

there can be no doubt but that their varying percentages represented
real differences in the hereditary composition of the first generation
plants. It would be hard to find a more conclusive case since there could
be no doubt as to the purity of the parents and what is more rare no
possible doubt as to whether a given kernel had a waxy or a homy
endosperm.

Mendelians are said to be unaware how fatal this phenomena is
to some of the chief tenets of modem theories of heredity, and
they are also accused, somewhat unjustly, I believe, of applying
the term ''imperfect dominance" to this and to the Citrus
phenomena.

In this case, both parents were undoubtedly homozygous for
their respective endosperm characters, so that heterozygosity will
not account satisfactorily for the deviations. But this is a dif-
ferent phenomena than Swingle found in his Citrus hybrids, for
here one is dealing with a fluctuation in a proportion or ratio
involving the same character, while in his experiments the diffi-
culty was the variation in presence and absence of distinct and
often new characters, indicating an extremely heterozygous
parentage.

As an explanation or working hypothesis for his own and
similar data. Swingle advances a somewhat new and suggestive
chromosome theory on the assumption that it fills an urgent need.
The theory of zygotaxis, as it is called, may be summarized as
follows :

Maternal and paternal chromosomes probably persist side by
side in the cells, unchanged in quality and number throughout
the whole development of the F^ organism. This being true,
Swingle, in order to explain his data, assumes that the influence
in character formation exerted by chromosomes on the F^ hybrids,
is in some cases due to their relative positions in the nucleus, and
that these relative positions result from accident or at least are
determined at the moment of nuclear fusion in fertilization, and
remain unchanged in succeeding cell generations. He further

Digitized by

Google

No. 567] NOTES AND LITERATURE 191

assumes that those chromosomes lying nearest the nuclear wall
(peripheral) are better nourished than those centrally located,
and hence they exert more influence in character formation, and
dominating synapsis, produce gametes similar in their hereditary
character to the cells of the first generation hybrids, whose char-
acter in turn was determined at fertilization by the configuration
the chromosomes took in the fusion nucleus. On this theory,
reversions, sports, etc., may result from sudden changes in the
nuclear configuration.

Three types of nuclear configuration are assumed to occur in
higher organisms, the character and effects of which are synop-
tically outlined below.

1. Interspecific Hybrids, — ^Usually sterile and intermediate.
Chromosomes repel each other and occupy opposite sides of the
Pi zygote nuclei, exerting equal influence in the ontogeny of P,
organisms, explaining why first generation hybrids of this char-
acter are always intermediate, little variable and usually sterile.
Synapsis often impossible.

2. Mendelian Crosses, — Abnormally inbred races of domesti-
cated animals and plants. Pj generation usually intermediate,
fertile, dialytic at synapsis. Dominance of certain characters in
these hybrids is due to the inherited potentialities of the chromo-
somes rather than to their nuclear positions.

3. Normal Cross-bred Species, — ^Probably normal in wild
species. Hybrids usually vigorous, fertile, and variable. Pree
intermingling of chromosomes in the fusion nucleus at fertiliza-
tion. Nuclear configuration permanent for each individual.
Synapsis normal.

This elaborate and attractive theory, based admittedly to a
great degree on assumptions, is advanced by Swingle in the belief
that it will help to clarify the problems of heredity, even though
he acknowledges it does not help one to arrive at satisfactory
explanations. In the reviewer's opinion, however, the field of
genetics is already burdened with enough theories of this par-
ticular type and the somewhat unnecessary but ever-increasing
new additions serve to confuse rather than clarify the ideas of
the average student of genetics. Besides, Swingle's assumption
that maternal and paternal chromosomes in the cells of Pj hybrids
repel each other and do not mingle in the Pi zygote cells is not
borne out by the few cytological facts at our command. Rosen-

Digitized by

Google

192 THE AMERICAN NATURALIST [Vol. XLVHI

berg V work on species hybrids of Drosera, Moeukhaus V investi-
gations of species hybricEs in fish and some work on certain
hybrids in the Echinodermata group give us facts that directly
oppose such an assumption. As a further criticism^ one may say
that most biologists who have had experience with pedigree cul-
tures would decidedly criticize the synoptic outline and the nar-
row sphere assigned to Mendelian phenomena.

Aside from the theoretical considerations, these two papers con-
tain djescriptions of CiirusAike species new to occidental horti-
culture, together with a somewhat detailed account of the various
Citrus hybrids and their hardiness and practical value, showing
the truly fine results achieved by the workers in this field toward
moving the Citrus belt northward and adding new varieties of
this genus to the world's horticulture.

Obland E. White

Bbookltn Botanic Oabden^
December 4, 1913

>Bosenberg, O., " Qytologische und Morphologische Studien an Drotera
Umgifolia X D. rotundifolia," Kungl. Svenska VetenskapBakademiens Hand-
linger., 43, N: on, pp. 1-64, 1909. 4 Tafn.

«Moenkliau8, W. J., "The Development of the Hybrids between Fwtduhu
heteroclitua and Menidia notaia with especial reference to the B^avior of
the Maternal and Paternal Chromatin," Amer, Jour, of Anatomy, 3: 29-65,
1904. Plates I-IV.

Digitized by

Google

The American Naturalist

MSS. intended for pobHcation and books, etc., intended for review sliould be
sent to the Editor of THE AiNERiCAN NATURALIST, Garrison-on-Hudson, New Yorlc.

Short articiee containing sunnarieo of research woric bearing on the
problens of organic evolution are otpecially welcome, and wiil be given preference
in pabllcatlon.

Ono hundred reprints of contributions are supplied to authors free of charge.
Farther reprints wiil be supplied at cost.

Sutoorlptioos and advertlsenents should be sent to the publishers. The
subscription price Is four dollars a year. Foreign postage Is fifty cents and
Caaadian postage twenty-live cents additional. The charge for single copies is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS
Lancaster, Pa. Qarrison, N. Y.

NEW YORK : Sub-Station 84

JCntercd ai Beeond-elaas matter, April 2, 1908, at the Post Office at Lancaster. Pa., under the Act ot

Ck>Dgres8 of March 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

>Vell Prepared Low Prtoes

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon.
7 or 8 volumes, please report, stating cash price, stat-
ing condition, binding and dates of volumes.

F. C. HARRIS,
Box 2244 Boston, Massachusetts

TO OOLOQISTS

and ENTOnOLOQISTS

W. F. H. ROSENBERG

Importer of Exotio Zoologioal Specimens
87. Haverstock HUl, London, N. W., Englaad

Begs to amiODDoe the publication of two new
Price lists : No. 18, Exotic Lepidoptera (over 8000
flpeoes), No. 19, Birds' Eggs (over 1100 species).

Thme wiU be mailed free on application, as well
mm aoj of the following lists:

Birdfikina (over 5000 species) ; Mammals (over
300 species); Septiles, Batraobians and Fishes (over
400 species).

Largest stook in the world of specimens in all
facmndiesof Zoology.

ALL MUSEUMS SHOULD
WRITE FOR THESE LISTS

All Specimens sent on approval.
flame stale which lists are required and give
mma of this periodical.

FINE LAND 5HELL5

I have In stock and for sale nearly ten thousand speciet
of land sbells. also a very sood stock of marine. All are ac-
eurately labeled. I am prepared to make up collections il-
lustrating every phase of the subject and from nearly all
conntries and islands. A fine illustrated booklet on the
beantiful shells of the Philippines, and other cireulars, will
be mailed free. Strictly scieDtific collections purchased for
cash. Correiipondence solicited with travelers and explorers.
Only the finest material handled. WALTEk P. WEBB,
202 Westminister Road. Rochester, N. Y., U. S. A.

The University of Chicago

Offers instruction during the Sum-
mer Quarter on the same basis as
during the other quarters of the
academic year.
L The undereraduate colleges, the
\ graduate schools, and the prafet-
I sional schoc^ provide courses in
9 Ari9, Litmratarm, Seimnemt
Conunmrcm and AaminUtra'
Hon, Law, Mmdieinm, £Jaca-
tion, andDMnHy. Instruction
is given by regular members of the
University Stan which is augmented
in the summer by appointment of
professors and iu^rocton from
other institutions.

Summer Quartar, 1914
Itt Term Jane 15-Jaly 22
' 2d Term Jahr 23-Aiis. 28

Detailed announcements will be
aem upon application.

The Unhrersity of Chkego

Mitchell Tower Oiicago, Illinoia

Digitized by

Google

The American Naturalist

with Special Rmfmrmmem to IIm Fmeton of Orgude ETolirtioo ud iUndOtw

OONTCNTS OF THE SCPTCMBER NUMBER
Thm Katona HUtorj of the Nine-lmnd«d Amadilto
otT«xM. Prolti8orH.H.Kew]ium.

GeiMtieal Btuditf on 0«iother». IV. Dr. BnAl%7
M.DaTit.

NotMon A Differential MortaUty obMrved between
TenebroolMciirifandT.nioUtor. Dr. Beta Aiken
Gortotr.

CONTENTS OF THE OCTOBCR NUMBCR

A OontribnUon towards an Analysis ef the Probmn
of Inbreeding. Dr. Raymond PeazL

The Inheritance of Coat Color in Hones. Professor
W. 8. Anderson.

The Variations in the Nnmberof Vertefacse and Ven-
tral Sontes in Two Snakes of the Qenos BeginiL
Professor Alexander G. Bnthren and dyftal

Shorter Articles and Beporti : The Stmnltaneooe

fessorB.A.JCmer8on. The Fourth Intematkmal
Genetie Conlerenee : Or. Frank M. SorfaM.

CONTENTS OF THE NOVEMBER NUMBER

The Effect on the OlTtpring of Intozieating ttie Mlile
Parent and the TranuniMion of the Defeeti to

with rcspeet to Seed Weight in theOerminaUon
of Garden Beam, Dr. r. Arthur Harrii.
Shelter Artldlet and Diicoision : Beeiprocal Crceiei
between Beeye't Pheasant and the Common
Ringneok Pheaiant prodndng Unlike Hybridi.
JohnCPhiUipe.

CONTENTS OF THE DCCCMBER NUMBCII

^ The Fixation of Character in Organisms. ByBdward
Sinnott.
Inheritance of Left-handedness. Professor Frauds

Supplementary Studies on the Differential Mortality
with Respect to Seed Weight in the OermlBatioa
ef Garden Beans, n. Dr. J. Arthur Harris.

Shorter Articles and Discussion: A Cross inToIring
Four Fain of Mendelian Charaeters in KiM. C
O.UUle.J.C.PhUUps.

Index to Volume XLVn.

OONTENTS OF THE «IANUARY NUMBER

Seleetton in Bxperimente with Tobaooo. Prou

fciMV BL M. SMt and H. K. Hayes.
Oynandxomorphout Anti, deicribed during the Dei*

cade. 1908-1918. Pxofesior WiUiam Morton

Wheeler.
Shorter Artielei and Diieaision: On the Beralti of

Inbreeding a Mendelian Popolatlon— A Corree-

tion and Bxtention of PreTions Condusione.

Dr. Raymond Pearl— Isolation and Beleetlon

alUed in Principle. Dr. John T. Qolick.

CONTENTS OF THE FEBRUARY NUMBER
SomeKewVarietiesofRatsandGnineapplgBandtbeIr

fessorW.B. Castle.
<• Dominant" and *' Recessire" Spotting in Mice. C.

C. Little. ^^
On Differential MorteUty with respect to Seed Weight

occurring in Field Cultures of Flsum satlTunu

Dr. J. Arthur Hairis.
The Inheritance of a Recurring Somatic Vailatiaa

in Variegated Ban of Maise. Professor B. A.

■merson.
Bestoration of Edaphosaurus crudger Copa. Pit^

f essor E. a Case.
Shorter Articles and* Discussion : Humidity ^a

Neglected Factor in Enylronmental Work. Dr.

Frank S.Lutz.

Single Number 40 CenU Yearly Subacriptiom $4i>0

THE SCIENCE PRESS

GerriBoii.N.Y. Substation 84: NEW YORK Lancaster. Pk.

Digitized by

Google

▼OL. XLYm, HO. 568 APBIL, 1914

THE

AMERICAN
NATURALIST

A XOHTHLT JOXnUTAL

Devoted to tbe Admioameiit of tbe Biologioal Soieneee with

Special Beferenee to tbe Faetors of Svolntioii

COHTEVTS

Pagt
I. The Oridn of X OapieUa Bnrta-paitorlt araolmoldM. Dr. Hbnbi Hub - Its
n. BltflogyofthoThsTtanoptera. n. Dr. A. Fbanklik Shxtll - - ase

m. Shortor Aztlclei and Dlioitision : Barriers as to Dutribation as regards Birds
and Mammals. Joseph Gbihnell. Yellow Varieties of Bats. Professor
W. E. Castlb -------.----247

ly. Hotel and Uteratnra : Woods on Heredity and the '* Influence of Monarchs."

THE SOnSNOE PRESS

LAVOASTEB, PA. GABBISOH, V. T«

HEW YORK : 8XTB-8TATI0N 84

Digitized by

Google

The American Naturalist

MSS. intended for publication and lioolcs, etc., intended for review should be
sent to the Editor of THE AMERICAN NATURALIST, Qarrison-on-Hudson, New Yorli.

Short articles eontaining sunnaries of researoh work bearing on the
problems of organio evolution are especially weieone, and will be given preferonoo
in publication.

One hundred reprints of contributions are supplied to authors ftet of oharge.
Further reprints will be supplied at cost.

Subscriptions and advertisements should be sent to the publishers. The
subscription price Is four dollars a year. Foreiqn postage Is fifty cents and
Canadian postage twenty-five cents additional. The oharge for single copies is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Qarrison, N. Y.

NEW YORK : Sub-Station 84

Bntered ai Beoond-clais matter, April 2, 1908, at the Post Office at Lancaiter, Pa., under the A«t ot

Congren of Maroh 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Ensrland

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon.
7 or 8 volumes, please report, stating casb price, stat-
ing condition, binding and dates of volumes.

P. C. HARRIS,
Box 2244 Boston, Massachusetts

TO OOLOQISTS

and ENTOnOLOQISTS

W. F. H. ROSENBERG

Importer of Exotic Zoological Specimens
57, Haverstock HUl, London, N. W., England

Begs to annonnoe the pablication of a new Price
List (No. 20) of Mammals, includlDg over 400 species
from various parts of the world.

These will be mailed free on application, as well
as any of the following lists:

BirdskiDS (over 5000 species) ; Birds' Eggs (over
1,100 species) ; Reptiles, Batrachiaus and Fishes
(over 400 species) ; Exotic Lepidoptera (over 8,000
species).

Largest stock in the world of specimens in all
branches of Zoology.

All Museums and Amateurs
should write for these Lists

All specimens sent on approval.
Please state which lists are required and give
name of this periodical.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up. Post

Cards 5 cents.

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU
89 West Bridge Street Oswego, N. Y.

The University of Chicago

Offen inatructkDdiirisg tiw Sias-
mer Quarter on the ame buit if
during the other qoMittn <d the
academic jeer.

The undmrmduate collcgea, the
graduate flchooU,jand the prale»-
sknal achocb proride courses is
Artm, Htmratnrmt ^
Cwnmmrcm and Ad
ftpfi* Law, Madieii

tiooB ana Mjioiaiiyt -

it given by regular mcmben of the
UniTerntygtafl which isaognmled
in the summer by appoimmest of
professars and instnicton bam
other institutioos.

Mitchell To\»ef

\ 1914
let Tom JwM 1 8-^lalF IS
2d Term lolv aS-^AiW. 28

Detailed announcemeiitB wfll \m
sent upon applicatioa.

Tbe University of Cliieege
CUcaco,

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVIII April, 29U No. 568

THE ORIGIN OF X CAPSELLA BURSA-PASTOEIS
ARACHNOIDEA

DR. HENRI HUS

University op Michigan

Since Jordan^ described a number of elementary spe-
cies of Capsella Bursa-pastoris, their constancy has been
a subject of cultural experiment. Herbarium material
demonstrates the existence of numerous apparently imde-
scribed forms. The finding of strikingly distinct forms,
such as Capsella Heegeri^ and, more recently, C. Viguieri,^
the work of Almquist* and that of ShuU have added to
the interest which this species holds for the investigator.
It was ShuU who determined the zygotic constitution of
various forms. To be able to demonstrate this with ex-
actitude is of the greatest value since Bateson and Lotsy
expressed their doubt as to the homozygocity of deVries's
(Enothera Lamarckiana. It was left to Nilsson^ to clearly
show its necessarily heterozygous character. The inter-
est aroused by this paper® leads me to believe that an

1 Jordan, A., ' ^ Diagnoses d 'esp^ces nouvelles ou m^connues pour servir
de mat^riaux k une ilore r^form^e de la France et des contr^s voisines."
Paris, 1864.

2 Solms-Laubach, H. Graf zu, ''Craciferen studien. I. Capsella heegeri,
eine nenentstandene Form der deutschen Flora/' Bot, Zeit., 55: 167, pL
7, 1900.

8 Blaringhem, L., ''Les transformations brusques des §tres vivants."
Paris, 1911.

^Almquist, E., "Studien fiber die Capsella Bursa-pastoris (L.)," Ada
Horii Bergiani, 4: No. 6, 1907.

s Heribert-Nilsson, N., * ' Die Variabilitat der (Enothera Lamarckiana und
dai Problem der Mutation," Zeitschr. /. ind, Abst, u. Vererb,, 8: 89, 1912.

• Lotsy, J. P., ' * Fortschritte unserer Anschauungen fiber Deszendenz seit
Darwin und der jetzige Btandpunkt der Frage," Progressus Bei BotaniccB,
4:361,1913.

198

Digitized by

Google

194 THE AMERICAN NATURALIST [Vol. XLVUI

account of certain cultures of Capsella, in which muta-
tions were simulated, would be of timely interest.

During the winter of 1908-1909, I collected in a green-
house at Ann Arbor, Michigan, and at the disposal of the
Botanical Department of the University of Michigan,
twelve rosets of Capsella Bursa-pastoris, the leaves of
whic'h showed certain more or less striking morphological
differences. With the hope of isolating certain biotypes,
the rosets were placed in pots and permitted to flower.
No measures were taken to prevent the accidental trans-
ference of pollen, but the pots were placed about six
inches apart. This, as will be shown later, is the only
precaution necessary to guard against cross-pollination,
provided the cultures are carried on in a greenhouse and
during the winter months. After a portion of the seed
had ripened, the plants, the majority of which retained
their climax leaves, became herbarium specimens. More
recently, after constant association has enabled me to
detect minute differences, it has been possible to identify
some of these plants with two of the biotypes described
by ShuU,^ to wit, rhomhoidea and simplex. At the time
of collection, the differences were sensed, but could not be
described technically, since the extent of the influence
wielded by fluctuating variability was an unknown quan-
tity. Never before had I so fully realized the truth of de
Vries's statement.®

We are trained to the appreciation of the differentiating marks of
systematic species. . . . Our minds are turned from the delicately
shaded features which differentiate elementary species.

The seed obtained was sown in sterilized soil during
the spring of 1910. From each seedpan 60 individuals
were transplanted to flats. As the plants grew older, it
was found that, with a single exception, the seedlings in
each of the flats were uniform, but that the seedlings in
the different flats were not alike, three types being dis-
tinguishable. The interest in these types, for the isola-

TShull, G. H., ** Bursa huraa-pastoris and Bursa Heegeri: Biot.ypes and
Hybrids,'' Publ. No. 112, Carnegie Institution of Washington, 1909.
8 de Vries, Hugo, *' Species and Varieties,'' 689, 1905.

Digitizei

Google

No. 568] X CAP SELLA BUBSA-PASTORIS ARACHNOIDEA 196

tion of which these cultures had been undertaken, soon
was overshadowed by the behavior of the seedlings bear-
ing the number 4,108.6 and which were the offspring
yielded by a plant of a type not described by Shull and
which I have named X Capsella Bursa-pastoris Setchelli-
ana, in honor of Professor William Albert Setchell.

Fio. 1. Appearance op a Linear-leaved Form among Seedlings op Capsella

Bursa-pastoHs.

During the time that the seedlings remained in the seed-
pan, no deviations from the expected course of develop-
ment were noted. However, after the seedlings had been
transplanted to flats and had remained there a week or
so, it became evident that some of the seedlings were not
making the expected growth. Their development ap-
peared most insignificant compared with that of the
majority. A closer examination showed the cotyledons
to be somewhat larger than normal and the leaves proper
to be exceedingly small and almost linear. Nor did they
attain the same length as the leaves of the rosets belong-
ing to other types.

Digitized by

Google

196 THE AMERICAN NATURALIST [Vol. XLVIII

An explanation of this peculiar development was songht
in a possible attack on the part of either fungi or bacteria
or in soil conditions. But the latter were uniform for the
entire flat. Neither fungi nor bacteria could be demon-
strated nor did the underground portion of the '* ab-
normal" plants look imhealthy or underdeveloped.

Fig. 2. Seedlixus ok x C. . . . SctcheUi and x C. , . . arachnoidef

At this stage the flat presented the appearance shown
in Fig. 1. At the time but three types were distinguished,
the first of these constituted by plants which showed an
incision of the blade, the second composed of those which
apparently had entire leaves, and a third, comprising the
small and linear-leaved rosets, which, because of the spider-
like appearance of the latter, has been designated xCap-
sella Bur sa-past oris arachnoidea. There also appeared a

Digitized by

Google

No. 568] X CAP8ELLA BURSA-F AST ORIS ARACHNOIDEA 197

single individual which, while closely resembling the form
arachnoidea, differed from it in having somewhat spatu-
late leaves. This plant, a plant of arachnoidea and two
of Setchelliana, are shown in Fig. 2.

After photographs had been taken, the plants were
potted and placed in the frames. None of the plants
made a growth as vigorous as that of the Capsellas grow-
ing in the open. The plants of the form arachnoidea de-
veloped leaves with a greatest length of 15 mm. and a
greatest width of a little over 1 mm., causing the plant to
retain its spider-like appearance. The roset with spatu-
late leaves appeared somewhat more vigorous, the aver-
age leaf measuring 22 mm. in length, with a greatest width
of 2.5 mm. In later generations I have been able to ob-
tain rosets of arachnoidea with a greatest leaf-length of
100 mm. and a greatest width of 6 mm.

In the frames, flowering shoots made their appearance,
those on arachnoidea being remarkable chiefly because of
their small size, reaching a length not exceeding 12 cm.
The flowers were small but well-formed. No well-devel-
oped pollen could be demonstrated. Seed did not form
and the capsules retained their original form, typical of
non-fertile capsules in Capsella Bursa-pastoris, remind-
ing one of the capsules of Capsella Heegeri. They do not
resemble the fertile capsules of C. procumbens. In the
next generation I saw a single capsule formed on arach-
noidea as the result of cross-fertilization, and in this case
it differed in no manner from the normal capsule such as
we know it in Capsella Bursa-pastoris.

The ''normal" plants, i. e., all those not belonging to
the form arachnoidea, matured a large amount of seed.
No measures were taken to prevent cross-pollination, but
no other plant of Capsella Bursa-pastoris, within a radius
of twenty feet, was in flower.

At this time, another attempt was made to group the
plants. It was foimd that the criterion used earlier, i. e.,
the incision of the blade, no longer could be relied upon,
since plants, which at the time of the previous coimt, had
shown an entire margin, now were more or less incised.

Digitized by

Google

198 THE AMEBIC AN NATURALIST [Vol. XLVIII

Unfortunately, after the seed had been collected, the plants
were destroyed, having lost their climax leaves. An attempt
to group them later with the aid of photographs failed, be-
cause photographs of all plants were taken during the
earlier stages only, i. e., before the appearance of the cli-
max leaves. Another classification, for which climax leaves
are not . essential, and which is based upon the relative
width of the first six or eight leaves, yields for 54 plants
the proportion: ''wide" 31, ''narrow" 16, "linear" 7, the
ideal proportion, as since worked out, being 33 : 16 : 16.
The fact that the number for "linear," which represents
the form arachnoidea, is too small by 9, may be ascribed
to various circumstances, among others the fact that the
last row in the flat did not appear in the photograph upon
which the count was based. It is in the last row of a flat
one ordinarily meets with the smaller or at least less vig-
orous individuals and it is very probable that in this last
row occurred a large percentage of individuals belonging
to arachnoidea. Furthermore, not all the seedlings, but
only sixty, were taken in each case. Almost unconsciously
one selects the largest individuals when transplanting
from seedpan to flat. It is probable that in this process
there were eliminated a greater percentage of seedlings
of the linear form than of any of the others. Hence no
great weight can be attached to the proportion obtained.
The collection of seed brought the work for 1910 to a
close. As far as I was aware, no forms similar to arach-
noidea had been either noted or described by any one who
had devoted his time to culture experiments with Cap-
sella. Neither Shull in America, nor Almquist in Swe-
den, nor Lotsy® in Holland, has made mention of such
forms in their publications. The fact that no seed was
produced by the aberrant form seemed to hold out little
hope for the continuation of the cultures, and the sole
trace left by this new form, if taxonomic form it was,
threatened to consist of but a few photographs and some
alcohol specimens. A single possibility presented itself.

»Lotsy, J. P., * * Vorlesungen iiber Deszendenztheorien, ' ' 1: 180, Jena,
1906.

Digitized by

Google

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 199

Whether the parent plant was of a hybrid character or
whether the parent plant was mutating, and the new form
or forms were to be looked upon as mutants, in either
case there existed the possibility, if not the probability,
that from the seeds obtained from those plants of the
second generation which appeared ''normal," a third gen-
eration might be obtained which would again present the
abnormal form. Such indeed proved to be the case.

Fig. 3. Early Stages in the De>*eia>pmext of Broad-i.eaved, Narrow-leaved

AND Li NEAR- LEAVED FORMS OF Capftcllu.

The seed for the next generation was obtained from 19
plants. The seed was sown separately in pots of steril-
ized soil. Certain of the parent plants, which we now
identify with ShuU's simplex and rhomboidea, produced
a uniform, broad-leaved offspring. Others behaved like
the parent, the form arachnoidea appearing in 197 indi-
viduals out of a total of 979, which does not include the
713 which bred true to the broad type. (For an illustra-
tion of these types see Fig. 3.)

It is unnecessary to go into details as to the various
theories which suggested themselves as a solution of the

Digitized by

Google

200 THE AMERICAN NATURALIST [Vol. XLVHI

origin of the linear-leaved form which, becatise of its
striking appearance, concentrated the attention upon it-
self. That perhaps we were dealing with a mutation was
a thought which most naturally obtruded itself upon
the mind of one who, for years, had fruitlessly tested
a large number of species in the hope of discovering a
case analogous to that of (Enothera Lamarchiana}^ The
possibility of a cross between a local form and either
Capsella Heegeri or C. procumbens, suggested itself.
However, the seedling stage of either of these two forms
does not bear the remotest resemblance to that of Cap-
sella arachnoidea. At the same time there was slight
reason for believing that either Capsella Heegeri or Cap-
sella procumbens ever had been grown in Ann Arbor.

During 1911 and the greater part of 1912, the problem
rested here, no satisfactory explanation being found.
But pedigree cultures were continued until, on the one
hand, we succeeded in placing the plants in optimum sur-
roundings for the production of climax leaves, and on the
other began to distinguish between the various biotypes.

The Biotypes

As has been noted previously, it was possible to use two
criteria for the classification of the rosets. Leaving out
of consideration the rosets of the linear-leaved arach-
noidea, it was f oimd that after dividing the rosets accord-
ing to the ''broad" or ''narrow" character of the earlier
leaves (Big. 3), it was possible to further subdivide each
group on the basis of the marginal indentation of the
leaves subsequently formed.

I. The ^' Broad'' Group. — Here the first four or five
leaves possess a blade which is approximately twice as
long as broad. Up to this stage the margin remains
entire. When the sixth leaf appears one ordinarily can
begin to distinguish between two types. These are :

Type 1. — In this, the first of the two broad-leaved
forms, the margin of the first eight leaves remains entire,

10 Hub, H., ''The Origin of Species in Nature," American Naturalist,
45: 646, Nov., 1911.

Digitized by

Google

No. 568] X CAPSELLA BURS A-P AST OEIS ARACHNOIDEA 201

showing at most a very slight crenation (Fig. 4). Usually
the ninth leaf, though sometimes it is the eighth and some-
times the tenth, shows a more marked indentation, though
seldom of a depth of more than 2 mm. on each side of
the leaf and slightly below the middle. Subsequent
leaves show an increase in the number and depth of the

Fig. 4. Dissection op Young Rosets of C. . . . simpler and C. . . . rhoin-

boidea. Showing the " Broad " Character op the Earlier Leaves

AND the Distinctive Character op the First Sinus.

indentations, the maximum for both being reached in the
climax leaves, which usually show five indentations reach-
ing about midway from margin to midrib. In those of
the earlier leaves which show a marked incision the lobes
are obtuse. In the later leaves the lobes become acute.
It may be stated as a general truth, that an increase in
the depth of the sinus carries with it an increase in sharp-
ness of the lobe. There is no secondary lobing, but some-
times the margin of the sinus shows a slight denticulation.
While in the earlier leaves the sinuses separating the
terminal lobe from the rest of the blade are the deepest,

Digitized by

Google

202 THE AMERICAN NATURALIST [Vol. XLVHI

the converse is true in the later leaves, where the sinuses
separating the terminal lobe are the most shallow. I
have identified this form with Shull's simplex}^ My
plants also agree fairly well with the illustration of onto-
genetic succession of leaf forms in Bursa . . . simplex,
shown by ShuU.^^

Type 2. — In the second of the two forms distinguished
because of the greater relative width of their first leaves,
the margin of the first five leaves remains entire, as in
the case of those of type 1 {simplex). The sixth leaf,
however, ordinarily shows a marked indentation, at least
3 mm. deep and slightly below the middle of the blade
(Fig. 4). This indentation may appear in one margin or
in both. The lower margin of the sinus ordinarily is at
right angles to the midrib, the upper margin making an
angle of 45 degrees with the midrib ( Fig.7, h ) . Even when
it has become difficult to distinguish between types on the
basis of relative width of the earlier roset leaves, it always
is possible to distinguish between tjT)e 2 (rhotnboidea)
and type 4 {Setchelliana and Treleaseana)^ by means of
the character of the sinus. In type 4, the lower margin
of the sinus makes an angle of 45 degrees with the midrib,
while the upper margin makes an angle of between 30
and 45 degrees with the midrib. Hence the first sinus in
C. . . . Setchelliana and C . . . Treleaseana is at least
90 degrees, while the first sinus in rhomboidea measures
seldom more than 45 degrees and frequently less.

The seventh leaf of plants belonging to type 2 ordi-
narily shows two indentations on both sides of the leaf,
dividing the blade into a lower portion, two central lobes
and a terminal lobe. The depth of the incision amounts
to about three-fourths of the width of the blade from mid-
rib to margin.

It is possible to delay the appearance of the first inden-
tations by transplanting from seedpan to flat either too
early or too late. In such cases, the indentations appear
in the seventh leaf only, or even later, and are rather

11 Loc. city 25, and PI. 2, Fig. 2.

"Shun, G. H., **Verh. d. naturf. Ver. in Bninn,'' 49, PI. 4, 191L

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST ORIS ARACHNOIDEA 203

shallow, reaching a depth of three-fourths of the width of
the leaf from margin to midrib in the eighth, ninth or
tenth leaf. However, once the indentations have made
their appearance, the leaf next produced ordinarily shows
two sinuses on both sides of the blade, usually the upper
set, rarely the lower, being the deeper of the two, and
almost reaching the midrib. The succeeding leaves show
an increase in the number of lateral lobes from two to
six. Since the incisions almost, if not quite, reach the
midrib, both lateral lobes and the terminal lobes are well
defined. Upon the lateral lobes secondary lobes appear,
both on the distal and proximal margins. It is to be
noted that only the climax leaves of well-grown specimens
of the homozygotic form distinctly show the lobing of the
proximal margin and this only on the middle lobes. The
lobing of the primary lobes results in the setting off of
a small terminal portion of each lateral lobe, which
possesses a more or less rhomboidal form. This terminal
lobe of the primary lobe can be observed to advantage
only in the climax leaves of well-developed specimens.

I have no hesitation in identifying type 2 with ShuU's
rhomboidea.^^

Capsella Bursa-pastoris simplex and C. Bursa-pastoris
rhomhoidea, described, respectively, as types 1 and 2,
agree in having the first five or six leaves twice as long as
broad, thus contrasting sharply with the plants to be de-
scribed under types 3 and 4, which constitute the *' nar-
row" group.

II. The ''Narrow'' Group. — In the plants belonging
here, the first five or six leaves possess a blade which is
from 2^ to 3 times as long as broad. Usually after the
appearance of the seventh leaf, sometimes not until the
appearance of the tenth leaf, it is possible, on the basis of
marginal indentation, to separate the plants with ** nar-
row'' roset-leaves into two groups, designated respec-
tively types 3 and 4.

Type 3. — Rosets of plants belonging to type 3 can not
be distinguished from those of iype 4, until after the

"ShuU, Verb., PL 2j Biotypes, PI. 1, Fig. 2.

Digitized by

Google

204 THE AMERICAN NATURALIST [Vol. XLVHI

seventh leaf has appeared (Fig. 5). It is to be noted that
for the first six leaves of type 4, the ratio between mean
length and width is 6:2, while for the corresponding
leaves of type 3, the same ratio is 5 :2. Once the seventh

Fig. 5. Dissection op Young Rosets of x C. . , . Setchelli and x C. . . .

attenuata^ Showing the " Nabrow " Character op the Earlier

Leaves and the Distinctive Character op the First Sinus.

leaf has appeared, a distinction readily can be made, since
in type 3, no sinuses appear, and the leaves, from the
seventh to the tenth, might be mistaken for those of
simplex (Fig. 5). Later leaves readily can be distin-
guished from those of simplex, by the pointed apex, the
very shallow sinuses, ending in a sharp tooth, and by the
fact that the greatest width of the blade lies above the
middle, about one third the length from the tip (Fig. 6).

This form, which because of its morphological charac-
ters on the one hand, and its behavior in breeding on the
other, can readily be distinguished from all others, I
designate X Capsella Bursa-pastoris attenuata.

Type 4. — Not only do the first leaves of plants, belong-
ing to this type, differ in relative width from the first
leaves of plants of rhomboidea and simplex, but there also

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST ORIS ARACHNOIDEA 205

is a difference in the apex of the leaf, the apices of leaves
of this type, like those of type 3, being decidedly pointed,
while those of types 1 and 2 are rounded.^*

At the sixth or seventh
leaf stage, the marginal
indentations make their
appearance, at first as
slight crenations, then as
long and shallow sinuses,
and finally, in the eighth
or ninth leaf, as a sinus
on one or both sides of
the midrib and about the
middle of the blade (Fig.
5 ) . The lower margin of
the first sinus ordinarily
makes an angle of 45
degrees with the midrib,
while the upper margin
makes an angle of from
30 to 45 degrees with the
midrib. This renders the
first sinus ordinarily
greater than 90 degrees
(Fig. 7, a). The depth
of the first sinus is ap-
proximately one half the
distance from margin to

midrib. In subsequent leaves the depth increases, so that
in the 11th leaf the sinuses almost reach the midrib. In
Treleasi, one of the two forms, which together constitute
type 4, the climax leaves show incisions to the midrib, and
a well-marked terminal lobe, while in the other the sinuses
are less deep but the terminal lobe still is well marked
(Fig. 8). The number of sinuses increases in propor-

^* It is to be noted that in mj cultures there appear, from time to time,
plants of rhomhoidea of which the leaves have sharply pointed lobes. What
relation these plants bear to others classed with them under rhomhoidea, I
am at present unable to say.

Fig. 6. Lateb Roset Leaves op w C,
, . attenuata and C. . . . simplex.

Digitized by

Google

206

THE AMERICAN NATURALIST [Vol. XLVIII

tion to their depth. If the seventh leaf has one sinus
in each margin, the eighth and ninth usually have two,
the tenth and eleventh, three, and so on, until the mean
of six is reached. As the lobes increase in number, they

Fio. 7. Early Roset Leaves of x
C. . . . Svtchein AND C. . . . rhom-
hoidea.

Fig. 8. Climax Leaves of x C.
. . SetchelH and x C. . . . Treleasi.

not only become narrower but the sinuses do likewise.
This is the result of a gradual increase in the angle
between the lower margin of the sinus and the midrib.
In the eighth leaf the lower margin forms an angle of
about 90 degrees with the midrib, causing the formation
of a primary lobe, triangular in shape and with an upper
angle of about 45 degrees, instead of the 90-degree angle
found in the first lobe. In older leaves the angle between
lower margin of sinus and midrib may increase to 110 or
even 120 degrees. The climax leaves therefore get to
resemble more and more those of rhomboidea, especially
since the distal margin of the sinus, from the tenth leaf
on, exhibits a number of denticulations which, in older
leaves, especially of one of the forms {Treleaseana)^
tend to become incisions, so that secondary lobes are

Digitized by

Google

No. 568] X CAPSELLA BVRSA-P AST ORIS ARACHNOIDEA 207

formed. However, the end of the lobes of early
leaves of type 4 always are sharply pointed (Fig. 9),
while the lobes of early leaves of rhomboidea are ordi-
narily rounded at the ends (Fig. 4).

Fig. 9. Upper Row : 3 Sets op Leav-es from as Many Plants of jr C. . . .

TreleasL Lower Row : 4 Sets of Leaves from as Many

PukNTS OF X C. . . . Setchclli.

From a morphological point of view these leaves are
entirely different from any form described by ShuU, the
differences being most marked and very readily recog-
nized once our attention has been called to them. But it
is especially the behavior of the plants on breeding which
leads me to recognize them as most distinct hybrid forms
and which I have designated X Capsella Bursa-pastoris
Setchelliana in honor of Professor William Albert
Setchell, and X Capsella Bursa-pastoris Treleaseana, in
honor of Professor William Trelease.

Type 5. — Capsella Bursa-pastoris arachnoidea. This
form, which readily is recognized from the first by its
linear leaves, does not require an elaborate description
at present, since it will be discussed in detail later. It has
been illustrated in Figs. 1, 2 and 3.

Digitized by

Google

208

THE AMERICAN NATURALIST [Vol. XLVHI

The above descriptions apply only to plants grown
under fairly uniform conditions, in a light soil in a green-
house, and treated in such a manner as to oflfer the plant
the most favorable conditions for development. By leav-
ing the plants too long in the flats, so that crowding re-
sults, by keeping them too moist and warm, etc., it is
possible to produce abnormal climax leaves in which the
typical diflferences can be recognized with difficulty only.
By leaving plants too long in the seedpans, by keeping
them too dry, it may be brought about that plants flower
without having produced climax leaves. There will be
doubtless many who, because of this, will refuse recogni-
tion to the segregates just described. **Qu8ecunque dixi,
si placuerint, dictavit auditor." Fortunately, the differ-
ences of behavior on breeding are such, we must recognize
their distinct genotypic constitution.

Genotypic Constitutions
ShuU, in the papers above quoted, made one of the most
important of recent contributions to science, since he de-
termined with exactitude the relations existing between
some of the lesser forms which, because of their alleged
constancy or inconstancy, have been a bone of contention
since the days of Jacquin. Making extensive cultures of
Capsella, Shull was able to distinguish four forms (Fig.
10), to wit, heteris, with leaves divided to the midrib, with

Fig. 10.

CL13IAX Leaves of C. . . .
AND C.

heteris, C. .
. . . simplex.

tenuis

rhomboidca

Digitized by

Google

No. 5d8] X CAPSELLA BURSA-PASTORIS ARACUNOIDEA 209

elongated primary lobes, a marked secondary lobe, in the
distal axil of the primary lobe and a well-marked terminal
lobe ; rhomboidea, with leaves divided to the midrib, with
an imelongated primary lobe, with an incision in the distal
margin setting oflf a secondary lobe and a corresponding
incision on the proximal margin of the primary lobe, set-
ting off, in well-grown specimens, a terminal portion of
each lateral lobe, generally of rhomboidal form; tenuis,
with the elongated primary lobe of heteris, but with a
sinus which usually does not reach the midrib, terminal
lobe clear cut; simplex, with lateral lobes obtuse, never
attenuated, the incisions being shallow and never reach-
ing the midrib.

Shull recognized here the presence and absence of two
factors, one (A) responsible for the sharp primary lobe
of heteris and the attenuation of the lobes in tenuis, while
the other (B) is responsible for the division of the leaf
to the midrib, the definite terminal lobe and the second-
ary lobes. On this basis Shull was able to represent the
biotypes by conventional Mendelian symbols, thus:
heteris, AB; rhomhoidea, aB; tenuis. Ah; simplex, ah.

That this conventional presentation gives us a reliable
working basis, my experiments have shown most satis-
factorily. With the aid of these symbols I have been able
to solve the origin of Capsella arachnoidea, the experi-
ments showing that, without question, forms presenting
the spider-like appearance of the rosets typical of this
plant are of hybrid origin.

The Zygotic Constitution of 4,108.6
The problem to be solved was that of the zygotic con-
stitution of the original parent, the plant which in my
notes is recorded as 4,108.6. Among its offspring neither
heteris nor tenuis made their appearance, while both
rhomhoidea (aB) ajid simplex (ah) were met with. Hence
the parent was homozygotic for (a), but heterozygotic
for (B). Therefore, its zygotic constitution, in part, must
have been aaBh.

Besides rhomhoidea and simplex there appeared two

Digitized by

; Google

210

THE AMERICAN NATURALIST [Vol. XLVHI

forms, referred to as types 3 and 4, the latter being ca-
pable of further subdivision. Neither of these was de-
scribed by Shull. At least one difference between rhom-
boidea and simplex, on the one hand, and types 3 and 4, on
the other, could be noted at once, i. e., the relative width
of the leaf. As has been shown above, the former have
their first leaves twice as long as broad, the latter three
times as long as broad. The idea suggested itself that
there might exist a factor which determined these charac-
ters. Since the original parent belonged to type 4, the
narrow character of the earlier leaves must be dominant
over the broad character. Also, since the original parent
produced both ** narrow ^^ and ** broad '^ types, it must
have been heterozygotic for this character. Using (N)
to indicate the gene, we get for the zygotic construction
of the parent plant oaBbNn.

aBN aBn

abN

abn

aBN

aBn

abN

abn

1
aBN
aBN

2
aBn
aBN

3
abN
aBN

4
abn
aBN

j aBN
aBn

aBN
, abN

aBn
aBn

7
abN
aBn

abn

aBn 1

10
aBn
abN

11
abN
abN

abn
abN

aBN

abn

aBn
abn

abN
<d>n

16 1
abn

abn 1

Pig. 11.

Diagram to Illustrate the Nature op the Offspring or m C. .
Setchelli (aaBhNn).

Since self-fertilization is the rule in Capsella, it was an
easy matter to test the validity of the theory. A form
aaBbNn, one with unelongated primary lobes, sinuses
reaching the midrib and with early leaves of a *^ narrow *'
type should yield, on self-fertilization, the following com-
binations: l.bbnn (square 16), a plant of which, accord-
ing to our definition, the earlier roset leaves should be

Digitized by

Google

No. 568] X CAPSELLA BUB8A-P AST ORIS ARACHNOIDEA 211

broad and of which the later leaves shall lack incisions
reaching to the midrib, a plant, in short, which should
have all the characteristics of Shull's simplex. Further-
more, on being self ed, it should yield a uniform offspring,
in all respects resembling the parent.

Such plants actually were encountered. Of the plants
grown to maturity, twelve were selected as seed-bearers.
All bore the simplex character. Ten of these plants were
selected from among the first generation of plants of the
supposed zygotic constitution BbNn, while one parent
(yielding No. 25,712) was derived from a plant bearing
the simplex character and another (yielding No. 31,112)
was derived from a plant which was shown to have the
zygotic constitution bbNn.

TABLE I
EviBENcx or HoMOZTOOTic Characizb of Simplex (hhnn)

Index
Nomber

Number of
Plants

Charmoter of

Index Number of Parent

Parent

Grand-
parent

25.712

G»»

S,n2BRl2P9

bbnn

26.912

S,212BRSPI

bbnn

BbNn

26.312

S,212CR5PI

bbnn

BbNn

26.612

187

S,212FRSPS

bbnn

BbNn

26.712

180

S,212HR7P7

bbnn

BbNn

30.012

276

8,212C/25P1

bbnn

BbNn

30.112

108

8,212^i22P6

bbnn

BbNn

30.212

8,212Gi26P8

bbnn

BbNn

30.312

162

8.2120/26P8

bbnn

BbNn

30.712

S,212HRSP6

bbnn

BbNn

31.112

26.012A/27P6

bbnn

bbNn

3.113

207

26M2BRIPS

bbnn

BbNn

1.399

This table offers an excellent illustration of the small
danger of an accidental cross, even if the plants are not
guarded, always, of course, when the proper precautions,
indicated above, are taken. Numbers 26,312 and 30,012,
as well as numbers 30,212 and 30,312, respectively, offer
instances of uniform inheritance in plants possessing
recessive characters only and of which the parents in the
one case were left tmguarded, in the other caged. Had

19 In tliis colmnn "G" indicates that the parent plant was guarded,
"O" that the plant was open-fertilized. In other tables the same abbre-
viation will be used.

Digitized by

Google

21 2 THE AMERICAN NATUBAUST [Vol. XLVIH

crossing taken place in the case of the nnguarded flowers,
this wonld^ because of the purely recessive characters
possessed by simplex, have become apparent at once. In
all cases the parents were checked by means of herbarium
specimens or photographs, or both.

2. hhNN (square 11). According to our hypothesis, a
plant of this zygotic construction should have the earlier
roset leaves narrow and the climax leaves should lack
incisions to the midrib. It also should breed true. A
plant fulfilling these conditions has not been encountered,
or rather, its recognition was delayed until the offspring
of the corresponding heterozygote hhNn could be observed.
As will be shown, the zygotic combination bbNN yields
a plant with the external characteristics of arachnoidea.

3. bbNn (squares 12 and 15). A plant of this zygotic
constitution should have narrow early leaves and the
climax leaves should lack incisions to the midrib. On
self-fertilization it should yield 25 per cent. bbNN, 50
per cent. bbNn and 25 per cent. bbnn.

bN 1 bn
bN 1 bN

bN bn
bn bn

Several plants were foimd which fulfilled the require-
ments as to leaf characters. Such plants, on being selfed,
yielded approximately 25 per cent, simplex, which we
know to have the zygotic constitution bbnn, while about
50 per cent, bore the parental characters, supposedly rep-
resented by bbNn. The remaining 25 per cent, clearly
belonged to the type arachnoidea. In all, 12 plants were
selected as seed-bearers, some being guarded, others re-
maining uncaged. The results are given in Table II.

The totals closely approximate the Mendelian ratio,
yielding, respectively, bbNN 24 per cent., bbNn 49 percent,
and bbnn 27 per cent. Having established the identity
of bbnn (simplex) and bbNn {attenuata)j we are forced
to recognize bbNN as the zygotic construction of arach-
noidea. It would be a comparatively easy matter to test

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST ORIS ARACHNOIDEA 213

this directly, prorided the form arachnoidea produced
seed. Though I have grown several hundreds of these
plants, I have obtained in all but eight seeds, and these as
the result of hybridization. Hence the test must be made
indirectly through crossing of forms yielding the desired
gametic combinations.

TABLE II
EviDBNci or Heteroztqotic Chakactie of attenuata (hbNn)

Number of Plants

Index Number
of Parent

Char,
of

O
or

Index

^yN

bblfh

bbnn

Char,
of

Num-

Grand-

ber

a
1

Parent

parent

26,012

25.50

51.00

25.50

S,212BR5PI

66Arn

BbNn

26,412

7.50

15.00

7.50

S,212DR10PS

bbNn

BbNn

31,212

14.75

29.50

14.75

2Q,012AR2PI

bbNn

31.312

15.00

30.00

15.00

26,012AR\P2

bbNn ' 0

bbNn

31,412

14.25

28.50

14.25

26.012il«lP3

bbNn 0

bbNn

31.512

25.50

51.00

25.50

26,012^«1P4

bbNn \ 0

bbNn

31.612

2.50

5.00

2.50

26,012^ie6P6

bbNn 0

bbNn

31.812

28.25

56.50

28.25

26,012J5«1P3

bbNn 0

bbNn

31.912

16..50

33.00

16.50

26,012B«1P6

bbNn 0

bbNn

3.213

64.25

130

128.50

64.25

26,912D/K2P4

bbNn 0

BbNn

3.313

61.50

117

123.00

61.50

26,912Jg:/?6P6

bbNn 0

BbNn

3.513

35.50

71.00

35.50

26,912/?iJ6P4

bbNn 1 0

BbNn

Total .

296

308.50

604

617.00

334

308.50

Of the twelve parent plants concerned in the above ex-
periment, five were selected from among the first genera-
tion of a plant having the supposed zj^gotic constitution
BbNn, while seven were the direct offspring of No. 26,012,
which had been shown to yield the three forms, arach-
noidea, attenuata and simplex, as indicated in Table II.

Tlhe simplex, obtained by selfing a plant of bhNn, breeds
true, as indicated in Table I, No. 31,112, a simplex, yield-
ing a uniform simplex offspring, consisting of 50 indi-
viduals.

4. BBnn (square 6). A plant of this supposed zygotic
constitution should resemble, in all respects, ShulPs
rhomb oidea, the earliest roset leaves being broad, and the
incisions of the climax leaves reaching the midrib. It
should breed true. Five lots, involving four parents, were
grown. Again it was shown, in the case of No. 26,812 and

Digitized by

Google

214

THE AMERICAN NATURALIST [Vol. XLVIH

No. 30,612, that the fact that plants are left unguarded
does not affect results. The parents, in all cases, were
selected from among the first generation of plants having
the supposed zygotic constitution BbNn. The results are
given in Table III.

TABLE III
Evidence or Homoztgotic Character of rhomboidea (BBnn)

Index Number

No.ofPUnta

Index Number of Parent

Char, of Parent

Gor 0

25.812

8,212J5iJlP6

BBnn

26,812

SM2BR3P2

BBnn

27.012

SA12ER10P6

BBnn

27.112

210

SM2ERISP12

BBnn

30,612

SA12BR3P2

BBnn 1

In all cases the off-
spring was uniformily of
the rhomboidea character.

bn
bn

5. Bbnn (squares 8 and
14) . Plants of this zygotic
constitution should resem-
ble those of the preceding
group, but on being self ed
should yield 25 per cent
homozygotic rhomboidea
{BBnn) J 50 per cent, het-
erozygotic rhomboidea
{Bbnn) and 25 per cent.
simplex {bbnn).

These three forms were
found to constitute the
offspring of a single plant,
8,212ffi?lP3, itself an off-
spring of a plant of the
supposed zygotic constitution BbNn. This plant, from

16 Bursa ... 39.

Fig. 12. Climax Lbavbs of a Hbtbb-
ozTOOTic C . . . rhomboidea and of a
HoMOZYQOTic C. . . . rhomboidea. . .

Digitized by

Google

No. 568] X CAPSELLA BUBSA-PASTOBIS ABACHNOIDEA 215

the first, was classified as a rhomboidea. At the present
time, a photograph of the young roset confirms this
classification. But two climax leaves, which, in the
earlier part of these experiments, were deemed suffi-
cient, show that the sinuses do not quite reach the midrib
(Fig. 12). Unfortunately, Shull, in the description of
his No. 054.28,^® does not mention this point, though he
does point out that **the later rosette-leaves had some
of the secondary lobes acutish, but not elongated. *' In
the older climax leaves, even of a homozygous rhom-
boidea, I find that the secondary lobes disappear. Shull,
in the description just referred to, is so specific as to the
typical rhomboidea character of the heterozygote that I
have hesitated to classify the heterozygotes and the homo-
zygotes. But the homozygotic rhomboidea, obtained as
the extracted recessive of a selfed plant of the supposed
zygotic constitution BBNn, always has sinuses which
reach the midrib. In other combinations, also, one can
distinguish between BB and Bb by the relative depth of
the sinus. For the present, then, we will rely upon this
character. In the case under discussion (26,612, the off-
spring of 8,212ffi?lP3, guarded) there were among the
39 plants 6 which clearly were simplex, the heterozygotic
rhomboidea was represented by 22 individuals, and the
homozygotic rhomboidea by 11 individuals, the calculated
ratio being 9.75 : 19.50 : 9.75. The percentage of simplex
is far too low, 15.4 per cent., instead of 25 per cent., but,
considering the small number of individuals concerned,
the total outcome is fairly satisfactory. It is almost tin-
necessary to add that in this, as in other cases, the off-
spring of the various plants is being tested as fast as
time and facilities permit.

Type 4. — ^Having shown the presumable correctness of
our supposition as to the zygotic constitution of the initial
plant {BbNn)y as far as the presence, appearance and be-
havior on breeding of simplex, rhomboidea and attenuata
are concerned, there remains to identify the major group of
combinations which, in a simple di-polyhybrid, constitutes

Digitized by

Google

216 THE AMEBIC AN NATURALIST [Vol. XLVIH

nine sixteenths of the total offspring and may be uniform
in appearance, the constitnents being separable only by
breeding, '*eine heillose Arbeit,'' as Baur has it. For-
tunately, in this case, it is possible to distinguish readily
between the various combinations.

One of the combinations, BBNN (square 1), should
breed true, being homozygotic for both characters con-
cerned. We would expect such a plant to have narrow
first leaves and climax leaves with incisions to the mid-
rib. Thus far I have not encountered such a plant, some-
thing which at one time led me to consider the possibility
of gametic repulsion, in this instance the gamete BN
being incapable of existence. This supposition seemed
the more plausible since the two genes B and N well might
be supposed to be antagonistic, the one being responsible
for an incision of the leaf to the midrib, the other tending
to make the leaf, especially the earlier leaves, narrow.
Were this assumption correct, none of the zygotic combi-
nations found in squares 1, 2 and 5, 3 and 9, and 4 and 13,
would be formed, though we would expect the same com-
bination as occurs in squares 4 and 13 to make its appear-
ance as the result of the fusion of the gametes bN and
Bn (squares 7 and 10).

Were this supposition correct, we should have a case
similar to that of the sweet pea ** Purple Invincible," and
we could not expect the gamete (bn) to be formed. Since,
however, simplex (bbnn) appears in our cultures, this
theory must be rejected. Eecently also, in culture No.
30,412, an instance was found in which the guarded parent,
supposedly of type 4, yielded, not simplex, rhomhoidea,
attenuata, arachnoidea as well as the parental type, but
only arachnoidea, rhomboidea and the parental type, and
in proportions closely approximating a ratio 1:1:2.

A plant which yielded 25 per cent, rhomboidea and no
simplex, must have been homozygotic for B, and since it
yielded also 50 per cent, of type 4, must have been hetero-
zygotic for N, its zygotic constitution therefore being
BBNn. Such a plant, on self-fertilization, should yield
25 per cent, rhomboidea. Provided the homozygote and

Digitized by

Google

No. 568] X CAPSELLA BUBSA--PAST0R1S ARACHNOIDEA 217

the lieterozygote have the same appearance, the remaining
75 per cent, should resemble the parent (Fig. 9, &).

BN
BN

Bn
BN

BN
Bn

Bn
Bn

But in one case (30,412), the parent being 8,4125i?9P9,
and open fertilized, the offspring consisted of 26.3 per
cent, rhomboidea, 46.2 per cent, of the parental type and
27.5 per cent, arachnoidea. If our supposition as to the
zygotic constitution of the parent is correct, then the
zygotic constitution of the arachnoidea in this offspring
must be BBNN. In the case of a selfed attenuata, we
found that approximately 25 per cent, of the offspring
was composed of arachnoidea of the probable zygotic con-
stitution bbNN. Is it possible that any Capsella, homo-
zygotic for N, would have the appearance of arachnoidea?
This seems more than probable, and other evidence, to be
adduced later, appears to support this view. The history
of the BBNn is as follows :

During 1912 I grew No. 8,412 from seeds of a plant
which resembled the grandparent 4,108.6. It was com-
posed of 1,079 individuals, among which various types,
such as '* broad, ^' ** narrow*' and ** linear,'' could be rec-
ognized. Not all plants were thus classified, a fourth
group of '* intermediates" being formed, indicating that
some of the plants, while in certain respects resembling
simplex and especially rhomboidea (deep lobing, second-
ary lobes), in other characters more closely approximated
the '* narrows," since their early leaves had been noted
as *' narrow." In the light of recent experience, it is easy
to see why the distinction was made, though at the time
the conception of the differences was most hazy. Several
of these ** intermediates" were grown, and of these a
single one yielded the seed for the next generation. This
plant had been permitted to flower unguarded, but after a
number of capsules had developed on the main stalk, this
was decapitated and the sideshoots were allowed to de-

Digitized by

Google

218

THE AMERICAN NATURALIST [Vol. XLVIII

velop. At this time the entire plant was caged. Subse-
quently the seeds of the open fertilized and of the guarded
flowers were sown separately, with the following results :

S0,412. Open Fertilized

Per Cent.

Planto

Found I Expected

30,512. Guarded
Plants

Per Cent.

Found I Expected

Aracknoidea. .
"Narrow"...
Rhomboxdea. .

27.5
46.2
26.3

40
67
38

36.25
72.50
36.25

21.15
36.15
42.70

105

61.50
123
61.50

The figures are given separately to again call attention
to the fact that open fertilization is no hindrance to pedi-
gree work in Capsella. Since the seeds came from the
same parent, we may add the results, which gives us
arachnoidea 23.50 per cent., ** narrow" 40 per cent, and
rhomboidea 36.50 per cent. The fact that the percentage
for ** narrow *' is too low and that for rhomboidea too
high, while the percentage for arachnoidea is within the
limits of probable error, is probably due to errors in
classification, since greater weight was laid upon lobing
of the adult leaves than upon comparative width of the
earlier ones. The value of this culture lay chiefly in its
suggestion of a zygotic combination BBNn, which prior
to that time, on account of the gametic repulsion theory,
was not supposed to exist. In consequence, a number of
cultures were made, with the following result :

TABLE IV
Evidence of Heterozygotic Char.\cter of Treleaseana (BBNn)

Index
No.

3,813
3,913
4,013
4,213
4.313
4,413

Total

Number of PlanU

1
1

BBXN

BBNn

BBnn

Found

Ex-
pected

44.76

89.50

44.75

15.26

30.50

15.25

28.60

45.25

102

90.50

45.26

173

196.75

400

393.50

214

196.75

Index No. of
Parent

30,412^«2P6
30,412^/?4P3
30,412^/26P3
S0A12BR2P6
30M2BR6P5
30,412J5/29P2

Char.

of
Parent

BBNn
BBNn
BBNn
BBNn
BBNn
BBNn

Char.

of
Grand-
parent

O BBNn

BBNn
BBNn
BBNn
BBNn
BBNn

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST ORIS ARACHNOIDEA 219

The *' narrows*' in question, then, fulfilled our expecta-
tion on the basis of a zygotic constitution BBNn. In
some cases the percentages are too high, in others too low.
The total yields fairly satisfactory results, to wit: BBNN
22 per cent., BBNn 51 per cent, and BBnn 27 per cent.
Two tests of the extracted recessive, a homozygotic rhom-
boidea, were made. The cultures. No. 3,713, from a
guarded rhomhoidea (30,412u4B2P3) and No. 4,113, from
an unguarded rhomhoidea (30,412u4iJ8P3), both derived
from plants of the supposed zygotic constitution BBNn,
yielded, respectively, 54 and 207 plants, all of which bore
the typical rhomhoidea characters.

In the cultures just tabulated, the plants of the sup-
posed zygotic constitution BBNn resembled the parent in
all respects. The form arachnoidea, in this case, must
have the zygotic formula BBNN. Unfortunately, in this
case also, it proved imf ertile.

A better acquaintance with plants of the zygotic consti-
tution BBNn led us to formulate certain differences be-
tween them and our original *' narrow.** Plants of the
BBNn character, readily can be segregated from those of
the BhNn character by somewhat narrower primary
lobes, split to the midrib and the development, in climax
leaves of well-grown specimens, of a secondary lobe, not
pronounced but recognizable (Figs. 8, 9).

On the basis of these morphological differences, as well
as because of the behavior of the plant on breeding, I
propose to segregate it from type 4 under the name
X Capsella Bursa-pastoris Treleaseana. This form is
homozygotic for B, while Setchelliana is heterozygotic
for B. Both are heterozygotic for N. They may be ex-
pected to look alike during the early stages. Later they
show a difference, since the form containing Bh does not
develop sinuses as deep as the form containing BB. The
form Treleaseana, when young, can readily be distin-
guished from a heterozygotic rhomhoidea (Bbnn) by the
relative width of the early leaves ; later such a distinction
is diflScult (Figs. 4, 5, 7). If any distinction at all is to
be made, it should be made on the basis of the rounding

Digitized by

Google

220 THE AMERICAN NATUBAZI8T [Vol. XLVUI

of the lobes, those of Treleaseana being sharp, those of
the heterozygotic rhomboidea rounded.

I am fully aware that in thus naming genotypes, I am
departing from all rules laid down by systematists. But
a rule is useful only as long as it serves a purposed For
the geneticist, the rules of systematists are of small value.
Subspecies, variety, form, are, after all, but very general
terms, almost incapable of definition because of too fre-
quent abuse. But once we have determined the zygotic con-
stitution of any plant, we have placed ourselves on a firmer
basis. Behavior in breeding is the proper criterion. And
while I recognize that this, for systematic purposes, is
impracticable, at the same time I assert the right to use
a trinomial for any organism of known zygotic constitu-
tion, this being, at the present time at least, the easiest
way of designating any particular form. Some day we
shall have formulas, corresponding to those of chemistry,
to designate the lesser forms.

The increase in the number of named forms, a neces-
sary consequence, need cause no alarm, since they concern
only him who occupies himself with one species exclu-
sively. But we must go even further than this. Squarely
facing the issue, we find ourselves placed in apposition
which necessitates the naming of heterozygotes. Obvi-
ously, numerous objections could be urged. But since it
has been shown, on the one hand, that certain forms can
exist only in a heterozygous form (Baur^s Antirrhinum)
and, on the other, that not only the difference between the
homozygote and the heterozygote is as great as that be-
tween many of our ** systematic '^ species (for instance,
attenuata, bbNn, and arachnoidea, bbNN)j but that a
homozygotic condition for a single gene gives the same
result, whatever the condition of the other known genes,
at least as thus far determined (arachnoidea occurs as
aaBBNN, aaBbNN and aabbNN)^ the advantage of nam-
ing all forms of different zygotic constitution must be
granted.

Thus far we have not encountered a plant of the zygotic
c<mstitution BbNN, at least as far as can be judged from

Digitized by

Google

Na 568] X CAPSELLA BURSA-PASTOBIS ARACHNOIDEA 221

breeding experiments. On being selfed such a plant
should yield :

BN
BN

bN
BN

BN
bN

hN
bN

It has been shown that plants of the zygotic constitu-
tion BBNN and bbNN exhibit the arachnoidea type. At
least 50 per cent, of the offspring then should show this
character. But if the suggestion made above is the cor-
rect one, i. 6., that all plants homozygotic for N exhibit
the arachnoidea type, then the parent and its entire off-
spring should bear this character. The unfortunate in-
fertility of arachnoidea prevents us from submitting this
hypothesis to direct experimental proof. But there exist
indirect means for establishing the probable truth of our
contention. In the first place, we may cross two plants,
the identity of which can be established beyond doubt, to
wit, attenuata (bbNn) and Treleaseana (BBNn). Such a
cross would yield :

Bn
bN

Bn
bn

Of these, we would recognize Bbnn because of its rhom-
boidea character, 50 per cent, would be recognized as
Setchelliana (BbNn), while the remainder, if our surmise
is correct, would consist of arachnoidea. Experiments to
determine this are under way. At the present we have
another, though by far less accurate, means of testing our
hypothesis. If the combination NN always results in a
form arachnoidea, the offspring of a plant of the zygotic
constitution BbNn would be composed of :

4 Setchelliana (BbNn),

2 Treleaseana (BBNn),

2 attenuata {bbNn),

4 arachnoidea (1 BBNN, 2 BbNN, 1 bbNN),

Digitized by

Google

222

THE AMEBIC AN NATURALIST [Vol. XLVHI

3 rhomhoidea (1 BBnn, 2 Bbnn),

1 simplex (bbnn).
Since BbNn, BBNn and bbNn, in the earlier experi-
ments, might have been confonnded in the later stages,
and since there is little doubt as to the earlier stages,
these three forms have been combined in Table V.

TABLE V
Besults from Selected Setchetliana (BbNn)

Index

"Narrow"

rhomboidea

simplex

No.

Foand

Expected

Found

Expected

Found

Expected

Found

Expected

26.912
3,613

134
94

167.6
89

94
45

78.8
45

68
33

59.1
33.75

19
8

19.7
11.25

This, especially in the case of No. 3,613, is a fairly close
approximation to what we might expect. When in No.
3,613 we attempt to distinguish between Setchelliana,
Treleaseana and attenuata, we get the following num-
bers, the expected numbers following in parentheses:
BbNn 39(45), BBNn 21(22.50), bbNn 34(22.50), the last
number being far too high. When the experiments were
begun, we distinguished only between '* narrow,"
'* broad'' and ''linear.'' To-day we know that the ''nar-
rows" include Treleaseana, Setchelliana and attenuata,
that the "broads" include rhomboidea and simplex, while
the linears are identical with arachnoidea. In this light
it is of interest to go back to the first generation of 1910.
Our data yield the figures given in Table VI.

TABLE VI

"Narrow"

"Linear*'

"Broad"

Index No.

Found

Expected

Found

Expected

Found

Expected

7,911

30.50

15.25

8,111

35.50

17.75

8.311

30.50

8.711

13.50

8.811

46.50

23.25

9,011

7.50

3.75

9,511

9.611

23.50

11.75

Total

328

308.50

133

154.25

156

154.26

Per cent

53.2

21.5

25.3

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST OBIS ABACHNOIDEA 223

It must be granted that the approximation is fairly
close, and that, taken in consideration with the others, it
offers ample support for the correctness of the diagnosis
of the zygotic constitution of the original plant. It at
least offers a working basis. One would be tempted to
accept it as a final solution were it not for the fortunate
appearance of a plant which does not fit into our scheme
and which, provisionally, has been nsuxiedCapsella Bursa-
past oris orbicularis.

Capsella Bubsa-pastobis obbiculabis
This form differs from any other plant encountered
in my cultures. While in a general manner resembling
simplex, it differs in being more robust, having larger
flowers (though not as large as those of C. grandiflora)^
and in having orbicular first
leaves (Fig. 13). All leaves
are covered with stout hairs.
It is a plant which tempts us
to draw a parallel between it
and (Enothera gigas, a name
which I have not used for the
sake of avoiding an impUed .'^;''^^,«^''"''*'' ''^ '^'
comparison.

The first plant of this type appeared in a culture of
attenuata (26,0125B3P5) and was of sufiSciently striking
appearance, though but four or five leaves had developed,
to call for a special note and a photograph. Later the
plant was potted and finally seed was gathered from the
unguarded plant. From this seed four seedlings were
obtained. At least three of them closely resembled the
parent, the fourth having somewhat narrower leaves.
Later the differences between these plants and those of
simplex became more apparent (Fig. 14). Those of my
students to whom the differences have been pointed out
have not the slightest difficulty in distinguishing between
the two forms. It is hoped that later, when by means of
prolonged cultures I shall have made myself more familiar
with this form, it may be made the subject of a distinct
paper where histological and cytological studies will find

Digitized by

Google

224 THE AMEBIC AN NATURALIST [Vol. XLVm

a place. One would be inclined to look upon orbicularis
as a mutation. But the fact that at first we classed
arachnoidea as such, later to prove it of hybrid origin,^^

Fio. 14. Four Seedlings of C . . . orbicularis and (the lower) Two Seed-
linos OF C. . . . simplex.

would tend to make us cautious, and lead us to attempt to
find a solution for the origin of orbicularis in the disso-
ciation or combination of certain * * units. ' ' While I should
not care to go quite as far as M. Heribert Nilson'® **das
ganze Mutations phanomen durfte unter einen gemein-
samen Gesichtspunkte : der Mendelschen Neukombination
eingeordnet werden konnen," yet it is probable that here
the majority of alleged mutations may be classed.

iTBaur's (Vererbungslehre, 189) narrow-leaved Melandrium album is
perhaps susceptible of the same explanation.

i^Zeitschr. f, ind. Ahst. «. Bcrerh., 8: 89, 1912.

Digitized by

Google

No. 568] X CAPSELLA BURSA-F AST ORIS ARACHNOIDEA 225

An examination of the herbarium material placed at
my disposal reveals the fact that plants, apparently
identical with C. orbicularis, occur in Europe. In the
Engelmann herbarium of the Missouri Botanical Garden
are two sheets (No. 3,661 and 3,664) containing specimens
which undoubtedly must be classed here. The latter sheet
bears the label : Thlaspi Bursa-pastoris humUe. Heidel-
berg. April 1828.

A culture of Capsella, derived from seed of a single
plant, unfortimately not preserved, escaped from culti-
vation in the Experiment Garden, and consisting of 182
individuals (Ehlers, No. 4,813), appears to be composed
entirely of orbicularis. And while I have never encoun
tered the plant in nature, these two facts lead us to another
possible explanation. Perhaps the appearance of orbic-
ularis in the original culture was due to an accidental
admixture, such as is almost impossible to guard against
when experimental plants are grown in a greenhouse used
for a variety of purposes.

The exact relation which orbicularis bears to the other
types of Capsella here described can, of course, be de-
termined only after a series of experiments has been car-
ried out. However, the delay in the completion of the
manuscript, caused by the unfortunate destruction, by
fire, of the botanical laboratories of the University of
Michigan, enables me to add that a third generation of
orbicularis, the parent being No. 32,012B1P3, shows at
least two and possibly three types, of which one is espe-
cially interesting in having rather narrow leaves, at least
as compared with those of typical orbicularis. The con-
trast between the two forms is increased by the fact that
in the narrow-leaved form the foliage is entirely glabrous,
while in the typical orbicularis the leaves are covered
with numerous stiff, almost bristle-like, hairs.

X Capsella Bubsa-pastobis abachnoidea
By this name is designated the linear-leaved form, the
appearance of which induced us to undertake the cultiva-
tion of Capsella Bursa-pastoris Setchelliana.

Digitized by

Google

226 THE AMERICAN NATURALIST [Vol. XLVIII

FKJ. l.'i. I{r)SKT.S iLI.rSTUATING Tl.K Two TYPKS

Already the leaves which immediately follow the coty-
ledons serve to distinguish plants of this type from all
others. At the ten-leaf stage even the casual observer is
able to segregate them at once from the other rosets.
The leaves are acicular and the cotyledons far larger
than those of the seedlings of the other forms. The
greater size of the cotyledons may be attributed to the
insufficiency of the subsequent leaves.

If one removes the terminal bud of seedlings of Atri-
plex hortensis or one of its color varieties, it will be found
that the cotyledons increase in length far beyond normal,
sometimes reaching a length of 8 cm. Under favorable
conditions the leaves of X C. arachnoidea may rfeach a
length of 100 mm., with a greatest width of 6 mm. (Fig.
16). The stem ordinarily is weak, having a diameter of
only 1 mm. It may reach a length of 30 cm. (Fig. 17).

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST OBIS ABACHNOIDEA 227

fiijs J^AK Recognized in C. . . . orbiculaHs.

flowers are small, the petals especially so. The

^^x^s shrivel up early and as a rule are devoid of polleu

^^^^'^^^^ Occasionally a few can be demonstrated. The

ovatr^,-^ though small, contains what appear to be ovules

.^^•^Xe of being fertilized. Thus far I have collected

f ^ :seeds contained in 6 capsules on unguarded plants

^ ^^^^^<;hnoidea (Fig. 18). Two of these germinated, the

^ ^ ^5^ ielding a plant which looks like simplex, though

ua^xx^^ a large amount of red coloring matter in the peti-

oVfe^^ "While the other is an arachnoidea. Attempts to arti-

\ tvc\^lXy fertilize arachnoidea have failed absolutely.

, ^^ has been shown above, one may distinguish, on the

"^ V^'^i.H of genotypic constitution, three forms of arach-

"^^ v^oidea, viz.: BBNN, BbNN and bbNN. Externally no

^'v^.

Digitized by

Google

THE AMERICAN NATURALIST [Vol. XLVffl

ifferences can be noted. A single exception perhaps may
e made to this statement. It had been noted that speci-
lens of arachnoidea frequently showed f asciation. This
isciation seems most marked in plants of the zygotic
institution BBNN (Figs. 19, 20, 21).

Fig. 16. Roset of x C. ,

arachnoidea.

While it is hoped that later a more extended report r^Sigr />
B made upon this plant, at present it may be stated tj^^
lere exists the probability, that it may throw some li^i^?
pon the nature of fasciations. In earlier publication^! »
have brought together some of the known facts bearing
pon this teratological character. Though a large por^

19 < < Fasciation in Oxdlis crenata and Experimental Production of Fascia.
ons," Bep, Mo. Bot. Gard., 17: 147, 1906; "Fasciations of Known
lusation, * ' American Naturalist, 42 : 81, 1908 ; ' ' Inheritance of Fascia-
on in Zea Mays," The Plant World, 14: 1911; "The Origin of Species
Nature," American Naturalist, 45: 641, 1911; " Frondescence and
Etfciation," Plant World, 14: 1911; "Fasciation in Oxalis crenata,"
otanical Journal, 2: 111, 1913.

Digitized by

Google

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 229

tion of the experimental garden is devoted to cultures of
fasciated races, nothing further has been determined than
that the fasciated character is inherited, that it is trans-
mitted through non-fasciated individuals, that its ap-
parentness depends upon nutrition, that it behaves as a

Fig. 17. Two Hebbarium Specimens of x C. . . . arachnoidea.

recessive character and that the fasciated character of the
stem appears to be associated with split leaves and cup-
shaped leaves. In a paper read before the Research Club
of the University of Michigan on March 16, 1910, and an-
nounced under the title ''The Identity and Inheritance of
Teratological Characters,^* I showed that split leaves,
ascidia, certain disturbances in the arrangement of the

Digitized by

Google

230

THE AMERICAN NATUBALIST [Vol. XLVin

flowers, supemumerary locules in the fruit, etc., may
safely be taken as an indication of the presence of the
fasciated character. More recently, Kajanus,2<> working

Fig. 18. Shoot of x C. , . . arachnoUleu, with a Large Number of Infertile
AND Few Fertile Capsules.

with different material, has fully confirmed the views
which I expressed at the time. This is of particular in-

20 Kajanus, B., "Polyphyllie und Fasziation bei Trifolium pratense L.,"
Zeiisch, /. ind, Ahst, u, Vererh., 7: 63, 1912; '*Ueber einige vegetative
Anomalien bei Trifolium pratense L., ibid., 9: 111, 1913.

Digitized by

Google

No. 568] X CAPSELLA BURSA-P AST ORIS ARACHNOIDEA 231

terest in connection with Capsella arachnoidea, since
many of the plants which do not show a fasciated stem
do show split leaves (Fig. 20) and a most peculiar whorl-
ing of the flowers (Fig. 21).

Fig. 19. Fasciated Plant of a* C . . . arachnoidea.

The spatulate condition of the leaves of the seedling
shown in Fig. 2 is believed to have been due to fasciation.

Capsella Bursa-pastoris arachnoidea, then, bears all the
earmarks of a fasciated race. All of the three zygotic
combinations which yield the arachnoidea type are homo-
zygotic for N. The recent work of East and Hayes, and
of Emerson on Zea Mays has shown that the fasciated

Digitized by

Google

232

THE AMERICAN NATURALIST [Vol. XLVIH

Fig. 20. Leaves of x C aruchnoidea.

Fig. 21. ABNORMAL Whorled Arrangement op the Flowers in Inplokbs-
CENCES OF w O. , , , arachnoidea.

Digitized by

Google

No. 568] X CAP8ELLA BUR8A-P AST OBIS ARACHNOIDEA 238

character is dominant, though Mendel, in his experiments
with Pisum umbellatum, has shown it to be recessive.
No fasciation, thus far at least, has been noted in the
other forms used in these experiments.

SUMMABY

A culture of Capsella Bursa-pastoris proved heterozy-
gotic, yielding certain new forms (X C. Bursa-pastoris
SetchelUana, xC. Bursa-pastoris Treleaseana, XC. Bursa-
pastoris arachnoidea and X C. Bursa-pastoris attenuata)^
as well as certain forms already described by ShuU (C
Bursa-pastoris rhomboidea and C. Bursa-pastoris sim-
plex) in the proportion 4:2:4:2:3:1. The distinction
between simplex and rhomboidea, both inter se and be-
tween them and the other forms, is readily made by any
one familiar with Shull's investigations. These two
plants agree in having the earlier leaves broad (Fig. 4).
The climax leaves of rhomboidea and simplex show
marked differences, especially as far as the incision of
the blade is concerned. These incisions, in simplex, reach
a depth equal to approximately one fourth of the width
of the blade (Fig. 10). In rhomboidea the incisions are
deeper, reaching the midrib in the homozygous form
(Fig. 12). The leaves of the latter also show marked
secondary lobes.

The distinction between X C. Bursa-pastoris Setchelli-
ana, X C. Bursa-pastoris Treleaseana and X C. Bursa-
pastoris attenuata is made with greater difficulty. They
agree in having long and narrow first leaves. The climax
leaves of Treleaseana and SetchelUana show marked
incisions, exceeding one fourth of the width of the blade,
and which may reach the midrib (Fig. 9). The latter
form also may show marked secondary lobes.

Besides the phenotypes here mentioned occur two
others, the one, X C. Bursa-pastoris orbicularis, with an
almost orbicular first leaf (Fig. 13) and a climax leaf
greatly resembling that of simplex (Figs. 14, 15), though
differing in texture. This form has not been sufficiently
studied, but is believed to be identical with one known to

Digitized by

Google

234 THE AMERICAN NATURALIST [Vol. XLVIU

occur in Europe. Finally there is X C. Bursa-pastoris
arachnoidea, a sterile, linear-leaved form, with a weak
stem and which frequently shows fasciation (Figs. 17-21).
To facilitate a distinction between these forms, a key is
appended :

a. Early leaves broad.

h. Early leaves orbicular. orbicularis,

&6. Early leaves twice as long as broad.

c. Climax leaves incised to midrib. rhomboidea.

cc. Early leaves not incised to midrib. simplex,

aa. Early leaves long and narrow.

b. Early leaves acicular. airachnoidea,

bb. Early leaves 2^-3 times as long as broad.

c. Climax leaves not incised to midrib. atienuaia.

cc. Climax leaves incised to or almost to the midrib.

d. Secondary lobes pronounced. Treleaseana.

dd. Secondary lobes absent. Setchelliana.

It was found that, besides the genes A, B, C and D,
whose existence was shown by ShuU, there exists another
gene, N, responsible for the narrow character of the
earlier leaves. For the various forms, mentioned here,
the following zygotic constitutions have been tentatively
determined: simplex, bbnn; rhomboidea, BBnn and Bbnn;
Setchelliana, BbNn; Treleaseana, BBNn; attenuata,
bbNn; arachnoidea, BBNN, BbNN and bbNN. The zygo;
tic constitution of orbicularis has not been determined.

As to the probable origin of X C. Bursa-pastoris Setch-
elliana, little can be said. It most probably results from a
cross between rhomboidea and attenuata {BBnn X bbNn).
This seems the most plausible explanation since, judging
from herbarium specimens, both attenuata and rhom-
boidea occur throughout the United States. Unfortu-
nately such an assumption necessitates an explanation of
the origin of attenuata.

My thanks are due to the regents of the University of
Michigan for the facilities placed at my disposal, to head-
gardener Adolph Weiner for his constant care of the ex-
perimental plants, to Messrs. J. H. Ehlers, A. Povah, C.
Oberlin and A. W. Murdock for assistance in classifica-
tion of the seedlings and to the director of the Missouri
Botanical Garden for the loan of herbarium material.

Digitized by

Google

No. 568] X CAPSELLA BURSA-PASTORIS ARACHNOIDEA 235

Conclusions

1. Besides the genes {A, B,C, D) discovered by Shull,
there exists in Capsella a gene N, responsible for the nar-
row character of the early leaves of certain forms.

2. Absence of the gene N results in the formation of
early leaves of a *' broad'' character.

3. The form designated arachnoidea is of hybrid origin,
as are the forms Setchelliana, Treleaseana and attenuata.

4. X Capsella Bursa-pastoris arachnoidea is formed
whenever the plant is homozj'gotic for N, whatever the
constitution of the remainder of the zygote {BBNN,
BbNN, bbNN), i. e., a homozygous condition for the pres-
ence of a single factor may overshadow the influence of
others.

5. Homozygocity for a single factor may be responsible
for total, or almost total, sterility.

6. A knowledge of the early stages, as well as of the
climax leaves, is essential for the classification of the
phenotypes of Capsella Bur sa-past oris.

Digitized by

Google

BIOLOGY OF THE THYSANOPTERA. H

DR. A. FRANKLIN SHULL
University of Michigan

II. SEX AND THE LIFE CYCLE

Intkoduction

From observations made on the abundance of males in
several species, Jordan (1888) was led to believe that
there might be among Thysanoptera, as in aphids, an
alternating life cycle ; that is, that there might be a series
of parthenogenetic generations during the summer, fol-
lowed by a generation of males and sexual females in the
latter part of the summer or in the fall. Coupled with
this he suspected that there were winged forms in the
parthenogenetic part of the cycle, and at least occasional
wingless individuals in the sexual phase.

Uzel (1895), however, was unable to detect any indica-
tions of such a cycle. He held that there could be no
question of parthenogenesis in a species in which males
were abundant all the time or at intervals. Only in spe-
cies in which the males were too rare to impregnate all
the females would he admit parthenogenesis. To prove,
in such a species, an alternating cycle like that of the
aphids, it must, in UzePs opinion, be shown that the males
are abundant only at certain seasons. As Uzel was ac-
quainted with no European species in which males were
plentiful at but one season, he rejected Jordan's sugges-
tion regarding an alternating cycle, and his view seems
to have been accepted by thysanopterists since that time.

To Uzers argument it may be objected that the pres-
ence of males, and even the occurrence of copulation, is
no proof that parthenogenesis is wanting. For among
the aphids and rotifers, the parthenogenetic and sexual
females exist side by side. Nor is parthenogenesis in
these two groups facultative (optional), as Uzel appears
to assume for Thysanoptera; a female is either only

236

Digitized by

Google

No. 668] BIOLOGY OF THE THTSANOPTERA 237

sexual or only parthenogenetic. Moreover, in the roti-
fers, females incapable of fertilization copulate as fre-
quently as do those requiring fertilization, as was first
shown by the work of Maupas (1890) on the rotifer
Hydatina.

Presence of males and occurrence of copulation are,
therefore, no proof of sexual reproduction. But even if
we accept, as Uzel does, this criterion of sexuality, Jor-
dan's view that there may be an alternating cycle would
receive some support if it could be shown that males are
more abundant at one season of the year than at other
times. Casual observations made by me several years ago
seemed to indicate this seasonal variation in the abund-
ance of males. As the data then available were meager,
no conclusion was drawn, but I subsequently undertook
to obtain such data on a larger scale, by making extensive
collections at all seasons of the year to determine the sex
ratio. The following pages give these data, along with
other observations bearing on sex or the life cycle.

I desire to acknowledge the assistance of my wife, by
whom much of the labor of determining species and
counting the sexes was done.

The Sex Ratio in Vabious Species of Thysanopteba
In making collections for the purpose of determining
the sex ratio, the food plants were examined very care-
fully, torn apart if necessary, and every individual cap-
tured. This precluded the possibility of obtaining an
erroneous sex ratio because one sex was more easily dis-
turbed than the other. A few individuals escaped, but
they could not have affected the sex ratio very greatly,
and it was known from their size that they were some-
times of the one sex, sometimes of the other.

The sex in the suborder Terebrantia is readily deter-
mined by the presence of an ovipositor in the female and
the rounded end of the abdomen in the male. In the sub-
order Tubulifera, the sex in Anthothrips verhasci was
determined by the presence of two short, heavy spines,

Digitized by

Google

238 THE AMEBIC AN NATURALIST [Vol. XLVm

one on each side of the abdomen of the male, near the end.
As the specimens, when placed on a microscope slide,
nearly always lie either on the dorsal or ventral side,
these spines are nearly always readily visible if present.
I used this criterion (mentioned in the re-description of
the species by Hinds, 1902) only after having taken eleven
pairs of this species copulating in nature, and observing
in every case that the male possessed these spines, and
that in the female they were wanting. In .other Tubu-
lifera, e. g., Anthothrips niger, sex was determined by the
longitudinal chitinous rod in the next to the last abdomi-
nal segment of the female. When the specimens were too
opaque to observe this rod, they were cleared by boiling
in caustic potash.

The data from these collections are given in the accom-
panying table. Unfortunately the collections could not
all be made in one year, nor in the same locality. Those
made from July 1 to September 18, 1912, were made at
the University of Michigan Biological Station, Douglas
Lake, Michigan; all others were made at Ann Arbor,
Michigan. It is not probable that the results are greatly
modified by collecting in two regions within the state.
In this table the larvae of all species are combined, as I
am unable to distinguish with certainty the larvae of
several of the species here mentioned.

The important facts contained in this table are, it seems
to me, the following :

Euthrips tritici appeared in spring at first only in the
female sex. Males were first collected nearly a month
later, and not until about the time fairly large larvae were
found elsewhere. Once the males appear, though their
number fluctuates in the individual collections, they fur-
nish a fairly constant proportion of the whole number
(about one third).

The males of Anthothrips verbasci appear in the earli-
est collection of this species, and in considerable numbers
throughout the season. The total proportion of males is
23 per cent., and the only considerable increases over

Digitized by

Google

No. 568]

BIOLOGY OF THE THYSANOPTERA

239

this percentage in individual collections are in the three
collections made in August, and on October 7. Consider-
ing the large majority of females taken September 12, the
abundance of males October 7 may be due in some way to

TABLE I

Showing Number of Males and Females of the Commoner Species of
thysanoptera captured at intervals during the active season

Date

Apr. 30, 1911
May 5

24
June

July

1912

Aug.

Sept.

Oct.

Nov.

7
15
21
29

5
11
16
17
19
26
27
29
30
31

9
12
13
18
20
21

2,
12
16
18. 1912

7, 1911
14
25, 1912

1911

Enthrips
tritici

tkripg
verbagk'i

Antho-
thrips
niger

9 cf

9 cf 9 cT

9; 0

7| 0

2| O!

17l 0

22 1

30 2

174

30 20 0

41 17

18 0

40 62! 52 10 42

21 27; 47 8 18

87|

11 gI

14! 25:

3
21

7
01
3!
2
1
29

0,
0

I 4

29 2

154;i51

81 1
16 13

43 25

58 15

30, 20

26' 18|

Total .

13
4

19
3

17i

ll
27

I I

879|441 64l'200'l62' 0 226' 0 530!l74 50

ro-

thrips

jnani-

Larvae,

catus

all

--_ - -

Species

5
46
72

187

16
3

5 0

3
2

2 103

21
13
10

2
18

203

191

Digitized by

Google

240 THE AMERICAN NATURALIST [Vol, XLVIH

the dying of their food plants ; but the greater proportion
of males throughout August is probably significant. It
should also be stated that I have collected adults of this
species, of both sexes, from dead mullein spikes in late
winter.

Anthothrips niger was found only in the female sex.
There are no records of males of this species, so far as I
am aware, in any published work.

Thrips tabaci was taken almost exclusively in the
female sex, the two males found September 2 being the
only ones I have ever collected.

In Anaphothrips striatus the total number of males is
less than 25 per cent. On August 20 and September 2 the
proportion of males is considerably greater than 25 per
cent., especially on the former date, while at other times
the proportion was nearly always less. The collection on
August 20 can hardly have been erroneous by chance, for
the figures given for that date are combined figures for
two collections from different localities. In one of these
collections there were 13 females and 14 males, in the
other 26 females and 34 males. This strengthens the
probability that the excess of males is significant.

Thrips physopus was collected in small numbers, but
shows a fairly constant proportion of males.

Chirothrips manicatus presents curious phenomena.
All the collections up to the end of July were made on
timothy heads in a small patch a few feet square near the
laboratory. On July 11 careful search revealed numer-
ous females, but only one male. By July 19 almost all
the thrips of this species were gone; only 5 specimens
were obtained, and these were females. Less than two
weeks later, however (July 30), on other timothy heads
in the same small patch, there were foimd 51 males and
but 1 female. No living thrips were taken here later, as
the timothy died; but subsequent collections elsewhere,
from timothy and bluegrass, show again almost ex-
clusively males.

Digitized by

Google

No. 568] BIOLOGY OF THE THYSANOPTERA 241

Additional Data Beabing on the Life Cycle and Sex

In view of the fact, to be discussed later, that Anapho-
thrips striatus has hitherto been known ahnost exclu-
sively in the female sex, and is known to reproduce par-
thenogenetically, and the fact that in the collections here
recorded the males constitute nearly 25 per cent, of the
total, the question arises, are these males functional f If
not functional in this species, are the males functional in
other species! A number of observations and experi-
ments I have made bear on these questions.

A single pair of Anaphothrips striatus was found copu-
lating in nature, which Uzel would have considered proof
that parthenogenesis did not occur. The testes of the
males are plainly visible without dissection. Suspecting
that they might not be fleshy organs at all, but chitinized
structures, perhaps vestiges of testes, I boiled a number
of specimens in caustic potash. The testes disappeared,
from which I judge they are not merely chitinous bodies.
I can say nothing of their cellular nature, owing to the
loss of material killed and fixed for that purpose. Nu-
merous sections of another species Anthothrips verhasci,
however, reveal well-developed testes. Cell divisions
(probably the spermatocyte divisions) and nearly mature
spermatozoa in bundles were observed in these sections.
Though the number of chromosomes could not be deter-
mined, it is an interesting fact that spindles in side view
usually showed a lagging chromosome.

Finally, with further regard to the functioning of
males, I have attempted to breed several species par-
thenogenetically. The results in thfe case of Euthrips
tritici were so far encouraging that two larvae appeared
on the plant on which virgin females had been previously
placed. But in these cases I could not be certain that the
food plant was uninfected. Experiments with Anapho-
thrips striatus and Anthothrips verhasci gave negative
results, but in each case failure to obtain young by par-
thenogenesis may have been due to the conditions.

Some observations on the place of pupation may also

Digitized by

Google

242 THE AMERICAN NATURALIST [Vol. XLVm

be here recorded. The rarity with which the pupae of
most species are discovered in collecting suggested that
they might not pupate on the food plant of the larvaB.
Some species of thrips, for example, the pear thrips
{Euthrips pyri), are known to pupate in the ground
(Moult on, 1912). Since many of the species included in
Table I may be found on white clover, which was abun-
dant at Douglas Lake, the place of pupation of these
species was tested in the following manner. A mass of
the flowers of white clover was collected. The flowers
were gently squeezed for some time to drive out all the
adults. They were then placed in a vessel under cover.
After two days, when the flowers were thoroughly dried,
they were again gently crushed to make sure that all
adults were driven out. At intervals from one to two
weeks afterward, 15 adult thrips appeared on the inside
of the glass cover. These were of three species, Euthrips
tritici, Thrips tabaci and Anthothrips niger.

I have also frequently observed the pupaB of Antho-
thrips verbasci in mullein spikes, those of Sericothrips
cingulatus on white clover, the pupa of Trichothrips tri-
dentatus under the bark of the white oak, where the larvae
and adults live, and that of an undescribed species on
willow galls along with larvae of the same species. I
judge from these observations that the majority of thrips
pupate on the plants on which the larvae live, and that
their rarity in collections is due merely to concealment
and sluggish habits.

Discussion of the Results in Relation to the Life

Cycle
From the data in Table I and the observations given
above it is evident that there is considerable diversity in
different species with regard to the life cycle, and diver-
sity within the same species at different times or in
different regions. First, as regards the mode of passing
the winter, it would seem that in Euthrips tritici only the
females survive that season. The reason for so believing

Digitized by

Google

No. 568] BIOLOGY OF THE THTSANOPTERA 243

is that males could not be found in the spnng until the
females had been active long enough to have produced
one generation of offspring. Males occur late in autumn,
but must perish before the end of winter. Likewise,
neither eggs nor larvae live over winter, or larv® would
appear earlier in spring. In Thrips physopus, on the
other hand, males were found as early as the females;
hence, in the absence of any collection earlier than May
22, and in ignorance of the time required for develop-
ment, I should assume that both sexes survive the winter.
Both sexes of Anthothrips verhasci have been seen on
dead mulleins in winter.

In species, like Euthrips tritici, whose males do not
survive the winter, if fertilization of the early spring
females takes place at all, it must occur in the fall. I do
not regard my breeding experiments as proof of par-
thenogenesis in this species, but it is by no means improb-
able that parthenogenesis occurs. More rigorous experi-
ments are needed.

As regards the mode of reproduction during the rest of
the year, there is nothing in the sex ratio, as given in
Table I, to suggest an alternating cycle in Euthrips tri-
tici. In other species, it would be possible to interpret
certain facts to mean that an alternation of partheno-
genesis and sexual reproduction occurs, or did once
occur. There is a well-marked increase in the proportion
of males in Anaphothrips striatus, for example, in Au-
gust. This is a particularly interesting species. Hinds
(1902) saw only the female of this species, though he
mounted and examined over a thousand specimens, and
he bred it parthenogenetically in the laboratory for
months. What purported to be the male was described
by Gary (1902), from Maine, but the specimens described
were evidently those of another species. The first males
ever recorded were described by ShuU (1909), two speci-
mens among probably two hundred females. It is re-
markable, therefore, that in the vicinity of Douglas Lake
there should be nearly 25 per cent, of males. Whether

Digitized by

Google

24^ THE AMERICAN NATURALIST [Vol. XLVHI

the presence of numerous males is dependent on climatic
conditions, or whether it is a racial difference, there is at
present no way of deciding. The weather was unusually
cold during the summer in which these records were
made, and it is desirable that the effect of temperature be
experimentally determined. The presence of males in
goodly numbers throughout the summer, the occurrence
of copulation in nature, and the failure of an attempt to
breed the species parthenogenetically, leave, as the only
reason for suspecting that it may have been partheno-
genetic at Douglas Lake, the fact that it is parthenoge-
netic elsewhere. But if the species is parthenogenetic in
one region and sexual in another, it is not diflScult to be-
lieve that it may be both parthenogenetic and sexual in
the same region. It is diflScult to decide whether the well-
marked increase in the proportion of males in August
and early September should be regarded as evidence of
such an alternation, or as due to a period of cold weather
or other climatic factor, or as a hereditary remnant of
the sexual phase of an alternating cycle once possessed
by the species. Only experiment, and perhaps cytological
study, can decide this question.

A similar but less marked increase in the number of
males is seen in Anthothrips verhasci, also in August. In
that month the proportion of males rose from about 20
per cent, to 40, or even nearly 50 per cent. In this species
the increase may be due to the late date at which the first
brood of larvfiB becomes mature. The life history of this
species is longer than that of most of the suborder Tere-
brantia, and may appear to be still longer because ene-
mies destroy many of the larger larvae. For these rea-
sons, in the region of Douglas Lake, the first generation
of larvae may not become mature until nearly August. If
this assumption is correct, the proportion of males found
prior to August is the proportion that survive the winter.
This explanation receives support from the cytology of
the germ cells. As stated above, there is a lagging chro-
mosome in the spermatocyte divisions, which suggests

Digitized by

Google I

No. 668] BIOLOGY OF THE THTSANOPTERA 246

the probability that there are two classes of sperm asso-
ciated with sex, as in the bugs and many other animals,
and that therefore the sexes should be approximately
equal in numbers. The 40 to 50 per cent, of males in
August accord fairly well with this explanation.

This explanation would not, however, account for the
increase in the number of males in late summer in a spe-
cies whose life Tiistory is much shorter than that of
Anthothrips verbasci. Thus, in Anaphothrips striatus,
Hinds states that the entire life history is passed through
in 12 to 30 days. Even in a cold season, such as that of
1912 at Douglas Lake, therefore, the life history can not
have been so long that the first adults would emerge in
the middle of August. The increase in the number of
males of Anaphothrips in August and September is not
to be explained, therefore, as due to the first appearance
of a new brood at that time.

Thrips tabaci likewise affords interesting, even if mea-
ger, evidence regarding the seasonal occurrence of males.
In this species males are exceedingly rare. Hinds (1902)
redescribed the male in quotation marks, from which it
is to be inferred that he did not have specimens. In my
own collecting, though the females were quite common, I
never saw a male until the summer of 1912. Then two
specimens were taken September 2, as shown in Table I*
These irregularly occurring males can hardly be func-
tional, so that Thrips tabaci is still probably to be re-
garded as wholly parthenogenetic. But their appearance
in late summer may be the vestige of a former sexual
phase, and may be caused now, as the sexual phase prob-
ably was in part formerly caused, by climatic conditions.

Chirothrips manicatus presented, at Douglas Lake, an
anomalous condition. As shown in Table I, and stated
more explicitly above, females were abundant in a given
small area early in July, but practically no males were
pr^ent. Then, so far as I could determine by painstak-
ing collections, the females disappeared ; almost no adults
of either sex, and not many larvae, were to be found. Two

Digitized by

Google

246 THE AMERICAN NATURALIST [Vol. XLVIH

weeks later, however, males were found in the same area
in large numbers. As these males were wingless, they
had probably not immigrated. The only other explana-
tion that occurs to me is that the larvae were present in
considerable numbers at the time of the earlier collec-
tions, but in the flowers, not among the spikelets of the
timothy, so that I did not discover them; and that the
female larvae reached maturity much earlier than the
males. In any case, it is diflScult to see how the males can
have been functional, when the two sexes occurred at dif-
ferent times. If such conditions recur frequently, Chiro-
thrips manicatus, eVen though it produces many males,
must be parthenogenetic.

SUMMABY

The principal conclusions reached in the second "part of
this work may be stated as follows :

Some species of Thysanoptera pass through the winter
in both sexes, in others the males perish. In none of
those studied does the egg or larva live over winter.

Pupation of most of the species of Thysanoptera stud-
ied occurs on the food plants where the larvae live, not-
withstanding that the pupae seldom appear in collections.

From the determination of the sex ratio, Euthrips
tritici shows no indication of an alternating life cycle. It
is probably sexual throughout the active season, though
this is not proven.

Chirothrips manicatus occurred abundantly in both
sexes, but the two sexes appeared at different seasons.
The explanation of this phenomenon is doubtful.

An increase in the number of males in Anthothrips
verbasci in late summer may be explained as due to the
great length of the life history and to selective mortality
during the winter, without assuming an alternating life
cycle.

Anaphothrips striatus, a species which has hitherto
been known almost wholly in the female sex, produced
about 25 per cent, of males at Douglas Lake. This may

Digitized by

Google

No. 568] BIOLOGY OF THE THYSANOPTERA 247

be due either to climatic conditions or to racial differ-
ences. Sexnal reproduction was not wholly proven, but
seems probable. An increase in the number of males in
late summer in this species and in Thrips tabaci might be
interpreted as indicating a sexual phase, or the vestiges
of a sexual phase that existed in the species formerly,
Jordan's belief in an alternating life cycle, which was
rejected by Uzel, thus receives some measure of jus-
tification.

BIBLIOGRAPHY
Gary, L. B. 1902. The grass thrips (Anaphothrips striata Osborn). Maine

Agr. Exp. Station, Bull. 83, June, pp. 51-82.
Jones, P. B. 1912. Some new California and Georgia Thysanoptera. U. S.

Dept. Agr., Bur. Ent., Tech. Ser. No. 23, Part 1, 24 pp., 7 pis.
Jordan, K. 1888. Anatomie und Biologie der Phjsapoda. Zeit, iviss,

ZooL, Vol. 47, pp. 541-620.
Hinds, W. E. 1902. Contribution to a monograph of the insects of the

order Thysanoptera inhabiting North America. Proc. U, 8, Nat.

Museum, Vol. 26, No. 1310, December 20, pp. 79-242.
Maupas, E. 1890. Sur la fdcondation de THydatina senta Ehr. Comp,

Bend. Acad. 8ci. Paris, Tome 111, pp. 505-507.
Moulton, Dudley. 1911. Synopsis, catalog and bibliography of North

American Thysanoptera. U. S, Dept. Agr., Bur. Ent., Tech. Ser. No.

21, 56 pp.
1912. Papers on deciduous fruit insects and insecticides. IV. The pear

thrips and its control. U. S. Dept. Agr., Bur. Ent., Bull. 80, Part IV,

pp. 51-66.
Shelf ord, V. E. 1911. Physiological animal geography. Joum. Morph.,

Vol. 22, No. 3, September, pp. 551-618.
Shull, A. F. 1909. Some apparently new Thysanoptera from Michigan.

Entom. News, Vol. 20, No. 5, pp. 220-228.
1911. A biological survey of the sand dune region on the south shore of

Saginaw Bay, Michigan. Thysanoptera and Orthoptera. Mich. Geol.

and Biol. Survey, Pub. 4, Biol. Ser. 2, pp. 177-231.
Uzel, H. 1895. Monographie der Ordnung Thysanoptera. KOniggratz,

privately published, 482 pp., 10 pis.

Digitized by

Google

SHORTER ARTICLES AND DISCUSSION

BARRIERS TO DISTRIBUTION AS REGARDS
BIRDS AND MAMMALS

The geographical range of any species of animal may be
likened to a reservoir of water in a mountain canyon. The con-
fining walls are of varying nature. A concrete dam, absolutely
impervious, may retain the water at one end. Along either side
the basin's walls differ in consistency from place to place. The
substratum varies in porosity, at some points being impervious
like the dam, at others permitting of seepage of water to a greater
or less distance from the main volume. The water continually
presses against its basin walls, as if seeking to enlarge its area.
And it may succeed in escaping, by slow seepage through such
portions of its barrier as are pervious or soluble, or by free flow
through a gap in the walls, if such ofiPers. The area occupied
by the water will extend itself most rapidly along the lines of
least resistance.

Every species has a center or centers of abundance in which
favoring conditions usually give rise to a rate of reproduction
more than sufficient to keep the critical area stocked. A tendency
to occupy a larger space results, because of competition within the
species: individuals and descent-lines multiply and travel radi-
ally, extending those portions of the frontier where least resist-
ance is offered. Such radial dispersal takes place slowly in some
directions, more rapidly in others, according to the degree of
passability of the opposing barriers. These barriers consist of
any sort of conditions less favorable to the existence of the
species than those in the center of abundance.

Theoretically, sooner or later and in all directions, every
species is absolutely stopped. But as a matter of undoubted
fact most barriers are continually shifting, and the adaptability
of the animals themselves may be also undergoing continual
modification ; so that perfect adjustment is beyond the limits of
possibility so long as topography and climate keep changing.
The ranges of species may thus be constantly shifting. Descent-
lines may move about repeatedly over the same general region,
like sparks in the soot on the back of a brick fireplace.

Yet, in all of our studies, of but a few years' duration, the

248

Digitized by

Google

Na 568] SHORTER ARTICLES AND DISCUSSION 249

time element is reduced almost to a negligible quantity, and we
may look upon the areas occupied by each species as, for the
time of our observation, fixed. We are thus enabled to compare
one with another, and because of the large number of the species,
we can infer a good deal as to the nature of barriers in general,
at least as regards birds and mammals. It is even conceivable
that ; with sufScient refinement in methods, the inquirer might in
time find himself able, from a comparative study of the ranges
of rodents, for example, to establish the identity of all of the
external factors which have to do with the persistence of each of
the species; in other words to analyze the ** environmental com-,
plex'' into its uttermost elements — as regards the existing species
of rodents in their recent development.

The most obvious kind of barrier to distribution is that con-
sisting of any sort of physical, or mechanical, obstruction. Such
obstruction aflFects directly the individuals of a species en-
countering it, either by stopping their advance or by destroying
outright such as attempt to cross it. As barriers of this nature,
are to be cited land in the case of purely aquatic mammals, and
bodies of water to purely terrestrial, especially xerophilous,
mammals. In each case the width of the barrier has to do with
the degree of impassability. Oceans and continents are most
perfect, and aflfect a large proportion of the species. The com-
paratively narrow Colorado River is a barrier of the first rank,
but only to a certain few desert rodents. Mechanical barriers,
where they exist at all, are clearly recognizable.

It is to be observed, however, upon considering the birds and
mammals of a whole continent, that by far the greater number
of species are delimited in range without any reference to actual
land and water boundaries ; more explicitly, their ranges fall far
short of coast lines. The barriers here concerned are intangible,
but nevertheless powerful. By their action the spread of species,
genera and families is held in check as surely as by any tangible
obstruction.

By these invisible barriers the individual may not necessarily
be stopped at all, as with animals of free locomotion; but the
species is affected. For example, the mocking bird in its Cali-
fomian distribution is closely confined to those parts of the state
IKMSsessing certain definite climatic features; but vagrant indi-
viduals, especially in autumn, occur far beyond the limits of
these restrictive conditions. Carnivorous mammals are well

Digitized by

Google

250 THE AMERICAN NATURALIST [Vol. XLVIII

known to be subject to sporadic wanderings on the part of indi-
viduals, but the species is kept in set bounds by some potent but
invisible set of factors. The very fact that individuals are quite
capable of temporarily transgressing these bounds and yet do
not overstep them en masse emphasizes all the more the remark-
able potency of this category of barriers as regards species and
higher groups.

Our geographic studies lead us to designate among these rela-
tively intangible barriers: (1) increase or decrease in prevailing
temperature beyond certain critical limits, according to the species
concerned; (2) increase or decrease in prevailing atmospheric
humidity beyond certain limits; (3) modification in food-supply
and appropriate breeding and foraging ground. The limits set
by each of these factors will vary with the physiological pecul-
iarities of the organism considered ; in other words the inherent
structural equipment of each animal figures importantly. In
these three sorts of barriers will be recognized what have been
called ''zonal/' ''faunal" and ''msociationaV delimitation, each
of which I will now try to define.

Two schools of f aunistic students are represented among Amer-
ican zoo-geographic writers of the present day. One, of which
C. H. Merriam is the most prominent exponent, sees in tempera-
ture the chief cause controlling distribution, and deals with the
ranges of species in terms of **life zones.'' The other school, of
which C. C. Adams, A. G. Ruthven and Spencer Trotter are
active advocates, assigns to temperature but a minor role, look-
ing rather to a composite control, of many factors, resulting in
ecologic ** associations," of which plants are essential elements,
and which are to be further explained on historical grounds.
The two sets of areas thus defined do not by any means corre-
spond. Yet the reviewer can not fail to note, here and there,
places where boundaries coincide, and such coincidences are so
frequent as to be suggestive of real concordance in some signifi-
cant manner. Is it not probable that both schools are approxi-
mately correct, the difference in mode of treatment being due to
different weights given the different kinds of evidence, or, in
other words, to difference in perspective?

Every animal is believed to be limited in distribution zonally
by greater or less degree of temperature, more particularly by
that of the reproductive season. When a number of animals
(always in company with many plants similarly restricted)

Digitized by

Google

No. 568] SHORTER ARTICLES AND DISCUSSION 251

approximately agree in such limitation they are said to occupy
the same life zone.

The observation of this category of distributional delimita-
tion is particularly easy in an area of great altitudinal diversity
like that comprised in the southwestern United States. The
writer is led to wonder if those authors who minimize the impor-
tance of temperature have ever been privileged to travel exten-
sively, and carry on field studies, outside of the relatively uni-
form eastern half of North America !

Study of any area which varies widely in altitude and hence
provides readily appreciable diflEerences in daily temperature
from place to place brings conviction of the very great eflEective-
ness of temperature in delimiting the ranges of nearly all species
of animals as well as of plants. Particular attention may be
called to the pertinent results of Merriam's survey of Mount
Shasta.

But temperature is not to be considered the only delimiting
factor of environment, though its possible overemphasis by the
Merriam school seems to have led some other persons to believe
that this view is held. In fact it becomes evident, after a con-
sideration of appropriate data, that very many species are kept
within geographic bounds in certain directions only by an in-
creasing or decreasing degree of atmospheric humidity. By
plotting the ranges of many animals as well as of plants coin-
cidence in this regard is found in so many cases as to warrant
the recognition of a number of **faunal areas" — on the causa-
tive basis of relative uniformity in humidity. It is probable that
every species is affected by both orders of geographic control.

The reader may enquire as to the grounds for employing the
widely used terms zone and fauna in the restricted sense here
prescribed. In reply, it may be said that this is not an inno-
vation, but is an adoption of a usage which has come about his-
torically among a certain group of workers in the geography of
vertebrate animals in North America. The writer recognizes the
fanlt in imposing restricted meanings upon old terms, but he
also hesitates at coining new words.

As to which is the more important, assembled data seem to-
show that more genera and higher groups are delimited by zonal
bomidaries than by faunal boundaries. The arresting power of
temperature barriers would therefore seem to be relatively the
greater.

Digitized by

Google

252 THE AMERICAN NATURALIST [Vol. XLVni

In the third category of distributional control there is a con-
spicuous association of the majority of so-called adaptive struc-
tures of animals (often of high taxonomic value) with certain
mechanical, or physical, features of their environment An
animal may thus intimately depend upon certain inorganic or
organic peculiarities, or both, of a given area, and be unable to
maintain existence beyond the limits of occurrence of those
features of the environment. Tracts of relatively uniform en-
vironmental conditions, including their inanimate as well as
living elements, are here called associations.

After a consideration of all the birds and mammals occurring
both within the state of California and elsewhere as far as the
writer's knowledge goes, associational restriction appears to be
governed by the following three factors, of relative importance
in the order named.

1. Kind of food-supply afforded, with regard to the inherent
structural powers of each of the animals concerned to make it
available.

2. Presence of safe breeding places, adapted to the varying
needs of the animals, in other words depending upon the respect-
ive inherent powers of construction, defence and concealment
in each species concerned.

3. Presence of places of temporary refuge for individuals,
during daytime or nighttime, or, while foraging, when hard-
pressed by predatory enemies, again correlated with the respec-
tive inherent powers of defence and concealment of each species
involved.

It is believed that the geographical distribution of any animal
is correctly diagnosed in terms of each of the three main group-
ings here suggested. In other words an animal belongs simul-
taneously to one or more zones, to one or more faunas, and to
one or more associations. No one of these groupings can be
stated in terms of the other, any more than a person can com-
pute liquids by candle-power, or weight in miles. The constit-
uent species within each of these groupings always belong to
the other two. To illustrate: the southern white-headed wood-
pecker inhabits the coniferous forest association of the San
Bernardino fauna of the Transition zone; the Abert towhee be-
longs to the mesquite and the quail-brush associations of the CJolo-
rado Desert fauna, of the Lower Sonoran zone ; the Pacific shrew
belongs to the upland riparian association of the northern coast
redwood fauna of the Transition and Boreal zones.

Digitized by

Google

No. 568] SHORTER ARTICLES AND DISCUSSION 268

Classification of Barriers to Species as Regards
Birds and Mammals
Barriers:

A, Intangible.

(a') Zonal (by temperature).

(6') Faunal (by atmospheric humidity).

(c') Afisociational.

(1) By food supply.

(2) By breeding places.

(3) By temporary refuges.

(Each of these three with regard to the inher-
ent structural characters of each species
concerned.)

B. Tangible (mechanical).

(a") Land to aquatic species.

(6") Bodies or streams of water to terrestrial species.

The above categories are believed to include all the factors
commonly involved in checking the spread of species of birds and
mammals. It is possible that inter-specific competition may
sometimes occur where associational homologues meet. But even
here it becomes a matter of relative associational fitness which
determines supremacy and consequent ultimate limits of inva-
sion of the forms concerned.

A mountain range, mechanically speaking, is no barrier at all,
per se, as frequently alleged. Only as it involves zonal or faunal
barriers does it aflfect distribution. The same is true of a valley
or a desert.

As far as contemplation of cases has gone, the writer's experi-
ence has led him to believe that the outlines of the ranges of all
birds and mammals may be accounted for by one or more of the
factors indicated in the analysis here presented. And as de-
tailed knowledge of the facts of geographical distribution accu-
mulates, the delimiting factors become more and more readily
detectable. By such a study, of comparative distribution, it
seems possible that the ranges of birds and mammals may become
subject to satisfactory explanation.

When considered in its historical bearing, the problem of
barriers concerns itself intimately with the origin of species. It
is believed by the writer that only through the agency of barriers
is the multiplication of species, in birds and mammals, brought
about.

Digitized by

Google

264 THE AMERICAN NATURALIST [Vol. XLVIH

The present contribution is abbreviated from-a general discus-
sion of certain distributional problems which forms part of a
paper to appear from the University of California press and
which treats in detail of the birds and mammals of the lower
Colorado Valley, in California and Arizona.

Joseph Gbinnell
Museum or Vertebrate Zoolooy,
University op Calitc>rnia

YELLOW VARIETIES OF RATS

In a recent number of the Naturalist I described a yellow
variety of the common rat {Mus norvegicus) which in recent
years made its appearance in England and is now a recognized
variety among fanciers. Dr. John C. Phillips and Professor L.
J. Cole have both called my attention to a fact which I had over-
looked; namely, the occurrence of a yellow variety in another
species of rat {Mus rattus) . Bonhote described the occurrence of
this variety in Egypt in 1910 and has since found by experiment
(1912) that the yellow variation of Mus rattics is recessive in
heredity precisely as it is in Mus norvegicus. The fact that the
yellow variation in mice is dominant in heredity, but can not be
obtained in a homozygous condition, stands, therefore, as a phe-
nomenon all the more singular and striking.

W. E. Castue.

BussET Institution,
March 3, 1914.

Digitized by

Google

NOTES AND LITERATUEE

HEREDITY AND ''THE INFLUENCE OP
MONARCHS''

In **The Influence of Monarchs'* (xiii and 422 pp., 1913, The
Maemillan Co., New York, $2.00) Dr. Frederick Adams Woods
makes a second and firmer step along the path entered on with his
interesting ''Mental and Moral Heredity in Royalty" published
in 1906. Dr. Woods's goal in beginning and continuing his an-
alysis of the character of royalties and the circumstances of their
reigns is one probably not immediately to be reached but also
probably one not impossible of attainment. It is indeed not one
goal that he has before him, but two, the ways to which lie close
together and parallel. One is the establishing of a new science of
history to be called historiometry ; the other is the making ap-
parent of the dominan'ce of heredity over environment in deter-
mining human fate.

That the methods and even the aims of most historical study
are not satisfying to all historical students is made obvious by
the constant complaining of historians to and of each other.
There are two conspicuous groups of these protestants, one de-
manding more interest, more imagination, a more literary treat-
ment of historical fact, and the other demanding a more signifi-
cant, more inductive, more scientific treatment. The former
wants more "humanity,'' the latter more biology, in history.
Dr. Woods is of the latter group.

But Dr. Woods is not primarily of any historical camp. He is
biologist, especially evolutionist and student of heredity. How-
ever, he marches very boldly into the ranks of the students of his-
torical human history — to distinguish thus the last few thousand
years of human history from the earlier many thousand years
of it — with the new methods and results of his historiometry,
just as Pearson, several years ago, invaded the biological camp
with his biometry. Something of historiometry in history there
has always been, just as there has always been something of
biometry in biology. But these reformers want to make history
and biology wholly, or, at least, most importantly, sciences of
measure. And each of them finds that his use of measure in them
leads him to discover that the facts that he is measuring offer, in
the new significance they are thus made to yield, a special argu-
ment for some particular one of the major factors in evolution.

255

Digitized by

Google

256 TRE AMEBIC AN NATURALIST [Vol. XLVIH

Biometry emphasizes the enormous importance and significance of
variation in all living things ; historiometry reveals the enormous
importance of heredity in human life and the affairs of society.

After an introductory chapter stating the need of a new inter-
pretation of history and of new methods of getting at this inter-
pretation, and a following general chapter further elaborating
and expanding his views concerning ''the philosophy of history
and historiometry,'' Dr. Woods plunges into a series of compact
histories of France, Castile, Aragon, United Spain, Portugal, The
Netherlands, Denmark, Sweden, Russia, Prussia, Austria, Turkey,
Scotland and England. In each of these he presents a swift sum-
mary of the economic and political conditions (success in wars,
increase in territory and prestige, prosperity, advance, failures in
war, loss of prestige, poverty, retrogression) of these nations in
the various reigns of a period of about 500 years for each country,
together with a statement of the personal traits of each monarch.
In all, three hundred and sixty-eight monarchs, regents or other
rulers, royal or non-royal, and correspondingly, three hundred
and sixty-eight sets, or periods, of national conditions, are pre-
sented.

Prom these data is derived the very positive and important
conclusion that the dominant causal influence in determining the
character of national, political and economic conditions has been
the personality of the monarchs, and that the prime determinant
of this personality is heredity and not environment.

A host of possible criticisms and objections to the method, its
results and their interpretation, leaps into every one's mind.
Well, they are all — or all that I have so far been able to formu-
late— ^anticipated, and ingeniously, and usually convincingly,
answered. At least they are anticipated and discussed. In this
the book reminds one of Darwin's ** Origin of Species."

To all who have read ** Heredity in Royalty" this new book of
Dr. Woods will need no recommendation of its interest and im-
portance. To those who have not, and are interested either as
historian, biologist, or natural philosopher in human history and
the bionomic factors that control it, *'The Influence of Monarchs"
may be strongly recommended as an original and very suggestive
treatment of the subject. To students of heredity the book is a
necessary library addition.

V. L. K

Stanford Univkesitt,
California

Digitized by

Google

The American Journal of Science

Established by Benjamin Sllllman In 1818

The Leading Scientific Journal in the United States

Devoted to the Physical and Nataral SdenoeB, with special reference to
Physics, and Chemistry on the one hand, and to Geology and Mineralogy
on the other.

Editor: BDWABDB.BAHA.

AsBoeUto Bdlton : Proftator OBOSOE L. OOODALE, JOHN TE0WBEU>0E. W. Q. FABLOW and WM. M. DAVIS

of Ommteldge ; ProfMtort A. B. VEBRILL. HENRY S. WILLIAMS and L. V. PIRSSON, of Now

Haren; Proftsaor JOSEPH S. AIMES, of Baltimore ; MB. J. S. DILLEB, of Waalilnfton.

Two volnmes annually, in monthly numbers of about 80 pages each.

This Journal ended its first series of 50 volumes as a quarterly in 1845 ; its second
series of 50 volumes as a two-monthly in 1870 ; its third series as a monthly ended Dec-
ember, 1895. A Fourth Series commenced in 1896.

SnlMcrlptioo price, $6 per year or 50 cents a ntunbert postage prepaid ia the United States
$6.25 to Canada ; $6,40 to Countries in the Postal Union. Back numiiers at reduced prices.
f^^en-Yolume Indexes, Vols. I-X XI-XX, fourth series, price one dollar. Address

The American Journal of Science
New Haven, Conn.

University Control

By J. MoElxBK GATTXLLy Professor of Psychology in CtolumUa XJnivendty

TofiUMr with a Mriet of Two Hundred and Ninoty-4iino Uniignad Lotten by TiwUHng Men of SeieBie
hbUi^ Academio Poritione and Artiolee by Josbvh Jamtbow, Onoaon T. IUim>, Joks J. Sramsov, J. B.
CmonoN, J. if oKmH CArrsLiH Gaoaoa M. Stbatios, SnwAVf Patov, Jon Jat Ghaimah, Jamxs P.
MvvBoa and Jaoob Qould Schubicav.

A great Tariety of Qoeetioni oonoerning general wdrenity adminletration an doatt with in an original and
hetpful way.^ — Nalur^

Theee QOotatiQiie and eauuni»lee are taken from Profeeeor CatteU'e informed and thorough dleooerion of the
Mibieot of tmifemity oontrol, a eubjeot upon which he hae had mnoh to eay of late, finding oooaeion for eanetie
oitkiim of exiBting Ameii an oooditione, and etanding ae the diami>ion of an aeademie demooaesr and a teaeb-
Ing prnfearion upon which a man may enter witi&oot forfeiting hie aeU-reepeot. — Th» IHoL

Sentencee and paragraphe that betoken the eipert, highly*traaned mind, the euggeetione that come to ••>
fnsh a d teQ ne that a new day ie about to dawn in eduoatioQal writing.«>nbe Boilon ^teaiiia IVaateripi.

SCIENCE AIND EDUCATION

A serien of rolumeafor the promotion ofaciontiAo reaeurcb and educational progroBB
VOLUME L The Fmmdations of Sdenee. By H. PonroABi. Containing the anthoriMd Bnglieh tranelatioB

hf Gaoaa Bsooa HALffran of "Soienee and Hypotheeie," "The Value of Soienoe," and "Scienea

■ad Method.** Price, $3.00 net.
VOLUMB n. Medical Reaearch and BdscatioB. By BiOBAan M. PiABcn, Wzluam H. Wblcb, W. H.

HowBLL, FBAjrxuv P. Maix, LawBLLTa F. Babxbb, Gkablbb S. Mmov, W. B. CAmioif, W. T.

GoimoiUitAM, Tbbobau) Sum, Q. N. SrawAsr, C M. Jaoxbom, E. P. Lton, Jambs B. HsBBiog, Joan

M. Dooeov, C. R. Babdbbn, W. OthOia, S. J. MaiAnn, Jambb Ewnio, W. W. Exbh. Hbiibt H. Dosa&d-

eoB, CSBBnnAir A. HaanB, and Hbnbt P. BownrrcH. Price, $3.00 net.
VOLUMB m. Uaivandty Control. ^o«e Ready. Po^eta+i^ FriM,%5MwtL

QARRisoN, N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84. NEW YORK CITY

Digitized by

Google

The American Naturalist

A MoatUr Jomul ••uUkli«d in 1867, D«TOtod Id th« AdTUie«neiit of th« Biologieal ScImcm

CONTENTS OF THE OCTOBER NUMBER

▲ Contribution towtidt an Analysis of ibe Problem
of Inbroedinf . Dr. Baymond PcarL

Tbo Inhtfltanee of Coat Color in Honet. Prof eMor
W.8.Andenon.

The Variations in the Nomberof Yertcbre and Van-
tral Scntas in Two Snakes of the Qenos Beffina.
Professor Alexander G. Rnthyen and Crystal
Thompson.

Shorter Articles and Beporta : The Bimnltaneoas
Modifloation of Distinct MendeUan Factors: .Pro-
fessor B. A. Smerson. TheFoorthlnternatiGnal
Oenetie Conferenee : Dr. Frank M. Snrfaoe.

CONTENTS OF THE NOVBMBEII NUMBER
The Effect on the Offspring of Intoslcatinc the Male

Supplementary Studies on the I>ifflerentlal Mortalitgr
with respect to Seed Weight in theQermlnatioB
of Garden Beans, Dr. J. Arthur Harris.

Shorter Arttdes and Discussion : Bedprocal Crosm
hetween Beere's Pheasant and the Common
Ringneck Pheasant producing Unlike Hyhrids.
John C. Phillips.

CONTENTS OF THE DECEMBER NUMBER

The Fixation of Character in Orranisms. By Edward

Sinsott.
Inheritance of Left-handedness. Professor Francis

Bamaley.
Supplementary Stndies on the DUTerential MortaUty

with Bespect to Seed Weight in the GerminaUon

•f Garden Beans, n. Dr. J. Arthur Harris.
Shorter Articles and Discussion : A Cross InTolTing

Four Pairs of Mendelian Characters in liiee. C.

CLlttte, J. C.Phillips.
Index to Volume ZLVIL

OONTENT8 OP THE JANUARY NUMBER
A Genetic Analysis of the Changes produced hj

fessor E. M. East and H. K. Hayes.
Gynandromorphous Ants, described during the De*

cade. lMa-1918. Professor William Ifdrton

Wheeler.
Shorter Articles and Discussion : On the Seeults of

Inbreeding a Mendelian Population— A OoRee»

tion and Extension of PreTlous Condurioos.

Dr. Raymond Peail-^Isolation and Belaotktt

alUed in Pnndple. Dr. John T. Qultek.

CONTENTS OF THE FEBRUARY NUMBER

Some New Varieties of Bats and Guinea-pifs and their

Belations to Problems of Color Inheritance. Pro-
fessor W. B. Castle.
«< Dominant" and *« BecessiTe" Spottlnff in Mice. C.

C. Little. ^^
On Difterential MortaUty with respect to Seed Weight

occurring in Field Cultures of Pisum sativum.

Dr. J. Arthur Harris.
The Inheritance of a Beeurrlng Somatio Variation

In Variegated Ears of Maise. Professor B. A.

Emerson.
Bestoration of Edaphosaurus crudger Cope. Pro-

fessor E. C. Case.
Shorter Articles and Discussion : Humidity— a

Neglected Factor in EuTironmental Work. Dr.

Frank B. Lutz.

CONTENTS OF THE MARCH NUMBER

Asa C. Chandler.
Biology of the Thysanoptera. Dr. A. Franklin BbnU.
Shorter Articles and Correspondence : The Endemic

CockereU.
Notes and Literature : Swingle on VarlatloB in li
Citrus Hybrids and the Thawy of Zygotasls.
Dr. Orland E. White.

Single Number 40 Cents YeaHy SulMcription, $4.00

THE SCIENCE PRESS

GarritomNeY. Substation 84: NEW YORK Lancaster. Pb.

Digitized by

Google

*, ' 4: 3

701. ZLVm, VO. 569 MAT, 1814

THE

AMERICAN
NATURALIST

A HONTELT JOTJBVAL

])«70ted to the Adyancement of the Biologioal Soiencea wttli

Special Beference to the Factors of EYoIntion

C0HTEVT8

Piige

I. Ectoparasite! of Mammals. Professor Ybknok Ltman Eblloqo - -267

n. Bogeneratloa, Variation and Correlation in Tliyone. Professor John W.
Scott -------------- aso

m. ffiborter Articles ana Discussion : Terms relating to Qenerio Types. Dr. O. F.

Cook --------- -----so8

lY. Votes and Uteratnre: Linkage in the Silkworm Moth. A. H. STiTBTBVAiffT.

Naboors's Breeding Experiments with Grasshoppers. John S. Dbxtbb - 815

THE 80IEN0E PRESS

IiAHOASTEB, PA. OABBIBOH. V. T.

in3W YORK : 8UB-STATI0H 84

Digitized by

Google

The American Naturalist

MSS. Intended for pnblioatlon and books, etc., Intended for review should be
sent to the Editor of THE AMERICAN NATURALIST, Garrison-on-Hodson, New York.

Short articles oontalning snnnaries of research work bearing on the
problems of organio evolution are eepecially weloone, and will be given preference
in publleatlon.

One hundred reprints of contributions are supplied to authors free of charge.
Further reprints will be supplied at cost.

Subscriptions and advertisements should be sent to the publishers. The
subscription price is four dollars a year. Foreign postage is fifty cents and
Canadian postage twonty-five cents additional. The charge for single copies Is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Qarrison, N. Y.

NBW YORK : Sub-Station 84

Entered ei leoond-oUM matter, April 2, 1908, at the Poit Office at Lancaster, Pa., under the Aet ot

Congreii of March 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Eriflrland

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Aodubon.
7 or 8 volumes, please report, stating cask price, stat-
ing condition, binding and dates of volumes.

F. C. HARRIS,
Box 2244 Boston, Massachusetts

TO OOLOGISTS

and ENTOnOLOGISTS

W. F. H. ROSENBERG

Importer of Exotic Zoologioal Specimens
57, Haverstock HUl, London, N. W., England

Begs to announce the publication of a new Price
List (No. 20) of Mammals, including over 400 species
from yariouB parts of the world.

These will be mailed free on application, as well
as any of the following lists:

Birdskins (oyer 5000 species) ; Birds' Eggs (over
1,100 spedes) ; Beptiles, Batrachians and Fishes
(oyer 400 species); Exotic Lepidoptera (oyer 8,000
species).

Laigest stock in the world of specimens in all
branches of Zoology.

All Museums and Amateurs
should write for these Lists

All specimens sent on approyal.
Please state which lists are required and giye
name of this periodical.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post
Cards 5 cents.
8x10 Enlargements 25 cents mounted.
Send for Samples and Price List.

FRANK R. BARBEAU
89 West Bridge Street Oswego, N. Y.

The University of Chicago

Offers inatntction during the Sum-
mer Quarter on the same basis aa
during the other quartefs ci the
academic year.

The undmnuluate colleges, the
graduate acLoola, and the proles-
siooal schools provide courses ia
Art9, Litmraturm, Sdmnemp
Coaunerem and A^btunimim*

MitdieU Tower

twn. Law, Mmdieinm, EA

tion» andDwinHy. Instruc

is given by tegular members of the
University staff which is augmented
in the summer by appomtmeiu of
professors and instmctocs from
other institutioos.

Swuner Quvter. 1914

r ad Tsrm July 23-Aa«. 28
Detailed announoesients wiU be
sent upon application.

Thm Unirersifr of Chicago

f^t«»»yp^ Illinois

Digitized by

Google

Digitized by

Google

258 THE AMERICAN NATURALIST [Vol. XLVIII

lated from the rest of the individuals comprising the par-
ticular species represented by it. This brings about cer-
tain striking conditions of abundant small variation and
subspecific (or intraspecific) distinction, which, however,
because of the general similarity of habitat, food and
habit, do not tend to grow rapidly into large (specific,
generic, family) diflFerences. The hundred or more species
of Mallophaga so far recorded from mammals have, until
very recently, all been ascribed to two genera, of which one
included nearly nine tenths of the total number of kinds.
There has been made a beginning — and not a particularly
convincing one — at breaking up this inclusive genus
(Trichodectes). It is a movement suggested more by the
needs of convenience than the needs of expressing a bio-
logical situation. Similarly, although not representing
so extreme a condition of likeness, the Anoplura, also
including about a hundred parasite species (occurring
only on mammals) have been, until recently, divided into
but half a dozen genera, with the great majority of the
species included in one. Certain aberrant forms found
on man, the monkeys, the elephant, and on seals and
walruses have always made necessary the recognition of
four or five quite distinct genera. Attempts, however, are
now being made to break up the unwieldly genus Hcema'
topinus.

As this paper is, in eflFect, a continuation of my paper
on ''Distribution and Species-forming of Ecto-parasites''
published in The American Naturalist in March, 1913,
which devoted itself to a consideration of the Mallophaga
(some 1,400 species as so far known) found on birds, and
to the problems presented by their conditions of life and
their host and geographic distribution, I can dispense
with any further account of the special biology of these
parasites by referring the interested reader to this
former paper. In it I have set out rather fully the spe-
cial structural and habit features of the Mallophaga.
Except that the Anoplura 'take blood, rather than
feathers and hair, for food, and have specially modified

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 259

mouth parts to do it with, and are perhaps even more
specialized in their physiological adaptations to their
host than the biting lice, most of the general remarks
made concerning the Mallophaga will apply to the suck-
ing lice also.

In their peculiar special relations to their hosts as per-
manent ectoparasites on them, wingless, and reluctant to
migrate even with opportunity, and so fitted physiologi-
cally to their parasitic life that they can not live for more
than a few hours (or, at most, and exceptionally, days)
off the bodies of their hosts, the Anoplura and Mallophaga
are alike. And hence the conditions and problems of
their distribution and species-forming are practically the
same for the two groups.

The thesis that I have maintained, on a basis of the
conditions presented by the bird-infesting Mallophaga,
I now wish to test by the conditions presented by the
mammal-infesting Mallophaga and Anoplura. This thesis
is, in fewest words, that the host distribution of these
wingless permanent ectoparasites is governed more by
the genetic relationships of the hosts than by their geo-
graphic range, or by any other ecologic conditions. The
fact, proved by abundant cases, that two host species of
wholly distinct geographic range and with no possible
opportunity for contact such as would permit of the
migration of wingless parasites from one to the other,
may have, nevertheless, one or more parasitic species
common to them both, is associated almost always with
the further fact that these common hosts are closely
related genetically. They are most often of the same
genus or of closely allied genera; they are almost cer-
tainly always of the same subfamily or family. The ex-
planation for the possibility and the reality of this inter-
esting host distribution I find in the hypothesis that the
common parasite species has persisted unchanged from
a common ancestor of the now divergent but allied host
kinds.

Also, if it be true that genetic relationship is the deter-

Digitized by

Google

260 THE AMERICAN NATURALIST [Vol. XL VIII

mining factor in accounting for the host distribution of
the parasites, then it is also true that the distribution of
the parasites will indicate in some measure the genetic
relationships of the hosts, and that occasional aid in
determining the genetic affinities of birds and mammals
of doubtful relationships may be had from a study of
their parasitic fauna. In my paper already referred to
I have pointed out some suggestive cases of this sort in
connection with the birds and their parasites.

In examining the conditions existing among the mam-
mals and their Mallophagan and Anopluran fauna, the
first necessity was the compilation of a complete record
or catalogue of mammalian hosts and their parasites,
together with the record of the actual locality of each
finding of parasites, together with a general record of the
geographic range of all the various hosts. This cata-
logue, or set of records, I have now completed, and
despite its meagerness compared with the similar cata-
logue of the bird hosts and their Mallophagan parasites
from which the notes for the former paper were drawn,
it contains enough records of interest to make worth
while a preliminary report on the condition obtaining
among the mammals and their parasites.

It is unfortunate that, although there are nearly one
fourth as many mammal species as bird kinds, only about
one hundred mammals figure in the Mallophagan host
list, while Mallophagan parasites have been taken from
over eleven hundred bird species. Also, only one hun-
dred different Mallophaga have been taken from mam-
mals, while about fourteen hundred have been taken
from birds. Of the Anoplura, which are found only on
mammals, records have been made from about one hun-
dred host species, these records referring to just about
the same number of Anopluran kinds. Thus the mam-
malian host catalogue with its list of parasites is a short
one ; as far as it goes, however, it is thoroughly interest-
ing and suggestive.

In working up the records I have used Trouessart's

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 261

'^Catalogus Mammalium" as an authority for the synon-
omy of the hosts, and my own judgment, based on a con-
siderable personal knowledge of the parasites and on a
careful consideration of all the more intelligible litera-
ture of the two groups, as a last court for the synonomy
of the Mallophaga and Anoplura species. The synonomy
of the parasites I have, however, not pushed far.

With so much of introduction and explanation we may
come to a swift resume of the results of a scrutiny of
these records, proceeding by sequence of the mammalian
orders, and referring to either or both groups of para-
sites as they may happen to be represented in the para-
site records of the successive host groups.

The Marsupialia are represented in the host list by
half a dozen species of kangaroos and wallabies (family
Macropidae) all from Australia, and a wombat, Phasco-
lomys ur sinus (family PhalangeridaB), from Tasmania
(also S. Australia?). From all of these hosts only Mallo-
phaga are recorded, no Anoplura having yet been taken
from a marsupial. The six species of kangaroos repre-
sent three genera (Macropus, Petrogale and ^prym-
nus), and their Mallophaga are of seven species, repre-
senting four genera. Four of the species belong to the
genus Boopia, and I strongly suspect are not all different.
In addition there is one Trichodectes, from Petrogale
penicillata, one Latumcephalum, from ** wallabies,** and
one Heterodoxus, which is recorded from Macropus
giganteus in Australia as well as from the same host in
the Jardin des Plantes, Paris. It is also recorded from
an undetermined wallaby in Victoria and one in Queens-
land, as well as appearing in three other records from
** kangaroo'* or ** wallaby** from Australia. The para-
site of the wombat is a species of Boopia, and it has been
twice recorded from the same host. It is interesting that
the kangaroo in the Jardin des Plantes harbored, even
after some period of captivity, only its own proper para-

Digitized by

Google

262 THE AMEBIC AN NATURALIST [Vol. XL VIII

sites without accepting new ones from its many, various
and closely pressing neighbors.

Of the four Mallophagan genera found on the kanga-
roo, three, namely, Boopia, Latumcephalum and Hetero-
doxus^ are peculiar to them. The third genus, Tricho-
dectes, is represented by but a single species which has
been recorded but once. This is the common Mallophagan
genus of mammals generally. The record is perhaps a
good one, but its lack of confirmation by being unrepeated
either for the same species or for any other species of
Trichodectes, is suggestive. Heterodoxus, Latumcepha-
lum and Boopia are two-clawed genera ; that is, they are
Mallophagan forms which belong to a family all the other
genera of which are confined to birds. The characteristic
structural difference between the mammal-infesting
Mallophaga and the bird-infesting species is the presence
in the first group of a single claw on each tarsus, and in
the second of two claws. This difference is plainly an
adaptive one concerned with the fitting of the foot for
the seizing of hairs and scrambling about among them,
on the one hand, and the manipulation of feathers and
moving about on them, on the other. In examining living
specimens under the microscope the special use and fit-
ness of the feet, in the one case adapted to hairs and in
the other to feathers, is obvious. However, Heterodoxus,
Latumcephalum and Boopia, and, in addition, perhaps
one other doubtful genus, represented by one species, and
perhaps two or three species of another two-clawed
genus, constitute exceptions to the general rule. It is of
decided interest to note that the only genera of two-
clawed Mallophaga found exclusively on mamtnals are
limited to the Marsupials. The antiquity and isolation of

1 The single valid species of this genus — the two or three that have been
named are undoubtedly all the same — ^has also been recorded from dogs 1 In
fact specimens in my own collection were received with the record *'from
Japanese dog. ' * And Enderlein has recorded it from a dog from China and
Neumann from a dog from Formosa. Yet dogs ordinarily do not harbor
this parasite, and kangaroos and wallabies do. It seems necessary to be-
lieve that the dog host records indicate cases of straggling from kangaroos
in zoological gardens or menageries.

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 263

this host group strongly suggests that the one-clawed con-
dition common to all other mammal-infesting Mallophaga
is a derivative from the original two-clawed condition
characteristic of the parasites of birds and of these ancient
manmials. The two-clawed condition is, of course, the one
common to insects generally and is characteristic of the
Atropids, in whom I am inclined to see the ancestors, or
near-ancestors, of the Mallophaga. All of the Anoplura,
it may be added, which are exclusively mammal-infesting,
are one-clawed.

In this connection the suggestiveness of the fact that
in face of the examination of many specimens of half a
dozen species of kangaroos and wallabies, no Anoplura
have yet been found on the Marsupials, may be referred
to. I am coming strongly to believe that there is no such
wide ordinal separation of the Mallophaga and Anoplura
as our clinging to the fetich of *' biting and sucking
mouth-parts'' as basis for radical classificatory separation
has led us to effect. I believe, with Mjoberg, that the two
groups of parasites have a fairly near genealogical
aflSnity, their differences, which are particularly those of
mouth-parts, being adaptive rather than palingenetic in
character. The Anoplura have gone on from the Psocid-
Mallophagan condition to a more specialized parasitic
habit, and are the extremes of a general line of ecto-
parasitic evolution. The absence of sucking lice from the
kangaroos may mean that the Marsupials are older than
the Anoplura! No other considerable group of mam-
mals, except certain families of strong-smelling Carni-
vora, is free from the blood-sucking parasites.

There are but two Edentates in the host list, one, the
Cape Ant bear, Orycteropus afer (family Orycteropo-
didae) of south and central Africa, harboring a sucking
louse, of genus and species peculiar to it, and the other,
the three-toed sloth, Brady pus tridactylus (family Brady-
podidae) of eastern South America, harboring a Mallo-
phagan of species peculiar to it but of the genus Gyropus
which is the less scattered, although still rather catholic,

Digitized by

Google

264 THE AMERICAN NATURALIST [Vol. XLVIH

genus of the two large ones characteristic of the
mammals.

The large order Ungulata, with its nnmerons domesti-
cated and semi-domesticated species, is a favorite host
group with both Mallophaga and Anoplura. Altogether*
about thirty Anoplura and two dozen Mallophagan spe-
cies are recorded from fifty host species representing nine
Ungulate families.

The family Elephantidae is represented by the African
and Indian elephants, recognized as distinct species of
distinct geographic range. They both harbor a common
Anopluran species, HcBmatomyzus elephantis, of species,
genus and family peculiar to the elephants. Fahrenholz
has given the varietal name sumatranus to specimens of
these sucking lice taken from an Indian elephant in
Sumatra. Records show that the parasites have been
taken from their elephant hosts not only in Africa and
Asia, but in various zoological gardens, as Paris, Ham-
burg and Rotterdam.

The small family of Hyracidae, or conies, is represented
in the host list by two species and perhaps a third one,
one of which, the Syrian coney of west and south Asia,
harbors one Anopluran and one Mallophagan, while from
the other, the Cape coney of South Africa, the same
Anopluran species is recorded as well as another of the
same genus. This record of a second species is from a
coney in the London Zoological Gardens. From the pos-
sible third species of Hyrax (taken in the African Congo
and perhaps, but not probably, also a Cape coney), a
second Mallophagan species is recorded of the same
genus, Trichodectes, to which that of the Syrian coney
belongs.

In the family Equidae three species, the horse, the
donkey and BurchelPs zebra, all suffer from the infesta-
tion of a common Anopluran species, Hcematopinus asini.
In addition, the horse and the zebra have a common

2 The sjnoDomy in the parasite records, and indeed in the host records
as well, is a vicious tangle. I have done the best I can, for the present.

Google

Digitized by

No. 569] ECTOPARASITES OF MAMMALS 265

Mallophagan parasite, Trichodectes parumpilosus, while
the horse and donkey have another common biting louse,
Trichodectes pilosus. Two varieties of Trichodectes
parumpilosus have been named by Piaget, one from the
zebra and another from * kittle horses of Java/*

The pigs (family SuidsB), of which three wild African
species besides the familiar animal of the barnyard are
fouBid in the host list, are infested by two (perhaps three)
species of Anoplura and one (a not too certain record)
Mallophagan. Hcematopinus suis is found on the domes-
tic Sus all over the world, while Hcematopinus latus of
Neumann, H. phocochoeri of Enderlein and H. peristictus
of Kellogg and Paine, which are almost certainly all one
species, are recorded from the wart hog, Phacochcerus
cethiopicus from Nyasa-land, Africa, and probably also
from another wart hog species from Africa, and the Red
River hog, Potamochcerus chceropotamus from Nyasa-
land, Africa. In addition Potamochcerus demunis (prob-
ably),rfrom German east Africa, is credited by Stobbe
with a Mallophagan parasite peculiar to it, Trichodectes
vosseleri Stobbe.

The peccary, Dicotyles tajacu (family Dicotylidae) of
Central wAmerica and southwestern North America, has a
Mallophagan species peculiar to it, belonging to the
smaller of the two large Mallophagan genera, namely,
Gyropus.

The dromedary, of north Africa and western Asia, and
the bactrian camel, of central Asia, harbor a common
sucking louse, Hcematopinus cameli. A doubtful second
species called H. tuherculatus (Neumann thinks it iden-
tical with cameli) has been recorded from a dromedary
imported from India into Australia. The ** South Amer-
ican camel,*' the llama, harbors an Anopluran species
peculiar to it, and two Mallophagan species, Trichodectes
hreviceps Rudow and T. incequalemaculatus Piaget. Al-
though Rudow's species are often suspect, I have just
had his hreviceps from a llama of Peru (collector C. H.

Digitized by

Google

266 THE AMERICAN NATURALIST [V0L.XLVIII

T. Townsend). With these llama Mallophaga there is
also a small Anopluran which I have not yet worked out.

The family Cervidae is represented in the host list by
about ten species. They are infested by three species of
Anoplura, each peculiar to its host, and six species of
Trichodectes (Mallophaga) of which T. tibialis is com-
mon to the roe deer of Europe and Asia Minor, an
African Capreolus, and our own black-tailed deer of the
western states. Trichodectes longicornis is common to
the red deer of Europe and Asia Minor and the fallow
deer of south Europe, Asia Minor and north Africa.

The giraffe (family Giraffidae) harbors a sucking louse,
Linognathus brevicornis, peculiar to it.

The great family Bovidse, with its many buffalo, buck,
sheep, goat and antelope kinds, is represented in the host
list by five or six species of Bos, four African bucks,
three or four sheep, the ibex, chamois and two or three
goats, and five or six antelopes, or gazelles. The domes-
tic ox, Bos taurus, harbors three species of Anoplura and
one Mallophagan. Curiously, none of these species is
recorded from any other Bos. On the other hand, the
zebu, the Indian buffalo, and the American bison all
have the same Anopluran species (and no other, nor any
Mallophagan), while the yak of central Asia and the
Kaffir buffalo each have an Anopluran peculiar to it.
The four species of African reedbucks and duikerboks
have, according to the records, each a peculiar species of
sucking louse. These records need scrutiny. One of
them is my own, but I had to describe the species without
seeing the types of the others. The domestic sheep
carries two Anopluran species and one Mallophagan.
The latter occurs also on at least two wild species of
Ovis, one of west Africa and the other of north Africa.
The fat-tailed sheep has a record from German south-
west Africa of a Trichodectes of its own.

The domestic goat harbors one Anopluran and at least
one Mallophagan, the latter being common also to the
Angora goat, the chamois, and a wild ( ?) goat of Guinea,

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 267

and a wild (?) goat of Java. A recent description of a
second Mallophagan species from the domestic goat is
not convincing. The chamois has also an Anoplnran, but
one, so far, peculiar to it. Three species of Gazella (or
Antilope) have three species of Trichodectes, of which
one is common to two host species, one of Arabia and
Syria and the other of north Africa and southwest Asia
generally. This same Trichodectes is also recorded from
the roan antelope, Hippotragus equinus, of east central
and south Africa. One species of Gazella carries an
Anopluran peculiar to it, as does also Tragelaphus grains
of west Africa.

The order Carnivora is represented in the host list by
eight families and a total of fifty-four species. Only one
species of Anopluran, the common sucking louse of the
dog (not found yet even on the wolf or fox, both of which
have other records) is recorded from a Carnivore, outside
of the two families Trichechidae (walruses) and Phocidie
(seals and sea-lions). From these two families, on the
other hand, only Anoplura are recorded.

The family FelidsB is represented by three species, the
domestic cat, tte California lynx and the tiger. The cat
and lynx have a common Mallophagan parasite, Tricho-
dectes subrostratus (and no other), while the tiger has a
biting louse presumably peculiar to it. The description
of this parasite is, however, very brief and unsatisfactory.

The family Viverridae, mongooses, ichneumons and
genets, is represented in the host-list by eight species, of
which five are of the genus Herpestes. Two of these
Herpestes species, one of southern Spain, north Africa
and Asia Minor, the other of west, east and south Africa,
harbor a common Mallophagan parasite. A record of
the finding of Trichodectes subrostratus, the familiar
biting louse of the cat, on Herpestes pluto, comes from
the Kameroons (Africa). It is probably a case o©
straggling, the mongooses being common enough in gar-
dens, and some of them fairly domesticated.

Of the family Canidae there are records from eleven

Digitized by

Google

268 THE AMEBIC AN NATURALIST [Vol. XLVin

species, including the domestic dog, a wild dog of South
America, two wild dogs of Asia, two foxes, and a wolf.
The domestic dog has a familiar sucking louse and is also
credited with that problematical normal or straggling
biting louse of a peculiar genus which I have referred to
in my account of the parasites of the kangaroos Tricho-
dectes latus, the common biting louse of the domestic dog,
is also conmion to the wolf, Canis lupus, of Europe and
Asia, and to the raccoon-like wild dog, Nyctereutes pro-
cyonoides, of Asia and Japan. The record of this last
came, it must be noted, from the Berlin Zoological Gar-
dens. There is no other record of commonness of para-
site to two hosts in the family. The English fox has a
single Mallophagan species, and the California fox has
another. The dhole, a wild dog of the Himalayas, has a
Mallophagan species, and the Magellan wolf of Patagonia
has another.

The family Procyonidae is represented in the host-list
by two raccoons, the California ring-tailed cat, and two
coatis of Central and South America, respectively. The
two raccoons, Procyon lotor of North America and Pro-
cyon psora of California, harbor a common Mallophagan
parasite. In addition a German record (from a zoolog-
ical garden?) credits Procyon lotor with carrying also a
Mallophagan which is the characteristic parasite of the
badger. On the California ring-tailed cat, Bassariscus
astuta, have been found two Mallophagan species, one of
which is the characteristic parasite of the skunks of
North and South America. The two coatis, Nasua narica
and Nasua rufa, one of southwestern United States,
Mexico and Central America, and the other of South
America from the equator south, both harbor a common
Mallophagan species.

The family Mustelidae, comprising the badgers, wea-
sels, martens, and skunks, an ill-smelling crew, offers no
attraction to blood-sucking parasites, but is represented
in the host-list by nearly twenty species from which
Mallophaga have been taken. The Old World badger has

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 269

a characteristic species, Trichodectes crassus. The mar-
tens, weasels and ermine have also a characteristic spe-
cies, Trichodectes retusus, which is recorded from the
pine marten of Europe and northern Asia, the beech
marten of the same range, still another Old World mar-
ten, the weasel of Europe and Asia, the ermine of north
Europe, Asia and America, and the weasel and mink of
North America, in all six or seven species of Mustela and
Putorius of very wide geographic range. The skunks of
North and South America have also a characteristic
Mallophagan species, Trichodectes nephitidis, described
by Osbom from the conunon North American skunk.
Mephitis mephitica, taken in Nebraska. I have found
this parasite on the western skunk, M. occidentalis, in
California, and on M. macrura of Arizona. It has also
been recorded from the spotted skunk, Spilogale inter-
rupta, of the southern United States, Mexico and Central
America, and I have examples from a *' skunk *' of Bolivia.
It is also recorded from a Chilian Mustelid, Galictis
quiqui, which ranges over South America from the Eiver
Plate south, and from another species of Galictis in
Brazil. Finally, examples of this ubiquitous pest are
recorded from Helictis everetti from North Borneo I The
last record comes from Neumann, a very careful and
weU-inf ormed student of the parasites, but his specimens
were taken from a skin in the Museum of Natural History
of Paris. The Old World otter, Lutra Intra, has a Tri-
chodectes of its own, as has also an African otter, L.
matschiei, and the North African Zorilla lyhica.
Mjoberg records a species of Boopia (typical kangaroo
parasite genus) from Lutra pruneri of India. As the
record is an extraordinary one, being the only case of a
Boopia found outside of Australia or on a mammal other
than a Marsupial, it is well to note the exact circum-
stances of the record. The parasites (several examples)
were got by Mjoberg from the Hamburg Zoological Mu-
seum where they were ticketed as having been taken
from a ^'soeben frisch angekommenes Thier" of the
species Lutra pruneri, the animal having been received

Digitized by

Google

270 THE AMERICAN NATURALIST [Vol. XLVHI

from India. There are to be considered in connection
with this extraordinary record, first, the possibility of an
exchange of labels in the course of the several handlings
of the Mallophagan specimens, and, second, the possibility
of a favorable answer to the question: Is Lutra pruneri,
which does not appear in Trouessart at all, only Lutra
lutra, the common Old World otter, and was the speci-
men from which the Mallophagan came a resident in a
zoological garden in which kangaroos or wallabies also
lived, affording a bare chance of straggling? The similar
aberrant records from dogs of the kangaroo parasite
Heterodoxus have already been referred to.

The bears (family Ursidae) have, so far, but one para-
site record to present, a Mallophagan species, Tricho-
dectes pinguis, having been described from the Thibetan
bear, Ursus thibetanus, a century ago.

The walrus (family TrichechidaB) harbors a strange
Anopluran parasite of species, genus and family peculiar
to its host, as, indeed, might be expected of any ecto-
parasite daring enough to brave comrade life with wal-
ruses. Examples of the parasite have been taken from
walruses from Spitzbergen, Frobisher Bay (Davis
Straits), the Hamburg Zoological Garden, and I have
recently had them from a ** Pacific walrus'' from ** south-
east of Siberia."

The family Phocidae is represented in the host-list by
at least five species of seals and sea-lions carrying an
equal number of Anopluran species representing three
different genera, all of them peculiar to the seals. A
single parasite species, Echinopthirius phocce has been
repeatedly taken from the fur seal, Proca vitulina, from
both Old World and New World shores. The harp seal
of the Arctic is credited with the same parasite, as well
as another. Hooker's seal of New Zealand and the Auck-
land Islands carries an Anopluran, Antarctopthirius
macrochir, of species and genus peculiar to it, while the
elephant seal of the south Pacific has another parasite
also of genus and species peculiar to it.

The large order Rodentia is well represented in the

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 271

host-list, representatives of thirteen families, summing
about sixty species, being listed. Both Mallophaga and
Anoplura infest the rodents, but certain families are
parasitized almost or quite exclusively by Anoplura,
while Mallophaga are the only parasites of others.

The SciuridsB (squirrels and spermophiles), for ex-
ample, with a dozen host species, are parasitized by a
dozen species of Anoplura with only a single Mallo-
phagan record; and a single record under such circum-
stances is always suspect. There is little commonness of
parasite species to two or more host species in this
family. Osbom's Polyplax montana is recorded from
the eastern and western North American gray squirrels,
and his P. suturalis has been taken from two Spermo-
phile species, both, however, of the same general range.
The well-differentiated parasite genus Acanthopinus is
represented by one species from the common Old World
squirrel, Sciurus vulgaris, and another from the eastern
gray squirrel of North America. These species, though
close together, really seem to be different. In addition I
have just found the Acanthopinus species of the eastern
gray squirrel on Douglas's squirrel in California, and
another (new) species on a California chipmunk. The
only Mallophagan species recorded from a Sciurid is
Gyropus turbinatus from the marmot, Arctomys mar-
motta, of the mountains of southern Europe.

From the beaver (family Castoridae) a characteristic
Mallophagan species, Trichodectes castoris, has been
taken in America. The beaver, it may be noted, is the
host of the only beetle (Platypsylla cast oris) that has
become a specialized permanent ectoparasite, passing its
whole life on the body of its host.

The Old World dormouse (family Gliridae or Myoxidse)
harbors a sucking louse, Polyplax pleurophcea.

The large family Muridae, including the rats, mice,
voles and lemmings, is represented by twenty host species
well scattered over the world. There are twenty-two
Anopluran species and two Mallophagan species in the
parasite list for the group. Both of these Mallophagan

Digitized by

Google

272 THE AMERICAN NATURALIST [VoL.XLVm

records are my own. One is a new species of Colpoce-
phalum (exclusively a bird-infesting genus) from a
''spotted rat,'* Uganda, Africa, sent me by Sjoestedt in
a collection made by the Swedish Zoological Expedition
to Kilimandjaro-Meru, Africa, in 1905-1906. It is un-
doubtedly a straggler from some bird taken at the same
time. The other is a poor specimen of Trichodectes from
Mus rattus, Canal Zone, Panama, sent me by Dr. Jen-
nings. It may be a good record — or it may be a deceiving
one. Both record and specimen need further scrutiny.
It is, perhaps, important to note that two specimens of
a wingless Psocid (Atropidae) were sent with the lot
labeled ** parasites from Mus rattus." It would be very
interesting if we could know that these Atropids were
really living on the rats, feeding on their hair or dermal
scales. I have found Atropids in rats' nests and birds'
nests living undoubtedly on the loose hairs, feathers and
dermal exuviae. It is my belief, based primarily on cer-
tain striking facts of morphology, that the Mallophaga
are degenerate descendants of the Rsocidae.^ Of the
murid Anoplura, two or three are common to several
hosts, as the well-known Polyplax spinulosa, recorded
from all over the world from the now cosmopolitan Mus
rattus and Mus decumanus, as well as from Mus syl-
vaticus of Europe and north Asia, and Mus alexandrinus
of south Europe and Asia Minor (perhaps only a variety
of Mus rattus) y and Polyplax affinis (perhaps only a
variety of P. spinulosa) recorded from Mus agrarius of
eastern Europe, and Mus sylvaticus of Europe and north
Asia. Polyplax (Hoplopleura) acanthopus, the common
sucking louse of the mouse has been taken from the now
cosmopolitan Mus musculus, and also from Lemmus tor-
quatus, the lenuning of Arctic Europe, Asia and America,
Microtus agrestis, the field vole of Europe, Microtus arvor
lis, another common vole of Europe and Asia, and Micro-
tus sp. from Iowa, U. S. A. The water rat, Hydromys
chry so g aster, of Australia, has a Polyplax species of its
own as has also Otomys bisulcatus of south and central

3 See Psyche, Vol. 9, 339, pp. 1902.

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 273

Africa, Hesperomys leucopus of North America, Epimys
aurifer of the Malay Peninsula, Gerbellus indicus of
northern India and Afghanistan, and Holochilus sciureus
of Brazil and Peru. The conunon Old World mouse, Mus
minutus, harbors three Anopluran species, while Mus
musculus has but two. The Old World water vole. Micro-
tus terrestris, has a parasite differing from the two in-
festing respectively the two Old World land species of
Microtus.

In connection with this resume of the Murid parasites,
I may say that I have now in process of working over
some two hundred vials of material collected last summer
from California mammals, which is going to add many
records to the Murid list of both hosts and parasites. It
will also add numerous records for the squirrels and
spermophiles (Sciuridae).

The family Geomyidae, gophers, is represented in the
host list by three North American and one Central Amer-
ican species. The Mallophagan species Trichodectes
geomydis occurs on all of these hosts. The North Amer-
ican hosts are Qeomys bursarius (Iowa), Thomomys
hott(e (California), Thomomys bulbivorous (California),
and the one Central American host is Macrotomys hetero-
dus (Costa Rica). T. bulbivorous may be a synonym of
T. bottcB. In addition, Qeomys bursarius has yielded an
Anopluran species of genus and species peculiar to it.

The pocket rats, family Heteromyidae, are represented
by a species of Perognathus (Baja California), and
Dipodomys merriami (Arizona). From both are re-
corded the same Mallophagan species, Trichodectes
californicus.

The jerboa, Dipus sp., is the sole representative of the
family DipodidaB. From it is recorded an Anopluran
species taken in Tunis.

The OctodontidaB are represented by three species
parasitized by one Anopluran and three different Mallo-
phaga. The three hosts are of three different genera, one
with an African range, the other two of South America.
The parasite species on each is peculiar to it. A third

Digitized by

Google

274 THE AMEBIC AN NATURALIST [Vol. XLVIH

record, crediting the characteristic Trichodectes pilosus
of the horse to a coypou of South America (in the menag-
erie of the Jardin des Plantes in Paris), is certainly
either a false record or one of rather extraordinary
straggling. The two Mallophagan species from these
South American tuco-tucos belong to the genus Gyropus,
which is the Mallophagan genus especially characteristic
of the related South American families, the Caviidae
(guinea-pigs), the Dasyproctidae (agoutis), and the Chin-
chillidae (chinchillas and vizcachas) (see following
paragraphs).

The guinea-pigs and mocos (family Caviidae) are repre-
sented by three species, and are strongly parasitized by
Mallophaga. They have no Anoplura. The domesticated
form, which is variously held to be a species distinct
from any wild one now known, or a variety of the wild
species, Cavia cutleri, harbors two well-known species of
Gyropus, namely G. ovalis and G. gracilis (this latter is
held by some students to be of distinct genus). In addi-
tion, Piaget has described a species of Menopon (bird-
infesting genus) from it, and Paine and I have described
another Menopon from it from collections we have had
from Peru and Panama. We have also found this latter
species on the wild guinea-pig, Cavia cutleri, from Peru,
and from this host Paine has described a species of
Gyropus peculiar to this host. From the Brazilian moco,
Kerodon moco, has been recorded a variety of Gyropus
gracilis, one of the familiar species of the domestic
guinea pig, as well as another species of Gyropus peculiar
to the moco. Recently Cummings has described a new
Mallophagan taken at Villa Rica, Paraguay, from the
wild guinea-pig, Cavia aperea. For this new species he
established a new genus called Trimenopon. As a matter
of fact the species is so much like Kellogg and Paine *s
Menopon jenningsi, except for its markedly larger size,
that I am not at all sure it should be added as a fourth
guinea-pig parasite.

A single agouti, Dasyprocta aguti, from Brazil, repre-
sents the family Dasyproctidae. From it have been de-
scribed two species of Gyropus peculiar to it.

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 275

The chinchillas and vizcachas (family ChinchiUidfiB,
or LagostomidsB) are represented in the host list by two
species, to which I can add another (perhaps two others)
on the basis of material recently received from Dr. C. H.
T. Townsend, of Peru. From Lagidium peruanum Gay
long ago described a peculiar Gyropns, and I have speci-
mens of a Gyropns which may or may not be different
from Gay's species. His description is very meager. In
addition I am about to describe, under the name PhUan-
dria toivnsendi, another species, representing also a new
genus, specimens of which have been sent me by Dr.
Townsend from the same host. Also in this Townsend
sending are specimens of a small Polyplax species (Ano-
pluran) from the same host.

The CercolabidsB or CoendidaB, American porcupines,
are represented in the host lists by five species, three of
Central and South America and two of North America.
They harbor no Anoplura, but are parasitized by two
Mallophagan species, of which one, Trichodectes setosus,
occurs on all the host species in the list. The second
Mallophagan is a Trichodectes recently described by
Stobbe from Cercolabes nova-hispanice of Mexico and
Central America. The other South American host porcu-
pines are Coendu (Cercolabes) prehensilis (northern
South America) and C. villosus (Brazil). The North
American hosts are Erethizon epixanthum (California)
and E. dorsatum (Nebraska).

Finally the family Leporidae, hares and rabbits, ap-
pears in the host list with six (perhaps only five) repre-
sentatives, of which four, namely, Lepus timidus, of cir-
cumpolar arctic regions, Lepus cuniculus, native to
Europe and north Africa but introduced over the whole
world, Lepus europceus of Europe and Lepus campestris
of western Canada and United States, harbor the same
species of sucking louse, representing a genus peculiar to
hares and rabbits. I must note that this species, Hcema-
topinus ventricosus Denny, is commonly referred to as
two species, of which one, H. ventricosus, is recorded
from the American host species and L. cuniculus, while

Digitized by

Google

276 THE AMERICAN NATURALIST [VouXLVm

the other, called H. lyriocephalus, is recorded from L.
timidus and L. europceus. But Neumann, an exception-
ally experienced student of the Anoplura, holds that the
two species are one. A deer-infesting Mallophagan, Tri-
chodectes tibialis, certainly a straggler, has been recorded
from Lepus europceus, and another Trichodectes (a very
old and uncertain record) from Lepus cannabinus.

The order Insectivora is represented by but two spe-
cies, the mole, Scalops argentatus, of North America,
and the shrew, Sorex araneus of Europe and Asia. Each
harbors an Anopluran species, that of the mole being a
curiously modified form and of species and genus peculiar
to its host, while that of the shrew is of a species not
found on other hosts.

The order Prosimiae, the lemurs, presents a single
record, that of a species of Mallophagan, Trichodectes
mjobergi Stobbe, described from the North Bomean
Nycticebus borneanus (family NycticebidsB).

The order Primates is represented in the host list by
four families, the CebidsB of the New World, the Cerco-
pithecidae, the single family of apes, SimiidsB, of the Old
World, and the family of man, Hominidae. The distribu-
tion of the ectoparasites of these groups is of unusual
interest to the special student and will likely prove equally
so to more general students.

The Cebidae, platyrrhine, tailed. New World monkeys,
are represented by two species, the spider monkey and
one of the howling monkeys of Brazil, members of differ-
ent genera, each with a Trichodectes species peculiar to
it. In addition three species of Ateles, one of Mexico and
Central America, another of Guiana and Brazil, and the
third an undetermined species of the genus represented
by a specimen in a traveling menagerie in Europe, have
yielded three species of the Anopluran genus Pediculus,
otherwise characteristic of man and the anthropoid apes.
These three Pediculus species have been recorded and de-
scribed by three different students of the group, all careful
workers, and there can be no doubt of the generic refer-
ence. But it is to be noted that the specimens of all three

Digitized by

Google

v^v

Ko.569] ECTOPARASITES OF MAMMALS 277

parasite species were obtained either from host skins in
a mnsenm (in one case the Zoological Mnsenm of Ham-
burg, in another, the Berlin Museum) or from a live host
in a menagerie. In no case, therefore, is the possibility of
a straggling record wholly excluded, but the coincidence
of three discoveries makes the records practically safe.
Finally, in this connection it is to be noted (a^ I have
already pointed out in a brief paper*), that, although
^i^\'\ Ateles is a tailed New World genus and presumably

^^^^v widely separated genetically from the anthropoids,

^. ^^ I'riedenthal has aflSrmed, on a basis of blood and hair

comparison, that Ateles shows unmistakable diflferences
'^e/t from other tailed monkeys, and resemblances with the

'^ anthropoids, and he suggests that in Ateles we should see

^^^ monkeys, which, in a certain sense, replace, in the New

TForld, the anthropoids of the Old. It is, in any event, a
t k strange thing that Ateles differs from the other Cebidae

my ^nd from the CercopithecidsB as well, in not harboring the

I'ij ^najylxiTan genus Pedecinus to which all monkey-infest-

rifi!;. %• -Arxoplura, except those of the simians, belong, but in

-,^ ^^^u^lly harboring parasite species of the genus found

,, ^/sex^rlxere only on the simians and man.

Tlx^ family CercopithecidaB, catarrhine. Old World

^ortfer^ys, is represented in the host list by a dozen spe-

^[^ ^^^s, f*:H*om which one Mallophagan species, viz., my Tri-

d ^Ao^^ ^^:^tes colobi from a guereza monkey, Colohus guereza

\\. ^^"■^- <^^::iudatus (East Africa), and ten Anopluran species

'^^^^^ TtDeen recorded. Of the Anoplura nine species be-

; l^^K t o the genus Pedecinus, long recognized as the char-

^ actox-i s^tic genus of the lower monkeys, as contrasted with

-' ^^^ S^^Tius Pediculus characteristic of the anthropoid apes

i^ ^^^ ^^^^t^an. For the tenth species, Fahrenholz establishes

t)^^ x^^^ genus Pthirpedecimis, just as for one of the

^^^^^■"^^ infesting species the separate genus Phthirius had to

^ ^^^ablished. There are several cases of the common-

^^^ of a single Pedecinus species to two or three hosts.

I '^^ ^'^eviceps Piaget is recorded from Macacus silenus of

Ectoparasites of the Monkeys, Apes and Man," Science, N. S., Vol.
^,v . ^^» Pp. 601-602, 1913.

Digitized by

Google

278 THE AMERICAN NATURALIST [Vol. XL VIII

India, Cercopithecus mona of west Africa, and a third
Cercopithecus skin in the Zoological Museum at Ham-
burg. P. longiceps Piaget is recorded from Macacus
cyclopis of Formosa, Semnopithecus maurus var. cris-
tatxis of Borneo, and Macacxis cynomolgus of the Malay-
sian region. P. eurygaster Gtervais has been recorded
from Macacus sinicus of India and on a macaque in the
Zoological Garden at Sydney, and another in the Zoolog-
ical Garden at Melbourne. A hamadryad (Paphio sp.)
of north Africa has a Pedecinus species peculiar to it, as
has a trachypithecus, of Malaysia, and the Barbary ape,
Macacus innuus, of northern Africa and Gibraltar. The
common Macacus rhesus carries one species of Pedecinus
peculiar to it, and that single species of Phthirpedecinus
already referred to. Macacus silenus also has recorded
from it two species both belonging to Pedecinus.

The family Simiidae, anthropoid apes, is represented
in the host list by three species, namely, the chimpanzee
and two gibbons. One of these gibbons is Hylohates
syndactylus of Sumatra; the other is H. leuciscus of
Borneo. A single species of Pediculus is common to
them both, and is not elsewhere recorded. The chimpan-
zee has also a single species of Pediculus which is pecul-
iar to it. No Pedecinus has been taken from a Simian.

Finally man, representing the fourth Primate family,
Hominidffi, is the host of three notorious Anopluran spe-
cies, two of which are species of Pediculus and the third
the only species so far known of another genus, Pthirius.
Neumann is inclined to see in Pediculus corporis only a
variety of Pediculus capitis. All of these parasites are
found on man in all parts of the world. Some curious
variations among the parasite individuals are shown,
perhaps the most curious being a plain tendency to a
darker coloration of the individuals occurring on the
bodies of men of the dark-skinned races. In my brief dis-
cussion elsewhere, already referred to, I have noted the
interesting significance of this possession by man and the
anthropoid apes of a common genus of Anopluran para-
sites, while the parasites of the lower monkeys belong to

Digitized by

Google

No. 569] ECTOPARASITES OF MAMMALS 279

a well-distinguished other genus. There is no doubt that
the close physiological fitting of parasites to host makes
their host distribution significant of genetic or ''blood'*
relationship, and this commonness of one type of parasite
to man and the apes, and its limitation to these hosts, and
replacement on the lower monkeys by another parasitic
type, is an added indication of the actual blood-likeness
of the Simians and man, a likeness apparently greater
than that between the Simians and the lower monkeys.

Ill

In the light of the plain statement in part I of this
paper of my belief gained from a study of the distribu-
tion of the bird-infesting Mallophaga, to the effect that
the host distribution of the permanent wingless ecto-
parasites of birds is determined more by the genetic rela-
tionships of these hosts than by geographic relationships
or any ecological condition, and the corollary of this,
which is that the distribution of the parasites may there-
fore often have a valuable significance as to the genetic
relationships of animals whose genealogic aflSnities are
in process of ascertainment, and in the light of the facts
of distribution for the mammal-infesting Mallophaga and
Anoplura as just set out in part II of this paper, I
hardly need to do more, in conclusion, than to point out
that the distribution conditions exhibited by the mammal
parasites, even in the face of the meager knowledge that
we yet have of the mammal-infesting forms, clearly, on
the whole, confirm this thesis. In fact, considering how
few mammal-infesting parasite species we yet know, it is
surprising how repeatedly the commonness of parasite
species to two or more related, although geographically
well separated, host species, is illustrated. All through
the order from Marsupials to Quadrumana this condition
is again and again exemplified. I am then, naturally,
made more certain of the essential truth of the thesis, and
can the more strongly recommend the attention of sys-
tematic zoologists to that practical application of it,
which I have stated in the form of a corollary.

Digitized by

Google

REGENERATION, VARIATION AND CORRELA-
TION IN THYONE

PROFESSOR JOHN W. SCOTT
University of Wyoming

It is well known that many Echinoderms possess a re-
markable power of regeneration, and the results given
here show some interesting phases of this process in
Thy one hriareus (Leseur). The problem was suggested
a few years ago in connection with class work in the
Marine Biological Laboratory at Woods Hole, Massachu-
setts. There it is a common practise for students who are
taking the invertebrate course to keep aquaria in which are
placed specimens brought in from various collecting trips
in the vicinity. Students are encouraged to study the
behavior of these animals, but their enthusiasm for col-
lecting frequently causes them to overcrowd their aquaria,
with disastrous results. After collecting Thyone, espe-
cially if they are kept in stagnant water, the student is
frequently amazed to find one or more of his specimens
that have undergone evisceration. In this process the
animal not only loses the principal feeding organs, the
tentacles, and the entire digestive system, consisting of
the esophagus, stomach and intestine ; but it also throws
out a whole series of organs surrounding the esophagus
including the circlet of calcareous plates, the nerve ring
forming the central nervous system, the portion of the
water-vascular system known as the ring canal with its
attached stone canal and Polian vesicles, and the muscles
which serve as retractors for the set of organs surround-
ing and attached to the esophagus. We shall refer to
these muscles as retractors of the esophagus.

The remainder of the animal after evisceration con-
sists, principally, of the dermo-muscular integument, the

280

Digitized by

Google

No. 569] REGENERATION 281

cloaca with its attached respiratory trees, the single
gonad, the radial canals of the water- vascular system and
the major portion of the dorsal mesentery by which the
intestine was suspended. Since this part of the animal
continues to give reactions, the student invariably raises
the question, ''Can Thyone regenerate the lost parts t*'
This question was the starting point of the following in-
vestigation. The work had not proceeded far when it
was discovered that important individual differences
occurred, and the question became, ''To what extent, or
how completely, may these individual variations be re-
produced in the process of regeneration f Curiously
enough, the most important differences between individual
Thyone involve structures which help to form the radial
symmetry of the animal. Consequently the problem has
a bearing on the phylogeny as well as the ontogeny of
Thyone.

In general, the results show that regeneration of all lost
organs may occur and that there is a decided tendency to
even reproduce individual variations. It was found that
the Polian vesicles varied greatly in number, size and
location. The retractor muscles in a single radius were
single or multiple, and for each individual this variation
was closely correlated with a corresponding variation in
the number of Polian vesicles. Whether one or more
Polian vesicles are present, there is a strong tendency for
these to occur on the left side of the animal, a fact which
undoubtedly has a phylogenetic significance. A more
complete statement and a discussion of these results will
be given in the following pages.

General Structure of Thyone
Thyone is functionally a bilateral animal. It has ante-
rior and posterior ends, dorsal and ventral surfaces, and
consequently right and left sides. The external opening
of the genital duct is located near the anterior end in the
mid-dorsal region. The structure and arrangement of
the tentacles is alike on both sides of the animal. Even

Digitized by

Google

282 THE AMERICAN NATURALIST [Vol. XL VIII

the feeding reactions, as Pearse has pointed out, indicate
a bilateral type. The single genital gland is median in
position ; the genital duct and the stone canal are in the
median dorsal mesentery ; a part of the intestine and the
stomach are supported by the same structure. The respi-
ratory apparatus is also a bilateral structure, one branch
arising from each side of the cloaca.

Fig. 1. Fio. 2.

Fig. 1. A Diagrammatic Drawing from a Dissection Made by Taking a
Longitudinal Cut in the Body Wall a Little to the Left of the Mid-vbn-
TRAL Line. Shows the arrangement of the chief organs concerned in eylsceration
and subsequent regeneration. B, w., body wall ; cl., cloaca ; c. p,, calcareous
plates ; i., intestine ; {. m., longitudinal muscles ; p., Polian vesicles ; r., ring
canal ; r. m., retractor muscles ; r. t., base of respiratory tree ; $,, stomach ; t.,
tentacles ; m. d,, mid-dorsal ; I. d., left dorsal ; {. v., left ventral ; r. d,, right
dorsal, and r. v., right ventral, interradial spaces.

Fig. 2. A Diagram to Show the Relation of Radial to Bilateral Sym-
metry. The esophagus (e) is shown in cross-section, cut Just anterior to the
stomach, and the view looks toward the anterior end. M., madreporite; r. r.,
ring canal. Other letters as in Fig. 1.

Notwithstanding this general tendency toward bilateral
synametry, the most conspicuous differences between indi-
viduals involve structures of the radial type. Fig. 1 is a
diagrammatic drawing of a dissection to show the general

Digitized by

Google

No. 569] BEGENERATION 283

arrangement of some of the more important structures
studied in this experiment. The dissection was made by
making a longitudinal cut in the body wall a little to the
left of the mid-ventral line, and then pulling the flaps
apart and pinning the animal down on its dorsal surface.
The Polian vesicle is shown attached to the ring canal in
the position where it is usually found when only one is
present, that is in the left dorsal interradial space. It
will be noticed that the retractor muscles are simply
branches of the longitudinal muscles, and hence are radial
in position. At the time of evisceration the body wall
breaks a short distance posterior to the tentacles, the re-
tractor muscles separate at the point where they join the
longitudinal muscles and the intestine breaks off just in
front of the cloaca.

A better understanding of the radial type of structure
will be gained by a reference to Fig. 2. This figure is a
diagram to show the relation of the radial to the bilateral
symmetry. The dorsal side of the animal is represented
toward the top of the page, the esophagus appears in
cross-section, cut just anterior to the stomach, and there-
fore one is looking forward to the other organs shown.
The retractor muscles, showing the position of the radii,
are much contracted and thickened, a condition in which
they are usually found after evisceration. The stone
canal ending in the small madreporite is located in the
mid-dorsal interradial space. Passing around in a clock-
wise direction, the other interradial spaces are designated
as right dorsal, right ventral, left ventral and left dorsal.
Polian vesicles may be found in any of the interradii ex-
cept the mid-dorsal space which always bears the stone
canal. Although only one Polian vesicle is represented in
this figure, the mid-ventral retractor muscle is shown
double, a split condition which is characteristic when two
or more Polian vesicles are present. This description
will be sufficient to show the general relation between the
radial and the bilateral sjnnmetry.

Digitized by

Google

284 THE AMEBIC AN NATURALIST [Vol. XLVIII

EVISCBBATION

Only one method of producing evisceration was used.
By placing a number of Thyone in a small aquarium of
stagnant sea water, the supply of oxygen is soon ex-
hausted. The animals become greatly distended, they
crawl up on the sides of the aquarium when possible, and
extend the siphon toward and frequently above the sur-
face of the water. All of their behavior, including the
pumping of the siphon, indicates that respiration is in-
adequate. In the course of a day or two the water be-
comes very foul ; soon some of the Thyone will eviscerate,
and a considerable percentage will do so as conditions
grow more unfavorable. Many, however, resist the un-
favorable surroundings and will not eviscerate though
kept for several days in foul water. But if the aquarium
is now placed where it will have a continuous stream of
water and air bubbles passing through it, the behavior of
the animals is somewhat different. They then tend to
contract to a minimal size, and sometimes assume a
volume not more than one fifth to one seventh of their
maximum distention. The respiratory movements are
practically discontinued; the animal seeks a position as
close as possible to the side and bottom of the aquarium.
Contraction does not always take place immediately. To
my surprise, after several hours I found Thyone which
had resisted the previous unfavorable conditions now dis-
charging their viscera. After remaining two or three
days in the running water, and the animals had appar-
ently become adjusted to this condition, I again set the
aquarium to one side partly filled with water. Then, by
repeating the conditions of the first experiment, as the
water became foul several more of the holothurians ap-
parently found life too strenuous to further retain their
internal organs. When the remainder of this lot of
Thyone was returned to running water, and again to
stagnant water, a few additional individuals underwent
self-mutilation. Out of a total of sixty-one specimens
used in this lot forty of them eviscerated. That is, autot-

Digitized by

Google

No. 569] REGENERATION 285

omy occurred in at least sixty-five per cent, of Thyone,
under the conditions described. Probably one reason why
this process did not occur in a still larger number is that
some animals occupied more favorable positions in the
aquarium. A discussion of the cause of evisceration will
be given later.

When evisceration occurs it is sometimes hard to see
just how the process takes place. Pearse ( '09) ascribes
the process to a '* structural accident''; that is, it is due
to a powerful contraction of the circular muscles at a time
when the calcareous ring is well forward. **But if the
tentacles are extended," he says, *'and the calcareous
ring is pushed forward a break may occur at &" (a point
in his Fig. 2 where the body wall joins the calcareous
ring) ''as a result of the strong contraction of the circu-
lar muscles at that point, and the visceral organs are
forced out. . . . Whether this autotomy takes place or
not depends upon the breaking of the inner branch of the
longitudinal muscle bands, whose normal function is to
retract the calcareous ring. When the strain brought
about by the contraction of the circular muscles becomes
too great these inner bands are torn asunder, usually at
the point a?'' (inner end of the retractors of the calcareous
ring). While it is true that muscular contraction and
consequent pressure undoubtedly plays a prominent part
in the process, close observation has convinced me that
this is not the only factor causing evisceration. Upon
several occasions I have watched carefully the breaking
of the body wall near its attachment to the calcareous
ring, and while there are times when the pressure appears
to be strong, especially when the animal is being irritated
mechanically, there are other times when the skin appears
to ''melt away" or separate with very little or no pres-
sure present. Indeed, after the skin once breaks at one
side and the viscera escape through the opening, the pres-
sure is relieved. But one may observe that the skin con-
tinues to break until the calcareous ring is entirely sepa-
rated. This, of course, would not happen if the process

Digitized by

Google

286 THE AMERICAN NATURALIST [Vol. XL VILE

depended entirely upon an accidental stmctural defect.
Another thing noticed is of interest in this connection.
When splitting open the body wall of an animal that was
eviscerating, and thus relieving any internal pressure that
might be due to contraction of the circular muscles, some
of the retractors were seen still attached to the longitu-
dinal muscles. Under these conditions it would not be
possible for the retractors to exert any pull against the
pressure produced by the circular muscles, yet the re-
tractors were observed to constrict off or break away
from the longitudinal muscles by what appeared to be
purely a local disturbance. It is hard to see how this
could happen, or how the skin continues to separate
around the calcareous ring after the first break is made,
if the process of evisceration depends solely upon the
breaking of retractors and internal pressure. Indeed,
the view that local changes take place in the tissues is
supported by other facts. Leptosynapta, if left in stag-
nant water or under other favorable conditions, under-
goes repeated autotomous fission as the result of local
constrictions, and Pearse states that autotomy depends
upon the presence of the anterior portion of the body,
and presumably upon the presence of the cireumoral
nerve ring. However, he found in Thyone that highly
irritating substances like acetic acid and clove oil did not
produce ejection of the viscera.

Nor were drugs like codene and atropine, which cause violent peri-
staltic waves of contraction to pass over the body, any more potent in in-
ducing autotomy. The same may be said of sodium chloride, atropine
and clove oil, although the injection of any of these substances was
often followed by a waving of the oral tentacles to perform feeding
movements, thus bringing about favorable anatomical relations for au-
totomy.

These results would indicate that the nervous system
is not primarily involved. Certainly the ejection of vis-
cera may occur in Thyone without any visible external
stimulus.

The parts eviscerated in Thyone have already been

Digitized by

Google

No. 569] REGENERATION 287

mentioned. However, sometimes evisceration is incom-
plete, as the following examples will show. On the morn-
ing of August 4, a Thyone, which we shall later speak of
as individual H, was found eviscerating in an over-
crowded aquarium jar. While the process usually re-
quires only a few seconds, or at most a few minutes,
the intestine in this case was not completely thrown out
until two or three hours later. This animal lived until
killed at the end of twenty-one days. In the afternoon
of the same day on which individual H eviscerated, an-
other Thyone was found with the process only partially
complete. Five hours later the intestine was still re-
tained, and scissors were used to cut it off at its anterior
end near the stomach. Though this Thyone received
equally good care it died at the end of two days without
further evisceration. A third specimen was found in-
completely eviscerated on the above date, but it was
allowed to stand until the next morning ; at this time the
injured end was open, the intestine was still within the
body cavity and a part of one of the branchial trees was
protruding. The intestine was pulled out and broken
oflF, after which the branchial tree was retracted and the
injured end partially closed. This animal also died at
the end of two days. A fourth Thyone was seized and
by squeezing was forcibly caused to throw off the usual
parts except the following: a part of the stomach, most
of the intestines, and some of the retractor muscles
which had broken off near their esophageal end. The
next morning it had expelled the remainder of the
stomach and intestine, two complete retractor muscles,
and some debris which had escaped from the intestine
into the body cavity. The anterior end of the part re-
maining appeared ragged and imperfectly closed. It
died on the third day. It is probable that the two re-
tractor muscles last expelled were broken off at their
posterior ends by local constriction, not when the body
was under pressure. A fifth animal, which we shall
designate as individual M, was found partly eviscerated

Digitized by

Google

288 THE AMEBIC AN NATURALIST [V0L.XLVIII

late on the afternoon of August 6. The next day it still
retained the stomach and intestine and at noon the diges-
tive tube was clipped off with scissors in the region of
the esophagus. Nothing peculiar was noted in its behavior
until four days later, August 11, when it discharged the
remainder of the digestive tube. It lived and was killed
at the end of eighteen days. These results are typical.
The animal dies unless it is itself able to eliminate all
organs concerned in the process of evisceration, and
therefore regeneration does not occur unless all these
organs are eliminated.

The eviscerated animals show comparatively a low
degree of mortality. In an attempt to raise twenty-five
mutilated Thyone seven died; three of these were un-
able to complete the process of evisceration as described
above, and two more, since they lived for fourteen days,
probably owe their death to other causes. The sixth
specimen to die lived three days and had been slow in
eviscerating. The seventh did not receive the best of
care and died after three days. So considering the
amount of injury the mortality is extremely small where
proper care is taken and -evisceration is complete.

It will not be inopportune to describe the subsequent
behavior of the different parts after evisceration. The
parts expelled lie on the bottom in a more or less inactive
condition untit they die, which happens usually in the
course of a few hours. At first the tentacles frequently
expand and contract. They are highly sensitive, as one
would expect, and if touched withdraw quickly into the
esophagus and at the same time the retractor muscles
will undergo strong contraction. By supporting these
parts near the surface of the water, so as to insure plenty
of oxygen, an attempt was made to keep them alive. In
some cases the parts remained alive for two or three
days, so this experiment appeared to be partially success-
ful. Death is probably due to the direct exposure of
tissues to the sea water and to the attacks of minute
organisms. The dermo-muscular portion of Thyone is

Digitized by

Google

No. 569] REGENERATION 289

much less sensitive than the expelled portion, just after
.evisceration. This is due to lack of a central nervous
system.

Behavior During Regeneration
After evisceration each specimen was placed in a sepa-
rate jar of fresh sea water. The injured end of the body
turns in and closes up tightly, and the entire body is
somewhat smaller than before evisceration. Respira-
tion is slower and not so vigorous. If the water is stag-
nant, within a few hours the animal usually climbs up
on the side of the aquarium by means of its tube feet.
This part of the animal therefore is capable of respond-
ing to a lack of oxygen, and the reaction is independent
of the central nervous system.

The observations upon the following individual, re-
ferred to in my notes as Thy one A, will serve to illus-
trate the general behavior during regeneration:

July 14, A.M. — ^Animal eviscerated itself in the usual way. In the
afternoon it climbed up on the side of the jar and clung there evidently
for the purpose of respiration.

July 15-16. — Acts as on the afternoon of the fourteenth. Keeps
closed and well contracted at the injured end. Entire body somewhat
smaller than before evisceration, due in part to organs lost. Respira-
tion slower and not so vigorous as normal.

July 17. — ^In the afternoon, after water was changed, Thyone took
up position on the sand against the side of the jar farthest away from
the source of light.

July 18. — The next morning it was half buried in the sand in same
position, with a few pieces of debris pulled over it. Remained so all
day.

July 23. — For some two days it has been slowly burrowing down until
only the two protruding ends of the body can be seen. When a piece
of debris that was being held over a part of the anterior end was
touched, this end retracted below the surface and the posterior end
withdrew until it could scarcely be seen. Later the posterior end re-
tracted when the shadow of my hand passed over it, the hand being held
about one foot away. The uninjured animal is even more sensitive to
shadow. The respiratory movements are growing stronger.

July 28. — For the past two or three days the Thyone has been slowly
moving through the sand in a posterior direction without uncovering
itself.

Digitized by

Google

290 THE AMEBIC AN NATURALIST [VOL.XLVIU

August 2. — It is now oriented with respect to the direction of the
light and has reached probably the darkest portion of the jar.

August 7. — Has advanced still farther. Came about half way out of
the sand to do this.

August 8. — ^Reacts quickly to shadows by withdrawing, and to jar-
ring the table. Evidently is recovering its normal behavior.

August 10. — Has again come up about half way out of the sand.
Reacts quickly to shadows as before.

August 11. — Came entirely out of the sand. Spent the day on the
sand or on the side of the jar. Appeared restless.

August 12, 4 P.M. — Has been clinging to the side of the jar and mov-
ing about more or less all day. Respiratory movements are strong and
apparently normal. Has just now expanded the anterior end suffi-
ciently for me to see the new growth of tissue formed around a penta-
gonal opening. Fifteen minutes later it was observed to extend a set of
minute tentacles and go through feeding movements. The tentacles ap-
peared to be slightly more than three eighths of an inch in length. Its
behavior continued apparently normal until it was killed twelve days
later.

The actions of other Thyone were studied under the
same conditions, and we shall now give a general sum-
mary of their behavior during regeneration. The earli-
est reactions after evisceration take the form of contrac-
tions resulting in the closure of the wound, and move-
ments in response to lack of oxygen. If the oxygen
supply is suflScient Thyone will draw itself closely into
the angle between the side and bottom of the aquarium,
or if the supply is deficient, it clings close to the side of
the jar near the surface. In from three to seven days
an instinct to burrow usually asserts itself. There is a
tendency for the body to contract very noticeably at this
time, and the whole organism becomes rather inactive.
This condition is probably necessary for the formation
of new tissue. Pearse makes the statement that in bur-
rowing the normal Thyone will cover itself in from two
to four hours. My observations on the mutilated ani-
mals indicate that they require from twelve to twenty-
four hours, in one case forty-eight hours, to complete the
reaction. The process frequently stops for some hours
and occasionally is never completed. In the Thyone de-

Digitized by

Google

No. 569] HEGENERATION 291

scribed above the animal did not begin to orient itself
with respect to the source of light until about the twelfth
day, but in another case the response took place on the
second day, which shows that this reaction does not de-
pend upon the central nervous system. It should be
stated that normal Thyone similarly placed were used as
controls. Thyone A was quite sensitive to shadows and
to touch on the ninth day, but it reacted more quickly on
the twenty-fourth day both to shadows and to mechanical
disturbances. Whether this was due to the regeneration
of a new central nervous system, or to a more highly
developed specialization of function in the old tissue, I
am unable to say. It is quite possible that both factors
were involved. Respiration is undoubtedly correlated
with the activity of the animal, and feeding movements
do not occur until the regeneration of all organs is well
established, at about twenty-seven or twenty eight days.
The internal changes that take place during regenera-
tion were studied in animals that were killed at different
stages in the process. Thyone N was killed nine days
after self mutilation. At the injured end there was a
very small plug of tissue representing the newly formed
esophagus; a thread-like continuation of this tissue, the
beginning of a new stomach-intestine, was also seen in
the mesentery. The calcareous ring and the ring canal
were not clearly defined. Another Thyone was killed at
about the same age after evisceration ; India ink was in-
jected into the cloaca and into the opening at the ante-
rior end in an attempt to demonstrate a cavity in the
newly formed thread-like, stomach-intestine. The re-
sults were negative and the esophagus was found to be
tightly closed. However, the interesting observation wa^
made that the anterior end of each of the longitudinal
muscles had split off a very slender branch to form a
new retractor muscle (see Fig. 3). These newly formed
retractor muscles were not more than one fourth inch in
length; their anterior ends were attached in a normal
position around the esophagus, but their posterior ends

Digitized by

Google

292

THE AMEBIC AN NATURALIST [Vol. XLVIII

were attached only a short way back, much in front of
the position of attachment of the full-sized retractors.
In another animal killed when a day or so older, the same
conditions held with reference to esophagus, stomach
and intestine. At least three of the radial canals belong-
ing to the water vascular system had branched and con-
nected at their anterior ends in such a manner as to
form a part of a new ring canal (cf. Fig. 4). I was un-
able to find the rest of the ring-canal and perhaps it was
not yet complete.

X /-v

FiQ. 3. Fig. 4.

FiQ. 3. Diagrammatic Drawing to Show that in Regeneration the Rr-
TRACTOR Muscles (r. w.) Arise by Splitting off from the Longitudinal
Muscles {U m.). Dissected a little to the right of the mld-yentral line; d., dor-
sal mesentery suspending the intestine {i. ) ; in,, integument ; e., region of
esophagus.

Fig. 4. To Show the Development of the Pentagonal Canal in a Thyone
about Nine or Ten Days after Evisceration, r., radial canal ; p., pentagonal
canal. The anterior ends of the radial canals fork dichotomously, and these
branches anastomose to form the canal which later assumes a circular shape
around the esophagus.

Thyone F, which was killed twelve days after eviscera-
tion, showed minute calcareous plates which formed a
very small esophageal ring not more than one millimeter
in diameter. The esophagus continued posteriorly in the
form of a small tube, the stomach-intestine, which was
suspended in the dorsal mesentery. This new digestive
tube was about 0.5 millimeter in diameter and contained
small, colored, movable particles that could be seen with
the unaided eye. The ring canal was completely formed.

Another specimen, Thyone 0, died at the end of four-

Digitized by

Google

No. 569] REGENERATION 293

teen days and was in bad condition when examined. The
stomach had begun to expand and retractor muscles were
present. Probably owing to the condition of the speci-
men, no calcareous ring, ring canal; or Polian vesicle
could be found. Another individual killed at about fif-
teen days showed the stomach slightly enlarged, and the
intestine, retractor muscles, calcareous ring, tentacular
canals, and ring canal well formed. Two small Polian
vescicles each about one millimeter in length were pres-
ent. The position of the new intestine was described in
my notes as follows :

From the stomach the intestine follows the ventral edge of the dorsal
mesentery, lying ventral to th'e gonaduct. At the gonad it turned ven-
trally with the mesentery and then forward for about one half inch to
the left interradial space; here it turns rather abruptly backward, con-
tinuing in the mesentery below the left branchial tree to the anterior
ventral part of the cloaca.

At a little later stage in another specimen the intestine
passed from the left ventral interradial to the right ven-
tral interradial space ; then posteriorly and again to the
left, following the ventral radial mesentery to the ante-
rior ventral side of the cloaca.

We see from the preceding description that all impor-
tant organs have been reproduced in form though not in
size, before the end of the fifteenth day. The first madre-
porite with its tiny stone canal was found some eighteen
days after mutilation. Twenty-one days after eviscera-
tion in one specimen the calcareous ring was about three
millimeters in diameter and the ampuUae at the bases of
the tentacles were well developed. Within a week after
this time the regenerating animal begins active feeding.
Thy one A, killed at 41 days, was practically a normal
animal both in behavior and appearance, except for the
fact that the regenerated organs had not yet reached full
size. The stomach was about one third normal size, but
the Polian vescicles were better developed. The intes-
tine contained a small amount of food material and was
nine or ten inches in length; most of this growth had

Digitized by

Google

294

THE AMEBIC AN NATURALIST [Vol. XL VIII

taken place posterior to the gonad. It was held in posi-
tion as previously described and had several additional
coils.

Individual Variations
To all outward appearances any two Thyone are as
much alike as two peas. It was not until the internal
organs were studied that important differences were ob-

FlO. 5. DiAQBAHS TO SHOW VABIATION IN POSITION AND SiZB OP THE POLIAN

Vesicles. P. v., Polian vesicles ; m., madreporite ; r. c, ring canal ; a-d, witb
one Polian yesicle, e-g, with two ; h-k, with three, I., with four ; c, d., f., g., k.,
with additional rudiments of these vesicles; j., with a branched vesicle.

served. While there are numerous minor differences,
the most conspicuous variations are found in the num-
ber, size and location of the Polian vesicles (cf. Fig. 5),
and in the number and arrangement of the retractor
muscles. On account of the radial structure of Thyone
not more than four Polian vesicles are present, since

Digitized by

Google

No. 569]

REGENERATION

295

a homologous structure, the madreporite and its stone
canal, occupies the dorsal interradial space. The num-
ber of vesicles varies in fact from one to four. By a
reference to Table I, it will be seen that out of 77 indi-
vidually examined, 41 had one, 20 had two, 14 had three,

TABLE I

To Show the Number of Polian Vesicles Peesent in a Given Number

OP Thyone. Also to Show thbib Location in the

interradlal spaces, with reference to the

Bilateral Symmetry of the Animal

Number of
PoliAD YesiolM

Number of IndiTld-
uala Eznmiued

Left
Dorsal

Left
Ventnil

Right
Ventntl

Right
Dorsal

1
2
3
4

41
20
14
2
77

38
17
14
2
71

3
19
15

2
39

. 0

1
1
2

Totals

and 2 had four Polian bodies. If one is to test the matter
of regeneration, of course it is important to know
whether the variations or individual peculiarities will
be accurately reproduced. Another striking character-
istic comes out when we note in the same table the loca-
tion of these organs. Of the forty-one individuals which
had a single Polian vesicle, all were on the left side of
the animal, and 38 were in the left dorsal interradial
space. In twenty specimens with two Polian bodies each,
36 were on the left side and only four on the right side
of the body. A similar asymmetrical distribution of
these parts was found when three Polian bodies were
present. In one specimen, however, two vescicles were
found in one space, the left ventral interradius, the only
instance of this kind observed; on account of this dou-
bling, the right side lacked one of the number to which it
was entitled in the table. Where four Polian bodies are
present the arrangement is, of course, symmetrical on
both sides. Still another interesting fact comes out when
we examine the totals in the last line. Out of the 77 indi-
viduals, 71 had a Polian vesicle in the left dorsal inter-
radial space, 39 vesicles were found in the left ventral.

Digitized by

Google

296

THE AMERICAN NATURALIST [Vol. XL VIII

17 in the right ventral, and only 4 in the right dorsal
space. That is, the total number on the left side com-
pared with the total number on the right side bears the
ratio of 110 to 21. Not only is there this tendency for
the vesicles to be more abundant on the left side of
Thyone, but the totals show that the chances of a given
Thyone having a Polian vesicle in any given interradial
space decreases in a counter-clockwise direction, begin-
ning with the left dorsal interradial position. Coincid-
ing with the number of individuals examined, the maxi-
mum number of chances is found in the mid-dorsal inter-
radius, where the stone canal is always present. That is,
the stone canal with its madreporite is a more funda-
mental and stable structure than each or all of the
vesicles.

The conditions are none the less interesting when we
compare the Polian vesicles with reference to size and
location, as will be seen from the examination of Table
II. The Polian vesicles are here divided arbitrarily
into three groups, designated as large, medium and
small, and their respective locations are shown. In addi-

TABLE II
To Show the Polian Vesicles with Reference to Size and Location

Sice

Left Donal

Left Ventral

Right Ventral

Right Donal

ToUl

Large

Medium

Small

Rudiment . . .

0
1
3

73
45
13
15

Total

146

tion some Thyone had the rudiments of other vescicles,
each too small to be considered a distinct pouch. These
are designated in the table as a *^ rudiment." It will be
noticed that all of the large, and most of the medium-
sized vesicles are on the left side; that all the small
ones, and most of the rudimentary ones are on the right
side. The table as a whole shows that not only does the
number of Polian vesicles diminish in a counter-clockwise

Digitized by

Google

No. 569] REGENERATION 297

direction, but their size diminishes following the same
law. These facts appear significant and without doubt
are suggestive of ancestral history.

If it is true that the radial symmetry of Echinoderma
is to be ascribed to a fixed stage in their ancestral his-
tory, we are led to suppose that the point of attachment
was on the right side of an originally bilateral animal.
The life history of Pentacrinus, the larval organ of Aste-
roidea, and a great many anatomical and embryological
facts support this view. While it is not within the prov-
ince of this paper to discuss the relative significance of
these matters, the evidence is so overwhelming that the
theory is generally accepted. It is also no doubt true
that some groups of Echinoderms took to a free-living
existence early in their ancestral history, and others re-
mained fixed until comparatively a late period. As proof
we may cite the embryological evidence that Holothurians
develop without any attached stage whatever, that the
Asteroids develop a larval organ and pass through a
Sessile stage for a brief period in their development,
while the crinoids usually remain permanently fixed
throughout life. At least we can best account on this
theory for the deep-seated and fundamental radial sym-
metry of some forms; the longer the attachment the
more deep-seated would become the type of radial sym-
metry. Now if this theory is correct we can use it to ex-
plain the conditions described above for Thyone. The
ancestors of this form must have broken away from the
fixed stage very early, for we find the radial symmetry
not well established on the right side of the animal as
evidenced by both the position and size of the Polian
vesicles. Out of 118 large and medium-sized Polian
vesicles, 112 were on the left side, while in a total of 28
small or rudimentary Polian bodies, 25 were found on
the right side. The arrangement of these organs in
Thyone adds one more bit of evidence to support the
following statement of Lankester.

Digitized by

Google

298

THE AMERICAN NATURALIST [VoL.XLVin

It therefore appears that the Holothurian stock branched off from
the Pehuatozoa before complete pentamerous symmetry of the hydro-
coele and associated organs had arisen, before any definite calcynal sys-
tem had developed, while the gonads were still a simple strand opening
to the exterior by a single posterior gonopore.

The muscles used as retractors of the oesophagus were
other organs in which there was considerable individual
variation. As a general rule each of the five retractor
muscles consists of a single band that takes its origin from
the longitudinal radial muscle about one third the way back
from the anterior end of the body and is inserted in front
into the wall of the esophageal ring. Such a retractor,
however, is frequently split up into several strands vary-
ing from two to five in number. A reference to Table III

TABLE III

To Show the Correlation between the Number op Polian Vesicles

AND THE Tendency poe the Retractoe Muscles to Divide

Number of Poli«n Vesicles

Retractor muscles, single

Retractor muscles, multiple

Average number retractor muscles, per individual

Average number retractor muscles, per radius

39
1

5.153
1.030

2
17
10.263

2.052

0 0

16 I 2
12.400 10-000

2.480| 2.000

shows that in 76 individuals examined, 41 had retractor
muscles all in single bands, while 35 specimens had these
muscles subdivided or multiple in character. This vari-
ation is especially interesting when considered with
reference to the number of Polian vesicles. For in forty
cases where one Polian body was present thirty-nine bore
the unsplit or single retractor and there was only one
specimen with these muscles showing a multiple number.
In thirty-six cases where two or more Polian vesicles
were present, all but two had the retractor muscles in a
split or divided condition. If we consider each strand
as a separate retractor muscle, we may then obtain the
average number of retractors per individual for any
definite number of Polian vesicles. By a reference to
the fourth horizontal line of Table III, one finds that the
average number in individuals with one Polian vesicle is

Digitized by

Google

No. 569] REGENERATION 299

just slightly in excess of five, the pentameric number,
and the average number when two Polian vesicles are
present is 10.263. This ratio is only partly maintained
when three vesicles are present, for the average number
is then 12.400, and in the two cases with four vesicles the
average was just twice the pentameric number. It is
therefore evident from the facts shown in this table that
with an increase in the number of Polian vesicles there
is associated a strong tendency for the retractor muscles
to take on a split character. If it were not for the fact
that the split character shows considerable variation in
the same individual one might suggest that the tendency
to divide is correlated with the greater functional activ-
ity of the water vascular system as evidenced by the in-
creased number of Polian vesicles and the location of the
longitudinal muscles that lie along and just internal to
the radial canals. About all one can say is that corre-
lated with a more complete radial symmetry with respect
to the Polian vesicles, there is a greater plasticity in the
retractor muscles, causing them to divide longitudinally
into separate muscle bands.

To what extent, or how completely, may these indi-
vidual variations be reproduced in the process of regenera-
tion! An answer was obtained in the following way.
First a close examination was made of all parts eviscer-
ated and a record was kept of all organs showing variable
structures. Special attention was given to Polian vesicles
and to retracter muscles. The mutilated specimens were
then placed in separate aquaria in which the water was
changed frequently to prevent it from becoming stale.
After a considerable interval these animals were killed
and the regenerated organs were compared with the lost
parts. Table IV shows several individuals compared in
this way. The number of retractor muscles found in each
radius is given in the order of the radii taken in a clock-
wise direction. A study of the table indicates that there
is a strong tetndency to reproduce individual peculiarities,
as shown by individuals B, E, G, H, M and 0. This does

Digitized by

Google

300

THE AMERICAN NATURALIST [VoL.XLVHI

not always hold true, for individual L reverted toward
the more radial type of symmetry. From these few cases
it would appear that individual peculiarities tend to pre-
dominate over ancestral influences in the process of re-

TABLE IV

To Illustrate the Relation between Beoeneration and Original Sym-
metry IN Thyone

IndiTiduAl

OrlglDAl Sjmmetrj

Regenerated Symmetry

Used

PoliAD Vetfcles

Retractor Muadea

Pollan Vesicles

Retractor Musclfs

B
E
0
H
L
M
0
W

2 +

2-2-2-2-2
3-3-2-2-2
1-1-1-1-1
1-2-2-2-1
1-1-1-1-1
1-1-1-1-1
2-2-2-2-2

2 +

3-3-2-3-3
2-3-2-2-2
2-1-1-1-1
2-2-2-2-2
2-3-2-2-2
1-2-1-1-2
2-2-1-2-2
2-2-2-3-4

2-2-2-2-2

2-2-1-2-2

generation. Specimens W, X, Y, are included in this
table to show further the correlation between Polian ves-
icles and retractor muscles.

Discussion and Summaby
There remains to be discussed the general bearing of
the foregoing experiments. First, the difference in the
number of Polian vesicles in different Thyone is partly
compensated by a variation in size, the fewer the number
the larger their size, though this ratio would, not be an
exact one. In other words the total volume of the Polian
vesicles in any given specimen bears a general relation to
the size and functional activity of the animal. Notwith-
standing this functional relationship since the actual
number varies so widely it would be interesting to com-
pare the number found in other species of holothuria with
the conditions in Thyone. The data secured on this ques-
tion were meager and not very definite. For example,
Packard in one of the older text-books says in speaking
of Thyone,

Digitized by

Google

No. 569] REGENERATION 301

There are three Polian vesicles, one fusiform and an inch in length,
the two others slenderer.

Clark ('02) gives the number for Thyone briareus
(Leseur) as usually one or two; for T. scahra (Verrill)
as usually single, and for T. unisemita (Stimpson) as
one. He also mentions six other holothurians found in
the Woods Hole region and all have a single Polian ves-
icle except Cucumaria frondosa (Gunnerus), which usu-
ally has one. He says nothing of the position in which
these vesicles are found. In another paper ( '01) Clark
mentions a large holothurian about 40-45 centimeters in
length {Holothuria mexicana Ludwig) in which there is
a great diversity in the number of tentacles and Polian
vesicles. The tentacles vary from 18 to 21, while the
Polian vesicles vary from 1 to 9. The number of speci-
mens examined, sixteen, was hardly suflBcient to obtain an
adequate comparison ; two had 1 vesicle each, two had 2,
five had 3, three had 7, one had 8, and one had 9. It is
probable that if one were to examine a large number of
individuals of each species, with reference to the number
and location of the vesicles, he would obtain further inter-
esting results. Lang ( '96) cites a number of groups of
holothurians in which only one vesicle has been observed ;
but states that there are a number of species in other
groups that have occasionally or usually more than one.

Where accessory vesicles occur they vary greatly in number, and ap-
pear to have very slight, if any, systematic significance. Where only one
Polian vesicle occurs it lies in the left ventral inten-adius, very seldom
in the left dorsal interradius. Where two or more vesicles occur, they
are also mostly formed in the ventral region of the circular canal.

Since Lang describes Cucumaria as the type specimen,
in which the Polian vesicle is said to be in the left ventral
region, it is possible that his generalizations were based
principally on this form. At any rate, the conditions in
Thyone seem to give a more definite significance to the
number and location of thB Polian vesicles.

Various explanations of autotomy and evisceration have
ieen suggested, many of them having a teleological char-
acter. The view that the holothurian offers up the better

Digitized by

Google

302 THE AMERICAN NATURALIST [VoL.XLVm

part of itself to appease the hunger of its enemy lacks
confirmation, since the viscera are distasteful to fishes
and to some other animals. It may be that the autotom-
ous elimination of the Cuvierian organs serves a defen-
sive purpose, as pointed out by Ludwig and Minchin, and
Minchin suggests that the viscera may also be lost in this
process and thus incidentally be associated with a pro-
tective response. In the case of Thyone, however, evis-
ceration can hardly be considered defensive, and certainly
it is not a process of self -division for only one part pro-
duces a new individual. Clark ( '99) in discussing self-
mutilation in the synaptas states the matter clearly in
the following terms :

I agree entirely with Cuenot ('91) in believing that autotomy is not
normal or defensive but is due entirely to pathological conditions. I
never saw a case of it in synaptas supplied with plenty of sand and an
abundance of sea water.

Lang ( '96) points out one of these pathological condi-
tions, and recounts the fact that

A Stichopus was observed to come entirely out of its skin, t. e.j the
whole integument dissolved into slime, so that only the dermo-muscular
tube enclosing the viscera remained.

In the present paper I have mentioned that Thyone at
times appears to undergo a similar softening of the
tissues in the region where the break occurs, and Pearse
('09) showed that autotomy is due, at least in part, to a
structural arrangement which he considers is accidental
in character. My observations further show that local
constrictions undoubtedly have an important part in sepa-
rating the retractors from the radial longitudinal muscles.
All of these factors are pathological and are due to exter-
nal or internal stimuli. The external (extra-cellular)
stimuli, mechanical and chemical, as tried by Pearse
('08), appear to be less effective in producing autotomy
than the purely internal (intracellular) stimuli such as
lack of oxygen and its associated phenomena. The chem-
ical (strychnine) that produced the largest percentage
of evisceration in Pearse 's experiments, probably affected
respiration, since it greatly increased the activity of the

Digitized by

Google

No. 569] REGENERATION 303

animal; therefore the need of oxygen would be propor-
tionately greater than the supply, and the Thyone ren-
dered more susceptible to evisceration. Now while autot-
omy undoubtedly enables the animal to maintain its exist-
ence for a considerable period on a smaller supply of
oxygen, the times when this would become necessary in
nature are probably rare, and it would be futile to specu-
late upon what evolution yet has in store for the process.

According to Lang, the retractor muscles of the oral
region have been derived by the splitting up of the ori-
ginally simple longitudinal muscles, and this specializa-
tion became more marked as the oral tentacles became
more highly developed and required increasing protec-
tion. Species are to be found in the Dendrochirotae in
which the separation and branching off of retractors from
the longitudinal muscles has not yet been perfected. In
regeneration the retractor muscles of Thyone are derived
in the same way, i. e., by splitting off from the longitu-
dinal muscles, and such progress is made that they are
fairly well developed by the time the tentacles take up the
function of feeding. The increasing sensitiveness and the
later activity of the regenerating animal are presumably
associated with the development of a new nervous system.

If we may regard the bilateral echinoderm larva as
representing an early phylogenetic stage rather than a
larval adaptation to a free-swimming existence, we will
now discuss the symmetry of Thyone. As stated above, it
is generally agreed that the radial arrangement of parts
of the echinoderm body is due to a fixed stage in its
ancestral history. Some holothurians and spantangoids,
show in their ontogeny first a free stage, second a radial
stage, and finally a bilateral adult. During the develop-
ment of asteroids that have a fixed embryonic stage, the
early bilateral symmetry is soon disarranged by the
development of organs on the left side of the animal.
For example, the left hydrocoele takes the form of an un-
closed water-vascular rosette which grows around the
esophagus to form the ring canal and its appendages, and
its connection with the dorsal pore gives rise to the stone

Digitized by

Google

304 THE AMEBIC AN NATURALIST [Vol. XL VIII

canal. Excepting the echinoids and erinoids in which
there is either no distinct Polian vesicle or else a simple
glandular structure, those echinoderms that have retained
the most distinctive type of radial structures have also
as a rule, retained the most symmetrical arrangement of
the Polian vesicles. Presumably these forms, the aster-
oids and ophiuroids, have quite recently abandoned the
fixed stage, and each individual usually has four Polian
vesicles and a stone canal, one in each interradius.
Among most of the holothurians a secondary bilateral
symmetry has become superimposed over the radial type,
and it is reasonable to suppose that there was a time in
the ancestral history of Thyone when the Polian vesicles
were symmetrically and radially disposed, or else the
animal quit its fixed habits before the radial symmetry
of the vesicles was thoroughly established. In the one
case we would have a regression, a sort of backward
retracing of the steps of evolution, or, which seems more
probable, the ancestors of Thyone began a free-living
existence before the radial arrangement of the Polian
vesicles had become complete. Also the fact that the
embryology of the holothurian egg is probably much
compressed and shows no trace of a fixed stage indicates
that the corresponding ancestral stage was compara-
tively short, or, very remote. Since the modem habits of
Thyone are bilateral, and since it is altogether improb-
able that such habits would produce the present arrange-
ment of Polian vesicles, the position of these organs must
be due to ancestral influence.

Now the Polian vesicles are capable of contracting and
expanding and their function when they are well devel-
oped is to act as accessory reservoirs of the water-vas-
cular fluid. Muscle and connective tissue in the wall of
the vesicle furnish the means to do this work. Of course,
if the ampullae are well developed there is little or no
need of Polian vesicles, as is the case in Asterias. But,
though the size and number of these vesicles is function-
ally correlated with the general development of the
water-vascular system, especially of the oral tentacles,

Digitized by

Google

No. 569] REGENERATION 305

and hence shows great variability in the different species
of holothurians, this does not in any way explain the
great excess of these vesicles on the left side of Thyone
briareus. In regeneration, probably through the influence
of functional correlation, there is a tendency for the old
tissue to reproduce the exact number and arrangement
of the lost vesicles, but it may reproduce a somewhat
more radial (ancestral) arrangement.

Enough has been given in this paper to show the need
of a more extensive and intensive reexamination of the
Polian vesicles. This would give a better idea of their
morphological and functional significance. The follow-
ing sunmaary and conclusions are based on the work
described :

1. Evisceration in Thyone includes the following or-
gans: Esophagus, stomach, intestine, calcareous ring,
nerve ring, tentacles, ring canal, Polian vesicles, stone
canal with madreporite, and the retractor muscles of the
esophagus.

2. The method used to produce evisceration was to
allow Thyone to stand in stagnant water imtil it became
foul. This was followed by treatment with rimning water
containing much oxygen. Alternating these processes
produced as high as 65 per cent, of self -mutilated indi-
viduals.

3. The structural accident theory of Pearse is inade-
quate to explain all of the conditions arising in the proc-
ess of autotomy. At times the skin appears to dissolve
away with little or no pressure present, and retractors
frequently break off by local constrictions instead of by
longitudinal pull.

4. The parts eviscerated are at first highly irritable,
and may be kept alive for some time. The part remain-
ing is less responsive, but reacts to touch, to lack of
oxygen, and probably to other stimuli.

5. Regeneration of all lost organs may occur, but it
takes place only when all parts concerned in evisceration
are completely expelled. Otherwise the animal dies.

Digitized by

Google

306 THE AMEBIC AN NATURALIST [VouXLVm

6. During the process of regeneration the behavior
gradually becomes more responsive and finally is like the
normal individual. This appears to be correlated with
the growth of a new nervous system.

7. Thyone is functionally a bilateral animal, but the
most conspicuous individual differences involve struc-
tures that have a radial arrangement.

8. The Polian vesicles vary greatly in number, size
and location. There is a strong tendency for these to
occur on the left side, and this arrangement is undoubt-
edly due to ancestral conditions, for the present bilateral
habits of Thyone could probably have no influence in
producing this asymmetry.

9. The retractor muscles in a single radius consist of
single or multiple strands, and this variation is closely
correlated with a similar variation in the number of
Polian vesicles. No explanation is forthcoming for this
peculiar plasticity of the retractor muscles, but the sug-
gestion is made that it may be functionally correlated
with the development of the water- vascular system.

10. It was found from the study of a number of speci-
mens that individual peculiarities of structure tend to be
reproduced in the process of regeneration. In this proc-
ess it would appear that individual variations tend to
predominate over generalized ancestral influence.

11. Autotomy enables Thyone to survive for a consid-
erable period on a smaller than normal supply of oxygen.
Nevertheless, the conditions which give rise to self -muti-
lation are seemingly in all cases pathological.

12. The conditions in Thyone afford some evidence for
believing that when this animal abandoned the fixed stage
the Polian vesicles conformed more or less to the radial
type. This is opposed to the statement of Lang that in
all cases where a multiple number is now present ** there
was originally only one vesicle.'' It is believed that the
present arrangement of Polian vesicles in Thyone can be
best accounted for on the theory of phylogenetic influ-
ence. That, in general, those vesicles have retained their
most complete radial arrangement in those species of

Digitized by

Google

No. 569] REGENERATION 307

echinoderms which have maintained to a high degree
the functional activity of the water-vascular system.

References
Bather, F. A., and Goodrich, E. S.

'00. A Treatise on Zoology. Ed. by E. Ray Lankester. Pt. 3. The
Echinodermata. London,
aark, A. H.

^09. The Affinities of the Echinoidea. Am. Nat., Vol. XLIII, pp.
682-686.
aark, H. L.

'99. The Synaptas of the New England Coast. Bull. U. S. Fish Com.,

1899, pp. 21-31.
'00. The Echinoderms of Porto Rico. Bull. U. S. Fish Com., 1900,

pp. 231-263.
'02. The Echinoderms of the Woods Hole Region. Bull. U. S. Fish
Cbm., 1902.
Cuenot.

'91. Etudes morphologiques sur les Echinodermes. Archiv, de Biol,
Vol. XI.
Gerould, J. H.

'96. Anatomy and Histology of Caudina arenata, Proc, Boston Soc.
Nat. Hi8t., Vol. 27, pp. 8-74.
Grave, C.

'03. On the Occurrence among Echinoderms of Larvae with Cilia ar-
ranged in Transverse Rings, with a Suggestion as to their Sig-
nificance. Biol. BiUl, Vol. V, pp. 169-186.
'05. The Tentacle Reflex in a Holothurian, Cucumaria pulcherima,
Johns Hopkins Univ. Circ, Vol. 24, pp. 504-^07.
Henri, V.

'03. Etudes des contractions rhythmiques des vaisseauz et du poumon
aqueaux chez les Holothuries. C. B. Soc, Biol., Paris, T. 55, pp.
1314-1316.
Johnson, R. H., and Hall, R. W.

'00. Variations and Regeneration and Synapta inhcerens. Science,
N. S., 1900, p. 178.
Lang, A.

'96. Text-book of Comparative Anatomy. London, Vol. II.
Ludwig, H.

'96. Echinodermen. Bronn's Klassen u. Ord. des Tierreichs. Bd. II,
Abt. 3, Buch 1.
Morgan, T. H.

'01. Regeneration. The Macmillan Co., 1901.
Packard, A. D.

'81. Zoology for High Schools and Colleges, 1881.
Pearse, A. S.

'08. Observations on the Behavior of Thyone hriareus (Leseur). Biol.

Bull, Vol. XV, pp. 259-286.
'09. Autotomy in Holothurians. Biol. Bull, Vol. XVIII, pp. 42-49.

Digitized by

Google

SHORTER ARTICLES AND DISCUSSION

TERMS RELATING TO GENERIC TYPES

In the field of biological taxonomy an important reform is in
progress. The change is from the method of concepts to the
method of types, in order that names may be applied with
greater precision and permanence. Under the method of types
we no longer think of the technical name of a plant or an animal
as attaching primarily to a concept embodied in a description or
definition, but as relating to the first representative of the group
that became known to science. In determining the application
of a specific name we go back to the original specimen or type
on which the description was based. The original description
has become secondary to the original specimen. In like manner
generic names are treated as relating primarily to groups of
species, with the original species as the generic type.^

Without waiting to appreciate the fundamental nature of the
change from concepts to types, many systematic workers took it
for granted that generic types were to be determined by elimina-
tion, in much the same way that generic concepts had been
treated, by gradual subdivision, restriction and removal of com-
ponent groups. The general results of elimination were the
same as under the method of concepts: The applications of
many of the older generic names did not become definitely fixed,
but remained dependent upon varying individual opinions of
the validity of the work of later authors. It often happened
that after elimination was accomplished only the doubtful or
unidentifiable species remained to serve as generic types. Grad-
ually it became apparent that the practise of elimination was
inconsistent with the method of types, and could not insure
stability in the application of names. Recourse was then had,
especially by zoologists, to the arbitrary designation of generic

1 Cook, O. R, 1898, "The Method of Types," Science, N. 8., 8: 513; also
1900, "the Method of Types in Botanical Nomenclature," Science, N. 8.,
12: 475, and 1902, "Types and Synonyms,'' Science, N. S., 15: 646.
Swingle, Walter T., 1913, "Types of Species in Botanical Taxonomy,"
Science, N. S., 37 : 864.

308

Digitized by

Google

No. 569] SHORTER ARTICLES AND DISCUSSION 309

types, the apparent object being to preserve the results of elimi-
nation, even though the theory had to be abandoned. Probably
it is only a question of time until the results of elimination will
be discarded, as well as the theory, and replaced by the actual,
historical types.

A plan for determining the historical types of genera was
adopted in 1907 in the American Code of Botanical Nomencla-
ture, and other applications of the method of types are being
rocognized by zoologists. Specialists in many groups are en-
gaged in the study of generic types, and the need of a special ,
terminology to facilitate work of this kind is becoming appar-
ent. Thus in Bulletin 83 of the U. S. National Museum, **Type
Species of the Genera of Ichneumon Flies,'' by Henry L. Vie-
reck, two new terms, **isogenotypic" and ** monobasic," are em-
ployed in treating of the application of generic names to type
species. The paper is of interest, not only to students of this
group of insects, but also as an example of the tasks that con-
front all taxonomists who appreciate the need of basing their
work upon types. The distinctions to which the special terms
refer are undoubtedly useful, and the possibilities of express-
ing them in more convenient form are worthy of consideration.

The word **isogenotypic,'' is used with reference to cases
where two or more generic names have been applied to the same
type species.^ For this purpose a new term is not needed unless
zoologists are unwilling to borrow from botanical nomenclature
a more convenient method of treating the same class of eases.
The botanical code provides a classification of synonyms, and
applies the word **typonym" to a name that has to be rejected
because an earlier valid name was proposed for the same type.
The formation and use of typonym are in accord with a familiar
analogy. As a preoccupied name becomes a homonym, it is easy
to remember that the use of a preoccupied type results in a

2 A diflferent combination might have been expected, such as * * autogeno-
tjrpic'^ or * * deuterogenotypic, " since isogenotypic suggests the notion
of equally good types or of equal numbers of types, instead of con-
veying the idea of one and the same type, or of a second use of the same
type. Genera have been termed * * isotypical * ' when they were described
from more than one species, but all truly congeneric, on the assumption that
such species would have equal standing as types. A still older use of the
word **i9otype'' had reference to equal representation of a genus by similar
or corresponding species in different geographical regions or geologic periods.
See Schuchert, Charles, 1905, U. S. National Museum Bulletin 53, Ft. 1 : 16.

Digitized by

Google

310 THE AMERICAN NATURALIST [Vol. XLVIH

typonym. A name based on a different type species, but eon-
generic with the type of an older genus, is termed a metonym.
A name rejected for lack of an identified type is a hyponym,
and one rejected for linguistic reasons, a caconym. All rejected
names fall readily into these five classes.

The other new term, *' monobasic," is used by Mr. Viereck to
indicate genera with only one species at the original place of
publication. In botanical literature the word ''monotypic" is
often employed in this sense, though also applied to genera that
consist of only one species. If previous use disqualifies mono-
typic, the same objection lies against monobasic. In addition to
an older chemical meaning, the same word was employed several
years ago in a biological sense, to describe a condition of descent
in simple lines.* Apart from being preoccupied, the word mono-
basic has a misleading implication, since under the method of
types each generic name is referred to a single type species.
The idea of a genus being based on many types is discarded
with the method of concepts. Appreciation of this incongruity
may explain why no such term as "symbasic" or *'polybasic'*
is used in contrast with monobasic, to indicate genera that were
first proposed in connection with more than one species.

Evidently there is need of a simple and consistent terminology
for indicating relations between generic names and type species.
The normal relation under the method of types is the designa-
tion of the type species at the original place of publication of
the genus. Genera provided with types by original designation
may be described as orthotypic, or normal-typed. With ortho-
typic genera there is no occasion to raise the question of how
many species were included at the original place of publication.

8 Cook, O. F., and Swingle, W. T., 1905, ** Evolution of Cellular Struc-
tures," Bull. 81, Bureau of Plant Industry, U. S. Department of Agricul-
ture, p. 20. Plants or animals with specialized habits of asexual reproduc-
tion, such as vegetative propagation, parthenogenesis or self-fertilization,
would be described as monobasic. The second edition of the Standard Dic-
tionary defines monobasis as follows: "The derivation of a stock from a
single parentage by inbreeding, or by propagation of buds or cuttings;
opposed to symbasis." Thus the danger of ambiguity in using monobasis
for nomenclatorial purposes is greater than in using monotypic, though it
must be admitted that the use of the word monotypic in two senses may
sometimes result in confusion. Genera that were monotypic in the strictly
nomenclatorial sense of being established in connection with one species may
not be monotypic in the more general taxonomic sense of including only one
species.

Digitized by

Google

No. 569] SHORTER ARTICLES AND DISCUSSION 311

Genera that are not orthotypic fall into the two classes al-
ready considered, those with a single species at the original place
of publication, and those with two or more species. It is now
generally agreed that when only one species was mentioned this
should be accepted as the type. Such genera may be called
haplotypic, or single-typed. When two or more species were in-
cluded in the original treatment of a genus, and no type was
designated, we have the problem of subsequent determination of
the type, resulting in what may be termed a logotypic genus,
that is, a genus with a rationally selected type species. The
object of selection is to determine the historical type of the
genus. Names must have definite applications, and historical
applications of generic names can be made definite by ascertain-
ing the historical types. The recognition of a new generic group
is usually based on one leading or dominant species, with the
others added as associate members.

In many cases the generic type is intimated by the original
author in dividing the genus into subgenera or sections, in illus-
trating one of the species or citing illustrations published in
earlier works, in naming the genus with particular reference to
one of the species, in recording economic uses, or in giving geo-
graphical or other indications of greater familiarity with one of
the species. If the application of these or other historical cri-
teria leaves more than one species eligible for selection, the first
of the eligible species should be taken as logotype. In this way
it is possible to develop a consistent system of type selection that
will commend itself as reasonable and give the same results in
the hands of different students.*

4 Simply taking the first species under a generic name as the type would
probably establish more of the generic names in their historical places than
the method of elimination, which accepts the last of the original species left
in the genus as the type. Either of these methods of selecting types would
result in many cases of separation of generic names from their historical
types, but these undesirable changes in the application of names can be
avoided by taking the historical considerations more directly into account,
as in the American Code of Botanical Nomenclature. Probably a more sat-
isfactory system for associating generic names with their historical types
could be developed by sufficient study of the problem. A policy of refusing
to revive generic names that were not directly associated with binomial spe-
cies to serve as types, would avoid many of the changes threatened by un-
mitigated priority. In proposing lists of **nomina utique conservanda" in
advance of any provision for the definite application of names, European
botanists have demonstrated one more way to put the cart before the horse.

Digitized by

Google

312 THE AMERICAN NATURALIST [Vol. XLVHl

In addition to the three ways of associating generic names
with their type species, there are many cases where generic
names have been applied to groups that do not include the type,
or any of the original species. Formal assignments of errone-
ous types also occur when generic names are not traced back to
their original places of publication, or when ineligible species
are designated as types. In dealing with the synonymy of
genera previously treated under names that belong to other
groups it will be convenient to have a distinctive term for this
class of cases. Such misplaced names, applied to groups that do
not contain the true type, may be indicated as pseudotypic, or
false-typed.^

It should be expected that more critical analysis of taxonomic
problems would lead to more definite distinctions and more pre-
cise terms. The older terminology was developed to facilitate
the study of names, whereas it is now apparent that provision
must be made for the study of types as another formal branch
of biological taxonomy. Nomenclature has a history of three
hundred years while systematic typology is only beginning. To
gain further insight into these typological problems is obviously
more important than to attempt premature applications of par-
tial solutions. It may take fifty or a hundred years to transfer

Failure to regulate the application of names is the fundamental defect of the
Paris and Vienna codes, and is hardly to be cured without thorough re-
casting.

5 Thus the palm genus Martinezia, as treated by Kunth, Martius, and
many later writers as relating to Martinezia caryotwfolia and its immediate
relatives, was pseudotypic, for this species does not appear to be congeneric
with any of the five species originally referred to Martinesia by Buiz and
Pavon. Hence it has been proposed to replace this pseudotypic use of Mar-
tinezia by a new generic name, TiXmia, (See Bull, Torrey Bot, Club, 28:
565.) The five original species of Martinezia belong to three natural groups,
now recognized as distinct families, the first two species to the Cocace®, the
third species to the AcristacesB and the others to the Chamsedoreacese. The
third species, M, ensiformis, should be taken as logotype of Martinezia be-
cause the figures used to illustrate the generic characters evidently represent
a member of the family Acristacese. Another reason for excluding the
cocoid species from consideration as type is that they are mentioned as
deviating from the ''essential characters of the genus,'' in connection with
the original description. The rule of the Vienna code, to the effect that the
name of a subdivided genus should go with the majority cf the species, would
carry the name Martinezia over to the family ChamsBdoreace®. The making
of such a rule shows that many European botan'sts were still working imder
the method of concepts, and were not accustomed to think of generic names
as inseparably connected with type Spec'es.

Digitized by

Google

No. 569] SHORTER ARTICLES AND DISCUSSION 313

the whole structure of biological taxonomy to the new founda-
tion of types. To suppose that any permanent adivantage can
be gained by elaborating defective methods under forms of legis-
lative enactments or judicial decisions is to show a limited ap-
preciation of the nature of the subject and of its historical de-
velopment. As long as legislation and interpretation are based
on inadequate study, they can represent, at most, only a tem-*
porary consensus of opinion, for it is of the very nature of
science to condemn and throw aside any doctrine or method that
has proven inadequate or fallacious.

Terms Relating to Synonyms

The following classes of synonyms were recognized in 1907,
in the American Code of Botanical Nomenclature:**

Homonym, — A name rejected because of an earlier applica-
tion of the same name to another genus.

Typonym, — ^A name rejected because an older name was based
on the same type.

Metonym, — ^A name rejected because an older valid name was
based on another species of the same genus.

Hyponym. — ^A name not associated with a type.^

9 Bulletin of the Torrey Botanical Club, 34: 167, 1907. .-r— .

7 Much confusion would be avoided by a consistent policy of withhold-
ing recognition of generic names that have not been associated with type
species. Thus the name Acoeloraphe, proposed by Wendland in 1879 in an
analytical key to genera of fan-palms (Bot, Zeitung, 37: 147), was not as-
sociated with a type, though evidently relating to a species mentioned in the
same paper as *'Bmhea serrulata," This Florida palm differs from the
Mexican type of Brahea in the leaf characters assigned to Acoeloraphe in
the key and in the seed character indicated by the generic name, the al-
bumen being solid instead of having a deep channel along the raphe. But
Acoeloraphe being left without a type, another name, Serenoa, was proposed
by Hooker f. in 18S3 for **Sabal aerrvXata R. & S.'' (Genera Plantarum,
3: 926). All subsequent writers have accepted Hooker's name, and Acoel-
oraphe should remain under Serenoa aa a hyponym. Nothing has tended so
strongly to bring the principle of priority into disrepute as the incontinent
revival of abortive names, to replace properly established names in current
use. No species was referred to Acoeloraphe until 1907, when Beccari
(Webbia, 2: 107) applied the name Acoeloraphe wrightii to a Cuban mem-
ber of a genus that had been described in 1902 under the name Paurotis,
a Bahaman species, Paurotis androsana, being the type (Mem. Torrey Bot,
Club, 12: 21). This transfer of the name Acoeloraphae to the genus
Paurotis was followed by Sargent in 1911 (Trees and Shrubs, 2: 117), but
Beccari 's genus Acoeloraphe is a metonym of Paurotis, and is also pseudo-

Digitized by

Google

314 THE AMERICAN NATURALIST [VOL.XLVni

Terms Relating to Type Species

Orthotype. — Type by original designation. A species desig-
nated as type in connection with an original publication of a
generic name. A genus whose type was formally designated at
the original place of publication is orthotypic.

Haplotype. — Type by single reference. A single species re-
ferred to a genus at the original place of publication, and on
this account accepted as the type. A genus proposed with refer-
ence to a single species is haplotypic.

Logotype, — Type by subsequent determination. The histori-
cal type of a genus, selected from two or more original species.
A genus whose type is selected from two or more original species
is logotypic.

Pseudotype, — Erroneous indication of type. A species erro-
neously indicated as the type of a genus. A genus treated on
the basis of an erroneous type, or so as to exclude the true type,
is pseudotypic.

0. F. Cook
BuEEAU OF Plant Industry,
U. S. Department or Agriculture,
March 13, 1914

typic, because of the original application of the name to Serenoa. Two
species of Paurotis are supposed to exist in Florida, one that is identi-
fied with the Cuban P. urightU (Grisebach & Wendland) and a local species,
P. arhorescens (Sargent).

Digitized by

Google

NOTES AND LITERATURE

LINKAGE IN THE SILKWOEM MOTH

One of the most striking recent developments in the study of
genetics is the discovery of linkage in many of those forms which
were supposedly thoroughly worked out. The most recent ex-
ample* is a very interesting paper by Y. Tanaka,^ entitled
** Gametic Coupling and Eepulsion in Silkworms." The data
presented in this paper demonstrate the existence in the silk-
worm moth of a group of four pairs of linked genes. Following
Tanaka's nomenclature we may designate these genes as follows:
N, which differentiates the larval color pattern known as
** normal'' from that called ** plain"; S, occurring in larvae
having the ''striped" pattern, and epistatic to N; M, the differ-
entiator for the **moricaud" larval pattern, also epistatic to
N; Y, the gene which differentiates caterpillars with yellow
blood and yellow cocoons from the recessive whites. Of the six
possible combinations of these genes, taken two at a time, all
but NM and SM were made, and all showed linkage. Fj
''coupling" tests, i. e., from matings where both dominants
entered the cross from the same Pi individual, were made for
SY and for MY, In each case there occurred cross-overs, or new
combinations of the characters, in such proportions as to lead
Tanaka to suppose the ratio of parental combinations to cross-
overs among the gametes to be about as 7:1. "Repulsion"
(where one dominant entered from each P^ individual) Fj
results were obtained for NS and for NY, In neither case did
any double recessives (cross-overs) appear, though over 3,000
caterpillars were obtained in the case of NY, and 224 in the case
of N8. From these data Tanaka concludes that the repulsion was
complete in these two cases. It has, however, been pointed out by
Morgan^ that such results will be obtained if the linkage is com-
plete in one sex only. In Drosophila such "repulsion" crosses
never produce double recessives in Fj, and it has been shown
that this is due to complete linkage in the male, crossing over
being frequent in the female between some pairs of genes. In
order to test this possibility it is necessary to mate doubly hetero-
zygous individuals to double recessives, when the gametic ratio
is obtained directly and without the complications present in
most Fg results. It so happens that Tanaka reports two such
crosses, one for each sex, though he does not recognize their im-

1 Jour, Coll, Agr,, Tohoku Imper. Univ., Sapporo, Japan, V, 1913.

2 Science, N. S., XXXVI, 1912.

315

Digitized by

Google

316 THE AMERICAN NATURALIST [Vol. XL VIII

portance in this connection. When a male heterozygous for S
and for Y, one dominant having been derived from each parent
(SysY), was mated to a doubly recessive {sysy) female, there
were produced 63 8y and 65 sY — no cross-overs. A female
heterozygous also for S and for Y, but having them ** coupled"
(SYsy), was mated to a male sysy, and produced 215 SY and
188 sy — again no cross-overs. Yet that crossing over may occur
between these two pairs of genes is shown by the fact that the
** coupling*' F2 results indicated a gametic ratio of about
7:1:1:7. We are, therefore, still left in the dark as to whether
crossing over occurs in only one sex, or in both. But it is certain
that the strength of linkage in this case is not always the same —
a point of great interest and importance. Similar cases have been
reported by Baur* in Antirrhinum, by Punnett* in the sweet pea,
and by the writer^ in Drosophila,

Tanaka refers to his case as differing from previously reported
cases of linkage in animals in that the sex differentiator is. not
one of the genes involved, and in that the linkage is sometimes
only partial. However, he refers several times to a paper by
Morgan* in which it is clearly shown that three of the sex-linked
genes in Drosophila also show partial linkage to each other, inde-
pendently of their sex-linkage. Punnett,^ in referring to the
same paper, has said, ** Morgan's experiments with Drosophila
suggest coupling of some kind between factors for eye color and
shape of wing, though both of these factors may show sex-limited
inheritance in other families." A study of the data referred tOy
or of any of the similar data on Drosophila since published, will
show that these genes always show sex-linkage, and that at the
same time they always show linkage to each other when both can
be followed in the analysis. The two phenomena are not mutu-
ally exclusive, but both are always present.

Both Tanaka (in a footnote) and Punnett refer to the latter 's
case in rabbits as the first example of linkage in animals not
involving sex. If the linkage between sex-linked genes is, for
some strange reason, not considered to belong in this category,
there are still at least two cases which antedate Punnett 's slightly.
A few months before Punnett 's paper appeared I had suggested*
the possibility of linkage in mice. It now seems rather probable
that -the relation in both mice and rabbits may really be that of

8 Zeits, f, ind. AbsUu, Vererh.-Lehre., VI, 1912.

^Jour, Genet., Ill, 1913.

5 Science, N. S., XXXVII, 1913.

ojaur, Exp. Zool, XI, 1911.

tJour. Genet., II, 1912 (Nov.).

8AMER. Nat., XLVI, 1912 (June).

Digitized by

Google

No. 569] NOTES AND LITERATURE 317

triple allelomorphism. For this reason I am inclined to assign
priority to Morgan and Lynch,® whose paper on linkage of genes
in Drosophila which are not sex-linked appeared after my own
paper and before Punnett's.
Columbia University A. H. Sturtevant

NABOURS'S BREEDING EXPERIMENTS WITH
GRASSHOPPERS

In a recent paper, Nabours ('14) describes breeding experi-
ments that he has been carrying on for some years with grouse
locusts of the genus Paratettix, His work is of special interest
in showing that in a wild species there exists a number of distinct
types that show alternative inheritance of a particular kind. His
paper may be summarized as follows :

1. Nine distinct, true breeding forms of Paratettix were col-
lected **in nature." These ** species" (as Nabours is inclined
to consider them) **are mainly distinguished by their striking
color patterns."

2. When an individual of one of these species is mated to one
of a different species the hybrid character of the oflEspring is
apparent at once, in that **all the characters of each parent are
represented in the F^ hybrid. ' * In other words, the hybrid is in
a certain sense an intermediate, and **the terms dominant and
recessive" are probably not ** applicable at all." This point,
while of little theoretic importance, has a practical value in that
the zygotic constitution of any hybrid can be recognized without
further breeding tests.

3. With one exception, each color pattern factor was found to
behave as an allelomorph to any other color pattern factor.

4. The various lengths of the wings and pronotum are appar-
ently not inherited, as such but are determined by environmental
factors, especially such as tend to prolong or to shorten the length
of larval life.

It appears that Nabours confuses the relation of the facts men-
tioned under 3, and that he supposes this to be the ordinary
behavior of "mendelizing characters," for he says:

The essential result of these experiments has been the extension of
this principle [Mendelian inheritance] to a considerable number of
tjrpes of a phylogenetically low group of ametabolous insects.

To be sure, he recognizes that other workers in genetics have
an attitude quite different from his, and he takes some little pains
to make clear his own point of view. To quote again (p. 142) :

• Biol Bull, XXIII, 1912 (Aug.).

Digitized by

Google

318 THE AMEBIC AN NATURALIST [Vol. XLVHI

The existence of unit characters in the De Vriesian sense does not
appear to have been as clearly demonstrated as that of alternative in-
heritance . . . and the interpretations are at great variance. Thus, one
group of authors [reference made to Bateson, Doncaster, and Tower]
recognize characters in organisms that can be replaced by other char-
acters when the proper crosses are made, . . . while on the other side
there are those [references to Whitman and Montgomery] who believe
that the organism as a whole is the only unit and that there are no
actual unit characters.

Again he says (p. 169) :

No character of one parent species is ever replaced in the F^ hybrid
by any character of the other parent. All the characters of each parent
are represented in the F, hybrid. It follows then that these grass-
hoppers do not exhibit characters which by crossing can be replaced by
other different characters; the whole pattern appears to be the only
unit.

There is no real conflict between Whitman's idea and the
accounts given by students of Mendelism, for the latter realize
that far-reaching somatic effects may result from a single factor,
and the composite character of the hybrid is not an uncommon
occurrence. Nabours identifies a particular pattern with the
* 'organism as a whole/' but since his evidence relates here to
color patterns only, nothing is gained by the introduction of such
a vague phrase as the ** organism as a whole." Specifically he
shows that the hereditary differences between any two types can
be explained on the assumption of a single differential for each
case.

With reference to the antithesis presented by Nabours, it must
be recognized that the modern literature of Mendelian heredity
affords innumerable instances where two or more characters
entering from one parent and their allelomorphs from the other,
reappear in the Fg generation in new combinations.

If we assume with Nabours that each of the eight color patterns
are represented by a characteristic condition of the "germinal
material," we may use his terms A, B, C, D, E, F, H or I to
symbolize this ** germinal material" for the various color
patterns. As Nabours uses the terms, an individual homozygous
for A is represented simply by A, and a hybrid between A and B
by AB. In ordinary usage, the homozygous form would be
represented as AA and its germ cells by A. This is a minor
matter. Ordinary usage has the advantage of being more
consistent.

According to Nabours, then, A mated to B gives AB; S mated
to F gives BF; C mated to E gives CE, etc. In gametogenesis
these factors segregate, so that, for example, BA gives germ cells

Digitized by

Google

No. 569] NOTES AND LITERATURE 319

A and B; BF gives B and F, etc. In other words, he treats the
matter as if he were dealing with a system of multiple allelo-
morphs, though he nowhere specifically calls them such. From
this point of view there are eight distinct allelomorphs con-
cerned with color pattern any two of which may constitute a
pair; in any zygote two allelomorphs (perhaps alike, perhaps
unlike) will be present, and in any gamete only one of the eight
will normally occur.

With one exception of which I will treat later, all of Nabours's
results can be explained by this hypothesis. This sort of explana-
tion is not new. (ShuU ('11), de Meijere ('10), Sturtevant
('13) and others have used it to explain results obtained in
Lychnis, Papilio, rabbits and other forms, and it will almost un-
doubtedly be shown to apply satisfactorily in still other cases.

The exception just mentioned occurred in the cross which
Nabours describes at the bottom of page 156 (e). Here a male
of the constitution CE was mated to a female of the constitution
BI, On Nabours 's theory, the gametes of the male should carry
C or E, but not both, and the gametes of the female should carry
B or /, but not both. The union of the two kinds of sperms with
the two kinds of eggs should give four classes of offspring, and
these were in fact obtained ; viz., 12 BC, 11 BE, 7 CI, 10 EL But
there appeared also on^ individual BEL Nabours 's explanation
of the case is that perhaps the BI * 'female parent gave at least
one gamete containing the factors for the patterns of both her
parents and that this double character gamete was fertilized by
one of the E gametes which came from the CE male." * Let us
see whether this is the most probable interpretation.

As Sturtevant has pointed out, for any case to which the idea
of multiple allelomorphism is applicable, an equally valid ex-
planation may be found in ** complete linkage" of the factors
concerned. To decide in any case between the two explanations
would be impossible.

If, however, linkage were not complete, a ** cross-over" class or
''recombination" class might occur, and this would suffice to rule
out the explanation based on multiple allelomorphs.

Such a "cross-over" class perhaps is furnished by the BE I
individual. The demonstration of this may be given by the use
of symbols, as follows :

Let us assume that A is the allelomorph of a, B that of b, C of
c, D of d, F of /, / of i, etc., making eight pairs of allelomorphs
altogether. Assume that each gamete of any individual carries

1 This explanation is essentially similar to that advanced hj Bridges ( '13)
to explain certain peculiar results in DrosophUa, Bridges assumed that in
garnet ogenesis the two X -chromosomes of a whiteeyed female failed to segre-
gate (in Bridge's terminology, non-disjunction occurred), and passed over
together into one gamete.

Digitized by

Google

320 THE AMERICAN NATURALIST [Vol. XLVin

one allelomorph of each pair, and that the eight factors thus
present in a gamete form a linked group, tending to segregate as
a unit in gametogenesis. Thus Nabours's form A would give
gametes of the form Abcdefhi, AB would give gametes of only
two forms, one corresponding to A and the other to B, viz.,
Abcdefhi and aBcdefhi, Two other forms are possible, formed
by the exchange of A with B, and of a with 6, but these will not
occur if linkage is complete. In dealing with the hybrid AB in
practise the factors cdefhi would not be put into the formulas,
as they are alike in all gametes.

These rules would apply similarly to all other species and
hybrids. Therefore in the case in which the BEI individual
occurred, we would represent the male parent, which Naboure
designated CE, by bCei-bcEi, and its gametes by bCei and bcEi,
The female parent, which Nabours designates BI, we would
represent by Bcei-bcel, and its gametes would be Bcei and
bcel if linkage were complete. If linkage were not complete there
would occasionally be formed gametes bcei and BceL One of
these latter (Bcei) was probably formed and fertilized by a
sperm of the type bcEi, thus giving rise to the BEI individuaL
No gametes corresponding to bcei appear to have been fertilized,
though of course we do not yet know what the appearance would
be of an individual so formed.

This matter would be easy to test, and it is to be hoped the
cross may be repeated. If then BEI forms should appear again
and in these when mated to other forms the factors B and I should
be found to stay together to the same extent as they before sepa-
rated, it would show that close linkage, rather than multiple alle-
lomorphism explains this particular instance.

It may be, too, that both linkage and multiple allelomorphism
play a part in the production of these phenomena. In any case
it seems that the test is at hand, and not difficult to perform,
excepting in so far as there are practical difficulties connected
with the rearing of the grasshoppers in sufficient numbers to
cover the point.

Literature Cited.

Bridges, C. B. 1913. Non-disjunction of the Sex Chromosomes of Droso-
phila. Jour, Exp, Zool,, Vol. 15.

de Meijere, J. C. H. 1910. Ueber Jacobsons Zttchtungsversuche bezttglich
des Polymorphismus von Papilio Memnon L. ?, etc. Zta, ind. Ahsi.-
Vererb.-Lehre, Vol. 3.

Nabours, B. K. 1914. Studies of Inheritance and Evolution in Orthoptera,
I. Jour, Genet., Vol. 3.

ShuU, G. H. 1911. Reversible Sex Mutants in Lychnis dtoica. Bat, Gaz^
Vol. LII.

Sturtevant, A. H. 1913. The Himalayan Rabbit Case with some Consid-
erations on Multiple Allelomorphs. Am. Nat., Vol. XLVII.
Columbia University John S. Dextbb

Digitized by

Google

o Important Biological Journals

The Journal of Qenetic3

Edited by
W. BATE80N and R C. PUNNETT
iodioal for ihe publication of reoorda of original retearoh in hereditj, Tariation, and allied topiot.
KAL also contains, from time to time, artioTes summarizing the existing state of knowledge in the
sndies of genetics. The numbers are fullj illustrated, and, when the snbjeot*matter demands it,
fttes are fredy employed. The Journal is issued in parts as material accumulates, and a volume
^ so far as possible, annually) consists of four such parts. Subscription price, $7.60 a Tolume ;
ies,^50.

3iometrilca

Edited by EARL PEAB80N
mal for the statistical study of biological problems, appearing about four times a year. A Tolume
\ about 500 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
ding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
ry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, 12.50.

I on back numbers of these journals vary, and postage from London is charged on back volumes
copies.

in America for the Cambridge University Press of England by

*he University of Chicasro Press .

CHicasTO^ Illinois J

University Oontrol

ij J. MoKebn Cattbll^ Professor of Psychology in Golombia TTniyeraity

her with a series of Two Hundred and Ninety-nino Unsigned Letters by T<eading lien of

lademie Positions and Articles by JoenpH Jastbow, Qbobob T. Ladd, Jomr J. Sramsov, J. B. /

r, J. MoKsBH Cattblih Osobgb M. Stbattov, Btbwabt Patov, Johm Jat CsAnuv, Jamss P. |

ind Jacob Gould BcHinacAN. t

it Tsriety of guestions oonceming general university administration are dealt with in an orifinal and |

y. — Nahir9, i

quotatioDS and atami^es are taken from Ptofeesor CatteQ's informed and thorough diseusrion of the I

anivenity control, a subjeet upon which he has had much to say of lata, finding occasion for eaustie }

r wristing Amen -an conditions, and standing as the champion oi an academic democracy and a teadn |

ion upon which a man may enter without forfeiting his self-respect.— TAs Ditd, t

IOCS and paragraphs that betoken the expert, highly-trained mind, the suggestions that come to i#- I

M us that a new day is about to dawn in educational writing. — Th» Bottom Kwoning Tnmooript* l

SCIBIVCe AND EDUCATION j

I of rolttmea for the promotion of Bcienti&o resesircb and eduomtional progroBM i

h The Foondatioiis of Sdeoce. By H. PonroAni. Containing the authorised English translatioB j |

■OBOT Bbocb Halstbd of "Sdenoe an^ Hypothesis,'! "The Value of Sdenoe,'! and !'8oi«nea •' r

lethod.- Price, $3.00 net. }
! n. Medical Research and Sdocatlon. By Bicraxd hL Pbabob.' William H. Waum; W. H.

BLL, FbAIIXLUI P. MaLL, LaWKXTS F. BaBXBB, CHABUU 8. MiNOT, W. B. CAMNOir, W. T.

qhjuv, Tbbobald Smith, G. N. Stbwabt, C. M. Jacksoh, B. P. Lton, Jambs B. Hbbbioe, Jorm
onsox, C R. BAitDBBTff, W. OphOlb, a J. MsLnBB, Jambs Ewmo, W. W. Kais, Hbbst H. Dovald-
^BBBTiAiv A. Hbbxbb, and Hbmbt P. Bownrrca. PrioCi 13.00 net.
: m. Uniraitity Control. 2fom BoaO^. Poqw »+^4. PHoo. %SM asf.

ON. N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84. NEW YORK CITY

Digitized by

Google

The American Naturalist

A MontUj Jounial. ettablished in 1867, Devoted to the Adrancement of the BiolofficiJ Sci— ret
with Special Reference to the Factors of Orvanie Evolution and HeredUF

OONTENT8 OF THE NOVEIMBER NUMBER

The Effect on the Oflipring of Intoxicating the Kale
Parent and the TransmiBElon of the Defects to
Subsequent Oenerations. Dr. Charles R. Stockard .

Supplementary Studies on the Differential MortaUty
with respect to Seed Weight in the Germination
of Garden Beans, Dr. J. Arthur Harris.

Shorter Articles and Discussion : Reciprocal Crosses
between Reeve^s Pheasant and the Common
Ringneck Pheasant producing Unlike Hybrids.
John C. PhilUps.

CONTENTS OF THE DECEMBER NUMBER

The Fixation of Character In Organisms. By Edward

Slnnott.
Inheritance of Left-handedness. Prof enor Frandi

Bamaley.
Supplementary Studies on the Differential MortaUty

with Respect to Seed Weight in the Germinatioo

•f Garden Beans, IL Dr. J. Arthur Harris.
Shorter Articles and Discussion : A Croas inTohriay

Four Fairs of Mendelian Characters In mee. C

C. Little. J. C. PhUllpi.
Index to Volume XLVII.

CONTENTS OF THE JANUARY NUIMBER

A Genetic Analysis of the Changes produced by
Selection in Experiments with Tobacco. Pro-
. fessor £. M. East and H. E. Hayes.

Gynandromorphous Ants, described during the De-
cade, 1908-1918. Professor William Morton
Wheeler.

Shorter Articles and Discussion : On the Results of
Inbreeding a Mendelian Population— A Correc-
tion and Extension of Previous Conclusions.
Dr. Raymond Pearl— Isolation and Selection
allied in Principle. Dr. John T. Gulick.

CONTENTS OF THE FEBRUARY NUMBER

Some New Varieties of Rats and Guinea-pigs and their
Relations to Problems of Color Inheritance. Pro-
fessor W. £. Castle.

" Dominant" and " Recessive" Spotting in Mice. C.

C. Little.
On Differential MortaUty with respect to Seed Weight

occurring la Field Cultures of Pisum sativvm.

Dr. J. Arthur Harris.
The Inheritance of a Recurring Somatic Variation

in Variegated Ears of MaUe. Profeisor R. A.

Bmerson*
Restoration of Edaphosaurus crudger Cope. Pro-
fessor K. C. Case.
Shorter Articles and Discussion : Hum idity— a

Neglected Factor In Enylronmental Work. Dr.

Frank £. Luts.

CONTENTS OF THE MARCH NUMBER

The Effect of Extent of Distribution on SpecUtlon.

Asa C. Chandler.
Biology of the Thysanoptera. Dr. A. Franklin Shull.
Shorter ArUdes and Correspondence : The Endemic

Mammals of the British Isle& Professor T. D. A.

Cockerell.

Notes and Literature : Swingle on Variation in Fi
Citrus Hybrids and the Theory of Zygotaxis.
Dr. Orland £. White.

CONTENTS OF THE APRIL NUMBER

Dr. Henri Hus.
Biology of the Thysanoptera. II. Dr. A. Fraakltn

Shull.
Shorter Articles and Discussion : Barriers as to Dia-

tribution as regards Birds and Mammals. Joaepb

Grinnell. Yellow Varieties of Rati. Piof^BMr

W. £. Castle.
Notes and Literature : Heredity and the InHiMiioe

ofMonarchs. V. L. K.

Single Number 40 Cento Yeariy Subwaription, $4.00
The NATUIUUST wffl U aenl to to^w •obMribew lor loi^

THE SCIENCE PRESS

Garrison, N. Y. Substation 84 : NEW YORK Lancaster. P^

Digitized by

Google

VI i i'M4

▼OL. ZLVm, HO. 570 JTHIE. 1914

THE

AMERICAN
NATURALIST

A HONTHLT JOUBVAL

Herotad to fhe Adyanoement of fhe Biological Scienoat wifh

Special Beference to fhe Faotore of BYolution

OOVTBHTS

Pag$
I. Sp«el6t-biilldl]if toy HytorUUsatton aad Kntatloii. Professor John H.

Gbboitld ------------- sai

XX. Httr«4it7 of BrlBtlet in the Oommon QreontootUe ny — A Study of Factors

gOTmdag Distritoution. Phikbas W. Whitino - - - - - 339

XXI« Pby tiolofleal Correlations and Climatic Reactions in AUUflt Breeding. Gbo.

F. Fbbbman -------------356

XV. Taxonomy and Byolutlon. ByX--------- 869

v. snorter Articles and Disenssion : Nabonrs's Qraashoppers, Multiple Allele-
morphism, Tankage and Misleading Terminologies in G^etics. Profiassor W.
E. CfASTLB ------------- 888

THE BOBNOE FBE88

fJUKIASTEB. PA. 0ABBI80H, H. t.

HEW TOBK: 8UB-8TATI0H 84

Digitized by

Google

The American Naturalist

MSS intended for publication and boolcs, etc., intended for review sliould be
sent to the Editor of THE AMERICAN NATURALIST, Garrieon-on-Hadeon, New York.

Short articles containing sannaries of research woric bearing on the
probiems of organic evolntion are especially welcome, and will be given preference
in publication.

One hundrea reprints of contrlbntlons are supplied to authors free of charge.
Further reprints will be supplied at cost.

Subscriptions and advertisements should be sent to the publishers. The
subscription price Is four dollars a year. Foreign postage is fifty cents and
Canadian postage twenty-five cents additional. Tne charge fbr single copies Is
forty cents. The advertising rates are Four Dollars for a page.

Lancaster, Pa.

THE SCIENCE PRESS
NEW YORK : Sub-Station 84

Oarrisont N. Y.

Bnterad as Moond-olaH matter, April 2, 1908, at tha Poit Office at Lancaster, Pa., under the Aet of

Confreti of March 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Enfirland

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon.
7 or 8 volumes, please report, stating cash price, stat-
ing condition, binding and dates of volumes.

P. C. HARRIS.
Box 2244 Boston, Massachusetts

TO OOLOGISTS

and ENTOnOLOGISTS

W. F. H. ROSENBERG

Importer of Ezotio Zoological SpecimenB
57, Haverstock Hill, London, N. W., England

Begs to annonnoe the pnblicatioD of a new Price
List (Na 20) of Mammals, including over 400 speoies
from Tarious parts of the world.

These will be mailed free on application, as well
as any of the following lists:

Birdskins (over 5000 species) ; Birds' Eggs (over
1,100 species) ; Reptiles, Batraohians and Fishes
(over 400 species); Exotic Lepidoptera (over 8,000
species).

Largest stock in the world of specimens in all
branches of Zoology.

All Museums and Amateurs
should write for these Lists

All specimens sent on approral.
Please state which lists are required and give
name of this periodical.

Photographic Films Developed

5 cents per Spool, Prints 3 cents up, Post

Cards 5 cents.

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU
89 West Bridge Street Osvrego, N. Y.

The University of Chicago

Offers iutnicUondoru^ the Ssm-
mer Quarter on the amie bus ai
during the other quaxten of the
academic Tear.
The nndoqnaduate coOegcs. the

i graduate achooU, and the proie^
siona] schools provide cooises in
ArtB, Litmratur; Scimmcm,
Cotnmmrcm and AdmmUtm'

tion. Law, Mmdiein^

Hon, and Divinity, Xnstnictioa
b given by regular members of the
Uoiveraity staff which is augmented
in the summer by appointment of
professors and instructors froB
other institutions.

Mitchell Tower

Snmmw QMitor. 1914

1st Term Jan« 1 S-JoIf 22

2d Term Julr 23~Avs.2S

Detailed aunouncexnents will be
sent upon applkatioii.

The Unhrerstty of CliieafO
Chicago. DibMia

Digitized by

Google

Digitized by

Google

322 TEE AMERICAN NATURALIST [Vol. XLVIII

distinct species. Arbitrary rules for species making, de-
signed to restrict the activities of the more vigorous
** splitters '' have been indeed laid down by experienced
and conservative systematists. The final test, however,
so far as any exists, is acknowledged to be whether a
group breeds approximately true to its kind and is ap-
proximately sterile with other closely related stock, and
yet in how few cases have both or either of these criteria
been actually applied by the describer of species !

As a matter of fact no stock that has been bred on a
vast scale, so far as I am aware, breeds absolutely true to
specific characters. In Morgan's Drosophila^ and De
Vries's (Enothera, numerous mutants appear, probably
through the absence of certain chemical elements, or by
unusual combinations of elements, in the chromatin of the
germ plasm. That this phenomenon has not been shown
for many other species is due, in all probability, to lack
of close attention to all the individuals in a huge proces-
sion of stock in the process of breeding. Any insect bred
as extensively as Drosophila ampelophila, the pomace fly,
has been would probably show as many mutants; some
would show more. C olios eury theme, the ** orange sul-
phur '' or alfalfa butterfly, is such an example. Though
this butterfly can not be bred on a scale comparable with
Drosophila, every thousand individuals yield many dis-
continuous variations : red eyes instead of green, tongue
uncoiled instead of wound in close flat spiral when at rest,
one antenna shorter than the other, the absence of certain
spots from the wings, gynandromorphism, caterpillars
with two longitudinal rows of large black dorso-lateral
spots or white dorso-lateral stripes upon a dorsal surface
usually unmarked, caterpillars with one proleg less upon
one side than the other. This is a partial list of points at
which the descendants of three females of Colias eury-
theme failed in a single summer to breed true to the char-
acteristics of the species, though bred under uniform
normal conditions. The fact that these discontinuous

1 Science, N. S., Vol. XXXIII, Nos. 847, 849, pp. 496-499, 534-537, 1911.

Digitized by

Google

No. 570] SPECIES-BUILDING 323

variations appear under uniform external conditions
leads one to be very skeptical toward most of the past
experimental work supposed to show the effects of the
environment upon insects in modifying the germ cells.
Any one wishing to try an experiment on the production
of variations by the influence of the environment, or upon
the inheritance of acquired characteristics, should deny
himself absolutely this privilege until he shall have bred
under normal conditions at least a thousand individuals
of the stock that he will subsequently employ.

That species necessarily breed true to the specific char-
acters ascribed to them by their inventors is an unveri-
fied dogma. At best the reporter picks out stray individ-
uals here and there from a vast procession of which he
can only see glimpses, and, trusting to the credulity of the
public in the established ideas about these matters, he
creates upon paper a new species. Doubtless the unit
characters of ^* specific ^^ grade in the stock of some spe-
cies are more generally constant or homozygous than
those of certain others, but it is reasonable to suppose
that, owing to dominance the heterozygous^ condition re-
garding certain characters is frequently masked and un-
noticed in apparently pure strains of wild stock. If the
heterozygote respecting a certain character be compara-
tively rare, or if it be a heterozygote based on several
interacting factors, like redness in the kernel of Nilsson-
Ehle's wheat,^ it may cross again and again with
the homozygous dominant, or with another heterozygote
of similar nature to itself, without the appearance in the
population of the recessive. That specific and varietal
characters do exist in heterozygous condition in wild
stock of * * pure ' ' species, unmasked by dominance and
easily detected, I have found to be the case in Colias at
several points. The color pattern as a whole apparently
fluctuates in variation, but these variations in detail are

2 The mixed Mendelian condition, I>(i2), producing germ cells D and B
in equal numbers.

8 Act. Univers. Lund, 1909.

Digitized by

Google

324 THE AMEBIC AN NATURALIST [VoL.XLVHI

strictly a matter of inheritance. Its *' fluctuation ** is not
due to a difference in environmental conditions surround-
ing different individuals, but evidently to the condition of
the germ plasm. The parents of any brood may be
heterozygous or homozygous for the determiners of color
pattern. If they come from a strain homozygous in this
respect and are alike in appearance, the offspring will
resemble the parents closely and show a narrow range of
variation, but if unlike and derived each from imlike
parents, a wide range of inherited '^fluctuation" occurs.
Such is often the case in the inheritance of a melanic
tendency so often attributed to the action of the environ-
ment, and of spots used in the diagnosis of species as, for
example, the conspicuous spot in the middle of the imder
side of the hind wing. This is commonly double in Colias
philodice and C. eurytheme, consisting of a chief and an
accessory spot, single in C. palceno, an arctic circumpolar
species, but it varies enormously. In eurytheme and
philodice the accessory spot may be absent ; in palceno, in
rare cases, it may be present. I have bred large families
of C. eurytheme in which both the chief and accessory
spots were, like those of the parents, almost uniformly
large and nearly equal in size. In other families, from
parents in which the accessory spot is nearly or quite
lacking, the offspring show a similar reduction. In
C. philodice I have found it possible by selection to estab-
lish a race devoid of the row of submarginal red-brown
spots of the under side of the wings. Thus, by selection,
strains, nearly or perhaps quite homozygous for definite
points of color pattern, may be established, derived from
a population which in the main is in an extremely hetero-
zygous condition. Yet species are named and distin-
guished on the basis of these features.

Another example of heterozygous condition of a char-
acter within a wild species is the white pigment in the
ground color of the '' albino " female of Colias, both in
the yellow species, philodice and the orange species,
eurytheme. The white female is regularly heterozygous

Digitized by

Google

No. 570] SPECIES-BUILDING 326

for this sex-limited character. Her daughters are white
or colored (yellow or orange, as the case may be) in equal
numbers. Still another interesting heterozygous feature,
though not of *^ specific '' grade, was seen last summer in
a pure strain of Colias eurytheme. A female appeared
that could not upon stimulation coil up her tongue.
Mated with a normal male, this abnormality was in-
herited in various degrees by half her offspring (37
uncoiled and 28 coiled). One of her daughters, abnormal
in this respect and mated with a normal of a different
strain, transmitted the abnormality to about 16 per cent.
of her offspring (29:151), showing that the possessor of
this abnormality is regularly heterozygous in respect
to it.

Whether (Enothera lamarckiana is or is not a complex
hybrid produced from two American species, is it not
certain that, like other wild and cultivated stock, it does
possess characters for which it is heterozygous, and that
the watcher for mutants frequently seizes upon rare com-
binations of recessive features as a part of his elementary
species f

But to breed true is only a secondary criterion of spe-
cies. Inbred strains of domestic animals and plants do
that to a certain degree. Varieties and races to a certain
extent may do the same. The real criterion (and the one
least often practically used by the systematist) is fertility
within the group and sterility with other closely related
groups. Here dogma holds sway among writers on or-
ganic evolution as well as among systematists, for we are
told by those who have been accustomed since childhood
to the idea of the objective reality of species that hybridi-
zation of species, that is, genuine species in good and
regular standing before the scientific public, has played
very little part in the origin of new species. This atti-
tude was entirely logical in view of the accepted ultimate
definition of a species. If the individuals of one species
are actually sterile with members of another, hybridiza-

Digitized by

Google

326 THE AMERICAN NATURALIST [Vol. XLVIII

tion of species can not play an important part in the manu-
facture of new wild strains. But only in comparatively
rare instances have attempts been made experimentally
to mate Linnaean species. The dogma of the objective
reality and uniform value of the species unit has diverted
us from seriously attacking this problem. Just as in the
nineteenth century the fixed idea of the immutability of
species blocked the progress of the doctrine of evolution,
so this dogma now stands in our way, and obstructs the
possibility of vision. We need now fully to recognize the
fact, which most biologists are ready to admit, that the
term species is applied to most heterogeneous groups of
individuals, groups of every conceivable size, based on
differences that are most diverse in number and impor-
tance, often separated from allied groups entirely by the
arbitrary judgment of the describer, and depending ulti-
mately upon his personal temperament. These groups,
as already stated, have been tested in comparatively few
instances by the only reputable criterion that can be
applied in the separation of closely allied groups, that of
sterility or fertility inter se.

To one who tries to divest himself of the accepted ideas
regarding species and is on the watch for evidence of
hybridization among imlike strains that we are accus-
tomed to call species, new cases of such hybridization
frequently come to light. Especially is this true among
the insects. In regions where the faunal areas of two
*^ good " species overlap or are contiguous, such crossing
not infrequently occurs.

A most interesting case is that of the four species of
the coccinellid beetle Adalia that occur in the same region
in Colorado, as worked out by Palmer.* These four forms
with clean-cut differences in color and color pattern had
been named and described by different authors as dis-
tinct species, yet three of them were found to be inter-
breeding with complete fertility but still respectively
maintaining their identity, forming a regular Mendelian

, ^'Annals Entom. Soc. America, IV,' 3, September, 1911.

Digitized by

Google

No. 570] SPECIES-BUILDING 327

epistatic series : a red-brown spotless form, melanopleura,
dominant at one end of the series, then annectans, a red-
brown, spotted type, and finally the recessive, melanic,
red-spotted humeralis with a color pattern different from
that of annectans or of Coloradensis, another red-brown,
spotted type of that locality. *' But ^^ says the upholder
of the present idea of species, '^ here we have a single
polymorphic species, not three or four different species.
The breeding experiments show that the describers of
these forms were wrong in ascribing systematic rank to
mere color varieties.'' It goes, of course, almost without
saying that the makers of these species did not before
naming their beetles, breed them to determine whether
they would breed true to type and were infertile
inter se. Indeed, in how few cases has this been done!
Even the larval stages of most known beetles are imper-
fectly unknown, much less the possible genetic relation-
ship of one type to another, as determined by breeding
them to maturity. BlaisdelP describes the case of two
Californian Coccinellidse which are found in winter in
small groups under the bark of eucalyptus trees. '* Usu-
ally there was one Olla plagiata with each of the groups
[of 0. abdominalis], irrespective of whether they were
made up of two or more individuals." The same author,
by selection of specimens of abdominalis representing
different types of color pattern, describes its range of
variation, but adds that his studies throw no light on the
relationship of the two species. Had he bred certain in-
dividuals of 0. abdominalis together, it is not at all un-
likely, in view of his observation of the regular occur-
rence of a few plagiata in every group of abdominalis^
that the former interbreeds with the latter and may be a
simple recessive in respect to it. Miss Palmer's work
on the allied Adalia certainly suggests this as a possi-
bility.

Another remarkable case is that of the nine true-breed-
ing species of grouse-locust, Paratettix, recently de-

^Entom, News, Vol. 24, No. 9, November, 1913.

Digitized by

Google

328 THE AMERICAN NATURALIST [Vol. XLVni

scribed by Nabours.® These nine color types, or species,
freely interbreed. The color pattern of the resulting Fi
hybrid in each case is a mosaic combination of those of
the two parents. The latter in subsequent inbreeding
may be extracted intact, each having been transmitted as
a distinct unit, without dominance.

In Lepidoptera, an order in which polymorphism is
notoriously common, hybridization between species has
been frequently observed. Standfuss^ devotes eight
octavo pages of his excellent *' Handbuch '' simply to the
enumeration of examples of such hybridization between
palaearctic species of moths and butterflies, and acknowl-
edges that he mentions only a fragment of all such cases
on record or preserved in collections* This list would be
greatly extended if American species were included.
Seven different hybrid combinations within the genus
Colias in the palsearctic region have been noted by Stand-
fuss.

Colias philodice, the clouded sulphur or clover butterfly
of the eastern and central United States, readily crosses
with C. eurytheme, the orange sulphur or alfalfa butterfly
of the western and central states. The territory of philo-
dice, according to Scudder extends like a wedge westward
from the Atlantic into the faunal area of eurytheme.
Overlapping thus occurs in the Mississippi Valley, though
philodice does not extend as far southward as the Gulf
States, Texas, Louisiana and Mississippi, in which
ei^rytheme is found.

These two species are fairly sharply distinguished by
the difference in the ground color, which in eurytheme is
orange, in philodice sulphur yellow. The middle spot of
the upper side of the hind wing is brilliant orange in
eurytheme, pale orange or yellow in philodice. The dark
border of the hind wing of the female is wider in eury-
theme than in philodice and broken with a row of large
yellow spots.

6 Journal of Genetics, Vol. 3, No. 3, February, 1914.

7** Handbuch d. palUarktischen Gross-Schmetterlinge/ ' 1896, p. 51-53.

Digitized by

Google

No. 570] SPECIES-BUILDING 329

It has long been known that these two species hybridize
in the Mississippi Valley, where both occur. By extended
experiments during the past summer and previous au-
tumn with eurytheme stock sent to me from Arizona
through the kindness of Messrs. V. L. Wildermuth and
E. N. Wilson and with philodice from New Hampshire, I
have found that the two species mate together readily, and
produce vigorous offspring. The species-hybrid males
were then mated with eurytheme females, and more than
half of the pairs (viz., four out of seven) were fertile.
Mated together, however, the species-hybrids showed
much sterility. Out of ten such matings, nine were in-
fertile. From the tenth pair, nineteen adult butterflies
were produced.

Orange in this cross is distinctly dominant over no
orange, or yellow, but the color of the heterozygote is a
pale orange overlying yellow, and is by no means as bril-
liant as the almost fiery orange of the large, summer
seasonal variety, the typical ** eurytheme.^* In broods
emerging the last week in August and the first three
weeks of September, when intense color may be expected,
the heterozygote is pale orange, corresponding approxi-
mately to the variety known as keewadin, whereas those
raised in the greenhouse and emerging early in December,
resemble the small orange-yellow winter type known asi
ariadne. Keewaydin, according to Wright,® occurs at all
seasons in California, though probably more abundantly
in spring and autunm. Hence he regards this as the
typical variety, rather than '^eurytheme/' It is inter-
mediate, however, in size and intensity of color.

In general, therefore, there is an incomplete dominance
of orange, the color of the heterozygote corresponding
either to that of the intermediate or to that of the winter,
seasonal variety of eurytheme, depending upon the time of
the year when, and the environmental condition under
which, the cross is made. The wide, spotted margin of the
hind wing in the female eurytheme, moreover, when pres-

8'*Butterflie8 of the West Coast of the United States," p. 119.

Digitized by

Google

330 THE AMERICAN NATURALIST [Vol. XLVIII

ent in marked degree, is dominant over the narrower
margin in philodice. This dominance of the orange mani-
fests itself quite as distinctly if the albino female of
eurytheme, instead of the orange female, is bred to the
yellow philodice male. The daughters of such a family
in one case (0, 1913) were 36 white, 35 orange; the sons,
numbering 72, were, of course, all orange. The white
species-hybrid (FJ is identical in appearance with the
albino eurytheme, the female color pattern of the latter
(wide marginal bands) being dominant, and the orange
middle spot both in pure bred albino eurytheme and in
the albino hybrid being usually paler than in their orange
sisters.

The second hybrid generation inbred (Fg) shows a well
marked segregation of the sulphur-yellow color of phUo-
dice, as a simple Mendelian recessive. Three out of the
sixteen colored (non-albino) individuals of the brood ob-
tained in December, 1913, are definite recessives of clear
sulphur yellow, with pale yellow middle spots on the hind
wing. The most highly colored individuals are four that
correspond in hue to pale examples of the light orange-
yellow winter variety, ariadne. There is no return, at
least in this winter brood (enclosed in a greenhouse in
New Hampshire in December), to the brilliant orange of
the grandparent al eurytheme stock. Nor do they even
return to the suffused light orange (intermediate) tint
of the heterozygous father (keewadin type), for the
ground color of all individuals of this brood (Fj) is
yellow, either flushed or spotted, except in three indi-
viduals, with orange.

An interesting case of probable hybridization in the
allied genus M eg ano stoma, or dog's head butterfly, is re-
corded by Wright® between the Califomian M. eurydice
and M. ccrsonia, common throughout the southern states.
The two species are remarkably different in color and
have different food plants. The male of eurydice differs
from that of ccesonia in having a violet luster and lackmg

9Loc, cit.f p. 116.

Digitized by

Google

No. 570] SPECIES-BUILDING 331

the black border upon the hind wings possessed by
ccesonia; in the female, etirydice is clear yellow with no
dark border, while in cccsonia the female has a wide
border similar to that of its male, though less well marked
on the hind wings. The probable hybrid called amorphce
is a female, intermediate in color between the typical
ccesonia and eurydice. That is, the border of ccesonia
crossed with no border (if my interpretation is correct)
is incompletely dominant. Wright says:

At one time I was of the opinion that AmorphcB was a hybrid between
Eurytheme and Caesonia . . . but of late years, as no male Amorphce
is known, I have concluded that Amorphce is simply a dimorphic female
[of eurydice].

Possibly it is both, an example of dimorphism produced
either by immediate hybridization, or by a mutation re-
sulting from some previous hybridization. That a male
appears to be lacking in this case would not be an argu-
ment against the possibility of hybridization, for by such
crossing the sex ratio is frequently upset, the product
being of one sex only. But it appears to be possible that
the male of this cross is that described as M. bernardino,
a variety of eurydice found in the mountains of the same
region where amorphce also occurs. It is an interesting
combination of the male coloration of both species, having
the violet hue of eurydice that is lacking in ccesonia and
having the dark border of the hind wings of ccosonia lack-
ing in eurydice. Its female is described as being smaller
than that of eurydice, but otherwise practically identical
with it. This case, as Wright has suggested, is a most
inviting subject for further study, and, judging by what
he says of the sexual instincts of the eurydice male — '* a
wooer . . . energetic and persistent, not hesitating to
ignore all rules of propriety, of species and of genera ''
— ^not difficult of experimental management.

The genus Basilarchia, the admiral butterflies, is well
known for the hybridization of its very unlike species,
B. arthemis the '* banded purple '' of the northern states,

Digitized by

Google

332 THE AMERICAN NATURALIST [VoL.XLVIII

B. astyanax the '* red-spotted purple '' of the southeast-
ern states. The hybrid species, B. proserpina, occurs
in a zone in which their two faunal areas overlap.
In this same group is the common *' viceroy '' B. ar chip-
pus, the range of which roughly covers that of both the
other species and extends further westward, touching the
Pacific coast in Washington ( Scudder ) . The experiments
of Edwards, and especially of Field, have shown that
these three well-differentiated pure species occupying
contiguous, or in respect to archippus overlapping,
territory are in some cases at least mutually fertile.
B. arthemis and astyanax regularly interbreed in the
narrow zone where proserpina occurs. Proserpina, the
hybrid, usually shows the general dominance of the
astyanax characters (lack of white band).

From eggs laid by a wild female proserpina Edwards^^
secured three arthemis, one proserpina. Field^^ raised
from a similar lot of eggs nine proserpina, seven arthe-
mis. Presumably in each case the male parent was the
recessive arthemis, and hence equal numbers of the two
types would be expected. Field has also succeeded in
crossing a 2 astyanax with a <? arthemis, and a 2 viceroy,
archippus, with a c? arthemis, the latter pair producing
nine males intermediate in color. Specimens of an ap-
parent hybrid, intermediate in color between astyanax
and archippus, have also occasionally been captured.

The complete overlapping of the faunal area of archip-
pus upon those of the two other species indicates that,
though crossing sometimes occurs, the resulting hybrids
are probably usually sterile, though this matter has not
yet been thoroughly investigated. Proserpina, however,
is a fertile and extraordinarily variable hybrid. In view
of its great variability it appears, by the way, not impos-
sible that archippus, the red-brown *' mimic '' of the mon-
arch, Anosia plexippus, may have arisen as a mutation
from the hybrid proserpina, though the wide-spread

10 Canadian Entomologist, Vol. IX, 1877.

11 Psyche, Vol. XVII, No. 3, 1910.

Digitized by

Google

No. 570] SPECIES-BUILDING 333

range of archippus at present and our ignorance of the
state of the Basilarchia stock at the time of the origin of
the ''mimic'' make any such specific historical guess
hazardous. It may, however, some time be possible by ex-
perimental breeding to extract from this red-spotted pur-
ple hybrid a red-brown type similar to archippus. If the
Basilarchia stock were as easily bred as Drosophila, one
might be very confident of accomplishing this. In any
event, the theory of the origin of mimicry by natural
selection is, in the opinion of the writer, entirely super-
fluous, though this celebrated monarch-viceroy case
should be exhaustively studied by experimental methods,
to determine whether natural selection now operates in
any degree in the matter.

Examples of clusters of interbreeding types may be
drawn in large numbers from various classes of animals
and plants. Bateson^^ has recently called attention to the
interesting case of the two American flickers described
by Allen,^^ the eastern Colaptes auratus and the western
and Mexican C. cafer, which hybridize in the zone in
which their faunal areas overlap, the American grackles,
the golden-winged and blue-winged warblers and their
hybrids, Lawrence's and Brewster's warblers, and others.

In reference to the common purple grackle, which
Chapman" regards as a hybrid between the Florida
grackle and the bronzed grackle, Kidgeway^^ says:

My own opinion in the matter exactly coincides with Mr. Chapman's
but since so many forms now ranked as sub-species are similarly in-
volved I prefer, at present, to leave the matter in abeyance.

This signiflcant statement from a master of ornithologi-
cal taxonomy indicates that hybridization among Ameri-
can birds is a promising subject for investigation.

Of the occasional mutual fertility of unlike strains dif-
ferent enough to be classed as unquestionable species,

12 "Problems of Genetics,'' 1913, Chap. VII.

i^BuU. American Mus, Nat Hist., Vol. IV, 1892.

" Ibid.

w Birds of North and Middle America,'' Part 2, p. 219, 1902.

Digitized by

Google

334 THE AMERICAN NATURALIST [VoL.XLYin

there also can be no doubt. '* We can only escape the
conclusion that some species are fully fertile when
crossed," wrote Darwin,^® ''by determining to designate
as varieties all the forms that are quite fertile/' and he
added that some plants exposed to unnatural conditions
are so modified '' that they are much more fertile when
crossed by a distinct species than when fertilized by their
own pollen. '*

The rareness of these crosses between unlike strains or
species and the partial sterility of the offspring are not
obstacles in the way of regarding occasional hybridiza-
tion as one of the chief sources of mutation and hence
eventually of new species, for, as my preliminary experi-
ments in hybridizing species of C alias have already
shown, there may exist within a strain of species-hybrids
certain individuals that ^re fertile, though the most of
their brothers and sisters, mated, respectively, in a similar
way, are sterile. Nature probably makes more random
experiments in hybridization than we imagine ; many fail;
some succeed; and in especially favorable stock Uke
Colias, judging from the numbers of closely aUied but
different types (species) occurring in the same localities
in western Asia or in northwestern United States and
British America, probably many succeed.

In seeking to determine how mutation, whether the re-
sult of hybridization or of possible climatic influences,
acts in the production of new species, it is possible from
cases already at hand to suggest possible steps in the evo-
lution of distinct, mutually infertile, types from one com-
paratively simple polymorphic species.

The well-known dimorphic European currant moth,
Abraxas grossulariata, in which the light-colored (reces-
sive) variety, lacticolor, is found in nature only in the
female sex, will serve as an example of an elementary
condition. Lacticolor males, as Doncaster*^ has shown,

16 *' Animals and Plants under Domestication/' Vol. II, Oiap. 19, p. 179.
IT* 'Report of the Evolution Committee, '* 4, 1908.

Digitized by

Google

No. 570] SPECIES-BUILDING 335

may readily be bred. When one of these males is mated
with a lacticolor female, there is produced in captivity a
pure lacticolor strain. If lacticolor males and females
should be segregated and allowed to breed together until
they have become as abundant as the typical form, this
case would then resemble that of the Colorado lady beetles
of the genus Adalia, described above, in that it would con-
sist of different types maintaining their identity while
freely interbreeding with complete fertility. The
Abraxas complex differs from the Adalia species-cluster,
however, in the occurrence of sex-linkage in the inheri-
tance of the lacticolor variety, whereas in Adalia the
factors for the different color patterns apparently are
distributed in the gametogenesis of a heterozygous indi-
vidual without sex-linkage, freely and at random.

A more advanced stage in evolution is that represented
by the Basilarchia species-cluster, in which partial steril-
ity between the viceroy and the two purple species, over
the faunal areas of which its own overlaps, and the differ-
ence in geographical distribution between the banded
purple and red-spotted purple, keep the three elements
apart.

By easy stages we may in imagination pass on to
groups composed of closely allied species which sterility
and local segregation completely separate from one
another, groups that probably have arisen from a poly-
morphic species that has broken up into its constituent
parts, and thus given rise to new elementary species.

The dimorphism of Colias differs from that of Abraxas
in that the color of the rarer type of female can not be
transferred in the ordinary course of breeding, without
further mutation, to the male. It is a sex-limited char-
acter, like the female color pattern in Colias, (i. e., a wide
dark border broken with spots) and not sex-linked like
the variety lacticolor of Abraxas.

The white female of Colias is regularly heterozygous
for color. She produces as many white daughters as

Digitized by

Google

336 THE AMERICAN NATURALIST [Vol

yellow, or orange, as the case may be. Evidently, i
to extract a pure white race from C. philodice or (
theme, it will be necessary by a mutation to obtaii
homozygous white female, and then by a further m
a homozygous white male. White males are kn
nature as rare aberrations, but, whether they an
zygous or heterozygous for color, it is impossible
Among the two thousand offspring of heterozygou
females of philodice and eurytheme that I ha^
since 1908, there has been not a single white mal
sons of a white female, though some are capable o
mitting the white, are always yellow or orange,
lately, however, raised a large brood in which
females were white. This was a '* back cross '' 1
a white female of the orange eurytheme and a m;
cies-hybrid (son of a white mother). Precisely
matings, however, gave both white and colored
offspring in equal numbers; hence in the produ(
this brood there was probably a mutation. Fro
stock as this the extraction of a pure white ra(
Colias at some time may possibly be accomplishec
In this connection it is interesting to note that ^
the testimony of a good observer, the late Mr.
Wright," who made the study of Californian bu
his life work, to the effect that the white variety oj
eurytheme ''is now quite common, though twe
years ago it was a great rarity, and it was accoi
feat to secure one of them, and if the present rat
crease of the blond form shall go on, in a few h
years the normal orange-colored female will be
and unknown." If this is a fact, and not an illus
to a general increase in the population of eur
owing to an increase in the cultivation of the foo<
alfalfa, in that region, it may be the result of \
mutations, whereby homozygous white females m.i
been introduced into the population. It will be o

iTLoc. cit., p. 117.

Digitized by

Google

No. 570] SPECIES-BUILDING 337

est to determine whether such true-breeding white
females actually occur in California.

Evolution in Colias is usually regarded, on the other
hand, as tending towards suppression of the white stock
rather than its further extension, inasmuch as Pieris and
other allied genera are white. It seems to be a reason-
able hypothesis that, by progressive mutations in Colias
affecting first the male then the female,^® white has be-
come yellow; yellow, orange; orange, red, or a fiery
orange ;^* or yellow may be transmuted into black, as in
an aberration of the male in C. philodice. By retrogres-
sive or degressive mutations, accordingly, we may hope
to isolate from C. philodice or C. eurytheme a pure white
race.

Summary and Conclusions

The erroneous idea that Linnsean species are homo-
geneous, well-defined groups of equal importance has
done much to retard progress in the experimental study
of evolution. The limits of a species are often arbitrary,
depending ultimately upon the temperament of the des-
criber, and frequently based upon ignorance of the near-
est allies of the individuals described, living in other
parts of the world.

The most definite criteria of species, viz., that '' spe-
cific *' characters are constant, and that hybrids of
Linnaean species are infertile inter se, are only approxi-
mately correct. Characteristics of species sometimes
occur in heterozygous condition. Hybrids of Linnsean
species, as has long been known, are often fertile. These
matters, owing to traditional, unwarranted respect for
described species, have received comparatively little in-
vestigation.

Examples of hybridization in Adalia, Colias, Meganos-
toma, Basilarchia and Paratettix among insects, in Co-
laptes, Quiscalus, and Helminthophila among birds are
cited.

18 In C. dimera of South America, for example, the female is yellow, but
in the male the /ore wings are orange.
1* As in the Asiatic eogene.

Digitized by

Google

338 TRE AMERICAN NATURALIST [Vol. XLVIII

Occasional fertile crossing of unlike strains that rarely
interbreed is a probable source of mutations and new
types.

A suggestion is made that a comparatively simple poly-
morphic species (like Abraxas grossulariata) may break
up into a cluster of mutually fertile elementary species
{e. g.y Adalia in Colorado). Further differentiation, in-
volving partial sterility, may be illustrated by the Basi-
larchia species-cluster. This may be followed by the
establishment, and isolation through complete sterility,
of distinct types, or species in the strict sense of the term.

Evolution of color in the yellow and orange butterflies
of the genus Colias involves white, which exists to-day in
heterozygous condition in certain females. If the an-
cestors of Colias were white, as in Pierids generally, we
have only to imagine a mutation in the male-producing
germ cells of the original white females, by virtue of
which white pigment was replaced by, or transmuted into,
yellow. This would make all the males yellow, leaving
all the females white, which is true of certain arctic
species to-day.

A similar mutation affecting the germ cells of these
white females, but introducing the factor for yellow into
only half of them, would produce the heterozygous condi-
tion found in C. philodice and C. eurytheme. Pure yel-
low strains may readily be bred from such mixed stock,
and hence, probably, it has come about that four fifths or
nine tenths of the females of C. philodice in eastern
United States are pure yellow.

Progressive mutations from yellow to orange and fiery
orange, affecting first the male, then the female, have
probably occurred in Colias in many part of the world,
especially in warmer climates. Climatic conditions deter-
mine the amount of orange pigment in the cross between
the orange eurytheme and the yellow philodice. This
hybrid is larger and contains more orange when raised in
summer than when bred in late fall and winter. C. philo-
dice in this cross is a Mendelian recessive.

Digitized by

Google

HEREDITY OF BRISTLES IN THE COMMON
GREENBOTTLE FLY, LUCILIA SERICATA
MEIG. A STUDY OF FACTORS GOVERN-
ING DISTRIBUTION^

PHINEAS W. WHITING
BussET Institution

In a previous paper^ I have given data showing that
variation in the number of posterior dorso-central and
acrostichal bristles of the common greenbottle fly, Lucilia
sericata Meig., is determined by hereditary factors.
Since the publication of that paper further evidence,
bearing upon the nature of the hereditary factors in-
volved, has been obtained.

Two general conclusions from the work may be stated
as f oUows :

1. Reduction in bristles tends to affect the males more
than. the females, while additional bristles are found more
often in the females.

2. Distribution as well as number of bristles is heredi-
tary.

On account of very high mortality in these flies it has
been impossible to make selections as might seem desir-
able. The results, however, furnish considerable evi-
dence for the foregoing conclusions, and throw light, I
believe, on the nature of factors governing distribution,
such as spotting factors, for example.

Fig. 1 shows the mesonotum of Lucilia sericata with
chaetotaxy normal. The bristles considered in my work
are those lettered A, B, C, the post-acrostichals, and A\
B\ C, the post-dorso-centrals.

1 Prom the Entomological Laboratory of the Bussey Institution, Harvard
University, No. 77.

2 Whiting, P. W., "Observations on the Chaetotaxy of Calliphorin®, "
AnncUs of the Entomological Society of America, VI, 2.

339

Digitized by

Google

340 THE AMERICAN NATURALIST [VoL.XLVm

riQ. 1.

It is evident that these bristles form a group of twelve
in four rows of three each.

This arrangement is recorded as 3, 3, 3, 3, the separation into rows
being denoted by commas.

When one or two of the anterior bristles of a row are omitted, the row
is denoted by 2 or 1, respectively.

In order to denote the omission of the second or third bristle wh«i
those anterior to it are not omitted, the normal positions of the bristles
are recorded as a, b, c, from anterior to posterior. Thus a row lacking
the second bristle would be called ac.

Addition of a supernumerary bristle ijito a row is denoted by ! in-
serted in the proper position between or in front of the letters denoting
the normal bristles. Thus addition of a bristle in front of a row would
be expressed by calling the row !abc.

Insertion of a supernumerary bristle between the normal rows is
denoted by parentheses enclosing a, b, or c, acording to the position
of the bristle from anterior to posterior. Thus a definition as 3, (a),
3, 3, 3, would denote the addition of a bristle between the first left post-
dorso-central and the first left post-acrostichal.

Additional bristles are usually smaller than the normal, but range all
the way from microchaetae to the size of the normal macrocluetflB. A
small bristle is denoted by italics.

The progeny of a few wild females have been bred and
counted since my previous paper.^ These have been

Digitized by

Google

No. 570] FACTORS GOVERNING DISTRIBUTION

341

O rS

•^ art

1:4

■s §

§ .2

5 <=i

COO

ill

■b

1-4

^"a

S
f^

•b

iHO

1-4

O 1-1

—

•

•b

1-1

iO^

•b

1-«

3^^

COl

&^°

»oco

1-4

eoa

•b

CO o

1-4

•p

t^»o

•b

O lO

a
1

•b

l-H

'^^

"■s"

-o

•
JO

1^^

§3

^t^^^B

c<> o

.-:

'ej

«H

.3 to
I I

a «H

*::' "3 rt

o o

Digitized by

Google

342 THE AMERICAN NATURALIST [Vol/XLVHI

averaged with those recorded previously and the results
given in Table I.

From this table it appears that progeny of normal
mothers show a certain degree of variation in the direc-
tion both of loss and of acquisition of bristles ; progeny
of reduced mothers tend more toward reduction; and
progeny of mothers bearing additional bristles tend more
toward the addition of bristles. It is also evident that in

3 It is thought desirable to put on record a detailed account of these fam-
ilies as they furnish in themselves a few points of interest. This record is
given below with the exception of the progeny of 1913-^, discussed in a
later part of this paper.

1913-5, L. sericata ? = 3, ab, ab, 3, taken at Bussey Institution, May 6,
1913, gave

11(J(J = 3, 3, 3, 3.
1 c? = 3, 3, 3, 2^

3c?c? = 3, 2, 3, 3. 17 $9 = 3, 3, 3, 3.

1 cf=3, 3, ac, 3. 1 5 = 3,2,3,3.

1913-C, L. sericata ? = 3, 3, 2, 3, taken at Bussey Institution, May 6,
1913, gave

49c?(?=3, 3, 3, 3. 5705 = 3, 3, 3, 3.

1 ^ = 3, 3, abc, 3. 1 ?= -'abc, a /be, 3, 3.

1 ^ = 3, albc, albc, 3. 1 ? = 3, a/bc, 3, 3.

1 c? = 3, a.'bc, 3, 3.
1 c? = 3, !abc, 3, 3.
In this case I attribute the additional bristles to the combination of fac-
tors introduced by the male. An example of this sort in which a reduced
female produces offspring abnormal predominantly by addition is very un-
usual. There are, however, occasionally flies with extra bristles in reduced
strains, a fact which may be explained by recombinations of factors or by
mutation.

1913-F, L, sericata ? = 3, 3, 3, 3, taken at Bussey Institution, March 19,
1913, gave

24c?c? =3, 3, 3, 3. 19?? = 3, 3, 3, 3.

from a mating of these were produced

92cfc? = 3, 3, 3, 3. 89?? = 3, 3, 3, 3.

1 <J = 3, 3, albc, 3. 1 ? = 3, albc, 3, 3.

1913-1?, L. cwsar ? = 3, 2, 2, 3 (the chaetotaxy normal for this species),
taken at Bussey Institution, May 5, 1913 gave

55c?(? = 3, 2, 2, 3. 34?? = 3, 2, 2, 3

4c?c? = 3, 1, 2, 3. 1 $ = ac, 2, 2, 3.

1 cJ = 3, 2, b, 3.

2<?(?=3, 1, 1, 3.

1 c? = 3, 2, 1, 3.

1 cJ = 3, b, 2, 3.
The flies of this mating are not averaged with the others, as it is possible
that this species may be different in its variability from L, sericata. It is
noteworthy, however, that here also reduction favors the male more than
the female.

Digitized by

Google

No. 570] FACTORS GOVERNING DISTRIBUTION 343

general reduction tends considerably to favor the males,
while addition favors the females to a slight extent.

In my previous paper (p. 264) is given in detail a
record of the progeny of a female L. sericata (1912-c)
lacking both of the first and the right second post-acros-
tichal (3, 2, 1, 3). These were inbred to the third gen-
eration, in all cases brother being mated with sister in an
attempt to analyze the stock as thoroughly as possible
and to reduce heterozygosis of factors.* Here again, due

« Mr. Harold D. Fish has kindly furnished me the following note :

''The importance of mating sisters with brothers for a long series of
generations in the experiments aimed to detect Mendelizing units of inherit-
ance and analyze groups of them, quite generally seems to have been over-
looked. As first shown by Castle ('03), random mating of the individuals
of successive generations beyond F^ tends to produce in each generation a
population with the same per cent, of homozygosis and heterozygosis as is
present in the F, generation, i, ^., 25 per cent, of the individuals are homozy-
gous for. one factor of a given allelomorphic pair, 25 per cent, homozygous
for the other factor, and 50 per cent, heterozygous for both. Such a system
of random matings often has been confused with the more restricted system
of mating sisters with brothers.

**It is evident that if A and B are an allelomorphic pair the F, zygotes,
resulting from a mating of AA with BB, will be AA, 2AB and BB.
Further, if these are all females and are mated in all possible ways with the
same number and kinds of males, one sixteenth of the matings will be AA
with AA, and one sixteenth will be BB with BB, One eighth of the matings,
then, will be homozygous and produce only homozygous young, which, be-
cause of the restricted system of mating only sisters with brothers, will pro-
duce, in turn, only homozygous matings. The remaining matings, seven
eighths of the total, will produce various proportions of homozygous and
heterozygous offspring and matings. It is rather natural to assume that one
eighth of these matings will be homozygous and seven eighths heterozygous.
This would mean that the proportion of heterozygous matings between indi-
viduals of the Fh generation would be (7/8 )*»"*. Accordingly one would ex-
pect an automatic increase in homozygosis. The expectation is justified al-
though the figures are misleading.

"Dr. Baymond Pearl first published the figures exactly expressing the
per cent, of automatic increase in homozygosis for patred allelomorphs, under
the restriction of mating only sisters with brothers. This article appeared
in the January, 1914, number of the American Naturalist. It is a correc-
tion of his paper in the October, 1913, number of the same periodical, in
which he states in no uncertain terms that an automatic increase in homo-
zygosis in obligate bisexual forms is impossible. When I read the October
paper I was naturally much surprised, since, nearly a year before, during
conversation with Mr. Whiting, the increasing per cent, of homozygous
matings resulting from successive matings of sisters with brothers had been
discussed. Of course, the per cent, of individuals in any generation, which
are homozygous for one or the other of a pair of allelomorphs, is the same

Digitized by

Google

344 THE AMERICAN NATURALIST [Vol. XLVIH

to high mortality, selection as might have been desired
has been impossible.

A detailed account of this strain is given in Table II.
In recording any mating of this strain the letter c denot-
ing the entire strain, is followed by Fi, Fg, etc., denoting
the generation from which the mated flies have been
chosen. This symbol is then followed by a, b, or c, denot-
ing the first, second, or third mating, respectively, of the
generation indicated. Thus mating cFgb is the second
mating of the second inbred generation of strain 1912 — c.
This method of recording matings has been followed
throughout my work.

Several points of interest are to be noted in this strain
but it is thought best to present the remaining data on
reduced strains before proceeding to a discussion of this
matter.

Strict inbreeding has been followed in the strain re-
corded below. In no case have there been either cousin-
matings or outcrossings.

1913-^, L. sericata ? = 3, ac, ac, 3, taken at Bussey Institution, Forest
Hills, Mass., May 6, 1913, gave

96c?c?=3, 3, 3, 3. 129?? = 3, 3, 3, 3.

1^5 = 3, ac, ac, 3. 1 ?=3, ac, 3, 3.

2c?cf = 3, ac, 3, 3. 2?? = 3, 2, 3, 3.

2(?c? = 3, 3, ac, 3.
1 c? = 3, 2, ac, 3.
4c?c? = 3, 2, 3, 3.
3c?c? = 3, 3, 2,3.

as the per cent, of the allelomorphic factors which are homozygous in the
average individual of that generation. Because Br. Pearl in his October
paper referred frequently to the paper by Dr. E. M. East ( '12) on "Hetero-
zygosis in Evolution and Plant Breeding. *' I gave Dr. East my figure ex-
pressing the per cent, of homozygosis in successive generations resulting
from matings of sisters with brothers. Dr. Pearl's correction followed a
letter from Dr. East which pointed out the error of applying the mathe-
matics of random matings in each generation to a case where sisters always
had been mated with brothers. The percentages, as computed, were pub-
lished by Dr. Pearl for the following generations: Pi — 100 per cent., Fx — 0
per cent., F^ — 50 per cent., F3 — 50 per cent., F* — 62.5 per cent., F^ — 68.25
per cent., Fo — 75 per cent., F, — 79.687 per cent., Fg — 83.594 per cent, F, —
86.719 per cent., Fjo — 89.258 per cent. Previous to giving these figures to
Dr. East I computed the number of generations necessary to reduce heterozy-
gosis to less than one half of one per cent, and found this condition first
realized in the F^s generation, which is 99.553 per cent, homozygous. The
importance of these figures in work of this nature is quite obvious. * *

Digitized by

Google

No. 570] FACTORS GOVERNING DISTRIBUTION

346

ea
o

u
00

•<

•»

i-«

N-^

r-l

oooooo

lis

• ^

•b

O O M 1-1 M

•gs

1-1

»H

PQ a

?«

CC*^ CO

8 1

? -

»o

o oo ^'i** CO NO'dl

■b

O o c^ ^ o

©|CO

•b

oo oo

oo|©

o o CO ^

N 1-1

•^ "C

•^

•b

o© o o

^ Ol'*

o © ot^

o ojt^

•

•b

oo^ «

iH O 00

a
3

3^&

cs» t* t* r>.

"^ coio

e S 5

N 1-1 00 CO

»^ os'^

HA**

c^-

oo© o

t^ <^

•b

O O C^ 1-1

r-l

o
,a

©O N OS

O r-l

•b

OONO

1-«

CO o ,00

1 '^

s
S5

a-'ss

gss

r-l CO

•b

225il!g

SSf2

i\ i

a .

. fl w o8

d 1 a
" §11

•31 •

C8 el

x-a| :

c w o « o

llil
5 « 2|o

s a s a a

2 sm

flw'^'H

p^pi^pt.

PhP^

Digitized by

Google

346

THE AMERICAN NATURALIST i^

from^F.a={|-||«.33.33-

• 24

cfd' = 3, 3, 3, 3.

53??:

= 3, 3, 3, 3

fJJ=:3, ac, ac, 3.

2?$;

= 3, ac, ac.

c?cf = 3, ac, 3, 3.

52?'

= 3,3,ac, 3

d'cf = 3, 3, ac, 3.

1 ?:

= 3, ac, Abe

(}(J=z3, a6c, ac, 3.

1 ?:

= 3, ac, 2, 3

.^ = 3, 3, a6c, 3.

2??

= 3, 3, a6c,

(? = 3, 1, 1, 3.

1 ?:

= 3, a6c, 3,

c^c^ = 3, ac, 2, 3.

1 ?:

= 3, a6c, 2,

5 ^^ = 3, 2, ac, 3.

2??:

= 3, 2, 2, 3

cr = 3, 2, 2, 3.

1 ?:

= 3, 3, 2, 3

5cfdr = 3, 2, 3, 3.

1 ?:

= 3, 3, a//c

cfc? = 3, 3, 2, 3.

1 ?:

= 3, a!bc, 3

^ = 3, ace, 3, 3.

from -4F,a = c? and ? = 3,

ac, ac, 3. Pair segregated Ji

July 25.

c?c?

(?(?

35 = 3, 3, 3, 3.

2=3, 2, a

18 = 3, ac, ac, 3.

4 = 3, 2, 1

18 = 3, ac, 3, 3.

7 = 3,2, J

12 = 3, 3, ac, 3.

7 = 3, 3, i

2 = 3, ac, 2, 3.

0 = 3, ac,

from ^Fjft =

c? = 3, ac.

,2,3.

S=3, a6c, 2, 3.

Pair segregated July 12; larvae

July 25.

c?c?

(?(?

45 = 3, 3, 3, 3.

1 = 3,3, 2

9 = 3, ac, ac, 3.

1 = 3, a6c

9 = 3, ac, 3,3.

3 = 3, 3, I

11 = 3, 3, ac, 3.

3 = 3, a6c

. 8

0 = 3, ac, 2, 3.

0 = 3, b, 3

1 = 3, 2, ac, 3.

0 = 3, ac.

0 = 3, 2, 2, 3.

1 = 3, a6c

1 = 3,2,3,3.

^from^F.a={|;3,ac,a,:

5, from AT^. Pair segregated
3, larvsB August 20.

c?c?

(?(?

32 = 3,3,3,3.

7 = 3,2,3

11=3, ac, ac, 3.

5 = 3,3, 2

20 = 3, ac, 3, 3.

0 = 3, ac,

13 = 3, 3, ac, 3.

1 = 3, ac.

6 = 3, ac, 2, 3.

0 = 3, ac,

7 = 3, 2, ac, 3.

1=2,2,2

4 = 3,2,2,3.

1=3, 3, a

from ^F,6 = c? and 9 = 3, 2,

ac, 3, from A¥^,

Pair segreg

13; larv8B August 20.

c?cf

18 = 3, 3,3,3.

2 = 3, 2, a

3 = 3, ac, ac, 3.

2 = 3, 2, 2

5=3, ac, 3, 3.

7 = 3,2,3;

2 = 3, 3, ac, 3.

4 = 3,3,2,

13 3 = 3, ac, 2, 3.

from ATiC=:f} and 2 = 3, 2, 2, 3, from ^F^. Pair segreg
13; larva? August 25.

Digitized by

Google

-5^^^>

No. 570] FACTORS GOVERNING DISTRIBUTION

347

6 15 = 3,3,3,3. 2 15 = 3,2,3,3.

10 7 = 3, ac, ac, 3. 3 8 =: 3, 3, 2, 3.

11 7 = 3, ac, 3, 3. 0 l = ac, 1, 2, ac.
10 12 = 3, 3, ac, 3. 1 0 = 3, 2, ace, 3.

8 3 = 3, ac, 2, 3. 2 1=3, ace, ac, 3.

7 7 = 3, 2, ac, 3. 1 0 = 3, ace, 3, 3.
5 4 = 3,2,2,3.

from ATffil = ^ and 2 = 3, ac, ac, 3, from A¥^, Pair segregated August

13; male died August 18 and another with same chaetotaxy put in; larvse
September 1.

(fcf 22 (f(f 22

0 3 = 3,3,3,3. 1 1 = 3, ac, 2, 3.

1 2 = 3, ac, 3, 3. , 1 0 = 3, 3, aabc, 3.
1 1 = 3, 3, ac, 3.

The record of 1913- A, recorded in tabular form is given in Table III.

We are now in a position to consider the nature of re-
duction of bristles in Lncilia sericata.

It is evident from Table t (record of first generation
flies), that reduction and addition of bristles are both
hereditary. It is further evident from Table III, (inbred
strain), that reduction yields readily to selection. This
eflFect may be expressed by making the number of bristles
lost the numerator of a fraction of which the denominator
is the number of bristles normal. We then have a ratio
for each generation of 1913 — A as follows :

18
2892

= 0.006 =*= .010,

^' ' 1788 " ^'^^^ ^ -^^^

435
^* ■ 4692 = ^-^^ * •^^'

^4 • TT?k = 0.104 =«= .003.

It may be readily seen by glancing at these figures that
selection has a very rapid eflfect. It also appears that as
we pass from Fj to F4 the effect of selection gradually
diminishes. This may be expressed by dividing the above
decimals for each generation by that of the preceding
generation.

0.055
0.006

= 9.16,

1}
F2

0.093
0.055

= 1.69,

F»

0.104
0.093

= 1.11.

The reason for this decrease in the effect of selection in
the later generations is that as the selection advances the
majority of the flies become reduced in two bristles only.

Digitized by

Google

348

THE AMERICAN NATURALIST [VoL.XLVm

o
o

1—1

SI'S

«5

I
&

1^1

S- 1

ll§3

> 1

o
o

^ o

O 00

0> CO

2 S

a ° "b

-a"

CO*^ ^CO . *

jj n ." ?i ■ CO $i I

§ SSJcocoSoScocoSJcoco^SJO^S

-S ^ n II J y II S

•d a^x

i-i^oiSl^ol-boSSg^ag

Digitized by

Google

No. 570] FACT0B8 GOVEBNING DISTRIBUTION 349

Barely does a fly occur lacking more than two. In the few
cases in which three or more bristles are lacking, the
absence of the third acrostichals or of the dorso-centrals
is as frequent as the absence of first and second acros-
tichals. Why this should be is difficult to understand, as
it would be expected that both first and both second post-
acrostichals might frequently be lacking in the same fly,
especially as flies asymmetrical for the loss of these
bristles are common.

A further point of interest lies in the fact that not only
is number of bristles a hereditary matter, but their dis-
tribution is also hereditary. Thus from Table I (first-
generation flies) we see that in general the first post-
acrostichals tend to be reduced more than the second.
This may be expressed as a fraction :

First post-acrostichals lacking 40.5 _ ^ . q

Second post-acrostichals lacking 34

It is possible that thig tendency to reduce the first post-
acrostichal more than the second is evidence of relation-
ship to L. ccesar Linn., in which the absence of the former
and the presence of the latter is the normal condition.
Strain 1913 — A (Table III), however, gives

First post-acrostichals lacking __ 329 _ ^ . «
Second post-acrostichals lacking 750

Considering the rieduction in the first post-acrostichals
separately, we may express the effect of selection as
follows :

Parents.
Matings.

l8t

post-acros.
lacking.

Offspring.
First post-acros. lacking.
First post-across, normal (2 per fly)

0(f)

^ = 0.021 ±.004.

AF^a

|«g = 0.087 ±. on.

AF^

-S^ = 0M1±.002.

AF,b

r\ = 0.093 ± .010.
356

Digitized by

Google

350

THE AMEBIC AN NATURALIST [VoL.XLVm

AF^
A¥nb
AFtC

64
346

60
190

74
292

= 0.185 ±.014.
= 0.316 ± .023.
= 0.253 ±.017.

From these figures it is readily seen that reduction m
the first post-acrostichals is not entirely consistent with
the direction of selection.

Let us test the same matter for reduction in the second
post-acrostichals.

Parents
Matings

AF^a

AF^

AFJb

AF^a

AFJ>

AF^

1st post-acros.
lacking

2(f)'

1
4
1.5

Offering.

First post-acros. lacking^

First post-acros. normal (2 per fly)

482
63

: 0.01 7 ±.004.

298
193

= 0.211 ±.016.

426

147

= 0.453 ± .016.

356

145.5

= 0.413 ±.018.

= 0.420 ±.179.

346

^^ = 0.363 ± .023

190
108

= 0.370 ±.019.

In this case also the results are not consistent witti the
direction of selection, although there is better agreeuaent
here than in the case of the first post-acrostichals. Ibis
is probably due to the fact that the numbers are larger.
As regards the irregularities that do occur, I consider
them as evidence of recombinations of multiple factors,
insofar as they are not due to probable error.

1912 — c (Table II) is a strain that especially tends to
lack the first post-acrostichals. Thus for the entire
strain

First post-acrostichals lacking __ 25 _
Second post-acrostichals lacking "" 11 ~

Digitized by

Google

- ' ■ ptfnI

No. 570] FACTORS GOVERNING DISTRIBUTION

361

In the 137 offspring of a single mating of this strain, cF^c,
there are 23 first post-acrostichals lacking, showing that
it is due to this mating especially that the strain is so
lacking in first post-acrostichals.

It can not as yet be said that the factors governing the
first post-acrostichals are altogether independent of those
governing the second. That a. certain degree of inde-
pendence obtains is evident from a comparison of the
ratio of reduction in first to reduction in second post-
acrostichals in flies in general (Table I), with the same
ratio for strain 1913 — A. In the former case we have
40.5/34 or 1.19. In the latter we have 329/750, or 0.43.
In order to establish the independence of the factors un-
derlying these two tendencies it will be necessary to
obtain, either by selection from a strain showing both
tendencies or by breeding from wild stock, two strains,
one tending to lack the first while retaining the second,
and the other tending to lack the second while retaining
the first.

A point of interest in strain 1913—^ is the presence of
twelve small second post-acrostichals in the progeny of
AFzb in which the female had one of these reduced to half
size. The progeny of AFza in which there was total ab-
sence of these bristles showed either presence or absence
of the same but no reduced bristles. In Fj, however, we
have eight reduced bristles. The occurence of these
small bristles in the progeny of certain matings is taken
as an indication of recombinations of multiple factors,
but the numbers are too small to establish this with cer-
tainty.

A glance at the tables shows that third post-acrosti-
chals are rarely lackhig. These are normally present in
all related species, while in a few, — Cynomyia mor-
tuorum, Musca domestica, Pseudopyrellia cornicina, and
others, there is normally but one post-acrostichal, and this
is always the last.

Posterior dorso-centrals are very rarely absent. Thus
in the 2,273 flies recorded in Table I only one had a single
post-dorso-central missing. Reduction in post-acrosti-

Digitized by

Google

352 THE AMEBIC AN NATURALIST [Vol. XLVIH

chals among these is 79.5. Among the 1,206 flies of strain
1913 — A there are but three post-dorso-centrals gone.
This latter is a highly reduced strain as regards post-
acrostichals, lacking 1,081. This great reduction in
acrostichals seems not appreciably to have affected the
dorso-centrals, a fact which argues for the independence
of the factors controlling the distribution of these two
sets of bristles.
Thus for flies recorded in Table I we have

Post-acrostichals lacking _ 79.5 _ ^ ^^
Number of FH^ " 2273 ^ ^'^^'
One post-dorso-central lacking.
For flies in strain 1913—^ (Table III) we have

Post-acrostichals lacking ^ 1081 _ ^ qq

Number of Flies " 1206 ~ ^'^^'

Three post-dorso-centrals lacking.

Among the 3,238 flies recorded in Tables I and in only
four post-dorso-centrals are lacking, while among the 810
flies of strain 1912 — c (Table II) there are 13.5 lacking.
The lack of post-acrostichals in this latter strain is 37.
There are 9.5 dorso-centrals lacking in the progeny of the
trio, cFga, among which there are only seven post-acros-
tichals lacking.

Thus we see that lack of post-dorso-centrals is in no
way correlated with lack of post-acrostichals, but is evi-
dently governed by distinct factors.

Variatiois by Addition of Bristles
A strain of Lucilia sericata, 1913 — E, showed some
interesting variations chiefly in the direction of addition
of bristles. The mother was normal (3, 3, 3, 3), taken at
the Bussey Institution, March 19, 1913. The detailed ac-
count of the strain follows :

c?cf ??

38 43 = 3,3,3,3.

1 0 = 3, 3, able, 3.
F,

from ^F.a={^ = 3;3^ab!c,3.

Digitized by

Google

No. 570] FACTORS GOVERNING DISTRIBUTION 35S

C?(f

76 = 3, 3, 3, 3.
0 = able, 3, 3, 3.

l=!abc, 3, 3, labc.

from EF^=Z(J and 2 = 3, 3, 3, 3.

cfcf

(?C?

318

251 = 3,3, 3, 3.

5 = 3, albc, 3, 3.

61 = !abc, 3, 3, !abc.

1= lalbc, 3, 3, labc.

4=!abc, 3, 3, 3.

0 = 3, abcl, abcl, 3.
0=labc, albc, albc, la

6 = 3, 3, 3, labc.

3 = 3 (a), 3, 3, 3.

l=labc, albc, 3, labc

2 = 3, 3, 3 (a), 3.

0 = 3, able, 3, 3.

5 = 3, 3, albc, 3.

l = albc, 3, 3, albc

0 = 3, albc, abcl, 3.

3= labc, 3, 3, lalbc

0 = 3, 3, 3, albc

1 = 3, albc, 3, labc

0 = 3, 3, 3, able.

l=labc, 3, 3 (a), labc

0=!abc, a!bc, 3, 8.

1= labc, albc, 3, lalbc

0 = 3, 3, 3 (b), 3.

l = albc, 3, albc, albc

1 = 3, albc, albc, 3.

0=lbc, 3,3, lalbc

0=lalbc, lalbc, !a!bc, lalbc

0 = 3, 3, 2,3.

1 = 3 (b),3, albc, albc.
2= lalbc, 3, 3, labc.

0 = 3, ac, 3, 3.

0 = 3, ac, ac, 3.

0
F.

from .

1 = labc, 3, albc, labc.

1 = 3, abc, 3, 3.

rro - f 0 ^^^ 3, 3, 3, 3.

^•«- is =3; 3; albc, 3.

Pair

segregated, July 22: larvae July 30.

c?c?

191

100 = 3,3,3,3.

0= labc, 3, abcl, labc

43= labc, 3, 3, labc.

0 = 3, abcl, 3,3.

3= labc, 3, 3, 3.

0=labc (a), 3,3.

1 = 3, 3, 3, labc.

0 = 3 (a), 3,3 (a), 3.

1= labc, albc, 3, labc.

l=labc, 3, 3 (a), labc

2= labc, 3, albc, labc.

l=lbc, 3, 3, 3.

2= labc, albc, albc, labc.

from .FF,6=<J and $ = 3, 3, 3, 3. Pair segregated August 22.

cSd 52 c?c? 9?

41 57 = 3, 3, 3, 3. 2 0 = 3, albc, 3, 3.

0 1 = 3, albc, 3, labc 0 1= labc, 3, 3, labc

1 0 = 3, abcl, 3, 3. 0 1 = labc, albc, albc, labc

A summary of this strain is given in Table IV.

The points of interest to be noted in this table are as
follows :

There are many supernumerary bristles in the flies of
this strain.

The number of bristles added in the progeny of any
mating is very variable and has no consistent relation to
the visible character of the parents.

Addition of bristles tends very much to favor the
females, reduction still affecting the males.

Despite the high ratio of bristles added, there are

Digitized by

Google

364

THE AMEBIC AN NATURALIST [Yoh. :S^x,^yj\S5^

CO*

£
i

o I
^ I

S 'I

CO
CO*
CO*
CO*

O NO

^00

1-1

||S

l-H fH lO

CO '^

1="

•b

CO a

|ii

0»'

o od

rH 1

CqL

^TIse

•s?i

•b

o o-^

«•"§

It-

00 ^G

Hft<*

•a

o oo

w^ >

•b

O OrH

O iH rH

•g

o»

»o

►

•b

rH iH ^

5
S

O O 1H

O '

•b

O O CO

5 S§

•b

X o> 00

CO «o ^

o»

v«

co-

co CO*

-a

CO'

CO cT

«l

§ II * CO II B

Co" OS

I Scoco^^

CO CO'

i i'boi i

CO CO 1

H 1

-bo

pKt

Digitized by

Google

No. 570] FACTORS GOVERNING DISTRIBUTION 355

nevertheless a few flies in the strain in which bristles are
lacking.

Bristles normally present may be lacking in individuals
having additional bristles.

Genebal Summaby and Conclusions

Taking a general summation of all the bred material of
Lucilia sericata, we find that reduction affects the males
while addition affects the females. Of the 5,367 flies bred,
2,708 are males and 2,659 are females, giving practical
equality.

Eeduction in the males is 748.5 bristles, while in the
females it is only 455.5 bristles. As has been noted before
the degree of reduction in the females is increased by the
later generations of strain 1913 — A, by reason of the fact
that reduction rarely goes beyond the loss of two bristles
in a single fly. Thus when most of the flies of a popula-
tion become reduced to this extent it is evident that reduc-
tion in the males would be but slightly in advance of that
in the females.

There are 210 bristles added in the males, while there
are 343 added in the females. Thus addition affects the
females more than the males. These figures for bristles
added represent number of bristles, and thus no distinc-
tion is made between bristles of large and bristles of
small size.

I wish to express my appreciation for the advice and
criticism offered me in this work by Professor W. M.
Wheeler, Messrs. H. D. Fish, S. G. Wright, and C. C.
Little.

Digitized by

Google

PHYSIOLOGICAL CORRELATIONS AND CLI-
MATIC REACTIONS IN ALFALFA
BREEDING^

GEO. F. FREEMAN
Arizona Agricui/tural Experiment Station

Climatic Reactions

To the worker who attempts to apply the recognized
laws of heredity to the actual operations of plant improve-
ment many difficulties arise which heretofore have been
largely avoided by students of pure genetics. Color and
form characters are but little affected by the immediate
ordinary environment and hence, for the sake of simplic-
ity, are usually chosen by investigators of heredity. To
the economic breeder, however, such characters are of but
little consequence except in so far as they indicate phyletic
relationships. Of greater importance to the breeder are
those differences in yield and quality which are the re-
sults of inherited, invisible, physiological powers within
the plants, whereby each variety may respond differently
in manner or degree to the same environmental stimulus.

Those hereditary units which have to do with vegetative
vigor, heat, cold and drought resistance, time of maturity,
chemical structure, reproductive strength, etc., are as yet
but little understood. This is largely due to the difficulty
of exact experiments concerning them. This difficulty is
occasioned by the complexity of the reactions of these
hereditary forces with the external environment, and also
by the direct influence of the development of one part of
the plant upon that of some other part The plant at ma-
turity presents the resultant of its environmental reac-
tions during development. The nature of these reactions

1 Bead before the American Breeders ' Association, Columbia, S. C, Jan-
uary 26, 1913.

356

Digitized by

Google

No. 570] ALFALFA BREEDING 357

is determined by the structure of the vital forces within.
These differences in vital structure may or may not be
accompanied by visible morphological differences. Such
cases of correlation have been known and used in selecting
for qualities which they were thought to indicate. The
much quoted example of the supposed correlation between
the short-haired rachilla and high brewing quality in bar-
ley is a case in point. It has been found, however, that,
whereas, in one strain or race the correlation may hold, in
another, the two characters are in no way related.
Another case of similar nature is the coupling of cob and
pericarp color in certain varieties of com and their com-
plete independence in others. Many other examples could
be adduced to show that the coupling of two characters in
a given race of plants is no indication that these same
characters are inseparably linked in all races of the same
species. These facts have greatly reduced the value for-
merly ascribed to gametic correlations in plant breeding.
Under our present knowledge, therefore, we must depend,
for the most part, upon direct experimentation, rather
than correlations, to discover the hereditary physiological
characters of the varieties with which we are working.
Any additional light, therefore, which may be had con-
cerning the nature of such characters, together with meth-
ods for the study of the behavior of the same in their rela-
tion to each other and to their physical surroundings, will
have not only a scientific value, but will also fill a distinct
practical need.

As an illustration of such a study we may now examine
the data concerning the development, yield and chemical
composition of forty-four regional varieties of alfalfa
which were grown on the Experiment Station Farm at
Phoenix, Arizona, during the season of 1910. In the case
of this plant, which occupies the ground throughout the
year and from which six or seven crops may be harvested
during the growing period, the climatic factors include a
long series of variations coincident with the changing sea-
sons. Now, since every variety consists of its own pecul-

Digitized by

Google

358

THE AMERICAN NATURALIST [Vol. XLVHI

iar complex of hereditary physiological forces, each sensi-
tive in its own manner and degree to the impinging ex-
ternal stimuli, it is not surprising that the resultant (the
gross climatic reaction) should be sharply different in the
several varietal groups.

The unequal effects upon the vegetative growth of the
different varieties brought about by the climatic changes
which occurred during the course of the summer may be
exhibited by calculating the place variation in yield. This
is best shown by correlating the first with each of the fol-
lowing cuttings throughout the season. The result is a
definite curve, beginning and ending high with a strong
sag in the middle.

TABLE I

Place Variation in Yieli>

Cattingi 1 and 2 1 and 3 1 and 4 1 and 5 1 and 6

Correlation. + .75 ± .04 + .68 ± .05 + .33 ± .09 + .36 ± .09 + .58 ± .07.

These figures indicate the presence of some disturbing
factor which reached its maximum intensity during the
fourth and fifth cuttings, and to which certain plots were
more sensitive than others. The average period through
which the growth of these two crops extended was June 22
to August 27. The fact that these dates include the hottest
portion of the summer strongly suggests temperature as
the disturbing factor.

The mean maximum temperature, mean minimum rela-
tive humidity and the correlation between yield and water
supplied are given in the following table :

TABLE II
Temperature, Relative Humidity and Water Supply

Cutting

Dates locIudlDg Arerage Periods
of Growth

From March 23 to April 23
From April 23 to May 23
From May 23 to June 22
From June 22 to July 23
From July 23 to August 27
From August 27 to October 5

Mean Max- iMeaDMioimum
imum Tempera- 1 Relative
ture o F. I Humidity

82.8
93.8
103.6
104.8
104.4
102.0

27.00
23.00
20.40
25.26
30.00
25.18

Correlation

Between Yield

and Water

Supply

— .09±.10
4- .05 dr. 10
4 .40 i. 09
-f .21 ±A0

— .04 dr. 10

Digitized by

Google

No. 570] ALFALFA BREEDING 359

That the relative humidity had little to do with yield is
shown by the fact^iat the highest averages for this factor
occurred on the first and fifth cuttings which were the
highest and lowest in yield, respectively.

Although it was intended to give each plot approxi-
mately the same amount of water for each cutting, uneven-
ness in tiie slope made this impossible. The average
amount of water applied to each cutting was 6.28 inches
with an average standard deviation of 1.54 inches. Now,
taking cognizance of this variation in the water supply,
we find that its effect upon the yield was only appreciable
in the fourth and fifth cuttings. Eecords were not made
of the water supplied to the first cutting, but after that
time they are complete. By reference to Table 11 it will
be observed that these correlations in the second, third
and sixth cuttings are so small as to be negligible, but in
tiie fourth and fiftb cuttings they are suflSciently large to
indicate that this factor was of some importance in gov-
erning the yields. These results may be interpreted as
meaning that approximately 6.28 inches of water were
ample for each cutting during the cooler weather of spring
and fall. That too much was not given at these seasons,
however, is shown by the absence of large minus correla-
tions. Factors other than water supply, therefore, gov-
erned the yields during these i>eriods. Hot, dry weather
came on during the growth of the third cutting, but the
amount of water supplied plus the winter and spring sur-
plus left in the soil was ample to mature the crop. With
the continued high demand for water during the hot
weather of July and August, the surplus having been
exhausted and the summer rains helping but little, six and
one fourth inches was not suflScient. There was, therefore,
marked suffering for water, which was reflected in the
yields of those plots that received slightly more or less of
irrigation than the others.

It would seem, therefore, that high temperature and a
slight deficiency of water were the disturbing factors in

Digitized by

Google

360 THE AMERICAN NATURALIST [Vol. XLVm

the relative yields of the varieties tested, and that certain
ones were more sensitive than others to these influences.

If we turn to the relation between stand and yield, we
shall again find a strong disturbance of the normal corre-
lation as shown in the following :

TABLE III

CJOREtELATION BETWEEN STAND AND YiELD
Cutting iBt 2d M

Correlation + .78 ± .04 + .55 ± .07 + .47 ± .08

Cutting 4th 6th 6th

Correlation + .54 ± .07 + .10 ± .10 + .70 ± .05.

The exceptionally low coefficient of the fifth cutting was
due to the low yields on the part of plots which had good
stands but were relatively inactive during the hot
weather and partial water famine which occurred at this
period. On the other hand, certain plots through their
resistance to heat and consequent activity at this period,
overcame to a large extent their handicap of poor stands,
and nearly obliterated the usual plus correlation between
stand and yield.

The data thus studied en masse indicate at least two
physiological groups which are unequally sensitive to the
climatic changes which occur in the course of a growing
season, and whose reactions were sufficiently strong to
change almost completely the order of the productivity of
the {dots. In order to test this conclusion let us turn to
the individual plots and endeavor to discover and classify
the physiological varieties indicated above.

If, now, we arrange the forty-four regional strains
according to their morphological characters and geo-
graphical origin, we shall have five more or less distinct
groups as follows: Mediterranean, Peruvian, European,
American and Turkestan. The behavior of these varietal
groups through the course of six cuttings during the sum-
mer of 1910 substantiates the conclusions already drawn
and illustrates the sharp differences in climatic reactions
which may be observed in the several varieties of a single
species.

Digitized by

Google

No. 570] ALFALFA BREEDING 361

Morphologically the Mediterranean and Peruvian al-
falfas are so distinct in type that any one at all familiar
with the different kinds of alfalfa would recognize them
at a glance, whether a whole field or a single plant be
observed. The presence of yellow or greenish blue flow-
ers also determines a variety to be of northern origin with
mixtures of falcata characters, which usually carry with
them resistance to cold and drought. Otherwise, the
Turkestan, American and European types are so nearly
alike that only an expert would recognize them in mass
culture. The individual variations within these three
types intergrade to such a degree that one could scarcely
assume to judge, from the observation of a single plant,
the type prevailing in the field from which it originated.
The three types, however, differ markedly in their phys-
iological reactions as we shall presently see. The distinc-
tions, in this regard, as exhibited on our plots, are not
nearly so marked between the American and Turkestan
alfalfas as between these two types, on the one hand, and
the European, on the other. However, in northern cli-
mates where winter resistance enters as a i)otent factor,
the Turkestan alfalfa exhibits greater hardiness than the
American form, and, therefore, is able to maintain a more
perfect stand through seasons of extreme frost.

When grown under Arizona conditions, the average
yields of each of these five type groups present seasonal
curves at once striking in their diversity and contrasts.
These differences are exhibited more easily by plotting
the average of all the plots as a straight line, and the aver-
age of the different groups as percentages of the total
average above and below the general average line.

In observing Fig.l, we are first impressed with a marked
similarity in the performance of the European and Medi-
terranean alfalfas, on the one hand, and the American and
Turkestan on the other, and also with the striking differ-
ences exhibited between the two groups. Although the
average yield of the European plots greatly exceed that of
the Mediterranean plots, the shapes of their respective

Digitized by

Google

362

THE AMEBIC AN NATURALIST [VoL-XLVDI

curves are almost exactly alike, tlie greatest relative yield
of each being in the heated part of the summer after the
beginning of the water famine. In like manner, the
American and Turkestan varieties made similar relative
yield curves, that for the Turkestan being slightly above
the curve for the American strains. Here, however, the

CUTTING

^ 3(
a. 31.

31
u
> J*

o
5 it

K ^^

5 i¥
u

u
Q. ia

if-

!•

AVE. YIELD 3 f i'J'LBa.

1^
U
It

Sf^iM

Relati\'g Yield of Regional Vabieties Based on tub AvEBAaB of All Plots

AS 100 Pbb Cent.

curves bend strongly downward in mid and late summer,
as if these types were much less resistant to the accumula-
tive effects of drought and heat. In fact, it would seem
that during the hot period included within the fifth cutting

Digitized by

Google

No. 670] ALFALFA BREEDING 363

(July and August), the American and Turkestan varieties
were comparatively inactive, yielding only about eight
hundred pounds of dry hay per acre, as against more than
1 ton and a half each on the first cutting. The relative
ield curve for the Peruvian type stands separate and dis-
tinct from the others. Although here, as with other varie-
ties, the yield declines with the advance of the season, the
persistence and vigor with which this strain resisted the
summer heat and drought caused it to gain rapidly on the
other varieties in relative yield throughout the season
until the very last cutting, when there was a slight decline.

Disregarding the shape of the curves we may now notice
the total yield for the season. In this respect the different
regional varieties take the following relative order: Peru-
vian, European, Turkestan, American and Mediterranean.
It is here noticeable that, though the European and Medi-
terranean varieties have similar seasonal yield curves,
they are not contiguous in the arrangement based on total
yields. This is a result of a marked difference in the
stand maintained by the two varieties which averaged
ninety-two per cent, for the former and seventy-four per
cent, for the latter. In their ability to maintain stand, the
Peruvian, European, Turkestan and American varieties
were about equal, averaging 92, 92, 93 and 94 per cent., re-
spectively. The lack of stand on the part of the Medi-
terranean alfalfas was not due to the poor quality of the
original seed, for all of these plots once had perfect stands.
This behavior is also in accordance with the records of
other fields of Mediterranean alfalfa in the southwest,
which have come under the observation of the writer. The
explanation of the weakness of the Mediterranean and
corresponding strength of the otherwise similarly reacting
European alfalfa in maintaining stand under Arizona con-
ditions is a subject for further careful physiological study.

The recognition, analysis, and calibration of these dif-
ferences of the physiological reactions of varieties are
thus seen to become a first essential in the study of cli-
matic adaptation^ and form the basis for rational pro-
cedure in the choice of varieties and in selective breeding
for the improvement of the same.

Digitized by

Google

364 THE AMERICAN NATURALIST [Vol. XLVIII

Correlations
In the improvement of varieties of plants, quality is
often as important as quantity of yield. This is especially
true in a forage crop, such as alfalfa. Since nitrogen, next
to fat, is the most expensive of the necessary food constit-
uents, it may be taken as the measure of quality. Com-
merical buyers judge alfalfa hay by its purity, odor, color
and percentage of leaves retained in curing and baling.
The value of the leaves lies in their relatively high nitro-
gen content and the consequent increased food value which
they impart to the hay. Expressed quantitatively, the cor-
relations between the nitrogen content of the hay and the
percentage of leaves for the six cuttings were as follows :

TABLE IV

CfOEBELATION BETWEEN NITROGEN CONTENT OP HaY AND PEB CENT. OP LEAVES

Cutting Iflt 2d 3d

Correlation + .46 ± .08 + .61 ± .06 + .72 ± .05

Cutting 4th 5th 6th

Correlation + .68 ± .05 + .61 ± .06 + .52 ± .07.

That the final value of the hay is markedly dependent upon
the composition as well as the percentage of leaves is
shown by the following high and fairly uniform correla-
tion between the nitrogen content of the hay and the nitro-
gen content of the leaves :

TABLE V

Correlation between Nitrogen Content op Hay and Nitrogen Content

OP Leaves

Cutting Ist 2d 3d

Correlation + .69 ± .05 + .73 ± .05 + .42 ± .08

CuttliiR 4th 5th 6th

Correlation + .67 ± .06 + .85 ± .03 + .74 dt .05.

If, now, we have shown that the quality of the hay de-
pends primarily upon the percentage and composition of
the leaves, we may proceed to investigate those factors
which indirectly modify the feeding value by influencing
the amount or character of these organs.

The factors most profoundly affecting the percentage
of leaves were yield, height and stage of maturity at
which the cutting was made. Local or varietal forces were

Digitized by

Google

No. 570] ALFALFA BREEDING 365

suflBciently constant to hold the place variation of this
character to the plus side of the equation for four out of
five determinations made, as is seen in the following
table:

TABLE VI

Place Variation in Pbbcentagi of Leaves

Cutting 1 and 2 2 and 8 8 and 4 4 and 5 5 and 6

Correlation. + 10 ± .10 + .23 ± .10 + .16 ± .10 + .46 ± .08 — .12 d: .10.

These correlations, however, are low and seem to indicate
that tihie natural varietal traits were being overcome and
obscured by other variable factors.

Contrary to expectation, the stand had little to do with
the percentage of leaves, as the following low and incon-
stant correlations show.

TABLE VII

COREELATION BETWEEN THE PERCENTAGE OF LEAVES AND STAND

CnUinK 1st 2d 3d

Correlation — .14 ± .10 — .02 ± .10 + .03 ± .10

CatUng . 4th 5tb 6th

Correlation + .10 ± .10 + .07 ± .10 + .24 ± .10.

On the other hand, the relation between height and
yield and percentage of leaves was constant and marked,
except in the last two cuttings.

TABLE vin

COBKELATION BETWEEN PERCENTAGE OP LEAVES AND HEIGHT AND YiELD

Catting l8t 2d 8d

Yield — .41 ± .08 — .60 ± .07 — .15 ± .10

Height — .48 ± .08 — .62 ± .06 — .68 ± .05

Catting 4th 5th 6th

Yield —.40 ±.09 + .20 ± .10 + .30 ± .09.

Height — .55 ± .07 + .09 ± .10 + .19 ± .10.

The sudden change from minus to plus in these correla-
tions should be noted. The average heights of the first
four cuttings were 32, 30, 28 and 27 inches, respectively.
The average height of the fifth and sixth, were 15 and 12
inches. This would suggest that at or below 15 inches the
mutual shading of the stems is not suflBcient to cause an
appreciable shedding of the lower leaves. Up to this
I)oint, moreover, growth usually takes place by an increase

Digitized by

Google

366 THE AMEBIC AN NATURALIST [Vol. XLVIH

in the number of nodes, each with its accompanying leaves
and side branches. Above fifteen inches, however, the
principal growth in height consists in a lengthening of the
intemodes and, consequently, a relatively greater produc-
tion of stem as compared with leaf tissue. In this phys-
iological correlation lies the core of the difficulty in breed-
ing at once for quality and quantity. The act of high pro-
duction within itself cuts down the quality of the product
by reducing the ratio between the leaves and the stems.

This difficulty, moreover, occurs in the composition as
well as the percentage of the leaves. The correlation ex-
isting between the nitrogen content of the leaves and the
number of days required to mature a cutting is shown in
the following table :

TABLE IX

COEBELATION

BETWEEN

THE Nitrogen Content of Hay and the Period
Eequtred roR Maturity

Cutting

Correlation .

l8t 2d 3cl
... —.33 ±.09 —.30 ±.09 —.27 ±.09

Cutting
Correlation.

4th 5tb 6th
... —.52 ±.07 —.50 ±.08 —.17 ±.10.

Quickly maturing varieties thus have leaves richer in
nitrogen than those which require a greater length of time
for completion of growth. When, however, we take the
average number of days required throughout the season
to mature a cutting for each plot and compare this with
the total seasonal yield we find a correlation of + .43.
Thus we are again confronted by a minus correlation be-
tween quality and yield which must be overcome if we
would make progress simultaneously in both lines.

As further examples of antagonistic correlations, a few
instances may be taken from the data furnished by forty-
three plots of pure races of alfalfa grown during the sum-
mer of 1910. The correlation between height and percent-
age of leaves was again constant and marked. The results
here paralleled those found for the regional varieties.
Whereas yield was uniformly correlated i)ositively with
both stooling capacity (av. No. stems per plant) and
height, it is interesting to note that there was also a uni-

Digitized by

Google

No. 570]

ALFALFA BREEDING
TABLE X

CORBBLATION IN PUBE BaCIS

367

Correlation Between

Cuttiogs

July

AugUBt

September October

Green weight and average
number stems

Green weight and average
height

+ .76 ±: .04
-h .01 d= .10

— .29 =b .09

— .39 ± .09

4- .42 rb .08
4- .44 ±.08

— .19 ±.10

— .15 ±.10

4- .62 ± .06;+ .50 ± .08
4- .22 ± .10 4- .33 ± .09

Average height and number of
stems

- .32 ± .091— .21 ± .10

Average height and per cent,
leaves

— .55 ± .07 -- .51 ± .08

form minus correlation existing between them. We thus
have two factors both making for yield, but seemingly
(probably physiologically) aatagonistic to each other. In
breeding for high yielding strains we are here again called
upon to overcome by selection an antagonistic physiolo-
gical correlation.

This brings us to the following final conclusion which
the writer wishes to emphasize :

In economic plant breeding one frequently encou/nters
physiologically negative correlations such as those, in
alfalfa, between height and stooling capacity, height and
percentage of leaves, and between yield and quality. In
seeking improvement, therefore, the breeder must recog-
nize and make use of these facts in the interpretation of
results obtained, and also search for races which violate
su^h naturally antagonistic correlations to the greatest
possible extent.

Geisteral Conclusions

That the complex of allelomorphs, which we call a va-
riety, may be definite as both to ultimate composition and
organization is not here questioned. When, however, we
consider that visible characters are only the expression of
the reactions of the vital forces of the plant with the en-
vironment, we can realize that the variety, as we see it, is
not a definite thing, but is a result of two independent
classes of factors. Change either and the result corre-
spondingly changes.

Digitized by

Google

368 THE AMEBIC AN NATURALIST [V0L.XLVIII

We are therefore to look upon the variety as a delicately
organized chemical compound. The various factors of
climate and soil may be compared to different physical
influences to which the original compound may be sub-
jected. As the chemist would expect reactions varying in
accordance with the physical stimuli used, so will the plant
react in agreement with the different environmental com-
binations. The extent to which this will change the nature
and appearance of plants is often far reaching. Cook,
working with cotton, has found that certain cultural condi-
tions at an early stage of growth will make profound dif-
ferences in the method of branching which determines the
whole subsequent development of the plant and affects
materially its economic value. Cultural and climatic reac-
tions often lead to error among those who assume them to
be mutative changes induced by the new conditions. That
these reactions may bring to light sub-races with heredi-
tary tendencies not hitherto called into expression and
which, by selection, may be secured as pure races, is the
probable explanation of many cases of supposed direct
climatic adaptation.

Thus, realizing the true nature of a variety, we can
draw further upon the analogy of the chemist who investi-
gates an imknown substance by testing its reactions with
a large number of known reagents. In like manner the
breeder can only understand the true nature of the hered-
itary vital forces within a plant after he has tested and
calibrated its reactions against a variety of soil and cli-
matic factors. These reactions are of interest to the
farmer only in so far as they affect the economic value of
the variety as grown in his own locality but to the breeder
and student of heredity their importance is fundamental.
This is so because they enable him to classify, coordinate
and interpret the experimental results that he obtains.
This ability finally must form the basis of all rational pro-
cedure, whether one be engaged in the study of pure gen-
etics or in the operations of practical plant improvement.

Digitized by

Google

TAXONOMY AND EVOLUTION

ByX

'' Some passages in this book, if taken alone and read hastily, may
appear to discourage systematic Zoology. This is far from my inten-
tion. No one can study the great naturalists of the seventeenth and
eighteenth centuries without feeling how seriously their work is impaired
by the defective systems of the time. It is not systematic but aimless
work that I deprecate — ^work that springs from no real curiosity in
Nature and attempts to answer no scientific questions." — L. C. Miall,
" Natural History of Aquatic Insects," Preface, p. L

Introduction

LiNN^us bestowing Latin names upon animals and plants
was simply tripping gaily across the back of a half submerged
Behemoth and mistaking it for dry land. Now the beast is
careering around, and in spite of zoological congresses and inter-
national rules nobody quite knows what to do with him. No
doubt when some zoological czar arises and issues his fiat a uni-
form system of nomenclature will be adopted and things will
begin to straighten themselves out. This can only be a matter
of time — ^the past can not be altered. On systematists to-day
necessarily devolves the dull, diflBcult and important duty of
going through the descriptive work of the early naturalists and
emending it; so that Spallanzani's derisive sobriquet of
** nomenclature naturalists'' was a little unjust, even in his time.

Whatever opinions may be held upon the genius of Linnaeus,
in justice to him it should be said that it was not until his ex-
ample had been followed by a crowd of other workers eager to
attain to immortality by way of the back door he had left open
that the fat was really in the fire.

Well knowing the confusion into which systematic work in
zoology was brought by the early naturalists, modem systemat-
ists in our opinion will be the authors of a similar confusion in
the future if some of the slipshod methods of modem syste-
matics are not corrected. Moreover, a confused nomenclature
is not the least of the evils which second-rate systematic work
brings in its train.

369

Digitized by

Google

370 THE AMEBIC AN NATURALIST [V0L.XLVIU

Systematisfts with a proud curl of tbe lip may tell us that the
work is not done now as it once was. Indeed, to those who are
not able to project themselves into the future it may seem in-
credible that the systematists of a later date will be able to find
much room for complaint in the elaborate descriptions and care-
ful figures of modem descriptive writers. For the moment, how-
ever, it €niffices us to point the parable by remarking that in
1780 Spallanzani was able to refer to the '* beautiful figures''
and ** careful descriptions" of a systematic worker on frogs.
We, of course, know without seeing them that the figures were
not beautiful nor the description, careful — any way in the sense
of being complete. We have therefore to reflect whether the
zoologists of a future generation will find the work of to-day any
freer of faults than that of the past centuries.

Systematic Work. General CoNsroERATioNs

It is necessary to insist at once that systematic work is not
merely a question of nomenclature, names and novelties. Sys-
tematists have only themselves to thank if such a narrow con-
ception of their province is very widely spread, especially
among morphologists and anatomists, who are ready to belittle
the value of the systematists' work. But science is measurement
and zoology — if you like — is description, and it is impossible to
dispense with the ^stematists' descriptive work. But we think
it possible to dispense with a good deal of stuff after this
fashion :

Metopidium high, suprahumerals rather long, acute, arcuate and
curved at the tips. Pronotum roughly punctured at the. bottom of fine
furrows. Color dark-ochreous. Posterior horn uniformly cylindrieal,
undulating or sinuous without rugosities. Underside, scutellum and
legs sordid-ochreous.

The phrase ''sordid ochreous " comes ready to hand and
makes it unnecessary for us to go in search of a suitable com-
ment.

**This is the 30th memoir" writes a systematist "on the
Zonitidae which I have published in this journal, describing in
all about 560 new species." We feel inclined to put our hands
resolutely on his shoulders and inquire if he ever saw a cteno-
phor swimming in the sea or watched the progress of an Asierias
towards its prey.

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION. 371

No one can look unmoved upon the Hymenopteran or Helicoid
specialist with head bent over a drawer full of shells or dried
insects on pins. It is not that we resent concentration or enthu-
siasm or even specialization, but the systematist has lost touch
with his own science of zoology.

Zoology, a cornucopia of marvels, lies at his elbow full to over-
flowing, but he is unmindful of it. It is as if a man should use
the Parthenon only as a convenient place on which to strike a
match for his pipe.

The divorce between systematic work and the rest of zoology
is the more regrettable because it is practically complete. It is,
we admit, expedient that zoology should be divided up into
anatomy, morphology and so on. But such a division is allow-
able only when it is expedient, while for intellectual purposes
such a division is and has always been a danger. To obtain
facts one must be an analyst, to consider them one must be a
j^ynihesist. Between the two there is all the difference between
a hodman and a natural philosopher.

But our contention is that not even the plea of practical ex-
pediency can justify the extreme state of specialization into
which systematic zoology has fallen, making itself manifeeft in
the concatenation of such purely artificial characters as that
*'the third joint of the antenna is longer than the second, that
the mesoscutellum is ovate and the color pink with blue spots.'*
All this simply makes one yawn, though there is this much to be
said in favor of this stamp of systematist, that nothing bores
him so much as the recitation of one of his own diagnoses or
being introduced to the systematist of another group.

Systematic work is a withered branch of the biological tree
which there is still hope of rejuvenating. Treviranus long ago
remarked that if we once regarded systematic work as a part of
biology and nomenclature as a means to an end rather than as
an end in itself, both might take their places in science. Let us
take every precaution against systematic work becoming one of
those unproductive and artificial pursuits which spring up like
mushrooms around centers of splendid endeavor and high
achievement. After Shakespeare came his commentators. Shall
it be said that after biology came the systematists t

We assume that the principal object of systematic work is to
discover the phylo-genetic classification of animals, for which it
is surely necessary that every animal as it passes through the

Digitized by

Google

372 THE AMERICAN NATURALIST [Vol.XLVIII

systematists' hands should be, as far as possible, thoroughly ex-
amined and described, no dependence being placed upon a few
superficial characters usually selected from the external parts t
That the systematist should concern himself, as he does, with
the external parts, leaving the anatomy to other workers, we
consider is as bad for the systematist himself as it is bad for the
science; for himself, he is doing work which can only keep his
soul alive with diflBculty — superficial clerical work which can be
"prompted by no real curiosity and attempts to answer no
scientific questions," and the results of the work itself is often
invalidated by the arrival of the destroying angel in the person
of the anatomist. For a superficial description often means a
wrong classification ; whence it follows that any zoo-geographical
deductions therefrom are invalidated; while a careless descrip-
tion usually ignores the possibilities of variation and shows no
evidence of pains having been taken to make identification easy.

Systematic work, then, is concerned with classification, geo-
graphical distribution, variation and identification, and there
would be no need for this paper, if it were more generally re-
alized that one thorough examination and description of the
whole animal assists those branches of the inquiry more than
twenty loose and superficial ones.

Of course systematic workers are not the only zoologists who
over-publish; yet they especially might cultivate a little of the
salutary reticence of C. L. Nitsch and Alfred Newton, who, with
no discredit to themselves, wrote and published little, yet it must
be admitted by those with an eye on the extravagant output of
others, to the advantage of zoology. The words "res non-verba*'
were the motto of Delle Chiaje, who, like Nitzsch, on his death
left behind many important discoveries unpublished and only
indicated in his drawings.

Classification in Genera i.

The coming of Evolution meant for systematic workers that
no system of classification would henceforth be considered as a
serious contribution to science, which was not constructed on
phylogenetic lines. It meant the final overthrow of such ideas
as Agassiz held, that the divisions of the animal kingdom were
instituted by the Divine Intelligence as categories of his mode of
thought — of such fantastic systems as those of Bafinesque and
Swainson and such strictly artificial ones as the arbitrary ar-

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION. 373

rangements of convenience which should be now used only in
those groups where, and for as long as, our knowledge of the
anatomy is so slight that some sort of temporary device for
sorting out genera and species has to be adopted.

The ideal system is now phylogenetic, i. c, it aims at recon-
structing in a genealogical tree the actual lines of descent.

Only those who have attempted the reconstruction of phylo-
genetic trees understand the intrinsic difSculties of the work.
There can be no doubt that the coming of Evolution has put
before the systematist a very difficult task. As to whether the
methods usually employed by him are adequate to the demands
placed upon them we are frankly sceptical.

Fortunately for the systematist the main lines of classifica-
tion in most groups are given him ready made by the morphol-
ogists who have laid down the foundations trusting to the ''s3n9-
tematist" to fill in the details. Such classifications — ^the main
phyla, classes and orders are of permanent value, because they
are founded upon a combination of characters of tried worth
judiciously selected after a careful survey of extensive embry-
ological and anatomical data.

Single Character Classification

On the other hand the minor systems — ^the families, genera
and species — the realm of the ''systematist*' — too frequently
consist of haphazard combination of a few character selected
because of their convenience in not entailing any anatomical
work, or selected on account of the ignorance existing of any
other — particularly internal — important characters. Ignorance
of their morphology has been the main reason for the difficulty
in classifying the Coleoptera. Entomologists are especially
prone to give their whole attention to what is visible without
the aid of dissection. In the Polyzoa the majority of forms are
only known by their external appearance and their classification
is proportionally unsatisfactory. In the MoUusca reliance is
placed on the shell ; in mammals the skull and the skin, in birds
the plumage are the articles of faith.

Single character classification or diagnosis by one or two
characters, as zoological history shows, has proved inadequate —
that it is unphilosophical is patent to all.

Such single character classification even when practised by

Digitized by

Google

374 THE AMEBIC AN NATURALIST [VoL.XLVIII

the great morphologists, men who, being acquainted with the
whole of the anatomy of the forms they were classifying, de-
liberately selected one or two characters after a survey of the
whole — was rarely a success. Huxley set out unabashed to
classify birds by their palate, and Agassiz fish by their scales —
^stems which have now shared the fate of most others which
set out to erect a classification on the modifications of a single
organ alone. Alfred Newton said that there was no part of a
bird's organization that by a proper study would not help to
settle the great question of its aflBnities.

The systematist who deals with the minor subdivisions of the
animal kingdom — families and genera — should be as much a
morphologist as the one who deals with the larger — the phyla
and classes.

Description

We have pointed out above that the adequacy of a system of
classification depends in great measure upon the thoroughness
of the description of the species and genera. Classification in
all groups has progressed in just proportion to the more exact
examination of the species considered in the classification.

The history of zoological research brings out this fact very
clearly, beginning with the work of Linnaeus, the originator of
the superficial diagnosis, passing on through Cuvier, who appre-
ciated the value of anatomical knowledge, to Von Baer, wbo
emphasized the importance of embryology.

It was not a "systematist" as we know him who first correctly
classified Lepas — ^the conchologists blindly accepted it as a
Mollusc. It was not a ''systematise' who first established Peri-
patus as an Arthropod, for the first describer of that animal
regarded it as a slug I

How rare it is to find in a description of a new species any-
thing more than an indication of the external parts. It is a
peculiarly arbitrary limit to a man's curiosity that restricts
his enquiry to the superficial aspect of an animal. A natural
philosopher ought never to be satisfied with the external ap-
pearance of things. The wisdom of the ancients bids us ** be-
ware of what things appear "; and the method of our modem
science is one of close and detailed observations. In scattering
names broadcast with liberal largesse upon species, varieties and

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION. 376

genera, (systematists have sometimes dropped into some carions
errors. Teratological specimens have been described as new
species and most zoologists have heard of the man who de-
scribed as a new species the longicom beetle, the head of which
having fallen off, had been fixed on upside down. His examina-
tion of a new species makes so slight an impression on his mind
that sometimes the same worker has described the same form
twice under different names.

The descriptive papers on Mollusca usually consist of short
descriptions of the shells, even written in a dead language. This
is conchology. Conchologists confine themselves to the pat-
terns and shapes of shells — ^nature's medallions — numismatics I
Much of this work — along with similar productions in entomol-
(^y and carcinology — we regard as positively flagitious.

Sir Ray Lankester in the article ** Zoology" in the Encyclo-
paedia Britannica (ed. XI.) remarks that museum naturalists
must give attention to the inside as well as to the outside of
animals and that to-day no one considers a study of an animal's
form of any value which does not include internal structure,
histology and embryology in its scope. Agassiz, too in his
famous ** Essay on classification" wrote that ''the mere indi-
cation of a species is a poor addition to our knowledge when
compared with such monographs as Lyonnet's Cossus, Bojanus'
'Turtle' Strauss Durckheim's Melolontha and Owen's Nauti-

**But," it will immediately be asked in chorus, *'do you
seriously suggest that a monographic volume should be devoted
to every new species!"

This is a leading question which brings us to the crux of the
whole matter, and can not be answered in simple '*Yea" or
**Nay."

The Provisionali Diagnosis

The amount of analytical study that may be given to «iny one
animal form in any one stage of its development is infinite.
The result is that in describing a new species for the purposes
of exact phylogenetic classification there must be a limit beyond
which it is unnecessary to go. Such a limit can not be otherwise
than arbitrarily selected according to the best judgment of the
fi^rstematic worker as to how much analysis is required to pl€u^
his new species, although at present, miserabile dictu, relatively

Digitized by

Google

376 THE AMEBIC AN NATURALIST [VoL.XLVni

very few animals have been thoroughly explored, yet in the dis-
tant future, in the millennium, it can not be doubted that every
genus, even every species will have been examined in toio in every
stage of its development and life-history as thoroughly as our
instruments and eyesight will allow, and perhaps a whole vol-
ume or several volumes will be devoted to every animal form.
At present, however, it is a waste of ink to consider a future so
far away. A more pressing duty is to consider how far modem
methods of superficial diagnosis fulfil the obligations placed
upon systematists not to give an exhaustive analysis of animal
forms, but to give suflScient data to meet the searching demands
of phylogenetic classification.

"We are aware of the fact that the convinced and determined
systematist does not maintain that the method of superficial
diagnosis does meet or is intended to meet the demands we have
been indicating. If he reads as far as this and does not throw
aside this paper in contempt, he is ready with eager forefinger
and glib apology to convict us of begging the question that sys-
tematic zoology can be ever anything, or should be ever any-
thing more than we have said.

It is often argued that the superficial diagnosis of the syste-
matic worker is simply a provisional diagnosis awaiting the con-
firmation of the anatomist. A plausible defence of the provi-
sional diagnosis is advanced by many workers in perfect good
faith which it is now necessary to anticipate and examine.

This argument defends the provisional diagnosis on two
grounds: (1) The advertisement theory; (2) the recognition
mark theory.

The supporters of these theories admit that the provisional
diagnosis in no way settles either an animal's systematic posi-
tion or its validity as a species. But it is alleged to be of value
and should be encouraged because it advertises the existence of
a presumptive new form which would otherwise renmin un-
known and overlooked in the store rooms of the museum and
laboratory, and because in giving an account of the external
parts, at all events, the systematist is describing those features
by which we are more or less easily able by a superficial exami-
nation to recognize summarily the form when it turns up again.

The first part of our answer amounts to a recapitulation of what
has been previously stated in general, viz., that systematics have
lost touch with the rest of the science. The output of systematic

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION. 377

work and the output of aiiat<Mnical and morphological work
nowadays move along completely different channels. The work
turned out by the systematic worker is scarcely, if ever, con-
ceived in the light of modem biological theory, is rarely couched
in terms of modem biology and rarely indicates a problem to be
solved or a question to be answered. It proposes distinctions
the anatomist sweeps away and hazards affinities the morphol-
ogist laughs at. It performs work that has to be done over
again, and instead of giving the morphologist what it claims to
give him — a sketch map of the country he is to traverse — cdl it
does is to bewilder him with a Will-of-the-Wisp's lantern, an
intolerable multitude of slipshod and untrustworthy directions
that he has come instinctively to suspect. We can not too often
ask the question, why should the work be done twice! Surely
it is time that something were done to stop this tremendous rush
for publishing provisional diagnoses that more time could be
devoted to the systematic study of animal forms, obtaining
thereby sound phylogenetic classification, sound deductions in
geographical distribution, valid species and a less confused
nomenclature.

Thus the systematist's protest that at least he ** advertises" pre-
sumptive new forms we can reply that he may do so, but that for
any purpose other than a dull census of the animal kingdom with
a very generous '*±" to it, he is a positive Benedick of zoolo-
gists, for ** nobody marks him.''

The upholders of the provisional diagnosis will say that at
any rate they are giving us a description of the external parts
and are increasing our knowledge by so much. True, but by so
inconsiderable an amount that when the anatomist comes along
with his scalpel he so quickly disposes of the external parts
merely by the use of his eyes that it is a matter of indifference
whether the former have been described or not. Moreover, the
great majority of the tens of thousands of descriptions that are
issuing from the press are of animals so closely related to pre-
viously described species that such descriptions really amount
to little more than a recitation of their distinguishing characters.

It is certainly useful to know that Caccabis rufa is to be dis-
tinguished from Perdix cinerea by its red legs and that the
LeporidaB can be discriminated by the character of their upper
incisors. But the question may well be asked, what is the use
of being able to distinguish one species from another without

Digitized by

Google

378 THE AMERICAN NATURALIST [Vol. XLVIII

being able to record at the same time anything about its bionom-
ics or anatomy which would give the distinction its real value.
A great deal is known about the partridges and hares, hence the
distinctions alluded to above are useful as an easy way of
quickly identifying them. But so long as nothing is known
about either of two species that are distinguished we are none
the worse oflP, if both remain indistinguishable.

Finally we would point out that of all people the systematist
should know that at present of the forms he advertises and
describes so copiously and summarily only a fractional part is,
or can be, dealt with by the laboratory worker. We are speaking
now of the anatomy pure and simple of new species and genera.
The laboratory worker proceeds slowly, is fewer in numbers
and has other problems — embryology (descriptive and experi-
mental), heredity, physiology (descriptive and experimental)
and morphology to attend to besides purely descriptive anatomy.
And yet anatomy — ^the very corner stone of the temple of
zoology — ^has to be restricted in output because none of the sys-
tematists will learn how to use a scalpel or look down a dissect-
ing-microscope — feats in themselves perfectly easy and calling
for no special training or faculties.

Possibly the upholders of the provisional diagnosis will main-
tain that by publishing his account of the difference between
closely allied forms the systematist is providing the biologist
with a stimulus to discover how much deeper such differences go.
But surely it is a strange perversion of a man's natural instinct
of curiosity that enables the systematist to rest content with
advertising problems instead of endeavoring to equip himself
for the task of undertaking them himself, who is eminently
suited to the work and whose occupation daily brings him into
close contact with them.

Finally we would point out that the enormous mass of species
which have been created upon superficial diagnosis so far have
remained unincorporated for the most part in the structure
it is designed to build up, viz., a clear comprehension of the
phylogeny of the lesser divisions of the animal kingdom. It is as
though a man were to set about building a house by making a
vast quantity of bad bricks and then to leave them scattered
about his site in the hopes that some one would come along and
make a house of them. Surely it is an economy of effort for the
systematist to take up the bricks and build himself, what time

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION, 379

the embryologist and morphologist are engaged upon their own
special tasks.

The Compabattve Value of Internal and External Parts

Briefly reviewing the discussion as far as we have carried it,
it will be seen that we are asking for sound phylogenetic classi-
fication of the smaller groups as well as of the larger ones, based
not upon single characters, but upon the whole of the characters
regarded collectively, for more careful and more thorough mor-
phological methods in description and for the discontinuation
of the provisional diagnosis. In view of the desirability of work-
ing up sounder schemes of classification from the enormous, un-
wieldy and superficially known mass of genera and species sys-
tematists can be rendering little service by continuing to turn
out indiscriminate provisional diagnoses.

It remains now to discuss in greater detail the proposal we
bring forward in the place of the provisional diagnosis.

The commonly accepted opinion is that while for the classifi-
cation of families and orders the internal parts must be taken
into consideration, for that of species and genera a summary of
the external parts is all that is required. On account of the
labor and diflSculty sometimes involved in dissection we are too
ready to assume that the internal parts in genera and species
present a dismal monotomy of character which it would be
profitless to investigate for systematic purposes.

If it is admitted that internal characters are of value among
the higher divisions of the animal kingdom, can the systematist
tell us at what precise point in the downward scale they cease to
have value, and at which reference need only be made to the
external parts? Even supposing for a moment that there is
such a limit, we are strongly of opinion that it does not come
before the genera.

A genus is of different value in different groups but as a rule
it presents so much difference in external form from other
genera as to warrant the inference that internal differences of a
like extent will be found if sought for. At the present moment
a genus is a perfectly arbitrary collection of species. We ven-
ture to prophesy that with more elaborate descriptions inter-
generic relationships will be more carefully defined and genera
will become less heterogeneous and more natural. But this is
by the way.

Digitized by

Google

380 THE AMERICAN NATURALIST [VoL.XLVHI

A priori it seems improbable that less variety will be found
among the various internal systems of organs than in the in-
tegumentary or exoskeletal parts. But an argument may be
put forward that the external parts in immediate contact with
the environmental forces would be the first to register change in
the modification of a species. The internal parts as stanchions
and bulwarks remain firm to give characters to orders and fami-
lies, while change makes assault without and gives characters
for species. For example, among the Asteroids it is said that
the internal organization is so uniform that the only method of
classification is to take the different ways in which the demands
of the external environment have been met.

But generally speaking a species depends for its survival not
simply upon the external front it presents to its environment.
An animal's form cannot arbitrarily be divided into external
and internal parts. It is an integral whole, and variation and
selection may occur anywhere, while the correlation of variation
is a text-book commonplace. As opposed to correlative variation
there is the law of the independent variation of parts. Not only
may variation occurring in one part cause a variation to take
place in another, but variation may take place independently in
some areas and be limited in another, so that in deciding upon
the comparative value of the internal and external parts in any
group consideration must be given to both these laws. In the
Asteroids, we assume that anatomists have taken the matter in
hand and found that the external parts vary as a rule independ-
ently of the internal which remain constant. But in how few
groups has such a precaution been taken I Is it not rather the
general rule simply to assume that the internal parts lack varia-
tion and are of no value systematically, as, for instance, in the
Lepidoptera, where the Lepidopterists expect that a classification
based upon the wing-markings or upon wing-neuration can ex-
press the true relationship of the various units!

Even in those groups where systematists have dissected and
found the internal parts valueless it still remains necessary, in
view of the law of independent and unexpected variation of
parts for them, to apply the scalpel to every new form.

It is impossible to deny that the external parts are often of
extreme systematic importance — ^they are exposed to the light
and develop color patterns (although color is usually an unsafe
guide if taken alone), and the external parts of such forms as

Digitized by

Google

No. 570] TAXONOMY AND EVOLUTION, 381

Arthropods and Molluscs being hard provide syptematists with
a sculpture on which it is easy to detect minute diflferences in
pattern. On the other hand we would remind the conchologist
that the external parts are by their very positions most liable
to exhibit lesions and weathering, and certainly in the case of
Mollusca where the dependence of the exoskeleton upon a spe-
cific article of diet (viz., lime salts) is very close, to register
** fluctuating variation" according to the constitution of the
medium or of the food ingested.

But here again if a more common practise were made in dis*
secting by systematists, variations would be found even in closely
allied species making the descriptions complete and often
even necessitating the erection of new genera. One of the
writers was dissecting an ordinary species when he discovered
that the epipharynx was so entirely different in form and struc-
ture from the usual type for the genus that, had it been an
external character it would long ago have been formed into a
new genus.

Karel Thbn^ has demonstrated how in HolothyricUB a single
internal structure is at variance with the other indications of
genetic aflBnity. A great many similar instances will be immedi-
ately called to mind by those who practise dissection.

Again, if systematists are convinced of the taxonomic value
of hard parts how comes it that they need to be reminded that
there are hard parts in the internal anatomy as well which they
so frequently and habitually leave unnoticed! The endoskele-
ton of Arthropods, gastric mills, pharyngeal ossicles and carti-
laginous supports are all systems which might be profitably
studied by the entomologist and carcinologist, while the con-
chologist generally proceeds as though the radula and jaw were
part of the **mush,'' as he so inelegantly terms the viscera.

Geographicali Distribution

The advent of the morphologist into the particular sphere of
systematics or the metamorphosis of the systematist into a mor-
phologist (it matters not how we put this desirable event) will
result in .the annexation not only of classification, but also of
questions of geographical distribution by anatomy and morphol-
ogy. How many pretty theories in geographical distribution

iZool lahh., Bd. XXIII, Syst., pp. 720-21.

Digitized by

Google

382 TRE AMEBIC AN NATURALIST [Vol-XLVHI

have collapsed. because they were built on the sands of an in-
correct classification! The similarity between the faunas of
South America and Madagascar is supported by many facts,
but the value of Solenodon in Cuba and Centetes in Madagascar
has been lessened by the recognition that the two genera re-
semble each other by convergence, and should now be classified
in different families.

The Dendrobatinse also are considered by Dr. Gadow as an
unnatural group, the two divisions — South American and Mas-
carene — ^having, according to him, lost their teeth independently.
Again, Dr. Gadow refers to the Ratitae as a heterogeneous as-
semblage of birds which is "absolutely worthless" for the
zoogeographer. There are scores of such artificial groupings —
the work of the systematist — which have led zoogeographers
astray.

The result is that systematic work as at present pursued is
of very little use to us in the study of geographical distribution.
It is hopeless nowadays for a zoologist to sit down with a
list of species and their range and trusting implicitly in sys-
tematic work to make maps of distribution and, as he so often
does, to draw deductions therefrom, for the validity of such de-
ductions must ultimately depend upon the anatomical and mor-
phological data. Moreover the study of geographical distribu-
tion is developing new methods of tackling its problems.

We do not consider it necessary to touch on the other remedies
that might be applied with a view to redeeming zoological taxo-
nomy from its present artificial state and to bringing it into line
with the rest of biology.

Such remedies — for instance, testing the validity of species
by genetic experiment and the intensive study of variation —
have been advocated many times before,* although with little
success. "We believe, however, that the reforms in descriptive
zoology we have advocated above are the more urgent.

«Cf. E. B. Poulton, '* Essays on Evolution," 2. ''What is a Species! "
and K, Jordan, **Novitates Zoologicie,'' 3, 1896.

Digitized by

Google

SHOETEE AETICLES AND DISCUSSION

NABOURS'S GRASSHOPPERS, MULTIPLE ALLELO-
MORPHISM, LINKAGE AND MISLEADING
TERMINOLOGIES IN GENETICS

In a review of Nabours's breeding experiments with grass-
hoppers,^ Mr. Dexter makes a distinction between an interpreta-
tion of Nabours's and his own, where I fail to see a difference ex-
cept in terminology. This is so typical of much recent Mendelian
work that I am tempted to call attention to it.

Nabours describes a cross between a female with characters BI
and a male with characters CE and comments on the production
of an individual with characters BEI, He says, as quoted, that
the ''female parent gave at least one gamete containing the fac-
tors for the patterns of both her parents (B and I) and that this
double character gamete was fertilized by one of the E gametes
which came from the CE male.**

Dexter prefers to call the supposed exceptional BI gamete of
Nabours Bcel, and the supposed E sperm which fertilized it hcEi,
stating that Nabours 's terminology would involve multiple allelo-
morphism, his own linkage. (Nabours uses, I think, neither ex-
pression.) Now what is the difference between the two interpre-
tations! Is it anything but verbal t Is there anything significant
in the small letters which Dexter has added to Nabours 's form-
ulae 1 If so, what is their significance t Do they mean any more
than the extra zeros in the expression 1.000 as compared with 1.0 1

Dexter proposes an experimental test, that the cross be re-
peated. '*If then BEI forms should appear again and in these
when mated to other forms the factors B and I should be found
to stay together to the same extent as they before separated, it
would show that close linkage, rather than multiple allelomorph-
ism explains this particular instance.*' How would it show it!
If we take Nabours *s assumption that B and I have exceptionally
gone into a single gamete and formed with E a zygote BI,E,
would it be counter to his assumption that they should subse-
quently hang together and that gametes should arise BI and E,
respectively? Would adding a few small letters to the formulae
1 Am. Nat., May, 1914.

383

Digitized by

Google

384 THE AMERICAN NATURAL

alter the case, changing it from multiple alle
age ? It seems to me that this is one more ca
ious conclusion is reached in consequence oJ
for absent characters in Mendelian formulfi
Wilson has pointed out others.

BussEY Institution,
Forest Hills, Mass.,
May 6, 1914

Digitized by

Google

Two Important Biological Journals

The Journal of Genetics

Edited bj
W. BATESON and R. C. PUNNETT
A periodioal for the pablioation of records of original research in heredity, variatioiti and allied topics.
The JoUBMAi. also contains, tiom time to time, articles summarizing the existing state of knowledge in the
▼arions branohes of genetics. The numbers are fully illustrated, and, when the subjecfc-matter demands it,
eolored plates are freely employed. The Journal is issued in parts as material accumulates, and a volume
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7-50 a volume ;
tingle oopiea, 52.50.

3!ometrik:a

Edited by EABL PEABSON
A journal for the statistical study of biological problems, appearing about four times a year. A volume
containing about 600 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
nine preceding volumes fonn a veritable encyclopaedia of biological statistics in the fields of anthropometry,
craniometry, medidne, heredity, etc. Subscription price, $7.50 a volume ; single copies, 12.50.

Prices <m book numbers of these journals vary, and postage from London is charged on back volumes
and single csopiee.

Published in America for the Cambridge University Press of England by

The University of Ohicasro Rress
Chlcasro, Illinois

University Control

By J. MoEjeen Cattbll, Professor of Psychology in CJolnmbia Uniyersity

Tocstlier with a series of Two Hundred and Ninety-nine Unsigned Letters by Leading Men of Belsnsi
boldins Acadnmie Positions and Articles by Josbpb Jasxbow, Gbobob T. Ladd, Jomi J. Svavuraov, J. B.
GunoBvov, J. MoEbbn Gattkll, GnoBoa M. Stbattov, Stbwabt Faxoh, Johoi Jat Chapmaii, jAioi P.
MvMMom and Jaoob Govxj> Schubican.

A great yariety of questions conoeming general uniTsraity administration are dealt with in an original and
Mpful way^ — Naturt.

Tbeeo quotations and examples are taken from Professor Gattell's informed and thorough disousslon of ths
SBbJeet of oniveraity control, a subject upon which he has had much to say of late, finding occasion for caustis
fritidsm of existing Ameri -an conditions, and standing as the champion of an academic democracy and a teach-
ing imrfcMlon iq)on wtdch a man may enter without forfeiting his self-respect. — The DiaL

Sentences and paragraphs that betoken the expert, highly-trained mind, the suggestions that come to va-
fresh a d tflQ us tiiat a new day is about to dawn in educational writing. — Th$ Boiion Bvemng TrameripL

SCIBINCE AND EDUCATION

A Beriem of volumes for the promotion ofacientiBe reaearcb and edncAtional progress

VOLUMB I. The Foundations of Science. By H. PoiNoudl. Gontaining the authorised English translatioa
fay GaoBaa Bbuob HAiaran of "Sdenoe and Hypothesis," "The Value of Sdenoe," and "Srienea
and Method." Price, $3.00 net.

VOLUBiB n. Medical Research and Bdncation. By Bighabd M. Pbabob, Wiluam H. Wbloh, W. H.
HowBLL, Fbaxxlih P. Mall, Lbwbllts F. Babkbb, Cbablbs S. Mxhot, W. B. Cannon, W. T,
CovTSOEUCAa, TBBOBALn SiinRB, O. N. Stbwabt, C. M. Jackson, K P. Lton, Jambs B. Hbbbiok, Johh
M. DonsQN, C R. Babubbn, W. OphOls, S. J. Mbuisbb. Jambs Ewxno, W. W. Ebbn, Hbnbt H. Dovald-
BOir. CHBiBnAH A. Hbbtbb, and Hbnbt P. Bowditgh. Price, 13.00 net.

VOLUMB HL TTnlravlty Control. iTotofiaidy. Paom9+484. Price, ISMutL

OARRi^oN, N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION M, NEW YORK CITY

Digitized by

Google

The American Naturalist

A Moatiar Joanml MtoblklMd in 1867. Devoted to tke AdwwuemmutA of fho Biologieal
with SpocUl Rof orooeo to tlio Foelon of Or«»aic ErotuOan and Horeditj

CONTENTS OF THE DECEMBER NUMBER

Tho Fixation of Chftnctw in Oxganiimi. By Edward

Sinnott.
Inheritance of Lefl-handedneM. Prof etior Frandt

Samaley.
Bapplementary Studlet on the Differential Mortality

with Reipect to Seed Weight in the Gennlnatlon

ef Garden Beam, n. Dr. J. Arthur Harrii.
Shorter Articlee and DiiooMion : A Croat inyolTing

Four Pain of Mendelian Charaeteri in Kiee. 0.

C. Little. J. C. PhilUpe.
Index to Yolnme XLVIL

CONTENTS OF THE FEBRUARY NUMBER

Some New Yarietiee of Rati and Gninea-piffi and their

Relatione to Problemi of Color Inheritanoe. Pto-

feMor W. E. Castle.
•* Dominant*' and " ReceMire" Spotting in Mice. C.

C. Little.
On Differential Mortality with respect to Seed Weight

oeonrrlng in Field Cultures of Pisum satlTum.

Dr. J. Arthur Harris.
The Inheritance of a Recurring Somatic Variation

in Variegated Ears of Maise. Professor R. A.

Restoration of Bdaphosaurus eruciger Cope. Pro--

fessor E. C. Case.
Shorter Articles and Discussion: Humidity— a

Neglected Factor in Enyironmental Work. Dr.

Flank S. Lutz.

CONTENTS OF THE JANUARY NUMBER

A Genetic Analysis of the Changes pcoduocd by
Selection in Experiments with TolweoQ. Pto*
fessor E. M. East and H. E. Hayes.

Gynandromorphous Ants, described during tte De*
cade, 1908-1918. Professor William Morton
Wheeler.

Shorter Articles and Discussion : On ttie BegaHs ef
Inbreeding a Mendelian Population— A Cotiee
tion and Extension of PiCTious ConclusioDs.
Dr. Raymond Pearl— Isolation and
allied in Principle. Dr. John T. Guliek.

CONTENTS OF THE MARCH NUMBER

The Effect of Extent of Distribution on I

Asa C. Chandler.
Biology of the Thysanoptera. Dr. A. Franklin SkoIL
Shorter Articles and Correspondence : The •g»vV"«i>

Mammals of the British Isles. Profeesor T. D. A.

OookerelL

Notes and Literature : Swingle on Varlatton in Wi
Citrus Hybrids and the Theory of Zygotazis.
Dr. Orland B. White.

CONTENTS OF THE APRIL NUMBER

The Origin of X Capsella Bursa pastoris arachnoidea.
Dr. Henri Hus.

Biology of the Thysanoptera. II. Dr. A. Franklin
Shull.

Shorter Articles and INscussion : Barriers as to Dis-
tribution as regards Birds and Mammals. Joseph
Grinnell. Yellow Varieties ol Rats. Professor
W. £. Castle.

Notes and Literature : Heredity aad the Influence
of Monarchs. Y. L. K.

CONTENTS OF THE MAY NUMBER
Eetoparasitesof Mammals. Profenor Vernon Lyman

Kellogg.
Regeneration, Variation and Correlation In Thyone.

Professor John W. Scott.
Shorter Articles and Discussion : Tenns ralntfng to

Generic Types. Dr.O.F.Oook.
Notes and Literature: Linkage in the Rilkwofm

Moth. A. H. Sturterant. Nabours*s Breeding

Bxperiments with Grasshoppers. JohnaDextsr.

Single Number 40 Cents
TIm HATURALIST wfll b#

Yeariy Subtcrqitiont
for four mositlia for Om DoOir

THE SCIENCE PRESS

GArrisont N. Y*

Sub^Ution 84: NEW YORK

LancastaTi Fm

Digitized by

Google

▼ox. XLVm, HO. 671 JVLT, 1914

THE

AMERICAN
NATURALIST

A HOHTELT JOXnUTAL

Devoted to the Advancement of the Biological Sciences with

Special Beference to the Factors of Syolution

COHTEVTS

Page
I. Pattern DeTdlopment In Mamwali and Birds. Qloybb M. Allen - - 886
IX. Inttmal BelaUoni of Terrettrlal AsBOolattona. Abthus G. Vbstal - - 418

m. Sliorter Artlolei and Diieawilon : Another Hypothetis to Acoonnt for Dr.

Swingle's ExperimentB with Citrus. A. G. and A. L, Hagbdoobn - - 446

THE BODSNOE FBEBB

&A1I0A8TEB. PA. OABBI80N. V. T.

HEW TOKK: 8UB-BTATI0N M

Digitized by

Google

The American Naturalist

MSS Intended for piblicatien and booke, ete,, Intended for review ehenld be
eent te the Editor of THE AMERICAN NATURALIST, Garrieon-en-Hndeon, New Yerk.

Short artiolee eontaittlng einmiariee of reeeareh work bearinf on the
preblenie of organle evelntlon are eepeeially weleonie, and will be fiven preferenoe
in pnblieatien.

One hnndreo reprlnte of eentrlbntione are enpplled to aatbore ft*ee of ebarie.
Farther reprlnte will be eupplied at ooet.

Subeerlptione and advertieeniente ehenld be eent to the pHbliehere. The
enbeerlptlon prlee ie four dollare a year. Fereion poetage ie iiny eents and
Canadian poetafo twenty-Uve oente additional. The ehargo for einfle oeplee Ie
forty oente. The advertieinf ratee are Four Dollare ftor a page.

THE SCIENCE PRESS

Lancaster, Pa. Qanison, N. Y.

NBW YORK : Sub-Station 84

Xnttred m Moond-dmit matter ,

, April 2, 1906, At the Poit Offlot At lAneaiter, Pa., nnder tb« Aet of
ConsMu of Mareh S, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Enflrland

WANTBD TO PURCHASE

a tet of BIRDS OF AMERICA by J. J. Audubon.
7or 8 volumes, please report, atatjng cash price, stat-
ing condition, binding and dates of volumes.

P. C. HARRIS,
Box 2244 Boston, MaesachnsetU

For Sale Entire

An important collection of Indian Birds'
Eggs, containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste
viridanus), Red-browed Finch (Calla-
canthis burtoni), Tibetan Sand-Grouse
(Syrrhaptes tibetanus), Spotted Sand-
Grouse (Pteroclunis senegallus). Ibis-
bill (Ibidorhynchus struthersi), White-
bellied Minivet (Pericrocotus erythro-
pygius), Bearded Vulture (Gypaetus
barbatus), numerous Cuckoos with their
hosts, etc., etc.

Particulars and full list may be had
from

W. F. H. ROSENBERG

57 Haverstock HiU London, N. W.

Photographic Films Developed

5 cants per Spool, Prints 3 cents up. Post

Cards 5 cents.

8x10 Enlargements 25 cents mounted.

Send for Samples and Price List.

FRANK R. BARBEAU
89 West Bridge Street Oswego, N. Y.

The University of Chicago

Oflfen imlzuctiao during theSoBi*
met Quarter on die aune bttsii «t
during the qUmt qoartcn ot Che
academic year.

The undeiKiadiiate coD^gca. Cha
giaduate achoola, and the prafes-
aiooal schools proride cooraes in
Art», Liimrmimr;

Hmu Lam, MmMdmrn, Bdmem-
Hon, omdDipimity. Instnictiaa
h nven hy lenilar memhen of tfar
UnivCTsityataflwhichkaugmrBtffd
b the summer by appoinaieatof
profesaon and instmcton fnsB
other institntiom.

•. 1914

lal Tana JttM 1 6-Jdb a
f 2dTanBMF2S-Aue^a9
Detailed annomucmenla «9 ba
aent apoo appliatka.

TheUeirew^tyolCycegs

Mitchell Tower

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVni July, 19U No. 571

PATTEEN DEVELOPMENT IN MAMMALS
AND BIRDS

GLOVER M. ALLEN,
Boston Society of Natural History

The particular coloring of mammals and birds is pro-
duced by two factors — pigmentation and the physical
structure of the hair or feathers. Both are often present
together. In certain mammals, for example the golden
mole (Chrysochloris) and the European Galemys, a
beautifully iridescent sheen is produced by the reflection
of light rays having a certain angle of incidence upon the
hairs which themselves contain pigment of a character-
istic color. In the duckbill (Ornithorhynchus) the same
thing is found. The peculiarity of feather structure that
causes iridescence is largely developed in certain families
of birds, as the hummingbirds and the pigeons (see
Strong, 1904, for an account of the feather structure).

It is not my purpose to discuss the use of this irides-
cence to the bird, beyond stating my belief thkt it is in
part at least for sexual display, as no one can doubt who
watches the male street pigeon strutting before his mate.
With amorous coos and lowered head, he confronts her
and, swelling out his throat feathers, turns about and
about, so that the light is reflected from his neck and
throat in a sparkle of rainbow hues. It has also been
suggested (Thayer, 1909) that iridescence may be a
strong factor in concealment, since from the variety of
the colors produced the bird is more diflScult to resolve
from its many tinted environment amid foliage and
flowers.

385

Digitized by

Google

386

THE AMERICAN NATURALIST [Vol. XLVIU

With many birds the characteristic coloration may not
be at all that of its pigment. Thus the blue of the male
indigo bird (Passerina cyanea) is due solely to the phys-
ical structure of its feathers which though pigmented
with brown, appear blue by reflected light. If, however,
a blue feather be immersed in oU and viewed under a
microscope by transmitted light, it is seen to be broivn-
pigmented. The physical feather-structure of the adult
male is thus in this species a secondary sexual character
chiefly developed during the breeding period.

The important point at present is, however, that the
color effects just described are none the less due to pig-
ment, quite apart from the fact that the apparent color
of the pigmented area may be different from the actual
color of the pigment (except that iridescence may some-
times be faintly seen in an unpigmented feather).

The use of pigmentation to its possessor is a matter
still under discussion and investigation. In many cases
it is doubtless the result of purely physical causes and
it is quite without the power of the animal to make use
of its coloration for outward effect. Thus the beautiful
colors inside the shells of some molluscs are never appar-
ent from an exterior view, and are supposed by some to
be in part a waste product, the result of metabolism
within the organism.

The present discussion has to do only with the external
pigmentation of the hair and feathers, respectively, in
mammals and birds.

The simplest cases of coloration are those in which the
body or its covering is everywhere of the same hue, or
nearly so — as in the elephant, the wild buffalo, or the
house mouse in which the hairy covering (or hide in the
elephant) is of a nearly uniform tone everywhere. So
too, the crow, the apteryx, and the nestlings of many
birds whose parents show a more highly differentiated
style of markings. Such mammals and birds, so far as
the development of pattern is concerned, I would con-
sider unspecialized, yet it does not follow that in this

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 387'

respect they are also primitive, though in most cases I
venture to think this 'may be true. The uniformity of
plumage is probably a derived condition in such a species
as the Cuban blackbird (Holoquiscalus assimilis) in which
the duller colored females have yet a yellow patch at the
bend of the wing, a style of marking widespread among
allied forms. The adult males, however, have lost this
and are wholly black. Gadow as well as Keeler (1893)
conclude that among related species in which there is a
tendency to differentiation of the coloring the end result
of the stages through which the species may pass is the
production of a wholly black bird. In general a wholly
black condition is no doubt to be considered as a derived
rather than a primitive state among birds whereas a uni-
formly dull plumage of a brownish or grayish tone is
probably in most cases primitive. Among mammals the
same is probably also true, for in both the black condi-
tion indicates either an excessive production of the black
over other associated pigments, or a loss of the power to
produce the latter, whereas the neutral gray or brownish
coloring is due to a more even mixture of such pigments.
As pointed out by Professor W. E. Castle, the ' kicked ^^
pattern of the hairs of mammals is probably primitive^
and it is certainly very widespread. It is well illustrated,.
for example, by the house mouse {Mus musculus) or the
wild guinea-pig (Cavia), in which three separate pig-
ments occur as granules in the individual hairs — yellow,
chocolate, and black. These three in their normal mix-
ture produce a neutral gray tint — mouse color — and an
examination of this type of coat usually shows that some
hairs are wholly black, others dark at base barred with
black and yellowish near the tip.

There are two ways in which patterns may be developed
from a uniformly tinted covering of hair or feathers : (1)
by a local change in the relation of the associated pig-
ments so that in certain areas only one or two sorts are
produced instead of three, or only one ; (2) by a failure to

Digitized by

Google

888 THE AMERICAN NATURALIST [Vol. XLYIII

develop pigment at all in certain places, so that a white
or unpigmented area is produced.

It is not rare among mammals to find that one or more
of the characteristic sorts of pigments are not produced
in certain individuals and probably the factor or factors
for these are lost altogether from the somatic and sex
cells alike. Such variations may be perpetuated through
inbreeding and so no doubt have arisen sundry domestic
color varieties of animals and plants. For example, in
the course of experiments with color varieties of the
house mouse (carried on some years since with Professor
W. E. Castle) we found that the chocolate-colored mice
which we bred as extracted recessives from black mice,
contained only chocolate pigment in their hair, whereas
in the black parents both black and chocolate pigments
were present, but the black masked a chocolate pigment.
Moreover, the chocolate mice always bred true to that
color, but if bred back to the black parents, gave black
young or both black and chocolate in Mendelian propor-
tions, according to the nature of the matings. The inter-
esting point here is that the chocolate mouse once pro-
duced, through the loss of its black-and-gray-pigment-
potentiality, can transmit no other pigment character but
the chocolate. What causes the occasional production of
an individual in which one or more of the characteristic
sorts of pigment is absolutely lacking is still unexplained.
Nevertheless it is of frequent occurrence not only among
domesticated species, in which the natural conditions of
life are so greatly modified, but also in species in a state
of nature.

A skunk normally marked, but chocolate instead of
black, a raccoon likewise of normal pattern but the pig-
mented areas yellow, are merely examples of the drop-
ping out of the factor for black pigment from the normal
combination of the two. Such specimens are of occa-
sional occurrence, and examples are in the museum of the
Boston Society of Natural History. Similarly are pro-
duced red woodchucks or muskrats, or wholly yellow field

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 389

mice (Microtus). Melanism commonly results through
an excess of black pigment which may mask a second
pigment. Thus the black hairs of the black variety of
fancy mouse commonly contain a considerable amount of
chocolate pigment as well, and so of the hairs of the
black-appearing skunk. A black mouse thus does not
contain the yellow pigment, while the chocolate pigment
is largely masked in general view by the black. In other
cases it may be that black pigment alone is present.

It is probable that many cases of dichromatism among
animals are explicable as similar cases in which one or
other of the pigments normally present becomes to a
greater or less degree inactive. Thus red forms of certain
blackish or dull-colored bats {e. g., the small Molossus of
Cuba) are apparently the result of the dropping out of
the factor for black pigment or its great reduction. The
red and gray phases of the screech owl {Otus asio) are
probably also explicable as a similar phenomenon.

It is only when this inactivity of one or more of the
pigment factors occurs locally on the body that a definite
color pattern is produced, in which neighboring areas of
the body are of contrasting hues. As an example may be
cited the variegated guinea-pigs, whose monotone ances-
tors are still abundant in a wild state in South America.
Professor Castle, through his studies of these patterns
in guinea-pigs, first suggested to me in 1903 that there
were definite areas of the body which, though contiguous,
are independent of each other in their pigment-producing
capacity. In this suggestion lies the key to the chief
investigation of this paper, namely, the defining of these
areas, and a study of their behavior in the development of
pattern by the second of the two methods previously
given — that is, through the failure of pigment to develop,
so that white or colorless areas result. This condition of
partial albinism is not uncommon among animals which
in their normal condition are completely pigmented. In
domestic species it is very general and in them tends to be
preserved. It also occurs normally in the shape of defi-

Digitized by

Google

390 THE AMERICAN NATURALIST [Vol. XLVHl

nite white markings in the patterns of many mammals
and birds. Magazines of natural history abound with
instances of total or of partial albinism among mammals
and birds, either of domesticated or of wild species.
Some writers have even recognized the fact that such
white markings tend to occur in certain parts of the
body, as at the tip of the tail or on the forehead. Darwin
speaks of the white forehead spot or star, and the white .
feet so common among horses, and implies that such
markings must be of some significance. His statement
on hearsay that white-marked horses are more suscep-
tible to poisoning from noxious herbs is, however, un-
corroborated. In 1882, W. H. Brewer gathered a number
of statistics as to the presence of white marks in horses
and cows, but reached no conclusion. He could find no
necessary correlation between the presence or absence
of white spots in forehead and feet, though it appeared
that white marks might be more frequent on one side of
the body than the other. But the tentative conclusion
that such animals habitually reclined on the side showing
the more white, is begging the question.

As briefly stated in my paper of 1904, the important
thing is not that white tends to appear at certain places,
but the converse, that pigment production is more intense
at certain definite centers on the body and the occurrence
of white or pigmentless areas is due to the restriction of
pigment formation at the periphery of these centers, so
that white occurs at their extremities or as breaks b<^-
tween contiguous color patches.

In mammals and birds these centers are typically five
on each side of the body, and a median one on the fore-
head. They appear to be homologous in both groups,
though in different species they show varying degrees of
modification in their behavior and development. When
a reduction of the pigment areas occurs, the appearance
is as it were a shrinking of the particular color patch
toward its definite center. The reduction may vary to
any degree, from that condition in which the break

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 391

between two adjacent patches is merely indicated by a
white streak to that in which it is reduced to a small spot
of pigment, or to zero, when the entire patch drops out,
leaving a white area. These patches are wholly independ-
ent of each other in the extent to which they may be
developed, so that a particular patch may be quite want-
ing on one side of the body, while its. fellow of the oppo-
site side is completely developed. Nevertheless, there is
often a marked tendency to bilateral synmietry in such
reduction. From a study of partial albinos in which the
pigment reduction is considerable, the location of the
ultimate centers of these patches becomes possible as well
as the determination of their normal extent. I have
studied several domesticated species in which white
marks are common, with the results briefly detailed below.
When all the centers are fully developed the animal is
completely pigmented; when none is developed, it is a
total albino. Between these extremes may be found every
conceivable degree of development. In an ideal case in
which each center is slightly reduced so as to be circum-
scribed by white, the animal would have a dark coronal
or crown patch and a series of five patches on each side
separated by a median dorsal and a median ventral
stripe. The anteriormost of the lateral patches center at
the base of each ear, and each in its greatest development
covers the side of the head from muzzle to behind the
ear. These I have called the aural or ear patches; the
next posterior are the two neck or nuchal patches each
of which pigments its proper side of the neck, and extends
from behind the ear to the shoulder and anterior edge of
the foreleg. When much reduced the patch, as it were,
contracts to a smaU area on each side of the neck, varying
slightly in its location among different species. Posterior
to these come the scapular or shoulder patches one on
each side of the body. Each pigments the shoulder area
and foreleg, except (usually) the front edge of the upper
part of that member. This patch shows interesting slight
variations in the extent over which it spreads in different

Digitized by

Google

392 THE AMERICAN NATURALIST [Vol. XLVHI

species. Centering nearly at the lower part of the back
are the pleural or side patches, each of which pigments
the area from the shoulder to the lumbar region and ante-
rior part of the hind leg of either side. Last of all, the
two sacral or rump patches, each of which on its respect-
ive side pigments the buttocks and tail. In most species
these two patches are so closely associated that they tend
to remain fused dorsomedially, so as to give the appear-
ance, when reduced, of a single median patch at the base
of the tail. Their frequent bilaterality, however, indi-
cates the dual origin of such median patches. Each of
the lateral patches in its complete development extends
from the mid-dorsal to the mid-ventral line or those of
opposite sides may overlap slightly. Reduction usually
first appears mid-ventrally.

It is probable that the retinas should also be considered
as an additional pair of patches, since morphologically
the eye is of dermal origin, and there is sometimes seen
a tendency to the formation of a small circumorbital
patch, which appears to break from the ear patch when
this is largely reduced.

Pocock (1907) has pointed out that in black-and-tan
dogs the tan appears about the muzzle, along the sides
and on the limbs, while the blacker portions are more
dorsal. It may be added that in tricolor hounds, in which
the several primary patches are reduced, these are often
tan color at their several peripheries and black centrally.
In both cases, the explanation is simply that pigment
formation is less intense the farther away from the pri-
mary centers.

The reason of the division of the body surface into
these independent areas of pigmentation does not here
concern me. It is no doubt the result of physiological
causes, and it is rather suggestive that the several patches
correspond externally to important nerve centers or
groups of nerves. Thus the eye pigment corresponds to
the optic nerve, the aural patch to the auditory nerve, so
that these two great external sense organs of the head
have each their corresponding pigment patch. The neck

Digitized by

Google

No. 571]

PATTERN DEVELOPMENT

393

patch corresponds with the group of cervical nerves, the
shoulder patch with the brachial plexus, the side patch
with the nerves of the trunk, and the rump patch with the
sacral plexus. It may be further suggested that the
median crown patch of the head corresponds to the pineal
eye, a suggestion that is strengthened by the fact that it
is more or less obsolete in mammals, just as the pineal
gland is vestigial, whereas in birds, which are more
reptilian in structure, the patch is usually well defined.
At all events it is a median unpaired structure, as are the
pineal and the interparietal bone.

Turning now to a more detailed consideration of these
pigment patches in sundry species of animals, we may
first examine a series of diagrams (Figs. 1-15) of the

Figs. 1-7. Diagrams Illustrating Pigmentation in the Domestic Dog.
Figs. 8-15. Diagrams Illustrating Pigmentation in the Domestic Dog.

Digitized by

Google

394 THE AMERICAN NATURALIST [VoL.XLVin

domestic dog, all of which are carefully drawn from
photographs or from living animals, and are selected
from a great number to show various conditions in the
reduction of the pigment patches. In these and the other
diagrams the black portions represent pigmented areas,
irrespective of the actual colors.

For convenience I have called the white stripes demark-
ing these chief or primary patches, *' primary breaks/'
since thiey are the first indications of a decrease in pig-
mentation such that two adjoining patches no longer
meet. Secondary or further breaks result in a general
disintegration of these primary pigment patches, and are
apparently more irregular in nature, though often they
follow certain fairly well defined lines. The first of the
primary breaks generally occur as white patches on the
chest or belly, about in the median line. These are not
shown in the dagrams, but in most cases should be under-
stood as present. In Fig. 1 the pigment areas show a
beginning in reduction. The two aural patches have
become separated and their failure to spread to the
normal limit in the median line has resulted in a white
nose stripe. A short transverse white marking indicates
a separation of the neck patch at its anterior edge from
the ear patch. Elsewhere the various patches are contig-
uous; but the extremities of th6 limbs and tail are pig-
mentless, as if pigment had failed to spread to the tips of
these members in its reduction. In Fig. 2 the same
primary break between the ear patches is present, and in
dogs it is one of the first and most frequent to appear.
The sam^ shrinkage of pigment from the extremities is
also seen. The neck patch of the left-hand side, however,
has completely dropped out, and its fellow of the right-
hand side is reduced posteriorly so that it fails to reach
the shoulder patch. Thus a white collar is formed. It
is also interesting to see that at its anterior end a distinct
constriction is present where the neck patch joins the ear
patch of the right side. Fig. 3 shows a somewhat similar
condition but the neck patch of the right side as well as

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 395

the ear patch is missing, while those of the left side are
fnlly developed. In Fig. 4 both neck patches are missing,
so that a white collar is formed. In dogs the neck patch
is usually the first to drop out altogether, so that a white-
collared dog is of very frequent occurrence. In fox
hounds this patch is shown unusually well, either wholly
or partly separated from neighboring patches. The sep-
aration of the ear patches, wholly or partially, so as to
produce a white blaze or line in the middle of the forehead
is about as frequent. In Figs. 8 and 12 a single neck spot
only (as it happens, in one on the right, in the other on
the left side) is still present but so slightly developed as
to be only a small island of pigment wholly separate from
the neighboring patches.

The crown spot is so often present in dogs as a little
oval island, always on the top of the head about in line
with the anterior bases of the ears (Fig. 4) that I am
convinced it is a primary patch. It is common in bull
dogs and bull terriers, and in other breeds is often seen
but is so commonly not indicated at all, that it seems
probable it is becoming lost, and its area is filled by the
ear patches, since these are often separated by a very
narrow median line only, which, as in Fig. 13, may con-
tinue posteriorly to separate the two neck patches
medially as well. In other cases (Figs. 1, 6) the failure of
the white nose stripe to extend farther posteriorly may
be due to the persistence of this patch.

The demarcation of the side from the rump patches is
indicated by the imperfect primary break across the
lower part of the back in Fig. 4, while in Fig. 5, a similar
primary break farther forward indicates the limits of the
shoulder and side patches. In each case the break is
incomplete transversely, with a narrow isthmus near the
median line. In dogs there is a marked tendency for the
ultimate centers of the side and rump patches to be close
to the median line, so that the corresponding patches of
opposite sides are confluent dorsally. This is especially
the case with the rump patches, with the result that it is

Digitized by

Google

396

THE AMERICAN NATURALIST [Vol. XLVIH

very rare to see the two rump centers separated, but
instead, as in Figs. 10 and 14, they appear, when much
reduced, as a small median spot at the root of the
tail. That they were originally paired, there is no doubt,
as there is frequently (as in Fig. 9) a deep median
notch indicating the median primary break between the
centers, or (as in Figs. 11, 12) one of the lateral centers
drops out, leaving its fellow of the opposite side. The
continued union of the side patches with the shoulder
patches is seen in Fig. 7, while in Fig. 9, though the union
is still present between these patches of the right side, on
the left side the shoulder patch has failed to develop, and
the side patch is so reduced that it does not meet its
fellow of the right. In Fig. 8 both shoulder patches are
present more or less bilaterally equal, and, as frequently,
are produced into narrow tongues on to the upper arm.
The two side patches in Fig. 8 are also reduced, so as to
be wholly separated from each other and from the neigh-
boring centers. They are further interesting in being
placed nearly median one behind the other instead of
nearly opposite. In Fig. 11, on the other hand, they are
far sundered, but this, in dogs, is a much less usual con-
dition. In Fig. 10 a single median dorsal patch repre-
sents the slightly developed side patches, but whether
this single patch corresponds to one or other of the two
centers, or whether the two are actually fused in the
dorsal line, I can not yet say.

The shoulder centers, when slightly reduced, are large
in dogs, and cover a considerable saddle-shaped area, as
indicated in Fig. 5, from near the center of the back for-
ward including the fore leg and part of the fore shoulder.
When further reduction takes place the pigment is drawn
away from the extremities and the saddle separates from
the neck patch (Figs. 2, 6) and then from the side patch
(Figs. 5, 9), and finally the shoulder patches separate
from each other (Fig. 8). One or other of the shoulder
patches may drop out entirely (Fig. 10) or be reduced to
a very small spot (Fig. 12) at what may be considered

Digitized by

Google

No. 571] PATTEBN- DEVELOPMENT 397

the ultimate center of the pigment patch, near the upper
part of the body, near or just back of the shoulder. The
ear patches seem to be the last to disappear, and these,
too, may be variously reduced or only one may be present
(Fig. 15). The approximate outlines of the patches when
fully developed are indicated by dotted lines in Fig. 15,
in which 1 is the crown patch, 2 the ear patch, 3 the neck
patch, 4 the shoulder patch, 5 the side patch, and 6 the
rump patch.

In dogs, there is seldom seen any tendency for these
primary patches to divide. . What has the appearance of
such a tendency is seen, for example, in the coach dog,
which is rather evenly flecked with rounded black spots,
with often in addition, black ears and more rarely reduced
rump patches. Fig. 9 shows such a dog in which both ear
patches, one shoulder, both side and both rump patches
are sharply indicated, though reduced. In addition there
are present on the white body areas between, many small
flecks of dark color, evenly distributed, which are clearly
not islands separated from the primary patches. Indeed
this spotting seems to constitute a wholly different cate-
gory of pigment formation, in addition to that of the
primary patches, which latter I have called ''centripetal"
pigmentation. As Professor Castle suggests to me, it is
probably homologous with the ''English" marking or
spotted condition of domesticated rabbits, and possibly
the dappling of horses is a similar phenomenon. When
these spots and the primary color patches are of the same
hue, it is not possible to distinguish the two in visual
appearance, unless the latter are reduced areally, when,
as is sometimes the case in the coach dog, one or more of
the primary patches is seen with the spots, as it were,
proliferating from its edge. This second element no
doubt enters as a factor in the color pattern when the
small spots are of a different color from that of the
general body surface, as in case of the cheetah {Cynce-
lurus) or the leopard and jaguar.

I am inclined to think that the excessive breaking up of

Digitized by

Google

398

THE AMERICAN NATUBALIST [Vol. XLVIU

the primary patches, to be considered under the cow, is
not a wholly similar phenomenon.

Five diagrams illustrating the domestic cat are shown
in Figs. 16 to 20, and are interesting to contrast with

Figs. 16-20. Diagbams Illustbatinq Piombntation in the Domestic Cat.

those of the dog, also a carnivorous mammal. The
demarcation of the primary patches is usually less sharp
than in dogs, but is in general similar. The most comnion
appearance is where the primary breaks occur in the
mid-line below, giving a white throat, chest or belly; or
the separation of the aural centers produces a white
streak on the nose or extends it up between the ears.
The ear patches in Figs. 17, 19, 20, show successive reduc-
tion, so that at first the hinder margin of the ears, as in
dogs, becomes white, then with further decrease in pig-
ment production, the inner bases only are colored. The
neck patch has its ultimate center farther back than in
dogs so that when much reduced, it is present as a pig-
mented spot at the very base of the neck or even at the
front of the shoulder (Figs. 16, 17). In Fig. 16 the neck
patch of the right-hand side is only slightly reduced and
is in contact anteriorly with the ear patch, while poste-

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 399

riorly it does not meet the shoulder patch. The left-hand
neck patch, however, is quite separate from the neighbor-
ing patches and is reduced to a small area at the junction
of the neck with the shoulder. It is absent in Fig. 17 from
the left side and is represented on the right side by a
similar small center, placed far back. In Fig. 20 the
neck patch or patches show a reduction to a single small
square median patch at the base of the neck, but whether
this represents a median fusion of the two lateral centers,
or whether one only has persisted and has shifted to the
midline, I do not attempt to say, though the former
hypothesis seems on the whole more probable.

The shoulder patch in house cats is relatively small,
and, as indicated by the indentations in Figs. 17, 18, is of
the fore side of the upper arm, but the shoulder patch
when fully developed seems to cover the rest of the leg
and a smaU scapular area. It is shown much reduced in
Fig. 19, on the right-hand side, and is altogether wanting
in Fig 20. The conjoined shoulder and side patches in
Fig. 18 are shown reduced laterally, so as to form a
broad median stripe which I take to mean that the ulti-
mate centers are closely approximated dorsally. The
neck patch is wholly absent, but both ear patches are
present and joined medially. The sacral patches, as
commonly, seem fused or at least very close together.
There is a small break midway on the tail, which sepa-
rates off a pigmented tip, a phenomenon which I shall
refer to under ** centrifugal pigmentation.^^ The side
patch is long comparatively, and extends forward to
cover the deficiencies of the shoulder patch, as in Fig. 17.
Here the left side patch has been reduced at its anterior
end, and its separateness from the patch of the right side
is indicated by the median indentations. It is often want-
ing in domesticated cats.

The sacral patches, pigmenting the buttocks and tail,
seem to be fused or closely approximated at the root of
the tail, as in dogs. I have seen no instance of the crown
patch being shown in the cat, though such may occur.

Digitized by

Google

400 THE AMERICAN NATURALIST [VoL-XLYIII

The approximate boundaries of the five bilateral patches
are indicated in Fig. 20 by dotted lines ; 2 is the ear patch,
3 the neck patch, 4 the shoulder patch, 5 the side patch,
and 6 the rump patch.

Among domesticated rodents the pigment patches have
been studied in rats, house mice, and guinea-pigs. In
all, the same patches appear except that in rats and mice
the median crown patch appears to be lost, though in the
guinea-pig it is often present. Diagrams of parti-colored
mice are shown in Figs. 21-24, and sufficiently indicate

Fios. 21-24. Diagrams Illustrating Pigmentation in Domestic Varibtibs
OF THE House Mouse.

the primary pigment areas. The white spot on the fore-
head of Fig. 21 indicates a primary break between the
two ear patches, and varies widely in different individ-
uals, from a few white hairs only to a large blaze The
inheritance of such a blaze has been studied by Little
(1914). The white mark at the base of the neck in Fig.
21 indicates the beginning of separation of the neck from
the shoulder patches and perhaps of the two neck patches
from each other, because of its longitudinal extension.
The white band across the neck in Fig. 23, however,
indicates probably only the beginning of a separation of
the neck from the shoulder patches, which in Fig. 24 has

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 401

wholly sundered these two areas, so that a white-collared
mouse results. The condition shown in Fig. 22 is similar,
except that the separation has taken place on the right
side only, between the neck and the shoulder patches of
but one half of the body. A break between the two neck
patches of opposite sides is further indicated in this
figure by the deep median reentrant back of the ears.

In all four diagrams the areal restriction of the
shoulder patches is shown, but in varying degrees. In
Fig. 21, the pigment has not spread to the feet, leaving
these white, and so in the other figures, but to a greater
degree. A median linear break between the shoulders
indicates the restriction of the patches of opposite sides
at this point, which in Fig. 22 is more clearly perceptible.
The posterior limits of the shoulder patch are further
shown in this diagram, by the beginnings of a break
between the shoulder and the side patches. In Fig. 24
this break is no longer interrupted, but clearly separates
the two areas. Further, the side patch has dropped out
on the left. In Fig. 23 an imperfect separation of
patches on the posterior part of the body has taken place.
On the right-hand side the shoulder patch, which in mice
is of considerable extent, has broadly separated from the
side patch, while on the left-hand side a long transverse
break has taken place between the side and the rump
patches, with two island-like white spots between, the
anterior of which probably marks the transverse line of
stress between shoulder and side patches, the posterior
the median line of breaking between the two side patches.
A slight indentation in the pigmented area far back on
the right side of Fig. 22 points to the beginning of restric-
tion between side patch and rump patch. The separation
of these patches by a transverse mid-dorsal break is
shown in Fig. 21, and their complete separation on the
left side appears in Fig. 23 (the transverse white mark),
while in Fig. 24, owing to the failure of the left-hand
pleural patch to develop, the two rump patches, both par-
tially separate from each other, are wholly disconnected

Digitized by

Google

402 THE AMERICAN NATURALIST [Vol. XLVIII

from the former except by a narrow isthmus on the right
side. The long tail is usually without pigment, or mainly
so where areal restriction is present, and it is seldom that
pigment extends far on to the base of this member when
the restrictive tendency appears. In the domesticated
varieties of rats, the same patches may be distinguished.
There is, however, an interesting variety known as the
''hooded" rat, in which the ear and neck patches appear
to be normal, but a narrow median dorsal area is pig-
mented for a varying length, sometimes quite to the root
of the tail. A separate factor seems here to be involved,
producing what may be called a ''centrifugaP^ type of
pigmentation, which in many forms of mammals causes a
black spine stripe {Sorex wardi, Tupaia tana, certain
forms of Apodemus, Equus caballus), and others.

Among guinea-pigs the typical primary patches are
beautifully shown and may be seen in sundry figures
published in papers by Professor Castle on heredity in
this animal. The guinea-pig is one of the few mammals
yet known in which the median crown patch is visibly
present, a character which I take to be primitive.

In guinea-pigs the breaking up of the ticked color
pattern has progressed under long domestication to an
extraordinary degree, so that not only are black, tawny
or grizzled animals produced in various shades, but even
in the same individual, the different primary pigment
areas may he of different colors. This fact is of much
significance, for it indicates liot only the mutual independ-
ence of the contiguous color areas, but further points to
the manner in which a variegated color pattern may have
been acquired. Among mammals the color pattern is in
general, not greatly developed in comparison with birds,
yet in many cases where some modification has taken
place, it is evident that this differentiation is confined to
the limits of one or two of the primary pigment patches.
Thus in the South American Tayra (T. harhara)^ the
head and neck are a grizzled gray, and the breaks occur-
ring in pied individuals show that the grizzled condition

Digitized by

Google

No. 571] PATTERN DEVELOPMENT

403

a.sr

3X
Figs. 25-32. Diaobams Illustbating Pigmentation in Horses.

Digitized by

Google

404 THE AMERICA^ NATURALIST [Vol. XLVHl

is confined to the aural and nuchal patches only, for else-
where the animal is black. In this case, too, the black
condition is probably derived, for youngish animals are
uniformly grizzled, and sometimes, apparently, this is
the adult condition as well.

Among domesticated ungulates the same primary
patches are to be distinguished in cases where partial
albinism renders their bounds apparent, with the excep-
tion that in horses, cows and deer I have seen no clear
indication of the median crown patch which in mammals
is probably obsolescent.

In both horses and cows the patches show interesting
and peculiar modifications. A series of diagrams (Figs.
25 to 32) show these patches in ''calico^' horses, though
not so fully as could be wished. The first indications of
areal restriction of pigment in horses appear in the shape
of a white ''star" or round spot in the center of the fore-
head. This is often accompanied by white at the base of
the hoofs, or sometimes the entire foot is white producing
the so-called ''white stockings." But there is no neces-
sary correlation between these white areas, such as
Brewer (1882) tried to show. The white on the forehead
may vary from a few white hairs to a broad blaze cover-
ing the entire front of the head between the eyes to the
muzzle. Sometimes the restriction of pigment is such as
to produce in addition to the white star on the forehead,
a white spot over each eye, and sometimes these three
spots are joined by a narrow unpigmented area. This
indicates that pigment production is weak at a spot
directly over the eye in comparison with neighboring
parts, and this no doubt accounts for the fact that
in black-and-tan or other dogs these are the pale spots
over the eyes where black pigment is not produced.
A white spot over the eye is also characteristic of many
rodents.

Next after the restriction of the ear patches and the
drawing away of pigment from the feet, the most comimon

Digitized by

Google

No. 571]

PATTERN DEVELOPMENT

405

white marking seems to be a primary break, as in Fig. 25,
from the shoulder back of the foreleg, which delimits the
posterior border of the shoulder patch. In the horse the
shoulder patch is large, and differs from that of any
mammal I have yet studied, in its great extent forward
along the dorsal side of the neck nearly to the head. In
Fig. 26 a small break at the back of the neck indicates the
beginning of separation between the ear and the neck
patches dorsally, and a long tongue of white running up-
ward from the forearm indicates the anterior limit of the
shoulder patch. This limit is marked still nearer the
dorsal line in Fig. 27 by a white spot on the side of the
neck near its base. In Fig. 28 the shoulder patch has
entirely dropped out and the white space outlines very
nearly its extent. The ultimate center is perhaps shown
by the small shoulder spot in Fig. 31.

The area covered by the ear patches extends well on to
the upper part of the neck, and in Fig. 29 is shown at its
greatest spread, or, as in Fig. 28, cut off by a narrow white
collar from the neck patch. The neck patch is remarkable
from the fact that in its areal reduction it becomes re-
stricted first dorsally, and the ultimate center of each side
is nearly ventral on the throat, so that, as generally seen^
the two centers form a single median patch on the front
or ventral part of the throat. In Fig. 26 the neck patch is
seen to pigment the anterior side of the forearm and is
partly separated from the shoulder patch by a long
tongue of white. It seems to extend up diagonally to
reach the mid-line of the neck for a short distance only,
as indicated in Fig. 28, where its bounds are only slightly
contracted. In Fig. 29 it is so far lessened as to be absent
from the forearm, though still in contact at the throat
with the ear patch where, however, a deep indentation
locates the dividing line between the two patches. In
Fig. 31 a median ventral division of the conjoined neck
patches is seen indicated at the upper part of the area,
which in this case no longer reaches the ear patches.
Still further reduction of both ear patches and neck

Digitized by

Google

406 THE AMERICAN NATURALIST [Vol. XLVm

patches is seen in Fig. 32, but, as commonly, the neck
patches seem fused in the midventral line. This shifting
of the neck centers ventrally is a rather remarkable
phenomenon which may have some relation to the manner
in which the head is held erect. For this reason it might
be expected also in antelopes, and is perhaps evidenced
in such a species as the oryx, in which there is a black
median line on the throat as though strongest pigment
production centered there rather than on the gray sides
of the neck. The median reduction of the shoulder
patches in horses is sometimes indicated by a white mane.

The rump patches in the horse appear to be much as in
other mammals, restricted to the tail and posterior part
of the buttocks and the entire foot. In Fig. 30 the patch
is shown at nearly its full development, except that it has
failed to extend to the entire hind foot. In Fig. 27 it has
drawn away still farther but remains in contact with the
side patch at one place. In Fig. 29 it is further restricted
to the tail and posterior border of the haunches, while in
Fig. 32 it covers only the root of the tail and that member.

The side patch is the largest of all and extends from
the shoulder to the fore part of the haimches and on to the
fore part of the hind leg nearly to the foot, as seen in Figs.
27 and 28, where it is still in contact with the rump patch,
or in Fig. 29 where it has become separated. In its fur-
ther reduction this patch may appear as a small spot
back of the ribs or, as often, a curious division takes
place, separating the patch into a dorsal area and a
lateral one. Occasionally this secondary break appears
in a horse which has most of its patches otherwise well
developed. In Fig. 31, the pigmented area of the tail,
buttocks and lumbar region consists of the conjoined rump
patch and a dorsal portion of the side patch, while the
ventral part of the side patch is present as the oval spot
at the groin. In Fig. 30 the latter spot only persists, but
in Fig. 32 the dorsal portion of the side patch alone is
present as a stripe along the entire back, except where it
breaks away posteriorly from the small rump patch.

Digitized by

Google

No. 571]

PATTERN DEVELOPMENT

407

This peculiarity of the side patch in horses is somewhat
paralleled in cows by a tendency to secondary breaking
up, though in a different way, as detailed below. It is
significant in this connection that in horses and donkeys
there is usually a black stripe along the spine from
shoulder to tail which may indicate that *' centrifugal pig-
mentation" is also present (see beyond). The dotted
lines in Fig. 32 indicate the approximate boundaries of
the several primary patches. The crown patch seems to
be wanting in horses ; 2 is the ear patch, 3 the neck patch,
4, 5 and 6 the shoulder, side, and rump patches,
respectively.

Of domestic ruminants I have studied the pigmentation
in the cow and show in Figs. 33 to 42 a few of the many

Figs. 33-36. Diaqbams Illustbatino Pigmentation in Domestic Cows, Side

View.

variations in partial pigmentation. These are all drawn
from photographs or from the animals themselves, and
are of cows in which, so far as I know, there has been no
attempt at breeding for pattern. Two types of spotting
may be distinguished in cows : first, that in which the pig-
mented areas are sharply outlined and solid or at least

Digitized by

Google

408 THE AMERICAN NATURALIST [Vol. XLUVQ

practically so; second, that in which there is a grefiLteTr <
less tendency for the primary patches to be much, bxro'k
np into small islands (as in Fig. 36) by secondary l3x*^ct'
though the main areas are still distinguishable. I tt
this second or fragmental type to be a different pXx^xxc
enon from the diffuse or dappled condition seen, ixx
coach dog or the dappled-gray horse.

In the cow, the ear patches as usual pigment ^^c^lx
proper side of the head to a short distance behimd

Figs. 37-42. Diaqbams Illustrating Pigmentation in Domestic C^^^ .
Seen Spread Out and from Above.

ears. The point of separation between ear patchy ^ ^
neck patches is indicated by a small break back ^:>^ -
skull in Fig. 38, while the posterior extent is sho^*^'"^'^ -.
the two ear patches in Fig. 42. These patches uS"*^^

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 409

draw apart first across the forehead making here a tri-
angular white mark, and on the muzzle, as in Fig. 34.
Further restriction broadens these white marks and joins
them by a narrow isthmus as in Fig. 35. In Fig. 40, the
two patches are still conjoined across the vertex, but are
much reduced, that of the right side more than that of the
left. In Fig. 42 they have failed to join medially, though
fairly well developed longitudinally. Still greater re-
duction, as in Fig. 37, confines them to the ears, the bases
of which appear to be the ultimate centers.

The neck patch in the cow is more extended posteriorly
than in the horse, and its center is strictly lateral rather
than nearly ventral. It is shown in Fig. 34 somewhat
contracted from the mid-line of the throat, but extends
squarely back against the f oreshoulder at the base of the
neck, and is fused near its ventral comer with the
small shoulder patch, itself much reduced. As in
other mamimals it appears to extend in its complete
development, to the front edge of the upper foreleg. The
animal in Fig. 41 shows a bilaterality in its pigmentation
that is rather unusual. What appear to be the reduced
neck patches are seen far back at the border of the fore-
shoulder. In Fig. 40 the left-hand neck patch has
dropped out, but that of the right side is still present,
though small, and in Fig. 42 it is reduced to a small spot
only.

The shoulder patch in cows is remarkably narrow,
and compressed between the neck patch and the body
patch, whence it extends as usual on to the foreleg. In
Fig. 33 a primary break back of the f oreshoulder marks
the nearly vertical posterior outline of the shoulder
patch. In Fig. 34 the separation of this area from the
neck patch is all but complete and the patch itself some-
what reduced. Its narrow vertical outline is thus indi-
cated, as well as in Fig. 39, in which there is a narrow
tongue-like extension down on to the center of the foreleg.

In its further reduction it appears as a small center
at the base of the scapula, as in Fig. 35, or in Fig. 40, in

Digitized by

Google

410 THE AMERICAN NATURALIST [Vol. XLVIK

which both shoulder patches are present, though small.
In Figs. 41 and 42 the shoulder areas are wanting. A
very common mark in cows is a white belt just back of
the foreleg. This is due to the development of a primary
break between shoulder patches and side patches, a con-
dition which is nearly realized in Figs. 33 and 38. It is
probable that this marking has been more or less fixed
through selection in breeding, and this has been the more
readily accomplished, since this break occurs in a place
which is one of the first in cowa to cease pigment
production.

The side patch is large and covers the entire lateral
region of the body from the scapula to the hips, and on
to the front edge of the hind limb. When only slightly
reduced, it appears as a blanket-shaped area across the
back as in Fig. 38, where it has not wholly broken away
from the shoulder and rump patches, or as in Fig. 33,
where it has become nearly separated. In its further
reduction this dorsal blanket shows a peculiar manner of
breaking up into more or less transverse stripes directed
slightly backward. The beginnings of these secondary
breaks appear in Fig. 39 in which are seen on each side
posteriorly two deep indentations at the edge of the
patch, whose points if extended would meet the white
pigmentless islands already present within the patch.
In Fig. 34 a similar series of indentations points to the
trisection of the side patch which is realized in Fig. 35.
Here is a characteristic which if developed might even-
tually result in the actual production of white stripes on
the body, such as are found, for example, in certain ante-
lopes as the bongo and the kudu. The tendency of the
side patch to divide into three, as in these diagrams, is
rather marked in cows, and even with further reduction
the three centers persist fairly well. The first of these
secondary centers is just back of the shoulder patch, the
second about over the last ribs, and the third over the
lumbar region. In Fig. 40 the first two are present on the
left side, with a small spot between, which has become

Digitized by

Google

No. 571] PATTERN DEVELOPMENT 411

separated from one or the other of them, while the third
or lumbar spot has dropped out. On the right side, the
first and second divisions are still fused dorsally, but
the lumbar division is distinct. The same three divisions
are seen in Fig. 35, better developed, whereas in Fig. 42,
the two lumbars are present, on6 on each side, and con-
siderably in advance of them, what seem to be the rem-
nants of the first division of the side area, the left one of
which has further broken up.

The rump patches show no especial peculiarities, but
cover the posterior part of the buttocks and hind legs,
and the entire feet and tail. Though frequently the two
patches of opposite sides are conjoined medially, they are
often, under considerable reduction, well separated. The
beginning of such a separation appears in Fig. 38, where
there is a deep median tongue of white anteriorly, mark-
ing the line of union. In Fig. 41 the reduction has pro-
gressed still farther so that the two patches are quite sun-
dered medially and do not extend to the tail. In Fig. 40
the patch of the left side has become inactive, and that of
the right si(jle is small.

A curious condition not infrequently seen is shown in
Pig. 37, in which all the patches are present, but those of
the right side are separated from those of the left by a
median dorsal white line, showing the distinct bilaterality
of these pigment areas. In the figure, the ear patches
are so restricted as not to reach the neck patches of
their respective sides, the shoulder patches do not extend
far on the forelegs, the side patches are reduced ven-
trally, and the rump patches, though in contact with the
side patches, do not pigment the tail or extremities of
the legs. A further reduction of pigment areas results in
Fig. 41, in which the paired centers of neck, side and
rump patches still appear.

The diffuse condition of pigmentation is illustrated in
Fig. 36, which is a photograph, inked in. The ear patch
is seen much reduced, but pigmenting the ear. The neck
patch is of most irregular shape, with several subsidiary

Digitized by

Google

412 THE AMEBIC AN NATURALIST [Vol. XLVHl

spots separated from its lower border. A clear line
separates the neck patch from the shoulder patch, which
is also of most irregular boundary. The side patch, at
its fore part, is broken into a series of small islands
which tend to arrange themselves in lines following the
direction of the ribs. The main part of the patch shows
a decided tendency to break into the usual three or per-
haps four portions. It is common for cows to have
patches with very irregular boundaries and tongues of
pigment, which may break oflf into isolated spots in a
most bewildering fashion, but even in such cases it is
possible to distinguish the main patches of which these
form part.

White patches occur in other domesticated ungulates
as the pig, the llama, the alpaca, the camel, the yak, the
reindeer, and the goat. In the water-buflfalo, occasional
animals seen in Egypt show a beginning of pigment re-
duction through the presence of white in the forehead or
on the tail. I have had no opportunity to study the mark-
ings of these species.

(To he concluded)

Digitized by

Google

INTERNAL RELATIONS OF TERRESTRIAL
ASSOCIATIONS

ARTHUR G. VESTAL
University of Colorado

Contents

I. Introduction.
11. Internal activities of the association, as determined by the con-
stitution of the individual organism.

A, Ecological constitution of the organism.

B, Constitution of the plant in relation to environment.

C, Constitution of the animal in relation to environment.

D, Internal activities of the association.

III. Belative influence of different organs with the association domi-
nance.

A. Factors of dominance among animals.

B. Criteria of dominance among animals.

C. Specialized and unspecialized animals.
IV. Distribution within the association.

A, Distribution in space.

B, Distribution in time.

V. Interdependence of terrestrial plant and animal commimities.

A, Geographic relations of terrestrial plants and animals.

1. Greographic range: the province.

2. Distribution within the province: distribution of plants

and animals into communities.

B, Local relations of plant and animal assemblages (relations

within the association).

1. Similarity of ecological type of plants and animals.

2. Belative dependence of plant and animal assemblages.

3. Correspondence in distribution within the association.

4. Uniformity of species composition of plant and animal

VI. Summary and conclusions
VII. References.

I. INTRODUCTION

The material here presented is based on the writer ^s
studies, during the past five years, of terrestrial associa-
tions of plants and animals, mainly in different parts of
the prairie region. The particular area chiefly used for
illustration in this paper is the sand prairie of the. Illinois
Eiver valley, plants and animals of which have been
studied by Hart and Gleason (1907) and by the writer
(19136). A later study has been made of the vegetation
of inland sand areas of Illinois (Gleason, 1910) ; the Ijake
Michigan beach area in northeastern Illinois has been
studied by Gates (1912) ; beach areas in Illinois and Indi-
ana by the writer (1914a). The chief representation of

413

Digitized by

Google

414 THE AMERICAN NATURALIST [V0L.XLYIII

the sand prairie is the bunch-grass association, well-
developed in parts of northwestern, central and north-
eastern Illinois, and in northwestern Indiana, in each
of which areas, as well as in the sandhills of Nebraska
and of eastern Colorado, the writer has studied. Discus-
sions of physical, vegetational and animal aspects of the
associations of the central Illinois sand prairie, together
with an annotated list of the animal species, with data on
food, habitat-relations, life-history, etc., are embodied in
the writer's paper (19136), to which constant reference
is made. Frequent citations to a more detailed study of
local distribution of grasshoppers, in a Michigan area
(Vestal, 1913a), and to the many associational studies
of Shelford, are to be found.

Th« data which have accumulated relate nearly equally
to the botanical and zoological aspects of associational
study, but since the subject of plant ecology is at present
more advanced than that of animal ecology, it has been
possible to treat the vegetational side of the problem
very briefly, so that more of the discussion relates to
animals and animal assemblages.

The writings most frequently cited are indicated by
italic capitals, the full titles appearing in the list of spe-
cial references at the end of the paper.

The writer wishes to thank Dr. Charles C. Adams, Dr.
Max M. Ellis and Dr. H. A. Gleason for suggestions and
criticism.

11. INTERNAL ACTIVITIES OF THE ASSOCIATION, AS DETER-
MINED BY THE CONSTITUTION OF THE
INDIVIDUAL OIWJANISM

The internal activities of the association may be said
to be the sum-total of the activities of all the plants and
all the animals which make up the association. Such a
sum-total of activities may well be thought of as an intri-
cate and complicated mass of dependencies. It wiU
simplify the treatment of the entire system of relations
if the chief dependencies of the individual organism are
first discussed. A knowledge of the ecology of the asso-

Digitized by

Google

No. 571]

TERRESTRIAL ASSOCIATIONS

415

ciation is built up largely from a knowledge of the ecol-
ogy of all the organisms which compose it.

A. Ecological Constitution of the Obganism

The constitution of the organism is the sum-total of
those of its characters which enter into relation with
environment. These are commonly classified as structural
and physiological. For the purposes of this discussion
it would seem preferable to subdivide physiological char-
acters, restricting the term physiological to denote those
characters concerned with ordinary metabolic processes
of the organism, and excluding those having to do with
life-history and rates of reproduction (these may be dis-
tinguished as biographical and numerical) and also, when
dealing with animals, those related to behavior {psycho-
logical characters). The constitution of the organism in
relation to environment will be discussed in terms of
these classes of characters.

B. Constitution of the Plant in Relation to
Envibonment

The environmental influences in the association are of
three kinds: (1) physical, (2) plant, (3) animal. Each
plant and each animal must obtain from each of these
three constituents of its environment certain necessaries ;
it has certain structural and physiological characters
which enable it to obtain these necessaries, and to with-
stand adverse environmental influences.

The environmental relations of plants are very differ-
ent from those of animals. A tabular comparison of
these relations has been made by Shelf ord {A: 593). As
therein pointed out, structural characters are of greatest
importance in the adjustment of the plant to the environ-
ment, and plants in a given habitat are likely to have a
common structure or growth-form, indicating common or
ecologically equivalent physiological conditions within.

Different plants (and different animals), within a com-

1 Based partly on Forbes ', classification of adaptation to food require-
ments (1909: 292).

Digitized by

Google

416 THE AMERICAN NATURALIST [VoL.XLVm

mon habitat, are similar in ecological constitution (eco-
logically equivalent) in so far as their presence is deter-
mined by the same environmental conditions. It should
be pointed out that there are local environmental differ-
ences within the area of the association which allow the
presence of differently constituted organisms, and that
the entire range of environmental conditions within the
habitat is usually much wider than that of the environ-
mental complex selected by a particular organism. The
environmental complex of the organism is not the same
as the sum-total of environmental conditions withm the
association. Each organism differs in greater or less
degree from others in ecological constitution, and thus
selects a different environmental complex.

The physical factors of the environment are of great-
est importance in the life of the plant. Plants influence
one another directly to only a slight extent. There is
usually very little of the social relationship among eco-
logically similar plants which will compare with such
relationships as seen in animals. Competition among
plants is mainly a struggle to determine which plants
are to be most favored by physical conditions, and it is
probably most severe for the physical factor present in
minimal quantity. In desert associations plant competi-
tion is almost exclusively for water, and extensive root
systems are developed. In grassland it is very largely
for above-ground space ; in forests it is principally for
light. The influence of the animal-environment is prob-
ably of greater importance than has commonly been
realized by plant ecologists ; the study of economic ento-
mology and of the effects of grazing upon grasslands is
helping to bring about a realization of the importance of
animal influence upon plant life.

The structures of plants show frequent and great
modification in response to the physical conditions of the
environment. These modifications are most frequent and
important with respect to the factor present in minimal
quantity. Characters which may be associated with

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 417

direct plant influence are infrequent. Certain plants
which become more abundant as a result of close grazing
are equipped with spines, or have acrid or pungent
juices; and many other characters may be correlated
with animal influence. The structural modifications are
most evident in the adjustment of the plant to external
conditions, though these are accompanied by physiolog-
ical characters which are also in harmony with the en-
vironment.

C. Constitution of the Animal in Relation to
Envibonment

The animal, like the plant, selects an environmental
complex which is of three kinds: (1) physical, (2) plant,
(3) animal. Different animals show extreme variation
as to the degree in which the different parts of the en-
vironment are important to their existence. Endopara-
sites, for example, are most directly concerned with the
animal part of their environmental complex.

The existence of any animal is dependent upon a num-
ber of physical factors, all of which must be present in
proper degree or quantity. Minimal and maximal quan-
tities of any one of several factors mark the limits of
existence of any animal {A: 598 — law of toleration of
physical factors). It is not necessary to consider these
factors in detail. The animal reacts to physical environ-
ment most evidently by its behavior : psychological char-
acters restrict activities more narrowly than do those of
other types. They are accompanied by structural and
physiological characters; hibernation, storage of food,
etc., are biographical characters correlated with seasonal
changes in physical environment. Animals which are
subjected to very severe physical conditions may produce
a larger number of offspring than those to which physical
conditions are favorable. This is an example of corre-
lation of a numerical character with the physical environ-
ment.

The plant environment reacts upon and modifies phys-
ical and animal environments, and has also direct influ-

Digitized by

Google

418 THE AMERICAN NATURALIST [VoL-XLVHI

ence upon the animal. In addition to its effect in the con-
trol of temperature, light and other physical factors, the
vegetation constitutes the basic food-supply for the ani-
mal community, and also provides shelter and materials
for abode {A: 601). Cases of direct association between
particular plants and particular animals are numerous,
but the majority of animals have no direct relation to
particular kinds of plants. Behavior characters are in
general of greater importance in the relation of the
animal to the plant environment, though such relations
are not confined to psychological characters.

There are two sets of relations between the animal and
its animal environment. These are: (1) social, and (2)
antagonistic. Social relations {inter-psychology and
inter-physiology of Shelf ord. A: 608, h) include those
between individuals of the same species, and between
animals of the same or similar mores^ (ecologically equiv-
alent animals), in so far as these relations are not
antagonistic. Breeding and family relations are the
principal activities which come under this head. Be-
havior characters are of greatest importance, as compared
with structural and other characters. The antagonistic
relations constitute the intermores-psychology and phys-
iology of Shelf ord {A : 608, c) . They are the antagonistic
relations between animals not ecologically equivalent, and
they are also antagonistic relations within a species and
between ecologically similar forms. These relations are
probably not greatly concerned with reproduction, but
center about the feeding activities of the animal. The
existence of the individual animal, in its relation to other
organisms, is dependent upon three conditions: (1) it
must obtain suitable and suflScient food, (2) it must be
free from destructive competition of animals of similar
requirements, (3) it must be able to escape or to with-
stand attacks of other animals (or, sometimes, of para-
sitic fungi or bacteria). The various characters of the

2 Mores (Latin for customs, habits) has been used bj Shelf ord (1911a:
30) to supply the need for a term including all physiological and behavior
characters of the animal.

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 419

animals are correlated with all three of these conditions.
The characters are both *' adaptive'' (fixed by heredity),
and regulatory (not fixed).

Following is a synopsis of correlations between the
various types of characters and the three conditions of
existence, in the relation of the animal to its antagonistic
animal environment.

(I) Characters Which Enable the Animal to Obtain Food

1. Structural Characters. — ^Animals of selective food-
habits often have specialized structures, as in the case of
the long tongue of woodpeckers. Animals of non-selec-
tive food-habits have mouthparts that are not so highly
specialized ; thus grasshoppers and cutworms have heavy
mandibles for cutting vegetation; tiger-beetles and
Chrysopa larvae have sharp piercing mandibles. The
whole structure of the predaceous animal, its *' action
system," is sometimes suggestive of the manner of pur-
suit or holding of its prey.

2. Physiological Characters. — The physiology of ani-
mals of different food-habits differs materially. Physio-
logical characters are not apparent, generally speaking^
and are secondary to psychological characters. The
range of food assimilable by the animal is usually much
wider than that selected by it, as is seen when animals of
selective habits take new kinds of food when the usual
food is exhausted, often thriving seemingly as well as.
before.

3. Psychological Characters. — Selection of food is
determined chiefly by behavior characters of the animal.
These may be so widely variable that the animal will be
virtually omnivorous, as in the case of crickets, or so
narrowly restricted that it eats only a single species of
plant or animal, as the leaf -beetle Blepharida, a sand-
prairie insect eating leaves of the three-lobed sumac, and
the pentatomid bug, Perillus, which feeds on Blepharida
(cf. E: 49, 30). Selection is only one of the many psycho-
logical characters relating to food. The behavior cnar-
acters manifested in obtaining food are of great variety.

Digitized by

Google

420 THE AMERICAN NATUBALIS I [VoL.XLVlIT

With these are accompanying structural and physiolog-
ical characters, which, however, play a subordinate part.

4. Biographical Characters. — These may consist in
timing the life-history of the animal with that of the
food-species (plant or animal) in such a way that the
period of greatest activity of the former coincides with
the period of greatest growth or abundance of the latter.
This feature may be incidental to seasonal change of
physical environment. Whatever its cause, it is very
general in an established association, so general that it
is seldom recognized. It is of advantage to both animal
and food species.

5. Numerical Characters. — The rate of reproduction
must be so adjusted to its food-supply (plant or animal)
*Hhat only the unessential surplus of this food shall be
appropriated, leaving the essential maximum product
undiminished'' (Forbes, 1909: 293). Species of re-
stricted food-habits must remain less numerous in indi-
viduals than general feeders, as the available food-supply
is very much less.

(II) Characters Which Remove the Animal from the
Competition of Other Form^

1. Structural Characters. — Structures which permit
animals to live in varied habitats, to take varied foods,
or to time their activities differently, remove each group
of animals from competition of all the others, resulting
in advantage to all. To that extent the f ossorial forelegs
of the mole, the long proboscis of the butterfly, and modi-
fications of the eyes of nocturnal animals, are characters
which do away with competition. The structural char-
acters are, however, accompaniments of modifications of
behavior, and are secondary to the latter.

2. Physiological Characters. — Ability to digest food-
materials unavailable to other animals is an advantage-
ous physiological character. Thus the leaf-beetle Chry-
sochus auratus, which lives on dogbane (Apocynum)^
and the ** skin-beetle " Trox, which eats animal tissues
in an advanced stage of decomposition, have few corn-

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 421

petitors for food. Physiological, as well as structural,
characters, are accompaniments to modifications of habit.

3. Psychological Characters. — Apparent preference
for certain activities, certain habitats, or certain foods,
together with peculiar behavior complexes, seem to be of
greater importance in removing animals from competi-
tion than structural and physiological characters.
Highly regulatory habits permit certain animals to ad-
just themselves to changing conditions of competition.

4. Biographical Characters. — ^Professor Forbes (1909:
295-298) discusses the alternative timing of the active
period among close competitors for food. (It so happens
that the animals mentioned, having almost identical
habits, compete with each other in many ways, besides
with respect to food.) In the sand prairie it has been
found that different species of certain genera, having
otherwise the same habits, differ greatly in life-history.
Evidence of this biographical adjustment is more or less
complete for two species of Arphia {E: 21), two or three
species of Hippiscus {E: 21), two species of the milkweed
beetle, Tetraopes {E: 47), and three species of P recta-
canthus, robber-flies {E: 55). In these genera the term
of activity of one species is abruptly followed by that of
another, the successive periods usually covering most of
the summer season.

5. Numerical Characters. — ^When a certain limited
food, place of abode, or other desideratum is used by two
or more kinds of animals at one time, a numerical adjust-
ment is likely to be found among these competing species.
The rate of multiplication of each species must be suffi-
cient to keep up its numbers, to allow it to hold place
with competing species. (Too high rates of multiplica-
tion, on the other hand, are disadvantageous because of
other influences.)

(Ill) Protective, Defensive and Concealing Characters
1. Structural Characters. — Animals have various de-
fensive, protective and concealing structures. Stings,
beaks, mandibles, teeth, claws, hairs, spines, resemblance

Digitized by

Google

422 TBE AMEBIC AN NATURALIST [Vol. XLVIH

to surroundings in color or form — all are of advantage to
animals which possess them. Certain of the interstitial
or blowsand animals resemble in color the sand on which
they rest (Cicindela lepida, Stachyocnends, Psinidia,
Spharagemon; ci. E).

2. Physiological Characters. — Malodorous and ill-
tasting animals are to a considerable degree exempt from
attack. This is essentially a physiological modification,
though a structural basis in the form of glands may be
present. In the sand prairie Chrysopa (lace-winged fly),
a number of Hemiptera, ladybird beetles, soldier bugs
{Chauliognathus) J blister-beetles (Epicauta), and others,
are ill-tasting (perhaps not to some animals). The
skunk's lack of caution is well known.

3. Psychological Characters. — Self-preservation in
animals depends more upon their activities and behavior
than upon special structures. The ordinary methods of
resisting or evading attacks of enemies are generally
known and need not be discussed. Many specialized in-
stincts have arisen, such as feigning death, or dropping
to the ground when disturbed, as seen in many herbi-
colous beetles.

4. Biographical Characters. — It is to the advantage of
animal species preyed upon by others if their period of
greatest abundance is timed with the period of greatest
activity of the animals which feed upon them.

5. Numerical Characters. — Animals, as well as plants,
must produce a normal excess in numbers which will pro-
vide food for other animals and still leave a sufficient
number of individuals to continue the species.

It will be noted that the various kinds of characters
usually accompany one another, all being parts of a
single modification. This modification may have rela-
tion to one or to several of the environmental influences
(physical, plant or animal) or to more than one kind of
antagonistic relation between the animal and others.
The modification is not necessarily advantageous to the
animal with respect to all or to any features of the

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 423

environment, though a large number of characters do
result in advantage. Characters advantageous in one
relation may be disadvantageous or indifferent in an-
other relation. The origin of the characters is not at
present a subject which can be treated in a study of inter-
relations of organisms (cf. Shelf ord, 19126: 342). Be-
havior characters appear to be of greatest importance to
the animal in determining its relations with other organ-
isms of the association, though usually these are accom-
panied by physiological or structural characters. The
animal is not adapted to a particular status in the asso-
ciation ; its ecological constitution determines what place
it shall be able to find among the other animals of its sur-
roundings. The relations among the various animals,
when a state of equilibrium has been reached, are the
result of mutual accommodation on the part of all the
animals involved.

D. Internal Activities of the Association

It has been indicated that the complex of activities
within the association is the synthesis of all the activities
of the individual organisms. Each plant and each ani-
mal is subjected to physical, plant and animal influences.
From the extreme complexity of the entire system of
relations within the association, it is hardly possible to
consider more than one or several of these at one time.^
It is possible, however, to see that each species finds a
status within the association, according to its particular
combination of internal and external relations. It con-
tinues in fairly constant numbers from year to year. A
change in these numbers, if at all great, may cause a dis-
turbance in the association, which is quickly regulated
by the activities of conflicting organisms (Forbes, 1880).
The entire association of plants and animals, by very

• Very helpful diagrams are given by Shelf ord (C: 167, 168) which il-
lustrate the food relations of land (prairie) animals. There are also dia-
grams showing food relations of aquatic animals (C: 70, 71). Food rela-
tions of animals of plains and mountain streams are discussed by Ellis
(1914: 122-127; diagram on p. 125). References to studies dealing with
interrelations of organisms may be found in the recent handbook of Adams
(1913: 123 et seq.).

Digitized by

Google

424 THE AMERICAN NATURALIST [Vol. XLVm

reason of the conflicting interests, the varying conditions
necessary for existence, and the varying methods of re-
sponse to these conditions, forms a self-contained and
self-regulating system of activities.

III. RELATIVE INFLUENCE OF DIFFERENT ORGANISMS WITHIN
THE ASSOCIATION— DOMINANCE

The plant ecologist determines which plants in an asso-
ciation are of greatest importance (dominant) by ob-
serving which species tend to increase at the expense of
others, which are most abundant, most frequent, largest,
etc. Competition among plants in a grassland associa-
tion is mainly for space, and the dominant species are
usually determined with considerable accuracy after some
study. With the animals the consideration of dominance
involves greater complexity. The important relations
between conflicting animal species are those in which
they obtain food, are removed from competition, or
escape enemies. These relations are in each case most
directly concerned with food. The plant-eaters of the
association thus form a dominant group within the asso-
ciation, since predaceous and parasitic animals, and
scavengers in large part, depend upon them for existence.
Individual species within the various food-groups, how-
ever, present such striking differences in importance,
that we can not speak of all plant-eaters as dominant
forms, or that all animals of other food-habits are un-
important. It is merely probable that the phytophagous
group will contain a larger proportion of dominant spe-
cies. This appears to be the condition in the bunch-grass
association.

A. Factors of Dominance Among Animals

The success of an animal species within an association
is due to the resultant effect of a large number of factors.
Among these may be mentioned number of individuals,
size, activity, voracity, concentration of food, rapidity of
growth, rapidity of reproduction, and wideness of dis-
tribution in space and in time. Dominance signifies more

Digitized by

Google

No. 571]

TERRESTRIAL ASSOCIATIONS

425

than mere ability of a species to thrive in its surround-
ings : the species of greatest influence are those on which
the greatest number of other animals depend ; thus domi-
nant species are successful, but successful species are not
always dominant. Species which are relatively free from
competition or which have comparatively few enemies
may be successful, but are not dominant, and are usually
not numerous. Species which are successful and at the
same time extremely abundant, usually form the food of
a large number of other animals, as it appears to be the
rule that no considerable source of food within the asso-
ciation is left unused. Dominance in a species, then,
would seem to include the dependence of other animals
upon it, plus the ability to thrive in spite of the drain
upon its numbers.

B. Critebia of Dominance Among Animals

The factors mentioned as contributing to the success
of a species, and the numbers of animals dependent upon
the species, are all indications of the degree of its domi-
nance. It appears that another criterion is available,
which perhaps expresses the summation of many factors
which contribute toward dominance. This is the degree
of specialization exhibited by the species in its adjust-
ment to a particular place in the association. Dominant
animals appear to be those of moderately specialized
habits rather than those of highly specialized, or rela-
tively unspecialized, habits.

C. Specialized and Unspecialized Animals

Each species may be referred to a position in the scale
of specialization in habit. The degree of specialization
of the species is well seen in the food-habits, though all
the habits are to be considered. The most abundant food
in the sand prairie is plant material, bunch-grasses. The
majority of the plant-feeders are adapted to eat herbage
of nearly any kind : they are not restricted to particular
species or particular parts of plants. They are non-
selective feeders. Grasshoppers,, cutworms and certain

Digitized by

Google

426 THE AMERICAN NATURALIST [Vol. XLVm

leaf -beetles are thus moderately specialized plant-eaters.
There are also non-selective predaceous animals, as tiger-
beetles and lycosid spiders, which eat any kind of small
animal. These are also moderately specialized. The
moderately specialized animals carry on the gross metab-
olism of the association; they constitute the dominant
group, and include the dominant species.

Selective feeders belong with the highly specialized
animals. In the bunch-grass association Languria hi-
color, an erotylid beetle, bores in the stems of the com-
posite Cacalia (Indian plantain), while Lygceus bicrucis
(hemipterous) feeds on the same plant; Perillus circum-
cinctus eats Blepharida rhois. Others of the associa-
tion eat selectively. The majority of parasites are
greatly restricted in their selection of hosts. Such ani-
mals are particularly dependent upon special kinds of
food, which in many cases are not available to general
feeders. Highly specialized forms are thus enabled to
avail themselves of opportunities denied to animals of
generalized type; but while they avoid competition by
the adoption of special kinds of food, or by special habit
of some other kind, they lack the versatility of the les§
specialized animals, being unable to adjust themselves
to changed conditions. They may, therefore, become
abundant at times ; but as they depend wholly upon one
variable condition (perhaps the presence of a particular
plant species, which may be quite infrequent) they never
can become dominant species. Absolute numbers of the
insects which live upon Cacalia, for example, are insig-
nificant in comparison with such animals as the grass-
hoppers.

On the other hand, animals of relatively non-specialized
habits would also be ineffective in the association, for
whatever field of activity they were to enter, they usu-
ally would find already occupied by some animal better
constituted for that activity. Such non-specialized forms
would assume particular importance only when some
animal on which they might feed should become unusually

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 427

abundant. Few animals are really non-specialized in
habits; many moderately specialized species, however,
may on occasion turn from their ordinary activities, per-
haps to appropriate a particularly abundant kind of food.
Many ants are thus habituated to certain ordinary kinds
of food, but are able to eat organic food of almost any
sort, and do vary their food with circumstance. When,
as frequently happens, some animal species becomes very
abundant,* the attacks of a great many species of flexible
habits becomes concentrated upon it, and the numbers of
the food-species are soon reduced to normal. Animals
with non-specialized habits, by taking whatever food is
easiest of access, act as regulators of disturbances within
the association. A clear exposition of the manner in
which species of generalized habits restore unbalanced
conditions to equilibrium is given in a paper by Forbes
(1883), in which the regulative action of birds upon
insect oscillations is discussed.

The animal's status within the association is deter-
mined not only by its food-habits, but by the sum-total of
its physiological and behavior characters (its mores).
The degree of dominance is indicated not merely by the
degree of specialization of food-habits, but in all habits,
by the degree of flexibility of behavior. An extreme
specialization in nearly any behavior character, as habit
of abode in the pit-digging ant-lion larva, prevents the
species from becoming dominant. The degree of spe-
cialization of behavior is thus a convenient criterion of
the relative influence of animals in the association. The
dominant animals are moderately specialized, and carry
on the ordinary work of the association. The highly spe-
cialized animals make use of space otherwise unoccupied
and food material not demanded by other species. Cer-
tain of the first group, with habits more highly regulatory
than is usual, with perhaps some few unspecialized forms

^ With some animale sudden abundance is a matter of seasonal periodic -
itjf as in the case of May-flies (Hexagenia) along the Illinois Biver (E:17).
The adults on emerging become a sudden source of food for animals of ad-
joining terrestrial associations, as the bunch-grass.

Digitized by

Google

428 THE AMERICAN NATURALIST [VoL.XLVin

in addition, tend, by following the path of least resist-
ance, to act in opposition to forces tending to destroy the
biotic equilibrium.

IV. DISTBIBUTION WITHIN THE ASSOCIATION

The association may be subdivided into minor groups
of organisms, both in space and in time. Each group,
being thus removed from the immediate influence of the
others, is to some extent self-contained, having its own
environmental conditions, its own assemblage of organ-
isms, and its own system of interrelations.

A. DiSTBIBUTION IN SpaCE'

Different parts of the space occupied by an association
present different environmental conditions. In the ver-
tical distribution, four strata, the air (cf. E: 73), the
plant layer, the surface layer and the underground layer,
are usually present. In forest associations, the plant
layer is complex, plants of various heights giving rise to
minor strata (cf. ^). In grassland associations the plant
layer is relatively uniform. Animals are most numerous,
during the feeding activity, in the plant layer. Others
find food at the surface or underground. Many of the
animals in the air or on the ground move about rapidly
from plant to plant. Predaceous animals (while active)
are frequently permanent members of air and ground
layers, depending for food upon the transient animals
and upon members of their own group. The ground
stratum is composed of the surface and subsurface layers
{E: 72), which are not, however, continuous horizontally,
but alternate to greater or less extent.

Local variability in horizontal distribution is due
partly to local discontinuity of the various strata. This
interruptedness is particularly conspicuous in open asso-
ciations, where the plants do not form a dense growth,
but are separated by open spaces. The subsurface area
is provided by cover of various kinds, which lies more or
less scattered about on the surface.

»Cf. Shelf ord, A, B, 1912&, C; aleo D: 167; also p. — of thin paper.

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 429

The motility of the animal allows change in stratum,
and to some extent and in some animals, in habitat, with
change in activity. The food-stratum and the food-
habitat are apparently of greatest importance in the rela-
tion of the animal to other organisms.

B. DiSTBIBUTION IN TiME

Physiological activities of the plants are subject to
diurnal variation, and are also greatly affected by varia-
tions in weather conditions. The greater part of the
animals of an association are active during the day.
Others are nocturnal. During the inactive period of the
day the animal rests in some more or less sheltered place,
perhaps in a burrow or nest. The inactive state is also
induced by unfavorable weather conditions.

Seasonal changes in the association are very great in
temperate climates, particularly in treeless regions,
where the winters are severe. Seasonal changes in
the vegetation are marked, certain groups of the
plants appearing in successive periods during a sum-
mer season, giving four or five successive aspects to the
plant cover. A corresponding seasonal distribution is
observed among the animals of the association (cf. D:
175).

Annual changes in the associations are indicated by
the very marked differences in the numbers of indi-
viduals, in certain species of plants and animals, in suc-
cessive years. This may be due (1) to fluctuation in the
numerical adjustment between different organisms, and
(2) to the effect of annually varying phenological condi-
tions upon the various organisms.

Oscillatory irregularities in the association take place
at indefinite intervals. The causes and nature of oscilla-
tions have been thoroughly treated in several of Forbes 's
writings (1880, 1883, 1887).

V. INTERDEPENDENCE OF TEBBESTBIAL PLANT AND ANIMAL

COMMUNITIES

The thesis of the following section is that, in terrestrial
climatic or extensive environments, the relations between

Digitized by

Google

430 THE AMERICAN NATURALIST [Vol. XLVTU

the assemblage of plants and the assemblage of animals
are intimate and regular of occurrence; so much so that
(1) the two are coextensive, (2) the two constitute to-
gether a community which may be called a biotic asso-
ciation, (3) neither plant nor animal assemblage usually
occurs independently of the other, (4) the geographic
distribution of many of the plant and animal species
which make up the assemblages are in general corre-
spondence, (5) the species composition of the association,
over its range, varies no more widely, relatively speak-
ing, than would an assemblage of plants alone. Perhaps
the single view-point of the botanist, on one hand, and
the zoologist, on the other, has tended to a neglect of the
dual character of the one problem. Probably most botan-
ists and zoologists agree that relations of animals and
plants within a habitat are most intimate, and there is
a tacit assumption that all the organisms in one place
constitute the true system of interrelations, but botanists
have spoken of plant communities, and zoologists of
animal communities. There are numerous disharmonies
and variations in agreement of plant and animal assem-
blages, but these must not be allowed to obscure general
facts of correspondence.

It is recognized that .plants and animals of an area of
essentially homogeneous physical conditions are inter-
dependent, the animals as a group being wholly depend-
ent upon the plants for food, and many of the plants
being directly dependent upon animals, as in the matter
of pollination. All are directly or indirectly affected by
animals in some way. It is also recognized that the
plants are a good index to conditions for animal life, the
plant assemblage affecting animals locally in modifica-
tion of the physical environment, and more directly in
providing food, shelter, etc. {A: 601). It is further ac-
cepted that plants and animals respond to general en-
vironmental conditions in similar manner (Craig, 1908).
Thus considered, the character of the plant population of
an area is an index to general character, or ecological

Digitized by

Google

^ifit^^ifwr

No. 571]

TERRESTRIAL ASSOCIATIONS

431

type, of the animal assemblage. These relations, how-
ever, are quite general, lacking detail. Detailed consid-
erations may be geographic, including geographic range
of species and of communities, and the distribution of
species and of individuals into communities; and they
may also be local, dealing with interrelations of plants
and animals within the area of the community.

A. Geogbaphic Relations op Tebrestbial Plants
AND Animals

1. Geographic Range: The Province. — ^If one were to
plot the geographic range of the plant species found to-
gether in a given climatic habitat, a general correspond-
ence in distribution would be made apparent, a large
number of the species ranging more or less continuously
over a common, rather definite area (cf. Transeau, 1905).
The similar ecological constitution of these plants and
their consequent selective distribution into similar envi-
ronmental complexes gives a uniformity to the vegeta-
tion over the geographic region in which these environ-
mental conditions are found, and the resulting vegeta-
tion unit is known as a vegetation province (Gleason,
1910 : 42). The area of the province is generally uniform
in physical conditions. This uniformity is only relative,
being subject to gradual geographic variation in climate,
perhaps giving rise to subregions in distant parts of the
province, and to abrupt local variations in soil, water-
content, exposure, etc., giving rise to local or edaphic
plant assemblages very different from those of the cli-
matic or geographic type. Thus the prairie province
occupies the winter-dry interior region of North America.
Environmental variations from east to west, climatic and
physiographic, divide the province into the three sub-
regions of Pound and Clements (1898). Certain plant
species range over one or all of these subregions, still
others establishing themselves over the whole area of the
province and also scatteringly eastward, in dry treeless
parts of the deciduous forest province, to the Atlantic
coast. These last are also typical prairie plants, though

Digitized by

Google

432 THE AMEBIC AN NATURALIST [Vol. XLVHI

extra-limital in parts of an adjoining province locally
approximating the prairie environment.

The habitat-selection of different animal species re-
sults, in precisely the same manner, in similarity of geo-
graphic range among ecologically similar animals. These
correspondences of distribution point to the existence of
definite areas characterized by general similarity of the
animal assemblages. As the physical factors of the en-
vironment are the same ultimately for animals as for
plants, and as the vegetational environment for animals
has the same range as the physical environment, we
might expect animal communities to have the same geo-
graphic distribution as plant communities, and we might
expect the area of the plant province to be characterized
by distinctive kinds of animals as well as by distinctive
kinds of plants. The province is thus not simply a vege-
tation province, but a hiotic province. This is not a
new notion. Euthven (1908: 388-390) has stated a cur-
rent viewpoint as follows :

Those who are acquainted with the literature of the field zoology of
North America are familiar with the fact that, since the time of the
Pacific Railroad surveys, naturalists have noted that there are in North
America well-defined biological regions. These have been pointed out
at various times by Allen, Cope, Merriam, and others, and the fauna of
each has been more or less investigated. . . . For example, we have forms
of birds, reptiles and mammals characteristic of the southeastern de-
ciduous forest region, and still others characteristic of the northeastern
coniferous forest region, etc.

Shelf ord {A: 604) bases his classification of animal
regions upon that of plant regions, as worked out by
Schimper (1903) and Transeau (1903, 1905).

How close the correspondence of distribution of par-
ticular animals with that of vegetation provinces may be,
is well shown in the case of North American rabbits
(Nelson, 1909). The distribution maps shown for certain
species and groups of these animals might almost serve
as maps of the provinces. Many other animals, verte-
brate and invertebrate, correspond in area with the plant
provinces. Among the insects listed by Hart (1907 : 205)

Digitized by

Google

No. 571] TEBBESTRIAL ASSOCIATIONS 433

as western species, those for which a number of locality
records are available are plainly to be assigned to the
prairie province, the range of most of them extending
west to the Eocky Mountains, north about as far as
Montana, east to Illinois or Indiana, and south to Texas.

Other animal species bear apparently no relation to
province boundaries. Such animals have been discussed
by Shelf ord {A: 606, footnote), who shows them to be of
three types: (1) Species of scattered but very wide
range, covering perhaps several plant realms (animals
of local associations of extreme habitats) ; (2) Species
occupying only a part of the plant realm in which they
belong (animals of such ecological constitution that their
range is restricted by some conditions unfavorable in
certain parts of the province) ; (3) Species occupying
intermediate ground between two realms — these are few
(Euthven). These exceptional species are found also in
plants, so that local associations are occupied by both
plants and animals of the scattered-but-wide type of
range, while certain subregions, as the Great Plains area
of the prairie province, contain associations with both
plant and animal species restricted to these less extensive
areas.

Associations of two adjoining provinces may inter-
grade^ if ecologically similar, or may alternate if dis-
similar. Similar associations of two provinces may con-
tain the same or closely related species, as with certain
grasshopi>ers which range in both northeastern and west-
em coniferous provinces {D: 173). But these same asso-
ciations contain also plant species in common, so that
irregularities of range are no greater in animals than in
plants.

2. Distribution Within the Province: Distribution of
Plants and Animals into Communities, — It is seen that
plant and animal species may correspond closely in geo-
graphic range. There may be also more local corre-
spondence in distribution. The plant community has
been found by the writer to be the convenient index of the

Digitized by

Google

434 THE AMERICAN NATURALIST [VoL.XLVni

area of the habitat for animals. It has been observed,
in an area in Michigan, that grasshopper species corre-
spond closely in local distribution with plant communities
(/)). There is evidence that local distribution of ani-
mals is seldom promiscuous as a result of motility (D:
159). It appears also that the local variability of envi-
ronmental conditions within the area of the climatic plant
community is suflSciently great, usually, to supply all
necessary conditions for a large number of animals, so
that the limits of the plant community need not be
passed, ordinarily.

The animal community of the area may be thus, in large measure, self-
contained, and coextensive with the plant community (D: 161).

One of the problems of plant ecology has been the
differentiation of plant communities or associations.
Mere comparison of lists of species is not sufficient; rela-
tive abundance of various species must be considered as
well. Animal assemblages in contiguous areas must be
separated in the same way. Given two adjoining habitats
differing in plant population, it has been found that, in
addition to differences of animal species,® there are also
differences of relative abundance in those animal species
common to the two areas {D: 154, 167).

The local area of a plant community is determined by
(1) local distribution of the physical environmental com-
plex, and (2) influence (competition, etc.) of adjoining
plant communities. Local area of the animal community
depends upon (1) local distribution of physical environ-
ment, and (2) local distribution of vegetational environ-
ment, the latter being uniform over the area of the plant
community. Contiguous areas differing in physical and
vegetational conditions will be expected to differ also in
animal population, in a degree comparable to that of the
differences in environmental conditions.

Physical habitats, and plant communities, sometimes
alternate, sometimes intergrade ; it is not unreasonable to
expect accompanying alternation or intergradation of

« Differences in species, both plant and animal, are accompanied l>y dif-
ferences in ecological constitution.

Digitized by

Google

No. 571] TERRESTBIAL ASSOCIATIONS .435

animal populations. Certain of the animal assemblages
of sand habitats, as studied in central Illinois, intergrade ;
others, as oak forest and bunch-grass, differ radically.

The above considerations, if correct, appear to signify
that, in ordinary climatic development of plant and ani-
mal life in temperate land environments, the area of the
animal assemblage is that of the plant assemblage, both
resting basically upon the physical environment. The
plant and animal assemblages are therefore coextensive
parts of a biotic association, composed of both plants
and animals, and this association as a whole constitutes
the real terrestrial community of living organisms.

B. Local, Eelations of Plant and Animal Assemblages
(Eelations Within the Association)

The more intimate relations between plants and ani-
mals are ^een in the detailed study of a single associa-
tion. The bunch-grass association of sand prairie is
selected for illustration (E: 68).

1. Similarity of Ecological Type of Plants and Ani-
mals.— Shelf ord has shown {A: 593-594) that animals
and plants may evince ecological similarity by similar
response to the same general environmental conditions,
behavior responses in animals^ corresponding to struc-
tural responses in plants,® so that mores of the animal
may be in accord with growth-form in the plant. Shel-
ford states {B: 87) that ''plants and animal communities
are in full agreement when the growth-form of each
stratum of the plant-community is correlated with the
conditions selected by the animals of that stratum.''

In the bunch-grass there is general agreement, ac-
cording to this criterion. The herbaceous stratum is oc-
cupied mainly by tuft and mat plants — bunch-grasses,
cactus and a few half-shrubs. Associated with the tuft or
mat growth-form is the sedentary mores of the plant-
inhabiting animals (leaf -beetles, stem-borers, ambush-
bugs, etc.). A considerable proportion of ground surface

. 7 Or motile organisms, cf. C: 305.
• Or sessile orjj^nisms.

Digitized by

Google

436 THE AMEBIC AN NATUBALIST [Vol. XLVEI

is bare sand; in the interspaces between the dommant
plants are slender annuals (interstitial plants), and here
are also found animals of the roving mores of the ground
stratum (interstitial animals). Many of these are swift-
running and predaceous (six-lined lizard, tiger beetles,
lyeosid spiders).

Correspondence in ecological type of plants and ani-
mals in the bunch-grass is not complete in several re-
spects. Shelf ord mentions types of disagreement {B:
88; C: 306-308), and there is a further important kind of
disharmony, in mixed associations, due to presence of
diverse types of plants and animals {D: 163). Mixed
associations are quite frequent in forest border regions,
and in the transition area between two provinces. The
plant and animal assemblages of a given habitat, partic-
ularly if climatic and extensive, are usually in general
ecological agreement, and the exceptions are likely to be
infrequent or temporary (Shelf ord, B: 88).

2. Relative Dependence of Plant and Animal Assem-
blages.— There is evidence that the agreement of plant
and animal assemblages of terrestrial associations is
often a matter of accommodation on the part of the
animal assemblage. In the early stages of development
of vegetation, local physical conditions dominate; in
later stages the vegetation assumes the type determined
by climatic conditions, and exerts nearly complete con-
trol over local physical factors. In established associa-
tions, therefore, the locally dominating environmental
feature is the vegetation. Shelford states that in the
several associations of a successional series, the domi-
nating animal mores are correlated with the dominating
conditions {B: 94) and that, as the forest increases in
density, the animals make use of the vegetation in in-
creasing degree, particularly for breeding-places, and as
places of abode {B: 90). Many grasshoppers of open
grassland depend upon a particular kind of soil for egg-
laying, while those of closed forest lay eggs in fallen
logs — a condition of the plant environment (D: 163).

Digitized by

Google

No. 571]

TEBBESTBIAL ASSOCIATIONS

437

The sand-prairie vegetation is in an intermediate stage,
certain animals depending chiefly on the presence of
loose bare sand, others on the bunch-grass vegetation.
With development of bunch-grass into closed grassland,
the interstitial animals are eliminated. The animals of
established associations, while in accord with climatic
physical conditions, are perhaps more intimately affected
by vegetation conditions. Since established associations
are very much more extensive than primitive associa-
tions, the importance of vegetation as a dominating part
of the environment for animals becomes apparent, and
we may conclude that the character of the plant assem-
blage determines, to a large extent, the ecological type
of the animal assemblage.

3. Correspondence in Distribution within the Associa-
tion.— The uniformity of physical and vegetational con-
ditions is only relative. There are spots in the bunch-
grass association in which local invasion of blue grass
has occurred, darkening and binding the soil. In such
partly humified situations, small colonies of the corn-
field ant, not occurring elsewhere in the bunch-grass {Er
57), have been found. There are also areas some few
feet in diameter in which the bunqjies of grass are few,
small and scattered. In these relatively bare patches the
abundance of interstitial animals is greatly increased.
More direct relations are seen in the case of animals
associated with particular species of plants. Within the
association, any animal species, like any plant species,
may be distributed generally throughout the area, or it
may be restricted to a part of the area characterized by
a slight environmental difference, or it may occur in
scattered parts of the association, characterized by
scattered local differences {D: 168). There is evidence
that, in so far as the vegetational environment is con-
cerned, distribution of animals within the association is
usually a direct function of similar distribution of plants.

4. Uniformity of Species Composition of Plant and
Animal Assemblages. — It has been seen that plant assem-

Digitized by

Google

438 THE AMEBIC AN NATURALIST [Vol. XLVHI

blages of definite ecological type, as regards growth-
form, etc., are regularly accompanied by animal assem-
blages of similar ecological type, as regards mores. In-
terest attaches also to the problem whether associated
plant and animal assemblages show definite species
relations.

One familiar with a certain association, who visits a
representation of that same growth in a different part
of the same climatic region, will be struck with the fact
that a large proportion of both plant and animal species
is well known, while a certain proportion, perhaps con-
siderably smaller, is new to him. The writer has been
impressed with the similarity of the plant and animal
populations of the sandhills of central Nebraska and of
eastern Colorado, to those of the sand prairie of central
and western Illinois, despite the fact that certain species
are not common to the two areas. Tiger-beetles, blow-
snake, grasshoppers, box-turtle, six lined lizard, western
meadow-lark, white-footed mouse, among the animals;
prickly-pear, lead-plant, bunch-grasses, sand-bur, sand
evening primrose, among the plants; are represented' in
the two areas either by the same or by closely related
varieties and species. There are no yuccas or sand-sages
in the Illinois sand prairie, no lizard Holbrookia nor
lubber-grasshopper Brachystola; and there are certain
eastern species not found in the western sandhills. But
on the whole the species (particularly the important
species) common to the two areas are more numerous.
This is the more remarkable in view of the fact that dis-
tribution of sand prairie is discontinuous, the largest,
nearly uninterrupted gap being several hundred miles in
extent. Many of the animals, as well as plant species, of
dry mixed prairie-grass in loamy soil, are the same along
the mountain-front in Colorado (Vestal, 19146) as in
north-central Illinois. The likenesses become much more
impressive as distance is decreased.

Absolute identity of species composition, where large
numbers of species are involved, is an ideal condition.

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 439

never actually attained. No one can say just what pro-
portion of species-in-common is necessary for two
growths to be said to represent the same association. In
addition to likenesses and differences of environment, of
aspect, and of history, which must be weighed, the differ-
ent plant and animal species vary so mtch in importance
in the association, in physiological variation and in range
of environmental tolerance, that associations can hardly
be separated or placed together on a statistical basis. A
comparison of species is fair if the following kinds of
plants and animals are left out of consideration; (1)
those of limited range within the climatic region or prov-
ince, including species belonging more properly to other
provinces; (2) those of very indefinite habitat-relations,
which are found in nearly any kind of habitat; (3) those
of special restricted habitats, which may be scattered
about in many kinds of associations, as moist dead wood,
in which particular fungi, beetles, perhaps snails, myrio-
pods and pill-bugs, are usually found ; or as excrement of
grazing animals, in which certain molds, certain dipter-
ous and scarabaeid larvae, etc., regularly occur, irrespect-
ive of surrounding conditions; (4) invaders from near-
by associations; (5) ruderal and introduced species; and
possibly one or two other groups. The second and third
groups may be called the irregular element; the fourth
and fifth may be known as the derived element. While
these groups make a formidable list, their representa-
tives constitute usually a very small proportion of the
organisms of the association. The other organisms, and
some of these, follow habitat-differences, as represented
in different associations, very closely.

Since hardly any two species are identical in habitat-
relations, geographic and even local variation must be
looked for, but since many species resemble each other
more or less closely in general ecological relations, there
come to be recognized certain ecological groups of spe-
cies, each characterized by a general type of growth-form
in plants, or by a general kind of mores in animals, and

Digitized by

Google

440 THE AMEBIC AN NATURALIST [Vol. XLVIII

these groups may be considered to be small or large,
according as we emphasize minor differences or general
likenesses.

Now within any limited region (let us postulate first
an area removed from the influence of an adjoining prov-
ince) there are only a limited number of ecological
groups, of growth-forms of plants, and of mores among
animals, each group represented by a limited collection
of species. Each habitat within this restricted area will
be characterized by definite physical conditions, and with
these will be correlated certain growth-forms of plants
and certain mores of animals, each represented by as
many of the species as can migrate into and survive
within the area, as determined first by capabilities of
migration and by habitat-selection, and second by inter-
relation of species and of individuals. It follows that
physical complexes which are alike will become populated
with similar complexes of ecological groups, represented
by similar collections of plant and animal species, and
that unlike physical areas will be occupied by different
combinations of ecological groups, and will be composed
of different species. Two areas within this region which
have similar physical conditions and similar plant
growths will be expected to have a large number of ani-
mal species in common, although direct relations between
species of animals and species of plants obtain only
rarely (between comparatively few associated plant-and-
animal pairs). It is to be noted that species composition
of the animal assemblage varies proportionately no more
widely than does that of the plant assemblage.

No terrestrial continental region is sufficiently isolated
to be free from influence of surrounding areas, and since
the influences are different from different directions, and
since there is continual change of physical conditions, and
of range and abundance of plant and animal species,
there must be more or less local and geographic varia-
tion of species composition within similar but separated
habitats. Geographic variation is wider with distance.

Digitized by

Google

J^p. 571]

TEBRESTBIAL ASSOCIATIONS

441

because the geographic and physiographic complexes
vary geographically, as well as the entire collection of
plant and animal species which may invade the habitat.
Within the area of the climatic province, however, or at
least within the area of a subregion of the province,
climatic, physiographic and biotic complexes are likely
to be relatively constant, that is, likenesses of two areas
are likely to be greater and more striking than differ-
ences. Within the province or. subregion, therefore, it
is to be expected that species composition of association
of closely similar habitats will be relatively constant
Particular plant and animal assemblages will be found
together, both associated with a particular habitat. Field
observation bears out these expectations.

Conditions within the transition zone between two
climatic regions or provinces are much more complex
than in an area in the middle of a sub-region or province ;
climatic and physiographic conditions vary to wider
extremes and are less stable ; the total number of species
near enough at hand to invade a given habitat is much
greater. Mixed associations, often transitional as re-
gards physical conditions, are composed of representa-
tives of both provinces. Animals of a particular associa-
tion of one province, may be found with plants of a
similar or equivalent association of the other province.
When three geographic elements are represented, as at
the southern end of Lake Michigan (cf. C, and Vestal,
1914a), the complication of conditions is extreme. Even
here, on the dry sand of old lake beaches, fairly typical
representations of sand prairie can be seen ; and though
such habitats are shared with deciduous forest associa-
tions, and with associations of the northeastern coniferous
forest province, and with mixed associations, the bunch-
grass growth can still be recognized in dry shifting sterile
sand, with bunch grass plant species, and bunch-grass
animal species. The tendency towards uniformity of
association of plant and animal assemblages is even here
to be made out.

Digitized by

Google

442 THE AMEBIC AN NATURALIST [Vol. XLVHI

If the foregoing considerations relating to relations
between plant and animal communities are correct, the
thesis mentioned at the beginning of part V would seem
to be justified, though the evidence is far from complete.
Plant and animal assemblages are mutually interdepend-
ent; the plant assemblage dominates in established
associations. Plant and animal assemblages correspond
in geographic distribution, in distribution into commu-
nities, and in more detailed distribution within the
habitat. They are made up of ecologically similar groups
correlated with the same physical conditions or with
each other. Though there are few direct relations be-
tween particular species of plants and animals, it so
happens that within any restricted region, particular
collections of animal species come into regular associa-
tion with particular collections of plant species, the spe-
cies composition within the habitat exhibiting a greater
or less degree of uniformity, except for minor irregular
and derived elements. The more restricted, or uniform
in biological conditions, this region is, the greater the
uniformity of the collection of species. Climatic and ex-
tensive associations, and established associations, show a
greater degree of uniformity than local or primitive
associations.

VI. SUMMARY AND CONCLUSIONS

The discussion is based principally upon the writer's
study of prairie associations, the bunch-grass associa-
tion of sand prairie in Illinois being chiefly used for illus-
tration. Internal activities of the association are a com-
plex of activities of all the organisms. Environmental
influences are of three classes, physical, plant and animal.
The characters of plants and animals are interpreted in
their relation to these influences. Characters of plants
may be classed as structural, physiological, biographical
and numerical. Animals have, in addition, behavior or
psychological characters. These groups of characters
are intimately related, one to another. The relations of
the animal to its animal-environment are of two kinds,

Digitized by

Google

No. 571] TERRESTRIAL ASSOCIATIONS 443

social and antagonistic, the latter relations being with
food-species, competitors and enemies. Correlations of
the various kinds of characters with relations involving
food, competition and enemies, are given. According to
its ecological constitntion, each organism finds a status
in the association, the whole being a self-contained and
self -regulating system of activities.

Dependencies within the association are concerned
mainly with sources and interchange of material and
energy. Dominant plants (the most influential species)
are those most intimately correlated with physical en-
vironment, as indicated by aggressiveness, abundance,
frequence, size, etc. Domiaant animals are most numer-
ous among phytophagous forms. Dominance in an ani-
mal species includes dependence of other animals upon it
(for food) plus the ability to thrive in spite of the drain
upon its numbers. The degree of specialization of be-
havior is a convenient index of the relative influence of
animals in the association. The dominant animals are
moderately specialized, and carry on the ordinary work
of the association. The highly specialized animals make
use of space otherwise unoccupied, and food material not
available to other species, or not taken by other forms.
Least highly specialized animals act as a check upon
undue departure from biotic equilibrium.

The association may be divided into minor groups of
organisms, both in space and in time. Space-division is
vertical, resulting in strata, and horizontal, resulting in
sub-habitats of greater or less magnitude. The strata
and sub-habitats present a larger or smaller degree of
discontinuity and of internal variability. Time-distribu-
tion is diurnal, seasonal and annual. There are also
time- variations produced by variability of weather condi-
tions and by oscUlatory disturbances.

The relations between plant and animal assemblages
have long been known, in a general way, to be intimate.
Plants and animals agree in similar response to common
environmental influence, and in types of geographic dis-

Digitized by

Google

444 THE AMERICAN NATURALIST [Vol. XLVm

tribution. Upon investigation, it begins to appear that
plant and animal assemblages are coextensive parts of a
biotic association, composed of both plants and animals,
and this association as a whole constitutes the real ter-
restrial community of living organisms. Plant and ani-
mal assemblages are mutually interdependent ; the plant
assemblage dominates in established associations. Plant
and animal assemblages correspond in geographic dis-
tribution, in distribution into communities, and in more
detailed distribution within the habitat. They are made
up of ecologically similar groups correlated with the
same physical conditions or with each other. Though
there are few direct relations between particular species
of plants and animals, it so happens that within any
restricted region, particular collections of animal species •
come into regular association with particular collections
of plant species, the species composition within the habi-
tat exhibiting a greater or less degree of uniformity,
except for minor irregular or derived elements. The
more restricted in area, or uniform in biological condi-
tions, this region is, the greater uniformity of the collec-
tion of species. Climatic and extensive associations show
a higher degree of uniformity than local or primitive
associations.

VII. REFERENCES

r> Special Befereiices

(A) Shelf ord, V. E.

19116. Physiological Animal Geography. Jour, Morph., 22: 551-
618.

1912a. Ecological Succession. IV. Vegetation and the Control
of Land Animal Communities. Biol, BuU., 23: 59-99.

1913. Animal Communities in Temperate America. Geogr. Soc.

of Chicago, Bull. No. 5, p. 362. Chicago.
(D) Vestal, A. G.

1913a. Local Distribution of Grasshoppers in Relation to Plant

Associations. Biol. Bull, 25: 141-180.

(E)

1913&. An Associational Study of Illinois Sand Prairie. Bull.

III. State Lab. Nat, Hist., 10: 1-96.
Other Articles Cited
Adams, C. C.

1913. Guide to the Study of Animal Ecology, p. 183. New York.

Digitized by

Google

No. 571] TEBBESTBIAL ASSOCIATIONS 445

Craig, W.

1908. North Dakota Life: Plant, Animal, and Human. Bull Am.

Geogr, Soc, 40: 321-332, 401-415.
Ellis, M. M.

1914. Fiahes of Colorado. Univ, CoXO. Studies, 11 : 1-136.
Forbes, S. A.

1880. On Some Interactions of Organisms. Bull. III. State Lab. Nat.

Hist., Vol. 1, No. 3, pp. 3-17.
1883. The Begulative Action of Birds upon Insect Oscillations. Bull.

m. State Lab. Nat. His., Vol. 1, No. 6, pp. 3-32.
1887. The Lake as a Microcosm. Beprint from Bull. Sci. Assoc, of
Peoria, III., pp. 1-15.

1909. The General Entomological Ecology of the Indian Com Plant.

Am. Nat., 43: 286-301.
Gates, F. C.

1912. The Vegetation of the Beach Area in Northeastern Illinois and

Southeastern Wisconsin. Bull. HI. State Lab. Nat. Hist., 9:

255-372.
Gleason, H. A.

1910. The Vegetation of the Inland Sand Deposits of Dlinois. Bull.

Ill State Lab. Nat. Hist., 9: 23-174.
Hart, C. A., and Gleason, H. A.

1907. On the Biology of the Sand Areas of IlHnois. Bull III State

Lab. Nat. Hist., 7: 137-272.
Nelson, E. W.

1909. The Babbits of North America. Bur. BioL Surv., U. S. Dept.
Agr., N. Am. Fauna No. 29, pp. 314.
Pound, R, and Clements, F. E.

1898. The Vegetation Eegions of the Prairie Province. Bot. Gas., 25:
381-394.
Buthven, A. G.

1908. The Faunal Affinities of the Prairie Begion of Central North

America. Am. Nat., 42: 388-393.
Schimper, A. F. W.

1903. Plant Geography upon a Physiological Basis. Oxford.
Shelford, V. E.

1911a. Ecological Succession. I. Stream Fishes and the Method of

Physiographic Analysis. Biol Bull, 21: 9-35.
1912&. Ecological Succession. V. Aspects of Physiological Classifi-
cation. Biol Bull, 23: 331-370.
Vestal, A. G.

1914(1, The Status of Prairie Associations in the Southern Beach Areas

of Lake Michigan. Jour. Ecology. (In press.)
1914. Prairie Vegetation of a Mountain-front Area in Colorado. Bot.
Gae. (In press.)
Transeau, E. N.

1903. On the Geographic Distribution and Ecological Eelation of Bog

Plant Societies. Bot. Gajg., 36 : 401 et seq.
1905. Forest Centers of Eastern America. Am. Nat., 39: 875-889.

Digitized by

Google

SHOETEE AETICLES AND DISCUSSION

ANOTHER HYPOTHESIS TO ACCOUNT FOR DR.
SWINGLE'S EXPERIMENTS WITH CITRUS

The results of the cross-breeding experiments with forms of
Citrus by Walter Swingle have given rise to quite a number of
different hypotheses, to account for the facts observed.

The facts are simply these. All the different forms of Citrus
used in the experiments, Citrus trifoliata, th€ lemon, orange and
other citrous fruits have, so far, proved to reproduce their own
type through seed.

Nevertheless, the plants raised from one single cross are ex-
ceedingly different among themselves. And yet, all these new
forms, for so far as tested, have proved truly to reproduce their
own kind only, if sown.

The theories offered to account for these facts are rather com-
plex. So far, we have not seen the simple hypothesis which we
want to add to the others.

The fact, that the Fi from almost every cross between types of
Citrus is multiform, can only be accounted for on the assumption,
that the parent plants are impure (heterozygous) for a number
of genes. The difficult question is this : how can a tree, impure
for a number of genes, produce seed which always only repro-
duces the type? We know, that if a plant reproduces itself
by an asexual method, all its daughter plants are pure for
those genes in respect to which it was pure, impure for those
genes for which it was impure. Is it possible that in these
trees the seeds normally produced are not derived from a union
between two normal gametes? In Citrus, with its adventitious
embryos, this is very well possible. If the forms of Citrus used
by Dr. Swingle are self-sterile, the seeds normally produced by
these trees, are not produced by the union of two gametes, but
as buds, asexually.

This hypothesis, that the Citrus used are self -sterile,- and that
the seeds normally produced, are produced asexually, fully ac-
counts for all the facts. All the daughter plants from un-
crossed seeds are genotypically identical with the mother plant,
as in all clones. On poUenization by another tree, normal seeds
are produced, each the result of the union of two real gametes.

446

Digitized by

Google

No. 571] SHOETEB ARTICLES AND DISCUSSION 447

These seeds contain different combinations of the genes, for
which the parent plants are impure, as normally. The Fj gener-
ation for this reason becomes as diverse as such generations
always are, if the parents are impure for numerous genes.

But these daughter plants^ although impure for a number of
genes, can, because of their self -sterility, in their turn only pro-
duce seed asexually and therefore their offspring will be like
themselves.

It should not be diflBcult to test our hypothesis. It seems
easier to find out, whether the seeds produced without crossing in
jCitrus contain the embryo formed by fertilization of the em-
bryo sac, or embryos formed adventitiously by the adjacent tis-
sue, than to test any of the other theories, which assume a pecul-
iar behavior of the chromosomes.

Our hypothesis, that a variable Fj, of only true-breeding
plants (from the union of two true-breeding forms), results
from habitual self-sterility and asexual production of seed, with
real fertilization in the case of a cross taking place, not only
accounts for the facts found by Swingle, but also for those found
by Kosen with Erophila vema. These facts were somewhat dif-
ferent. The Fi plants were all identical, and somewhat inter-
mediate. They gave rise to a variable Fj generation of which
all the plants bred true to their type. These facts can be ex-
plained on the assumption, that Erophila vema is self-sterile,
and that, in the absence of cross-fertilization, unfertilized egg-
cells develop parthenogenetically. Such F^ plants, which are
impure for a number of genes, should therefore produce as many
different kinds of F^ plants, as there are female gametes pro-
duced, and in the same proportions. In the case of such a plant
being impure for two genes, we should expect it to produce
plants of the four different types, not in the usual proportion of
9:3:3:1, but in equal proportions, 1:1:1:1. The F^ plants
from such seed could only be pure for all the genes present.

It would be possible in Erophila vema to find out whether F^
plants, impure for two genes, produced daughter plants of the
four kinds, AB, Ab, aB, and ah, in the proportion of 9 : 3 : 3 : 1,
or in proportion 1:1:1:1, and thus to test our hypothesis.

To find out, whether it is possible, that aC plant, impure for a
number of genes, produces a variable F2 generation of only
completely homozygous plants, we have begun a series of experi-
ments with squashes. Some hybrid plants have not produced a

Digitized by

Google

448 THE AMERICAN NATURALIST [Vol. XLVIU

single fruit from carefully sealed female buds, others have given
plenty of empty fruit, but some hybrids have produced several
fruits, full of viable seed. If this seed is formed by the par-
thenogenetic development of unfertilized normal egg-oells, as
we have reason to believe, we expect to raise a variable P, gene-
ration of exclusively homozygous plants. If these seeds have
developed by apogamy, or any other asexual process, we expect
to obtain a second generation consisting exclusively of plants
like the original hybrids. Thus we will have a non-cytological
test to decide between apogamy and true parthenogenesis.

A. C. Hagedooen,

A. L. Hagedoobn

BussuM, Holland,
March 18, 1914

Digitized by

Google

Two Important Biological Journals

The Journal of Qeneticfi

Edited by
W. BATESON and R. C. PUNNETT
A periodical for the imblioation of records of original reeeaieh in heredity, yariation, and allied topioe.
The JoUBNAii also oonlains, from time to time, articles summarizing the existing state of knowledge in the
▼arions branches of genetics. The numbers are fully illustrated, and, when the subject-matter demands it,
oolored plates are freely employed. The Journal is issued in parts as material accumulates, and a volume
(appearui^, so fiur as possible, annually) consists of four sack parts. SubsGdption price, $7.50 a volume ;
single copies, |2.60.

Biometrilca

Edited by KABL PEARSON
A Journal for the statistical study of biological problems, appearing about four times a year. A volume
containing about 600 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
nine preceding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
craniometiy, medicine, heredity, etc. Subscription price, $7.60 a volume ; single copies, 12.50.

Prices on Imm& numbers of these Journals vary, and postage from London is charged on back volumes
and single copies.

PuUishcd in America for the Cambridge University Press of England by

The University of Ohiica^ro Press
Chilcasro^ Illinois

University Control

By J* tloKiaax CattblL| Professor of Psychology in Oolnmbia Uniyendty
TocsUmt with a series d Two Hundred and mnotj-nine Unsigned Lstteis by Tissding lien d Beisoss
( Agadenio Positions and Artides by Josnni Jastbow, Gnonon T. LAnn, Jokv J. Bm?— sow, J. B«
J. lidEani GAVimuL, Gaoaan 11. Svnancni, Stbwasv Pavoh. Jom Jat GBAniAJV, Jaksb P.
I and Jaoob Gould ScBinacAH.

A great Tsriety d questions conoetning gsnwal anivsnity administration an dealt with in an original and
hstoftd waar.— JVolMri.

Thesa quotations and essmples are taken from Professor Cattell's inf osmed and thoroogh dlseosrion of the
snbjeet d tmirersity oontrol, a sobjeet upon which he has had much to say d late, finding occasion for eamtis
wltfeism off esisling American conditions, and standing as the champion d an academic demoeracsr and a teael^
lag ptufession vpon wbkh a man majr enter wHhoot forieitins his self-respect. — Tk$ DidL

BsntsBcea and paragraphs that betokso the espert, highly-trained mind, the suggestions that come to fs-
ftmk a d teU ns that a new day is about to dawn in eduoational writing.— Tfte BotUm EtmSng TtomteHpL

SCIBISCB AND EDUCATION

A seriem of irolnmes for the promotion of aeienti&e reseMreb and eduemtional progreaM
YOLXJMR I. The lymndations of Science. By H. PoofOABi* Containing the anthoiised English translation

by Gaoaoa Bsnoa Hautbd d "Bdenoe and Hypothesis," "The Value d Beienoe," and "Seionee

and Method." Price, $3.00 net.
YQLCJMB n. Medical Sesearch and Education. By BioHABn 11. Piabob, Wiuaam H. Wnum, W. H.

HowsL&to FBAXXLai P. Mall, Lbwblltb F. Bingnn, GBABLas 8. Minor, W. B. Camkoii, W. T.

OonmcnjfAV, TwoaALP Smiir, G. N. SfawASV, C. M. Jaoksoh, B. P. Ltom, jAiaa B. HannoK, Jobs

M. DonaoM, C. R. Babdbbb, W. OpbOls, & J. Mblisbb, Jambs Ewimg, W. W. Ebbb, Hbbbt H. Doxalp-

aoir, CSnoRiAjr A. Haana, and Hbmbt P. Bowpitob. Price, $3.00 net.
TOLXJMBIIL Uairenity Control, i^ow Eaad^. Fo^Ms+i^i. Prie$. %$M ntk

aARRisoN, N. Y. THE SCIENCE PRESS Lancaster* pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The American Naturalist

A MoBllilr Joonial. MtaUialied in 1867, D«TOted to the Adrancemeiit ol the BiologiMi Sriiiim
with Special Raf ereaco to tfao Faetors ol Orvanic EvolotioM and HaraditF

CONTENTS OF THE JANUARY NUMBER

A Genetic Analyiii of the Changes produced by
Selection in Sxperlments with Tobacco. Pro-
fessor B. M. Bast and H. E. Hayes.

Gynandiomorphous Ants, described during the De-
cade, 1903-1918. Professor WiUiam Morton
Wheeler.

Shorter Articles and Discussion : On the Results of
Inbreeding a Mendelian Population— A Correc-
tion and Extension of Preyious Conclusions.
Dr. Baymond Pearl— Isolation and Selection
alUed in Principle. Dr. John T. Gnlick.

CONTENTS OF THE FEBRUARY NUMBER

Some New Varieties of Bats and Gaine«-idgs and ttieir
Belations to Problems of Color Inheritaaee. Pro-
fessor W. B. Castle.

•• Dominant" and " Kecess.ye" Spotting in lOce. C.
c Little.

On Difler^tialMortaUty with respect to Seed Weight

occurring in Field Coltures of Pisum sattTom.

Dr. J. Arthur Harris.
The InheritaBce of a Beeorring Somatic Varlatloa

In Variegated Bars of Maise. ProfeMor B. A.

Bmerson.

Restoration of Edaphosftunis erndger Cope. Pro.

fessor B. a Case.
Shorter Articles and Discussion : Humiditj— a

Neglected Factor in BnTironmentftl Work. Dc

Ftank B. Luts.

CONTENTS OF THE MARCH NUMBER

The Effect of Extent of Diitrlbntlon on Speeiation.

Asa C. Chandler.
Biology of the Thysanoptera. Dr. A. Franklin Shull.
Shorter Articles and Correspondence : The Endemic

Mammals of the British Isles. Professor T. D. A.

OockerelL

Notes and Literature : Swingle on Variation in Fi
Citrus Hybrids and the Theory of Zygotaxis.
Dr. Orland E. White.

CONTENTS OF THE APRIL NUMBER

The Origin of X Capsella Bursa pastoris araehnoideab
Dr. Henri Hus.

Biology of the Thysanoptera. n. Jh, A. Fraskltn
Shull.

Shorter Articles and Discussion: Barriers as to Dis-
tribution as regards Birds and Mammals. Joseph
Grinnell. Yellow Varieties ot Bats. ProflBSSQr
W. E. Castle.

Notes and Literature : Heredity and the Tnfhieiioe
of Monarchs. V. L. K.

CONTENTS OF THE MAY NUMBER

Betoparasitesof Mammals. Professor Vernon Lyman

Kellogg.
Begeneration, Variation and Correlation in Thyone.

Professor John W. Scott.
Shorter Articles and Discussion : Terms relating to

Generic Types. Dr. O.F.Cook.
Notes and Literature: Linkage in the Silkworm

Moth. A. H. Sturterant. Nabours^s Breeding

BxperlmentB with Grasshoppers. John a Dexter.

CONTENTS OF THE JUNE NUMBER

Species-bailding by Hybiidintion and Matati<m. Pro-
fessor John H. Gerould.

Heredity of Bristles in the Commtm Greenbottle Fly—
A Study of Factors goTemingDistiibatioiL FhineM
W. Whiting.

Physiological Correlations and Climatic ReaotionB in
Alfalfa Breeding. Geo. F. Freeman.

Taxonomy and Erolution. By X

Shorter Articles and Diacosrion: NabonnPs QnMS-
hoppera, Multiple Allelomorphism, Tan^«e and
if^fiiAiiing Terminologies in Genettos.
W. ELCsstle.

Single Number 40 Cents Yeariy SulMcriptioii, $4.00

Tlie NATUIt/aJST wffl U Mill to iB«w MbMribMTS lor foor mcmtibs Ib^

THE SCIENCE PRESS

GarriMm, N. Y. Sub^Station 84 : NEW YORK Uneastcr. I^u

Digitized by

Google

TOL. ZLVm, VO. 672 ATTOTTST, 1914

THE

AMERICAN
NATURALIST

A HOHTHLT JOmUTAL

Itofotad to fhe Advancemdnt of fhe Biolog^eal Soieii06f wift

Special Beferenoe to fhe Paotors of Erolutiom

COVTBVTS

Pag9 * 1

I. Koltlple AlUlomorpha In Mloe. Profeisor T. H. Mob0an - - - - 449 J

IL Tldrteen Tears of Wheat 8«leotloii. T. B. Hutohbson - - - - .459 ;

m. Pattern Derelopment In iraininala and Blrda. Glotbb M. Allxn - - 467
IV. The Meadow Jumping Honae. Dr. H. L. Babcook ..... 430

y. Shorter Artlelea and IMienaslon : Studies on Inbreeding. Dr. Batmomd J

Pbarl. Parallel Mntations in (Enothera biennis L. Dr. J. STOMPS, Dr. ' i

Bbablby M. Davis. The Theoretical Distinction between Multiple Allelo- ;

morphs and dose Linkage. ProfSsssor T. H. Mobgan, Professor W. E. {

Castlb -------------- 491 I

VI. Motes and Literature: Biometrics. Dr. Baymond Pbabl. A New Mode ef \

Segregation in Gregory's Tetraploid Primulas. HBBMAKif J. Mullbb - 500

THE BODSNOE PBEBB

IiAHDABTEB. PA. OABBI80 V. V. T. |

i
HEW TOBK : 8UB4n ATIO V 84

Digitized by

Google

The American Naturalist

1188 iBteRded for poblloation and books, etc., iotendod fbr roviow thoald bo

•oat to the Editor of THE AMERICAN NATURALI8T, fiarriooa-oa-Hadooa. Now York.

Short artlolet eontalalag ouamarieo of roooaroh work boariag oa tho

J»robleaio of orgaaio ovolatioa aro oopeelally welooafe, aad will bo fllvoa proforoaoo
a pablioatioa.

Oao huadroQ roprlato of coatribatloat aro oappllod to aathoro froo of ehargo.
Farthor roprlato will bo oappllod at coot.

8aboorlptloao aad advortlooaioato oboald bo ooat to tho pabllohors. Tho
oaboorlptlon prioo Is foar dollars a yoar. Foroiga postago Is fifly ooats aad
Caaadlaa pootago twoaty-flvo ooats addltloaal. The ohargo for singlo eoploo lo
forty ooats. Tho advortislag ratos aro Foar Dollars fbr a pago.

THE SCIENCE PRESS

Lancaster, Pa. Qarridon, N. Y.

NEW YORK : Sub-Station 84

Sntered u teeond-olui mattor, April 2, 1906, at the Post Office at Laneaiter, Pa., under tlie ▲•! of

Oontrew of Maroh 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Particulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a aet of BIRDS OF AMERICA by J. J. Auddxn.
7 or 8 volumes, pleaae report, atating cash price, stat-
ing condition, binding and dates of Tolumes.

P. C. HARRIS,
Box 2244 Boatoa, Maasachiisetta

For Sale Entire

An important oollection of Indian Birds'
EggS; containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste
viridanus), Red-browed Finch (Calla-
canthis burtoni), Tibetan Sand-Grouse
(Syrrhaptes tibetanus), Spotted Sand-
Grouse (Pteroclurus senegallus), Ibis-
bill (Ibidorhynchus struthersi), White-
bellied Minivet (Pericrocotus erythro-
pygius), Bearded Vulture (Gypaetus
barbatus), numerous Cuckoos with their
hosts, etc., etc.

Particidars and full list may be had
from

W. F. H. ROSENBERG

57 Haveratock HiU London, N. W.

Marine Biological Laboratory

Wooda Hole, Mass.

INVESTIGATION

MmlOfYmmr

TmeffMbt fai reMsnh in godtK,
Bmbc70loc7t Phjsiology aad B«^
an^. 0eTei&ty-«ix ptirata labors

INSTRUCTION

Mr 1 to Aacmt 11.

1914

tofie^ $100 each for aoi arm ^tam
months. THrtj tables are svafr^
able for beginiiwB in researoh aiw
aedra to work under the dhaedcn
of members of the staff. Tba les
f or sDob a tabis is asOOO.

Comaes of laboratory

with leotnree are olfared to lasarts

olocy, Mo ,

of & Alna, Physieloo i
Boolocr onEtrina SIfaad aai L^
Vegstsdoo. Each ooims raq^ns
^ fuH tims cf ths stod«tt>rnF£
1.00. A lecture ooans
' Aspects of
isi

SUPPLY

DEPARTMENT

OpMi tha Eatlra Year

IMag, sad to

Preeenred material „
antouds and of Afgas, 1
worta and Moseea '

<Jasiworiu or for ^_^

Livtog mateiisl f nmiibad in l.
ss ordered. Pri«e liats of
lofiesl and Botanfaal

sent on appHeation. Stc..

is desbed. For prioe I e*s aad afl
informatioa r^srdinc msfM.
addreei

GEO. M. GRAY, Curator. Wooda Hoto. w— ,

Ths annuel announoement will be sent on appHontioa to

Tha Diisctor» Ifadaa Biologiesl Ubocatoty, Woada

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVIII August, 19U No. 572

MULTIPLE ALLELOMORPHS IN MICE

PROFESSOR T. H. MORGAN
Columbia University

Some breeding experiments with mice that I have been
carrying on during the last two years have shown that
yellow, gray gray-belly, gray white-belly and black are
allelomorphs. To this series a fifth allelomorph may pos-
sibly be added which for the present naay be called new
gray. This quadruple (or quintuple) system of allelo-
morphs fulfils the conditions of a multiple series in that
only two of the allelomorphs can exist at the same time
in any individual. In other words, a mouse may be pure
for any of these genes (except for yellow, in which the
pure form is not viable), or a mouse may be heterozygous
in any two of the genes, but never in more than two.
The evidence that establishes this series of allelomorphs
may be briefly stated as follows :

In 1911, I pointed out that if yellow mice (producing
yellow and chocolates) are bred to agoutis (grays), and
their yellow offspring mated, they should produce not
only yellow and agoutis (as they did) but some choco-
lates (or blacks) also; but no chocolates appeared. I
stated that the results obtained were explicable if yellow
and agouti are allelomorphs.^

1 The discussion in the same paper of the presence of chocolate yellow and
black bars in the ticked hair in relation to the occurrence of chocolate, yel-
low and black color in domesticated races may only confuse the ontogenetic
production of characters with the gametic inheritance of factors. The

449

Digitized by

Google

460 THE AMEBIC AN NATURALIST [Vol. XLVl.

Sturtevant (1912) showed that the results are also
consistent with the hypothesis that there is close or com-
plete linkage (genetic coupling) between yellow and
agouti. In principle this is the same as saying that when
yellow and agouti enter from different sides (mother and
father) they separate in gametogenesis, or in other words
they '* repel '' each other and behave, as I said, like
allelomorphs.

The numerical results would be the same whether
yellow and agouti arei treated as though completely
linked or whether they are? treated as allelomorphic.
What I had vaguely seen in my 1911 paper was clearly
explained in the following year by Sturtevant 's treat-
ment of the same data, to which he added that of Little
and Miss Durham.

Sturtevant showed, from an analysis of Miss Dur-
ham's results, in which she used ordinary gray (gray
'* gray-belly '0 niice, that her results are consistent with
the hypothesis of absolute linkage, or, on my interpre-
tation, with the hypothesis of allelomorphism. Sturte-
vant's conclusions were promptly contradicted by C. C.
Little on the evidence furnished by some of his earlier
experiments, in which he obtained yellow, grays and
black (or chocolates) in offspring from yellow to black
(or chocolates). Such a result would be inconsistent
with Sturtevant's hypothesis. Little also appealed to
certain experiments of Miss Durham, in which, he stated,
results like his own are given. Since Little has been
unable to get again his former results, but has obtained
evidence in favor of Sturtevant 's view, and since it is
clear that he misunderstood Miss Durham's evidence,
his contradiction ceases to have any weight.

factorial hypothesis relates to those differentials that serve to separate
different types in inheritance and is not concerned with the problem as to
how those differentials produce their effects. Breeding experiments show
that gray differs from black by one differential, from yellow by another, and
from cinnamon by a third. So far as Mendelian segregation of these dif-
ferential genes is concerned it is of no consequence that the gray hair is
made up of a black, a yellow, and a chocolate band.

Digitized by

Google

No. 572]

MULTIPLE ALLELOMORPHS

451

After the publication of my own and of Sturtevant^B
paper I set to work to obtain crucial evidence in favor
of, or opposed to, the view that yellow and gray are
allelomorphic. Little, also, it appears, has carried out
some new experiments which he has recently published,
with the results just stated. My own data have been
ready for some time> but I have withheld them in order to
get a sufficient body of evidence to make the case con-
vincing, especially in the light of the possibility that the
crossing over might occur in one sex and not in the other.
For, if no crossing over occurred in the male, there
might be crossing over in the other sex, which would not
reveal itself unless the experiments were deliberately
planned so that both sexes are tested. This consideration
seems to have been overlooked by Little, for he has
omitted in his confirmatory paper to give the sexes of the
animals used. Without a knowledge of this relation even
his confirmation fails to confirm (as he supposes) the
view that he formerly combated.

Since Miss Durham worked with common gray and I
with gray white-belly, and both are ** repelled ^^ by yellow,
i. e., both are allelomorphs of yellow, it follows that these
two grays are also allelomorphic to each other.

The evidence that black belongs to the same series of
allelomorphs is obtained in the following way : If a given
yellow is mated to black, and yellow and gray offspring
are obtained, and if then the yellow offspring are mated
to black again and now give yellow and black only, the
proof is furnished; for in the first mating yellow and
agouti have repelled each other, and the yellow-bearing
gametes have united with the black gametes of the other
sex to give the yellow offspring. The second mating
shows that black is now repelled in turn by yellow and is
tiieref ore allelomorphic.

This may be illustrated in the following way: Let
B^ = yellow, b = black and B = gray. These three
factors may be treated as allelomorphs, then:

Digitized by

Google

452 THE AMERICAN NATURALIST [Vol. XL VIII

Yellow B^B by black bh,

Gametes of Pj yellow B^-B.
Gametes of Pi black b-b.
p B ''6= yellow.
* Bb =gray.

Gametes of F^ yellow B''-6.
Gametes of pure black b-b.

^ 5*'6 = yellow.

^'bb = black.

But if yellow and black and gray are not allelomorphic
the same matings should give the following results :

¥' = yellow, y' = not yellow, b = black. B = * * gi av ' ' ( not black ) .
Yellow YYBB by black l/ybb^^
Gametes ofP, yellow Y'B-i/B,
Gametes of Pi pure black y'b-y'b.
p Y'B 3/'6=7SIow.

^ y'B y'b=zgT2Ly.

Gametes of Fi yellow Y'b-Y'B-y'b-y'B.
Gametes of pure black y'b-y'b.
Y'¥ ~y'b = yellow^
„ Y'B y'b= yellow.
^» y'6 i/'6 = black.
y'B i/'6=:gray.

On the second assumption yellow, gray and black
should appear in the back cross. The former and not the
latter view is therefore consistent with the actual results.

The Symbols Employed
It is, of course, a matter of secondary importance what
system of symbols is followed. The requirements are
simplicity, consistency and suggestiveness, but one can
not always arrange to have all three at the same time.
The simplest scheme, for a system of allelomorphs like
these, would be to have some common letter to indicate
their relation and an exponent to suggest the different
characters for which each stands. If we take the symbol
b (black) for the common letter, and use capitals for
dominance, the allelomorphs will be:

b = black.
B^= gray gray belly.
5"'=: gray white belly.
B^ = yellow.

If one preferred to take Y (yellow) as the common letter
the series would be y'^, y'^, y'^, Z'; or, if one preferred

Digitized by

Google

No. 572] MULTIPLE ALLELOMORPHS 453

to take G (gray), as the common letter, the series would
be g^, g^y G, g^\ On the whole the first series seems to
me somewhat preferable.

The factor for cinnamon is entirely independent in
heredity of the preceding series of allelomorphs. This
factor may be represented by ci and its normal allelo-
morph by Ci. The formula for the wild gray would then
be Ci Ci, and that for cinnamon would be ci ci. Black
would be h b, and the double recessive cinnamon black
(or chocolate) would be bb ci ci. Chocolate is one of the
commonest types of domesticated mice and since I have
used it very extensively in my matings, its relation to the
other types may be further stated. It is known that if
chocolate is bred to wild gray, and if the gray offspring
that are obtained are then inbred, they give, in Fo, the
following classes: 9 wild gray, 3 cinnamon, 3 black, 1
chocolate.

It is clear that chocolate is the double recessive type.
Of the two genes, that differentiate chocolate from wild
gray, chocolate has one in common with cinnamon and
the other with black. In other words, chocolate is cin-
namon black, and technically should receive this name.

The Experimental Evtoence
Is There a Separate Factor for White-belly?
The first series of experiments was made in order to
determine whether the peculiarity of white-belly, shown
by the wild race of white-bellied grays, is due to a factor
that may be separated from the gray white-bellied mice,
or whether it is completely linked to gray (or allelo-
morphic to it). As wild gray house mice offer some
drawbacks in breeding work, I used cinnamon blacks
(chocolates). Gray white-bellied mice were bred to

2 It is not possible to make a system of allelomorphs (in which the
"compounds'^ are serially epistate to each other) consistent entirely with
the system of nomenclature that I have suggested for the usual cases in
which mutant allelomorphs are contrasted with the normal allelomorphs of
the wild (or supposed original) type.

Digitized by

Google

464

THE AMERICAN NATURALIST [Vol. XLVHI

chocolates.^ The gray white-bellied offspring were se-
lected and these were bred again to chocolate. The cross,
in regard to sex, was made both ways. If there is an
independent factor for white-belly that can separate from
the factor for gray gray-belly, then some gray gray-
bellied mice should appear. None were obtained, as the
following table shows. We must conclude either that
there is one factor that gives the gray white-bellied coat,
or else that the postulated factor for white-belly is so
closely linked to the gray factor that it has not sepa-
rated once in 100 times. Therefore unless such a sepa-
ration occurs it is simpler to assume one factor for gray
white-belly that is allelomorphic to black and to gray
gray-belly, etc.

TABLE I

Mating

Gray or Cinoa-
mon White-
belly

Black Chocolate White

9 1 cf

9 d^

Gwb 9 by Ch cf . . . .

Ch 9 by Gwb cf

Totals

7
2
9

21
14
35

2
3
5

9
10
19

4
9

11 3
10 1
21 3

1
1

Taking both crosses together, there are 44 grays to 54
blacks and chocolates, which approximate at least to
expectation. To these numbers I may add the follow-
ing data taken from similar experiments made for other
purposes in which one parent was, as before, gray white-
belly.

Gray-white Belly. Black or Chocolate,

c? 9 c? S

17 25 20 20

Presumably, therefore, the results may be treated as
though a single gene for gray white-belly exists. It will
be observed that the experiment has been made in two
ways, for at the time I was aware of the possibility that
crossing over, if it occured, might be limited to one sex.

• At the time when the experiment was made all the gray white-bellied
mice were heterozygous for black and for agouti (including some with the
factor for cinnamon).

Digitized by

Google

No. 572] MULTIPLE ALLELOMORPHS 455

We are justified, therefore, in treating gray white-belly
as an allelomorph of gray gray-belly, the former domi-
nating. If crossing over should occur, it might perhaps
only be realized in the gray or cinnamon mice, since it is
possible that the ticked condition of the hair (that is,
common to gray and to cinnamon) is necessary to realize
this condition. The expected crossover that would be
observed would be gray gray-belly mice. The contrary
class would then be black or chocolate mice that carry
the factor for white-belly that might or might not show
the influence of the supposedly separable factor.

My white-bellied stock of mice had been killed after
my earlier results had been published, but Mr. B. B.
Horton had kept some of my original stock alive, and
from him I obtained a few of these mice in 1912 to carry
on the above experiments.

An extraordinary sex ratio appears in the next to the
last table, where there were 26 males to 76 females, ap-
proximately 1:3. The mice were entered when about
three weeks old. The sex was noted, but no special atten-
tion given to the determination. There is some chance
of mistaking the sex of young mice, but one familiar
with these animals can determine with certainty the
sex at three weeks if sufl5cient care is taken. I have no
reason to suppose that I made such errors which would
have to be frequent to give these results. If taken, then,
at their face value, the data seem to show that there is a
sex-linked lethal gene present here. It is not linked to
any of the factors involved, and this is not expected,
since neither black nor agouti is sex-linked. If further
work confirms this conclusion (and I hold it as a provi-
sional conclusion until it can be further studied) we have
here the first evidence of a sex-linked gene in mice. A
sex-linked lethal should give a sex ratio of 1(?:2$.

The Allelomorphism of Yellow, Gray and Black
The allelomorphism or '* repulsion '* of yellow and
agouti (gray) may be tested in various ways. One of the

Digitized by

Google

456

THE AMEBIC AN NATURALIST [Vol. XLVIII

simplest tests is the following: Yellows were bred to
chocolates. The combination gave yellow and agouti off-
spring, when certain yellows are used, and yellow and
chocolate offspring when other yellows are used. Mixed
litters of yellow, agouti and chocolate do not appear.
Now when yellow and agouti appear in a given litter (as
above) the yellow parent must have carried agouti. If
her yellow gene ** repels'' the agouti gene, then none of
the yellow daughters should contain agouti genes, con-
sequently if such yellows are next bred to chocolate the
offspring should be only yellow and chocolate (or black)
and never yellow and agouti. This, in fact, is what my
experiments have shown. In the two following tables the
results of crossing yellows by chocolates are given by
litters. The yellows that were used at first were for the
most part heterozygous for gray white-belly, hence in
the. earlier litters yellows and grays were generally ob-
tained. The yellow offspring of these earlier litters were
for the most part used in the later experiments, hence
the later litters are made up of yellows and chocolates.
The records (not given here) showed in every case that
yellow mice from litters of yellow and gray gave, when
bred to chocolate, only yellows and chocolates.

TABLE II

Yellow ^ by Chocolate $

Litters

Yellow

2 .

. 1

3 13 24

4 Q^^^^'s

5 4!3 5 7!2l2'2 2

34|. . .|. .1. .

34 3 2412 2

.i. .j. 4 . .1.

Gray

422 .
! ! ! 2

4 4 55

4'.
.1.

Chocolate

White

. .2343.33

. .I.l.i.l. . .1.

. 2 3.2 . 34 2

TABLE III

Yellow $ by Chocolate c?

Litters

Yellow.... 414

Gray 5 ..2

Chocolate

4|3

5!2

4
1

4'2'4
12..
. . . . 1

1 4 88
3 5 13

2I2 8
.. .. 3
22 ..

3
1

1
3

1
4

4
2

3
3

4
1

Black 1. . . .

White 1 . . . .

1 2!..

* Probably two litters combined.

Digitized by

Google

No. 572] MULTIPLE ALLELOMORPHS 457

TABLE IV

Summary op Litters

Yellow and Gray Yellow and Chocolate

Yel. Gray Yel. Choc.

101 78 70 67

The experiment is not demonstrative, however, unless
both the yellow daughters and sons are bred to chocolate,
for it might be that yellow and agouti are linked and
crossing over might occur in one sex and not in the other
sex. For instance, if we start again with yellow by choc-
olate, then if their yellow offspring contain agouti linked
to yellow that does not cross over in one sex, let us say in
the males, it follows that a yellow male bred to chocolate
would give only yellows and chocolates, for the agouti
gene would go with the yellow. Therefore, both sexes
must be tested. This essential element in the proof has
been overlooked by Little, for he fails to state whether his
test experiments were made with both sexes. In my
main experiments I have used yellow sons only, and the
tables are based on those data, but in a few cases I have
mated the yellow daughters (whose brothers were agouti)
also to chocolate and have found that these females give
only yellows and chocolates, which shows for both sexes
that no crossing over of yellow and agouti occurs.

A specific case will illustrate this point. A yellow
male was bred to a chocolate female and gave 5 yellow
and 7 gray offspring in two litters. One of the yellow
daughters was bred to chocolate and in four litters pro-
duced 11 yellows and 9 chocolates. A yellow grand-
daughter gave 9 yellows, 7 chocolates and 4 whites.

A yellow female bred to chocolate gave 8 yellows and
16 chocolates, but as I have no record of the preceding
generation, I can not be sure that this result is compar-
able to the last. It shows at least that a yellow female
gave only two kinds of offspring.

A **New Gray" Factor
A word may be added about the *'new gray." In the
original stock obtained from Mr. Horton there was a

Digitized by

Google

458 THE AMERICAN NATURALIST [Vol. XLVHI

gray female with a not-pure-white belly. She was not
used in the main lines of the experiments described above.
But she was kept in stock and bred with chocolates.
About a year later I noticed in the offspring of a pair of
cinnamon white-bellied mice a few mice that looked like
chocolates, but which showed, on closer inspection, dis-
tinctly ticked hair. One of these new grays bred to
black (heterozygous) gave some chocolates, blacks, new
grays, and one very dark, almost black, mouse with
ticked hair.** The female was bred next time to a house
mouse (gray gray-belly) and produced all gray gray-
bellied oflfspring that had a dark coat, but not nearly so
dark as that present when the new gray is heterozygous
for black. Until further tests have been made it can not
be stated whether or not the new factor belongs to the
yellow-black system of quadruple allelomorphs.

8 The resemblance of this mouse to the rabbit "agouti -black" homozygouB
for black is very striking (Punnett, Jour, of Genetics, II, 1912).

Digitized by

Google

THEBTEEN YEAES OF WHEAT SELECTION

T. B. HUTCHESON
Associate Agronomist, University of Minnesota

Intboduction

In 1901 the Minnesota Agricultural Experiment Sta-
tion planted a number of varieties of wheat from the
polonicv/m, spelta, turgidum, d/urum and wlgare types in
foundation beds in order to have specimens of these diflfer-
ent types always on hand for class work, hybridiza-
tion or demonstration purposes. Six of these varieties —
hedgrow {turgidum), Eussian {vulgare), common speltz
(spelta), kamouka (durum), and Polish (1) and Polish
(2) (varieties of polonicum) — have been grown continu-
ously since that time and an effort has been made to
improve them by selection. The method followed was
that introduced at this station by Professor W. M. Hays
and called the ''centgener^' method.

The centgener method consists, briefly, in starting with
individual plants, planting one hundred selected kernels
from each plant at equal depths and at equal distances
apart in separate plots. A plot of one hundred plants is
called a centgener. Careful notes are taken on the plants
in each centgener and at harvest time five or more of the
highest yielding plants are selected from which the seeds
for planting the next year are taken. From these five
best plants from five to ten of the best heads are selected
and thrashed together. One hundred of the largest and
plumpest kernels are then selected out of the seed ob-
tained by thrashing these selected heads, and these are
planted in the centgener test the next year. This work is
continued from year to year, each season the hundred
best kernels from the five or more best plants being
planted in succeeding centgeners.

469

Digitized by

Google

460 THE AMERICAN NATURALIST [VOL.XLVIII

In 1908 an experiment was planned with the object of
developing a strain of wheat which would have a minimum
amount of culm exposed between the base of the spike and
the upper leaf sheath, or in other words, to produce a
short-necked variety of wheat. The ultimate purpose of
reducing the neck lengths was to reduce the area of the
stem exposed to the black stem rust. Since this rust
ordinarily does little damage to that portion of the culm
enclosed in the leaf sheath, it was thought that a short-
necked wheat would be more likely to escape serious
damage from stem rust than a long-necked kind. For
this work individual plants were selected which had short
necks and the seed from these were planted in separate
centgeners. Each year at harvest time ten or more plants
which appeared to the observer to have the shortest necks
were selected from each centgener and measurements of
their neck lengths were made and recorded. One hundred
kernels were saved from these shortest necked plants
each season for subsequent centgeners, thus making a
continuous selection for short neck lengths.

The data derived from the above experiments seems to
throw some light upon the much-discussed question as to
whether or not selection within a pure line can increase
yield or change type enough to make it a desirable prac-
tise from the practical breeder's standpoint. In both of
the experiments, we have the requirements for a pure
line satisfied. Wheat is a normally self -fertilized plant.
Each centgener was started from a single head in 1901
and these heads have bred true to type ever since.

The long period of years over which this experiment
has extended makes the data particularly valuable. One
of the adverse criticisms to most pure line work is that
it has not extended over a long enough period of time.
Thirteen years are about as long as any practical breeder
would be apt to keep up selection on one pure line and
covers the longest period of continuous selection for a
self-fertilized plant yet reported.

Another criticism to pure line investigations is that in

Digitized by

Google

No. 572] THIRTEEN YEARS OF WHEAT SELECTION 461

many cases it has not appeared certain that the material
studied was a pure line. Since the plants have bred true
to type throughout the whole period of study, it is obvi-
ous that this criticism will not hold for the data herein
presented.

The work haS been conducted at this station under the
direction of Professor W. M. Hays from 1901 to 1905,
under Professor E. C. Parker 1905 to 1908, under Pro-
fessor Andrew Boss from 1908 to 1911 and under Pro-
fessor C. P. Bull 1911 to 1913.

Plate I. Average yield per plant for all varieties. X-X, fitted straight

line.

Selection to Increase Yield
The varieties studied, the average annual yield of each
variety and the average yield per plant for the six vari-
eties under test are shown in Table I. In the years 1903
and 1904 weather conditions were unfavorable, making
it impracticable to obtain correct average yields per
plant, so data for these years were omitted. However,
selections of the best plants were made in these two
seasons as in the others and the best seed from them
were kept for planting, so the continuous selection for
increased yield was uninterrupted.

Digitized by

Google

462

THE AMERICAN NATURALIST [Vol. XLYIH

TABLE I
Showing Yield Pee Plant — ^Yeaes 1901-1913

Name of Variety

Yield per Plant in Grs.

1901 1902 1905 ' 1906 1907 1908 1909 1910 | 1911 1912 1918

Hedgrow. .
Russian. . .

Speltz

Kamouka. .
Polish (1) . .
Polish (2)..

3.10
1.00
2.40
1.60
.80
1.10

2.80
1.70
1.80

3.69 2.48
3.67, 1.96
3.99 2.99

2.60 1.99 2.(

1.30
.95

2.52| 2.04
2.83' 1.97

1.27
1.71
1.38
1.39
1.03
1.26

3.75
2.74
3.38
3.31
1.48

2.49 2.65 2.02 .99,3.67

1.95 1.37 2.70

2.01 2.14 2.59

1.67 1.36! 2.16

1.91! 1.70 1.66 1.12 1.74

1 1.61 1.31 1.78! .51 1.33

2.71 2.17
2.40 2.86
2.19! 2.48

Average

1.65

1.84 3.10! 2.35* 1.34 2.93| 2.22 2.18 1.83 1.24 2.36

Selection to Incbease Height
The average height of the plants for each year of the
test is shown in Table II. Though no attempt was made
to select for increased height, since a number of workers
have shown that height in the small grains is distinctly
correlated with yield, it is natural to suppose that the
selected plants were among the tallest as well as "being the
highest yielders of each year's crop. When this experi-
ment was begun, it was not known that height and yield

table n

Showing Average Height Per Plant — 1901-1913

Height in Inches

Name of Virietj

1901

1902

1905

1906

1908

1909

1910

l$»ll

1912

1913

Hedgrow

36
34
34
36
40
28

41
37
38
34
38
30

42
40
37
34
41
37

43
36
47
38
38
37

46
44
44
40
42

41
41
41
40
42
35

38
33
39
32
33
31

41
36
42
38
39
38

36
35
39
36
38
31

Russian

Speltz

Kamouka

Polish (1)

Polish (2)

33
33
32

Average | 35

were correlated, so the figures on height were kept merely
as a matter of general interest and with no idea that they
would have bearing on the problem. Among those who
later found height correlated with yield are Deneumostier
(10),^ Love (ai),2 Myers (12),8 Leighty (^2)* and

1 Deneumostier, C,
1910.

'Correlations in Wheat," Ann, Oemblouan, 20, No. 5,

Digitized by

Google

No. 572] THIRTEEN YEARS OF WHEAT SELECTION 463

Selection to Decbbasb Nbok-Lengths
The result of the selection for short neck-lengths is
shown in Table III. This is a clear illustration of how
misleading short-term experiments may be. Had the
experiment been discontinued at the end of the third year,
the figures would have indicated that it was possible to
modify this character very rapidly by selection. How-
ever, in the following two years the neck-lengths seemed
to revert to the mean of the pure lines, and the last year
they were actually longer than when the experiment was
started. The reduction in the first three years was prob-
ably due to growing conditions.

TABLE ni
Showing Besult of Selection tob Shobt Necks

Averago Neck L ngtb

in Curre

1909

1910

1911

1912

1918

Series A

7.4
6.1
6.8
6.2

1.86
1.12
1.66
2.08

.24
.79
.66
.69

7.34
8.13
7.63

10.47

9.64

Series B

11.6

Series C

8.21

Series D

13.82

Discussion
From the data presented in these tables, it is evident
that there has been no permanent gain for these thirteen
years of selection either in yield per plant, height of
plant, or shortening of neck-lengths. The expected sea-
sonal variations occur. A comparison of the yield of
Haynes Blue Stem, which is grown extensively in Minne-
sota, and was continued in the variety test without any
attempt at selection throughout the whole period, with
Hutcheson (13).«*

2 Love, H. H., *'A Study of the Large and Small Grain Question," An,
Bep, Am, Br, Asso., 7: 109-118, 1911.

» Myers, C. H., * * Variation, Correlation and Inheritance of Characters of
Wheat and Peas," Cornell University Thesis, 1912.

4 Hutcheson, T. B., "Correlated Characters in Avena sativa, with Special
Reference to Size of Seed Planted," Cornell University Thesis, 1913.

«Leighty, C. E., "Studies in Variation and Correlation of Oats, Avena
saiiva," Cornell University Thesis, 1912.

Digitized by

Google

464

THE AMEBIC AN NATURALIST [Vol. XLVIII

the average yield of the selected varieties, is shown in
Table IV. The average yield in bushels of the Haynes
Blue Stem is also platted in comparison with the average
yield of the selected varieties in Plate II. In 1912 a
severe hail storm injured the variety plats so much that

Plate II. Comparing seaj^onal fluctuations in selected lines with unse-
lected Blue Stem. Solid line, yield per plant in grams for selected lines;
dashed line, yield in bushels per acre for Blue Stem.

it was thought best not to include the yield of the Haynes
Blue Stem for that year. This gives an incorrect appear-
ance to the curve, as it was extended just as if this year
was present and midway between 3911 and 1913 in yield.
It will be noticed from Table IV and Plate II that the

TABLE IV

Comparing Seasonal Fluctuations in Selected Lines with Unsblbcted

Blue Stem

1901

1902

1905

1906

1907

Yield in grs. per

plant for selected I

lines I 1.65 1.84 3.10 2.35| 1.34 2.93

Yield in bu. per

1908

acre for bluei
stem 22.9

1909 I 1910 1911

2.22

23.9 30.4 24.00 21.00 26.00 26.6

1912 , 1918

2.18 1.83 1.24

I I

2.36

24.6 24.2 1 123.2

Digitized by

Google

No. 572] THIRTEEN TEARS OF WHEAT SELECTION 465

fluctuations from year to year agree very closely. These
data indicate that increased yield is due to favorable
environmental factors and not to improvement by
selection.

A comparison of the yield of each variety for the first
five years of the test with that of the last five years is
shown in Table V. The data in this table show that
there is no significant difference in yield for these two
periods; In Eussian and Polish (1) there is a slight
increase in favor of the latter period, but in the other
four varieties there is just as much decrease for this
period. However, there is not enough difference in any
case to indicate either permanent improvement or de-
crease in yield. As far as these varieties are concerned,
it seems that selection has brought about no permanent
improvement.

TABLE V

GOMPAUNa THE YiELD OF THE FlEST-YEAB PERIOD WITH THAT OF THE LAST

FivE-YEAB Period

IstS-year Period

Last 6-year Period

Name of Variety

Height

Yield

Height

Yield

Hedgrow

41.6
38.0
40.0
36.4

2.67
1.90
2.61
2.01
1.64
1.62

38.4
36.4
39.2
36.8
37.4
33.4

2.34

Kiimriftn

2.18

Spelta

2.40

Kamouka

1.97

Polish (1)

39.8
33.4

1.61

Polish (2)

1.31

Average ..1 38.2

2.06 ! 36.5

1.97

A curve of the yields of the six varieties under con-
sideration for the thirteen years of the test was plotted
and a straight line was fitted to it, by the method of the
least squares, to indicate the trend of the yield. This
curve is shown in Plate I. There is a slight downward
tendency in this straight line, but it is not enough to indi-
cate a tendency toward decrease in yield. The line fitted
to the curve of height (Plate III) also shows a slight
tendency downward.

The data herein cited are not sufficient for definite con-

Digitized by

Google

466

THE AMERICAN NATURALIST [Vol. XLVIII

elusions. However, the indications are that from a prac-
tical breeder's standpoint permanent improvement in
pure lines in small grains, if possible, is certainly not
rapid or apt to be very marked. Thirteen years of selec-

A
V

9«

_J^

\ 7

7 — —

J /l >v M

"" \~"--;

15 ^

\ I

1ZE2

"5 "3

«i r

~* "•!

"f "IB

Plate III. Average height of all varieties. X-X, fitted straight line.

tion covers considerable time and expense, and, as far as
can be seen from the varieties reported in this paper, it
has resulted in no permanent improvement. This would
suggest that some other line of improvement must be
sought. It is probable that much more rapid progress
could be made by segregating pure lines from mixed
populations and combining the desirable characters of
these lines by hybridization.

Digitized by

Google

PATTERN DEVELOPMENT IN MAMMALS
AND BIRDS

GLOVER M. ALLEN
Boston Society op Natural History

Partial Albinism in Wild Mammals

Partially albinistic individuals of species that normally
are wholly pigmented, occur frequently in a wild state,
and almost any large series of a given species may con-
tain a few. I have examined many such, in which it was
perfectly evident that the white mark was due to areal
restriction of some one or more of the primary pigment
areas just as described in the various domestic species.
It is apparent that the white markings in both are quite
comparable, but in species under domestication no agency
seems present whereby such pied individuals are elimi-
nated, whereas in a wild state the sudden acquisition of a
large amount of white in an individual would not only
render him too different from his fellows, but might put
him at a disadvantage because of a conspicuousness to
which as a species he had not yet become accustomed.

There are many other species in which, as we now see
them, white markings form a permanent and normal part
of the pattern. Among those in which these white mark-
ings are few or simple, it is often evident that they are
merely prjjpary breaks between the pigment patches that
have become more or less fixed by long periods of selec-
tion, whether natural, sexual or otherwise. As I shall
endeavor, tg^^how, there are species in which a beginning
has alreaa^ been made towards the development of a
pied pattern, though it has not yet become well fixed*
Still other species show a more complicated white and
pigmented pattern, the white portions of which can not
readily be derived from primary breaks alone. Such I
take to be highly developed patterns and make no attempt

467

Digitized by

Google

468 THE AMEBIC AN NATURALIST [Vol. XLVIII

to analyze them here. Examples of this type are seen in
the zebra, the spotted skunks {8pilogale)j the striped
weasel {Ictonyx), Probably more than one factor is
responsible for some of the combinations of stripes and
spots seen, for example, in certain spermophiles (Citellus
13-Uneatus) J but I shall not now attempt a discussion of
these.

One of the most frequent manifestations of pigment
reduction in mammals is the presence of a white spot in
the normally pigmented forehead. This is due primarily
to the reduction of the ear patches, which fail to meet at
their median edges. Perhaps, too, the apparent loss of
the crown patch in some mammals still further tends to
lessen the amount of pigment production at this point.
Babbits and hares very often have more or less white in
the forehead, but none of the species has developed this
sufficiently to make it a permanent mark. Moseley in his
^'Naturalist on the Challenger,'^ speaks of a ''black
variety'' of wild rabbit — doubtless introduced — "with a
white spot on the forehead" as occasionally found on
Teneriffe, Canary Islands, but this mark is common,
and I have seen it in such widely sundered species as the
eastern varying hare of New Hampshire and the black-
necked hare native to Java. A specimen of Leisler's bat
{Nyctalus leisleri) in the Museum of Comparative Zool-
ogy has a white spot in the middle of the forehead and
another on the mid- ventral line of the abdomen — ^the first
a primary break between the ear centers, the second
probably a ventral primary break between tjig^e of the
sides. Among the Insectivora, the West ]pdiah Solen-
odon paradoxus has a white patch at th6 nape of the
neck which has become a permanent part of *^s pattern.
It is clearly the enlargement of a primary D'reak sepa-
rating the ear patches and neck patches on the median
dorsal line. It is a fact of much interest that in a con-
siderable series of this species in the collection of the
Museum of Comparative Zoology hardly two have it
developed alike, but it varies from a few white hairs to

Digitized by

Google

■?^.i

No. 572] PATTERN DEVELOPMENT 469

a large patch 15 X 10 nun. wide. Evidently it has not
yet become precisely defined in its limits, though now a
permanent mark of the species.

White marks in the forehead are common among the

species of the Mustelidae or weasel family. A narrow

^hite median line is present in the Javan mydans and in

*he skunks {Mephitis) as part of the permanent pattern.

In the badger (Taxidea) a white line is not only pres-

^^t on the forehead, but it is often extended medially so

1^® to separate the pigment patches of both sides of the

^ ^y. In the New York weasel (Mustela noveboracensis)

^^ the eastern United States a few white hairs are often

/>7tesent on the forehead, and other instances could be

xnixlti plied. Among monkeys, a white spot on the nose is

pr&s ^nt in some species of Lasiopyga, and in an allied

^en.xi.s Rhinostigma, it is elongated vertically to form a

wrZiit^ streak.

-A^ 3ret more illuminating case is that of the Muskeget

^e^c^Ila mouse {Microtus breweri) a derivative of the

^^^t^^oM^ignmOiOn brown meadow mouse of the New England

^a^^ctHand. On this island of white sand oflf the Massa-

jix^^ ^^tts coast, a pale variety has developed which is very

.^^^izMzuct from that of the neighboring shores. Not only

^i* .^a paler race, but albinism also has begun to appear,

^ "fcii^^ ,«it occasional individuals have a white fleck between

te ^^ ,^rs, showing the drawing apart of the ear patches.

Of ^^i^ series of 62 specimens in the collections of the

M^-^^-^^^xma of Comparative Zoology and the Boston Society

0^ -^^"^ .^^tural History, no less than 13 had such white flecks,

^^^^ ^Z3ne had in addition a white spot just in advance of

i^^ ^shoulders, marking the line of separation between

^^c^'lsi and shoulder patches. In our studies on the hered-

\^^^ ^i>f coat colors in mice, Professor Castle and I dis-

c^^r^ored (Allen, 1904; see also Little, 1914) that the pied

wyxi^ition is recessive in the Mendelian sense towards

^OcLfe self colored, so that partial albinos bred to wholly

^Ygicaented mice produce in the second generation, if

mterbred, 25 per cent, of spotted young. The figures

Digitized by

Google

470 THE AMEBIC AN NATURALIST [Vol. XLVIH

given above (13 in 62) are near this in case of the Muske-
get mouse, but the matings are of course more promiscu-
ous. The case is interesting in connection with the
studies of Ramaley (1912) and Pearl (1914), tending to
show that in a mixed population the recessives may in-
crease so as to exceed the dominants. Although the
spotted mice do not, in case of this species, exceed the
unspotted individuals, they nevertheless are of far more
frequent occurrence than they are in the mainland repre-
sentatives of the species. This accords with the fact that
island-living mammals are very conunonly albinistic, and
the cause is doubtless that the population is much more
inbred, so that the recessive condition of partial albinism
is more likely to be propagated than if successive genera-
tions have a wider range over which to spread. It seems
probable that heredity will tend to increase the propor-
tion of spotted mice of Muskeget, and that if this condi-
tion is disadvantageous, a large part of the spotted indi-
viduals will be killed off, yet in the course of time they
may become adjusted to this condition and will survive
in increasing proportion till the white mark becomes
characteristic of all the animals. Cory (1912) records
the capture of seven muskrats at Hayfleld, Iowa, all of
which were uniformly marked, having a white ring around
the neck and the entire underparts, feet, and end of tail
white. I can think of three causes influencing the status
of such white markings. These markings may be in-
herited in a purely automatic way as unit characters;
but if thus inherited they may be (1) increased through
selection, natural or sexual; or (2) eliminated by the same
agent; or (3) they may be, at first, of no influence at all
in the economy of the animal and persist or not, accord-
ing as they are heritable.

I have mentioned that island mammals tend to be more
albinistic than their mainland representatives. Other
cases may be mentioned, as the common squirrel (Sciurus
vulgaris leucurus) of Great Britain, which differs nota-
bly from that of the continent in having frequently a

Digitized by

Google

No. 572] PATTEBN DEVELOPMENT 471

white or whitish tip to the tail, often for one half its
length. A similar white tip is occasionally seen in our red
squirrel {8. hudsonius) as an albinistic mark, and is due,
of course, to the terminal restriction of the rump patches.
The deer of Whitby Island, Puget Sound, are said to be
much marked with white, and sundry marsupials of
Papua as well as the monotreme Zaglossus are subject to
white markings. In the cuscus {Pseudochirus) the pig-
ment is sometimes restricted to small patches and round
spots scattered on the back, those in the region of the
shoulder of a different color from those of the side and
rump patches. Another instance is that of the white-
footed mouse of Monomoy Island, Massachusetts, the
mid- ventral parts of which are pure white to the roots of
the hairs, an albinistic condition to be clearly distin-
guished from that in which the belly appears white, but
only because of the white tips to the hairs whose bases
are dark-pigmented.

The restriction of the rump patches so as to produce
a white tail-tip is common among mammals. It is found
in occasional specimens of many species as the shrew mole
{Blarina)y Brewer's mole (Parascalops), the meadow
jumping mouse {Zapus)^ the white-footed mouse {Pero-
myscus)y and squirrels (Sciurus). In some it has be-
come developed as a permanent and characteristic mark,
as in the woodland jumping mouse {Napceozapus)j the
red fox {Vulpes)^ such genera as Hydromys, Tylomys,
the Virginia opossum (Didelphys virginiana), the tree
kangaroos {Dendrolagus). In many others a pure white
beUy is developed through ventral restriction of the
shoulder and side patches.

Among ungulates the break between the ear patches
has been developed to form a broad white blaze from
forehead to nose in case of the blesbok (Damaliscus
albifrons) of South Africa and in related species in
East Africa. The chevron-mark on the forehead of cer-
tain antelopes is possibly a specialized development of
the same thing.

Digitized by

Google

472 TRE AMERICAN NATURALIST [Vol. XLVIU

White buttock patches are present in several unrelated
ungulates — as the pronghorn {Antilocapra), the wapiti
{Cervus canadensis) ^ and the Rocky Mountain sheep
(Ovi'S canadensis). Probably these are the result of
restriction or total inactivity of the pigment patches
covering the rump.

Fio. 42a. Diagram Showing the Pigmented Patches op a Partiallt ALBi!fO

Dbbb.

Among the deer family white is generally confined to
the under surfaces and the primary white breaks have not
been developed to form patterns. Albinistic deer are
fairly common, however, and in Fig. 42a I have made a
tracing from a photograph showing the side of a par-
tially albino doe in which areal restriction of pigment has
taken place in such wise that the primary patches are
all indicated, and separated from those of the opposite
half of the body by a median dorsal white line. The ear
and the neck patches are joined, but a few small islands
of pigment are left here and there, much as in cows.

In the young of many deer and in the adult of such
species as the axis deer, a spotted pattern is developed.

Digitized by

Google

No. 572] PATTERN DEVELOPMENT 473

There is an obvious tendency for the spots to become
arranged in longitudinal rows, and intermediate stages
may be found in which they coalesce to form broken lines.
There is little doubt that the complete white stripes
occurring in part of this pattern were formed originally
through the coalescence of rows of white spots. In the
tapir a somewhat similar spotted pattern is found in
the young, while the adult Malayan tapir has lost the
shoulder and side patches, producing thus a white-bodied
animal, pigmented to the back of the foreleg and on the
buttocks and hind legs. Among the ground squirrels
(Citellus) a beautiful series can be picked out showing
the transition from a uniform grizzled mixture of ticked
hairs to indistinct spotting, then rows of white spots, and
finally broken and complete longitudinal stripes. The
production of these stripes I believe to be due, not to the
development of breaks between the primary pigment
patches, but to the action of a factor which is the negative
of the so-called ''English'' marking in rabbits, so that
instead of the development of scattered small pigments
spots there are formed, instead, spots without pigment.
That it is possible to evolve a striped pattern from spots
through selection, I have no doubt, and indeed, it is gen-
erally believed. On the other hand, it is quite possible
that the converse may happen, and spots result through
the breaking up of stripes. According to the experiments
of Professor Castle and Dr. MacCurdy, however, it seems
to be a diflScult matter to fix a given marking by rigid
selection, yet it must be admitted that a few years' work
even of careful breeding is nothing in comparison with
the age-long selection that may have been at work on the
species. That it is a difficult matter to produce a given
pattern is further evidenced by the fact that in many
species in which white markings regularly occur as part
of the pattern, these are subject to great individual
variation in their extent, showing that they are even yet
not wholly definite.
It was formerly urged against evolutionary doctrine

Digitized by

Google

Digitized by

Google

No. 572]

PATTERN DEVELOPMENT

under side of the specimens. In the large coastal race of
mink found from southern Maine to the Carolinas {M, v.
lutreocephalus), the entire pelage is usually brown, ex-
cept for the chin which is white. Occasional white marks
are present in some specimens along the mid-ventral
line of the throat and chest, and between the hind legs.
In the smaller typical M. vison of northern New England
northward the white marking is apt to be more extensive,
and in no two individuals exactly alike. The diagrams
show the ventral markings of a few specimens from New
England and Nova Scotia. In Fig. 43 the amount of
white is very small. The chin spot, which represents the
beginning of a break between the two ear patches at their
antero-ventral extremity, is always present and has be-
come now a fixed mark of the species, though variable in
extent. A slight break in the center of the chest shows
where the two shoulder patches have failed to meet, and
a white spot at the anal region indicates a like restriction
of the rump patches. Similar spots appear mid-ventrally
in Fig. 44, with the addition of a few white hairs, medially
at the upper throat, where the ear and neck patches join,
and a few more on the lower throat at the line of union
of the neck patches of opposite sides. In Figs. 45 and 46
no break is present on the abdomen, but in the former
figure, a large transverse break has appeared on the
upper throat where the ear patches fail to unite with the
neck patches and with each other, and a median line runs
forward to join the white of the chin, showing the greater
restriction of the ear patches ventrally. An imperfect
separation of these patches along the center of the throat
has taken place in Fig. 47, and a more considerable break
occurs in the same place in Fig. 46. In the Pacific Coast
mink {Mustela vison energumenos) a well-developed
white patch on the chest is rather characteristic, some-
what larger than in Fig. 45. This is due to the ventral
restriction of the shoulder patches which fail to meet
below. In Fig. 46 this white area is seen with a tongue
extending upon the center of the lower throat, and on to

Digitized by

Google

476 THE AMEBIC AN NATURALIST [Vol. XLVm

one fore leg, as well as in the mid line of the thorax, mark-
ing nearly the anteroposterior limits of the shoulder
patch. The neck patches are not separated in this figure
but have become so in Fig. 48, so that a continuous line
of white runs from chin to chest. In Fig. 47 the shoulder
and the side patches have both failed to join venfrally,
and thus a broad white line is formed down the center of
the belly from the conjoined neck patches to the rump
patches. If all these breaks were to be present in a single
animal, there would be a narrowed white area along the
entire ventral side of the body from chin to anus, extend-
ing on to the lower side of the fore legs. Practically
this condition exists in another species of the same
genus, Streator's weasel (Mustela streatori) of the
Pacific Coast, in which the throat, chest and belly are
white but the width and boundaries of the white area are
very variable in different individuals. It is therefore in
a stage beyond that which the minks have reached, yet it
has not attained the stage in which the white area is of
definite and rather constant bounds, as in certain other
weasels, for example Mustela noveboracensis, in which
the white, of the belly extends nearly or quite to the
lateral border of the body, but in different individuals
varies slightly, and M. cicognanii, in which the white area
of the belly constantly extends to the lateral boundary of
the venter from throat to anus. This is the condition
toward which the mink is tending.

Another interesting case in which a pattern mark ap-
pears to be evolving through the fixation of a primary
break between pigment patches is that of the so-called
tayra of South America (Tayra barbara) a large Muste-
lid. The Central American race (biologice) of this animal
is wholly black, but the typical subspecies of Brazil and
northern South America is subject to a varying amount
of reduction in pigmentation. Curiously, this takes place
at the posterior end of the neck patches or at the anterior
part of the shoulder patches. Three of five specimens in
the Museum of Comparative Zoology are marked in this

Digitized by

Google

No. 572]

PATTERN DEVELOPMENT

477

way. All have a triangular patch of white at the base of
the throat ventrally, as a break between neck and shoulder
patches and a partial separation of the neck patches
from each other. Each has a dorsal mark of white ; in the
first a narrow linear break between the shoulders ; in the
second a broader transverse mark, and in the third a
square patch of white occupying nearly the width of body
between the shoulders to the base of the neck. The white
throat marking increases in e^^tent from first to third,
just as does the dorsal marking. Probably in time this
white mark, now of irregular size and appearance indi-
vidually, will become a permanent part of the pattern. In
this animal the entire head and neck are a grizzled gray
as far back as the posterior limit of the neck patches, and
the rest of the body is black. This, then, shows that the
pigment patches of head and neck are differentiated in
color as well, from the patches of the rest of the body.
The occurrence of white markings in the back is relatively
uncommon in mammals, though white on the under sur-
faces is common, and, as shown by Mr. Abbott H.Thayer,
may be of real service to the animal as a factor in con-
cealment.

In the development of white pattern-marks, the evi-
dence seems to show that these come in at first as small
and fluctuating spots, which may be of little effect in the
economy of the animal. Their further development might
lead to the extinction of the species if they render it too
conspicuous to enemies, unless the species at the same
time makes use of them or accommodates itself to their re-
vealing effect. Often, no doubt, they may not be a source
of danger at all. A case in point may be that of Sciurus
finlaysoni, a Malayan squirrel, most of the individuals of
which are largely marked with white, and of which speci-
mens may be found side by side, varying from an almost
entirely pigmented condition to one of completely white
coat and black eyes. White squirrels are occasional in
other species, as albinos, but these rarely survive more
than a generation in the cases I have known, whereas

Digitized by

Google

478 THE AMERICAN NATURALIST [VoL-XLTDS.

Finlay son's squirrel seems to have accustomed itsell \yj
gradual stages to the white condition, so that it is prol^
ably not at a great disadvantage by reason of its
whiteness.

Pigment Patches in Bibds

In birds the same primary pigment patches seem to bo
present as in mammals, and they are homologous in tbe
two groups. In defining the extent of the pigment patolies,
however, allowance must be made for the fact that the long
feathers may cover a part of the body remote from tlieir
origin. The distribution of the feathers or the pterjrlosis
of the species in hand must also be remembered. In
order to arrive at the true interpretation of the patoTies,
it is necessary to consider the pigment as projected back
from the vanes of the feathers to the part of the l>ody
at their bases. By so doing, it becomes evident thL^t a
feather variegated with pigmented and unpigmented ( or
white) areas indicates none the less that the tesLt^TcL^T
arises from a place of pigment formation. It is orxly ^
wholly white feather or patch of feathers that caxx ^^ ^^

considered albinistic in the sense here intended. ^-^J^^ ''^

factor determining the intermittent formation of '^p^^
ment in the individual feather is probably a v^Yimq/^
different one from that determining the presence or ^i
sence of pigment formation at certain places on t^J
body, though not necessarily different except in its inter,
mittent action.

In the domestic pigeon of our streets and buildings,
we have a species that in its wild state is normally fully
pigmented except .for a white rump patch. Under semi,
domestication it has developed partial albinism to a large
degree, so that it is possible to obtain a complete series
representing on the one extreme a totally pigmented
bird without a trace even of the white rump patch, and
on the other extreme a bird of pure white plumage. A
few of the intermediate stages in areal reduction of pig-
mentation are shown in Figs. 49 to 53, selected from
birds raised for the market and, so far as known, not

Digitized by

Google

No. 572]

PATTERN DEVELOPMENT

479

bred for pattern. The first steps in reduction are shown
in Fig. 49. Here there is seen first a crescentic band of
white feathers passing from eye to eye around the occi-
put. This is a primary break marking off the crown
patch posteriorly. This patch in birds, in contrast to its
development in mammals, is the main patch of the head,

s-o

6^/

Figs. 49-63. Diaobams Showinq Pigmentation in the Domestic Pigeon.

covering the area from the base of the bill to the eyes
and occiput. In Fig. 50 its posterior limit is similarly
defined by a primary break separating it from the neck
patches, and although it does not extend forward quite
to the eye in this specimen, it shows a beginning of sepa-
ration from the more lateral ear patches by virtue of the
indentations on each side posteriorly. In Fig. 52, the
crown patch is shown slightly reduced in extent and
wholly separate from the ear patches, which have become
inactive altogether. In Fig. 53 it has dropped out with
the latter. It is evident then that by greater or lesser

Digitized by

Google

480 THE AMERICAN NATURALIST [Vol. XLVHI

reduction of the crown patch alone it is possible to pro-
duce a pigeon with a mere white spot at the back of the
head, one with a white stripe from the base of the beak
through the eyes to the back of the head (or some part
of such a stripe) to a pigeon in which by the total reduc-
tion of the patch, the entire top of the head is white.
Such specimens can be found in most any miscellaneous
flock. There is a tendency often for the patch to be irreg-
ularly broken, sometimes divided almost into two parts,
a result of the pterylosis to some extent.

The ear patches in pigeons, and probably in all birds,
are rather insignificant, and the smallest of all the pri-
mary pigment areas. They include the feathers from
the posterior angle of the lower mandible to the angle of
mouth and thence back, including the ear coverts. I do
not feel sure that the patches of opposite sides may not
join on the chin, but the present evidence tends to show
that the chin is pigmented by a forward extension of the
neck patch, which, under reduction, often leaves a small
island of pigment between the mandibular rami. In Fig.
50 the neck patches are seen to have broken away ante-
riorly from the crown and ear patches and the separa-
tion of the latter from the crown is indicated by deep
reentrants along the line of the separation. In Fig. 51 a
remnant of the ear patch of the left side alone remains
in dorsal view, consisting of a small tuft of pigmented
feathers' at the fore end of the aural area and a single
pigmented feather just behind it. In this specimen there
are a few pigmented feathers on the chin as well, which I
take to be an isolated bit of the neck patches.

The neck patches are bilateral in origin, and pigment
the entire throat and neck back to a point corresponding
to the base of the neck vertebrae. They meet the crown
patch and separate the ear patches at the occiput. In the
domestic pigeon the neck patches correspond very closely
to the area of differentiated feathers that give the metal-
lic reflections. In the reduction of this area it is common
for the anterior part of the throat to be white, and then a

Digitized by

Google

No. 572]

PATTERN DEVELOPMENT

481

break occurs between the neck patches and those of the
head as in Fig. 50. Posteriorly the neck patch under
reduction may become separated by a white ring at the
base of the neck, from the shoulder patches as in Fig. 50.
The ultimate centers of these patches seem to be in the
pigeon well back on the base of the neck. These are
shown, of small extent, in Fig. 51, as two small areas of
pigmented feathers, one on each side of the base of the
neck. In Fig. 53, further reduction has taken place, so
that the patch of the left side only remains as a small
center. In Fig. 52 there is a large median dorsal patch,
which, as in mammals, may represent the two centers of
opposite sides which even under much reduction have not
in this individual become divided medially.

A very common manifestation of pigment reduction in
pigeons is to have the primaries or some of them white,
as in Figs. 49 or 50. This indicates a failure of pigment
to develop at the extremities of the shoulder patches,
just as in mammals white forefeet mark a slight reduc-
tion of the same areas. It is a fact of much interest that
in the guinea fowl {Numida)^ which has been under
domestication but a short time comparatively, a distinct
breed has arisen in which this same reduction of pigment
is present, resulting in a speckled bird with pure white
primaries and often a pure white area on the breast. In
the pigeon, further reduction cuts off a narrow ring of
pigment encircling the breast, or, it may be, broken in the
mid-ventral line. This ring represents the reduced
shoulder patches, and is to be seen in many wild species
as a permanent part of the pattern. The white collar at
the base of the neck in Fig. 50 marks the separation
between the neck and the shoulder patches at the ante-
rior border of the latter. In other specimens the patches
are separated medially by a white area down the back.
The ultimate centers of these patches seem to be near
the elbow or on the upper arm at the base of the tertiaries,
as seen in Figs. 52 and 53.

Digitized by

Google

482 THE AMERICAN NATURALIST [Vol. XLVIII

The side patches are rather small and seem to center,
as in Fig. 51, near the groin on either side. They pig-
ment the belly back of the breast area included by the
shoulder patches, and extend on to the hind legs as well.
In a specimen before me, the shoulder patches pigment
the bases of the wings and the entire breast correspond-
ing roughly to the length of the sternum, and tend to be
separated by encroaching white feathers midventrally.
The side patches ara much more reduced, and are con-
fined to a small area at the top of each thigh. The re-
mainder of the patches has become inactive, so that a
completely white belly and back result. A very common
occurrence is the white rump patch due to the restriction
of the side patches, so that a break occurs between them
and the tail patches. The rump patches in birds are
situated far back, as in mammals, and pigment the tail
coverts and the rectrices as in Figs. 49-51. The bilater-
ality of the two patches is often indicated in pigeons by
the occurrence of a few pure white rectrices in the center
of the tail. Other birds show pure white feathers at
either side of the tail, with a tendency ta bilateral sym-
metry, a most important fact, since it indicates restriction
at the outer extremes of these centers. In the restriction
of pigment formation, the rectrices are the first to be-
come white, as one would expect, since they are situated
at the extremity of the body and farthest from the center
of the patch. In Fig. 52 these centers are seen to be at
the base of the tail above, and include the upper tail
coverts. They are still joined medially, but that of the
left side is more extensive than the patch on the right side.
The approximate boundaries of the several pigment
patches are indicated in Fig. 53 by dotted lines ; 1 is the
crown patch, 2 the ear patch, 3 the neck patch, 4 the
shoulder patch, 5 the side, and 6 the rump patch, as they
appear in a dorsal view. Ventrally the neck patch runs
forward to the symphysis of the mandibles.

In a flock of domesticated mallard ducks which I
studied, the same patches were found indicated, and

Digitized by

Google

No. 572] PATTERN DEVELOPMENT 483

some of the details of these are shown in Figs. 54-56.
In the male wild mallard there is no white in the pattern
of the head and neck except a white ring at the base of
the neck. In one of the domesticated breed, shown in
Fig. 54, the crown patch was very beautifully marked
off, as in the pigeon (Fig. 49), by a white band from eye
to eye passing about the occiput. This duck was further
interesting ii^ showing the median division of the two
neck patches, ^s a narrow white line running down the

rt Si

Sf.

Figs. 64-^7. Diagbams Showing Pigmentation in Domesticated Mallard
Ducks and in the (Wild) Labrador Duck (57).

back of the neck medially, from the occipital stripe. An-
other duck shown in Fig. 55 had lost the nfeck patches
entirely, but showed the same occipital stripe bounding
the crown patch posteriorly, and the ear patches dorsally.
The ear patches still adjoin the crown patch anteriorly.
In Fig. 56 is represented another of these ducks in which
both ear patches are distinct and separate on either side
of the head. The crown patch appears as two narrow
lines of pigmented feathers which are not quite in con-
tact posteriorly. I have not obtained a satisfactory
explanation for the apparent tendency of this patch to

Digitized by

Google

484 THE AMERICAN NATURALIST [Vol. XLVHI

divide medially. Probably for some reason the forma-
tion of the pigment is more intense at the sides of the
crown than in the center where the nerve and blood
supply is less. In the pterylosis of this area the develop-
ment of feathers is seen to be greater at the sides also.

The neck patches and the side patches are absent en-
tirely, but the shoulder patches are both present, in Fig.
56, that of the right side covering the scapulars and
middle of the upper back, that of the left side including
a few only of the scapulars.

The tail patches are both present, and separate from
each other, as shown by the median white rectrices.

In this same flock of mallards was a female which had
a white ring at the base of the neck in the same situation
as the white ring which in the male is a part of the per-
manent pattern. It was not quite complete dorsally,
however, in this female, and was somewhat broader than
regularly in the male. Nevertheless, it is apparent that
this white collar in the male is merely a primary break
between neck and shoulder patches that has become
developed as a part of the normal pattern.

Stone (1912, p. 318) in his paper on the phylogenetic
value of color characters in birds, hints at the existence
of these patches. He says, in part :

In matters of pattern there seems to be a deeper problem involved,
i. e., the determination of the cause governing the appearance of a dif-
ferently colored patch on corresponding parts of the plumage of birds
belonging to wholly different groups ... or the presence of a mystacial
stripe, a superciliary stripe, a light rump patch. ... In fact if a bird
exhibits a bright or contrasting patch of color, it is, in the vast major-
ity of cases, found on one of several definite portions of the plumage,
as the crown, the throat, the bend of the wing, the rump, «te.

These contrasting areas are due to the development of
one or more of the primary patches, or of breaks between
them, or again paler areas, as at the bend of the wing or
on the rump, indicate often a lessening of pigment inten-
sity at a distance from the respective primary centers.

(To he concluded,)

Digitized by

Google

NOTES ON THE MEADOW JUMPING MOUSE

(ZAPUS HUDSONIUS) ESPECIALLY RE-

GARDING HIBERNATION

H. L. BABCOCK, M.D.
Dedham^ Mass.

The jumping mouse is the only one of the wild mice of
this region (Massachusetts) which exhibits the habit of
regular hibernation. Regarding this habit there are a
number of references in the literature on the subject.
Barton* was one of the first to refer to the fact that this
mouse became dormant in winter. He says, in describing
the actions of one he had in captivity :

On or about the 22d of November it passed into the torpid state. It
is carious to observe that at the time it became torpid the weather was
unusually mild for the season of the year, and moreover the animal was
kept in a warm room, in which there was a large fire the greater part
of the day and night ... It was frequently most active while the
weather was extremely cold in December.

This was in Philadelphia, Pa.

Audubon and Bachman* regret that they live in a region
where the species does not exist and can not speak from
personal observation on the subject.

Godman,^ Thompson* and Kennicott^ speak of its habit
of hibernation.

Tenney* gives an account of a specimen of this species
taken alive on January 18, 1872, near Vincennes, Ind. It
was dormant, coiled up tightly, **the nose being placed
upon the belly, and the long tail coiled around the ball-like

1 ' ' Some Accomit of an American Species of Dipus or Jerboa, ' ' bj Ben-
jamin Smith Barton, MJD., Translatiana of the Am. Philosophical Society,
Vol. IV, No. XII, 1799.

* Viviparous Quadrupeds of No. America/' Vol. 11, 1851, p. 255.
« Godman, "Am. Nat. Hist.,'' Vol. I, 1842.

4 Rev. Zadoc Thompson, "Nat. and Civil Hist, of Vermont," 1842.
sKennicott, Patent Office Report for 1857.

• Tenney, "Hibernation of the Jumping Mouse," Am. Naturalist, June,
1872, Vol. VI, No. 6, pp. 330-332.

485

Digitized by

Google

486 THE AMERICAN NATURALIST [Vol. XLVm

form whicli the animal had assumed. " It was taken from
a nest about two feet below the surface, made of bits of
grass. The mouse showed no signs of life at first, but on
being held in his hand, soon became feebly active, and on
being placed in a warm room, came out of its dormant con-
dition entirely. It again became dormant that night, but
was aroused twice again by the application of heat, within
the next few weeks, in spite of very cold weather.

Merriam^ tells of taking an active male at Easthamp-
ton, Mass., on February 11, 1872, and states that during
the mild winter of 1881-82, in Lewis County, Northern
New York, he saw jumping mice active several times.

Seton^ speaks of finding a Zapus Hudsonius on Sep-
tember 27, 1888, at Carberry, Manitoba, in a nest of leaves
under the roots of a stump, nearly torpid. He says :

In the country near Carberry, I never saw it active after September
first.

Stone and Cram® believe that this mouse passes six
months or more of every year hibernating underground.
They speak of seeing a family of them turned up by a
plough in May and exhibiting not the slightest symptom
of life, on being handled or breathed upon.

Burroughs^^ tells of a female jumping mouse in cap-
tivity that began hibernating early in November and con-
tinued until May, with several intervals of activity, espe-
cially after warm weather came on.

Preble^ ^ says:

Hibernation varies with the locality, but usually begins about the time
of the first heavy frosts and lasts until Spring. The fall pelage is
usually assumed and the animals become exceedingly fat before entering
winter quarters. Although they often lay up stores of food in nests or
burrows during summer, it is not known that they use this food during
winter. The animals are generally found singly (sometimes in pairs)
in nests at a depth varying from a few inches to two or three feet below
the surface. Hibernation sometimes takes place above ground.

7 C. H. Merrian, M.D., ** Mammals of the Adirondack Region," 1884.
8E. T. Seton, ** Life-histories of Northern Animals," Vol. I.
» Stone and Cram, ** American Animals," p. 103-104.
10 John Burroughs, '* Squirrels and Other Fur Bearers," pp. 121-124.
11 E. A. Preble, "Revision of the Jumping Mice of the Genus Zapus,'*
U. S. Dept. Agr. N. A. Fauna Series, No. 15, 1899.

Digitized by

Google

No. 572] NOTES ON MEADOW JUMPING MOUSE 487

On June 25, 1912, a female Zapus Hudsonius was taken
alive, by the writer, on the edge of a small pond in eastern
Massachusetts. It was placed in a small wire cage, and
after a few frenzied efforts to escape, became quite tame.

On July 5 it gave birth to five young, blind and hairless ;
but when the family was transferred to a larger cage, the
mouse deserted the young and they soon died. One dis-
appeared mysteriously, and may have been eaten by the
mother. The young measured at birth: total length 33
mm. ; taiil 9 mm. ; hind foot 4 mm.

Throughout the summer the mouse ate chiefly rolled
oats and shredded wheat, and was also very fond of straw*
berries and blueberries. It refused most of the common
fruits and vegetables.

It was almost wholly nocturnal in its activity, although
when disturbed during the day it would immediately begin
to eat and remain active for half an hour or more.
Toward the latter part of the summer, it seemed to grow
quite fai Rhoads^^ says in this connection:

When going into winter quarters they are exceedingly fat, as I can
testify from experience in removing this tenacious yellow blanket from
the skips of them. This fat is their fuel. By spring it is nearly gone.

During the latter part of August there were several
very cool nights (49'' F. minimum) and on the night of
August 28 it did not come out. This fact was apparent
from the clean drinking dish, which was placed in such a
position that the mouse could not approach without scat-
tering saw-dust in it The absence was repeated on Au-
gust 30, and September 1. Throughout September its
actions were irregular. Every night until the 21st, with
the exception of the 12th and 17th, it was active, but on
the 22d disappeared for four nights. It was then active
for two more nights (26 and 27) and following that,
inactive for six (September 28 to October 3). From Oc-
tober 4 to 28 it was out every night, although not as vigor-
ous as formerly, neither did it eat as much. When ap-
proached it seemed to pay no heed, as if in a sort of
stupor.

12 S. N. Bhoads, ' ' The Mammals of Pennsylvania and New Jersey. ' '

Digitized by

Google

488 THE AMEBIC AN NATURALIST [Vol. XLVin

OrPlc/AL Observations

Fffon US' W^ATHM/r Bvftau, Bosroft Sr^mN-

Chmt SHow'm NJHiMt/^ htf^HTir

• » /fl6HTS OH WMtCH THE /iOOSe WIS NoTACTir£.

There was no evidence of any attempt at storing away a
supply of food, although there was ample opportunity.
This habit of storing food is mentioned by Homaday**
who says :

In the autumn it stores in the ground quantities of food for winter
use, but despite this fact, under certain conditions, it becomes so thor-
oughly dormant in winter that it seems to be quite lifeless.

According to Seton,^

It is quite ready to respond at any time to any spell of unusually fine,
unseasonable weather, even in the depths of winter, and it is probably
for these arousing times, as much as for the spring time famine, that it
lays up its abundant stores of food.

Preble^ ^ also mentions this habit, but Shufeldt" denies
it He says, in speaking of the deer mouse {Peromyscus
Leucopus) :

Is it to meet the requirements of his condition that this mouse lays up
a goodly stock of food during the autumn t Something the Zapus does
not do.

Following the period of activity through October, the
mouse was inactive on the four nights of October 28, 29,
30 and 31, and reappeared for the last time on the night
of November 1, after which it retired for the winter. The
cage was placed by an open window of an empty box stall
in a stable where the temperature was practically that of
out doors. The mouse built its nest in the side of a large
sod placed in one comer of the cage.

In spite of a very mild winter, the lowest oflScial tem-

18 Hornaday, * * The American Natural History. ' '

i*R. W. Shufeldt, M.D., "Chapters on the Natural History of the United
States."

Digitized by

Google

No. 572] NOTES ON MEADOW JUMPING MOUSE

489

October NovE/fBEFr «^^

perature for this section being only 3° F. (February 10,
1913,) the monse did not survive the cold weather, and
was found dead, when the cage was opened on June 17,
1913.

The nest was found to be located in the extreme end of
the sod, only IJ inches from the top and about 1 inch from
the edge. It was roughly oval in shape, being hollowed
out of the loam and lined with a few blades of grass. It
measured roughly IJ inches by 1^ inches and was just
large enough to contain the mouse when curled up into a
ball. The opening was on the side. Death was probably
caused from exposure to continued cold owing to the un-
protected location of the nest.

The poor judgment shown in not building the nest
securely in the middle of the large sod, and other similar
instances of poor management, have led the writer to
believe that the intelligence (if tiiat term may be used) of
the Zapus Hudsonius is of comparatively low grade, much
lower, for instance, than that of the deer mouse {Peromys-
cvrS Leucopus).

The accompanying chart, which is a record of minimal
nightly temperatures, according to the official observa-
tions of the TJ. S. Weather Bureau for this section (Bos-
ton), shows the activity of the mouse in relation to the
temperature during August, September, October and No-
vember, and brings out some rather interesting facts. For
example, on October 15, 16 and 17, with the minimum
nightly temperature 42°, 36° and 42°, respectively, the
mouse was active, while on September 22, 23, 24 and 25,

Digitized by

Google

490 THE AMERICAN NATURALIST [Vol. XLVIH

with the minimum nightly temperature of 48°, 49°, 56°
and 53° F., respectively, a much warmer series of nights,,
it remained inactive. And again, after November 1, dur-
ing a warm spell in which the minimum nightly tempera-
ture for November 7 was 64° F., the mouse did not appear.

A study of this chart suggests the question as to how
much the temperature has to do with this habit of hiber-
nation.

It is a somewhat general belief that temperature regu-
lates the degree of torpidity.

Barton^ maintains that

the torpid state of animals *is altogether an accidental circumstance and
by no means constitutes a specific character. The same species becomes
torpid in one country and not in another. Nay, different individuals
of the same species become torpid or continue awake in the same neigh-
borhood or even on the same farm.

Seton® believes that

while torpor is more or less controlled by temperature, the habit of tor-
pidity, like the changing pelage of the white-hare, is so deeply ingrained
constitutionally that there is a strong tendency to torpify at a given
time without regard to the original cause.

It is evident from this chart that torpidity develops
gradually, at first for only one night at a time. Whether
this process is explained by a cerebral anemia, a slow
toxemia of the brain centers or some other of the theories
regarding sleep, it seems to require about two months in
which to become sufficiently developed to control com-
pletely voluntary body functions. During that interval
the animal occasionally awakens, probably from hunger
and habit as much as from any effect in change of tem-
perature.

After torpidity is thoroughly established, changes of
temperature may be important external factors, as has
been demonstrated on numerous occasions in producing a
temporary activity by the application of heat. It is safe
to say, however, that the temperature is not the only ele-
ment which influences the length of the period of hibema^
tion.

Digitized by

Google

SHOETER ARTICLES AND DISCUSSION
STUDIES ON INBREEDING— IV

On a General Formula for the Constitution op the nTH
Generation op a Mendelian Population in which

ALL Ma^TINGS are OP BROTHER X SiSTER ^

I. In a former paper in this series^ the constitution of a Men-
delian population in which all mating was of the brother X
sister type was worked out empirically. The results there pre-
sented may be put in the form of a general formula, by means
of which the constitution of any generation may be written down
from a knowledge of the preceding generation; that is from a
knowledge of the n — 1th generation the nth generation may be
at once written down.

II. This general formula may be developed as follows. A
single character pair will be considered, A denoting the dominant
character and a the recessive. Equal fertility for all matings
is assumed, the number of individuals per family being taken as
25, of which s are males and s are females. One family will then
make s matings and produce s families in the next generation.
Each mating is, by hypothesis, of a brother with his sister.

Starting as before with a pair from a population in which all
individuals are of constitution Aa there will be in the next
generation one family of the AA + Aa + aA + aa type. In all
succeeding generations there will be six types of families, viz. :

(1) A A families.

(2) AA-\-Aa families.

(3) Aa families.

(4) 'Aa-}-2Aa-\- aa families.

(5) Aa^ aa isjmlies,

(6) aa families.

1 Papers from the Biological Laboratory of the Maine Agricultural Ex-
periment Station, No. 66.

It seems desirable to publish sa a general series of ** Studies on In-
breeding" the results of certain investigations now in progress in this
laboratory. The three papers which have already appeared in this series,
without the general title, are: 1. **A Contribution towards an Analy-
sis of the Problem of Inbreeding,'' Ameb. Nat., Vol. XLVII, pp. 577-
615, 1913. II. ''Tables for Calculating Coefficients of Inbreeding," Ann.
Eept. Me. Agr. Expt Sta, for 1913, pp. 191-202. III. ''On the Results
of Inbreeding a Mendelian Population: A Correction and Extension of
Previous Conclusions," Ameb. Nat., Vol. XLVIII, pp. 57-62, 1914.

2 Amir. Nat., Vol. XVLIII, pp. 57-62, 1914.

491

Digitized by

Google

492 THE AMERICAN NATURALIST [Vol. XLVIH

The proportionate number of each of these types of families
will change in successive generations according to the following
system.

Let o»_i denote the number of A A families in the n — 1th

generation, and
Pn-x denote the number of AA'\'Aa families in the

n — 1th generation, and
gn-i denote the number of Aa families in the n — 1th

generation, and
r„_i denote the number of AA and 2Aa and aa families

in the n — 1th generation, and
Un-x denote the number of Aa-}-aa families, and
Vn-i denote the number of aa families.

It will be possible to write down u and v in any case without
calculation because of the synunetrical relations of a Mendelian
population, since always under normal conditions such as are
assumed in the general treatment, we have

^-l=Pn-i,

Vn.i = On.iy

Vn =0n.

It is necessary, therefore, to consider only the coeflScients for
the first four types of family. In the nth generation the consti-
tution of the population in respect of families (not individuals)
will be as follows :

Families in nth generation

= s(On.t + l/4tpn.i + l/16r,..JAA famiUes
+ 5(l/2p„.i + l/4trn.x)AA +Aa families
+ s(l/8rn.i)Aa families
,+ 5(l/2pn-x + qn-x + l/4trn.,)AA

+ 2Aa + aa families
'^8(v^)Aa + aa families
j'^s{vn)aa families.
Or, taking coefficients alone we have

On = On.t + l/4pn-i + l/16rn.i,

Pn = l/2p,.i + l/4rn_i,

gn = l/8rn.i, ...

rn = l/2pn-x + gn-x + V^n.^, ^ ^

Un = l/2w»_i + l/4rn_i = Pn,

Vn = Vn.i + l/4Wn-i + l/16r„.i = On.

Digitized by

Google

No. 572] SHORTER ARTICLES AND DISCUSSION 493

III. Let us see how this formula works out in a concrete case.
Assume the same conditions of fertility as in the former paper,
that is, put 2s = 32, or s = 16. Start with a single -4.-4. + 2Aa
-[- aa family.

Then On-i = 0,

^"-1 = 0,
r„.i = l.
Then in the next generation we shall have
16(0 + 1/4(0) +1/16(1)} = 1AA family
+ 16(1/2(0) + 1/4(1) }=:4(4A + 4a) families
+ 16(l/8(l)} = 24a families

+ 16(1/2(0) +0 + 1/4(1)} = 4(44 + 24a + aa) families
+ 4(4a + aa) families
+ laa family.

This is the fact.

In the next generation we shall have

16(1 + 1 + 1/16(4)} =3644 families

+ 16(1/2(4) +1/4(4)} =48(44 + 4a) families

+ 16(1/8(4) } = 84a families

+ 16(1/2(4) +2 + l/4(4)} = 80(44 + 24a + aa) famiUes

+ 48 (4a + aa) families

+ 36 (aa) families.
This is the fact.
In the next generation we shall have

16(36 + 1/4(48) +1/16(80)} = 16 X 53 = 84844 famiUes
+ 16(1/2(48) + 1/4(80) } = 16 X 44 = 704(44 + 4a)

families
+ 16(1/8(80) } = 1604a families
+ 16(1/2(48) + 8 + 1/4(80) } = 16 X 52 = 832(44 + 24a

+ aa) families
+ 704 (4a + aa) families
+ 848aa families.

Succeeding generations follow the same law and need not be
worked out in detail.

IV. So far the discussion has confined itself to families, as this
must be the basic unit in the theory of any form of inbreeding.
Turning to individuals we have the following simple relations to
pass to individuals.

In the nth generation the number of

Digitized by

Google

494 THE AMERICAN NATURALIST [Vol. XLVIH

AA (or oa) individuals = 2s (o„) V4-5(p«) +l/2s(r„),
Aa {or a A) individuals = 25 (gn) +s{l/2pn) + 1/25 (rn).

The first of the above expressions multiplied by 2 gives the
total homozygotes, and the second multiplied by 2 gives the total
heterozygotes.

Raymond Peael

PARALLEL IViUTATIONS IN OENOTHERA BIENNIS L.

In the summer of 1912 I cultivated pure strains of 0. biennis
L. and of the 0. biennis cruciata de Vr. of our Dutch dunes, as
well as of their hybrids, made with the purpose of studying the
behavior of the cn/ciafo-character in crosses. In one of these
cultures I unexpectedly obtained two mutants, which because of
their similarity to corresponding variants derived from 0.
Lamarckiana have been called 0. biennis nanella and 0. biennis
semi-gigas. The first mutant, 0. biennis nanella, occurred in the
second generation of the cross 0. biennis X 0. biennis cruciata
and differed from 0. biennis in all those points which separate
0. Lamarckiana nanella from 0. Lamarckiana, The other vari-
ant, 0. ^biennis semi-gigas, appeared in the second generation of
the reciprocal cross, 0. biennis cruciata X 0. biennis, suggesting
immediately by its much more vigorous habit and especially by
the larger size of its buds and flowers the differences between O.
Lamarckiana and 0. gigas, A count of its diploid number of
chromosomes proved it to deserve the name semi-gigas, 21
chromosomes being shown by nuclear plate-stages in the meris-
tematic tissue of young buds. From these facts, showing that
0. iiennis is in a mutating condition, I drew the conclusion that
the phenomenon of mutation in the genus (Enothera is older than
the species 0. Lamarckiana — 0. biennis generally being consid-
ered to be an older species than 0. Lamarckiana — and further,
that the mutations in this group can not be the result of hybridi-
zation, as was assumed by some authors at that time — ^nobody
doubting of the purity and constancy of 0. biennis. As a mat-
ter of fact, both of my mutants have been derived from crosses
between 0. biennis and 0. biennis cruciata. But I laid special
emphasis on the fact that 0, biennis and 0, biennis cruciata have
exactly the same germinal constitution except for the factors that
determine the shape of the petals, 0, biennis cruciata being prob-

Digitized by

Google

No. 572] SHORTER ARTICLES AND DISCUSSION 495

ably a mutant from 0. biennis itself. Therefore, hybrids between
these two forms can be looked upon as pure 0. ^biennis except for
floral characters.

With this conception Bradley Moore Davis does not agree.^
He thinks that the 0. biennis and 0. biennis cruciata of our dunes
are not so closely related types, that a cross between them can
be treated **as though it were the combination of forms within
the same species which have similar germinal constitutions." He
says:

It should be made clear that the form '^ 0, biennis cruciata " is recog-
nized in the more recent taxonomic treatments as a true species sharply
distinguished from types of biennis by its floral characters. ... 0.
cruciata is found wild in certain regions of New England and New
York and is consequently a native American species. . . . Whatever
may have been the origin of 0. cruciata or its possible relationship to
O. biennis, a cross between these t3rpes must certainly be regarded as a
cross between two very distinct evolutionary lines and its product a
hybrid in which marked modifications of germinal constitution are to
be expected.

From Davis's point of view I '* really made a cross between
two rather closely related species'' and obtained in the second
generation 'Hwo marked variants due to some germinal modifica-
tions as the result of the cross. " In so far as my observations bear
upon the problem of mutation Davis's interpretation is exactly
the reverse of mine. To him they further illustrate the same
phenomenon which he is obtaining through his ** hybrids of
biennis and grandiflora, namely, that behavior by which these
hybrids in the F^ generation throw oflf variants that in taxonomic
practise would be considered new species readily distinguished
from the parents of the cross and from the Fj hybrid."

It will be shown in the following lines that the objections made
by Davis are not sufficiently justified. My argument consists of
two points.

In the first place, Davis is mistaken as to the nature of the 0.
biennis cruciata de Vr. of our dunes. This strain is in reality
quite another type than the different forms of the American 0.
cruciata Nutt., called by some authors 0. bienrm cruciata. With
this species it has in common only the character of the narrow

1 Bradley Moore Davis, ''Mutation in (Enothera biennis L.t" Tmc Ameri-
can Natuealist, Vol. XLVn, 1913, pp. 116-121; ''Genetical Studies on
(Enothera/' IV, The American Naturalist, Vol. XLVII, 1913, pp.
546-671.

Digitized by

Google

496 THE AMERICAN NATURALIST [Vol. XLVTEI

petals, all other features of the stem, foliage, flowerspikes and
fruits being exactly those of the Dutch 0, biennis L. It must
certainly be looked upon as a mutation from the 0. biennis L. of
our sand dunes. Until now it has only been found a couple of
times in single individuals in the midst of the ordinary 0. biennis,
the first time in 1900 by Dr. Ernst de Vries in the dunes in the
neighborhood of Santpoort, Holland, in one individual — and
from this one specimen all the subsequent generations of 0. bien-
nis cruciata in the cultures grown by de Vries and by myself have
been derived. Besides this, our 0. biennis and 0. biennis cruciata
are so similar to one another except for floral structure that
plants of both types can not be separated before the flowers open.
Therefore we have the right to assume that the crossing of these
two forms is concerned alone with the floral characters and that
with respect to all other characters parents as well as hybrids are
mere biennis. Therefore the two variants which arose in my cul-
tures from crosses between 0. biennis and 0. biennis cruciata
obviously prove the faculty of mutation in 0. biennis.

In the second place I have found now that it is not necessary
to cross 0. biennis with 0. biennis cruciata in order to obtain the
above named mutants, as Davis seems to believe. Already in his
new book Professor de Vries figures a dwarf derived from O.
biennis cruciata grown in pure line. Shortly afterwards I myself
obtained six mutants from the 0. ^biennis of our sand dunes grown
also in pure line. A few details about these cultures of last year
may be given here. In all they counted 920 individuals, 430 of
which belonged to the third and 490 to the fourth generation of
a pure line, the point of departure for which had been one in-
dividual brought into the experimental garden in the rosette
stage from the dunes near Wyk aan Zee in the beginning of 1905
and self -fertilized in the same year. The six mutants which ap-
peared in these pure cultures of 0. biennis were the following.
First a dwarf, then a "biennis semi-gigas having 21 chromosomes
and finally four individuals of the 0. biennis sulfurea, a pale-
flowered form of 0. biennis, which had been found already several
times in our dunes in the midst of the ordinary biennis, but was
not with certainty known to be a mutant from the latter form
until now. The two flrst named mutants and one sulfurea ap-
peared in the third generation of our pure line, the nanella and
the semi-gigas coming from the same mother. The three remain-
ing ^/7/wrea-individuals appeared in the fourth generation, all

Digitized by

Google

No. 5721 SHORTER ARTICLES AND DISCUSSION

497

descending from the same motherplant. Of these mutants the
nanella and semi-gigas are especially valuable because similar
forms have been produced by 0. Lamarckiana, It will be seen that
the MenniS'dwaris seem to be somewhat rarer than the dwarfs of
Lamarckiana. Whilst for the latter the mutation coefficient is about
1 per cent, our 0. biennis nanella appeared as the only dwarf among
920 individuals. The above cited 0. biennis cruciata nanella
was the only dwarfed individual in a culture of 500. And the
dwarf which I got in 1911 was the only one among about 600
plants. In this connection I wish to recall the conclusion
reached by de Vries that in 0. Lamarckiana the pangen for tall
stature must be assumed to be present in the labile condition on
both sides, in 0. biennis, however, only in the male sexual type,
whilst in the female sexual type active alta-pangens have to be
supposed. The way from biennis to biennis nanella might there-
fore possibly be somewhat longer than the one from Lamarckiana
to Lam4irckiana nanella. The biennis semi-gigas which appeared
in the last sunmier corresponded in all points exactly with the
mutant of 1911. Moreover a count of the chromosomes, as shown
by nuclear plate-stages in the meristematic tissue of young buds,
determined them to be 21 in number. Even as the specimen of
this type, that appeared in 1911, and as the semi-gigas mutants
produced by 0. Lamarckiana, the plant of last year proved to be
almost absolutely sterile.

In his second above-mentioned paper Davis says about the 0.
biennis of our dunes: **No species of (Enothera is perhaps so free
from suspicion as to its gametic purity. If Stomps can obtain
mutations from tested material of the Dutch biennis grown in
pure lines he will have the basis of a strong argument. ..."
Portimately the experiment asked for by Davis, has been made
in the same year as his criticism. The Dutch biennis L., culti-
vated in pure line, has produced a dwarf, a semi-gigas and some
stdfurea-mdiyidnsls, proving its mutability beyond all doubt. I
therefore trust that the conclusions arrived at in my first paper,
concerning this mutability and its consequences, may now be ac-
cepted as thoroughly valid.

Theo. J. Stomps

Amstebdam, Holland

Digitized by

Google

498 THE AMERICAN NATURALIST [Vol. XLVni

In a recent review^ of Stomps 's studies on CEnothera biennis
L.2 from the sand dunes of Holland I protested against his desig-
nating as mutants a naneUa type and a semi-gigas type which
were obtained in the second generation of crosses between (Eno-
thera 'biennis LinnsBus and its variety 0. biennis cruciata de
Vries. The criticism was presented on the general ground that
however close the possible relationships between the two parent
forms, they nevertheless constituted lines so far apart as to
render unsafe a conclusion that marked variants obtained from
their crossing are mutants in the sense of de Vries and Stomps.
Such variants, it seemed to me, might have been the result of
hybridism between two lines suflSciently divergent to upset the
similarity of germinal constitution shown in their vegetative
morphology, for the species biennis and its variety cruciata are
said to differ only in their flower structure.

In that review I incorrectly associated 0. biennis cruciata de
Vries with 0. cruciata Nutt., an American species entirely dis-
tinct from the variety cruciata of de Vries, which has been found
only once (in the year 1900) on the sand dunes of Holland
among plants of 0. biennis, I greatly regret my confusion of
these two types, since I was led in my criticism to regard Stomps 's
crosses between biennis and biennis cruciata as though they were
crosses between two distinct although possibly closely related
species. In this I was clearly mistaken, since all of the evidence
short of experimental proof, which Stomps may yet obtain, indi-
cates that biennis cruciata de Vries is a variety of biennis L. and
arose as a mutation on the sand dunes of Holland. The crosses
of Stomps are, therefore, to be regarded as between a species and
its mutant variety. I trust that the mutationists will accept this
acknowledgment of an error.

There is, I believe, a body of naturalists for whom the value
of evidence for mutation rests fundamentally upon the unques-
tioned purity of the parent stock, and to them any cross, no matter
how close, is open to criticism. Stomps has justified his first con-
clusions by obtaining in later studies the same mutants biennis
nanella and biennis semi-gigas from lines of the pure species O.
biennis Linnaeus. Had he waited for these later results before

1 Davis, B. M., "Mutations in CEnothera biennis L.f American Nat-
uralist, Vol. XLVII, p. 116, 1913.

2 Stomps, T. J., "Mutation bei CEnothera biennis L.," Biol. Centralb,, VoL
XXXII, p. 521, 1912.

Digitized by

Google

No. 572] SHORTER ARTICLES AND DISCUSSION 499

publishing on the first there could have been no objections to his
main contention that 0. biennis from the sand dunes of Holland
is capable of giving rise to true mutants.

Stomps is continuing his studies on this same Dutch biennis
with the view of determining its possible powers of mutation, and
it is a pleasure to review his second paper* which presents some
extremely interesting data, a paper in which no important criti-
cism can be based on the source and character of the material em-
ployed. No wild species of evening primrose has been so long
under experimental and field observation or is better known to
the workers with cenotheras than this plant. The species has
proved uniform in culture to a remarkable degree and it would
be difficult to find a type of (Enothera so free from suspicion of
gametic purity. The species appears to have been in Holland
since pre-Linnaean days and is therefore very old. As material
for experimental studies on mutation the Dutch biennis seems to
the writer the best of all the Oenotheras so far brought into the
experimental garden.

The starting point of Stomps 's cultures of (Enothera biennis
was a plant transplanted from the sand dunes in 1905. From
seed of this plant, self -pollinated, a second generation was grown
in 1910, three selfed plants of which gave the seed for a third
generation of 430 individuals, and a fourth generation of 490
plants was grown from two selfed plants of the third generation.
Thus in all 930 individuals were observed in the third and fourth
generations from the plant that gave rise to these pure lines.
It is true that these lines have not been under selection for many
generations, but, considering the stability of the species and its
habit of close pollination, it is very improbable that the source of
the cultures should have been a plant not representative of the
type. Furthermore, Stomps presumably will continue indefi-
nitely the lines now established and thus determine through
later generations whether their mutating habits remain constant.

Among the 430 plants of the third generation there appeared
1 biennis naneUa, 1 biennis semi-gigas and 1 individual of biennis
svlfurea; the first two came from the same mother plant. Among
the 490 plants of the fourth generation appeared 3 individuals of
biennis sulfurea, all from the same selfed mother. The variety
sulfurea differs from the species biennis in having flowers of a

» Stomps, T. J., ' * Parallele Mutationen bei (Enothera biennis L., ' ' Ber.
deut lot. Gesell, Vol. XXXII, p. 179, 1914.

Digitized by

Google

500 THE AMERICAN NATURALIST [VoL.XLVm

lighter yellow, and is reported by de Vries to be not uncommon
in the wild state mixed with the species proper. Sulfurea has
been held systematically to be a variety of biennis but this is the
first time that it has appeared in the experimental garden as a
derivative of that species. Thus out of a total of 920 plants
there were 4 individuals of the color variety s'tUfurea, 1 nanelk
and 1 semi-gigas, in all 6 mutants, a showing that may well
gratify Stomps.

The mutant biennis nanella differed from typical biennis in
much the same way that Lamarckiana nanella differs from
Lamarckiana and like the latter dwarf showed evidence of a
bacterial infection. Certain selfed flowers set no seed because
of diseased stigmas. Other flowers pollinated from pure biennis
set good fruit. The ratio of the appearance of biennis nanella
is much lower than the mutation coeflScient of one per cent, which
de Vries has reported for Lamarckiana nanella. It should also
be remembered that de Vriies* obtained a crticiata nanella in a
culture of 500 plants from 0. biennis cruciata.

The mutant biennis semi-gigas in comparison with typical
biennis showed a stronger habit, broader leaves, thicker buds,
larger flowers, supernumerary stigma lobes, and the presence
of 4-cornered pollen grains. Counts of the chromosomes in
meristematic tissue determined the number to be 21. Therefore
in this plant, as in the biennis semi-gigas obtained by Stomps
from the cross cruciata X biennis, there is clear cytological evi-
dence that one of the gametes which formed the zygote contained
14 chromosomes, i, e,, double the number characteristic of the
gametes of (Enothera, This is another case of triploid mutants
in (Enothera to be added to the list of Stomps and Miss Lutz.
The plant was self sterile, but set fruit when pollinated by
biennis, although the yield of seed was very poor.

Stomps is justified in calling attention to the agreement of his
second biennis nanella with the plant derived in 1911 from the
cross biennis X cruciata, and of the agreement of his second
biennis semi-gigas with the plant from the cross crudataX
biennis. It is to be hoped that he will next obtain the crumk
variety as a direct mutant from the Dutch biennis and thus
establish its relationship and origin beyond all possible doubt.

Stomps has before him the opportunity of making through the
study of (Enothera biennis very important contributions to our

* See ''Gruppenweise Artbildung, " p. 299 and Fig. 108.

Digitized by

Google

No. 572] SHORTER ARTICLES AND DISCUSSION 501

knowledge of the frequency of mutations and their importance
in organic evolution. That retrogressive mutations take place is
not likely to be seriously doubted by any one who has followed
thel experimental work of recent years both botanical and
zoological. The loss of characters through germinal modification,
even in what seem to be **pure lines,'' appears to be not
uncommon.

Most of all is desired information on the possibilities, fre-
quency and character of progressive mutations. Can the muta-
tion theory satisfactorily explain progressive advances in organic
evolution or must amphimixis chiefly carry that responsibility?
Mutants of the tetraploid gigas-like type would appear to be pro-
gressive, and we can see the reason in their doubled chromosome
count which gives larger nuclei, larger cells and modified tissues.
6H^as-like forms are however very rare and in 0. Lamarckiana
gigas the fertility is relatively low. More common are the trip-
loid semi-gigas forms, but these seem to be sterile or almost
sterile when selfed, and the work of Geerts indicates that the
triploid number in OEnothera returns to the normal through
the elimination of supernumerary chromosomes. Very inter-
esting is the recent paper of Gates and Thomas'^ which offers
evidence that lata-like characters are associated with the pres-
ence of a single additional chromosome.

And what of the series of forms which diflfer from the
(Enothera parent types with as yet no evidence of peculiarities in
their chromosome count, hrevistylis, IcBvifolia, rubrinervis, oho-
vata, scintUlans, etc. Will forms similar to these and perhaps
others in addition be represented in a series of derivatives from
(Enothera biennis? The mutants biennis nanelia and biennis
svlfurea belong to this group and have already been obtained
by Stomps. One may almost envy him his opportunity for an
intensive study of this species.

Bradley Moore Davis

Univeesity of Pennsylvania,
June, 1914.

5 Gates, B, R., and Thomas, N., * ' A Cytological Study of (Enothera mut.
lata and (E. mut. semilata in Relation to Mutation," Quart, Jour, Mic, Soi.,
Vol. LIX, p. 523, 1914.

;

Digitized by

Google

502 THE AMERICAN NATURALIST [Vol. XLVHI

THE THEORETICAL DISTINCTION BETWEEN

MULTIPLE ALLELOMORPHS AND

CLOSE LINKAGE

Professor Castle's diflSculty in understanding the distinction
made by Mr. Dexter^ is owing to his iinfamiliarity at first hand
with the phenomenon of linkage. The distinction between allelo-
morphs and close linkage has already been given several times
elsewhere and need not be repeated ; but if Professor Castle has
failed to note it, or to see its significance, it is probable that
others may have done the same. I may be pardoned, therefore,
for attempting once more to show why, for clear thinking, it is
important to keep in mind the difference between allelomorphs
and close linkage. Furthermore, since we have here one of the
newest developments of Mendelism, it seems to me that it may be
worth while not to let Professor Castle's criticism pass un-
challenged.

Dexter pointed out that the mode of treatment that Nabours
followed in the analysis of his results is the procedure of multiple
allelomorphism, although Nabours does not seem entirely con-
versant with the fact, but treats the results as though they were
regular phenomena. In one case, however, Nabours got an un-
conformable individual. Dexter points out that if this case is
not due to non-disjunction (a known process that will cover
such cases) it shows that here at least the factors involved are
not allelomorphs, but must be treated as though closely linked.

How could the matter be put more directly? I confess I am
at somewhat of a loss to discover why Professor Castle is con-
fused. Perhaps it is the subsequent development of Dexter 's
explanation that has troubled him. Let us again try to make the
distinction clear.

If the factors B and E are not allelomorphic to each other then
each must have another allelomorph. This is nothing but pure
Mendelism, which no one will, I suppose, dispute. It is entirely
irrelevant whether we use small letters or none at all (as Castle
prefers) for the allelomorphs. If they are a part of the Men-
delian machinery, who cares very much what we call them?

If then we have here two pairs of allelomorphs, crossing over
may take place, as it does in other cases where two pairs of linked

1 The Ameeican Naturalist, June, 1914, p. 383.

Digitized by

Google

No. 572] SHORTER ARTICLES AND DISCUSSION 503

genes are involved.^ This is all there is to the matter. We need
not dwell, therefore, at length on Professor Castle's statement
that here is another case of an erroneous conclusion reached in
consequence of using small letters for ** absent" characters,
except to remark that Dexter did not use small letters for absent
characters, and that the erroneous conclusion has been drawn by
Professor Castle himself.

T. H. Morgan
Columbia Uniyxbsity

Professor Morgan has called my attention to the fact that in
criticizing a single point in Mr. Dexter 's review I have given the
impression, to some at least, that I regarded Dexter 's views as
erroneous. Such was not my intention, and I wish to correct the
impression, if I may. I do not for a moment question the reality
of '* unit-character" inheritance or indorse the idea of **the or-
ganism as a whole" as the only inheritance unit. I agree here en-
tirely with the view which I understand Dexter to hold. If Na-
bours has encountered nothing but simple allelomorphs among his
grasshoppers (which I neither assert nor deny), this by no means
proves that only simple allelomorphs exist even among said grass-
hoppers. An organism which seems to have only one variable
''gene" may nevertheless possess any number of other genes
which are not varying so far as we can discover, and in which con-
sequently all zygotes are homozygous and all gametes similar to
each other.

It is only in Dexter 's discussion of the significance of the ex-
ceptional ^^B E 7" individual that I should dissent from any part
of his excellent review. Nabours's explanation of this case, ac-
cording to Dexter, is essentially that of ** non-disjunction," in-
stead of which Dexter himself offers the explanation of ** link-
age," and proposes a repetition of the experiment to decide
between them. Now I do not question for a moment the genuine-
ness of either "non-disjunction" or "linkage," as they occur for
example in Drosophila. Through the kindness of Professor Mor-
gan I have been able to demonstrate both these phenomena re-
peatedly to classes in genetics in the course of their laboratory
work upon Drosophila. The point which I wished to make in com-

2 CroBBing over would not take place if the factors in question were aUelo-
morphic. If the case is one of non -disjunction the subsequent generation
would also give a different kind of result from that of linkage. (See
Bridges, Jour, Exp, Zool, 1913.)

Digitized by

Google

504 THE AMERICAN NATURALIST [Vol. XLVIO

menting on Dexter 's review (and this is the only point in which
I dissent from his opinions) is that the repetition of the experi-
ment, provided it had the outcome suggested by Dexter, would
leave us as much in the dark as we were before concerning the
correct interpretation of the result. Very likely, however, addi-
tional facts might be observed which would give some clue, so that
I quite agree with Dexter 's suggestion that the case should receive
further study. But I can not see that at present linkage has
more in its favor as an interpretation than non-disjunction.

The ** demonstration" which Mr. Dexter gave of his argument
by introducing duplicate ** symbols" instead of the single set used
by Nabours, seemed to me quite superfluous and possibly to have
been a real stumbling block in the logical process. This is why I
raised the question as to the significance of the small letters. The
terminology is that of the ** presence-absence" hypothesis, as
commonly understood, but Professor Morgan assures me that
such is not the significance which Dexter attaches to the symbols
used. It seems to me therefore that the significance attached to
the symbols is vital to the argument in the ** demonstration."

I quite agree with Professor Morgan, however, that symbols
are a matter of small consequence. Suppose we omit the ** dem-
onstration" by means of symbols altogether. Should we then
have any reason to favor linkage as an interpretation rather than
non-disjunction ? I can not see that we should have. It seems to
me quite possible that neither explanation will prove adequate.

When albino mammals are crossed with cplored ones, piebalds
sometimes are obtained in later generations. So far as we know,
these result neither from ** non-disjunction" nor from ** cross-
overs." Perhaps the 5 ^ 7 individual also is a tertium qmd.

W. E. Casti^

Digitized by

Google

NOTES AND LITERATURE

BIOMETRICS

An Important Contribution to Statistical Theory
One of Pearson's most valuable contributions to statistical
theory is his test for goodness of fit^ It enables one, with the
aid of Elderton V tables, easily to determine the pirobability that
a given system of observed frequencies does or does not differ
significantly from a series of theoretical frequencies supposed to
graduate the observations. The significance of this criterion in
Mendelian woA has recently been pointed out by Harris.*

Hitherto this criterion has found an important limitation in
the fact that, as originally developed by Pearson, it was appli-
cable only to frequency systems. It could be used to test good-
ness of fit only where the observations were counts of the number
of times particular classes of events occurred. But, of course,
frequency systems comprise only one kind of observational data
to which one has occasion to fit curves. Much more often there
is need for a criterion of goodness of fit where the observations
are of the nature of true ordinates, rather than frequencies.
Such cases include all data of the sort where a mean y is deter-
mined for each x, as in a growth curve; or in the regression
observed in a correlation table, where for each successive value
of one of the variables the mean value of the correlated variable
is calculated. There has been no method of testing the good-
ness of fit for such curves. From a visual inspection of the
plotted regression line one has been compelled to form his judg-
ment as to whether it was or was not a good fit.
Recently a Russian statistician, E. Slutsky,* has extended

1 Pearson, K., * * On the Criterion that a Given System of Deviation from
the Probable in the Case of a Correlated System of Variables is Such that
it Can be Beasonably Supposed to Have Arisen from Bandom Sampling,"
PW. Mag., 5th Series, Vol. L, pp. 157-175, 1900.

2 Biometrika, Vol. I, pp. 155-163.

• Harris, J. A., **A Simple Test of the Goodness of Fit of Mendelian
Ratios,'' Ameb. Nat., Vol. 46, 1912, pp. 741-745, 1912.

* Slutsky, E., * ' On the Criterion of Goodness of Fit of the Regression Lines
and on the Best Method of Fitting Them to the Data, ' ' Jour, Boy, Stat, Soc,,
Vol. LXXVII, Part I (December, 1913), issued 1914, pp. 78-84.

605

Digitized by

Google

506 THE AMERICAN NATURALIST [Vol. XLVIH

Pearson's theory to cover the class of curves, formerly not
amenable to such test. The result forms an extremely valiiabl^
extension of biometric theory.

Briefly Slutsky's essential result may be put as follows. "H^
finds (the complete proof is not given in this paper) that

<t)-

where x^ is the quantity denoted by the same letter in Peaxrsorx
original work, and is the argument in Elderton's table; 'W.^^
the frequency in the x^ SLTraj, i. e,, the number of observa^Liox
on which each observed ordinate is based; ep is the differexic
between the observed and the calculated mean y for eacb ^
array ; and a^,^ is the standard deviation of each Xp array ; ». <
the standard deviation of the group of observations from iwrliic
each particular y was calculated. 8, as usual, denotes suxzu

tion. Knowing x*, P is read directly from Elderton's tables.

Slutsky gives a couple of examples of the application of "tb
method in his paper. For illustration here I have preferr^^ 1
take an example from my own unpublished data. The oh^^ir^^^
tions (t/^p) in this case are the mean butter productions o
American Jersey cattle, based on seven-day tests.®

The theoretical points Y^^ are calculated from the equatioi
y = 14.21098 + .02503? — .0038a;» + 3.0104 log x,

the constants of which were determined from the ob8erva.t:ioxj
by the method of least squares.

The test for goodness of fit is carried out in Table I. It slioixl
be said that, following the suggestion given by Slutsky in fai
paper, I have used in the a«^^ column the graduated rather -tytxs.
the observed values. In the present case the scedastic curv^^ i
hopelessly far from a straight line. It is, in point of f ^d
logarithmic.

From this table we have x* = 32.115. This is beyond the rsux^
of Elderton's table. By a rough, but sufficiently accurate, gr«i.l>li
ical extrapolation, I find for present values of n' and x*»

P = .417 about.

In other words, if the butter production of Jersey cows chan-g^e
with age according to the curve given, we should expect; *

«For data see *' Jersey Sixes and Their Tested Daughters," publish©^ ^^
American Jersey Cattle Club, New York, 1909.

Digitized by

Google

No. 572]

NOTES AND LITERATURE

507

get a worse agreement between observation and theory in 42 out
of every 100 random samples on which the point was tested. In
other words, the fit may be considered suflSciently good. As a
matter of fact, the fit is extraordinarily close over most of the
curve. Four (only) out of the 32 ordinates contribute more than
50 per cent, of the value of x*-

TABLE I

^elJ?

Obserred

Calc. Batter

SUndard

Bntter Prodoc-

Prodaetion

Errors

Frequency

DeT. of

iioD in Lb«.

inLb*.

Arrays

^-p

^xp

K-I^x,)

'-X,

1.25

14.26

14.23

.02

.04

.500

1.76

15.15

16.16

.00

.97

.000

2.25

16.57

16.69

.12

273

1.49

1.771

2.75

16.96

16.06

.10

312

1.83

.932

3.25

16.38

16.35

.03

645

2.07

.114

3.76

16.72

16.57

.15

611

2.25

2.271

4.25

16.92

16.74

.18

704

2.38

4.027

4.75

17.09

16.89

.20

632

2.49

3.432

5.26

17.01

17.00

.01

656

2.66

.008

5.76

17.07

17.09

.OJi

382

2.62

.022

\6.25

16.98

17.16

.18

419

2.66

1.933

6.75

17.04

17.21

.17

277

2.68

1.114

7.26

17.09

17.25

.16

286

2.68

1.016

7.76

17.48

17.27

.21

190

2.68

1.167

8.25

17.30

17.28

.02

166

2.67

.009

8.75

17.17

17.27

.10

121

2.64

.174

9.25

17.66

17.26

.31

109

2.61

1.616

9.75

16.67

17.21

.54

2.57

4.194

10.25

17.06

17.17

.12

2.62

.143

10.75

17.42

17.11

.31

2.46

.619

11.25

16.95

17.06

.10

2.40

.094

11.76

17.00

16.97

.03

2.33

.005

12.26

17.05

16.88

.17

2.26

.113

12.76

16.54

16.79

.25

2.18

.092

13.25

16.34

16.68

.34

2.09

.291

13.75

18.14

16.66

1.68

1.99

6.673

14.26

15.89

16.44

.66

1.88

.599

14.76

16.16

16.30

.16

1.77

.036

15.25

16.37

16.16

.21

1.65

.065

16.75

16.76

16.00

.26

1.53

.063

16.25

16.42

16.84

.42

1.40

.117

16.75

16.76

15.67

.08

1.27

.016

Totals . . .

6,781 1

32.115

It may be said, in conclusion, that Slutsky's contribution is
one which will be highly valued by all investigators who have a
critical interest in the graduation of observational data, whatever
the field in which they may be working.

Raymond Pearl

Digitized by

Google

08 THE AMERICAN NATURALIST [Vol. XLVXIX

A NEW MODE OF SEGREGATION IN GREGORY'S
TETRAPLOID PRIMULAS

In a recent paper^ Gregory reports a very interesting cctse i
which two different races of Primulas suddenly gave rise to ^Isu
tetraploid forms, having double the usual number of ctix-oxri
somes, and apparently having the factors doubled also (indi^v^d
ally), for this was true of all the factors which could be foUo^^w-
in his hybridization experiments. It is important to kno^'v 1x4
segregation will take place in such individuals, as there ar^ f c
allelomorphs of each gene present.

Let us suppose that a tetraploid jform pure for the doan-ins

(A A\
and therefore of composition -r t) ^ crossed ^^^itli

pure recessive giant f--j. Gametes AA and aa will m^^±^

A A

fertilization, forming the hybrid — — (the maternally de^x^^v

a a

genes are represented on one line, say the upper, the patem::^^^^
derived genes on the other line).

Now, if this were an ordinary case of * 'multiple factors" * ii^
diploid organism, although the two dominant factors, whi(3ti. ^
may again call A's, may produce the same effect upon the or'g'a.
ism, yet they arei not interchangeable, and the same is true o:f "t^
recessive factors. That is, if we call both dominants A, we xxxti
designate one of them as A^, and the corresponding recesMsi'V"
must also be designated as a and a^, for A will always &egr^^S^
into a different gamete from a, and A^ from a^, there beings 't^
distinct allelomorphic pairs. On the chromosome view o:f h
redity, we would say that A and a always lay opposed 1
each other, in homologous chromosomes, on the spindle of "th
reduction division, as did also A^ and a^, but neither A nor
lay in chromosomes homologous to those of either A^ oxr a-
and assorted independently of them. The line-up of factors o
the spindle in the reduction division in this case would t"

AA^ A a^

equally likely to be 1 or — v-^, depending merely upon ^wlxic

a a a a.

iR. P. Gregory, '*On the Genetics of Tetraploid Plants in Trii"^'^^
sinensis/' Proceedings of the Boyal Sooiety, 1914.

2 i, e., a case where two (or mare) independent pairs of factors pro^^*^'
similar effects, upon the same character. Many examples of this are
€. g.j the inheritance of red flower in flax.

Digitized by

Google

No. 572] NOTES AND LITERATURE 509

way the pairs are turned with reference to each other. The first
alignment gives gametes AA^ and aa^, the second gives Aa^ and
aA^ Thus three gametes with a dominant factor to one pure
recessive would on the average be produced, the ratio being
lA A : 2Aa : laa, omitting primes.
In a tetraploid form, however, A and A^ are alike and inter-

A A^

changeable, as also are a and a^. In the hybrid j-, therefore,

a a

there would be at least one other mode of pairing of allelomorphs
possible, giving two new modes of line-up on the reduction

spindle, and they would occur just as frequently as the two

A A^
previous kinds. The two new arrangements would be -^ — ,

a a

giving gametes AA^ and a^a, like those in the first of the two

A a
previous cases, and — jj, giving gametes Aa and a^A^^ These
a A.

latter gametes would be indistinguishable from the Aa^ and aA^
gametes given by the second of the two usual arrangements un-
less A could be distinguished from A^ and a from a^. This could
happen only if the allelomorphs were of four different kinds or
if there were linkage of these genes with other genes for which
the plant was heterozygous. Unless, therefore, linkage or mul-
tiple allelomorphism were involved, we could not distinguish
between this mode of pairing of allelomorphs and the usual kind ;
both would give three gametes containing at least one dominant,
to one pure recessive (i. e,, lAA:2Aa:laa, omitting primes).

Still a third type of pairing of allelomorphs is possible in a
tetraploid plant, however. There seems no a priori reason, on
the chromosome view, why, in a tetraploid plant, a gene should
have to segregate from one of the -allelomorphs derived from the

A A*

opposite parent. That is, in a plant of composition ^ ,

a a

paternally derived genes being indicated on, say, the upper line,

maternally derived ones on the lower, there is no apparent reason

A a
why the line-up of chromosomes at reduction should not ^^TiTT

A a^

-r, — or as often as it is one of the other types, since all four

A^ a

chromosomes are homologous. Thus we should get gametes Aa,
A^aS Aa^ and A^a.

8 If linkage with other genes could be followed, we should with this mode
of pairing obtain crossing over between the chromosomes containing A and
a*, respectively, and between those containing A* and a, respectively; this
would not occur on any other mode of pairing.

Digitized by

Google

510 THE AMERICAN NATURALIST [VoL-XLVHI

We could distinguish such gametes individually from those
obtained by the ordinary arrangements only if linkage were in-
volved, for then we should sometimes obtain results indicating
that the chromosomes containing A and A^ had crossed over
with one another, and so had probably behaved as homologous
chromosomes at the reduction division. However, we could also
determine whether this mode of pairing occurred or not merely
by determining the relative numbers of the different kinds of
gametes formed. For, if the third type of pairing occurred, we
should obtain 4Aa gametes in addition to the 2AA, 4Aa and 2aa
derived from the other two types of pairing. The ratio of
gametes would then be five containing a dominant to one reces-
sive, there being 1 AA : 4Aa : laa, as opposed to the ratio
lAA : 2Aa : laa obtainable on either of the other modes of segre-
gation.* The latter or more usual ratio is the only one considered
by Gregory, who apparently takes it for granted that in so far
segregation must be of the same sort as in diploid forms.

Let us see which ratio is more in accord with his experimental
data. As the ratio of oflfspring in a back-cross is the same as the

A A

gametic ratio, it will be seen that a back-cross of — — by a reces-
sive should give 3 A : la plant on Gregory's view, the 3A's consist-

A A A a

ing of 1 : 2 . On the other view, a back-cross should result

® a a a a '

A A Aa

in 5A : la, the 5 A 's consisting of 1 : 4 — . On inbreeding an

a a a a

plant, however, owing to the random fertilization of

a a

gametes, Gregory's 3:1 gametic ratio would result in a 15:1
ratio among the offspring (which correspond to Pj) and our
own 5 : 1 gametic ratio would give a 35 : 1 ratio of A to a among
the oflEspring.
A summary of his back-crosses of Pj heterozygous thrum-eyed

plants of the type to recessive pin-eyed plants - - gives

a a a a

the result 61 thrum: 6 pin (10:1, as compared to the two ex-
pectations 3:1 and 5:1). Among the F^ thrums there should

^Counts of cbromosomee in the maturation divisions of the tetraploid
plants show that the chromosomes sjnapsed in pairs, not in groups of four.
Synapsis in fours would be, in effect, the same as pairing of the random sort
suggested in this paper, so far as any one set of allelomorphs are concerned,
but it might give different linkage results.

Digitized by

Google

No. 572]

I^OTES AND LITERATURE

511

on Gregory's view be 1 :2 ,on the other view 1 :4 — .

aaaa' aaaa

Tests of twenty-one Pj thrums, by mating them to themselves

and also to recessives, showed that there was only one which was

A A Aa

certainly and 15 which must have been — . (A few gave

a a a a

numbers too small to be significant, and one or two were of doubt-
ful composition.) This result is within the limits of probable error
on the 4 : 1, but hardly on the 2:1 expectation. The one Pi

thrum plant which was of composition — — gave, on back-

a a

crossing, 67 thrums : 18 pins, a ratio of 3.7 : 1, to correspond with

Gregory's 3:1 or my 5:1 expectation. On inbreeding it gave

44 thrums : 2 pins, a ratio of 22 : 1, to correspond with Gregory's

15 : 1 or my 35 : 1 expectation. The other Pj thrums, being of

Aa "

composition (aside from the few doubtful ones), gave, on

a a

the average, 1 thrum: 1 pin on back-crossing, and 3 thrums : 1 pin

on inbreeding; these results would be expected on either view.

Crosses were also made involving the character green versus

A A

red stigma (green being dominant). Here the forms, on

a a

back-crossing, gave a total of 114 green : 30 red (3.8 : 1 instead of

3:1 or 5:1), and on inbreeding they gave 75 green: 2 red

(37.5:1 instead of 15:1, as on Gregory's expectation, or 35:1,

on my own).

It will be seen that the numbers in the above crosses are too

small to be very significant, individually, for a settlement of the

question at issue, but if summed up they become more decisive.

Thus, a summary of the offspring of all back-crosses of the

form to the recessive gives 242 dominants ( A) : 54 reces-

a a

sives (a), or 4.5:1, as compared with the 3:1 expectation of

Gregory and the 5:1 of the view advocated in this paper.

"Where the dominants among these oflBspring were tested they

Aa A A

were found to consist of 1 and 15— — , as compared with the

a a a a' ^

1 : 2 expectation of Gregory, and ours of 1:4. Pinally, a sum-

A A

mary of the cases where forms were inbred shows that 119

•^ a a

dominants: 4 recessives resulted, a ratio of 30:1 where Greg-

Digitized by

Google

512 THE AMERICAN NATURALIST [Vol. XLVTEI

ory's expectation would be 15:1 and our own 35:1. Moreover,
the individual records fluctuate in both directions about the
ratios to be expected upon our point of view, but practically all
vary in the same direction from the expectation of Gregory,
namely, in the direction of the other expectation.

There is reason, then, to believe that in these Primvlds the
factors derived from the same parents may segregate from each
other as allelomorphs, while allelomorphs derived from opposite
parents meanwhile assort at random. For although the allelo-
morphs exist in sets of four they must pair two by two for segre-
gation, as do the chromosomes, and two derived from the same
parent may happen to pair with oniB another. The chance that
this should occur is one third, since there are three possible
modes of pairing. Such a result is difficult to explain except on
the chromosome view of heredity. It would give ratios different
from those theoretically expected by Gregory, but more in
accord with his expqrimental data. The principle upon which
our own expectation is founded may be briefly summed up by
saying that where more than two factors which are normally
allelomorphic to each other are present, the pairing of these
allelomorphs with each other preparatory to segregation usually*
takes place at random.

Hermann J. Mulleb

8 That this is not always true is shown by Bridges ' case of * * non-dis-
junctional" females of Drosophila, which contain one Y and two X chromo-
somes. Any two of these chromosomes normally act as homologuee to each
other in the reduction division of the normal fly, which contains only two
of them. But where all three are present together they do not pair at ran-
dom, for they oftener undergo the segregation X-XY than XX-Y, pre-
sumably because the two X's are much more like each other than like the
Y, and so more apt to act as homologuee.

Digitized by

Google

wo Important Biolosfical Journals

The Journal of Qeneticfi

Edited by
W. BATESON and R C. PUNNETT
A periodioal for the pabliofttion of reoords of original researoh in heredity, yariation, and allied topiof .
JouBVAL also contains, from time to time, articuee rammarizinfj: the ezumng state of knowledge in the
>iisbranohe8 of genetios. The numbers are fully illustrated, and, when the subjeot-matter demands it,
«d plates are freely empl<^ed. The Journal is inued in parts as material aoonmnlatea, and a volume
earing, so far as possible, annually) consists of tour such parts. Subscription price, $7.50aTolume;
e oopies, 12.60.

3iometrik:a

Edited by KABL PEABSON
k journal for the statislieal study of biological problems, appearing about four times a year. A Tolume
lining about 500 pages, with pitkeB and tables, is issued annually. Volume X begins in 1914. The
preceding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
ometry, medicine, heredityy etc. Subscription price, $7.60 a volume ; single copies, 12.60.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes
dnglc copies.

iflbed in America for the Cambridge University Press x»f England by

Th© University of Ohicasro Rress
Ctkloago^ Illinois

University Control

By J. MoKbbit CAXTSLLy Prof eesor of Psychology in Oolambia XTniTenity

rogether with a series of Two Hundred and mnety-nine Undgned Letters by Tiiading Men of I
Dg Aeadeodo Positicnis and Articles by Josvh Jastbow, Obosob T. Ladd, Jorv J. Braf— so», J. B.
9BToir, J. lioEBBH Gattbuh Obqbob M. SrmAnoirp BrmwAsat Pavov, Jobs Jat CKAmASt Jakw P.
BOB and Jacob Gould Schubkar.

L great Tariety of qoestions ocmeerning general iiniiF«nity administration are dealt with in an original aad
111 way.— ^oliiri.

rbese quotations and examples are taken from PkofesMr Catteil's inf<«ned and thorough disousrion of the
ot of nniyenity eontrol, a subject upon whioh he has had muoh to say of late, finding oeearion for eaustfs
ism of earisting American oonditicms, and standing as the ehampion of an aeademie demoeiaey and a teaeb*
rofisfon upon whioh a man may enter without forfeiting his self^espeot.— Tfts DiaL
ientenoes and paragraphs that betoken the expert, highly-trained mind, the suggestions that oome to Be-
ad tell its that a new day is about to dawn in educational writing.— 7!fcs BoffoB Swmnimg Tfom^eripi.

SCIBINCe AND EDUCATION

^riem of Tolumeafor the promotion ofseientifio reaetweb and educational progremB

[JMB L The Foundations of Science. By H. PonroABi. Containing the auth<»iaed English translatloB
ly Obobob Bboob Hau»bd of "Soienoe and Hypothesis," "The Value of Soienoe," and *'Brienea
ind Method." Price, $3.00 net.

[JME IL Medicgl Research and Sdncatlon. By Bxobabd M. F^iabob, Wiiaum H. Wblob, W. H.
lowBUi, FaAjnodH P. Maui, Lbwblltb F. Baw¥Bb, Chabubs 8. Mznot, W. B. Gamnoh, W. T.
'youmaHMMM, Tbbobalo Smith, Q. N. Stbwabt, C. M. Jaoksok, E. P. Ltok, Jambs B. HsBsiaK, Jens
tf. Donsow, C. R. Babdbbv, W. C)PBOLB,a J. BIbutibb, Jambs Ewnro. W. W. Ebbv, Hbbst H. Dobaia-
K>B, CHBiBnAB A. HBBfBB, and Hbbbt p. BownrroB. Price, $3.00 net.
[JMEm. UaiTsnlty ContreL NawBmOy, Paom»+484* Pr<e$.9SMfMk

tR150N, N. Y.

THE SCIENCE PRESS

SUB-STATION 84» NBW YORK CITY

LANCASTBRt PA.

Digitized by

Google i j

The American Naturalist

CONTENTS OF THE FEBRUARY NUMBER

Some New Varieties of Rftig mad Ghiii)e»-pif i and tbeir

BelatioDf to Probleim of Color Inheritance. Pio-

feMorW.E. Castle.
• Dominant" and '* BecessiTe" Spotting In Mice. C.

C. Little.
On Differential Mortality with respect to Seed Weight

occurring in Field Cnltnres of Fismn saUvum.

Dr. J. Arthur Harris.
The Inheritance of a Recurring Somatic Variation

in Variegated Ears of Maise. Professor R. A.

Bmerson.

Bestoration of Edaphosaums crudger Cope. Pro-

fessor E. C. Case.
Shorter Articles and Discussion : Humidity— a

Neglected Factor la Enylronmental Work. Dr.

Frank £.Luts.

CONTENTS OF THE MARCH NUMBER

The Effect of Extent of Distribotton on SpeelaftlaQ.

Asa C. Chandler.
Biology of the Thysanoptera. Dr.A. B^ankllnSbiili.
Shorter Articles and Correspondence : The Endemie

Mammals of the British Isles. Professor T. D. A.

Cookezell.

Notes and Literature Swingle on Variation in W^
Citms Hybnds and the Theory of Zygotazlsu
Dr. Orland £. White

CONTENTS OF THE APRIL NUMBER

The Origin of X Capsella Bursa pastorls araohnoidea.'
Dr Henri Hus.

Biology of the Thysanoptera. U. Dr. A. Franklin
ShuU.

Shorter Articlei and Discussion : Barriers as to Dis-
tribution as regards Birds and Mammals. Joseph
Grinnell. Yellow Varieties of Rats. Professor
W.B. Castle.

ofMonarchs. V. L. K.

CONTENTS OF THE MAY NUMBER

Kellogg.
Regeneration, Variation and Correlatton In Thyooe,

Proiessor John W. Scott.
Shorter Articles and Discussion : Terns relating to

Generic Types. Dr. O.F.Cook,

Moth. A. H. Stnrteyanfc. Nabonrs's Breading
BzperlmentBWlthGraasboppen. JohA&Dextec

CONTENTS OF THE JUNE NUMBER

Spedes-hoilding by HybridizaUon and Mutation. Pro-
fessor John H. Gexoold.

A Study of Factors goremingDistribuUon. Phineas

W. Whiting.
Physiological Correlations and Climatic Reactions in

Alfalflft Breeding. Geo. F. Freeman.
Taxonomy and Erolution. By X.
Shorter Articles and Discussion* Nabonrs's Graas-

hoppeis, Multiple Allelomorphism, Tankage and

Misleading Terminologies in Genetics. Professor

W.E.Ca8Ue.

CONTENTS OF THE JULY NUMBER

Pattern Dereiopment In Mammals and Birds.
QloTer M Allen

Arthur G.VestaL
Shorter Articles and Dlscnssioii- Another Hy-
pothesis to Account for Dx Swingle's Szpcri*
ments with Cltnu. A. C. and A. L. Hagedoonv

Single Number 40 Cents Yearly SuUcription, $4^00

THE SCIENCE PRESS

Garrtton,N.Y. Sub-Station 84 : NEW YORK Lancastert Pb.

Digitized by

Google

▼OL. ZLVm, VO. 673 SEPTElfBES. 1914

THE

AMERICAN
NATURALIST

A MOHTELT JOXnUTAL

BtToted to the Adyanoement of the Biologioal Soienoos wltt

Special Boferenca to the Factors of Brolntioii

OOHTEHTS

Pag€

I. StadiM on Inbreodlng. Dr. Raymond Pbabl - 618 '

n. THe Ohroiiiosome HypotlMili of Linkage appUod to Oases In Sweet Peas and

Frlmnla. Calyiw B. Bbidgbs 624 j

m. THe Reduplication HTPOthesls as applied to Drosoplilla. Dr. A. H.

Stubtevant -------- 686 !

IV. Pattern DeTeiopment in Iffaniinals and Birds. Dr. Qloyeb M. Allbn - 660 ;

▼. Shorter Arttdes and Oorrespondence : The Bearing of the Selection Experi. j

ments of Caatle and PhiUipi on the Variability of Qenes. Hbbmann J. I

MuiJiKB ------------- 507

THE BOIENOE PBBBB

I
IJUraASTEB, PA. OABBI80H. V. f . |

HEW TOBK: SDB-BTATIOH 84

I
; j

Digitized by

Google

The American Naturalist

MSS Intended for publication and iMNiko, etc., Intended for review etionld be
eent to the Editor of THE AMERICAN NATURALIST, Garrleon-on-Hudeon, New York.

Short artiolee containing eunmariee of reeearch work bearing on the
probleme of organic evolution are eepecialiy weloenie, and will be given preference
in publication.

One hundreo reprinte of contributione are eupplied to authore free ef charge.

Further reprinte will be eupplied at coet.

Subecriptione and aovertieeniente ehould be eent to the pobllehere.

The

eubecrlption price ie four doliare a year. Foreign poetage le fifty cents and
Canadian poetage twenty-five cents additional. Tne charge for single copies is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Garrison, N. Y.

NEW YORK : Sub-Station 84

Entered m Beoond-cUn matter, April 2, 1908, mt the Poit Offlee at Leacaiter, Pa., under the Aet of

Congreif of Mareh 8, 1870.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prioes

Partioulars of

G. DINESEN, Bird Collector

Husavllc, Nortli Iceland, Via Leidie, England

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audubon.
7or 8 volumes, please report, stating cash price, stat-
ing condition, binding and dates of volumes.

P. C. HARRIS,
Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds'
Eggs, containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste
viridanus), Red-browed Finch (Calla-
canthis burtoni), Tibetan Sand-Grouse
(Syrrhaptes tibetanus). Spotted Sand-
Grouse (Pteroclurus senegallus). Ibis-
bill (Ibidorhynchus struthersi), White-
bellied Minivet (Pericrocotus erythro-
pygius), Bearded Vulture (Gypaetus
barbatus), numerous Cuckoos with their
hosts, etc., etc.

Particulars and full list may be had
from

W. F. H. ROSENBERG

57 Haverttock HiU London, N. W.

Marine Biological Laboratory

Woods Hole, Mass.

FMOities foe rMMTch is Zookflr.
Bmbiyologj, Pliyaiolocy sad Bs^
snj. 8ev«nty-«iz privsta labor*-
toneiL $100 eaoh for not orw time
months. Thirty tabfes are araS-
abla for basumera in maaaroh wka
desire to work under the dfaaetioa
of membera of the etaff. The Im
f or SQoh a table if SfiOOa

Courses of labocatoty inshuetfos
with leotoree are offend In Inverte-

INVESTIGATION
Eadra Yaar

INSTRUCTION

JolrlloAvgaatll,

1S14

BcolofyoC]_

Vegetatioa. Eaoh eoursa iagidm

the fan time of the stadent. Tta,

^~1>X)0. A loetore oourae on the

-11 Aspects o( HolocF

Sciences Is also offecedi

SUPPLY -

DEPARTMENT p««^ed matoriaTcrf aU troei of
Op«B iIm Entirs Y«ar animala and of Algae, FanAiiver-
worta and Moeaea fumwied for
daaiwork, or for ths aansaom.
living material f undahad in aeaaoo
aa ordered. Price Hats of So5-
logioal snd BotanSeal sMtarial
eent on appfioation. State wUch
is deeired. For price 1 sta and all
information regarding material,
sddreaa

QEO. M. GRAY, Curstor. Woods Hole, Maa

The annual annooneement wQI be eent on appBestian to

Tte Director, Marine Blologicsl Lsbocstoqr, Woods Hele

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVIII September, 19 U No. 573

STUDIES ON INBREEDING. V

Inbreeding and Relationship Coefficients ^

Dr. RAYMOND PEARL
University of Maine

In the discussion of inbreeding coefficients contained in
a series of recent papers from this laboratory^ no mention
has been made of an important consideration which arises
in connection with such coefficients. The further problem,
to which we may now turn, may be stated in the follow-
ing way.

The pedigree of an individual consists of two halves.
One of these halves is made up of the sire and his an-
cestors ; the other of the dam and her ancestors. Follow-
ing the conception of inbreeding set forth in detail in the
earlier papers of this series it is plain that the values of
the coefficients of inbreeding for a particular pedigree are
composed of the following elements.

1. The occurrence of the same individual animals more
than once on the sire's side of the pedigree only.

2. The occurrence of the same individual animals more
than once on the dam's side of the pedigree only.

1 Papers from the Biological Laboratory of the Maine Agricultural Ex-
periment Station, No. 69.

« Pearl, R., ** Studies on Inbreeding. I. A Contribution Towards an
Analysis of the Problem of Inbreeding," Ame«. Nat., Vol. XL VII, pp.
577-614, 1913; "The Measurement of the Intenwty of Inbreeding," Me.
Agr. Expi. Sta. Bui, 215, pp. 123-138, 1913. Pearl, R., and Miner, J. R.,
"Studies on Inbreeding. III. Tables for Calculating Coefficients of In-
breeding," Me. Agr. Expt Sta. Ann. Rept. for 1913, pp. 191-202, 1913.

513

Digitized by

Google

514 TRE AMEBIC AN NATURALIST [Vol. XLVIII

3. The reappearance of animals which appear first on
one side of the pedigree (either the sire's or the dam's)
on the other side.

If only 1 and 2 are to be found in the pedigree it means
that the sire and the dam are totally unrelated (within the
limits covered by the pedigree in the particular case). On
the other hand, the occurrence of 3 means that sire and
dam are in some degree related, and that a portion of the
observed inbreeding arises because of that fact. Now
the coefficients of inbreeding, in and of themselves, tell
nothing about what proportionate part has been played
by these three elements in reaching the final result. It is
a matter of great importance to have information on this
point, because of its genetic significance. It is the pur-
pose of this paper to describe a general method for ob-
taining this desired information.

The first step in the method, stated briefly, is to break
up the pedigree elimination table formed to get the suc-
cessive values of pn+i — gn+i, in our former notation, into
four different parts. One of these parts will include the
primary reappearance on the sire's side of the pedigree of
such animals as appear first on the same side. This may
be called the ''male only" table. The second part will
include the primary reappearance on the dam's side of
such animals as first appear on the same side. This is
the ''female only" table. The third part will include the
primary reappearance on the dam's side of such animals
as first appear on the sire's side. The fourth part is the
reverse of the third. These last two may be called the
"cross tables." The sums of the totals of these partial
tables will give the total pn+i — gn+i values for the succes-
sive generations.

The formation of the tables on this plan may be illus-
trated with some examples. These examples will also
show the skeleton method of writing pedigree elimination
tables, which saves much labor. This was referred to,
but not significantly illustrated, in the earlier papers. It
consists simply in doubling the total of the column for
each generation rather than the separate items.

Digitized by

Google

No. 573]

STUDIES ON INBREEDING

515

TABLE I

Paktial Pedigree Elimination Table fob King Melia Bioteb 14th Show-
ing THE Pbimaby Eeappearances on the Sire's Side op the
Pedigree of Animals which first Appear on that Side

Generation | 2 3

10 11

Melia Ann's Son

(2)»
1

Melia Ann 3d

(6)»
3
2
1

1
1
1
1

1 !""•■

Lucy's Stoke Pogis . . .

Melia Ann

St. Lambert Boy

. .[

Lettv Rioter

Allie of St. Lambert . . .

1 j

Lord Aylmer

..L...

Amelia 2d

(32)»
1
1

1 1

Victor Hugo

Oakland's Nora

1 1

Stoke Pogis 3d

1
1
1
1
1

bert

Sir George of St. Lam-
bert

i 1

Diana's Rioter

Orlofif

j 1

Ix)me

...J

Huiro's Victoria

....1--..

(82)»
1
1

1
1
7

1
4
2
1
1
1

::::::i::::::':::::'

Victor Hugo

. . 1 . .

Pauline

....

I..:::'.:::::

Canada's John Bull. . .

..1.--.

Oakland's Nora

1 '

Stoke Poids 3d

Kathleen of St. Lam-
bert

1 1

liOrd Lisgar

. . 1 . .

Lucy of St. Lambert. . .

1 1

Pet of St. Lambert . .

Orlofif

Bachelor of St. Lam-
bert

... 1 ' .

Ida of St. Lambert . .

(210)'
2
2
3

. i

Victor Huc^

Bt-nk^ Pogifl 3d . . . , , .

::i::i::::'::::':::::

. . 1. .

JjOTue

1 ■

' ::::

Amelia

'Y'"!""'

(438)»

Lord Lisgar

' 1

1 I

Pride of Windsor

Laval.

I i

Amelia

■:i::::':::::::::

: 2 1.

Victor Hugo

3 (894)»
1

Laval

::::*i::::::

Amelia

..../::::::! i

Lisette

1 , ^

L . . .- . J 1

Berthe.

1 1 ■.

^ 1 1

1 1

Totals

3 1 16 1 41

105

219 1 447 ■ 898

1.796

^ Li this and the following table the numbers in brackets are in each case
twice the sum of the numbers in the preceding column. They represent the
accumulated ancestral reduplication up to the generation in question.

Digitized by

Google

516

THE AMERICAN NATURALIST [Vol. XLVHI

The pedigree for 12 ancestral generations of the Jersey
bull King Melia Rioter 14th (103901) may be taken as the
first illustration.

TABLE II

Partial Pedigree Elimination Table for Kino Mella. Rioter 14th Show-
ing THE Primary Reappearances on the Dam's Side of the
Pedigree of Animals which first Appear on that Side

Generation 2 3 4 5 6 7 8 9 10 11 12

King's Rioter Lad — — — 1 2 4 8 16 32 64 128

Table III is clearly the one which demands special
attention. As will shortly appear, it is the most important
for the theory of inbreeding. Let us attempt its analy-
sis. Just what does the first entry mean genetically f It
states that King Melia Rioter, an animal which first ap-
peared on the sire's side of the pedigree, reappeared in
the second ancestral generation on the dam's side. What
this clearly means is that at least one half of all the dam's
ancestors, in the third and higher ancestral generations,
are identically the same animals as are ancestors of the

iOO

f=8=l

1
>

r/ ,

M^'

1,'

» M

f J*

C£JVCffA TlOftS

Fig. 1. Diagram showing (a) the total Inbreeding (heavy solid line) and (5)
the relationship (heavy broken line) curves for the Jersey bull, King Meila
Rioter 14th. The high order of the Inbreeding and relationship between the sire
and dam In this case Is evident by comparison with the lighter lines, which give
the maximum values for continued brother x sister, parent x offspring and cousin
X cousin breeding.

Digitized by

Google

No. 573]

STUDIES ON INBREEDING

:>r

sire. The next entry in Table III indicates that in the
fourth and higher ancestral generations at least 5/8 of
all the dam's ancestors were the same individual animals
as were also ancestors of the sire. One half of them were
the same before the reappearance of St. Lambert's Rioter
King. He makes up the additional 1/8 of the dam's
ancestry.

TABLE III

Partial Pkdigeee Elimination Table for King Melia Eioter 14th Show-
ing THE Primary Reappearances on the Dam's Side op the
Pedigree or Animals which first Appear on the Sire 's Side

Generation 12

3 1 4

5 6

9 10 : 11

King Melia Rioter 1

St. Lambert's Rioter Kins

(2) (4)

■y\\

(10)

1
1

■ 1

King of St. Lambert

St. Lambert Bov

(?4)

St. Lambert Boy 1 . .

2
1
1

....

Oakland's Nora. 1

St. Lambert's Rioter Kinar . . . . ! . .

(56)
1

1
1

St. Lambert Bov

King of St. Lambert

St. Lambert's Lettv . . .

' ■ ■ r ■ "

(118)

Lettv Coles 2d

. . . L__ _

(238) ...

King of St. Lambert

...1. ...!....

1 . ..

Liouise's Grace L .

.. .1 '

1 ...

Totals |l

12 1 28

119

240 4809601,920

From these tables it is obvious that a very considerable
portion of the inbreeding shown in the pedigree of King
Melia Rioter 14th arises from the fact that his sire and
dam were closely related. Furthermore, both sire and
dam are closely inbred in their own lines. The curve of
total inbreeding in this case is shown in Fig. 1, along with
the curves for continued brother X sister, parent by off-
spring, and cousin X cousin mating.

TABLE IV
Summarized Pedigree Elimination Table for Kino Melia Bioter 14th

Generation

cT only

1
5

105

119

219 1 447

898

960

1,796

9 only

16
240

480

128

Cro68-over

12* 28

1,920

Together

1 2

6 16 146 1104

232 1 475 959

1,922

3,844

From this we have, for the inbreeding coefficients,

Digitized by

Google

518 THE AMERICAN NATURALIST [Vol. XL\Tn

Zo = 0
Zx =25.00
Z, =25.00
Z, =37.50
Z4 =50.00
Z. =71.88
Ze =81.25
Zt =90.63
Z. =92.77
Z, =93.65
Zi»= 93.85
Zu= 93.85

These facts will possibly be made clearer to those not
actually working much with pedigrees by Table V, which
gives the first four ancestral generations* of the pedigree
of King Melia Rioter 14th.

Gteneralizing the above reasoning we get the following
result.

In A3, and higher ancestral generations, 2/4 = 50.00 per

cent, of the dam's ancestors are animals whidi are also

ancestors of the sire.
In A4, and higher ancestral generations, 5/8 = 62.50 per

cent, of the dam's ancestors are animals which are also

ancestors of the sire.
In Ag, and higher ancestral generations, 12/16 = 75.00 per

cent, of the dam's ancestors are animals which are also

ancestors of the sire.
In Aq, and higher ancestral generations, 28/32 = 87.50 per

cent, of the dam's ancestors are animals which are also

ancestors of the sire.
In A7, and higher ancestral generations, 59/64 = 92.19 per

cent, of the dam's ancestors are animals which are also

ancestors of the sire.

4 In tbe study of pedigrees stress is naturally laid on tbe ancestral genera-
tions, rather than on the filial, as in breeding experiments. It beoomes very
convenient to have a brief designation for ancestral generations, in the same
way that Fi, F,, etc., are used to denote filial generations. I would suggest
the use of the letter A with sub-numbers for this purpose. We then have Ai
denoting the parental generation. A, the grandparental, At the great-pa-
rental, etc.

Digitized by

Google

No. 573]

STUDIES ON INBREEDING

519

In Ag, and higher ancestral generations, 119/128 = 92.97
per cent, of the dam's ancestors are animals which are
also ancestors of the sire.

In A9, and higher ancestral generations, 240/256 = 93.75
per cent, of the dam's ancestors are animals which are
also ancestors of the sire.

In Aio, and higher ancestral generations, 93.75 per cent,
of the dam's ancestors are animals which are also an-
cestors of the sire.

In All, and higher ancestral generations, 93.75 per cent,
of the dam's ancestors are animals which are also an-
cestors of the sire.

In A12, and higher ancestral generations, 93.75 per cent,
of the dam's ancestors are animals which are also an-
cestors of the sire.

TABLE V
Pedigree fob Four Ancestral Generations or Kino Melia Bioter 14th

O O

No. 63200 &

Marjorie
Melia Ann's
Son.

No. 56581 c^
Melia Ann's King.

No. 22041 cT
Melia Ann's Son.

No. 100775 9
Lottie Melia Ann.

No. 157263 9

Marjorie MeUa
Ann.

No. 22041 c^
• Melia Ann's Son.

No. 905883 9
Mary Melia Ann.

No. 181544 9

No. 58169 cf
King of All Kings.

No. 54896 cf
St. Lambert's Rioter King.

Letty
Silver
Hair.

No. 114804 9
St. Lambert's Letty.

No. 148456 9

Exile's Silver
Hair.

No. 32559 cf

Exile of St. Anne's.
No. 60449 9

SUver Hair 4th.

No. 73104 &

#King Melia
Rioter.

No. 63200 cf

<g> Marjorie Melia
Ann's Son.

No. 56581 d*
<8) Melia Ann's Eling.

No. 157263 9
0 Marjorie Melia Ann.

No. 181544 9
<8) Letty Silver Hair.

No. 58169 d
<S) King of All Kings.

No. 148456 9
^ Exile's Silver Hair.

No. 219360 9

Dula
Riotrees
Maid.

No. 62098 cf

King Rioter's
Lad.

No. 54896 c^
# St. Lambert's Rioter King.

No. 142296 9
Eling's Riotress Nora.

No. 218796 9

St. Lambert's
Dula Riotress.

No. 57778 c^
St. Lambert's Boy.

No. 174761 9
Rioter Lad's First Daughter.

Digitized by

Google

520 THE AMERICAN NATURALIST [Vol. XLVIU

•
These percentages are quantities of a good deal of

interest. They measure the degree in which King Melia
Rioter 14th 's sire and dam were related to each other.
Community of ancestry is the basis of kinship.

Percentages derived in the way shown above, from
cross pedigree elimination tables, I propose to call co-
efficients of relationship, and to designate by the letter
K, with appropriate sub-numbers referring to the genera-
tion. These relationship coefficients are, with some limita-
tions, independent of the inbreeding coefficients in the
values they may take, though the two will usually be cor-
related to some degree. It is, however, possible to have a
high value of Z with K=0.

TABLE VI

Comparing the Maximum Possible Values or the Coefficients of In-
breeding (Z) when the Coefficient of Relationship K
Equals (a) Zero, and (ft) 100

Generation Maximum Possible Value Maximum Possible Value

otZyfhenK = 0 of Z when fi: = 100

Ai 0 0

A, 0 * 50.00

As 50.00 75.00

A^ 75.00 87.50

As 87.50 93.75

A, 93.75 96.88

A, 96.88 98.44

A, 98.44 99.22

A, 99.22 99.61

A,o 99.61 99.80

The most important feature of the relationship coeffi-
cients is found in their genetic implications. This can be
indicated best by an illustration. Let us consider the case
of the maximum possible degree of inbreeding with K = 0.
This will be found when the sire and the dam are each
inbred to the highest possible degree (continued brother
X sister mating) but are in no way related to each other.
Such a case would be afforded, for example, if a Jersey
bull, the product of continued brother X sister mating,
was bred to a Hoi stein cow, which was also the product

Digitized by

Google

No. 573] STUDIES ON INBREEDING 521

of a continued brother by sister breeding. Clearly K
would be 0, since no animal on one half of the pedigree
could even appear on the other. The values of the suc-
cessive coefficients of inbreeding (Z's) in such a case are
shown in Table VL where they are compared with the
coefficients of inbreeding in complete continued brother
X sister mating, where K = 100.^

From this it appears that an individual^may he inbred
in 10 generations to tvithin two tenths of one per cent, as
intensely, measured by the coefficients of inbreeding, if
his sire and dam are in no way related, as he would be if
his sire and dam were brother and sister. But clearly the
germinal constitution of the individual produced would,
except by the most remote chance, be quite different in the
two cases. This point is so evident as to need no elab-
oration. It has been brought out by East and Hayes.^

The values of the K^s for a particular pedigree evi-
dently furnish a rough index of the probability that the
two germ-plasms which unite to form an individual are
alike in their constitution. This will follow because of the
fact that the probability of likeness of germinal constitu-
tion in two individuals must tend to increase as the num-
ber of ancestors common to the two increases. Just what
is the law of this increase in probability is a problem in
Mendelian mathematics which has not yet been worked
out. The general fact, however, seems quite sure.

From the above discussion it seems plain that in reach-
ing a numerical measure of the degree of inbreeding it is
not sufficient to consider coefficients of inbreeding alone.
The coefficients of relationship must also be taken into
account.

It is suggested that the two constants be written to-
gether for each generation, the coefficient of inbreeding
being followed by the coefficient of relationship in brackets.
Thus we have

fi Since, of course, all of a sister 's ancestors are identical with her
brother 's.

• U. S. Dept. Agr. Bur. Plant Industry, Bui. No. 243, pp. 1-58, 1912.

Digitized by

Google

522 THE AMERICAN NATURALIST [Vol. XLVIII

Inbreeding and Relationship Coefficients of King Melia Biotes 14th

Zo (£:.)= 0 (0)

Z, (£:,)= 25 (0)

Z, (£:,)= 25.00 (50.00)

Z.(^4)= 37.50 (62.50)

Z4 (£:.)= 50.00 (75.00)

Z, (X. )= 71.88 (87.50)

Z^{Kr)=Sl,25 (92.19)

Z, (£:,)= 90.63 (92.97)

Z,(i:,)= 92.77 (93.75)

Z, (£:«)= 93.65 (93.75)

Z,o(Ku) =93.85 (93.75)

Zu(£:„) =93.85 (93.75)

The physical meaning of these expressions is simple
and straightforward. Z^{K^) tells us that in the 5th an-
cestral generation of King Melia Rioter 14th he had only
one half as many different ancestors as was possible for
that generation, and of his ancestors three fourths were
common to his sire and his dam. However one looks at
the matter there can be no denial that King Melia Rioter
14th is a closely inbred animal.

In Fig. 1 the heavy broken line gives the relationship
coefficients for King Melia Rioter 14th. It will be instruct-
ive now to consider another example by way of contrast.
Again a Jersey bull, Blossom's Glorene (102701), will be
taken. Only the final result need be given.

Inbreeding and Relationship OoErnciENTs op Blossom 's Glcrene

Z.{K,)= 0 (0)

Z, (£:,)= 0 (0)

Z,(i:.)= 12.50 (0)

Z,(B:,)= 12.50 (0)

Z,(i:5)= 25.00 (0)

Z»(i:e) =29.69 (0)

Z.(JS:t) =35.94 (0)

ZtCBTs) =40.23 (0)

The total inbreeding and the relationship curves are
given in Fig. 2.

The difference in the breeding of this bull and the one
considered in the former example is striking. In the 8th
ancestral generation Blossom *s Glorene has but 60 per

Digitized by

Google

No. 573]

STUDIES ON INBREEDING

523

cent, of the number of different ancestors possible in that
generation, but not one single animal in the ancestry of
his sire occurs in the ancestry of his dam (within the
limits Ai to Ag). The probability is that Blossom's Glo-
rene is heterozygous in respect of most of his characters,
while King Melia Rioter 14th is homozygous.

dOf

^-^'

1 /

/
/

(-■•'""■

)

v»^

/■.

Z. A 4

►-K>-<

;

9 A

f A

Fig. 2. Diagram showing the total inbreeding (heavy solid line) and the
relationship (heavy broken line) curves for the Jersey bull Blos8om*s Qlorene,
oyer a period of eight ancestral generations. Compare with Fig. 1.

Summary

The object of this paper is to call attention to the fact
that inbreeding of considerable degree may exist in the
entire absence of any kinship between the two individuals
bred together, and to bring forward a method of sepa-
rately measuring what proportion of the observed in-
breeding in a particular case is due to kinship of the pa-
rents, and what to earlier ancestral reduplication. A pro-
posed coefficient of relationship is described, and its ap-
plication illustrated by concrete ciases.

Digitized by

Google

THE CHROMOSOME HYPOTHESIS OF LINKAGE
APPLIED TO CASES IN SWEET PEAS
AND PRIMULA

From the Zoological Laboratory, Ck>lumbia University.
CALVIN B. BRIDGES

Theke are two views as to the nature of linkage. The
earlier view, developed by Bateson and his co-workers, is
that this phenomenon is an expression of symmetrical
reduplications in the germ tract A more recent view,
developed by Morgan and his co-workers, treats linkage
on the basis of a linear arrangement of genes in the
chromosomes and of the history of these genes during
normal gametogenesis. The advocates of the reduplica-
tion view have rarely applied their principles to the re-
sults on Drosophila on the ground that the results for
Drosophila are complicated by sex-linkage. That sex-
linkage is simply an additional, but wholly independent,
phenomenon, is proven by the many cases in DrosophUa
in which sex-linkage is not involved, yet in which the link-
age of the genes to each other is of the same type as the
linkage of sex-linked genes to each other.

In this paper I shall attempt to show that the theory of
linkage which we have successfully applied to all cases in
Drosophila, whether involving sex-linked genes or genes
which show no sex-linkage, applies equally well to the
non-sex-linked cases occurring in sweet peas and primula.
The only serious drawback to such an application lies in
the nature of the data which have been collected for these
cases. The least satisfactory form of data from which to
determine a linkage value is that presented by Fg results.
In cases in which two r^cessives enter from opposite par-
ents (** repulsion"), the excessive smallness of the douWe
recessive class in Fg renders any calculation subject to great
error. Slightly better are the F2 results from coupling,

524

Digitized by

Google

H-

i;*!

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 525 ^ -

but here there is no direct parallelism between the

zygotic and gametic ratios. In determining what gametic l> "^

ratio underlies the Fg results given by an experiment, the
practise has been to compare by the eye the given result
with a series of Fg results calculated from selected gametic
ratios. Collins has shown^ that this practise has led to
serious error. In F2 coupling cases in which there has
been no crossing over in one sex (autosome genes in [t.

Drosophila), there is a direct relation between the gametic
and zygotic series, but only in certain classes which com-
prise from one fourth to less than one half of the indi-
viduals of an experiment. While such data are more s
accurate than the usual F2 results, yet the percentage of , r
individuals which can be used directly is so low that we I '•
avoid the use of such a method. In F2 results involving JT,
only sex-linked genes, the efficiency is at least 50 per cent., ii[
for here there is always a direct relation between the
gametic and zygotic ratios in one half the flies (the males).
However, half the total number of flies (the females) are
useless unless the cross is made in such a way that Fa
becomes a back cross. These different kinds of F2 results
(the two most advantageous of which are not generally
applicable) are separated in effectiveness by a wide gap
from the back cross which we use equally well in all cases,
which gives a zygotic ratio directly proportional to the
gametic ratio, and in which every individual occurs in the
most advantageous relations.

Perhaps the least unsatisfactory method of dealing
with such F2 series as are available in the case of the
sweet peas, is by means of the coefficient of association as
derived by Yule. Yule^s coefficient of association is calcu-
lated from a zygotic series of the form AB:aB:Ab:(ib by
the formula:

^ «: . . . ... (ABXab)-'(aBxAb)

Coefficient of association = / aj^-^ t (-7-/ t> ^ . >i t ( .

(AB XoJ)) + (aB X Ab)

To find the gametic ratio corresponding to this coeffi-
cient, use is made of a table which gives the coefficients

1 Am. Nat., '12.

Digitized by

Google

526 THE AMERICAN NATURALIST [Vol. XLVIII

calculated from the zygotic series corresponding to such
gametic ratios as 2.5:1, 3:1, 3.5:1, etc. For the same
ratio in the coupling and repulsion series the coefficients
are slightly different, so that two tables should be made.
Upon the chromosome basis the best method of express-
ing the amount of linkage is in terms of percentage of
crossing over. The gametic ratio ml found through the
coefficient of association, when expressed as a percentage

, 100

becomes — — r.

n + 1

Aiccording to the chromosome hyi>othesis, all genes
which are linked to each other lie in the same chromosome.
In sweet peas the first case in which linkage was observed
was that of round pollen^ and red flower color. Later it
was found that hooded standard was linked to round and
to red. The genes for these three characters, then, may
be treated as though carried by the same chromosome,
which we may call chromosome I, of the sweet pea.

The relative distances of these genes from one another
in the chromosome can be determined from the degrees of
linkage. The farther apart in the chromosome any two
genes lie, the greater will be the amount of crossing over
between them. If two genes lie very close together, then
the percentage of crossing-over will be very small (the
gametic ratio very large).

Fortunately Punnett has recently collected the data,
upon these linkage cases in sweet peas. In the table which
follows, I have summarized the data given by the various
tables of Punnett. In the first column to the right of the
data appear the coefficients of association. In the next
column appear the corresponding gametic ratios calcu-
lated by interpolation to the nearest tenth. In the last
column are the equivalent percentages of crossing over,
found from the gametic ratios.

We may use one per cent, of crossing over as our unit
of distance in measuring the space between two genes.

2 I have used a terminology here like that used for the cases in DrasophilOj
naming the gene after that member of the pair of allelomorphs which may
be considered as the mutant from the wild type of pea.

Digitized by

Google

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE

527

The gene for red is then about eleven units from that for
round, and the gene for hooded is nearly one unit from
thai for red.

TABLE I

Chromosome I

Bound Pollen and Bed Color

Wild Type

Coefficient
of

ASIOCift-

Round Red Round Red tlon

Qftmetlc
Ratio

overt

Coupling ... 7,897

583 614 2,197 .9596
Bed Color and Hooded Standard

7.9:1

11.C

Wild Type

Red Hooded Red Hooded

Coupling ...2,568

16 17 857 .9998
Bound Pollen and Hooded Standard

125. :1

Wild Type

Round Hooded Round Hooded

Coupling ... 626

74 83 174 .8932

4.7:1

18.

Bepulsion ...3,140

1,413 1,438 14 .9577

8.7:1

10.3

The order of arrangement of these genes in the chromo-
some can be discovered from a comparison of the linkage
values found above. The linkage value (11.2) for round
and red is the most accurately determined of those in-
volved, so that we may lay this down as our initial or
base line:

0 . 11.2

DiAOBAM I. Ro = round pollen, R = red flower.

The next most accurate value is that for red and
hooded, namely, 0.8. Hooded lies therefore only about
one unit from red, but if these two values only, namely,
round red and red hooded, were given, we should be un-
able to decide whether hooded lies between round and red
at a position near 10 (that is, 11.2 — .8) or beyond red in
a locus at 12 (that is, 11.2 + .8). In order to determine
whether hooded lies to the left or to the right of red the
data for the third value, round hooded, need only be accu-
rate enough for us to decide between these values of 10

Digitized by

Google

528 THE AMEBIC AN NATURALIST [Vol. XLVIII

and of 12 units. The data from the coupling experiments
(which even though less extensive then those from the
repulsion experiments are probably more accurate) give
a value of about 18 units. Since the repulsion data give
10 units, 18 is probably too high, and an intermediate
•position correct. The higher (12) of the two possible
values is then the correct value. The position at 10 is not
excluded by these data, but is far less probable. In a case
in which one of the two first values is very small, as here,
the accuracy demanded of the remaining or third value is
much greater than in cases where neither of the values
are small, and one has only to decide between two very
different values by aid of the third. There are other ways
of arriving at this order of genes which are independent
of the size of the values. One of those methods, such for
example, as that of double crossing over, would definitely
settle the order of these three genes, but unfortunately
such data have not yet been published.

If hooded lies beyond red at 12, the complete first chro-
mosome diagram will be as follows :

Ro RH

0 11.2 12.

Diagram II. . Chromoeome I, Sweet Pea. Ro = round pollen, R = red flower,

H = hooded.

In the above diagram Rq indicates the locus of round
(and also of long). The symbols in the diagrams are
used to designate loci which may be occupied by either
allelomorph of the pair.

It has been observed that hooded flowers have always
a uniform color in standard and wings, instead of having
these two regions colored differently as in the normal or
bicolor type. Bateson assumed that this unicolorism was
only another somatic effect of the hooded gene. However,
an alternative explanation is that the unicolor is caused by
a specific gene which is very closely linked to hooded. If
this should be found to be the case, then this fourth gene
also will be located at about 12 units from round.

Digitized by

Google

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE

529

There is one other gene which probably belongs in the
first chromosome, namely, the intensifier found in the
'* black knight'^ race. The linkage data of red color and
intensity of color have been given in Report II to the
Evolution Committee, page 90.

TABLE n

Bed Color and Intense Color

WUd Tfpe

Coefficient
of

Alioclft-
Red Interne Redlntenie tlon

Gametic
Ratio

Percent-
age of
Crofi-
overi

Coupling ... 149

29 35 22 .527

1.9:1

35.

If these data are significant, then intense is in the first
chromosome at a locus about 35 to the right or left of red.
It should give about 24 (35 — 11) or 46 (35 + 11) per cent,
of crossing over with round, depending on whether it lies
about 24 to the left of round or 35 to the right of red.

The Second Chromosome op Sweet Peas

In the case of the second chromosome in sweet peas,
the linkage values are based on smaller numbers, but the
order of genes is more certain.

The first linkage case of this chromosome was that of
sterile anthers and light axils. Later the cretin form of
flower was found to belong to this linkage group. As in
the case of the first chromosome, I have summarized the
tables of Punnett in Table HI.

Wild Type
Coupling . . . 1,170 .
Eepulsion . . 1,335

TABLE III

Chromosome II

Sterile Anthers and Light Axil

Sterile

643

Coefficient

of
Assoc! A-
Light Sterile Light tion

30 379 .9945

714

.988

Percent-
age of
Oftoietic CroM-
Ratio oven

22. :1 4.4

20. :1

4.9

Wild Type
Coupling ... 282
Bepulsion . . 48

Light Axils and Cretin Flower
Light Cretin

Light
49
22

Cretin
52
27

59
3

.734
.610

2.6:1
2.7:1

28.
27.

Digitized by

Google

530 THE AMERICAN NATURALIST [Vol. XLVIII

Sterile Anthers and Cretin Flower

WUdType Sterile Cretin Sterile Cretin
Coupling ... 165 58 58 78 .556 2.:1 33.

ttepulsion .. 764 355 345 25 .683 2.6:1 28.

The linkage value for sterile and light, namely, 4.4 units,
is the most accurately determined of those in the second
chromosome. The value for light and cretin is about 28
units. Using the distance 4.4 between sterile and light as
our base line, then, we should find that cretin lies at 4 + 28
or 32 from sterile if the order of genes is sterile, light,
cretin ; but if the order is cretin, sterile, light, then cretin
should lie at 28 — 4 or 24 from sterile. The value for
sterile cretin should approximate either 24 or 32. There
is no very small value here as there was in the first chro-
mosome, and not such great accuracy is required of the
remaining value, since it should be easy to distinguish be-
tween 24 and 32. The coupling data for this value gh^
33 units, which enables us to fix the order of genes as
sterile, light, cretin. The following diagram of chromo-
some II expresses these relations more clearly.

0 4.4 32.

Diagram 111. Chromosome .II, Sweat Pea. S = sterile, L = light, C = cretin.

When crossing over is as free as in the case of sterile
and cretin and of light and cretin there should be some
double crossing over. That is, crossing over might occur
in the section of the chromosome near sterile and light
and at the same time another crossover could occur in the
section between light and cretin. This occurrence would
be readily seen if normal plants heterozygous in any com-
bination of these three genes were back-crossed to plants
purely recessive in all three. A relatively few plants from
such a test would give very valuable information on sev-
eral points, while an experiment of a few thousand indi-
viduals from such back-cross tests would enable one to
discover, through the phenomenon of interference, much

Digitized by

Google

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE 531

as to the character of the chromosome, the average length
of the intemode, and the i)ercentage of chiasmas per node.

Independence of Chromosomes I and II of Sweet Peas

If two groups of genes are carried by separate chromo-
somes, we may expect to obtain free assortment and
typical 9:3:3:1 ratios in F2, when any two genes from
different groups are involved. There are rather extensive
data for three such cases in sweet peas, and in each there
is practically complete independence. The data given in
Table IV are summarized from Report III to the Evolu-
tion Committee (page 37) and Report IV (page 17).

TABLE IV

Independence op the First and Second Chromosomes

Round Pollen (1st) and Light Axil (2d)

Coefficient Percent-

of a^e of

Associa- Gametic CroFs-

Wild Type Round Light Round Light tion Ratio^ overs

1,246 341 399 142 .131 1.15:1 47.

Red Color (1st) and Light Axil (2d)
Wild Type Red Light Red Light

1,563 545 506 232 .136 1.16:1 47.

Red Color (Ist) and Sterile Anthers (2dO
Wild Type Red SteHle Red Sterile

838 403 265 147 .071 1.07:1 48.

The greatest departure from the 50 per cent, of cross-
ing over expected from independent assortment is only
to 47 per cent.

There are several other characters whose genes seem to
be independent of those in the first and second chromo-
somes. This is interesting from the point of view that
each independent gene or group of linked genes requires
a distinct chromosome as a carrier.

Digitized by

Google

532 THE AMERICAN NATURALIST [Vol. XLVHI

Linkage Cases in Pkimula

In the case of primula, linkage was first found between
red (versus green) stigma and red (versus magenta)
flower color. Long style (versus short) and dark stem
(versus light) were found to be linked with red stigma.
Indications were observed that still a fifth gene, a domi-
nant which reduces the color of the flower to a tinge in the
corolla tube, belonged to this group.

A back cross involving the three genes, red stigma, red
flower and long style was made. Credit is due to Gregory
for the use of this method for obtaining linkage data.
Unfortunately many of the individuals were useless for
the linkage of red flower color, because of the occurrence
of white ; and the numbers are small.

In Table V, I have summarized the data given by
Gregory.*

table V
The First Chromosome op Primula

Bed Stigma and Red Flower
Non-crossoYera Crossovers

tl i| II §1 iU If III

Coupling
back cross ... / 28 39 17 18 1.9:1 34.6

Wild Red Red Red Stigma .

Type Stigma Flower Red Flower

Coupling F, 1,174 305 289 232 .511 1.8 : 1 35.3

Bed Flower and Long Style

Non-crossovers Crossovers

Red Long Wild Type Red Long

Coupling back cross. 40 53 6 5 8.4:1 10.9

Wild Type Red Long Red Long
Coupling Fa 38 2 4 12 .966 8.6 : 1 10.4

Bed Stigma and Long Style

Non-crossovers Crossovers

Red Stigma Wild Red

Long Type Stigma Long

Coupling back cross. 44 64 35 30 1.6:1 37.

^Jour, Genetics, '11, Vol. I; Proc. Bay. 8oc,, '11, Vol. — , 84.

Digitized by

Google

No. 573] CHROMOSOME HYPOTHESIS OF LINKAGE

633

Repulsion

Bed Stigma and Dark Stem
Wild Type Red Stigma Dark Red Stigma Dark
....137 66 62 0 —

The three values are— red stigma red flower 35, red
flower long style 11, and red stigma long style 37. Of
these, red stigma red flower is based upon the most data,
and may therefore be taken as our base line. The value
■ for red stigma long style should be 35 — 11 or 24, if the
order of genes is long, red stigma, red flower; but 35 + 11
or 46, if the order of genes is red stigma, red flower, long.
The value shown by the table is 37. This means that long
lies to the right of red at a locus 46.

35.

46.

Diagram IV. Chromosome 1, Primula. R,= red stigma, R = red flower, L = long

style.
58 59 > 60 61 62

The apparent discrepancy between the values 46 and
37 is due in most part to double crossing over, the effect
of which is always to lower large values disproportion-
ately more than short. When the discrepancy is known,
the amount of double crossing over can be calculated
approximately. Here the amount of double crossing over is

46-37
2

= 4.5.

That is, 4.5 per cent, of all the gametes are the result of
double crossing over. A somewhat larger amount of data
from a back cross in which all the individuals are effective
would give by direct experiment a true value for the
amount of double crossing over.

A chromosome diagram should be built up of values
independent of double crossing over. According to our
experience with Drosophila, if there is not more than ten
per cent, of crossing over between two genes, the double
crossing over is negligible. Thus in the first chromosome
in sweet peas, the values obtained from the experiments
are not changed by double crossing over. However, in the

Digitized by

Google

534 THE AMEBIC AN NATURALIST [Vol. XLVm

case of the second chromosome, where the total percentage
of crossing over is about 32, there is probably one or two
per cent, of double crossing over. The diagram of the
second chromosome is in this respect only tentative, and
the plotted position of cretin will be moved a little farther
to the right when the amount of double crossing over
between light and cretin has been found. The value 4.4
for sterile anther light axil is not aflfected by double
crossing over, since the section of chromosome between
these two loci is so short that a double break would prob-
ably not occur between them at all. The amount of double
crossing over between any two loci can only be found
when there is a gene between them. Thus if a gene
should be found which lies between light and cretin, either
by indirect calculation or, better, by direct experiment,
the amount of double crossing over could be found. The
more genes which can be worked with in the same chromo-
some, the more accurate becomes the diagram.

All the values found for these cases in sweet peas and
primula are based upon such small numbers that they
can be used only as illustrations of the way in which one
would apply to new cases certain principles worked out in
Drosophila. While they serve as examples in line with
these principles, they are entirely inadequate as proof.
A very interesting case of variation in linkage is pre-
sented by some of the families involving chromosome 11
of the sweet pea. In this article I have avoided such data
as far as I could, but it is possible that the order in which
I have aligned these genes will be found to be incorrect
when data upon all three genes in a back cross are ob-
tained. Such data would show, through the phenomenon
of double crossing over, what the order of genes is, even
though variations in the linkage should occur.

Columbia University,
May, 1914

Digitized by

Google

THE REDUPLICATION HYPOTHESIS AS
APPLIED TO DROSOPHILA

Dr. a. H. STURTEVANT
Columbia University

A NUMBER of papers developing the reduplication hy-
pothesis of linkage have recently appeared in the Journal
of Genetics. They are based almost entirely on the
experiments of Gregory ( '11) on Primula and of Punnett
( '13) on the sweet pea. The data are not entirely satis-
factory because of the relatively small number of genes
involved, and because in most cases the gametic ratios can
be only approximately determined. This is due to the
fact that most of the data concern Fj counts, from which
gametic ratios can not be calculated directly. In Gregory's
best case a much more satisfactory method was followed—
the heterozygous plants were tested, not by mating to
others of their kind,#>ut by crossing with plants recessive
with respect to all the genes involved, which gives the
gametic ratio directly. In this case, however, we have
only a relatively small series of data involving as many as
three pairs of linked genes. It is obvious that from such
data no adequate test of the reduplication hypothesis can
be made.

The phenomena of linkage have been very extensively
studied, by Morgan and others, in the fly Drosophila. In
this animal there are many genes belonging to the same
linkage groups, and these have been studied on a large
scale. In the case of the sex-linked group there is never
any difficulty in calculating the gametic ratio from F2
results, since the Fj males from any cross always give it
directly. I have recently published a paper (Sturtevant,
'14) giving a complete summary of the published results
obtained from studies of the linkage of these genes. In that
paper I have adopted the chromosome explanation of link-

535

Digitized by

Google

536 THE AMERICAN NATURALIST [Vol. XLVIII

age proposed by Morgan ( '11). Here I shall use the same
data for a test of the reduplication theory. It may be of
value to contrast the two views by making a rigorous
application of them to the same facts. Since the data
concerning the sex-linked group of genes in Drosophila
form the simplest and most extensive series now avail-
able, I ^hall deal more especially with them. The reader
is referred to my other paper for the detailed data, for
references to original sources, and for a full treatment of
the chromosome hypothesis as applied to these and other
data.

It may be well to give first a brief catalogue of the
sex-linked genes discussed in this paper. The nomen-
clature is that suggested by Morgan ('13). This may be
confusing to those accustomed to the '* presence and ab-
sence" system, but this should not be a serious objection
here, since a clear conception of the somatic appearance
of the animals discussed is not essential for our present
purpose. The relations would be as clear if hieroglyphics
were used for symbols.

Y is the gene which differentiatefe the wild **gray''
bodied fly from the yellow mutant, y.

V diflferentiates the wild red-eyed fly from the ver-
milion-eyed mutant, v.

M diflferentiates the ^'long" wing of the wild fly from
that of the miniature-winged mutant, m.

R is another gene aflfecting the wings. The wild fly
has R, the rudimentary- winged mutant has r.

Br^ occurs in a dominant mutant form having a narrow
eye known as barred. The allelomorph present in the
wild fly is designated fer'.

The other characters concerned bear such a relation to
one another that the genes involved are considered as
forming a system of quadruple allelomorphs. The alter-
native to this view is the assumption of complete linkage,
but I have given elsewhere (Sturtevant, *13) my reasons
for preferring the multiple allelomorph interpretation.
The eye of the wild Drosophila is red in color. A single

Digitized by

Google

No. 573]

REDUPLICATION HYPOTHESIS

537

mutant obtained from it had white eyes (Morgan, *10),
and this character proved to be a simple sex-linked reces-
sive. From the white-eyed form arose a fly with eosin
eyes (Morgan, 12). This new character was found to be
a sex-linked dominant to white, and a sex-linked recessive
to red. Finally, there arose a form with cherry eye color
(Safir, '13). This has the same relation to red and to
white as has eosin. Mated to eosin it gives an inter-
mediate color, which splits up into cherry, intermediate,
and eosin in Fg. The nomenclature adopted in this case
is as follows :

Allelomorph present in the red -eyed fly, W.
Allelomorph present in the white-eyed fly, tr.
Allelomorph present in the eosin-eyed fly, w^.
Allelomorph present in the cherry-eyed fly, iv<^.

Trow ( '13) has suggested the possibility of an asym-
metrical reduplication series, giving a gametic series of
wAB:xAh:yaB:zah, where w need not equal z, nor
X equal y. It should be noted that an actual demonstra-
tion of such a ratio, or of its non-existence, is almost ex-
cluded for the reason that it would be practically impos-
sible to be sure one was not dealing with a case involving
diflferential viability. However, perhaps the most stri-
king general fact brought out by the study of linkage is
that each pair of linked genes (allelomorphs), considered
separately, follows a perfectly regular Mendelian course.
I think we are, therefore, justified in assuming that the
number of gametes bearing A is always equal to the num-
ber bearing a, and similarly for B and 6. Then, in Trow's
asymmetrical series,

w + x^y + z, .

w '\-y=^X'\-z.
Hence, «

w=^z and x=^y.

In all that follows I shall assume that the reduplication
series are always symmetrical. On this assumption it
becomes unnecessary to consider the two halves of the

Digitized by

Google

538 THE AMERICAN NATURALIST [Vol. XLVIU

series separately, and I shall therefore use only two terms
in speaking of gametic ratios. By adding together the two
halves of the series larger numbers are obtained, so that
chance deviations are relatively smaller. Differential
viability is also partially overcome in this way. Of
course on the reduplication theory both terms of the
gametic ratio must be integers, since they represent num-
bers of cells, but nevertheless it has seemed to me more
convenient for purposes of calculation to express them
always in the form nil. Thus a gametic ration of 3:2
may be written 1.5 : 1.

It was suggested by Bateson and Punnett ('11) that
the intensity of coupling and of repulsion between the
same two pairs of genes may be identical. That this is
substantially the case has been shown again and again in
Drosophila, and has become a truism among those work-
ing on that form. Before presenting data on this point I
wish to bring up another matter on which the same data
have a bearing. Punnett ( '13) has said, * * But where three
[pairs of] factors are concerned . . . the value of the
primary reduplications is evidently altered, and there
would seem to be some process whereby these reduplica-
tions react on one another." Bailey ('14) has suggested
that the nature of this interaction may be such as to cause
the two primary series to be o£ equal intensity. It may be
categorically stated that there is no interaction effect in
Drosophila, The best data for a test of the relative inten-
sity of coupling and repulsion, and of ''fundamental,"
''primary'' and "secondary" reduplication series, in-
volving the same allelomorphic groups, is that furnished
by the relations of the various forms of W {W, tv, tv^, tv^)
to the M pair of allelomorphs {M and m). Table I is a
summary of the data on this case. In computing the
fundamental series I have used only the data from such
of my own experiments as involve only two pairs of genes,
since that from other sources is for the most part made tip
of primary series in which the other primary series in-
volved is masked.

Digitized by

Google

No. 573]

REDUPLICATION HYPOTHESIS

539

TABLE I

Fundamental Series

Nature of Cross Actual Numbers

Gametic Ratios

WM X wm

777: 470

1.6 + : 1

Wm X wM

93: 221

1 : 2.4 —

WM X w^m

634: 348

1.8 -f: 1

Wm X weM

46: 110

1 : 2.4 —

Wm X w^M

461 : 855

1: 1.9 —

w^M X wm

4,171: 1,858

2.2 -f : 1

w^m X wM

891: 1,898

1 : 2.1 -h

woM X wm

75: 47
Primary Series

1.6:1
Other Primary

Nature of Cross Actual Numbers Gametic

Ratio Series Involved

WM X w^m

178: 85 2.1—:

MBr"

w^m X wM

69: 122 1:1.8

—

MBr"

WM X wm

5,838:2.911 2.0 -|- :

Wm XwM

1,111:2,493 1:2.2-1-

WM X wm

2,261:1,011 2.2 -h:

Secondary Series

Primary Series

WM X w^m

719:407 1.8 —

: 1

WV, VM

Wm X W7«M

' 227:509 1:2.2

—

WV, VM

It will be noted that in all these cases the gametic ratio
approximates 2 : 1, or 1 : 2, according to the nature of the
cross. There are only four cases showing a noticeable
deviation from this value, and of these two involve only-
small counts. The most serious is the first. In this case
there is a deviation of 54.3 from the 2 : 1 ratio, and the stand-
ard error is 16.7[V1/3X 2/3 X (777 -f 470) = ± 16.7-].
Since the deviation is slightly over three times the stand-
ard error, it is perhaps significant, especially since there
is at least one other rather large deviation (the second
ratio in Table I). For our present purpose, however, it
is probably not significant, since similar deviations occur
in different experiments of exactly the same type. I have
recorded elsewhere (Sturtevant, '14) the results of a num-
ber of tests of individual females heterozygous for these
two allelpmorphic groups. Taking only those cultures
which produced 100 or more flies, we find the following
results :

Digitized by

Google

540 THE AMEBIC AN NATURALIST [Vol. XLVHI

Seven females of the constitution w^mwM gave gametic
ratios ranging from 1.5 : 1 to 2.7 : 1, with the modal class
at about 2.0:1.

Seventeen females w^Mwm gave ratios ranging from
1.5:1 to 3.4:1, with a single individual at 4.2:1. The
modal class was at about 2.2 : 1.

It seems highly probable that all these deviations from
a 2 : 1 ratio, not due to insuflficient numbers, may be satis-
factorily explained on the basis of differential viability,
which is known to occur here (for a discussion of the
vagaries of diiferential viability see Bridges and Sturte-
vant, '14). I do not wish to be understood as arguing
that the gametic ratio for any two pairs of genes is abso-
lutely constant, but only that it is in most cases uninflu-
enced by the way in which the genes are combined and by
heterozygosis for other genes. That it may sometimes
show marked differences is now well established. I have
myself studied two cases of this sort, and I have good
evidence (not yet published in detail) that there are defi-
nite genes which cause great differences in the ganietic
ratios for whole linkage groups. In one case this gene
itself shows linkage to those in the group it affects. But
even here the intensity of coupling and of repulsion is
affected alike, and it makes no difference how few or how
many genes a fly is heterozygous for ; the linkage is strong
or weak according to the form of the linkage-affecting
gene which the fly happens to carry. In each of these
cases I have been able to obtain about the same extreme
values both for coupling and for repulsion.

In what follows I shall assume that the intensity of the
reduplication series is not affected by the way in which
the genes are introduced, nor by the number of linked
genes involved in the cross. The obvious corollary of
this is that reduplication occurs even in homozygous indi-
viduals, and that the nature of the series of divisions is
in general independent of the constitution of the indi-
vidual. This conclusion is directly opposed to the point
of view expressed more especially by Punnett, in the

Digitized by

Google

No. 573]

REDUPLICATION HYPOTHESIS

641

passage quoted above and elsewhere. If reduplication
occurs at all it is the same in the wild fly as in the most
complex linkage experiment we have yet carried out.

If it is assumed that the intensity of coupling and re-
pulsion is identical, it becomes unnecessary to consider
them separately. I shall therefore lump together all the
data involving the same groups of allelomorphs, regard-
less of how they were put into the cross. When three
pairs of genes are involved there are eight possible com-
binations of them in F2, but only four if we add together
the two halves of the reduplication diagram. There are
the two original combinations, which I shall designate
ABC. Then there are three combinations derived from
each of these by a shifting of one gene, which I shall
designate ABc, AhC and aBC, the small letters referring
to those pairs which have been shifted. Thus, to take an
imaginary case, if we cross LMn by ImN, the gametes
produced by the Fi individuals will be classified as
follows :

ABC ABc AhC aBC
LMn LMN Lmn IMn
ImN lmn IMN LmN

In the following tables I shall reduce all data to this
form. In each case the genes will be arranged so that
AB and BC will be the primfiry reduplication series.^

Table II contains such a summary of all the crosses in-
volving three pairs of sex-linked genes. Table III shows
the gametic ratios derived from these data, and also the
values for the secondary series calculated on the basis of
Trow's ^* special" hypothesis. For the sake of brevity
only one term is used : a gametic ratio of 3 : 1 is written 3 ;
a ratio of 3 : 2 becomes 1.5, etc. With the simplifications
introduced here Trow's formula becomes

AC =

(ABXBC)+1

AB + BC '

1 As was pointed out by Punnett ('13), in a system of three reduplica-
tion series the one with the lowest intensity is to be regarded as the second-
ary series.

Digitized by

Google

542

THE AMERICAN NATURALIST [Vol. XLVIII
TABLE n

Allelomorphic Groups

ABC

ABc

AbC

aBC

YWM

8,212

4,013

119

YWR

278

160

YVM

1,082

665

YVR

315

138

196

YVBr"

WVM

194

102

WMR

1,726

535

139

872

WMBr'

220

129

TABLE III

Gametic ratios

Experiment

Observed

Calculated

AB \ BC AC

YWM

95.5 2.1
438.0 1.74

1.7 22.0

1.8 2.6
2.0 3.4
2.0 24.7

2.0-

1.72

1.5

1.7
1.3
1.2

2.0 -f

YWR

1.74

YVM

1.6

YVR

1.3

YVBr"

1.4

WVM

1.9

WMR

WMBr"

2.2 3.9
1.9 3.6

1.6
1.4

It will be seen that in every case the calculated value
for the secondary reduplication is higher than the ob-
served value. The same relation comes out in two experi-
ments which I have done involving genes of another
group in Drosophila (see Table VIII, Sturtevant, '14).
Punnett's case is so involved that calculations accurate
enough for our present purpose can not be made. In
Gregory's experiment one of the genes (M) could not be
followed in all the plants because masked by another gene.
We are not given the data for S and G in those plants in
which M was classified separately from those in which it
was not. The data are therefore not available for exact
calculations, since the numbers are too small to overcome
chance deviations. The data for my own two experi-
ments appear in Table IV.

The same relation comes out more strikingly in another
way. If we let m equal the intensity of the AB series and
n that of the BC series, then on Trow's special hypothesis

Digitized by

Google

No. 573]

REDUPLICATION HYPOTHESIS

543

the four kinds of gametes should occur in the following
proportions :

ABC — mn

ABc — m

aBC — n

AhC-l

TABLE IV

Observed

Calculated

Experiment

BVgCv

3.4
2.6

2.4
1.0

2.7

BCvSp

1.4

That is, l/(m + l) of the gametes should have A and B
interchanged. Of these, l/{n-\-l) should have B and C
also interchanged. If N represents the total number of
gametes, then the size of the AhC class should be repre-
sented by the expression

AbC =

(m + l)(r^+l)

Table V shows the relation between the size of this class
as observed and as thus calculated, in the ten experiments.

TABLE V

ABC

AUelomorphic ^ a ^

Groups Observed Calculated

YWM 9 42

TWB 0 0

YFM 22 30

YVB 55 69

YVBr" 10 15

WFM 1 4

WMB 139 208

WMBr" 25 34

BVgCv 2 7

BCvSp 12 20

Thus it appears that in all ten experiments Trow's
formula gives values for the AC series and for the AbC

Digitized by

Google

644 THE AMERICAN NATURALIST [Vol. XLVIU

term which are too large. Moreover, this feature appears
in a more complex cross which I have carried out, in-
volving four pairs of linked genes {YWVM)^ and in each
separate part of all these experiments, regardless of how
the crosses were made. It may, then, be taken as a con-
stant relation. It can only mean that there is some rela-
tion between A and C besides that resulting from second-
ary reduplication. In other words, to use Bailey's terms,
Trow's ''special'' hypothesis is not valid.

Let us then examine what Bailey calls Trow's ''gen-
eral" hypothesis. Suppose the primary series to be of
the following values :

AB = l:l,
BC = m:l,
AC = n:l.

Trow's general formula for calculating what should be
the observed value of the AC series is

.^ lmn-\-n

The special formula is derived from this by assuming
n = lj when the formula becomes

Since this always gives a value which is too large, it
follows that n is always less than one. This means that
the AC primary series is reversed— that the combinations
present in the parents tend to be reproduced in fewer
numbers than the new combinations. I have worked this
out for the case of BCvSp (see Table IV), and find the
primary series there to be 0.6:1, though the observed
series is 1.0. The "fundamental" AC series has been
obtained for most of the cases in Table III, and has
always been found to be of the usual form {i. e., n:l,

Digitized by

Google

No. 573]

REDUPLICATION HYPOTHESIS

545

where n>l). (See Table I, Sturtevant, 14.) In fact,
as stated above, the fundamental series always approxi-
mates the secondary (observed) series.

There are two hypotheses as to the mechanics of re-
duplication series where more than two pairs of genes are
involved. The first was suggested by Bateson and Pun-
nett ( '11), and consists in the assumption that when three
pairs are involved eight cells are formed by three succes-
sive divisions, each of which segregates one pair of genes.
The eight cells then represent the eight possible kinds of
gametes, and are supposed to reduplicate independently
until the proper proportions are reached. Bailey sup-
poses that if it be shown that two primary series do not
interact on each other this scheme will be more likely to be
correct than will Trow's, which I shall discuss next. It
seems to me, however, that this hypothesis begs the ques-
tion. It is derived entirely by working backwards from
the observed results; it affords no basis for predictions;
and it does not offer a simple mechanical explanation of
any of the observed results. For pragmatic reasons I
believe we should adopt it only as a last resort.

Trow supposes that two cell divisions occur, segregating
two pairs of genes. The four resulting cells then go
through with their reduplication, which is a primary one.
When this is finished there occur divisions which segre-
gate the other pair, and the other primary reduplication
is carried out. On Trow's general hyjx) thesis, which I
have tried to show is the only one which can hold, it is
supposed that the second series of reduplications is
affected by both of the first two pairs of genes. C is re-
duplicating more if with B than if with b, less if with A
than if with a. This scheme of Trow's has one great
advantage in that it accounts for the fact that tiie class
which I have called AbC is always the smallest one.
Reference to Trow's calculations will show that this rela-
tion should always occur, and Table n shows that it does
occur. On the octant scheme there is no explanation of
this relation — ^we simply have to assume that it does occur
somehow.

Digitized by

Google

546 THE AMERICAN NATURALIST [Vol. XLVHI

It will be noted that several of the gametic ratios in-
volved here closely approach 2 : 1. YV, YM, WV and WM
are the most conspicuous examples. It may seem that
such a simple ratio is due to a very simple reduplication
series, but I do not think such an assumption can be suc-
cessfully maintained. The tables given above show that
YM and WM have approximately this same value when
they appear as secondary series, and the data for the
combination YWVM show the same thing for YV (see
Sturtevant, 14).

If, as I have maintained above, the same series of redu-
plications must occur in all flies, whether we can follow it
or not, then it follows that in these three cases the 2 : 1
ratio is never due to a simple series, but always to a long
and complicated one, since in all three one of the primary
series is of high intensity.

It was pointed out by Trow that the intensities of the
reduplication series afford a method of calculating the
number of cell divisions necessary to complete the series.
If we assume that approximately the same series is
occurring both in homozygous and in heterozygous flies,
we have the following series in Drosophila as a basis for
such calculations.

Sex-linked

Group

YW =

90.1

wr=

2.1

VM =

31.8

MR =

5.0

RBr =

21.7

Second Group

BVg =

3.6

VgCv =

10.4

CvSp =

2.8

SpBa =

10 +

Third Group

PEb =

100 ±

Digitized by

Google

No. 573] REDUPLICATION HYPOTHESIS 547

All of these series must be considered as either primary
or secondary and therefore involving primaries of higher
intensity. In fact there is unpublished evidence that
many of them can not be simple primaries. A num-
ber of series of very high intensity are known, and will
appear in future publications. Therefore all the calcula-
tions that follow give results which are far too small.

According to Trow, the minimal number of successive
cell divisions required to complete the series is given by
the expression mnp • • • where m, n, p, etc., are the larger
terms of the primary series involved. In the present case
the value of that expression is something over 76,000,-
000,000. However, Trow's formula seems to be wrong.
If a be the number of cell divisions required to produce
m cells, then 2* = m. If this expression gives a value of
a which is not an integer, then the next higher whole
number is to be taken. In the case of the first series two
divisions are necessary to segregate the genes, and in the
following series one is required. The number of succes-
sive cell divisions required then is (a + l)-f(& + l)
-f-(c?-f-l)-f----f-l, where b, c, etc., bear the same rela-
tion to n, p, etc., that a does to m. In the case of Droso-
phUa the value of this expression is 56. As pointed out.
however, this value is certainly far too small.

The total number of cells required is given by the
expression 2mnp • • • + 2np • • -j- 2mn • • + 2mp • • •
-f 2m • • • -f 2w • • -f 2p • • • -f • • • + 2mnp -f 2mn -f- 2mp
+ 2np+ • • • +2m + 2n + 2p,-\-l.

This gives a value considerably above 600,000,000,000— a
manifest absurdity. However, it is not necessary that all
these cells should be produced, since the ratios would not
be appreciably affected by some lines becoming crowded
out. It is necessary, on the other hand, that all of the
series shaU be completed in every line which does live,
since every female Drosophila,^ which is of the proper
constitution to be tested, shows linkage for every pair of
genes tested.

2 The results discussed here deal only with the linkage in female flies.

Digitized by

Google

548 THE AMERICAN NATURALIST [Vol. XLVIH

Thus we are forced to assume an enormously complex
series of cell divisions, many of them differential, pro-
ceeding with mathematical regularity and precision, but in
a manner for which direct observation furnishes no basis.
It seems to me that it is not desirable to assume such
a complex series of events unless we have extremely
strong reasons for doing so. I can see no sound reason
for adopting the reduplication hypothesis. It apparently
rests on two discredited hypotheses : somatic segregation,
and the occurrence of members of the 3 : 1, 7 : 1, 15 : 1, etc.,
series of gametic ratios in more cases than would be ex-
pected from a chance distribution.

The chief advantage of the chromosome hypothesis of
linkage which has been proposed by Morgan ('11), and
which I have followed elsewhere, seems to me to be its
simplicity. In addition it appeals to a known mechanism,
and a mechanism toward which the experiments of Boveri,
Herbst, Baltzer and others point as the correct one. It
explains everything that any of the forms of the redupli-
cation hypothesis does, and in addition offers a simple
mechanical explanation of the fact that ** secondary
series" are always smaller than Trow's *' special hypoth-
esis ' ' calls for them to be. On the reduplication hypoth-
esis this fact must merely be accepted, for, I think, it
<?an not be explained.

Columbia Unr^eesity,
May, 1914

LITERATURE CITED

Bailey, P. G.

'14. Primary and Secondary Reduplication Series. Jour, Genet. , III.
Bateson, W., and R. C. Punnett.

'11. On Gametic Series Involving Reduplication of Certain Terms.
Jour, Genet., I.
Bridges, C. B., and A. H. Sturtevant.

'14. A New Gene in the Second Chromosome of Vrosophila, etc. Biol,
Bull, XXVI.
Gregory, R. P.

'11. On Gametic Coupling and Repulsion in Primula sinensis. Proc,
Boyal Soc.y 84. B.
Morgan, T. H.

'10. Sex Limited Inheritance in Brosophila, Science, XXXII.

Digitized by

Google

No. 573] REDUPLICATION HYPOTHESIS 549

'11. An Attempt to Analyze the Constitution of the Chromosomes on
the Baeis of Sex-limited Inheritance in Drosophila, Jour, Exp,
Zool, XI.
'12. Further Experiments with Mutations in Eye-color in DrosopMUi.

Jour, Acad, Nat, 8oi, Philadelphia, XV.
'13. Factors and Unit Characters in Mendelian Heredity. Amer. Nat.,
XLVII.
Punnett, B. C.

'13. Beduplication Series in Sweet Peas. Jour, Genet., III.
Safir, S. B.

'13. A New Eye-color Mutation in Drosophila, Biol, Bull, XXV.
Sturtevant, A. H.

'13. The Himalayan Babbit Case, with Some Considerations on Mul-
tiple Allelomorphs. Amer. Nat., XLVII.
'14. The Behavior of the Chromosomes as Studied Through Linkage.
Zeiis, f, ind. Ahst,- u, Vererb,-Lehre,
Trow, A. H. )

'13. Forms of Beduplication — Primary and Secondary. Jour,
Genet,, II.

Digitized by

Google

PATTERN DEVELOPMENT IN MAMMALS AND

BIRDS. Ill

GLOVER M. ALLEN
Boston Museum op Natural History

Partial Albinism in Wild Birds

In birds under natural conditions of wild life partial
albinism is fairly common. Lists of species of which
aibinistic specimens are known were published by Ruth-
ven Deane (1876, 1880) some years ago, and by others.
Scattered instances are in all the bird journals or maga-
zines of general natural history. In most cases in which
the white markings are clearly defined against the pig-
mented parts of the plumage, these may be referred to
their particular primary breaks between the several
areas of pigment formation. In other cases the pigment
reduction is of the diffuse type, tending to form spots.

A few instances follow in which the several primary
patches have been observed in wild birds, either as acci-
dental marks or as permanent parts of the pattern.

The Crown Patch. — In 1908, a pair of robins nested
near Lowell Park, Cambridge, one of which showed a
partial separation of the crown patch, through the pres-
ence of a white band, as broad as the eye^s diameter,
passing from one eye around the back of the head to the
other eye. In the Wilson Bulletin (Vol. 2, p. 45, 1908)
W. E. Saunders records the capture of two robins each
with a white collar about the neck, probably marking: the
separation of the neck patches from the shoulder patches.
Coues (1878) records a brood of black robins at St.
John's, N. B., one of which was kept in captivity by the
late G. A. Boardman. In September, after moulting, it
was still pure black, except for white wings and tail,
which seems to indicate an areal restriction of the
shoulder and rump patches, though the pigment, where

550

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 551

produced, must have been superabundant. Ward (1908)
has described a case of a black robin becoming albinistic
and reviews a number of such cases. The ability of the
same feather follicles in different moults to produce
feathers with different sorts or amounts of pigment is
thus evidenced and has lately been carefully studied by
Pearl and Boring (1914) in the hen.

In addition to the case of the robin above mentioned,
the white line marking off the crown patch from the ear
patches is sometimes found abnormally in other birds.
Thus Sweet (1907) records two slate-colored juncos
{J unco hy emails) taken in March, 1903, at Avon, Maine,
in which there was a white line above the eye, and the
black throat patch was absent, owing no doubt to the
ventral restriction of the neck patches, as often seen, for
example in pigeons. Maynard^ figures the head of a
young female black-poll warbler {Dendroica striata) in
autumn, showing an inclination to assume a white super-
ciliary stripe. I am convinced that this mark so common
in many birds, is merely a development of the primary
break marking off the crown patch from the ear patches
so that it has become a permanent part of the pattern.

The failure of the crown patch to develop at all, as is
sometimes the case in the domestic pigeon, results in a
white-crowned bird. In the West Indian Columba leuco-
cephala, exactly this modification has taken place and the
entire top of the head is permanently white. The same
condition is found in sundry other genera, including a
humming bird, a heron, and others. It would be inter-
esting to discover by experiment if it were not easier to
produce a definite white marking through selecting for
the non-development of a certain patch or patches, than
to try to restrict a certain pigment patch to definite
bounds as in the experiments of Dr. MacCurdy and Pro-
fessor Castle (1907).

The crown patch as a separate unit in pigmentation, is
often of a different hue from the surrounding patches.

1 ''Birds of E. North America/' 1896, p. 585.

Digitized by

Google

552 THE AMERICAN NATURALIST [Vol. XLVIII

Thus in the case of the terns, the black-capped chickadee,
the black-crowned night heron, and other birds, a black
crown patch is noticeably marked oflF.

The Ear Patches. — The ear patches in birds are small,
yet often specially marked out by white boundaries, which
are permanent parts of the pattern. Yet there is no
doubt but that the acquisition of such white boundaries
is a derived character. It is common for the ear patches
to be colored differently from the surrounding parts,
forming as in some species of tanagers a black auricular
area contrasted with the blue of the head and neck. Of
particular interest in the present connection, however,
are those cases in which a pigmented ear patch is more
or less clearly marked off by a white line above it or
below, or both. The superciliary stripe, so common in
birds, is of course a development of a primary break
above the patch, separating it from the crown patch.
Where the stripe is narrow it is hard to say which patch
has begun to be restricted, though often no doubt both
are more or less involved. Thus the Garganey teal has
a very wide white eye stripe, and in other species of
ducks the whole side of the head may be white, indicating
much greater restriction of pigment formation in con-
tiguous patches. A beautiful example of the develop-
ment of a white stripe at the lower border of the ear
patches is found in the Inca tern, in which a line of white
feathers runs from just above the gape along the lower
side of the auricular patch and separates it from the
dark throat. But not only is the white line developed,
but the feathers composing it are specially elongated and
recurved, as if the mark were one of particular decora-
tiveness. The dark ear patch is noticeable in many
hawks, separated above and below by white areas, as in
the duck hawk and the osprey, though differing in the
size of the white areas.

An instance in which the white line separating the
crown patch from the ear patch, is even now in course of
becoming established as part of the permanent pattern,

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 553

is afforded by the common guillemot {Uria troille) of the
northern Atlantic. The other related species of the genus
have the head and neck uniformly pigmented, but in U.
troille a considerable proportion of specimens show a
narrow white eyebrow and a postorbital line, in exactly
the situation of the stripe in the albino robin previously
noted, though not so broad nor so extended. Birds so
marked were formerly considered a distinct species — the
ringed murre {Uria ^^ringvia") — or perhaps a plumage
of U. troille, and much effort has been made to determine
their exact status. Both plumages are f oimd in the same
colonies and the two sorts of birds are known to have
mated together (Miiller, 1862). Verrill estimated that
about 40 per cent, of the nesting birds he saw on the
Labrador coast were of this variety, but this is probably
a rather high estimate. I am convinced that the true
explanation of this puzzling variation is that incipient
albinism has gained a foothold, of such nature that areal
restriction of the ear or crown patches is developing, so
that a white line results between them. In the crested
auklet i^thia) a member of the same family, of the
Pacific Coast, such a line has become fixed so that it
now forms a characteristic mark of the species. In the
case of the ''ringed murre," I should expect to see the
eye stripe in the young as well as in the adult stage of
those individuals which are to have the mark — in other
words it is a permanent trait. No doubt the heredity of
this white stripe is of some definite sort, and if a reces-
sive character, it may nevertheless in time become com-
mon to an increasing number of birds, as this is a
colonial species and the possibility of inbreeding is thus
increased.

The Neck Patches. — In birds the neck patches extend
forward from the breast to meet the crown patch at the
occiput and the ear patches at the sides of the head,
thence ventrally to include the throat and chin. A study
of albinifetic pigeons, as previously noted, indicates that
the neck patches are two separate areas of pigmentation,

Digitized by

Google

554 THE AMERICAN NATURALIST [Vol. XLVHl

one on each half of the part covered, with an ultimate
center at the base of the neck, usually the last spot to re-
main when the area is much reduced.

In albinistic individuals, that is, those in which restric-
tion of the pigment areas has taken place, the neck
patches are usually first reduced at the upper part of the
throat, so that a white patch appears from the chin to
upper throat, as commonly seen in street pigeons; in
others, however, the restriction may be at the posterior
end of the patch, so that a white ring develops at the
bas.e of the neck.

In many birds the neck patches have been much devel-
oped as characteristic pigmented areas. Two general
categories may be here distinguished: (1) those in which
the neck is rather uniformly colored all about, and (2)
those in which the ventral portion is heavily pigmented
and the dorsal portion much less so. In the latter belong
such birds as the black-capped chickadee {Penthestes
atricapillus) with a black throat but a pale neck. So,
too, the golden- winged warbler (Vermivora chrysop-
tera). In this latter category it is probable that a sec-
ond factor is present, comparable to that producing a
centrifugal type of pigmentation in mammals, such for
example as in the Himalayan breed of rabbit, which has
the end of the nose and the feet black-pigmented, contrary
to the usual rule of normal areal reduction where the
extremities are the first to become white. That this is
a separate category from a physiological standpoint is
indicated by its behavior in heredity as worked out so
admirably by Faxon (1913) in the case of the Brewster's
warbler. He discovered that the black throat as present
in the golden-winged warbler is recessive in the cross
with a related species, the blue- winged warbler (Ver-
mivora pinus)y a yellow- throated bird. The oflFspring
of this cross have white throats,— the so-called F. leuco-
bronchialis. The black throat patch may be evidence of
*^ centrifugal" pigmentation as defined farther on (p. 53).
The essential bilaterality of such a throat patch is

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 555

further shown by the fact that one half only may be pres-
ent as in the golden-winged warbler recorded by Dr. C.
W. Townsend (1908).

The first category, in which the neck is uniformly pig-
mented is illustrated by many of the duck tribe, and
probably involves the normal primary patches only.
The primary patches are usually restricted first antero-
ventrally producing a white throat. Often this is carried
dorsally so as to form a white ring around the upper
part of the neck by the separation of the neck patch from
the crown and the ear patches. Again, if the neck patches
are restricted posteriorly a white ring is formed at the
base of the neck, a comnion permanent character in many
species. The peculiar little goose-like hird—N ettapiis, of
India— has developed this type of marking so that its
white neck is encircled by a narrow black ring, and the
Labrador duck (Camptorhynchus) has a nearly similar
mark (Fig. 57). Other ducks, e. g., the mallard, have the
white ring at the base of the neck, only.

In an interesting paper on the geese occurring in Oali-
fomia, Swarth (1913) has pointed out that in the cack-
ling goose {Branta c. minina) there is much variation in
the amount of white on the head and neck. Figs. 58 to
62 are traced from a series of photographs illustrating
this paper and show the throats of five specimens. The
wide range of variation in these specimens indicates to
my mind that this goose is in process of reducing the
neck patches, and thereby developing a white collar, such
as is present in the mallard, and perhaps also a white
throat. The usual condition seen in Branta canadensis
and in so-called normal specimens of B. c. minina is seen
in Fig. 58. The white cheeks have been developed long
ago in the history of the species, in part perhaps by the
depigmentation of the ear patches. Now a second change
is taking place in one of its subspecies. Thus in Figs. 59,

61 and 62, the neck patches have been reduced poste-
riorly, a varying amount in each case. In Figs. 60, 61 and

62 these patches have been restricted anteriorly pro-

Digitized by

Google

556

THE AMERICAN NATURALIST [Vol. XLVIII

ducing a white throat, and as sometimes in the pigeon,
imperfectly, so that a little island of pigment is cut off
just at the chin. It is also obvious from these figures,
that reduction may take place either at one end or tbe
other, or at both ends in different individuals. The ul*^^-
mate development of this line of reduction will prodix<^^

r8^

Figs.. 58-62.

(•

Vabiations in the Development of the Neck Patches in
Cackling Goose (after Swarth).

TH«

the narrow black collar seen in Nettapus previously rrx^xi-
tioned. It is worth noting also that in this goose tl-ie
limits of the neck patch are by their black color shax-i:>ly
defined posteriorly from the gray of the breast whic^lzt is
pigmented from the shoulder patches.

The Shoulder Patches. — The shoulder patches apE>^-ar
to center near the base of the wing, and in reduction
produce white remiges, such as appear in a domestics. tecJ
race of guinea fowl, as well as a white breast. The
domesticated guinea fowl often shows this white area, iri
the midline of the breast as the pigment areas fail to
spread ventrally. In the normal pattern of wild biirds,
however, white wings are seldom seen except among cer-
tain sea birds. White wing patches are often developed,
but tliese are frequently only bars on pigmented featla-eT*
as in the goat-suckers. Probably among small land toi^^
much white in the large wing feathers is a disadvantaj^

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 557

and so not much developed. It is noticeable that white
patches in the wing are often of such a nature that they
are concealed through the folding of the wings when the
bird is at rest. This accords with my belief that while
in flight the bird is unavoidably conspicuous by reason
of its motion, and that white patches showing at such
times add little or nothing to the disadvantage. In the
hairy and the downy woodpeckers (Dryobates), a white
stripe down the back is developed as part of the pattern,
and no doubt as in many mammals, marks the separation
between the pigment areas of opposite sides. Centrifugal
pigmentation is seen in some species as the kittiwake in
which the outer primaries are black.

The side patches are conmaonly continuous with those
of the shoulders, and when ventrally restricted, give a
white abdomen. Their median separation dorsally, is
seen in the hairy and downy woodpeckers as above noted.
I have not studied any special developments of these
areas, and they are commonly small.

The Rump Patches. — In birds as in mammals the two
rump patches pigment the posterior extremity of the
body. Their ultimate centers are dorsal and so close to-
gether that it is much less common for them to be sepa-
rated medially than to be restricted laterally. With a
slight areal reduction, a separation takes place between
them and the side patches dorsally, so that a white area
on the rump results. Often this white area represents
doubtless a slight restriction of both sets of pigment
patches which by drawing farther apart increase the
white area along the lower part of the back. In the
domestic pigeon much variation may be found, from a
condition in which the lower back is wholly pigmented
to one in which it is mostly white. The primary break
which causes this white patch has been much developed
in many groups of birds as a particular mark in the
pattern. In many species it is simply of a paler hue than
the surrounding parts as in the yellow-rumped warbler
iDendroica coronata) or the pine grosbeak (Pinicola).

Digitized by

Google

658 THE AMERICAN NATURALIST [Vol. XLVIII

In others the tendency to albinism thus expressed has
gone farther so that a pigmentless spot is formed. This
white rump patch is present in many unrelated groups
of birds in which it has independently arisen through
parallel development. Thus it is seen in many of the
smaller petrels, in the palm swift, the flicker woodpecker,
the white-rumped and other sandpipers, the white-
rumped shrike, the European house martin and others.
The tail feathers are pigmented by these patches, and
among various species show many steps in the process
of pigment reduction. As in the domestic pigeon, occa-
sional albinistic individuals show white outer tail
feathers, in accordance with the rule that the first pig-
ment reduction takes place at those parts of the primary
areas that are farthest removed from the pigment centers.
I have seen a white outer tail feather in wild specimens
of song sparrows and Lincoln ^s sparrow and it is occa-
sional in other species. In others again this mark has
become developed and fixed as a species character. Thus
in the bay-winged bunting (Pocecetesgramineus) there
is a single white outer feather on each side, in the junco
{J unco hy emails) there are two. A white central tail
feather is much rarer, but a pure white tail is found occa-
sionally as in the hummingbird, Leucuria phalerata, the
bald eagle and certain gulls, due to the permanent reduc-
tion of the pigment area of the rump at this extremity.
I once examined an albino ruffed grouse (Bonasa) which
was entirely white except for a single feather among the
upper tail coverts at the left side of the rump. This
blemish in the otherwise pure white bird seemed inexpli-
cable to those who examined it with me, but it merely
represents the last remnant of the left-hand rump patch,
still persisting though all the other pigment centers were
inactive.

It is very interesting that the white rump mark, so
commonly found in unrelated groups of birds, is one
which is conspicuous in flight only, and the same is true
of many of the white tail marks, such as outer white

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 559

feathers that disappear when the tail is shut. This points
to the conclusion that the development of a white mark
which is ever conspicuous is allowed in nature in such
cases only where it may be no detriment to the species
through rendering it too conspicuous by contrast. Thus
the bald eagle or the black-backed gull have nothing to
fear from such a banner mark. For small weak-flying
birds, however, the case may well be different. Yet even
these often show much white and I believe that it would
be possible for a species in its phylogeny to develop more
and more white if at the same time its habits of watchful-
ness or other actions developed equally to counteract any
disadvantageous result that might accompany the in-
crease. No doubt also a psychic factor is involved, com-
parable to what among ourselves we call '* fashion."
Thus a change in action or dress which departs too far
from the accustomed appearance is apt to be disliked at
first, though in time it may if persisted in, be tolerated
and at length accepted. In the development of white
markings, for example in the feathers of the tail, it
seems likely that a series of small steps must have been
made rather than too great and sudden changes. So in
the rock pigeon the white of the tail is limited to the outer
vane of the outer tail feather. In the turtle dove the
outer vane of the outer feather, and the entire tips of the
four outer feathers are white. The next step would be
to develop an entirely white outer feather and then two
(as in the passenger pigeon) and so on. In the sparrows
similar steps are shown by the lark sparrow {Chon-
destes) in which the tips only of the outer feathers are
white, the bay-winged bunting which has practically all
the outer feather white, and a little of the tip of the
second, the junco with two outer feathers and part of a
third white. No doubt steps such as these must have been
passed through by many white-tailed species.

It is difficult to say how disagreeable to their normally
colored neighbors, albino birds may be. I have seen an
albino robin in the fall of the year with a flock of other

Digitized by

Google

560 THE AMERICAN NATURALIST [Vol. XLVHI

robins and a white-spotted bee-eater with a flock of its
brethren, in both cases wholly at peace. This of course
was in flocking time when the social spirit is strong. The
song sparrow (Melospiza) with white outer tail feathers,
previously mentioned, was attacked and driven off by
another song sparrow. In the Journal of the Maine
Ornithological Society (Vol. 6, p. 48, 1904), C. H. Clark
writes of a pair of albino eave swallows (Petrochelidon
lunifrons)j at Lubec, Maine, .

among a large colony of the common ones who seemed greatly annoyed
at the albinos' presence and fought with them until they finally killed
one ... or rather injured it so badly that it died soon after.

I also have a note of a white robin at Montclair, N. J.,
which in early July, 1909, was seen to be much beaten and
driven about by another robin and eventually flew at full
speed against a tree and was killed.

Centrifugal, Colobation

In addition to the primary pigment patches which I
have discussed at some length, and the speckled condition
or ** English" marking, there is, as I have already inti-
mated, a third condition in which pigment is developed
at the extremities or points. It may be called a centrifugal
type and is almost the reverse of the centripetal or ^'pri-
mary-patch'' class.

The two latter types of pigmentation may both be
found in the same individual, but ordinarily this is not
evident except in cases where the primary patches are
somewhat restricted in area. It then may become appar-
ent that pigment is present at exactly those points where,
in the centripetal type of coloring, it is first to be lacking.
Moreover it persists strongly, even though the primary
areas are much reduced or largely absent. Curiously
this sort of pigment seems almost always to be black.
Apparently centrifugal pigmentation does not occur in
all species. I have never seen any trace of it in dogs.
In the house cat it is frequent, however. Thus in Figs.

Digitized by

Google

" I

No. 573] PATTERN DEVELOPMENT 561

18 and 19 it appears at the end of the tail. In the former
figure the sacral patches are mnch reduced, though pres-
ent, and together spread nearly half the length of the tail.
The terminal half, or less, of the tail, however, is dark-
pigmented, and a break occurs between the two sorts of
markings, due to the failure of the centripetal patch to
spread so as to unite with the centrifugal area. In Fig.

19 the sacral patches have wholly failed to develop but
the centrifugal patch still covers the distal half of the tail.
Possibly the dark heel marks in Fig. 16 are patches devel-
oped in the same way. In the house cat, a dark or
** smutty'' nose is often present in contrast to an other-
wise white face, or with the ear patches only slightly
reduced. In the breed of rabbits known as * * Himalayan, ' '
the centrifugal pigmentation remains, though the centri-
petal markings have disappeared, so that it is pure white
except for the black nose, ear tips and toes. No doubt,
however, it would be possible for the two types of pig-
mentation to appear in a single individual. This is sug-
gestive of the winter phase of the Arctic hares, in which
the black ear tips contrast strongly with the otherwise
white pelage. The physiology of the process whereby
certain animals acquire a white winter coat is not yet
fully worked out. It is curious that in occasional melan-
istic individuals of the eastern varying hare, the black
color is retained throughout the winter, instead of being
replaced by white — again a persistence of hlaclc pigment.
In dappled gray horses a black patch sometimes appears
on the bridge of the muzzle, usually the first place to show
white in the restriction of centipetal pigmentation. The
feet may also be black. Among certain antelopes a black
muzzle mark is similarly present, and in Hunter's ante-
lope {Damaliscus hunteri) a white border partly sur-
rounds such a mark. This, I believe, is due to a slight
restriction of the ear patches, sufiicient to prevent them
from reaching the muzzle, and of about the same nature
as seen in the blesbok {Damaliscus albifrons) in which,
through the absence of a centrifugal nose patch, the entire

Digitized by

Google

562 THE AMERICAN NATURALIST [Vol. XLVIII

front of the muzzle is white. The white chevron on the
muzzle of several antelope {Strepsiceros, Taurotragus)
is probably the result of a similar restriction of ear
patches combined with a centrifugal nose patch, leaving
a white line between. The black dorsal stripe seen in
many mammals and the black tail tip are probably mani-
festations of centrifugal pigmentation. The latter mark
is common in stoats (Mustela) and among those that
change to a white coat in winter, as the ermine, the tail
tip still remains black. In sundry other genera, as
Genetta, a black tail tip is part of the normal pattern.

In their paper on albinistic negroes, Simpson and
Castle (1913) published some highly interesting photo-
graphs of *' piebald" individuals. In four persons of one
negro family the hair over the median part of the head
from the occiput to forehead is pure white, as though due
to a restriction of the aural pigment patches. In addi-
tion, more or less of the median area of the back, as well
as the hands (including much of the forearms) and feet
(including the lower part of the ankle) are pigmented.
These latter areas may represent centrifugal pigmenta-
tion, but it should be noted that this is present in the
dermis. Possibly there is a close relation between dermal
pigment and that produced in the centrifugal style of
pigmentation.

Among birds, the black of the outer tail feathers of the
ptarmigan (Lagopus) may be comparable. A black area
is also sometimes present on the middle of the throat, or
as in certain gulls the outer primaries may be black.

This form of pigmentation is not found universally and
the conditions governing its appearance are unknown,
though its heredity in the ** Himalayan'' rabbit has been
somewhat studied by Professor Castle.

Summary

The principal points of this paper may be summed up
as follows :

1. In mammals and birds that normally are corn-

Digitized by

Google

N'O.STS] PATTERN DEVELOPMENT 663

pletely pigmented, there are certain definite points of
the body from which as centers the tendency to develop
pigment in the epidermal structures may become less
and less. Outward from each of these centers pigment
formation spreads to include very definite areas which in
wholly pigmented animals overlap slightly at their
borders or are at least contiguous.

2. A reduction in the area covered by any of these
primary patches results in a white mark at the line of
junction of two contiguous color patches, where no pig-
ment is produced. These white marks between the pri-
mary patches are spoken of as primary breaks.

3. Through a study of the breaks in pied individuals
of domesticated species of mammals and birds, the
boundaries of the primary patches have been determined.
These are homologous in the two groups and subject to
a certain amount of variation in different types. They
are: a median crown patch unpaired, and five paired
patches on the opposite sides of the body, which are
nanoed from the general areas they cover, the ear, neck,
shoulder, side and rump patches. Their limits are more
precisely defined imder the different species treated.

4. These patches are physiologically independent of
each other and may be differently colored in the same
individual.

5. Pied patterns among many wild species have been
brought about through the areal reduction of these pig-
ment patches in a definite way so that the white markings
resulting as breaks between the reduced patches have
become fixed and form a permanent part of the normal
pattern.

6. In several wild species this development of white
markings is shown to be even now taking place, but the
amount of pigment reduction is still fluctuating so that
the white markings vary much in extent with different
individuals.

7. The development of such white markings takes place
probably by little and little, so that the departure from

Digitized by

Google

564 THE AMERICAN NATURALIST [Vol. XLVHl

type is not so great as to arouse antagonism against the
varying individual on the part of others of its species.
Also, the gradualness of the change allows the species to
become accommodated to any disadvantage that might
concomitantly arise.

8. The converse of this centripetal style of pigmenta-
tion is present in many species, and results in pigmenta-
tion (commonly black) at the extremities or along lines
where primary breaks occur in the centripetal form,
namely at the tip of the nose, ears, tip of the tail or the
toes ; possibly the black dorsal stripe is due also to centri-
fugal pigmentation. Patterns may develop as in certain
antelopes by a white break between patches of the two
types.

In conclusion, I wish to express my indebtedness to
Professor W. E. Castle for much helpful criticism and
advice, and to the Museum of Comparative Zoology for
permission to make record of specimens .in its study
collection.

REFERENCE^

dg^f

Allen, G. M. , '

1904. The Heredity of Coat Color in Mice. ProA Amer. Acad. Arts and
Set., Vol. 40, pp. 61-163.
Brewer, W. H.

1882. On the Disposition of Color-markinglHIf Domestic Animals.
Proc. Amer. Assoc. Adv. 8ci., VoL 30, pp. 246-251.
Butler, A. W.

1888. Notes Concerning Albinism among Birds. Jour. Cincinnati Soe.

Nat. Hist., Vol. 10, pp. 214-216.
1888a. Albinos in the C\i\ier Club Collection. Jour. Cincinnati 8oc. Nat.
Hist., Vol. 10, pp. 216-217.
Castle, W. E. See MacCurdy, H. ; also Simpson, Q. I.
Cory, C. B. . ♦

1912. The Mammals of Illinois and Wisconsin. Field Mus. Nat. Hist,
Zool. Ser., Vol. 11.
C[arpenter], F. H.

1884. Some Phases of Albinism. Ornithologist and Oologist, Vol. 9, p. 48.
Coules, E.

1878. Melanism of Turdus migraiorius. BuU. Nutiall Om. Club, VoL 3,
pp. 47-48.

Digitized by

Google

No. 573] PATTERN DEVELOPMENT 665

Deane, B.

1876. Albinism and Melanism among North American Birds. BuU,

Nuitall Om. aub. Vol. 1, pp. 20-24.
1880. Additional Cases of Albinism and Melanism in North American
Birds. Bull. Nuttail Om. Club, Vol. 5, p. 25 (also 1879, pp. 26-
30, Vol. 4).
Faxon, W.

1913. Brewster's Warbler (Helminthophila leucohronchialis) a Hybrid

between the Golden-winged Warbler (Helminthophila chryaop-
tera) and the Blue-winged Warbler {HelminthophUa pinus).
Mem. Mu8. Camp. Zcol, Vol. 40, pp. 309-316.
Hoffman, W. J.
1878. Remarks upon Albinism in Several of Our Birds. Ameb. Nat.,
Vol. 12, pp. 474-476.
Keller, C. A.

1893. Evolution of the Colors of North American Land Birds. Occa-
sional Papers Calif. Acad. Sci., No. 3, xii + 361 pp., 19 pis.
Lawrence, G. N.

1889. Bemarks upon Abnormal Coloring of Plumage Observed in Several
-Species of Birds. Auk, Vol. 6, pp. 46-50.
Little, C. C. ^

1914. '^Dominant" and ''Recessive'* Spotting in Mice. Amee. Nat.,

Vol. 48, pp. 74-82.
M 'Galium, G. A.

1885. Albinism. AuTc, Vol. 2, pp. 113-114.
MacCurdy, H., and Caetle, W. E..

1907. Selection and Cross-breeding in Relation to the Inheritance of
Coat-pigments and Coat Patterns in Rats and Guinea-pigs.
Carnegie Inst. Washington, Publ. 70, iii + 50 pp., plate.
Mliller, S. H. C.

1862. Faerornes fuglefauna med bemaerkninger om fuglefangsten.
Vidensk. Meddelels. Copenhagen, 1862, pp. 1-78.
Pearl, R.

1914. On the Results of Inbreeding a Mendelian Population: a Correc-
tion and Extension of Previous Conclusions. Ameb. Nat., Vol.
48, pp. 57-62.
Pearl, B., and Boring, Alice M.

1914. Some Physiological Observations Regarding Plumage Patterns.
Science, New Ser., Vol. 39, pp. 143-144.
Pocock, B. I.

1907. On the Black-and-Tan Pattern of Domestic Dogs (Cants fa-

miliaris). Ann. Mag. Nat. Hist., Ser. 7, Vol. 19, pp. 192-194.
1909. On the Colors of Horses, Zebras, and Tapirs. An^. Mag. Nat.
Hist., Ser. 8, Vol. 4, pp. 404-415.
Bamaley, F.

1912. Mendelian Proportions and the Increase of Recessives. Ameb.

Nat., Vol. 46, pp. 344-351.
Simpson, Q. I., and Castle, W. E.

1913. A Family of Spotted Negroes. Ameb. Nat., Vol. 47, pp. 50-^6,

Figs. 1-4.

Digitized by

Google

566 THE AMERICAN NATURALIST [Vol. XLYin

Stone, W.

1912. The Phylogenetic Value of Color Characters in Birds. Jour, Acad.

Nat, Sci, Phila., Ser. 2, Vol. 15, pp. 311-319, pi. 27.
Strong, B. M.

1904. The Metallic Colors of Feathers from the Sides of the Neck of the

Domestic Pigeon. Mark Anniv. Vol., New York, pp. 263-277,
pi. 20.

1905. Causes of Blue and Green in Feathers. Biol, Bull., VoL 8, pp.

237-238.
Swarth, H. S.

1913. A Study of a Collection of Geese of the Branta ocmadensis Group

from the San Joaquin Valley, California. Univ. of Calif. Publ.,
Zool., Vol. 12, pp. 1-24, pL 1-2, 8 text-figs.
Sweet, D. A.

1907. Notes from Avon [Albinistic j uncos from Maine]. Jour, Maine

Omith. Soc, Vol. 9, p. 82.
Thayer, G., and A. H.

1909. Concealing Coloration in the Animal Kingdom. New York.
Townsend, C. H.

1883. Some Albinos in the Museum of the Philadelphia Academy. BuU.
Nuttall Om. Club, Vol. 8, p. 126.
Townsend, C. W.

1908. On the Status of Brewster's Warbler {Helminthophila leuco-

bronchialis). Auk, Vol. 25, pp. 65-68.
Ward, H. L.

1908. A Rapid Melanistic and Subsequent Partial Albinistic Change in
a Caged Robin. Bull. Wisconsin Nat, Hist. Soc, Vol. 6, pp.
43-47.
Worthen, C. K.

1897. Albinism, Melanism and Hybridism. Osprey, Vol. 1, pp. 23—24.

Digitized by

Google

SHORTER ARTICLES AND CORRESPONDENCE

THE BEARING OF THE SELECTION EXPERIMENTS

OP CASTLE' AND PHILLIPS ON THE

VARIABILITY OP GENES

Castle and Phillips have recently reviewed the results of six
years' work in which they selected for and against **hoodedness''
in rats.^ In ** hooded'* or ** piebald" rats only part of the coat
is pigmented; the area of dark (versus white) coat varies greatly
in different animals, but tends, in those of medium grade, to
cover the head, shoulders and middle of the back, like a hood.
Starting with a strain which was probably hybrid, although of
unknown ancestry, and selecting during thirteen generations
for a larger extent of colored coat (**plus" selection), they suc-
ceeded in obtaining animals with a greater and greater area of
pigmentation. The average, the mode, and the extremes were
raised. Conversely, selection for less pigmentation (''minus'*
selection) was accompanied by a gradual but decided and
continual dimunition in the dark area. ** Return" selection
also succeeded ; that is, plus selection was effective even in a line
which was already lighter than the average on account of a
previous minus selection, and, vice versa, minus selection caused
a lightening of a strain that had been made exceptionally dark
by a prior plus selection.

Certain crosses proved that more than one factor affecting
hoodedness is involved in the difference between the different
races. Therefore the production of animals of desired grade by
selection may perhaps be explained as a mere sorting out, into
different lines of descent, of different combinations of the various
factors for hoodedness originally present in the heterozygous
ancestors. It is the opinion of Castle and Phillips, however, that
this explanation will not suffice to account fully for the continued
eflficacy of selection in their experiments, and they believe it
probable that a factor or factors for hoodedness are undergoing
variation of a fluctuating nature.

1 Castle and Phillips, * * Piebald Bats and Selection, An experimental test
of the effectiveness of selection and of the theory of gametic purity in
Mendelian crosses. '* Published by the Carnegie Institution of Washing-
ton. See also Castle's "Pure Lines and Selection" in American Breeders'
Magazine, 1914.

567

Digitized by

Google

568 THE AMEEICMJ NATURALIST [Vol, XLVIII

A conclusion so radical and so opposed to previous work
should not be accepted, however, as long as it remains at all
reasonably possible to use instead an explanation in harmony
with the results of Johannsen and other investigators. Johann-
sen dealt with a character — dimensions of seed — ^which must be-
yond any doubt have been partially dependent upon a very great
many factors, yet he found that selection had no effect whatever
after he had separated the diflPerent genotypes from one another.
Thus he proved the constancy of a great many genes **at one
blow'* — namely, of all the genes appreciably concerned in seed
size. Of course, if there had been a chance for cross-fertilization
in his experiments, he, like Castle, would have obtained a result
from selection, but this would have been due to recombination,
not variation, of genes. All our evidence points to the conclusion
that the vast majority of genes are extremely constant, although
they differ somewhat in that very slight amount of variation
which they do show. For example, in Drosophila, although in
the case of most genes not more than one mutation has been
found, yet in one case (possibly in two or three cases) a locus has
mutated three times, each time in a different way, thus giving
rise to a system of multiple allelomorphs containing four mem-
bers. This gene evidently is more subject to mutation than the
others, yet this formation of a series of multiple allelomorphs can
not even remotely be compared to fluctuating variability, for the
three mutations were all large steps (much smaller could easily
have been detected) , and they were found only during the exami-
nation of some millions of individuals in the rest of which the
locus, was not observed to mutate at all. Some few genes are
known, however, which really do change frequently (e. g., that
for ** variegated" corn), but these cases are extremely rare;
moreover, here the degree and nature of the change are fixed,
and also, after the change has once occurred the instability of the
gene is lost. Thus, in no known case do the variations of a gene
among, let us say, several thousand immediate descendants of the
individual passessing it, form a probability curve, as neo-Dar-
winians might perhaps suppose, nor even are any cases known
where genes can undergo frequent changes that may vary at all
in kind or amount or occur successively.

Let us then inquire into the probability and adequacy of that
explanation of Castle and Phillips's results which does not require
the assumption that a gene or genes involved change compara-

Digitized by

Google

No. 573] SHORTER ARTICLES AND CORRESPONDENCE 669

lively frequently and successively, but which assumes a sort-
ing out of numerous factors. It is now pretty generally ac-
cepted by Mendelians that the germ plasm of any of the higher
organisms contains a large number of genes, which play vari-
ous roles in the numberless processes and reactions of devel-
opment whereby the egg is transformed into the adult indi-
vidual. The exact nature and intensity of any one characteristic of
this adult organism (e. ^., hoodednessin rats) is dependent upon
the nature of each of the various reactions which were involved
in producing this character, and thus dependent upon all the
genes (and environmental factors also) involved in any of
those reactions. Now, in an ordinary Mendelian cross, all the
individuals are usually homozygous and alike in respect to all
but one of the pairs of genes that noticeably affect the character
concerned. In such a case, then (so far as differences in environ-
mental influences do not obscure the outcome), one obtains the
simple Mendelian results derived from the segregation, at reduc-
tion, and recombination, at fertilization, of but this one pair of
allelomorphs.

The strain of hooded rats, however, was probably a hybrid
between two races of rather remote relationship. When two such
races are crossed, the individuals often differ in more than one
pair of those factors that affect the character studied, especially
if the character is such as to be influenced by a relatively large
number of genes. It can not be questioned that some characters
are thus determined or influenced by a much larger number of
developmental reactions than are others, and such characters
will therefore vary more in inheritance, since if a difference
exists between two individuals in respect to any given gene,
these characters are more likely to be affected than others. Gross
size, for example, is a character dependent in this way upon an
exceptionally large number of genes, for any gene which influ-
ences the size of any organ must affect to some extent the total
size. In some other cases in which characters are found to be
influenced by relatively many genes, the reason for this is not
so evident, e, g.y in the case of the red flower-color of flax, or the
truncated condition of the wing in some races of Drosophila,
Here the production of the character may be conceived to be
dependent upon some reaction that can be easily modified by
various means.^ For our present purpose we must assume that

2 It is conceivable that differences in respect to numerous genes have
sometimes arisen even in the case of characters not naturally very easily

Digitized by

Google

570 THE AMERICAN NATURALIST [Vol. XLVHI

the character **hoodedness" belongs in this class and that the
ancestral hooded rats used by Castle and Phillips were the de-
scendants of a cross involving many genes for that character.

The results of such a cross are of course complicated, for the
different pairs of allelomorphs generally can undergo recombina-
tion at the reduction division of the hybrid, so that in P, or
subsequent generations as many different genetic types of indi-
viduals are formed as there are possible different combinations
of those factors wherein the ancestors differed. Not all these
genetic types, of course, will fall into different phenotypes, yet
generally there will be a large number of overlapping pheno-
types among the progeny.

The larger the number of factors in which the two ancestral
lines differed, the larger will be the number of different passible
combinations of these factors, and accordingly the smaller will
be the chance of any individual having one of those particular
combinations necessary to a relatively high or a relatively low
intensity of the character. In other words, the larger the num-
ber of factors (for one character) for which a population is
heterogeneous, the more numerous are the possible different
grades of intensity of this character funong the different indi-
viduals, but the fewer will be the individuals which approach the
more extreme grades theoretically possible in such a population.*
Suppose, for example, that two parents differ in five pairs of
factors for hoodedness, which are partially dominant* to their
allelomorphs and summative in their action. Then in Pj not one

influenced by diverse means, merely because one of the two races had been
subjected to a very long and drastic selection, so that any of those rare
mutations which affected that character in the desired direction had in this
race been preserved. Selection in such a case, however, would have to in-
volve many millions of individuals.

8 One extreme, e. g., the **plus,'» will be rather frequent, however, if all the
**plus" factors dominate completely. But in the case of the hooded rats
we must assume either that dominance is generally incomplete or that in
the case of some factors the *' minus'' allelomorph dominates in the case
of others the "plus," since F, rats from a cross of the plus by the minus
strain are on the average intermediate in type between these two extremes.

* It is of course by no means necessary to assume incomplete dominance
of the factors. If dominance is complete (in some cases the "minus" fac-
tor may dominate, in others the "plus"), the rigor of selection wUl be di-
minished, since heterozygous forms can not be distinguished from homozy-
gous. Therefore, although a somewhat greater number of individuals wiU be
found having the limiting values, it will take longer to bring the average up
to the limit.

Digitized by

Google

No. 573] SHORTER ARTICLES AND CORRESPONDENCE 571

individual in a thousand will have the most extreme dark or
light grade of hoodedness possible. However, by selecting the
more extreme individuals, and mating them together, a still more
extreme grade of hoodedness may be obtained in F3 (both as to
average and limiting values), and the same process may be con-
tinued for a good many generations. The number of generations
during which effective selection is possible depends on the num-
ber of factors concerned, the rigor of selection, and the amount
of inbreeding of brother to sister.

In regard to the latter point, since brother and sister are much
more apt to be alike in their genetic constitution than are other
individuals, oflFspring from such a mating are more apt to be
homozygous and alike, or, we may say, such offspring will tend to
be homozygous and alike in a larger number of factors; then,
mating two individuals homozygous for these factors together,
there will be much less variation and so less opportunity to con-
tinue selection among their progeny. In the case of Castle and
Phillips's experiments, however, no such attempt at inbreeding
was reported. Here, then, the individuals mated together would
be more apt to differ genetically, even though they looked alike
(thus, one might be A A bb, the other aA bB), and their
descendants would therefore present a larger number of different
combinations of factors for the selector. Often a greater effect
may be eventually produced in this manner than by inbreeding,
for a larger number of combinations of factors are thus pro-
duced, some of which may be of more extreme type. The effect
would usually be slower, however, since such matings tend to
keep the strain heterozygous and are often steps backwards.
Cross-breeding, then, will help to explain the relatively slow but
long-continued and eventually large effect of selection in Castle
and Phillips's experiments, although such a result could also be
obtained without cross-breeding if the factors were numerous
enough.

The ''return selections" also are easily explicable on the
multiple factor view. Due to the original difference in so many
factors, and the fact that cross-breeding diminishes the tendency
to homozygosis which selection favors, the rats were presumably
heterozygous even after generations of selection. They would
not be as heterozygous as before, of course, and, correspondingly.
Castle and Phillips did find less variation in the rats after selec-
tion. Yet there would still be a good chance for recombination,

Digitized by

Google

572 THE AMEBIC AN NATURALIST [Vol. XLVIII

and an alteration in the race could therefore be produced by
further selection or by return selection. As we have seen, this is
especially true if certain factors are completely dominant, al-
though dominance is by no means a necessary condition.

As a very simple illustration, let us suppose that the **plus"
factors A and B dominate over the ** minus" factors **a" and
**b," respectively, and each increase the pigmented area to about
the same extent. To begin with, two moderately hooded indi-
viduals, Aa bb and aa Bb, were mated together. They produced
laa bb — light-hooded, laa Bb and lAa bb — both moderate, and
lAaBb — dark. We first select for dark; mating the dark
rats together, 9 darks, 6 moderates, and 1 light, would be pro-
duced (Fg). The average color of the oflEspring has thus been
increased by selection (the limiting color, too, if dominance is
incomplete). It can be still further increased in subsequent
generations. On the other hand, the color can be made lighter
again by a ** return selection,'* for if, instead of mating the F^ or
F3 darks together, we mate the moderates or mate darks with
moderates, many of the matings will give offspring lighter, on the
average, than in the preceding generation ; e. g., Aa Bb by Aa bb
gives 3 dark, 4 moderate, 1 light, as compared with the previous
9 dark, 6 moderate, 1 light. In subsequent generations, the
average could be brought still lower.

Let us now see whether there is any experimental evidence
in support of the multiple factor explanation of Castle and
Phillips's results, aside from the fact that it is adequate and is
the only one consistent with other work. One point of evidence
we have noted — the variability of the rats continued to decrease
as a result of selection in either direction. This we should of
course expect on the multiple factor view, for selection gradually
tends towards homogeneity in a population, even though it may
require a long time to produce complete homogeneity. The
second and strongest evidence is from crosses.

The crosses show that one of the factors concerned in differ-
entiating hooded rats from wild rats, which are pigmented all
over, or from ** Irish" rats, which are almost completely pig-
mented, is ** hypostatic." In other words, a rat having the
normal allelomorphs of this factor will always be self-colored, or
nearly so; one having the other allelomorphs will always be
distinctly hooded, although the amount of the hoodedness varies.
**Self," as it happens, is dominant, in this case, over hooded.

Digitized by

Google

No. 573] SHORTER ARTICLES AND CORRESPONDENCE 573

Thus, on crossing a hooded to a wild or Irish rat, all the F^ are
self (or nearly so) ; in Fj there are three selfa to one hooded, but
the hoodeds vary in intensity. The question then is, does this
variation (so far as it is not due to **environmentar' differences)
depend upon what other "epistatic'' or ** modifying" factors for
hoodedness may or may not be present, or is there evidence that
it depends instead, or in addition, upon a variability of one or
more of the factors for hoodedness ? As will be shown below, it
can be proved that different combinations of modifying factors
do occur in the different hooded indiviuals : this being true, there
can be no ground for making the unusual postulate that in this
case or in the selection experiments a factor or factors concerned
undergo variation.

The proof is that when light hooded rats from the minus
strain are crossed to wild or Irish rats the hooded rats in Fj vary
much more than did the original strain of hooded rats and aver-
age much darker. Obviously, the Pi hooded rats differed from
the wild or Irish in a number of modifiers as well as in the hypo-
static factor; moreover, as we should have expected, this differ-
ence consisted chiefly in the fact that the wild or Irish rats con-
tained **plus'' allelomorphs in place of some of the ''minus''
modifiers present in the Pi strain that had undergone minus
selection. Thus the Fj hooded rats, containing various combina-
tions of these modifying factors wherein the two strains differed,
varied much more than did the parental strain of hooded rats,
and were on the average much darker.

In order to escape this conclusion that modifying factors were
involved, Castle and Phillips at first postulated that the reason
that the Fj hooded were darker than the original ** minus" strain
was because the factor for hooded had in many cases become con-
taminated by its allelomorph j;the factor for self) in the Fi rats.
This is violating one of the most fundamental principles of
genetics — the non-mixing of factors — in order to support a vio-
lation of another fundamental principle — the constancy of fac-
tors. The refutation of their supposition came unexpectedly
soon. It would be expected, on the view of multiple factors, that
the wild or Irish rats (containing the allelomorph for self in
place of the hypostatic factor for hooded) would not possess as
many ** minus" modifiers as the hooded strain which had been
specially selected to contain as many of these as possible ; neither
would these **self " rats contain as many **plus" modifiers as the

Digitized by

Google

574 THE AMERICAN NATURALIST [Vol. XLVIII

hooded strain which had undergone plus selection (and which so
contained nearly all of the plus modifiers originally present in
either the self or the hooded ancestors) . Thus it was to be ex-
pected that, just as a cross of self with the minus race gave Fj
hooded rats darker than the original minus strain, so a cross of
wild or Irish rats with hoodeds resulting from the plus selection
would give Fj hooded rats lighter than those of the plus strain.
This result was actually obtained. It was fatal to the idea that
the difference between the Pi strain of hooded rats and the F^
hoodeds was due to contamination of the allelomorph for hooded
with that for self, since such contamination should have resulted
in Fg hooded rats darker than those of Pj, not lighter. For wild
and Irish rats are both much more extensively pigmented than
hoodeds even of the plus strain.

The change in hoodedness from Pi to Fj was therefore due
to recombinations of the modifying factors wherein the two
strains differed. That many such modifiers were concerned is
indicated by the evenly distributed variability of the Fj hoodeds
and the fact that very few were as extreme as the hooded grand-
parents. The same fact is brought out in a cross of the minus
with the plus race ; here no clear-cut ratios were obtainable, the
classification into different genotypes being rendered impossible
by the multiplicity of factors (no one of which was hypostatic
as in the other crosses). Of course, this knowledge of so many-
factors being concerned in the crosses helps our interpretation
of the selection results decidedly, for the more numerous are the
factors concerned, the longer would it be possible to continue an
effective selection on the progeny of the hybrids, and the oripr-
inal hooded rats of the selection experiments were admittedly in
all likelihood descended from just such hybrids. The exact num-
ber and effect of the different factors can not be determined from
Castle and Phillip's data, since to do this very special crosses
must be made and individual pedigrees kept. Selection experi-
ments can be of little value so long as there are factors for which
the individuals may be heterozygous, unless these factors can be
accurately followed in inheritance.

Of course, it is quite possible that in the course of these long-
continued experiments mutations affecting the hoodedness occa-
sionally happened to arise, especially since it seems likely that
this character is dependent upon an unusually large number of
genes, for then, as a matter of mere chance, any mutation which

Digitized by

Google

No. 573] SHORTER ARTICLES AND CORRESPONDENCE 575

occurred would be more likely to affect it than it would be to
affect most characters. It is interesting to note that one such
mutation, of a very marked and unquestionable character, was
in fact observed. The mutant factor proved to be a strong
**plus" modifier, which was almost completely dominant, and
itself showed no contamination or variation, so far as could
be determined. It arose, as it happened, in the plus strain.
A part of the effectiveness of selection may therefore have been
due to the occurrence and sorting out of such occasional muta-
tions, but there is no way of telling how many of these took place,
or any need for assuming them at all in explaining the result.
These rare mutations, however, would form a very different phe-
nomenon from such fluctuating or frequent and progressive vari-
ation of a gene or genes concerned as Castle postulates. Although
the academic possibility of variation of the latter type can not
be denied, there is no experimental evidence which can be used
to support it, and there is good evidence against it in many
individual cases.

It is diflftcult to believe that this suggestion of Castle and
Phillips was not made in a spirit of mysticism, when we con-
sider also their suggestion that the genes may undergo contami-
nation, and especially when we consider the following passage,
with which their paper concludes :

It seems to us quite improbable that the plus mutation could have
arisen in the minus selection series. We believe that the repeated se-
lection which was practised had something to do with inducing this
change in the plus direction. If one can increase at will the " modi-
fiers " which make the pigmentation more extensive, it does not seem
strange that after a time a readjustment should occur within the cell
which should incorporate modifiers in that part of the cell which is re-
sponsible for the unit-character behavior of the hooded pattern. This
would amount to a quantitative change in the unit-character for hooded
pigmentation.

To thus suppose that independent genes fuse or induce changes
in one another, merely because they happen to produce similar
end effects upon the organism, and in spite of the fact that they
usually lie in different chromosomes and are apt to differ from
each other as much as do other genes, is utterly teleological.

A paper by A. L. and A. C. Hagedoorn criticizing Castle's work
and conclusions, appeared at the same time as the paper of

Digitized by

Google

576

THE AMERICAN NATURALIST [Vol. XLVUI

Castle and Phillips,^ The Hagedoorns champion the multiple
factor hypothesis as an explanation of Castle's results, and also
cite certain rather inconclusive experiments of their own to sup-
port this point of view. They err, however, in supposing that
the factors concerned must be incompletely dominant; as we
have seen, this is not a necessary assumption, if we admit that
in the case of some modifiers the ** minus'' allelomorph dominates,
in others the **plus." They also err in denying the possibility,
on the multiple factor view, of successful *' return selection," if
inbreeding be strictly followed. In fact they offer this as a test
of their point of view. As we have seen, ** return selection'*
would be possible in some cases, even if the animals were inbred;
and in Castle and Phillips's experiments, where inbreeding was
not followed, ** return selection" was certainly very effective.

Finally, papers have recently appeared by MacDowell,* in
which he gives evidence that certain other cases of inheritance
{e. g,, head size in rabbits), formerly considered by Castle to
support the idea of genie variation and contamination, are
probably best interpreted on the view of multiple factors instead.
His evidence consists in the fact that the characters concerned
are somewhat more variable in the offspring of back-crosses than
in Pj, as we should expect on the basis of recombination of
multiple factors, but which he believes could not plausibly be
explained otherwise.

Hermann J. Muluer

« A. L. & A. C. Haagedorn, ' ' Studies on Variation and Selection, ' ' Zeit, f.
ind, Abst. u, Verdb,, 1914.

oE. C. MacDowell, "Multiple Factors in Mendelian Inheritance," Jour,
Exp. Zool, 1914, and Carnegie Inst, of Wash., 1914.

Digitized by

Google

Two Important Biological Journals

The Journal of Qenetic3

Edited by
W. BATESON and R C. PUNNETT
A periodioal for the pablioation of records of original reaearoh in heredity, variation, and allied topics.
The JoTTBJTAL also contains, from time to time, articles snmmarizing the existing state of knowledge in the
▼arions branches of genetics. The numbers are folly illustrated, and, when the subject-matter demands it,
colored plates are freely employed. The Joubnal is issued in parts as material accumulates, and a volume
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7.50 a volume ;
■i^le oopies, $2.50.

3iometrik:ci

Edited by KAEL PEABSON
A Jonmal for the statistical study of biological problems, appearing about four times a year. A volume
oontaining about 500 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
niiie preceding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
cnuiiometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, 52.50.

Prices on back numbers of these journals vary, and pos^e from London is charged oa back volumes
and single oopies.

Published in America for the Cambridge University Press of England by

Xhe University of Ohicagfo Rress
Chleasro^ Illinois

University Control

By J. MoKebk Oattbll, Professor of Psychology in Oolnmbia University

Tofstliar with a series of Two Hundred and Ninety-nine Unsigned Letters by Leading Meo of ScisoM
hfiMfag Aeadeokio Positions and Articles by Josbph Jasvbow, Qbobgb T. Ladd; Johv J. Sravnnov, J* B.
CaaiwTua, J. MoKunr Gattbll, Qwmam M. Scbatvov, Stbwabt Patov, Jobs Jat GHAncAH, Jamwb P.
linmoB and Jaoqb Gouu> Sc&ubmah.

A great variety of questions eonoemlng general uniyenlty administration ara dealt with in an original and
belpfol way w— iy^alurs.

These quotations and examples are taken from Proleesoi Cattail's informed and thorough diaenasion of the
■objeet of nrnverflity oontrol, a sabjeot opon which he has had much to say of late, finding occasion for oaustie
^fitiotsm of existing American oon^tions, and standing as the ohaminon of an academic demooracy and a taaeh-
ing profession upon which a man may enter without forfeiting hii Belf-respeot. — 2^ Dial,

Sentences and paragraphs that betoken the eocpert, highly-trained mind, the suggestions tiiat come to m-
Ijredi a.d teU us that a new day is about to dawn in educational writing. — Th* Bottom Svmiao TfomtcHfL

SCIBIVCB AND EDUCATION

A, meries of rolnnteBfor the promotion ofBcienU&o reBearcb and educational progresM
VOLUME L The Fonndatioss of Science. By H. PonroABi. Containing the authorised EngjUsh translation

hy Qmowam Bsuoa Halstbd of "Science and Hypothesis," "The Value of Science,'! and f Science

and Method." Price, t3.00 net.
TOLT7ME n. Medical Research and Education. By Biobabd M. PlABon, Wzluam H. Wblgb, W. H.

HowsbL, Fwunodjs P. 1/Ull, LnwaLLTS F. Babxbb, ChabiiU S. Minot, W. B. Cahhoh, W. T.

ComraiLifAV, Thbobald Siovb, G. N. Stbwabt, G. M. Jaoxson, & P. Lton, Jaicsb B. HirawTflg, Jobs

M. DoDflov, C. R. BABDnnR, W. Opnthji, a J. Vtmuraai, Jamss Ewino, W. W. Esnir, Hbsbt H. Dosaad-

SQV, CBBmriAM A. HanrBB, and Hutbt P. Bowditgb. Price, t3.00 net.
TOLUMB in. UniTSfstty Control. If aw Btadw^ Pooet s+^^ PKoe. tSM Mt

QARtusofi, N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The American Naturalist

A HoBthlr Joamal. MUblidied in 1867, Devoted to Um ildvuiMmMt «f Um Biolosieal grtwrM
with SpMaid RtiTmmcm to the Factors of Orcaaie EvolotiM and HwStr

CONTENTS OF THE MARCH NUMBER

The Effect of Extent of Dlftributtoi/on SpeeifttloiL
Asa C. Chandler. i

Biology of the Thysanoptera. Dr. A. Franklin Shall.
Shorter Artioles and Correspondence : The Endemic

Mammals of the Britishlslesr Professor T. D. A.

CockerelL

Notes and Literature : Swingle on Variation in Fi
Citnifl Hybrids and the Theary of Zygotaxis.
Dr. Orland E. White.

CONTENTS OF THE APRIL NUMBER

The Origin of X Capeella Bursa pastoiis araohnoldea.

Dr. Henri Hus.
Biology of the Thysanoptera. XL Dr. A. VrankUn

Shall.
Shorter Aitielei end Discussion: Barriers as to Dla-

tribntion as regards Birds and Mammals. JoMph

Orinnell. Yellow Varieties of Bats. rroflMsoc

W.E. Castle.
Notes and Literature: Hcrsdlty «ftd the Influence

ofMonarchs. V. L. K.

CONTENTS OF THE MAY NUMBER

Eetoparasitesof Mammals. Professor Vernon Lyman

Kellogg.
Begeneration, Variation and Correlation in Thyone.

Professor John W. Scott.
Shorter Articles and Discussion: Terms relating to

Generic Typeif Dr. O.F.Cook.
Notes and Literature: Linkage in the Silkworm

Moth. A. H. SturteTant. Naboun's Breeding

BzperimentiwithQrMshoppers. John a Dexter.

CONTENTS OF THE JUNE NUMBER

Spedes-lKiilding by Hybridisation and MutadaL Pro-
fessor John H. Geroold.

Heredity of Bristles in the Common Greeobotfle Pty~
A Stady of Factors goremingDistiibiition. Phineaa
W. Whiting.

Physiological Correlations and Climatic 'Rcarttoiis la
AlfUik Breeding. Gea F. Freeman.

Taxonomy and Evolution. By Z.

fiborter Artides and Discosslon* NaboorsPs Cirsis
hoppers, Multiple Alldomorphlsm, Tankage and
Misleading Terminologies in Qenetifis.
W.R Castle.

CONTENTS OF THE JULY NUMBER

Pattern Development in Mammals and Birds.
GloTer M Allen

Internal Belations of Terrestrial Assoeiations.
Arthur G.VestaL

Shorter Articles snA Discussion* Another Hy-
pothesis toAecounI for Dz Swingle's Experi*
menti with Citnu* A. C. and A. L. Hsgedoom*

CONTENTS OF THE AUGUST NUMBER

Multiple Allelomorphs in Mice. Prtrfessor T. H*
Morgan.

Thirteen Years of Wheat Selection. T. B. Hatch-
eson.

Pibttem DeTelopment in Mammals and Birds.
Glover M. Allen.

The Meadow Jumpiof Mousa. Dr. H. L. Babeock.

Shorter Articles and Inseussion : _ _

Studies on Inbreeding. Dr. Baymond PieaiL
Parallel Mutations in (Enothera Mamtt L. De;
J. Stomps. Dr. Bradley M. Davis. The T
ical Distinction between Multiple AUe..^
and Close Linkage. Professor T.H.Mofgitt.
fessor W. E. Castle.

Notes and Literature:

Biometrics. Dr. Raymond Peazt. A New Mod^
of Segregation in Gregory's Tetrapioid Prinndae.
Hermann J. Muller.

Single Number 40 Cents

The NATURALIST wfll IM I

Yeaily SulMcription, $4.00

for four mo&tha for Om Dolbr

THE SCIENCE PRESS

Garrison, N. Y.

Sab.StatM» 84: NEW YORK

Digitized by

Google ^

roi. ZLTin, 10. (74 OCTOBEK, 1U4

THE

AMERICAN
NATURALIST

A MOHTHLT JOTTBHAL

Devoted to fhe Adyanoement of the Biological Soienooi with

Special Beference to the Factors of Evolution

COHTESTS

Page
I. Bex-limited and Sex-Unked Inberltanoe. Professor T. H. Morgan - - 67T

n. UUierltanoe of Bndoiperm Texture In Sweet x Waxy Hybrida of Malie. G. N.

Collins and J. H. Kempton --------- 684

zn. ▲ Study of yarlation In the Apple. W. J. Younq - - - - - 696

IV* Sliorter Article! and Dlaensalon : Variation and Correlation in the Mean Age
at Marriage of Men and Women. Dr. J. Abthtjb Habris, Boxana H.
YiVL/LN. Duplicate Genes. Sbwall Wright ------ eso

V. Hotea and Literature : A Study of Desert Vegetation. Professor Chablbs £.

BBSSBT ------------- 6*9

THE 80IEN0E PEE88

{■AEOABTEB, PA. OABEI80N, V. T.

HEW TOEK: SUB-STATION 84

Digitized by

Google

The American Naturalist

MSS. intended for pnblication and books, etc., intended for review should be
sent to the Editor of THE AMERICAN NATURALIST, Garrieon-on-Hndeon, New York.

Short articles containino ennnaries of research work bearing on the
prohlems of oroanic evolntion are especially welcome, and will be given preference
in publication.

One hundrea reprints of contributions are supplied to authors f^e of charge.
Further reprints will be supplied at cost.

Subscriptions and advertisements should be sent to the publishers. The
subscription price Is four dollars a year. Foreign postage is fifty cents and
Canadian postage twenty-flve cents additional. Tne charge for single copies Is
forty cents. The advertising rates are Four Dollars for a page.

THE SCIENCE PRESS

Lancaster, Pa. Qarrison, N. Y.

NEW YORK : Sub-Station 84

Bntarad ai aeoond-dan matter, April 2, 1908, at the Post Office at Lanoafter, Pa., under the A«t of

GongrcBs of March 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Partloulars of

G. DINESEN, Bird Collector

Husavik, North Iceland, Via Leidle, Engrland

WANTED TO PURCHASE

a set of BIRDS OF AMERICA by J. J. Audaboo.
7 or 8 volumes, please report, stating cash price, stat-
ing conditicm, binding and dates of volumes.

P. C. HARRIS,
Box 2244 Boston, Massachusetts

For Sale Entire

An important collection of Indian Birds'
Eggs, containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste i
viridanus). Red-browed Finch (Calla- |
canthis burtoni), Tibetan Sand-Grouse \
(Syrrhaptes tibetanus). Spotted Sand- •
Grouse (Pteroclurus senegallus). Ibis-
bill (Ibidorhynchus struthersi). White-
bellied Minivet (Pericrocotus erythro-
pygius). Bearded Vulture (Gypaetus
barbatus), niunerous Cuckoos with their
hosts, etc., etc.

Particulars and full list may be had
from

W. F. H. ROSENBERG

57 Haverstock HiU London, N. W.

Marine Biological Laboratory

Woods Hole, BAass.

INVESTIGATION
Eatif Ymmr

INSTRUCTION

My 1 to A««ast 11,
1914

FRoflities for rewsreh in Zoslsgy.
Bmbryolocy* Phjsioloc7 sad Bo^
say. Seventy-abc private Isbora-
torW $100 eaoh for not ovw ^ns
months. Thirty tablea aza a^aB-
able for beghmera in rosesroh vbo
dadre to work under the dtiwiioii
of membera of the staff. Ibe fM
f or raoh a tabto k $5aoa

Couim of laborstofy
with leotaree are offered in L
brata Zoolocf , BmbfyoloeT;
oloc7> MorphotofT and Taa
oftLe AW^hyaiolocy aaa
Boologj of Mariae Straad aad B«f
yegetatioa. Each eoorae laqofaee
the foU time of the B^idflnL F^
$50.00. A lecture oouxae ob &•
Philoeoplikal Aspects oC f
•ad Allied Sdencea is also (

and

SUPPLY
DEPARTMENT ^i^JS^S^tSTtju^ u
Opaa Ike Eatlra Yaar animals and of Algae, Panclf livar-
worts and Mosses fom' '

dasswork, or for tiie

living material furoiribed in i ..
as ordered. Price Usta of Zolk
logical and Botanical material
sent oa appHoation. State whatk
is denred. For price 1 sts aad al
information regarding

QBO. M. GRAY, Ciuator, Wooda Hole. 1

The annnal annotmoement will be sent oa appUeaftioa to

The Director, Marina Biolegical Laboratory, Woada Hela

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVIII October, 19U No. 574

SEX -LIMITED AND SEX-LINKED INHERITANCE

PROFESSOR T. H. MORGAN
Columbia Univbrsity

Darwin used the expression * inheritance as limited by
sex'' to include all cases in which a character is peculiar
to one sex. His list of such cases covers in the main the
group of secondary sexual characters. Darwin's expres-
sion has been contracted to sex-limited inheritance, and is
widely employed to-day in the same general sense in which
Darwin used the expression. For instance, Bateson in
his book *^ Mendel's Principles of Heredity" includes
both horns in sheep and color blindness in man as sex-
limited characters.^

Now that the inheritance of several of these cases has
been definitely worked out, it has become increasingly evi-
dent that such characters as color blindness, and haemophi-
lia in man, the twenty-five ^'sex-linked" characters in Dro-
scphila, and certain characters in birds and in butterflies
follow a law of inheritance that is essentially different
from that followed by some of the other cases. It has
become necessary, therefore, to recognize two groups of
cases that differ fundamentally in regard to their heredity.
To one of these groups I have applied the term sex-linked
inheritance, and, for the present at least, we may still make
use of the older expression sex-limited inheritance (and

1 See pp. 169-174 in section headed ''Heredity Limited bj Sex; the Horns
of Sheep/' where the term sex inheritance limited descent (p. 172) also
appears.

577

Digitized by

Google

578 THE AMERICAN NATURALIST [Vol. XLVHI

sex-limited character) to cover that class of cases (obvi-
ously a very mixed one which will be broken up as our
knowledge regarding it becomes more certain) that in-
cludes largely, as originally intended, the secondary sexual
characters.^ In those cases of sex-linked inheritance, in
which the male is heterozygous for the sex factor, the
grandfather transmits his peculiarity, through his daugh-
ters, to half of his grandsons only; and reciprocally an
affected female transmits her peculiarity to all her sons,
and, through her sons bred to her daughters, to half of
her granddaughters and to half of her grandsons^. More-
over the appearance of the character in the female is not
exceptional or abnormal, as is sometimes implied in cases
like color blindness in man, for, the character can always
be transferred from the male to the female by suitable
crosses.

On the other hand, there are cases in which a character
appears in one sex only— the character is limited, there-
fore, to the male or to the female. Such cases may be
properly called sex-limited, and were so called by Darwin.
As typical examples I may cite the horns of certain races
of sheep that are present in the ram and absent in the

2 G. H. Shull has recently said {Zeit, Ind, Ahst, und Vererh., XU, 1914,
p. 160) that, in his opinion, it would be better to retain the term sex-limited
for those cases that I call sex-linked and call other cases secondary sexual
characters. This view is not historically in accord with Darwin's usage of
the term ' ' limited by sex. ' ' This fact, in itself would be a sufficient argu-
ment for rejecting ShulPs suggestion, but, in addition, the term sex limited
is an actual misnomer for the class of cases to which he proposes to apply it.
There are cases like the eosin eye of DrosophUa that differ in the male and
female in the same way as do many secondary sexual characters (in fact they
are such in a descriptive sense) but nevertheless show sex-linked inheri-
tance. Since a new name is required to express our fuller information in
regard to some of the characters that were originally included under the
older term, why not begin by adopting suitable and expressive ones.

8 In those cases in which the female is heterozygous for a sex factor, as
in birds and in butterflies, the same principle is involved but the sequence is,
in a sense, reversed; thus the grandmother transmits, through her sons, her
peculiarity to half of her granddaughters; and reciprocally, the affected
male transmits his peculiarity to all of his daughters, and, through his
daughters bred to his sons, to half of his grandsons and to half of his grand-
daughters.

Digitized by

Google

No. 574] INHERITANCE 579

ewe (or else more developed in the ram than in the ewe) ;
the color of butterflies like Papilio Memnon, with three
types of females ; and the dark spot on the abdomen of the
male of the bug Euchistus variolarins. These characters
can not be transferred through the gametes to the female
of their own race by any known combination.

Whether one likes or does not like the particular terms
used to denote these two classes of cases, the fact remains
that there are two such categories, and to ignore their
existence is only to make obscure a distinction that is per-
fectly plain.

Concerning the mechanism involved there is something
more that may be said. It has been suflSciently shown in
the case of sex-linked inheritance that the sex-linked char-
acter follows the known distribution of the sex chromo-
somes. It is imnecessary to repeat here the abundant
evidence in support of this statement. The simplest inter-
pretation of this known relation is that the character ia
dependent for its realization on the sex chromosomes. I
do not mean, of course, that the sex chromosomes alone
produce the character but that something in these chromo-
somes, some ^^ factor,'' acting in conjunction with the rest,
of the cell, conditions the character.

On the other hand, in the case of sex-limited characters:
the facts can not be explained on the assumption that the
characters follow the sex chromosomes. It is clear that
they do not do so. But we can give a consistent interpre-
tation of the facts if we assume that sex-limited characters
follow the distribution of the ordinary chromosomes.

Since this relation has recently been not understood
and misinterpreted I may be pardoned, I hope, for taking
up the question once more.

Wood crossed homed Dorset sheep with hornless Suf-
folks. The sons had horns, the daughters lacked them.
Inbred these gave in the Fg generation— homed c?, 3^
hornless c?, 1; homed $, 1; hornless % 3. Bateson and
Punnett have shown that the results are explicable on the
basis that one factor for horns in the male produces

Digitized by

Google

580 TEE AMERICAN NATURALIST [Vol. XLVIII

horns but one factor is insuflScient in the females. This
conclusion was put to the test by breeding an Fj hornless
ewe to a hornless ram. The Fj ewe should be hetero-
zygous for the factor for horns, and, therefore, when she
is bred to a homozygous hornless ram, half of her off-
spring should be heterozygous for homlessness and half
homozygous for homlessness. Since half of her sons
should have a factor for horns they are expected to
develop horns, and this is what occurred. Half oi the
daughters also should have a factor for horns, but should
not develop horns, and this also was true.

It has been recognized for several years that this and
related cases can not be explained on the assumption that
the factors involved are carried by the X or by the Y
chromosomes. But we can interpret the statement that
one factor for horns is sufficient in the males to call forth
horns, but not sufficient in the female ^*in terms of chromo-
somes,'' if a factor for horns is carried by one of the
•chromosomes other than the sex chromosome. In other
wrords we need only appeal to a mechanism with which we
are familiar to cover the results.

The second illustration is furnished by the recent
experiments of Foot and Strobell, and since the authors
have rejected the chromosome hypothesis as inapplicable
to their results, and since in the case of insects the condi-
tions are simplified because castration experiments have
shown that the sex glands are not themselves responsible
for the secondary sexual characters, we may profitably
consider this case even more fully.

In one of the bugs, Eiichistus variolari/us, the male has
a black spot on the abdomen. The female lacks the spot.
A female of this species was crossed to a male of another
species, viz., Euchistus servus, having no spot in eitiier
sex. The daughters had no spot, the sons had a spot
fainter than that of variolarius. Inbred these gave, in F„
249 females without a spot, 107 males with a spot (devel-
oped to different degrees) and 84 males without a spot
The Fi results show that one factor for spot in the male

Digitized by

Google

No. 574] INHERITANCE 581

suflSces to call forth in some degree the spot in the hybrid.
Its intensity varies from a condition approaching that in
pure variolcurius to a faint spot (possibly even to no spot
at all). The Y^ results show also that a single factor in
the female fails to cause the spot to develop in that sex.
In the Fi male the failure of the spot to reach in most
cases its full development shows obviously that the same
conditions that produce a male that is perfect so far as
his sex gonad is concerned, do not suflSce to cause the full
development of the spot, although the factor for the spot
is present in one dose at least. The only confusion that
is liable to arise is that in none of the Fg females did the
spot appear, although in some of them there must have
been a double dose of spot. But the difficulty is imaginary
as a little thought will show. In the first place the female
of E. variolarius herself does not show the spot, yet this
female must have a double dose of spot if spot is in the
X chromosome or in any other chromosome {except the Y).
Foot and Strobell by an elaborate analysis of the case
show that the factor can not be carried by either the X or
the Y chromosome. It is unnecessary to repeat their
argument ; for, if the factor were carried by the X chromo-
some, only half of the grandsons should show it, while, in
fact, many more than half of them show it; and it could
not be carried by the Y chromosome because the Y chromo-
some of variolarius is not present in the female, hence
could not have entered the cross as made. We are con-
cerned then only with a third possibility, viz., that there is
something in the female condition itself that is inimical to
the development of the spot. Since neither X nor Y
carries the factor in question it must be present in duplex
in the female of variolarius {if every gamete mu^t have it
in simplex and the e:fperiment shows that this is the case),
cmd since the spot does not show in the female of vario-
larius, it is obvious that it can not appear in thai sex even
in duplex. If it be grafted that the character is like other
Mendelian characters, cmd the authors' evidence show thai
it is inherited as are Mendelian characters, the conclusion

Digitized by

Google

582 THE AMERICAN NATURALIST [VoL.XLVni

is self evident; for, in demonstrating thM all of the
gametes of variolarius carry spot the authors actually
destroy their own argument.

It only remains to point out some of the different ways
in which a factor being present in duplex both in the male
and in the female produces its effect only in the male. In
some cases it has been shown that the ovary produces
some substance that is inimical to the production of cer-
tain characters. For instance in fowls and in ducks the
presence of the ovary suppresses the development of
the male plumage. That the factors for the male plumage
are present is shown by its development when the ovary
is removed. But in some insects it has been found that
neither the ovary nor the testis produces these kinds of
substances; for, when the testis or the ovary is removed
the secondary sexual characters are not affected. Here
the mode of explanation must be different But the con-
ditions, or complex, or factors that produce the ovary in
the female are acting in every cell of the body, and con-
sequently an effect, that is indirectly caused in the fowl
or duck, might be directly caused in the insect. For, each
cell is a chemical factory. Such a factory may help to
produce an ovary and the ovary produce a substance that
demonstrably suppresses the male plumage, or the same
kind of factory may do similar work through the activity
of some other part of the body, or conceivably it may do
its work in every cell of the body. This it seems to me is
the most reasonable view to take of the matter in the case
of the variolarius-servus cross. "We can express the same
thought in symbols by representing the female of vario-
larius by XXAABBCCDDSS, etc., and the male by
XTAABBCCDDSS, etc. The chemical interaction be-
tween two X 's and the rest of the cell is of such kind that
it produces a female, and the female complex, as such, is
inimical to the development of a spot and favorable for
the development of the accessory organs of reproduction
and of all secondary sexual characters of the female, while
XY and the rest of the cell is inimical to the development

Digitized by

Google

No. 574] INHERITANCE 583

of the acessory organs and of the secondary sexual char-
acters of the female, and favorable for the development of
the accessory sexual organs and of the secondary sexual
organs of the male. This view is of course compatible
with the idea that there may be special factors for these
organs in chromosomes other than the sex chromosomes,
and the view holds both in a general way and on the
special chromosome hypothesis as well.

To assume that all the factors for characters that are
shown by the male or by the female must be carried by a
sex chromosome of some kind, if carried at all by chromo-
somes, is a travesty of the point of view of those who hold
to the chromosome hypothesis as a reasonable working
hypothesis to account for Mendelian inheritance. Just
as it has been shown that there are factors in the sex
chromosomes that affect many parts of the body, that are
not concerned with differences of sex; so, on the other
hand, the evidence shows that there are factors in other
chromosomes that are influential in producing secondary
sexual characters.

Digitized by

Google

INHERITANCE OF ENDOSPERM TEXTURE IN
SWEET X WAXY HYBRIDS OF MAIZE

G. K COLLINS AND J. H. KEMPTON
U. S. Department of Agricui/tubb

Introduction

In a previous publication,^ the first and second genera-
tion of crosses between sweet and waxy varieties of maize
were reported and a tentative explanation of their be-
havior was suggested. It is now possible to add the
results of the third season, which to some extent afford a
test of the explanation proposed in our first publication.

The immediate result of crosses between the Chinese
variety of maize having a waxy endosperm and varieties
with sweet endosperm was the production of seeds having
a horny endosperm indistinguishable from that of ordi-
nary field varieties of maize. In the second xenia genera-
tion all three kinds of endosperm reappeared in the pro-
portion of 9.20 horny, 3.95 sweet, and 2.85 waxy. This
ratio was accepted as a 9:4:3 dihybrid ratio. For al-
though the deviations of the individual ears, individual
families and the totals were too large to be ascribed to
chance, the deviations were not consistently in one direc-
tion and to predicate more complicated formulae would
have necessitated different assumptions for different ears.
The only interest in treating the problem in this way would
be that of solving a mathematical puzzle, for it would be
practically imi>ossible to secure individuals enough to test
adequately the validity of the assumptions which it would
have been necessary to make.

Admitting, then, that the ratios were only an approxi-
mation representing a general tendency, it became of

1 Collins, Q. N. and Kempton, J. H., ' * Inheritance of Waxy Endosperm in
Hybrids with Sweet Corn," Circular 120, U. S. Department of Agriculture,
Bureau of Plant Industry, 1913.

684

Digitized by

Google

No. 574]

INHERITANCE

685

interest to learn whether predictions were still possible.
For the purpose of maMng comparisons easy, the
origuial diagram representing the second xenia genera-
tion is here repeated. (See Fig. 1.) The meaning of the
symbols is as follows : S is the factor for sweet, and X the
factor for waxy. When both S and X are present the seed
is expected to be horny. Small letters indicate the absence
or latency of the factors.

5a 5x sX

mmr\

Msm

5X^

m[£r

5X*^5?(

«X

urn

WXf

fHom

Sim

ta.

sx
sx

Fig. 1. Diagram showing the gametic composition of second-generation hybrids
between waxy and sweet varieties of maize.

Since in both sweet, and waxy the alternative factor
necessary to produce horny is assumed to be lacking, the
gametes produced by sweet varieties are represented by
Sx and the gametes produced by varieties with waxy endo-
sperm by sX. The synthetic horny produced by crossing
waxy and sweet is then represented by a combination of

Digitized by

Google

586 THE AMERICAN NATURALIST [Vol. XLVIH

these, or SxsX. Assuming a chance recombination of
these factors in the gametes derived from these synthetic
horny seeds, the gametes will be of four kinds. Both the
sweet and the waxy may be present (SX) or the sweet
may be present without the waxy {Sx)j or the waxy with-
out the sweet (5X), or both may be absent (sx). At ferti-
lization each of these kinds of gametes may unite with
any one of the four corresponding kinds derived from the
other parent, producing 16 zygotic combinations. In the
diagram the four classes of gametes from one parent are
given in the horizontal row at the top, and the same four
classes from the other parent in the vertical row at the
left. Each gametic combination from the top is repeated
four times in the squares below, while each combination
at the side occurs four times in the corresponding hori-
zontal row of squares. Thus each of the squares repre-
sents the result obtained by combining the gametes repre-
senting the horizontal and vertical rows that intersect at
that point. In all cases where both S and X occur together
the seed should be homy, where only S occours the seed
should be sweet, when only X occurs it should be waxy,
and in one square (No. 16), where neither S nor X occurs
there is a new combination which the results have shown
to be a new type of sweet seed, indistinguishable from
ordinary sweet seed but behaving differently when crossed
with other types of endosperm.

In accordance with the above analysis the expected re-
sults in the third xenia generation were as follows :

Self -pollinated horny.

Self-pollinated sweet.
Self -pollinated waxy.

Proportion-
ate No. of
Ears.

Proportions
of Seed
Classes.

1
2
2
4

All horny
3 horny: 1 sweet
3 horny: 1 waxy
9 homy: 4 sweet:

All sweet

1
2

All waxy

3 waxy: 1 sweet

3 waxy

Digitized by

Google

No. 574]

INHERITANCE

587

Crosses between different plants
from homy seeds.

Crosses between different plants
from sweet seeds.

Crosses between different plants
from waxy seeds.

Crosses between homy and sweet.

Crosses between horny and waxy.

Crosses between sweet and waxy.

25
20
20
16

5
4

3
6
1
2
2
4

1
1
2
2
2
4

All homy

3 homy: 1 s'weet

3 homy: 1 waxy

9 homy: 4 sweet: 3 waxy

AH sweet

All waxy
3 waxy: 1

sweet

All horny

1 horny: 1 sweet

1 homy: 1 waxy

3 horny: 1 waxy

1 horny: 2 sweet: 1 waxy

3 homy: 4 sweet: 1 waxy

All homy

3 horny: 1 sweet

1 homy: 1 waxy

3 horny: 2 sweet: 3 waxy

All homy

All waxy

1 homy: 1 sweet

1 homy: 1 waxy

1 sweet: 1 waxy

1 homy: 2 sweet: 1 waxy

Third Xenia Oeneration

Four of the ears bearing second xenia generation seed
were selected for planting in 1913, one self- and one cross-
I>ollinated ear from each of the two hybrid families Dh 216
and Dh 221. These families were selected because in 1913
the family Dh 221 showed the greatest deficiency of sweet
seeds and Dh 216 was the only family that showed sweet
seeds in excess of the expected. The three classes of
seeds from each of the ears were planted separately.

Unfortunately as the result of an accident crosses were
not made between the plants grown from the different
classes, but a total of 77 selfed ears were obtained, a num-

Digitized by

Google

588 THE AMERICAN NATURALIST [VoL.XLVin

ber sufScient to indicate whether the initial assumption
regarding the gametic compositions was of value in
arranging the observed facts.

Pkogeny of Sweet Seeds

Sweet seeds were assumed to result from squares 6, 8,
14 and 16. It will be seen that in none of these is there
any factor other than 8 and since the absence of both
factors, as in square 16, is also assumed to produce sweet,
we should expect nothing but all sweet ears from self-
pollinated plants grown from sweet seeds.

Seventeen self -pollinated ears were secured from plants
grown from sweet seeds. All the seeds of these ears were
sweet with the exception of one waxy seed. This one waxy
seed was colored and since it occurred on an ear from a
white sweet seed that otherwise produced only white
sweet seeds, the exception may reasonably be ascribed to
accidental foreign jwllen.

Pkogeny of Waxy Seeds

Waxy seeds were assumed to have resulted from the
combinations shown in squares 11, 12 and 15. Seeds from
square 11 should produce only waxy seeds. Squares 12
and 15 should produce ears with waxy and sweet seeds
in proportion of 3 waxy to 1 sweet. There should, there-
fore, be one all waxy ear to two with both waxy and sweet
seeds. There were in all 29 ears from waxy seeds, 11 of
which were all waxy and 18 with both waxy and sweet
seeds. The numbers are small but at least both kinds of
ears were secured and the proportion does not violate the
original assumption. The 18 ears with both waxy and
sweet seeds all produced them in approximately the 3:1
ratio. The numbers are given in Table I. The totals with
3,154 seeds indicate that if there is a deviation, it is almost
certainly less than 2 per cent.

All the sweet seeds that occur on ears grown from waxy
seeds are assumed to belong to the new class of sweet
seeds corresponding to that represented in square 16.

Digitized by

Google

No. 574]

INHERITANCE

589

Plantings of such seeds axe being made for comparison
with the ordinary class of sweet seeds having the same
ancestry. These are represented by the sweet seeds occur-
ring on ears having homy and sweet seeds.

TABLE I

Waxy Seeds SELr-poLLiNATED. Ears Showing Waxy and Sweet Seeds.

Expected: 25 Per Cent. Sweet

Parent Ear

PediRree
Number

Total
No.

No.
Waxy

No.
Sweet

Per Cent, of
Sweet Seeds

Devla-

tiou -1-

, Prob.

Seeds-

Seeds

Error

[1938

301

216

28.2 =b 1.7

+1.9

1939

112

24.1 ± 2.7

- .3

Dh 216-1

(Crosa-Pollinated)

1940

264

202

23.5 =b i:8

- .8

1942
1943

18
349

14
258

4
91

22.2 ± 6.6
26.1 =b 1.6

- .4
+ .7

1949

149

100

32.9 ± 2.6

+3.0

1950

138

103

25.4 ± 2.5

- .2

1972

389

302

22.4 ± 1.4

-1.9

1973

187

136

27.3 ± 2.2

+1.0

Dh 216-2

(Self-PoUinated)

1974

174

138

20.7 ± 2.1

-2.0

1975
1976

85
34

67
21

18
13

21.2 ± 3.0
38.2 =fc 5.6

-1.3

+2.4

1977

313

232

25.9 ± 1.7

+ .5

1978

109

27.5 =b 2.9

+ .9

ri994

136

105

22.8 ± 2.4

- .9

Dh 221-2

1 1995

155

116

25.2 =fc 2.3

, + -1

(Self-Pollinated)

1 1996

39.2 =t 4.6

1 +3.1

1 1997

190

146

23.2 ± 2.1

1 - .9

Total...

3,154

2,351

803

25.5 =fc .5

+1.0

Pkogeny of Horny Seeds
From the horny seeds the expected results are more
complicated. They may be tabulated as follows :

1 ear (Square 1) with seeds all horny

2 ears (Squares 2 and 5) with seeds 3 homy: 1 sweet
2 ears (Squares 3 and 9) with seeds 3 horny: 1 waxy

4 ears (Squares 4, 7, 10 and 13) with seeds 9 horny: 4 sweet: 3 waxy.

Ears were, therefore, expected in the projwrtion of 1 all
homy ear, 2 with homy and sweet seeds, 2 with homy
and waxy seeds and 4 with all three classes. Thirty ears
were secured from seed classed as homy. These ears
were distributed as follows : 1 all homy, 5 with homy and

Digitized by

Google

590 THE AMERICAN NATURALIST [Vol. XLVm

sweet seeds, 3 with homy and waxy seeds, 19 with horny,
sweet and waxy seeds and 2 all sweet.

The two all sweet ears are entirely outside the expected.
Their appearance may be explained on the assumi>tion
that seeds classed as homy in 1912 were in reality sweet
No microscopical examination of the starch was made
and the seeds were classified on their appearance,
wrinkled seeds being classed as sweet and smooth seeds
as homy. The separation of homy from sweet seeds is
more difficult to make than waxy from either homy or
sweet.2

There were, however, very few doubtful seeds in the
second xenia generation and in suggesting this interpre-
tation, we may with some propriety be accused of attempt-
ing to explain away ^^ green balls.*'*

The two all sweet ears were descendants of an ear Dh
221-2, which showed an excess of homy seeds and a
deficiency of sweet. The expected number of sweet seeds
in Dh 221-2, which had a total of 493 seeds, was 123 and
only 106 were classified as sweet. If this deviation re-
sulted from a faulty classification, that is, if some of the
sweet seeds failed to show the characteristic wrinkled
exterior, we might exi)ect that about 17 of the 300 seeds
classed as homy would produce ears with all sweet seeds.
Eleven of the ears secured from homy seeds in 1913 were
descendents of this ear.

The remaining 28 ears from horny seeds are distributed
among the 3 classes in reasonably close agreement to the
expected. Measured by Pearson's formula for the good-
ness of fit,^ it appears that such a deviation might be
expected once in about twenty times.

2 The difficulty of distinguishing between sweet and starchy seeds in
crosses where the starchy variety has small seeds has been pointed out by
East and Hays, ''Inheritance in Maize/' Bull. 167, Ck)nn. Ag. Exp. Sta.,
1911, p. 40.

» Pearson, K., and Heron, D., "On Theories of Association," BiometrQca,
IX, pp. 309-314.

-*P/n7. Mag,, Vol. L, 1900, pp. 157-175. The application of Pearson's
formula to data of this kind was called to our attentioki by Mr. G. Udney
Yule.

Digitized by

Google

No. 574]

INHERITANCE

591

The three ears with homy and waxy seeds produced
these classes in the expected 3 : 1 ratio. The numbers are
given in Table 11,

TABLE n

Horny Seeds Selp-pollinated. Eaes Showing Horny and Waxy Seeds.

Expected: 25 Per Cent. Waxy

Parent Ear

Pedigree ToUl

No.
Horny
Seeds

No.
Waxy
Seeds

, Deria-
Per Cent, of | tion +
Waxy Seeds Prob.

1 Error

Dh 216-2

(Self-PoUinated)
Dh 221-2

(Self-PolUnated)

1962 1 327
/2000 312
\ 2007 121

247

235

80
77
39

24.5 ± 1.6 ' .3
24.7 =b 1.6 .2
32.2 =b 2.9 2.5

Total... 760 i 564 1 196

25.8 ±1.1 .7

Four of the five ears that produced horny and sweet
seeds were also as close as could be expected to the 3 : 1
ratio. The fifth, however, Ped. 1965, with 249 seeds, had
only 19 sweet seeds or 7.6 per cent. The numbers are
given in Table m. The only explanation that can be

TABLE ni
Horny Seeds Self-pollinated. Ears Showing Horny and Sweet
Expected: 25 Per C?ent. Sweet

Seeds.

Parent Ear

Pedigree
Number

ToUl

No.

Seeds

No.
Horny
Seeds

No.
Sweet
Seeds

Per Cent, of
Sweet Seeds

Devia-
tion-*.

Prob.

Error

Dh 216-1

(Cro8&.Pollinated)

Dh 216-2

(Self-PoUinated)

Dh 221-1

(Cross-PolUnated)

Dh 221-2

(Self-PoUinated)

1965
1979

1988

/2003
\2008

249
442

160

175
179

230
344

121

134
141

19
98

41
38

7.6 ± 1.1
22.2 ± 1.3

24.4 ± 2.3

23.4 ± 2.1
21.2 ± 2.1

16.0
2.2

.8
1.8

j Total... 1 1,205 970 ' 235

19.5 di .8

6.9

offered in connection with this exceptional ear is that
suggested for the occurrence of the two all sweet ears
among those grown from seeds classed as homy, namely,
the existence of sweet seeds which failed to show a

Digitized by

Google

592 THE AMERICAN NATURALIST [Vol. XLVin

wrinkled surface. This explanation is rendered less prob-
able, however, by the unusual behavior of the aleurone
color in this same ear. In the previous discussion the
aleurone color has not been considered. To treat of the
aleurone color would naturally lead to the question of
correlation between that character and endosperm texture,
a subject which in these crosses is very complicated and
for the treatment of which the results thus far obtained
are inadequate. It may be said, however, that with the
exception of Ped. 1965 the proportions of colored to white
seeds in all the ears bear out the assumption that the in-
heritance of the aleurone color is governed by two factors,
both of which must be present to produce color. In Ped.
1965, however, which was grown from a colored seed, only
23 of the 249 seeds were white. The colored and white
seeds are beautifully distinct with no intermediate or
doubtful seeds. The ratio of 9.2 per cent, white might be
explained as an approximation to the dihybrid ratio of
6.25 per cent, but we must then admit that instead of both
factors being necessary for the development of color
either factor alone may produce color.

The 19 ears from horny seeds that showed all three
classes are assumed to have the same gametic composi-
tion as the original second xenia generation, previously
reported. The numbers are given in Table IV. The last
column of the table gives the odds in 1,000 that deviations
equal to those observed are not chance deviations from the
expected proportions, as calculated by Pearson's formula.
Thus in Pedigree 1953 the odds are 809 to 191, or practi-
cally 4 to 1, that the deviation is not the result of chance.

As in the original ears, the approximation is sufficiently
close to render futile any attempt to predicate a different
arrangement of factors, but many of the deviations are
too large to be ascribed to chance. In the totals the sweet
class is too low and the waxy too high, in fact there is no
significant difference between the totals for these two
classes. The deviation from the expected is, however,

Digitized by

Google

No. 574]

INHERITANCE

593

largely the result of two ears Ped. 1954 and 1967, and if
the explanation suggested for the two all sweet ears from
horny seeds is admitted, it may also account for the devia-
tion in these two ears. In both ears the deficiency of
sweet seeds is accompanied by an excess of homy seeds,
while in neither ear is there a significant excess of waxy
seeds.

TABLE IV

HoBNY Seeds Self-pollinated. Eabs Showing all Thbee Glasses.

Expected: 56.25 Per Cent. Hobnt, 25 Pee Cent. Sweet

18.75 Per Cent. Waxy

Pedi-
gree

Total
No.
Seeds

Horny

Seeds

Sweet Seeds

Waxy Seeds

Cbancei
Id 1000

that the
Devia-
tion
is not
Accl-
denUl

Parent Ear

No. Ex-
pected

No. Ob-
served

No. Ex-
pected

No. Ob-
served

No. Ex-
pected

No. Ob-
served

1953

809

1954

360

197

236

999-1-

1955

198

111

103

470

Dh 216-1 1

(Cross- ,

PoUinated)'

1956

777

1957

148

664

1958

176

100

133

1963

540

304

289

135

133

101

118

832

1964

1 39

983

1966

170

1 ^

101

524

,1967

158

1 «^

108

995

Dh 216-2 '

(Self- 1

Pollinated)' 1980

486

Dh 221-1 1

ri985

258

1 145

131

784

(Cross- ,

1986

120

' 67

443

Pollinated)-

1987

375

211

201

716

[ f 1999

' 43

676

Dh 221-2 1

2001

. 15

(Self. 1

' 2002

118

i 66

426

PoUinated)!

2004

134

' 75

455

,2009

1 44

979

Total

3,141

1,767

1.800

785

681

589

660

999.99

With these two ears excluded the deviation in the total
for the remaining 17 ears may be ascribed to chance.
Tested by Pearson *s formula such deviations might be
expected once in about 50 times.

Digitized by

Google

594 THE AMEBIC AN NATURALIST [Vol. XLVHI

Conclusions

The immediate (xenia) result of crossing varieties of
maize having sweet and waxy endosperm was the produc-
tion of seeds with a homy endosperm resembling that of
ordinary field varieties. In the second xenia generation
all the ears contained seeds of the three classes, sweet,
waxy and horny, in fairly definite ratios. The data were
arranged in accordance with the Mendelian formula corre-
sponding most nearly to the observed numbers.

The third generation, like the second, gave results suffi-
ciently close to dihybrid ratios to render unprofitable the
assumption of more complicated ratios. There are, how-
ever, deviations from the expected numbers of too great
magnitude to be ascribed to chance.

The ratios of waxy to non-waxy seeds were regular as
far as the conditions of the experiment could determine,
except for a slight excess in the number of waxy seeds in
nearly all the ears in which all three classes appeared
(Table IV). A deviation in number of waxy seeds as
large as that shown in the total would not be expected to
occur as the result of chance more often than once in one
thousand times.

The ratios between sweet and homy, while approxi-
mating the predicted ratios, show numerous irregularities.
Wherever there is a significant deviation in the number
of sweet seeds, the observed number is below the expected.
Eeasons are advanced for believing that the deficiency of
the sweet class may result from a failure of some sweet
seeds to develop a wrinkled exterior rather than from any
irregularities in segregation.

The results show the value of representing the char-
acters by gametic factors. This method provides an
orderly arrangement of the facts of heredity thus far
observed with respect to these characters and makes pos-
sible fairly accurate predictions regarding the genetic be-
havior of the various seed classes.

Washington, D. C,
July, 1914

Digitized by

Google

A STUDY OF VARIATION IN THE APPLE
W. J. YOUNG

Assistant Horticulturist, Washington Experiment Station

As a rule the subject of variation in the several char-
acters of the apple has been given but incidental attention,
and that usually in connection with the study of other
problems. As a result the literature on the subject is of
a fragmentary character consisting usually of a few ob-
servations here and there in papers dealing with other
subjects.

It is perhaps worth while to note a few of the investi-
gations which have thrown some light in an incidental
way upon the causes of variation in apples. In fertilizer
tests which were made at the New York Geneva station^
and elsewhere, no well-defined and uniform influence of
the various elements of plant food upon the color could
be detected, though the New York station reports more
decided results in seasons when the natural conditions
were unfavorable to the development of highly colored
fruit. In the comparison of tillage and sod mulch in an
apple orchard, also conducted by the New York Q-eneva
station,^ it was found that the fruit from an orchard in
sod was more highly colored and matured one to three
weeks earlier than that from the tilled plot, though the
latter was better in quality and kept four weeks longer in
common storage. The influence of the stock upon the
character of the fruit is a matter of much obscurity, the
investigation of which presents such difficulties that it
has received little attention. The effect of pollination
also is still far from settled. It was thought at one time
that the characters of the fruit were profoundly modified
by the pollen received by the blossom. The data on this

iBiiU. 289.
« Bull. 314.

595

Digitized by

Google

596 THE AMEBIC AN NATURALIST [Vol. XLMH

subject have been collected by Munson,^ who found that
evidence that the pollen has any direct effect upon the
fruit is largely lacking. Aside, then, from indirectly
modifying the size of the fruit, the influence of the pollen,
in so far as our present knowledge goes, may be left out
of account in a study of apple variation.

Without doubt the most noteworthy contributions to
the knowledge of apple variation are the recent papers
by Shaw, of the Massachusetts station, and Stewart of the
Pennsylvania station. Shaw's first paper, which ap-
peared in the Massachusetts station report for 1910, deals
entirely with the variation of the Ben Davis apple. In
comparing specimens grown in a number of widely sepa-
rated localities it was noted that variations due to eU-
matic condition were strongly marked and affected prac-
tically all characters of the fruit. Modifications of form
were especially noticeable The depth of coloration was
looked upon as correlated with latitude, being pink in the
specimens from Arkansas and deep crimson in those
grown farther north. The amount of overcolor seemed
to be controlled by local conditions. The color was espe-
<;ially good in the apples from the Pacific coast and those
from Colorado, Pennsylvania, and Indiana. In a given
orchard temperature appears to be the most influential
factor governing size. The flesh was notably white in the
fruit from Colorado. The apples from Colorado and
California were less firm than those from other localities.
The southern-grown specimens were more juicy and of
better quality than those from the north, which were apt
to be dry, hard, flat, and sometimes astringent. It ap-
pears that a mean temperature of at least 60° F. for the
growing season is required for the satisfactory produc-
tion of the Ben Davis. The poor quality of the northern-
grown specimens is apparently due to a lack of sufficient
heat to properly develop the fruit.

In Shaw's second paper in the Massachusetts station
report for 1911 the fact is emphasized that the grower

»Me. Sta. Rept. (1892), pp. 29-32.

Digitized by

Google

No. 574] VARIATION IN THE APPLE 597

should choose those varieties which he can grow to the
highest degree of perfection under his conditions of soil
and climate. The causes of variation are summarized,
giving special attention to the influence of temperature as
a factor in the distribution of apple varieties. The north-
em limit is regarded as fixed by the lowest temperature
which the tree will stand, while the effect of summer
heat upon the development of the fruit is looked upon as
limiting the distribution southward. The elongation
of the fruit was found to be correlated with a low tem-
perature for two or three weeks after blooming. A low
summer temperature produces greater acidity, higher
content of insoluble solids, greater astringency, smaller
size, and scalding in storage. The extent of coloration
was regarded as decreasing from the center of distribu-
tion in passing either north or south, while the intensity of
coloration was considered greatest in high latitudes and
altitudes. Excessive summer heat results in uneven
ripening, premature dropping, rotting on the tree, poor
keeping quality, lack of flavor, mealiness, less intense
color, and smaller size. For each variety there is a mean
summer temperature at which it reaches its highest de-
velopment.

It will be noted that Shaw's method of investigating
the problem consisted in securing fruit for comparison
from widely separated localities and attempting to corre-
late the various characters with the conditions of produc-
tion. Stewart, on the contrary, confined his study to
apples grown in one locality and noted the effect of modi-
fying one at a time those factors within his control. This
is the more scientific method of procedure, but has the dis-
advantage that the variations are far less striking and a
smaller number of factors can be studied. An account of
Stewart's experiments and the results so far attained is
found in the reports of the Pennsylvania station since
1907. These papers deal largely with the effect of fer«
tilizers and different cultural methods on the yield, color,
size and growth of the apple. The various factors influ-

Digitized by

Google

598 THE AMERICAN NATURALIST [VoL.XLVin

encing these characters are enumerated and the results
are given of the studies made of them. It is noted that
the factors are so interrelated that the best conditions for
producing one effect are often injurious in some other
direction and that the chief problem in orchard manage-
ment is a proper balance of the various factors. An
** optimum principle " is recognized, according to which
plant growth and development increase as the most dis-
tant essential factors approach the optimum. The factor
farthest from the optimum, therefore, whether below oi
above, may control the results from a crop

Outline of the Expekiment

Since the season of 1912 was one of full crop in nearly
all centers of apple production, conditions were especially
favorable for the study of variation in this fruit. The
writer accordingly obtained specimens for study and com-
parison grown in a number of localities under quite dis-
similar conditions. The method employed was therefore
that of Shaw, as pointed out in the last paragraph, rather
than that of Stewart. The study has been pretty largely
confined to Washington-grown apples, though a few have
been obtained for purposes of comparison from the east
and middle west. The formal investigation of the prob-
lem has been carried on but a single season, which is en-
tirely too brief a study to demonstrate conclusively all
points touched upon. The conclusions reached, however,
are strongly supported by many observations in various
localities extending over a number of seasons, and are so
suggestive of further lines of study as to justify a report
at this time.

In carrying on this investigation the aim has been to
secure as much information as possible regarding the con-
ditions under which the fruit was grown. The endeavor
has been to get into communication with the growers and
obtain from them through correspondence data regarding
the character of the soil, rainfall, irrigation, elevation,
exposure, temperature, age of trees, fertilization and

Digitized by

Google

No. 574] VARIATION IN THE APPLE 599

cultivation. The chief line of observation had to do with
the variations which occur in the different samples of the
same variety as obtained from different sources. To get
at this side of the problem, careful observations were
made as to the condition of the apples, and their various
characters were recorded in a complete technical descrip-
tion of each sample for the purpose of making a compara-
tive study of the samples of the several varieties. In ad-
dition to this written description, photographs were made
showing typical specimens in various positions and when
cut in cross and longitudinal sections. In general it may
be said that variations are found in the form, size, color,
internal structure, texture, flavor, quality, specific grav-
ity, chemical composition, time of ripening, and keeping
quality. The attempt is made to correlate these char-
acters with the conditions of growth in so far as they are
known and to work out the law of the relation of environ-
mental factors to the characters of the fruit.

The following apples were made use of in the study:
Arkansas, one sample; Arkansas Black, one sample;
Baldwin, eight samples; Ben Davis, nine samples; Deli-
cious, three samples; Esopus, seven samples; Gano, seven
samples; Grimes, seven samples; Jonathan, eight sam-
ples; Lawver, one sample; Mcintosh, one sample; North-
em Spy, seven samples; Rhode Island Greening, three
samples ; Rome, eleven samples ; Stayman, five samples ;
Tompkins King, four samples; Wagener, six samples;
White Peamain, three samples; Willow, one sample;
Winesap, ten samples; Winter Banana, one sample; Yel-
low Bellflower, four samples; Yellow Newtown, seven
samples; and York Imperial, two samples, making a total
of 117 samples embracing 24 varieties. These apples
were obtained from fourteen localities in the state of
Washington and also from one locality in each of the fol-
lowing states: New Hampshire, Indiana, Missouri, New
York, and West Virginia.

Before leaving the preliminary portion of this paper
the writer wishes to express his appreciation of the aid

Digitized by

Google

600 THE AMERICAN NATURALIST [Vol. XLVHI

received from those who have helped in various ways in
the investigation. Thanks are due to the members of the
staff of the department of horticulture for suggestions
and encouragement, to the members of the library staff
who have rendered aid in the study of the literature of the
subject, to Mr. Geo. A. Olson, chemist of the experiment
station, who has analyzed the various samples of Grimes,
Jonathan, Yellow Bellflower, and Winesap, and finally
to the various fruit growers and others who have cooper-
ated in seculring the fruit and have furnished notes on the
conditions of production. To all these the writer takes
pleasure in acknowledging his gratitude and indebted-
ness.

Environmental Factors

Aside from small individual differences, better called
fluctuations than variations, and other more striking
modifications of comparatively infrequent occurrence
and obscure origin, which it is customary to explain as
bud variations, if, indeed, the application of a name to a
phenomenon can pass as an explanation, it is quite gener-
ally recognized that variation in any variety of fruit is
due to the operation of external influences. A knowledge
of the various factors which make up the environment
and their influence upon plant life is necessary to an in-
telligent study of variation. It should be noted, however,
that this influence is not necessarily the same with plants
propagated vegetatively as with those grown from seed.
In the latter case certain modifications of an adaptive
nature which enable the plant to fit in more perfectly with
its surroundings are apt to persist, while less favorable
modifications tend to disappear by the elimination of the
individuals possessing them. In the former case, on the
other hand, the modifications observed are the direct re-
sult of the conditions, unaffected by selection, and whether
desirable or not they persist as long as the environment is
unchanged and the vegetative propagation is continued,
unless, indeed, the environment is so unfavorable that the

Digitized by

Google

No. 574]

VARIATION IN THE APPLE

601

changes induced are pathological in nature and the plant
can not survive.

Perhaps the most important factor to which plant life
is subjected is the moisture relation. This may be deter-
mined by the amount of moisture actually present or by
the modifying influence of other coexistent factors which
interfere with the availability of the moisture and the
capacity of the plant to make use of it. Among such in-
direct influences may be noted the modifying effect of
temperature upon the rate of absorption and transfer of
moisture, the presence in the soil of certain salts or humic
acids which interfere with the osmotic activity of the
roots, and certain atmospheric conditions favorable to
rapid transpiration. In such cases car^ is necessary to
determine which is the direct and which the indirect cause
of the modifications. If it is borne in mind that many
factors cause variation through' their influence on the
moisture supply confusion may often be avoided.

The temperature relation is much more obscure than
the moisture relation in its effect upon plant growth.
Heat, being a molecular phenomenon, acts directly upon
the protoplasm and its effects are therefore physiological.
It is now pretty well understood that heat alone is in-
capable of modifying plant structure, but acts indirectly
through othQr factors and the functions of the plant. The
direct effect of temperature is limited very largely to its
influence upon the rate and amount of development. A
slight difference in the average temperature of the grow-
ing season influences greatly the relative development of
apple varieties. The accompanying table gives the mean
monthly temperature during the growing season at
Geneva, N. T., and Pullman, Wash., since the establish-
ment of the experiment stations at those points, as well as
the mean for two years at White Salmon, Wash.

Locality.

April

May

June

July

Aug.

Sept.

Oct.

Aver.

Geneva. N.Y

Pullman. Wash

White Salmon, Wash. .

44
47
50

57
52
56

67
59
63

71
66
70

69
66
67

68
60

50
48

! 63

60
67
60

Digitized by

Google

602 THE AMERICAN NATURALIST L V^OL. XLVIII

The difference in the development of certain varieties
of apples at these places will be noted later. It will be
seen that the season opens slightly earlier in Pullman
than in Geneva and closes at about the same time. It
would appear, therefore, that the better development of
most varieties at the latter station is due rather to the
higher temperature than to a difference in the length of
season. At White Salmon the season is considerably
longer than at either of the other stations, while the tem-
perature from May to September is intermediate.

Latitude and altitude are frequently mentioned as im-
portant factors in the modification of varieties. These,
however, are not primarily factors, but depend for their
influence upon the effect of other factors, which in turn are
influenced by the location. Differences in altitude espe-
cially result in marked changes in climate often in places
geographically near together.

The light relation is of much importance to the fruit
grower. It is clearly evident that the development of
color in apples is largely dependent upon the sunshine,
and quality also may be affected through the production
of sugars. Both intensity of insolation and duration of
the daylight must receive consideration. In general, trop-
ical, arid or alpine situations are characterized by high
insolation, while a long period of daylight during the sum-
mer months is a factor in northern latitudes.

The effect of atmospheric influences is largely indirect.
It has already been noted that the condition of the air
may modify the moisture relation through its effect upon
transpiration, thus dryness, high temperature, and rar-
ification all favor evaportion, and this effect may be in-
creased in windy situations. Atmospheric pressure is a
factor of importance in high altitudes.

The soil may be of importance as a factor in causing
variation through either its chemical composition or its
physical properties. The former leads to a consideration
of the influence of fertilization, the latter to the effect of
different methods of culture. Here again other factors,

Digitized by

Google

No. 574] VARIATION IN THE APPLE 603

and especially the moisture relation, have an important
bearing, since one of the primary results of cultivation is
the conservation of the soil moisture. There is no doubt
that the nature of the soil greatly affects the crop and the
matter has been given much study. The intimate associa-
tion of other factors, however, makes it somewhat diflScult
to pick out those influences for which the nature of the
soil is directly responsible.

The influence of other organisms includes not only a
consideration of the effect of insect and fungus pests but
in the broad sense embraces such items as pollination,
pruning and thinning, intercrops, cover-crops and plant-
ing distance. Human agencies, including all operations
of orchard management, might properly be included here.
Many of these are, of course, indirect, exerting an influ-
ence through theii^ effect upon some other factor.

The Law of the Optimum

Having enumerated the chief external influences to
which plants are subjected during their period of develop-
ment and to which variation is largely due, the question
naturally occurs whether there can be formulated any
basic principle or law which will express the manner in
which plants react with the environment. Such a law
would be of use not only in the study of variation, but
would shed m^uch light on the adaptation of plants to new
environments. It would constitute a unifying principle
whereby isolated facts and disconnected observations ap-
pear in proper relation and perspective. Though a dis-
cussion of this subject might logically be delayed until
after the characters of the several varieties and their
modifications have been noted, it is thought most fitting to
introduce the statement at this point and examine the
fruit in the light of such generalizations as it has been
possible to make.

A plant can live and perform its functions only within
certain intensities of the various factors of the environ-
ment. The degrees of intensity beyond which activity

Digitized by

Google

604 THE AMERICAN NATURALIST [Vol. XLVIII

ceases are known as the zero points. The plant does not
necessarily die at once, but passes into a dormant state.
If the intensity becomes still more unfavorable a point is
finally reached at which death occurs. The minimum de-
gree of intensity of a factor at which the plant may re-
main active is known as the lower zero point, while the
greatest intensity is called the upper zero point. With
some factors these points are wide apart, so that, other
conditions being favorable, the plant will continue to de-
velop after a fashion at any but the most extreme intensi-
ties of such factors. With other factors the limits are
comparatively narrow. A plant will reach that degree of
development only which is permitted by that factor which
is in the least favorable degree of intensity. Such factors
are called limiting factors. In passing from one zero
point toward the other, a point is finally reached at which
any given function of a plant reaches its highest state of
activity. This point is known as the absolute optimum for
that function and may not correspond to the most favor-
able intensity of that factor for the performance of the
other functions of the plant. The point of intensity of a
factor at which all the functions of the plant are per-
formed to the best advantage is termed the harmonic
optimum. If each factor is of an intensity corresponding
to the harmonic optimum, the plant is in a condition of
equilibrium known as the ecological optimum and will
reach the highest state of activity of which it is capable.*
As the life of a plant is made up of various functions,
so its structure is made up of a number of organs having
various characters. These characters are the result of
development, which in turn is dependent upon the per-
formance of the several functions of the plant under the
influence of those external conditions which make up the
environment. If a factor of the environment is modified
in its intensity, the balance of the functions of the plant
is disturbed and the plant reacts to its changed environ-
ment by a modification of its functions which may result

4 Schimper, A. F. W., '^ Plant Geography.*'

Digitized by

Google

No. 574] VARIATION IN THE APPLE 605

in a different kind of development, or in other words a
variation. Having observed the close connection between
the characters and the functions of the plant, we may now
inquire whether the former maintain a relation to the
environment similar to that maintained by the latter.
Putting aside generalizations for the present and confin-
ing attention to the apple, it is to be noted that both Shaw
and Stewart foreshadowed such a relationship in the
papers already noted. Neither, however, carried the
analysis far enough to formulate a rule of general appli-
cation, though Stewart came near doing so. Sha^ recog-
nized that the highest perfection in any given variety
could be attained only under the most favorable summer
temperature. Stewart applied this idea to other factors
than temperature in his ''optimum principle," which is
'' that plant growth and development increase as the most
distant essential factors approach the optimum." His
failure to recognize, the connection between the various
factors of the environment, on the one hand, and the sepa-
rate characters of the apple, on the other, may be ac-
counted for by the fact that his investigations dealt only
with fruit grown under slightly modified conditions, which
resulted only in such slight variations that the independent
modification of the separate characters escaped notice.

In examining various samples of apples produced under
the influence of quite dissimilar combinations of environ-
mental factors, the writer has many times noted the modi-
fication of certain characters more or less independently
of others. It is true that characters are often found to
vary together through a relationship of direct or inverse
correlation. Such cases, however, are possibly as often
due to the response of the various characters to the same
factor of environment as to any direct connection between
the characters, though the latter no doubt exists in many
cases. Keeping in mind these facts and also the close
relationship of function and character, the writer has
formulated a principle which he believes is of general ap-
plication not only to apples but to other horticultural

Digitized by

Google

606 THE AMERICAN NATURALIST [Vol. XLVIH

crops and perhaps in a degree to all plant life. For this
generalization, which expresses the relationship of char-
acters to environmental factors the name '^ Law of the
Optimum " is proposed.

This law may be stated as follows: For any given
variety there is for each character a certain intensity of
each essential factor of the environment at which, other
conditions remaining the same, that character reaches its
highest development. When all essential factors are in a
condition of optimum intensity for any character, that
character will reach the most perfect development of
which it is capable. A modification of the intensity of any
such factor either above or below the optimum will be ac-
companied by a less perfect condition of the character
concerned. The optimum intensity of a factor may be
wide or narrow in its limits and the optimum for one char-
acter may or may not overlap the optimum for others.
A variety will be at its best when grown in an environ-
ment the factors of which are as near as may be to the
optimum intensity for all characters. Under such cir-
cumstances the variety is in a state of balanced adapta^
Hon to its environment. If removed from such an en-
vironment to one in which certain factors are distant from
this state of average optimum intensity for all characters,
the equilibrium is destroyed and the variety is thrown
into a state of unbalanced adaptation, in which those
characters farthest removed from their respective optima
are injuriously affected, while others may be bettered by
being placed in a combination of factors of an intensity
nearer their optima. A discussion of the practical appli-
cation of this law and its bearing upon apple culture in
the northwest will be deferred for the present and taken
up in a later section.

A COMPABATIVE StUDY OF THE SAMPLES

A close study of the various lots of apples used in this
experiment brings to light variations in practically all
characters. Many, however, are modifications of charac-

Digitized by

Google

No. 574] VARIATION IN THE APPLE 607

ters inconspicuous in themselves or are slight in amount
and so do not attract attention. A complete account of all
variations noted would comprise a full technical descrip-
tion of each sample which would far exceed the limits of
this paper. For this reason it is thought best to append
only some brief comparative notes regarding the more
conspicuous variations noted in each variety. In this
connection it is well to note the origin so far as known of
the varieties included in this study. Arkansas and Arkan-
sas Black, Arkansas; Baldwin, Massachusetts; Ben
Davis, probably Kentucky or Tennessee ; Delicious, Iowa ;
Esopus, New York; Gano, probably Kentucky or Mis-
souri; Grimes, West Virginia; Jonathan, New York;
Lawver, possibly Kansas; Mcintosh, Ontario, Canada;
Northern Spy, New York; Ehode Island Greening, Ehode
Island ; Eome, Ohio ; Stayman, Kansas ; Tompkins King,
New York ; Wagener, New York ; White Pearmain, prob-
ably Eastern States; Willow, Virginia; Winesap, New
Jersey; Winter Banana, Indiana; Yellow Bellflower, New
Jersey; Yellow Newtown, New York; York Imperial,
Pennsylvania. It will be observed that all originated in
the east or middle west. Most no doubt appeared as seed-
lings and were selected and propagated because of their
excellence and value when grown under those conditions
of environment which prevail at their places of origin ; in
other words they were individuals which happened to be
in a condition of balanced adaptation to that environment.
Their behavior under other environments could be deter-
mined only by actual tests, and some notes on the subject
are included in the following paragraphs.

ArkansorS {Mammoth Black Twig). — ^As only one sam-
ple of this variety was examined its behavior can be com-
pared only with what is known of the variety in other
localities. The fruit was more elongated and conical in
shape, smaller in size and less highly colored than that
produced in the warmer apple-growing sections of the
east. The flesh was inferior in texture, indicating poor
development. The variety seems not at all adapted to

Digitized by

Google

608 THE AMERICAN NATURALIST [Vol. XLVUI

the location where grown, but might do better at lower
altitudes and in warmer situations in the state. Never-
theless, the quality is not good enough to recommend the
variety for dessert, and it is to be hoped that it will not be
planted extensively in the northwest. The keeping qual-
ity was excellent.

Arkansas Black. — This variety of the Winesap group
attains a deeper color than the Winesap and equals that
variety in size and quality. The specimens examined
were not especially well colored though, it is known to
color well in the irrigated valleys. It seems to be better
adapted to the conditions of the state than the Arkansas.
In keeping quality it was among the best.

Baldwin. — The Baldwin attains its highest perfection
in New York and New England, where it is a great favor-
ite in the markets and is produced more largely than any
other variety. As grown in this state the fruit is snaailer
and more elongated than the eastern product and Jms a
more deeply furrowed basin. As grown at Pullman the
color lacks intensity, though the fruit is well covered. In
the western part of the state the fruit is well colored,
especially in the northern part of the Puget Sound Basin.
The lots from White Salmon show a good many poorly
colored fruits mixed with those of better color, while the
quality is rather better than in those examined from other
parts of the state. It is, however, inferior to the eastern-
grown Baldwin and is evidently poorly adapted to the
conditions of the northwest. All of the Washington-
grown fruit displayed a tendency to wilt in storage and
some of the lots from the western part of the state rotted
seriously as a result of fungous infections not apparent on
the fruit at the time of storage.

Ben Davis. — Though displaying considerable lack of
balance in the adaptation of the different characters to
conditions in certain parts of the state, this variety seems
on the whole to reach a good degree of development in the
warmer valleys. In quality the lot from Missouri was
superior to those from any part of Washington, though

Digitized by

Google

r .574] VARIATION IN THE APPLE 609

many of the Washington-grown apples of the variety were
equal to those from most sections of the east. Striking
variations in form were displayed by the fruit from dif-
ferent localities. Those lots from the more elevated and
cooler sections of the state were of an oblong, conic form
and usually had shallow irregular basing, while those
from the warm valleys were less elongated and had deep
and usually quite regular basins, being more like the fruit
from the Ben Davis belt of the east. The fruit developed
better texture and quality also in the valleys though it was
coarser and more spongy than the eastern fruit. Most of
the Washington grown samples of Ben Davis were more
decidedly striped than those from the east. This effect is
produced by the clearer yellow ground color, which in the
eastern-grown fruit is more or less suffused with red.
The apples from the elevated localities of Pullman,
Cloverland and White Salmon were relatively small in
size and poorly colored. Because of its low dessert qual-
ity, the planting of this variety for shipment to the east
can not be recommended. The most desirable feature of
the Ben Davis fruit is its good keeping quality. A tend-
ency to mealiness late in the season was observed in
some of the fruit from the irrigated valleys, while those
grown at Pullman and Cloverland wilted badly toward
the close of the season.

Delicious. — This is one of the newer varieties and when
well grown is a dessert apple of fine appearance and high
quality. In many of its characters, but especially in flavor
and aroma, Delicious resembles the White Pearmain,
though in color it bears a likeness to the Winesap group.
In moderately elevated situations in some parts of the
state it displays a well-balanced adaptation and attains
excellent size, color, texture and quality, though none of
those examined were quite equal in quality to the Deli-
cious from New York. When grown in too low and warm
a location the fruit has a tendency to become overripe and
when stored tends to soften in the center, after which it
loses greatly in quality. The sample from Clarkston had

Digitized by

Google

610 THE AMERICAN NATURALIST [V0L.XLVUI

a beautiful dark red color, while that from Cloverland
was dull in color and poor in texture.

Esopus (Spitzenburg). — This is almost the only variety
which the writer has examined that attains the first rank
as a dessert apple in this state. In certain sections it dis-
plays a better balance of adaptation so far as flesh charac-
ters are concerned than any other variety. The samples
obtained from White Salmon and the irrigated valleys
were of excellent quality as dessert apples, though of
scarcely as good texture as the variety attains in the east.
Overgrown apples are especially coarse in texture. West
of the Cascades and in the more elevated locations the
Esopus does not reach as high quality as elsewhere. This
variety is inclined to wilt in storage unless well grown.

G(mo.— This is an apple of the Ben Davis tyi)e, but of a
more uniform red color. Practically all the remarks in-
cluded under Ben Davis, aside from those dealing with the
distribution of color, apply equally well to the Gano. At
its best, the Gano is of slightly better quality than the Ben
Davis, which fact, together with its more handsome ap-
pearance, renders it a more desirable variety to plant, yet
neither can be recommended in a section desirous of build-
ing up a reputation and market for dessert apples. It is
interesting that both the highest color and the best as well
as the poorest quality was attained by apples from the
east and middle west.

Grimes (Grimes Golden). — This variety, like the Ben
Davis, displays considerable variation in form, depending
on the locality of production. The specimens from the
middle west were roundish to decidedly oblate, while those
grown in Washington were all more or less elongated.
Those grown west of the Cascades displayed a greater
tendency to a conical shape than those from the eastern
part of the state, and were also poorer in quality. When
grown in the more elevated sections, as at Pullman,
Grimes appears poorly developed and immature and is
inferior in size and quality. Those from Grandview dis-
played the best balance of characters and it seems prob-

Digitized by

Google

No. 574] VARIATION IN THE APPLE 611

able that this variety is better adapted to the irrigated
valleys than to other sections of the state. All samples
were more or less wilted by midwinter, except the fruit
from Grandview, which remained firm but showed some
tendency to rot. Scald was very bad in the latter part of
the season.

Jonathan. — ^Although rather extensively grown in a
number of localities in Washington, none of the fruit
which the writer has examined gave evidence of a well-
balanced adaptation to the conditions of growth which
prevail in the state. All were inferior in color to the fruit
obtained from the east and middle west. The apples from
Clarkston and the Yakima Valley were of good size but
lacked both richness of flavor and aroma. The same lack
was evident in the fruit from the western part of the state.
At Pullman a pretty good quality is attained, but the
fruit does not come up to the requirements as to size and
gives other evidence of imperfect development. At
Cloverland and in other elevated locations fruit of a poor
texture and deficient coloring is produced. Jonathan
seems to reach its highest development in certain sections
tributary to the Ohio valley and the Washington-grown
Jonathans can not compete with fruit from that section
when well grown. The samples from Morgantown, West
Virginia, were of a beautiful clear dark red color, good
size, fine tender flesh, and very high quality. In storage
these specimens remained firm and retained their flavor
until April. The others wilted considerably after mid-
winter.

Lawver. — This variety attains good size and fine color
in the irrigated valleys, but the quality is not good enough
to recommend it to the fruit growers of the northwest.
The variety ordinarily keeps well but the specimens
stored proved to have poor keeping quality — owing to
fungous infection.

Mcintosh. — The Mcintosh is deserving of attention as a
variety of high quality which appears to have a fairly
well-balanced adaptation to certain sections of the north-

Digitized by

Google

612 THE AMERICAN NATURALIST [Vol. X:X:^"VIII

•west. At Pullman the elevation is too great for th^^ \^%t
development of the variety, but the Spokane Valleji;^ -pto-
duces Mclntoshes of a high degree of excellence. 1IBI?l:iere
is good reason to believe that the valley of the nor: — tlxem
and northeastern sections of the state can rival the in^itter
Eoot valley of Montana in the production of this v^t^zari^ty.
The fruit stored wilted badly by midwinter and lost ii^cxuch
of its flavor soon after.

Northern Spy. — Of all the varieties examine <^:3. the
Northern Spy seems least adapted to the conditicz:^ :i3. s of
growth in this state. As produced in New York an<:3^ JSew
England this fruit is a dessert apple of the highest qi^^^o^ality
when well grown and properly colored. In Washi^angton
east of the Cascades the color fails to develop aMizi-ci the
quality is much inferior to that of the eastem-growraL fruit.
In the western part of the state the color develops ea^ss well
as in the eastern states, but the quality is no bett^ zar- than
elsewhere in the state. The unsurpassed cooking ci "U-^lity
of this variety seems to be largely retained, ho^v^^^ver,
which is its only redeeming feature. It may be ^^v^orth
planting to a limited extent as a culinary fruit iowr liome
use, but can not eoini>ete in the markets with the e^^ stem-
grown Northern Spys. The specimens from the v^^ stern
part of the state were largely infected with fungi, :r*^ suit-
ing in much decay early in the season. Those fronci. I^uU-
man and Clarkston kept fairly well, though the f oXT^aer
wilted badly late in the season.

Rhode Island Greening. — This variety, togethexr ^with
Baldwin and Northern Spy, constitutes the most j^-x'Oini-
nent and successful apples in the orchards of New "STork
and New England. They are also amcng the varieties 1^35t
adapted to the conditions found in this state. Their per-
fect balance of adaptation to eastern conditions is jpx'cih
ably to a large degree responsible for their "popxHsLirH^y ^^
the east and may also account for the lack of "b^J^^^
which they display in the northwest. As grown at Wbite
Salmon and at Pullman the Greening reached a good ^^^/
but was decidedly inferior in quality to the specim^^^

Digitized *by

Google

No. 574] VARIATION IN THE APPLE 613

from New Hampshire. At Pullman the fruit was rather
flat and strongly ribbed, while at White Salmon the apples
were oblong in shape and had, as a rule, rather small
cavities. It can not be recommended for Washington,
except possibly for local use as a culinary fruit. This
variety is a fairly good keeper. Those grown at Pullman
wilted badly late in the season, while the lot from White
Salmon gave evidence of considerable fungus infection.
Ro7ne {Rome Beauty). — This is one of the most popu-
lar varieties grown in the state east of the Cascade Moun-
tains and is about the only commercial variety which
reaches good marketable size in the high uplands of the
Inland Empire. The Eome reaches its highest develop-
ment in the Jonathan belt of the middle west. The best
specimens examined, all characters considered, came from
Morgantown, West Virginia. They were of a nearly uni-
form deep red color, of good size and attractive form, and
of pretty good quality for the variety. In many parts of
Washington the Eome fails to color well. The specimens
from White Salmon and Grandview were especially poor
in color. The latter were overgrown and of poor quality,
while the former were among the best of the variety. The
usual form of the variety is round or nearly so, varying
to somewhat roundish conic or roundish ovate. The form
of the cavity is subject to quite a little variation. As pro-
duced at Pullman and other elevated sections of the state
the cavity is very shallow^ but becomes deeper in the val-
leys. The specimens from West Virginia had fairly deep
cavities. Indeed it seems probable that those localities
which produce Ben Davis of the elongated type also pro-
duce Eomes with the shallow cavities. The Eome is by
nature a culinary apple. In quality it is but little better
than Ben Davis. It seems unfortunate, therefore, for the
lasting reputation of the industry, that it should have be-
come so firmly established in northwestern horticulture.
It is to be earnestly hoped that it may in time be replaced
by a variety of better quality. In its adaptations to the
conditions of the state, the Eome seems to be fairly well

Digitized by

Google

614 THE AMERICAN NATURALIST [VoL.XLVin

-balanced in most of its characters. The balance, however,
is not the same in all sections and is nowhere quite so per-
fect as in certain localities in the middle states. Most
samples kept well until the latter part of the season and
then became mealy. The overgrown specimens from
Grandview were the first to break down in this way.
Those grown at a greater elevation showed a slight tend-
ency to wilt late in the season. None of the samples dis-
played an inclination to rot until late in the season.

Stay man Winesap,— In both size and quality the Stay-
man is the best of the Winesap group. Its most serious
fault is a rather dull color which often fails to cover the
fruit well. The samples obtained from the middle west
were of better €olor and texture than those grown i?5v
Washington, though the lot from Indiana were very coar ^
in texture. Those grown at Pullman were small and v>:x
ferior in every way. The fruit from Grandyiew v?^^^
especially large, flat, and fairly well colored, while tlcx^^v
from White Salmon was more elongated, slightly X^^^
colored, and rather more aromatic in flavor. These 'tr^^^
lots retained their firmness in storage much longer fcTtx^^
the others and those from White Salmon scalded l>.^3^.<iiy
late in the season. It is very similar to the Winesa. ^^> in
its adaptations.

Tompkins King, — This variety is popular in the
'ern part of the state, where it attains a large size and
color, though the latter character develops well at
man. None of the samples equaled in quality the va^
as grown in New York. Those grown at Pullman \m^ ,^a^ ^
very good flavor, though the flesh characters were tho ^o O;^ '"

poorly matured fruit. The fruit from the western pa^ -mrt. o^p 1 1

the state was of a fairly elongated conic form, whil^ ±l3.s^^ • ,/ ^

grown at Pullman was shorter and strongly ribbed. rTiix^ ^ ■

variety appears to be but poorly adapted to Washima-^^o^
conditions. The fruit grown at Pulhnan wilted badly" la*Q
in the season, while that from western Washington irottec^
considerably owing to fungus infections.

Wagener. — Though of the Northern Spy clasf^, tb^

.1 ^

Digitized by

Googl(

No. 574] VARIATION IN THE APPLE 615

Wagener displays a much better balance of adaptation to
the conditions of the state than the Northern Spy. It
seems to reach its best development in the cooler regions
of the state. The specimens from Grandview were of
good size and very juicy, but were poor in color, coarse
in texture, and deficient in flavor. Wagener develops
especially well in the Spokane Valley. The specimens
from Opportunity were large, well colored, and of excel-
lent quality, though somewhat coarse in texture. Those
grown at Pullman were more aromatic but possibly not so
rich in flavor and did not develop sufficient size. This
variety does well west of the Cascades and especially in
the northern part of the Puget Sound Basin. The speci-
mens from Eastsound were large, highly colored, and fine
in texture, but less aromatic than the eastern Washington
fruit. The samples obtained from West Virginia gave
evidence of having been grown too far south. They were
poorly colored and of rather poor texture, but of good size
and excellent flavor. In form the fruit from Opportunity
was roundish, that from Eastsound roundish conic, while
the remainder was decidedly flattened and all samples
were more or less strongly ribbed. This variety shows
very little tendency to wilt in storage. The fruit from the
highlands keeps well, but that from the irrigated valleys
shows a tendency to physiological decay. Scald is serious
after midwinter.

White Pearmain {White Winter Pearmain). — In gen-
eral appearance this variety often closely resembles the
Yellow Newtown, but is usually more elongated and more
largely blushed. Moreover, it is quite different in flavor
and is remarkable for its fine aroma. It is a variety of
high quality and attractive for a yellow apple, moreover,
it attains its good qualities in the irrigated valleys better
than on the highlands, the specimens from Cloverland be-
ing dull and green in color and poor in texture, but well
blushed and highly aromatic. Its worst fault is suscepti-
bility to the apple scab. It would seem to be better
adapted to growing in the state than some of the more

Digitized by

Google

61 6 THE AMERICAN NATURALIST [Vol. XLVm

popular varieties. The fruit from the Yakima Valley
retained its firmness much better than that from Clover-
land, but lost somewhat in flavor toward the close of the
season.

Willow {Willow Twig). — The writer has examined this
variety only as grown in the elevated portions of eastern
Washington. In such locations it does not develop espe-
cially well in either size or color and is of too poor quality
to be worthy of consideration. Moreover, it wilts badly^
in storage, though when well grown the fruit has excellent
keeping quality. It is evidently poorly adapted to thi^
section.

Winesap. — In some of the irrigated valleys this varietj?^
is one of the most popular apples grown. It attains e^
good marketable size and an attractive color, though non^^
of the samples examined were equal in color or quality tc:^
the Winesaps from Indiana and West Virginia. In el^ —
vated localities, as at Pullman, Cloverland and Whit^^
Salmon, the fruit is small and poorly colored and has fleslzm.
characters indicating imperfect development and matur^ —
ity. As grown in the irrigated valleys the fruit is apt t ^i^*
be deficient in flavor, and, if large, coarse in texture. Th. ^^
lot from Cashmere showed the best balance of character* a^
of any Washington, grown specimens, but these were in ii.«3
way superior to the Winesaps from West Virginia. It i- ^3
probable that the better grown fruit from the easter^Hria ^

Winesap districts is equal to that grown in Washington f ^y^ ^

all respects, with the possible exception of size, which, irff
large, is, as noted, apt to be accompanied by deterioratio ^ac*
in quality. It is evident then, that the balance of adapts*-—
tion of this variety to northwestern conditions is impe:Kr"—
feet at best and that the planting of Winesaps in Waslr»-—
ington may easily be overdone. This variety proved to t^^
one of the best in keeping quality. Those from Pullma^:»3
and Cloverland wilted late in the season, though most o:^
the other lots were in excellent condition in April and ^»
few were held in storage until July.

Winter Banana. — ^As only a single lot of this variety

Digitized by

Google

No. 574] VARIATION IN THE APPLE 617

was examined in detail, it is difficult to make very positive
statements regarding its behavior in the state. Though
less desirable than a red apple, it is a variety of handsome
appearance and is fairly good in quality. It is perhaps
rather better adapted than the average to certain sections
of the state and appears to develop best in fairly elevated
situations. It is especially well liked in the Spokane Val-
ley, and fruit grown there is said to have good keeping
quality, though the specimens from western Washington
were past season by midwinter. They wilted badly and
showed much scald.

Yellow Bellflower. — This variety appears to be better
adapted to the western part of the state than to the irri-
gated valleys. The apples from Clarkston were coarser
in texture, milder in flavor and poorer in quality than
the samples received from the east. There were no very
striking differences in form, structure or appearance ex-
cept that the eastern Bellflowers were more often blushed
than those from Clarkston. The apples from Puyallup
were overgrown specimens from young trees, were coarse
and spongy in texture, and inferior in quality. As this is
a tender fruit, easily injured by careless handling, • and
does not appear to be especially well balanced in its
adaptations, it is not desirable to plant extensively for
shipping. Moreover, it is not a good keeper. The speci-
mens from Puyallup were practically past season when
received and those obtained from the east were more or
less injured and such specimens decayed quickly. Some
of the lot from Clarkston, however, kept sound and firm
until past midwinter, but deteriorated in flavor toward the
last.

Yellow Newtown. — ^When at its best, this variety has
few equals. It is narrow in the limits of its adaptations
and its successful culture in the eastern states is confined
to small areas, where, however, it is in nearly perfect
equilibrium with its environment. In many places in the
northwest it is grown successfuUjj, though it scarcely
equals in quality the best eastern product. The fruit
from White Salmon and some of the irrigated districts

Digitized by

Google

618 THE AMERICAN NATURALIST [VoL.XLVm

was of excellent quality, but coarser and less delicate in
texture and of not quite so good flavor as the apples from
West Virginia. The specimens from Cloverland were
hard and green and gave evidence of imperfect maturity.
Evidently the elevation is too great for its proper devel-
opment. The single sample from western Washington
consisted of well-colored, extensively blushed fruit, but
was inferior in quality. Owing to its limited area of suc-
cessful production in the east, it is worth planting in
Washington wherever its characters give evidence of a
fair degree of balance of adaptation with the environ-
ment. This variety is perhaps a better keeper than Wine-
sap. Some of the fruit from White Salmon kept in good
condition until July, though overgrown fruit and that
which has been exposed to heat before storage showed
signs of physiological decay late in the season. Under-
developed specimens wilted in storage.

York Imperial. — ^In sections of Virginia and neighbor-
ing states the York Imperial occupies the place of su-
premacy held by the Baldwin farther north. This is
doubtless due to its perfect balance with the environmental
conditions of that region, and, like the Baldwin and other
sorts perfectly adapted to their eastern habitat, this
variety finds itself out of equilibrium when moved to the
northwest. The apples from western Washington were
of good size and color, but were coarse and undesirable
in texture and poor in quality. The specimens grown
at Pullman were smaller, more elongated, and less com-
pressed than the others, and the axes were less oblique.
They were somewhat better in quality, though not good
enough to justify more extensive planting. The fruit
wilted in storage, and that from western Washington
gave evidence of fungous infection and scalded badly after
midwinter.

Discussion of the Effect of Environment Upon Apple

Chabactebs
Size. — Size is the direct result of development. An
apple will reach its maximum in growth when all factors

Digitized by

Google

No. 574] VARIATION IN THE APPLE 619

are at the variety optimum for the physiological proc-
esses upon which development depends. A departure
from this optimum, whether toward a greater or less in-
tensity, means a decrease in size, as is observed in ap-
proaching either the northern or southern range of a
variety. It has been frequently noted, however, that the
optimum for growth is not the best combination of fact-
ors for the development of certain other desirable char-
acters, so that it is well to choose an environment having
certain factors in a somewhat less degree of intensity,
being content with fruit of fair size but superior in other
respects. Since the apple contains about 85 per cent, of
moisture it is evident that the water supply is a factor of
prime importance in determining size. It is possible by
excessive irrigation to force an abnormal growth of the
fruit, though always apparently at the expense of text-
ure, flavor, and keeping quality. It is evident, then, that
if fruit of good quality is expected, irrigation must be
moderate in amount, especially with vigorous young
trees. Thinning may result in increased size owing to
the larger amount of moisture available for each fruit.
Temperature and length of season are of importance in
determining, respectively, the rapidity of growth and de-
gree of development attained.

Form. — One of the striking features revealed by the
study of a number of varieties from several localities is
the fact that the modification in shape due to the differ-
ence in environment is by no means uniform for the
several varieties. Some varieties are quite constant in
shape while others are much more plastic in this respect.
Moreover, certain varieties are much more easily in-
fluenced than others which respond in the same way,
while still others respond differently to the same factors.
One of the most frequently observed and conspicuous
modifications of form consists of the elongation of the
axis of the fruit relative to the horizontal diameter. This
character has been especially studied, in the case of the
Ben Davis, by Shaw, who found the elongation most
noticeable in fruit from the northeastern states, the mari-

Digitized by

Google

620 THE AMERICAN NATURALIST [VokXLVm

time provinces of Canada, and the Pacific coast. Shaw's
papers dealing with this subject have already been noted.
Upon studying the climate in these localities, it was found
that the temperature for two or three weeks after the
blooming season was notably lower than in the sections
where the Ben Davis assumes its normal shape. Since
this appeared to be the only factor constant for the
several localities, it is suggested as the explanation of
this variation. It has been shown, however, that temper-
ature is incapable of influencing form except by its action
through the functions of the plant in modifying the effect
of some other factor. It is the writer's opinion that the
elongation is due to the relative moisture supply of the
different parts of the apple at this period of develop-
ment as influenced by the temperature ; that it is primar-
ily a modification due to the moisture relation rather
th^n to the direct effect of temperature, the latter being
a secondary cause. The rapidity of circulation of the sap
and therefore the supply of moisture to the organs of the
plant is greatly influenced by the temperature. It is a
well-known fact of plant physiology that much less moist-
ure passes through the plant in the cool days of spring
than during the warmer weather of midsunmaer. A re-
duction of the temperature at this time results in a still
more sluggish movement of the sap. In the period im-
mediately after blooming the energy of the plant, so far
as the development of the fruit is concerned, is directed
primarily to the proper nourishment of the growing seeds
and the adjacent parts. If at this time the circulation
of the sap is retarded by a temperature unwontedly low
for the variety, the moisture supply of the fruit is
lessened and a relatively larger amount goes to the seeds
and adjacent parts, while the pulpy portion of the fmit
receives a more scant supply. As a result, the axillary
development is proportionately greater than the sweUing
of the fruit due to the accumulation of moisture in the
superficial tissues. After some two or three weeks the
form of the fruit becomes fixed and is not noticably in-
fluenced by the moisture supply thereafter.

Digitized by

Google

No. 574] VARIATION IN THE APPLE 621

The elongation of the fniit is usually accompanied by
a constriction of the apex resulting in a conical form.
This may be due to the greater development of the basal
portion, which is adjacent to the point where the sap
enters the fruit and may therefore be better supplied,
though the physiology of fruit development is in need of
further study. In the Grimes, however, an oblong form
results. The Mcintosh, as grown at Pullman, is often
decidedly obovate, a variation which the writer ascribes
to the same influences that produce the elongated conic
form of the Ben Davis and other varieties, though in this
variety the response is somewhat different. The Ehode
Island Greening, Willow and Wagener, as a rule, fail to
assume an elongated form in localities where it is well
marked in some other varieties. Also in certain varie-
ties which are naturally conic in form and considerably
elongated, as Delicious and Yellow Bellflower, this effect
is not evident. The larger number of varieties, when
grown in this state, have a more ribbed form than the
same varieties in the east. This seems to be due to a lack
of balance in adaptation, though the particular factor
which gives rise to the variation has not been determined.
Some varieties, like the York Imperial and the Yellow
Newtown, are compressed in form, that is elliptical in
section, and have an oblique axis when grown in certain
environments. These characters seem to be in some way
related to the better development of the fruit, as they are
less evident in fruit from the elevated and unfavorable
. sections of the state. Beach has noted in the *' Apples of
New York'' a similar difference between the Newtowns
of western New York and those of the Hudson Valley, the
latter having a more oblique axis and elliptical form.

Stem. — The stem is one of the most variable structures
of the apple, and, owing to the fact that stems of different
lengths, diameters and shapes are commonly found in
any lot of apples grown under practically uniform con-
ditions, it is difficult to associate such variations with the
environment. The writer has noted, however, in the case
of some short-stemmed varieties, like the York Imperial,

Digitized by

Google

622 THE AMERICAN NATURALIST [VoL.XLVin

that those lots grown under less favorable conditions had,
on the average, longer stems than others grown under a
more favorable environment.

Cavity. — The most conspicuous variation in the cavity
is in its depth. This is of especial note in the Rome, which
has a very shallow cavity in most parts of the state. This
is doubtless due to the same cause which produces the
elongated form of the fruit in many varieties, namely the
elongation of the axis resulting from a deficient moisture
supply incident to a low temperature after the blooming
season. In this variety the elongated axis obliterates the
cavity instead of modifying the general outline of the
fruit. The same variation is also noted to a less degree
in a number of other varieties. An especially furrowed
cavity is often observed associated as a rule with the
ribbed form of fruit.

Calyx. — The writer has failed to observe any modifica-
tions of importance in the calyx lobes of the fruit. The
size of the calyx cup or '' eye '* of the apple is influenced
by the development of the fruit. In large fruit this open-
ing is apt to be large, so that the lobes are separated, re-
sulting in an open or partly open calyx. Small or poorly
developed apples, on the other hand, usually have the
calyx closed.

Basin. — The depth of the basin seems to depend upon
the same factors as that of the cavity and seems to be
much more readily influenced than the latter. The width
is often associated with the form of the apple, a very con-
stricted apex resulting in a narrow basin. A much fur-
rowed basin results from a combination of factors un-
favorable to the best development of the fruit.

Skin. — Statements have often appeared in regard to
the effect of various climatic factors upon the thickness
and toughness of the skin. Estimates of these characters,
however, appear to be based entirely upon sense impres-
sions of the observers, although it would seem that exact
measurements would not be especially difficult. In the
absence of such accurate data, an expression of opinion

Digitized by

Google

No. 574] VARIATION I!^ THE APPLE 623

would be premature. Dry air and sunshine are favorable
to the production of clear, smooth skin.

Color. — There seems to be no doubt that the coloration
of apples depends upon the influence of several factors of
which light is usually the most important. The impor-
tance of light is easily demonstrated by covering the fruit
during development either wholly or in part. The in-
tensity of illumination is also, evidently, quite narrow in
its limits, so that a point is soon reached at which the
color begins to pale owing to excess of illumination. It
has been frequently noted that apples grown near the
southern limit of the range of a variety are paler than
those grown farther to the north. This effect appears to
be the result of an excess of the two factors, heat and
light. It has been mentioned in the discussion of the
characters of several varieties that, contrary to the gen-
eral impression, those grown in Washington east of the
Cascades where insolation is intense were less highly
colored than those from western Washington or the east-
ern states. The most marked example of this kind which
the writer has observed is the Northern Spy. Again,
contrary to the general impression, most of the samples
from elevated locations were poorly colored, a fact which
may be attributed partly to the strong insolation and
partly to the poor development due to the low summer
temperature. It appears, therefore, that either too
strong or too weak illumination may result in poorly
colored fruit and that the best color is developed under a
condition of optimum intensity of the light.

It is suggested above that temperature may influ-
ence color. This is most commonly observed in the
case of apples grown under conditions of too
high summer temperature, though a deterioration
in color also results if the temperature is much
below the optimum for the variety. It is often
stated that apples become more highly colored the farther
north they are grown. This is only true in part. Those
varieties which are adapted to the most northerly por-
tions of the apple belt are able to develop their highest

Digitized by

Google

624 THE AMERICAN NATURALIST [Vol. 5^x^-^n.

color at the limit of winter hardiness of the tree* ^\e

southern varieties, on the other hand, require for tlx.^^ "best
development of color a higher summer temperatur-^^ -fciau
is experienced in the northern localities. The Wi:B=^^sa^^
for example, when grown in Central New York is :E3«i.rV5^
covered with a pale red. At Pullman the major^il^^r of
varieties color poorly, due at least in part to th.^^ <^<yol
climate. That the temperature and not the shortn^^^ss of
the season is the factor involved is shown by the fa^:^ t i:laat
most of these varieties color well in central New^ "5E"orlf
which has a season of about the same length fcTfcioxi^k
averaging several degrees warmer.

Cultural conditions may influence the color to ^3l. cer-
tain degree. In general those processes of orcharc3^ xman-
agement which favor the early maturity of the fr-^jia.it; re-
sult in improved color, especially in localities ha^^^ixig a
short growing season. Pruning and wide planti»:ziL.gr are
regarded as favoring high coloration by admittin^^^ light
into the tree, though it is possible that in regions ^^^^here
the light is intense these factors may not be of so .^great
importance in their effect upon color as in less sunn. -^^^ loca-
tions. Something has been said of the influence '^z^rC the
soil in the discussion of the literature and it ha- g=^ been
noted also that studies of the effect of fertilizers upi^ .«>xi the
color have not yielded satisfactory or uniform r:--^^^ults.
The influence of iron compounds is worthy of brL '^3^:C dis-
cussion in this connection. It seems evident, frcz^:*^^"^ the
chemical studies which have been made, that the r^^^* <^ PSf-
ment includes iron in its composition. This has- some-
times been assumed to mean that the chief requixT'^^^^^^Dt
for highly colored fruit is the presence of plenty of ^a^v^u-
able iron compounds in the soil. As a matter of^ :t^et,
iron is also necessaly to the formation of chlorophyJ^ ^-^^
most soils contain an abundance of that element fox* ^ie
purpose. From the chemical data compiled by Ste^^'^^^^^
it appears that the ash of the fruit contains a, ZOT^<^i
smaller proportion of iron than that of the leaves. It^ is
logical to conclude, therefore, that soils containing stx^'

6 Pa. Sta. Eept. for 1910-11.

Digitized by

Googl

No. 674] VABIATION IN THE APPLE 626

cient iron for the development of chlorophyll in the leaves
are also fully supplied for the formation of the red pig-
ment of the apple.

Internal Structure. — ^The form and relative develop-
ment of the core and associated structures are subject to
numerous variations, which, however, are seldom so con-
spicuous as to attract attention unless closely studied,
and appear to be of little practical importance to either
the grower or consumer of the fruit. The number of
seeds may be mentioned as an indication of the thorough-
ness of cross pollination and in most varieties the pres-
ence of one or more well developed seeds is a requisite
to the proper development of the fruit. Small or poorly
developed fruit, the result of too short a season or too
low a temperature, is apt to have the core closed and axile,
or nearly so, while in the same varieties good develop-
ment is usually associated with a more open abaxile core.
The carpels of such poorly developed fruit are usually
entire and smooth, while those of the better-grown fruit
are more or less cleft and often tufted.

Flesh Characters. — From the standpoint of the con-
sumer, these are by all odds the most important charac-
ters of the fruit, though lost sight of through the empha-
sis placed on external characters, and no grower who
has at heart the permanent prosperity, extension and
normal development of the industry can aflford to look
upon quality as a secondary consideration. Neglect in
this matter is sure to result sooner or later in a bad repu-
tation for the fruit among a considerable proportion of
buyers, which appearance and advertising will not be
competent to overcome. The fact can not be denied that
the great majority of varieties fail to attain as high
quality in the northwest as when grown in the eastern or
middle states where nearly all of them .originated, while
^t the same time they may excel in other important char-
acters. This is especially true of most of the choice
dessert apples. Such unequal development can have no
other interpretation than that these varieties are in a
state of unbalanced adaptation to the environment. This

Digitized by

Google

626 THE AMERICAN NATURALIST [Vol. Xt^^^v^j^

fact being recognized, the main question is, How cam^ fliis
disadvantage be overcome T Evidently the solution ^ioes
not consist in a steadfast refusal to face the situ-^^^^-t^Q^^
and vehement declaration that the fruit of any parti ^^z^xilar
district is the best that can be produced. Such ta^ -^z^tics,
though well meant, can be permanently successful only

when the statements are justified by the facts. If ^^^.^^pie
(Culture in Washington is to be maintained upon a ^ '^z>T3iid
basis it will be necessary first of all that growers aslaall
exercise great care in planting to choose those var*i_^^ties
most nearly in equilibrium with the environment Lzbzx the
various sections of the state, at the same time avc^ iciing
over-irrigation or other errors in orchatd manag^^ zaacieiit
which may tend to an unequal development of the ^3 liar-
acters of the fruit, usually at the expense of qc»..«ility.
Even this, however, may be but a temporary mak^ ^laift^
since few if any of the better varieties possess tl=^ -^ re-
quisite power of adaptation. It will be necessary fi^^Hcrst of
all to determine if the variations which appear -'^v^heii
apples are grown from seed in the northwest are ^^aote
favorable in character than those which are display^^<3 by
introduced varieties. If such should prove to be tk^3 oase
the writer is under the conviction that the apple (*-irM-T ttire
of the northwest should ultimately be largely mad ^3 C3ver
on a basis of new varieties of local origin. A num"li>^r ^^
such varieties have already appeared, but unfortu-:*^^^*'*^'^
some of them have been chosen with little regamr-cfl ^^]
quality. No work of greater value to the future laorti-
culture of the region can be undertaken by the expezKr^i^"^^^^'
stations of the northwestern states than the develoX^-'^^f^''
of apple varieties of high quality and perfect adap>*"^*^^°
to the various sections of their respective states.

The apples of high quality which show a fair deg'X*^^
adaptation to the irrigated sections are Esopus, "ST^J^^^
Newtown, Delicious and White Pearmain. The last^ ^^^^
found by Lewis, of the Oregon station, to be one o:f ^^
best poUenizers on every variety tested. Joaatl^^/
Winesap and Stayman, though largely grown, sho^i^^ ^^
general a poorer balance of characters. In the more

Digitized by

Google /

No. 574] VARIATION IN THE APPLE 62T

elevated valleys Wagener, Delicious and Mcintosh are
doubtless most worthy of culture. The highlands of
eastern Washington are very poorly adapted to the grow-
ing of winter apples, though some of the early apples do
fairly well, among which may be mentioned Oldenburg,
Gravenstein and Yellow Transparent. On account of
the abundance of sunshine the Oldenburg develops a high
sugar content for the variety which counteracts its natural
acidity and results in an apple of pretty good dessert
quality. Of the winter apples, Eome reaches good
marketable size but the quality is not high and the east-
em market should not be jeopardized by shipping this
variety. The Palouse, an apple of local origin, is of
much better quality, but has little standing in the market
as yet. The Dutch Migonne, a variety from western
Europe, shows a better balance of characters in eastern
Washington than in most other sections of this country.
It is of good size, fairly well colored and excellent in
quality.

Many varieties popular in the eastern states color
better west of the Cascades than in eastern Washington,
though there is usually manifest a lack of balance in
other characters. In certain respects the environment
resembles that of western Europe and many of the va-
rieties of cherries, plums, prunes, and other fruits of
that country do very well here and, indeed, in other sec-
tions of the state as well, though in a number of instances
varieties of northwestern origin are gaining in favor
rapidly. Apple breeding, however, requires more time
for its accomplishment and further importations of
fruits, especially apples, adapted to the mild climate of
western Europe would no doubt prove an advantage
through the possible discovery of sorts adapted espe*-
cially to the western part of the state.

Quality is not in itself a simple character. It depends
upon all the characters of the flesh which determine the
desirability of the fruit for eating, such as texture, juici-
ness, aroma and flavor. Fineness of texture evidently
depends upon a proper combination of favorable factors.

Digitized by

Google

628 TRE AMERICAN NATURALIST [Vol. XLVIH

Conditions favoring rank growth result in coarse texture,
as was observed in several instances in the case of apples
grown under irrigation, especially if the fruit was over-
grown. Some of the fruit from young trees also was
overgrown and coarse. Tenderness depends upon the de-
velopment. Poorly grown, under-developed fruit grown
where the temperature is too low or the season too short
for the variety has hard flesh which becomes spongy
rather than mellow toward the end of the storage season.
Overgrown fruit of certain varieties, on the other hand,
often shows lack of coherence between the cells, often ac-
companied apparently by larger intercellular spaces, and
jsuch fruit tends to become mealy as the season pro- |

gresses. Juiciness is primarily a manifestation of the J

amount of moisture in the fruit, but is also associated »

with the tenderness of the cell walls and their tendency
to break rather than to separate. In general an ^^"^l^g^
dance of moisture results in juicy fruit though the ^ ^ '■

ness is not in proportion to the moisture supply. Th^ •^^^^^^^^' '

stances which give the apple its aroma are pres^^^^^^ ^^^
such small amounts that their investigation is dl^jfj in

They are volatile compounds and affect the flavor '€z>:f ^^^'
apple largely by their action on the sense of sm^Xl,
cool climate is favorable to their production and i-fc -^^
often observed that they were most strongly develo :^=^^<3^ :
the apples from elevated situations. Flavor d^^>^j,^
upon the kinds, amounts and relative proportions -^z^f the
soluble solids, especially the balance between suga:BC!^ ^ and
acids, and will be given further consideration in tl::». ^ dig- i

cussion of the chemical composition. Immatur^-^^ and
under-developed apples contain some tannic acid, ^^iT^hich
is often sufficient in amount to give an astringent (^Tti^^rac- •

ter to the fruit. '

^ Keeping Quality. — In its relation to the enviTo:M:^TX2entf .

keeping quality evidently follows the same rule as other
variable characters of the apple, namely, that for ^ny
variety the keeping quality depends upon the optimum
intensity of the various external factors. Apples grown \

where the temperature is too low or the season too short

Digitized by

Google

No. 574] VARIATION IN THE APPLE 629

to develop the fruit to a proper stage to keep well, soon
wilt, lose flavor and scald, or show other evidence of de-
terioration as was frequently observed in the fruit from
high altitudes. On the other hand, too great excess of
certain factors results in overgrown or overripe fruit
having a tendency to rot, mealiness, or physiological de-
cay, as in the case of the Yellow Bellflowers from Puyal-
lup and some of the fruit from the warm valleys. The
balance of factors favorable to good keeping quality does
not appear to differ much from that which produces the
fruit which is most desirable in other characters, though
it is possible that the required intensity of some factors
may be slightly lower. It appears, therefore, that a good
balance of the other characters of the fruit and perfect
adaptation to the environment will be accompanied, as a
rule, by good keeping quality, provided that the fruit is
properly handled and not infected with disease, while aa
unbalanced adaptation of characters to environment ia
likely to result in poor keeping quality. It seems prob^
able that irrigation in itself does not result in poor keep-
ing except when improperly applied or carried to excess:
or associated with other factors in such a way as to de-^
stroy the equilibrium of the environment. The relation;
of specific gravity to the keeping quality is discussed in
a succeeding paragraph.

Specific Gravity.— It has long been understood that
varieties of apples differ in their relative weights; thus
Wolf River is comparatively light and Baldwin is gener-
ally regarded as a heavy apple. The only record found
of the determination of specific gravity of apples is that
of Howard's work in the National Bureau of Chemistry,
Bulletin 94, in which it is noted that the specific gravity
diminished 3 per cent, to 5 per cent, during storage.
From the account it is not clear whether the determina-
tions at the different dates were made with the same
apples. The decrease of specific gravity is ascribed to the
increase of air spaces between the cells due to the soften-
ing of the middle lamella. In the specific gravity determi-
nations made by the writer a number of points was noted.

Digitized by

Google

630 THE AMERICAN NATURALIST [VoL.XU'-^V^rj-ji

The different lots of a variety may differ considerabB^-^y- \^
specific gravity, though as a rule running somewhat ci^-Tl^oae
together, thus Ben Davis and Gano are apples of low s^;t)e-
cific gravity, while Grimes, Stayman, Wagener, and Y^=^ ^^ ^^
Newtown run rather high and Baldwin and Rome m^_ ^^^^ ]^
classed as medium in this respect. Overgrown afz:^ 3|p>^eg
were low in specific gravity, probably owing to mor^^^r ^[^
space between the cells. This is more apparent 'm.zi^.^f^on
examining the results for individual apples than "mj=».^poii
comparing the average for different lots, as in the l-^^^^tter
case the extremes are modified by averaging with tlm. -^^ le-
sults for more normal specimens. On the other tzM^-^i^nd,
small and rather undeveloped apples are apt to ha^^^^?-c a
high specific gravity on account of their solid flest*^ land
usually closed core. Juicy apples, if not overgrown, ZMoM^SLve
a high specific gravity when the juiciness is due to £u Hxigh
moisture content.

The relation of specific gravity to keeping quality i s of
interest. While some late keeping varieties have xior-

mally a low specific gravity, those lots of a given v^ik: jriety
having a high specific gravity for the variety are u^ -•J-Silly
the best keepers. This is in line with the fact that c^ ^mr±. ain
■causes which give rise to fruit of poor keeping qmii«-^iJity
also produce a low specific gravity. This is shown^ ^v^ery
strikingly by a comparison of the specific gravities am- ^s <5al-
culated month by month through the season. A^ ^3 the
calculations were made at the time the fruit was z^" ovinia
fit for use, the monthly averages show the steady inc^^^*^^^^
in specific gravity with the better keeping quality ^z^^^ ^^^
fruit, though modified somewhat by the peculiarit i ^ ^ ^'
the different varieties which happened to be in sea^^ ^:>xi at
different times. These averages are as follows: N<z^^^^^^'
ber and December, 0.787; January, 0.787; Feb^r^'*^-^^'''
0.810; March, 0.831; April, 0.852. Though these j-^^tJ^s
may seem to be at variance with Howard's obserir.^t'*^^^^
it is possible that if the same specimens had been -fc^^^^
at intervals a decrease in specific gravity would ii^ve
been noted.

Chemical Composition.— In order to throw some HS'^^

Digitized by

(

Google y

No. 574] VARIATION IN THE APPLE 631

if possible, upon the relation of chemical composition to
the other characters of the apple and to determine
whether the composition is influenced by the environment,
the juice of the various samples of Grimes, Jonathan,
Winesap and Yellow Bellflower was analyzed by the de-
partment of chemistry.

The juice of the Grimes and Winesap contains, as a
rule, a decidedly higher percentage of total solids than
that of the Jonathan and Yellow Bellflower. It is also
generally higher in specific gravity and has a greater
viscosity. In Grimes and Yellow Bellflower the juice of
the eastern-grown fruit contains a large proportion of
total solids than that of the Washington grown fruit,
though this rule does not hold good in the other varieties.
The apples from the irrigated valleys and western Wash-
ington were low in total solids with the single exception
of the Winesaps from Cashmere. The analyses fail to
show any constant difference in sugar content in favor of
the fruit produced in the sunny climate with long hours
of daylight characteristic of the apple-growing sections
of the state.

In Grimes the total sugars are fairly high and the pro-
portion of sucrose is especially large. The acid content,
on the other hand, is low as a rule. The result is a rich,
mild or nearly sweet flavor. A sample from Puyallup
showed the lowest sucrose content combined with the
highest acid content, and this was the least rich as well as
the most acid in flavor.

, Jonathan, on the other hand, displays a low content of
total sugars and especially sucrose, while the acid content
is slightly higher than in Grimes, indicating a subacid
apple, lacking in richness. The lots from Missouri and
Indiana were highest in sucrose but were of scarcely as
good quality as the Jonathans from West Virginia. The
latter were low in both sucrose and acid, but displayed a
good balance between these constituents, indicating an
apple with rather thin juice, not very rich, but pleasant
and refreshing. Its evident superiority resulted largely

Digitized by

Google

632 TEE AMERICAN NATURALIST [Vol. XLVIH

from the fine texture and well-developed flavoring con-
stituents not shown by the analysis,

The Winesaps, though high in total sugars, are low in
sucrose, indicating a heavy juice rather lacking in rich-
ness. The comparatively high acid content corresponds
to the sprightly subacid character of the fruit. The high-
est acid content was found in the fruit from Cloverland,
where it is associated with a total lack of sucrose result-
ing in a comparatively poor fruit. The apples from
Cashmere and White Salmon were also devoid of sucrose
in the juice, but the acid content was low and the flavor-
ing principles well developed, as a result of which the
quality was fairly good. The poorly developed Winesaps
grown at Pullman were deficient in sucrose, acid, and
flavors and were correspondingly poor in quality.

The Yellow Bellflowers, though low in total sugars,
were rather high in sucrose and also in acid. The bal-
ance between these constituents is good and results in a
moderately rich, pleasant, subacid flavor.

SUMMABY

The opportunity for the study of apple variation was
unusually good, owing to the facilities afforded for the
examination of fruit from various localities and different
environments, and it has been possible to work out the
fundamental principle upon which variation resulting
from external factors depends and to apply it in the study
of environmental adaptations. This principle, the Law of
the Optimum, states that, for any given variety there is
for each character a certain intensity of each essential
factor of the environment at which, other conditions re-
maining the same, that character reaches its highest de-
velopment.

In the application of this law to varietal adaptations,
the essential point is the proper balance between char-
acters and environmental factors, that is, all factors
should be of such an intensity as to permit a good all-
round development of the fruit. In the absence of such

Digitized by

Google

ilo.574] VABIATION IN THE APPLE 633

a balance certain characters may fail to reach a proper
degree of development while others develop to excess.

The failure in quality and other respects of many of
the best dessert varieties of apples when grown in Wash-
ington is due to such a lack of balance. Practically all of
them originated under a much different environment and
were selected and came into prominence owing to their
perfect balance of adaptation in localities having a set of
external conditions similar to those under which they
originated. The hope of northwestern apple culture in
the future lies in the careful selection of varieties and the
origination locally of varieties of high quality showing
adaptation to the conditions of growth in the various sec-
tions. In the meantime plantings should be made from
those varieties of high quality which show the best
adaptation. These are Esopus, Yellow Newtown, White
Pearmain and Delicious for the irrigated valleys, and
Wagener, Delicious and Mcintosh for the higher valleys
of northern and eastern Washington. Jonathan, Stay-
man and Winesap show a poorer balance and should not
be planted too recklessly. The climate of the Pacific
coast resembles that of western Europe more than that
of the eastern states, and further importations of Euro-
pean varieties is desirable especially for testing west of
the Cascades.

The moisture relation is probably the most important
factor in inducing variations, and is doubtless responsible
for certain variations which have been ascribed to other
causes which act indirectly by modifying the moisture
supply. The elongation of the fruit following a cool
period after blooming may result from a diminished cir-
culation of the sap, giving rise to an insuflBcient supply to
provide for the simultaneous development of the fleshy
portion and elongation of the axis. Variation in the
depth of the cavity and basin in certain varieties is prob-
ably to be explained in a similar way.

Color modifications depend to a great extent upon, the
light relation and somewhat upon development as influ-
enced by temperature. The optimum intensity for the

Digitized by

Google

634 THE AMERICAN NATURALIST [Vol. XLVin

production of red pigment is quite narrow in most varie-
ties and poor color may result from either deficiency or
excess. Latitude and altitude affect the color only as
they modify the factors upon which color depends, caus-
ing them to approach or recede from the optimum. The
influence of elements in the soil is not well understood.
It is probable that soils containing sufficient iron for the
proper development of chlorophyll contain an abundance
for the production of red pigment in apples.

Aside from such differences as depend upon the hand-
ling of the fruit, variations in keeping quality appear to
follow the law of the optimum in the same manner as the
other characters of the fruit. Conditions which favor the
best all-round development result, as a rule, in good keej)-
ing quality. Apples grown under irrigation are said to
keep poorly probably because of their unbalanced adapta-
tion to the environment. Certain factors which favor de-
velopment and maturity are present in excess, resulting
in overgrown or overripe fruit.

Varieties differ in specific gravity according to the
extent of intercellular spaces in the flesh and the open-
ness of the core. Overgrown specimens are low in speci-
fic gravity. As a rule, those lots which kept best in any
variety had the highest specific gravity.

Chemical composition is associated somewhat with
quality. High sucrose content results in richness of
flavor. Fruit of high quality has the sugars and acids
well balanced and the flavoring constituents well devel-
oped. A heavy juice is usually associated with a high
content of soluble solids. Fruit grown under irrigation
is ordinarily rather low in soluble solids. There seems
to be no constant relation between the amount of sunlight
and the production of sugars, and flavors appear to de-
velop best in a relatively cool climate.

Digitized by

Google

SHOETEE AETICLES AND DISCUSSION

VARIATION AND CORRELATION IN THE MEAN AGE
AT MARRIAGE OF MEN AND WOMEN

Somewhere in sociological literature we have met with the
statement that whereas the mean age at marriage of men diflfers
from district to district because of social and economic conditions,
the mean age at marriage of women varies but little because of
these factors. In view of the high ''assortative mating'*
coefficient^ for age of bride and groom, this statement seemed so
remarkable as to be open to question.

Its validity can be very easily tested provided the mean age at
marriage of men and women from a series of districts differing
in economic and social conditions are available. If the mean age
of women is independent of these conditions, or far less depend-
ent upon them than that of men, one should find (i) that the
variation of mean age of brides is lower than that of mean age
of grooms, and (ii) that for a series of districts the coeffiicent of
correlation between the mean age of brides and grooms is very
low.

The only suitable series of data that we have been able to find
is that given by A. Dumont* for the average age in years and
months at first marriage of the males and females of the 87
departments of France. Grouping his data in classes of five
months' range, we find, in terms of months:'

1 See Lutz, Science, N. S., Vol. 22, pp. 249-250, 1905. For a general re-
view of the literature of assortative mating see Harris, Pop. 8ci, Mo., Vol.
80, pp. 476-492, 1912.

2Dumont, A., Bev. Ecole Anthrop, Paris, Vol. 14, p. 163, 1904.

8 The results given by the ungrouped data are :

For Men For Women

Mean 337.87 ± .83 284.45 ± 1.01

Standard deviation 11.49 ± .69 14.00 ± .72

Coefficient of variation 3.40 ± .17 4.92 ± .25

Considering the shortness of the series, the results are in as good agreement
as could be expected.

For Men For Women

Mean 337.76 ± .80 284.43 ± 1.03

Standard deviation* 11.03 ± .56 14.25 ± .73

Coefficient of variation 3.26 ± .17 5.01 ± .26

-* Without Sheppard's correction for the second moment.

635

Digitized by

Google

636

THE AMERICAN NATURALIST [Vol. XLVHI

-3I(

y^ f

X 1

/^ /

-30

^-7^. /

-J-

/ W

-»•<

/^ 1

/ TT

-M^<

-aT(

-am

•

I— i

ift

3:;

Ift

3|]

Mean Age of Men.

We note that the women many on an average about four years
and five months younger than the men. Contrary to what we
have been told, their mean age at marriage both absolutely, as
measured by the standard deviation, and relatively, as measured
by the coeflBcient of variation is more variable than that of men.
The diflferenee in standard deviations for the ungrouped material
is 2.51 It .93 and for the grouped records 3.22 db .92. These are
2.71 and 3.49 times their probable error, and hence perhaps
significant. For the coefficient of variation, the differences by
the two methods are 1.52 it .31 and 1.74 ±l .31. These are 5.69
and 4.96 times their probable errors and their significance is
even more probable than those for the standard deviations.

The correlation coefficient from the grouped data by the prod-
uct moment method, using the means and standard deviations
given above, is

• Digitized by

Google

No. 674] SHORTER ARTICLES AND DISCUSSION 637

rmf = . 781 ±.028.^

Thus on a scale of — 1 to +1 the interdependence of mean
ages of men and women is very close indeed.* Expressing the
same relationship in terms of regression by the well known
formula

\ <ri» / (Tin

where the bars indicate population means and the sigmas popula-
tion standard deviations of m= males and /= females,

/=— 56.474 + 1.009 m.

Thus we see that each month's increase in average male age is
followed by a month's increase in mean female age. The fit of
the straight line to the empirical means as shown in the diagram
is excellent — considering the small number of the district means
from which the equation is deduced.

Thus the available data show that the mean age at marriage of
women instead of being less variable from district to district than
that of men is actually more variable — ^both absolutely and rela-
tively.

In short, there is, as far as our data go, no evidence for the
assertion that while the time of marriage of men is closely de-
pendent upon the complex of social and economic conditions that
of women is practically independent of them.

We have published this note in the hope that it may suggest
to some one with the opportunities of obtaining really adequate
data an investigation of the problem which has several rather
important points of interest.

J. Arthur Harris,
RoxAKA H. Vivian

Cold Spbino Habbob

s Tbe difference method applied to the ungrouped material gives

r=. 763 ±.030.

The difference is of no significance.

« Possibly, however, the relationship is in part spurious. The mean of
males and females were taken on the basis of the same N, or approximately
the same N, for the various districts. Data for investigating this question
are not available. The point should be borne in mind by a subsequent
worker.

Digitized by

Google

638 THE AMERICAN NATURALIST [Vol. XLVIII

DUPLICATE GENES

Some interesting questions are raised by a recent article by
Gregory: **0n the Genetics of Tetraploid Plants in Primula
sinensis,"^ Reciprocal crosses of two races of P. sinensis were
made. One cross gave entirely normal results in Fg as regards
chromosome number and hereditary characters. The reciprocal
cross gave an Fj generation which was sterile with the parents
and produced only a giant variety in Fj. This proved to have
the tetraploid chromosome number. Experiments indicated that
the genetic factors had also all been doubled, a very significant
parallelism.

Gregory uses the nomenclature A AAA, AAAa, AAaa, Aaaa,
and aaaa to represent all the possible conditions as regards a
pair of Mendelian factors. He states that heterozygotes of the
form AAAa should give gametes AA and Aa, and should pro-
duce, on selfing, the zygotes AAAA, 2AAAa and AAaa, and that
the last class selfed should produce recessives. On the chromo-
some theory of heredity, this assumes that the four chromosomes
concerned are equally likely to pair in synapsis in any of the
possible ways, a very interesting phenomenon if the assumption
proves correct. But it is conceivable that two independent
synaptic pairs may be formed. It may be that only chromo-
somes from the same original race pair in synapsis. It is true
that the first of the original crosses shows that the chromosomes
of the two races can enter into normal mitosis and presumably
into synapsis with each other. But the reciprocal cross indicates,
perhaps, that in the environment of the cytoplasm of this cross,
they can not enter into synapsis. If this condition continues in
later generations, we should represent the zygotes as AAA' A',
AAA'a', AaA'a', etc. This is the way in which duplicate genes
have been represented previously as by Nilsson-Ehle, East and
ShuU. With this representation, heterozygotes of the form
AAA'a' could neVer give rise to recessives after selfing for any
number of generations.

Which hypothesis is true in this case could easily be deter-
mined by experiment. The published results are not sufficiently
explicit on this point. If the original cross were of the type
AA X a'a', producing in Fi Aa', the Fg, AAa'a', would be a
homozygote on the second hypothesis, and recessives should never

1 Proc. Boy, Soc, B 87, 1914.

Digitized by

Google

No. 574] NOTES AND LITERATURE 639

appear. On Gregory's Tiypo thesis reeessives should appear in
later generations. On the second hypothesis, homozygous races
of the types AAa'a' and aaA'A' would be obtainable, in appear-
ance like heterozygotes. These would breed true indefinitely
when self ed, but should give reeessives in Fg after crossing, as in
a case proved by Nilsson-Ehle.

Sbwalii Wright
BussEt^ Institution,
Forest Hills, Mass.,
June 19, 1914

NOTES AND LITEEATUEE
A STUDY OF DESERT VEGETATION^

Between three and four years ago Dr. W. A. Cannon, of the
Desert Botanical Laboratory at Tucson, Arizona, visited southern
Algeria in order to become acquainted with the more obvious
features of the plant physiological conditions of the desert, and
to make detailed studies of the root habits of certain desert
plants. Prom Algiers the journey proceeded nearly due south
about three hundred miles to Ghardaia, thence east about one
hundred miles to Ouargla, and another hundred miles to Toug-
gourt, returning through Biskra, and Batna to the northern coast.
Throughout this long and wearisome journey the vegetation was
studied in connection with the geographical and climatic environ-
ment and the results are brought together in a volume of some-
what more than eighty pages of text and thirty-seven plates, one
of which is an outline map of the region visited.

Dr. Cannon speaks of the similarity of the flora of Algeria to
that of southern Spain, France and Italy, where one is reminded
of the yegetation of portions of California. Once in the desert on
the way south low-growing shrubs on the plain become char-
acteristic, including species of Tamarix, Zizyphus and Artemisia,
Where water is available for irrigation, oases occur with their
luxuriant vegetation of date palms, apricots, figs, mulberries,
peaches, pears, oranges, as well as artichokes, beans, carrots,
melons, peas, potatoes, squashes, etc. Further south the plain

1 Botanical Features of the Algerian Sahara. By William Austin Cannon,
Washington, D. C. Published by the Carnegie Institution of Washington,
1913.

Digitized by

Google

640 THE AMEBIC AN NATURALIST [Vol. XLVHI

is covered with small stones and pebbles and ''not a tree, shrab,
or herb appears to hide the bare ground. The moimtains are
naked rock, while the harsh outline of desert ranges and the
distant low sand ridges give no evidence of plant life. But a
closer examination of plain, dune and mountains reveals the
presence either of living forms or of the dried remains of plants
of a preceding moist season, in numbers and in kinds not at first
suspected.'' All of which might well describe the desert condi-
tions in our own southwest. This similarity is emphasized by the
resemblance of many of the plants to those found in our Arizona
deserts. Thus the **quidad'' (AcanthylUs tragacanthoides) **has
a very close resemblance to small specimens of 'ocotillo' (Fott-
quieria splendens) of the southwestern United States." And
this resemblance extends to the structure of the spines and the
return of the foliage after rains. It is interesting to note that
the natives bum off its numerous spines, after which the stems
'*are eaten with avidity by camels,*' reminding us of the similar
treatment and use of some cactuses in Arizona. Further to the
south the vegetation is still more sparse and xerophytic, includ-
ing Ephedra, Retama, Haloxylon, and among grasses, Arisiida
pungens. Near Ouargla, the southern point reached, there are
places where no vegetation is present, as on the dunes, and yet
on ^ the fixed sand nearby were found Euphorbia guyoniana,
Retama retam and Oenista saharce.

Much attention was given to the root habits of the plants
encountered, and in the general summary which follows the
account of the journey comparisons are made with the root habits
of Arizona plants.

With this meager introduction we must refer the reader to the
volume itself, which it is quite impossible to summarize in these
pages. One thing impresses itself forcibly upon the reader, and
that is that a desert is a hungry place in which the permanent
vegetation maintains itself against plant-eating animals by a
thorny or spiny protection. Yet Dr. Cannon points out that in
this character of spininess the American desert plants excel those
of the plants of the Sahara region.

Charles E. Bessey

The Univebsity of Nebraska

Digitized by

Google

Tvv^o Important Biological Journals

THe Journal of Qenetics

Edited by
W. BATESON and R C. PUNNETT
A periodical for the pnblioation of records of original research in heredity, variation, and allied topioe.
The JoiT&KAL also oontains, from time to time, articles summarizing; the existing state of knowledge in the
TarioQB branches of genetics. The numbers are fully illustrated, and, when the subject-matter demands it,
colored plates are freely employed. The Journal is issued in parts as material accumulates, and a volume
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7.50 a volume ;
■ingle eoines, $2.50.

3iometrik;a

Edited by KABL PEARSON
A journal for the statistical study of biological problems, appearing about four times a year. A volume
containing about 500 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
nine preceding volumes form a veritable encyclopaedia of biological statistics in the fields of anthropometry,
oraniometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, 12.50.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes
and single copies.

Published in America for the Cambridge University Press of England by

The University of d^icasro Press
Chlcasro^ Illinois

University Control

By J. MoKebn Cattsll, Professor of Psychology in OolombiA Uniyendty

Tocsthsr with a series of Two Hundred and Ninety-nine Unsigned Letters by Tieading Men of Bdit
hnKHng Aoademio Positions and Articles by Jossfh Jabtbow, GnoBOS T. Laod, John J. Orrnvwanam, J. B.
CamuHTOM, J. MoEsair Oattsll, Qsobos M. SvjuTToy, Stswabt Patoh, Jobs Jat Chapkah, Jakm P.
liinnK>B and Jaoob Oould Sghitbican.

A great Tariety of questions oonoeming general univenlty administration are dealt with in an original and
kelpfol way. — Nature*

Tbese gnotations and examples are taken from Professor GatteU's informed and thorough disoossion of tha
flobjeot of university control, a subject upon which he has had much to say of late, finding occasion for oaustia
crftioiflm of eristing American oouditions, and standing as the ohampion of an academic demoorao7 i^d a teaeh-
ins profession upon which a man may enter without forfeiting his self-respect. — Th§ DiaL

Sentences and paragraphs that betoken the expert, highly-trained mind, the suggestions that oome to M-
iwmmt a d teU us that a new day is about to dawn in educatioual writing.— TAs Badon Sunimo TnmeeripL

SCIBINCB AIND EDUCATIOIN

A series of volumes for the promotion ofacienti£o research and eduostionul progress

VOLUME L The Poimdatioiis of Science. By H. PomoAai. Containing the authorised English translatioB
1^ QaoBaa Banoa H.4Lstbd of "Science and Hypothesis," "The Value of Scienoe," and "Sdenet
and Method.'* Price, $3.00 net.

VOLUME n. Medical Research and Sducation. By Rioeabd M. Pbabcb, William H. Wbloh, W. H,
HowBLL, Fbahklih p. Mall, Lbwbllts F. Babkbb, Chablbs S. Minot, W. B. Camnon, W. T«
GeuMOtucAir, Thbobald Smith, G. N. Stbwabt, C. M. Jackson, E. P. Lton, Jambs B. HBRBinB, Jobs
M. DonsoN, 0. R. Babdbbn, W. OphOlb, S. J. Mbltxbb, Jambs Ewino, W. W. Kbbn, Hbhbt H. Dovauh
00V, Cbbimiah a. Hbbtbb, and Hbmbt P. Bowditch. Price, $3.00 net.

VOLUME m. nnirenlty Control. Now Beady. Paget x-\-4S4> Price, tSM mL

QARRisofi. N. Y. THE SCIENCE PRESS lancaitbr. pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The American Naturalist

A Moathlr Joanal, MtaUkbed in 1867.
wi& Spwaal Rmtunmem to IIm

to Um AdvancMBMit of tk
of Orsasie Erololioa and

CONTENTS OF THE APRIL NUMBER

The Origin of X CapMlla Bursa pastorif araohnoldea.

Dr. Henri Hub.
BlolofyoftbeTbyfMK^tera. II. Dr. A. Fruiklin

8hull.
Shorter Artielet and Difonition : Barritn m to Dit-

tribntlon m regard! Blrdi and Mammalt. Joseph

Orinnell. Yellow Varletiee of Bate. Profeetor

W.E. Castle.
Notes aM Literature : Heredity and the Influence

ofMoaarchs. Y. L. K.

CONTENTS OF THE MAY NUMBER

Xetoparasitesof Mammals. Professor YemoDLTmaa

Kellogg.
Begeneratlon, Variation and Correlallon in Tbyoaa,

Professor John W. Soott

Shorter Arttcletand Discussion: Vmna xelalfag to
Generic Types. Dr.O.F.Oook.

Notes and Literature: Linkage in th» SUkwonn
Moth. A. H. Bturterantw Nabonie's Breedli«
Brpertments with Qraeshoppera, Johna Dexter*

CONTENTS OF THE JUNE NUMBER

Spedes-lmildlng by Hybridization and Mutation. Pro-
fessor John H. GeronM.

Heredity of Bristles in the Common Oreenbottle Fly—
A Study of Factors goremlng Distribution. Fhlnees
W. Whiting.

Physiological CorrelationB and Climatic BeacticDS in

AlfiBlfk Breeding. Geo. F. Freeman.
Taxonomy and Evolution. By X.
Shorter Ajrtidles and DucuBslon* Nabours's Grass-

hoppen, Multiple Allelomorphism, Tankage and

Misleading Terminologies in GenetioB. Profiessor

W.E. Castle.

CONTENTS OF THE JULY NUMBER

Pattern DeTeiopment in M^"**«^'« and Birda.
Glorer M Allen

Internal Belations of Teneetrial JjMWiafeleiH.
Arthur G.VestaL

Shorter Articles and Discnsslott* Another Hy-
pothesis to AooounI for Di Swingle's Szpert-
ments with dtnu. A a and A. U Hagadoocn,

CONTENTS OF THE AUGUST NUMBER

Multiple Allelomorphs in Mice. Professor T. H
Morgan.

Thirteen Years of Wheat Selection. T. B. HuUh-
eson.

Pattern Deyelopment in Mammals and Birds.
GloTorM. Allen.

The Meadow Jumping Mouse. Dr. H. L. Baboock.

Shorter Articles and Inscussion :

Studies on Inbreeding. Dr. Baymond Pearl.
Parallel Mutations in (Enothera IriamU L. Dr.
J. Stomps, Dr. Bradley M. Dayis. The Theoret-
ical Distinction between Multiple AUelomorj^
and Close linkage. Professor T.H. Morgan. Pro
fessor W. E. Castle.

Notes and Literature:

Biometrics. Dr. Baymond Peait. A New Mod-
of Segregation in Gregory's Tetrapioid Primulase
Hermann J. Muller.

CONTENTS OF THE SEPTEMBER NUMBER
Studiet on Inbreeding. Dr. Raymond PearL
The Chromosome Hypothesis of Linkage applied to

Cases in Sweet Peas and Primula. CalTln B.

Bridges.
The Reduplication Hypotheaia ae applied to Droao-

phila. Dr.A.H. SturtOTant.
Pattern Development in if^mn^u uid Birda. Dr.

GlOTcr M. Allen.
Shorter Articles and Ocrrespondenoe :

The Bearing of the Selection Ezpedmenti of

Castle and PhiUips on the YarlabOity of i

Hermann J. Muller.

Sngle Number 40 Cents
TIm NATURALIST wffl b*

Yeariy SalMcriplioii, $4.00

lor four mooflia for Ob« DoBar

THE SCIENCE PRESS

GurmmifN. Y.

Sub-Sutkm 84 : NEW YORK

.Pb.

Digitized by

GooQle

LIU. irO. 575 HOVEMBEB. 1914

THE

AMERICAN
NATURALIST

A MOHTHLT JOTnUTAL

]>«?otad to the Advanoenient of tho Biologioal Soienoai wifli

Special Eefereneo to the Factors of Evolution

C0VTEVT8

Page
. A Comparlion of the BMponit a of Sosiilt and Motile Plants and Anlmali.

Professor VioroR £. Shblford 641

[. An Aptoroni Droioplilla and its Oenttio Beliaylor. Chablbs W. Mstz - 676

I. SliOTter Artlcloi and DiscnMion : Formnlse for the Eesults of Inbreeding :
Professor H. S. JxNNiNOS. A Short-oat in the Computation of Certain
Probable Errors : Howabd B. Fbost. Qalton and Disoontinuity in Vari-
ation : Dr. B. BuGGLBS Oatbs. The Osteology of a Donble-headed Calf :
ProfeBSor A. M. Bbesb --------_- 698

THE SOIENOE PEESS

LiVOASTEE, PA. QASBI80H, H. T.

HEW TOSK: BUB-BTAnON M

Digitized by

Google

The American Naturalist

MSS intended for publioatfon and beoke, etc., intended for review elionid bt
eentto tlie Editor of THE AMERiCAN NATURALIST. GarHson-on-Hndoon, New York.

Short articiee oontalnlng eaniniarlee of retearch work iiearing oa the
probienie of organic evoiotion are etpeelally weloome, and will be given preference
in pnbilcatlon.

One hundrea reprints of contributioae areenppiied to antboro free «if charge.
Further reprlnte will lie euppiled at coet.

Subeoriptione and advertieenients ehonid be sent to the pHbllshers. The
subscription price is four dollars a year. Foreiqn postage Is fifty cents and
Canadian postage twenty-five cents additional. The charge for single copies is
forty cents. The advertising rates are Fonr Dollars fbr a page.

THE SCIENCE PRESS

Lancaster, Pa. QarrlsoDt N. Y.

NEW YORK : Sub-Station 84

Bntered ai Moond-elaaa matter, April 2, 1908, at tha Poft Offloa at Lanoattar, Pa., ondar tha Aet ot

Congran of Maroh 8, 1879.

FOR SALE

ARCTIC, ICELAND and GREENLAND

BIRDS' SKINS,

Well Prepared Low Prices

Partloularo of

G. DINESEN, Bird Collector

Huoavlk, North Iceland, Via Leidle, England

WANTED TO PURCHASE

a aet of BIRDS OF AMERICA by J. J. Auauboii.
7or 8 volumea, pleaae report, atatinff caah price, atat-
ing condition, binding and datea of vohnnea.

P. C. HARRIS,
Box 2244 Boston, MatgachusetU

For Sale Entire

An important collection of Indian Birds'
Eggs, containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste
viridanus), Red-browed Finch (Calla-
canthis burtoni), Tibetan Sand-Grouse
(Syrrhaptes tibetanus), Spotted Sand-
Grouse (Pteroclurus senegallus), Ibis-
bill (Ibidorhynchus struthersi), White-
bellied Minivet (Pericrocotus erythro-
pygius), Bearded Vulture (Gypaetus
barbatus), numerous Cuckoos with their
hosts, etc., etc.

Particulars and full list may be had
from

W. F. H. ROSENBERG

57 Haverttock HiU London, N. W.

Marine Biological Laboratory

Woods Hole, Mast.

INVESTIGATION
Yaar

TtucOMm tat jaaaaroh !■ ga<aPt

INSTRUCTION

JalFltoAaawtll,

1914

PlfeyaiolQgj L
aw. BrnnaHj^mx privata labofa^
torie^ SlOO aaoh for not orar ihnm
montfaa. Thiitr ♦ahlaa ara avaS-
aMa for bogiiiDan in taatardb vbo
derira to woA under tba dJiaetiaa
of mambaia of tha atafT. Tba Im
f or auflfa a ttOila ia $50U)0.

Coonea of laboratocr

with laotorea ara offarad ia
brata Zooloaj, Sml

BcotofF of kadaa
Vagatation. Eaoh oooraa i
tha fun tima of tha atadani
$50.00. A laet«ica eoona
PhOosoiiliieal.A^actB af
aMl AuSdidaneaa ia afaa

SUPPLY

DEPARTMENT

Opaa tha Eatira Yaar

living, and la

PrMezrad matarial _
animalf and of Alcaai,

living matarial f unfahad in i
M ordarad. Prioe liM of
iogieal and Botaaloal

aont on appKeatkm. 8ta

la deairad. For gxrioa I ota and al
fakfonnation

GEO. M. QUAY, dnator. Woods Hole,
Tiia annual announoamant will ba anxt oi
Tha Diractor, M aiina Biologkal tabofatoay, Wi

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVni November, 19U No. 575

A COMPAEISON OF THE EESPONSES OF

SESSILE AND MOTILE PLANTS

AND ANIMALS

PROFESSOR VICTOR E. SHELFORD,

University of Illinois

I. Introduction 642

II. Basis of Discussion 642

1. Responses 643

2. Sessile and Motile Organisms 644

3. The Individual and its Relation in Colonies and Groups 644

(a) Animals and Plants made up of Single Individuals .... 644
(&) Colonial or Multiple Individualed Plants and Animals. 645

i. Numbers of Individuals 645

ii. Stems 646

iii. Metabolic and Reproductive Relations of Individ-
uals 647

i. Movements 649

ii. Structural Response 650

(<J) Response of Sessile Organisms 651

i. Structural RespouFes 651

ii. Movements 653

(/) Response and Taxonomy of Sessile Organisms 653

III. Parallelism between Sessile and Motile Organisms with reference

to Ecology 654

1. Breeding 654

2. Comparison of Sessile and Motile Elements of the Biota 655

3. Sessile Motile Organisms in Ecological Succession 656

rV. Influence of Response Phenomena upon Biological Theory and Con-
troversy 657

1. Topological View 657

2. Natural Selection View 658

3. Supposed Non-inheritance of Response and the Grerm Plasm

Doctrine 660

4. Influence of the Study of Response on Present-day Biological

Theory 661

5. Aspects of the Untenability of the Germ Plasm Doctrine 662

6. Measure of Values in Biological Science 664

7. Summary and Conclusions 669

V. Literature Cited 672

641

Digitized by

Google

642 THE AMERICAN NATURALIST [VoL.XLVni

I. INTRODUCTION

During the past few years the attention of hiologists
has turned more and more from those phenomena which
were supposed to be comparatively fixed, to responses to
stimuli. Physiologists have long been concerned with
the mechanism of response ; psychologists are interested
in its modification. Geographers, climatologists and
ecologists have recently turned their attention to re-
sponses in natural environments and zoologists have
become interested in response, particularly from the point
of view of its specificity. In these quite independent
investigations and compilations there has been little
attempt at analysis with a view to determine legitimate
lines of comparison among the exceedingly diversified
types of organisms which have been investigated, and
some confusion has resulted. For example, since the
more obvious responses of plants are structural, persons
not familiar with comparable phenomena among animals
have made erroneous comparisons of sessile plants and
motile animals. This paper is written to present in as
nearly uniform terms as practicable (a) analysis of kinds
or aspects of response, (6) justifiable kinds of compari-
son, and (c) the bearing of response phenomena on
biological theory and controversy. It aims to show that
the numerous kinds of response are reducible to a few
simple types common to both plants and animals, and that
the failure to consider all types has been responsible for
confusion and various one sided theories. It further aims
to show that study of response during the past few years
has led to an unusual broadening of our conceptions.

II. BASIS OF DISCUSSION

As a basis for discussion we must first have a clear
understanding of the character and definition of response.
Secondly, we must determine what constitutes an indi-
vidual in those plants and animals that are made up of
repetitions of parts. Thirdly, we must note whether or
not the organism is sessile or motile, capable of playing
the part of either, or colonial pelagic.

Digitized by

Google I

No. 575] RESPONSES OF PLANTS AND ANIMALS

1. Responses

The word response is used in various slightly different
senses. In general it refers to more complex and time-
requiring phenomena than * * reaction. ' ' In geography the
term has been used (Goode, *04) to cover all changes in
culture supposed to be produced by climate or other
geographic conditions. It is also applied by geographers
and geologists to changes in the physical characteristics
of man (evolution) which Goode ('04) has stated are
slower than the cultural responses. In general botanists
have used the term to cover changes of plant structure
and function induced by external conditions. Cowles
('11), however, uses the word ** reaction" to cover these
phenomena. Coulter ('09) used the term response as
synonymous with adaptation in plants. Zoologists have
used the term to apply to changes in animals due to exter-
nal conditions, but with little agreement as to what is to
be included. We will use it here to include reactions,
changes in functions, structure, color, 'induced by external
conditions either directly or indirectly, without regard
to how simple or how complex the processes involved
may be.^ The length of time required to bring the
changes about may arbitrarily be taken as not exceeding
the time required to breed five to ten generations of the
species concerned. All organisms respond to stimuli
because each stimulus acts upon some internal process.
Strictly speaking, the response is the change or changes
in the physical or chemical processes of the organism (or
the part or parts concerned) which results from the
disturbance.

Those things which we commonly see and term response
are often the later and less important phases of the dis-
turbance. The striking phases of responses of motile
organisms are usually movements which follow closely
upon stimulation. In sessile organisms the noticeable
responses often appear only after a considerable period.
In both sessile and motile organisms some responses are

iFor good representative bibliography see Adams, '13, Ch. VIII and
IX.

Digitized by

Google

644 THE AMERICAN NATURALIST [VoL.XLVin

not evident because they concern internal, chemical and
physical processes which aflfect neither form nor move-
ment. Changes in the enzymes secreted by digestive
glands, which accompany changes in food (Jennings, '06,
p. 347), are examples. While thus recognizing that re-
sponses are concerned primarily with internal processes,
we must of necessity refer chiefly to the external phases.

2. Sessile and Motile Organisms

Sessile organisms are those which are sedentary in
habit, whether attached or possessing slight powers of
locomotion. Motile organisms are those that habitually
move about. Vagile or creeping forms as well as swim-
ming, walking, flying, burrowing types are included.
Most sessile animals are capable of moving their parts,
while only a few sessile plants possess this capacity, and
these only to a slight degree.

There is no sharp distinction between sessile (seden-
tary) and motile organisms. Every possible gradation
exists between fixed non-motile types as trees on the one
hand and the pelagic fishes on the other. It is the
extremes which we will compare.

3. The Individual and Its Relations in Colonies
AND Groups

The following comparison of animals and plants is an
attempt to distinguish potential or incomplete individuals
in colonial organisms and compound organisms which,
while not comnjonly recognized as colonial, are made up of
incomplete individuals.

(a) Animals and Plants made up of Single Individuals
The vast majority of animals belong here. Most pro-
tozoa, solitary sponges, solitary hydroids, sea anemones,
worms not preparing for asexual division, echinoderms,
moUusks, arthropods and vertebrates. Only single-celled
plants, young seedlings and possibly a few adults of multi-
cellular plants which possess but one grotving point

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 646

(exclusive of roots) belong in this group. Single indi-
viduals as described here are the basis for determining
what shall be called individuals in colonial and compound
types.

(6) Colonial or Multiple Individualed Plants and Animals

A number of animals and the vast majority of the
plants belong here. The group can be roughly divided
into two types, (a) those having a chain or plate arrange-
ment of incomplete individuals and (6) those having a
branching or tree-like arrangement. The groups of in-
complete individuals of type a occur among the Protozoa,
worms undergoing asexual reproduction, many of the
Bryozoa and some of the Tunicates; both sessile and
pelagic (plankton) forms occur. On the plant side type
a includes plate-like colonies of algae, filamentous algae,
some thallose plants and probably some of the fungi,
though the great multiplicity of forms makes the separa-
tion of this group from the branching tree-like types,
diflBcult.

Type b includes some of the colonial Protozoa, the
majority of the sponges, hydroids, corals and the branch-
ing Bryozoa. The algae, fungi, mosses, ferns and flower-
ing plants are all represented. The colonies are usually
attached to the substratum (sessile).

i. Numbers of Individuals. — Among the animals the
number of so-called zooids is the number of incomplete
individuals. In the sponges there are as many zooids as
there are excurrent openings (osculae) (Minchins, '00,
p. 91). Zooids usually possess a mouth opening and
organs for securing food, though in some cases they may
be specialized for reproduction, defence or locomotion as
in some of the Coelenterates. Among the colonial plants
there are as many incomplete individuals as there are
buds or growing points (vegetative regions). There are
no regularly occurring organs in animals, strictly com-
parable to leaves. However, any organs such as tentacles,
gills, etc., which secure or absorb nutriment may be re-

Digitized by

Google

646 THE AMERICAN NATURALIST [VoL.XLVm

garded as analogous to leaves. Each potential fcwd with
its leaf may be compared to a zooid. In comparing plants
and animals, roots can perhaps be compared luitk the
holdfast organs of hydroids. In both groups, roots and
root-like organs are individuals of a very low order of
individualization and of a type not well represented
among animals. The holdfast organs of animals are not
important absorbers of food and water.

ii. Stems and Other Connecting Organs (Conducting
Tissues). — The most striking difference between the in-
complete individualed or colonial plants and colonial ani-
mals is the presence in the former of specialized stems and
highly complex conducting tissues (Cowles, '11; Piitter,
'11, pp. 361-66). The conduction of food materials from
the root to other parts of the plant and from the leaves to
the root is a functional necessity not paralleled even in
those colonial animals showing the greatest division of
labor. In animals stems are relatively undifferentiated
and are often made up of living, relatively unspecialized
zooids, as, for example, in many Bryozoa such as Crisis,
The tendency to cauliflory in some plants and the ability
of cambium to produce shoots and of the stems of most
hydroids to produce individuals indicates that such a con-
dition may be potentially present in all. In stalked
Protozoa the stems are solid, while in most Coelenterates
they are tubes, usually simple though sometimes complex,
made up by mere elongation and branching of the stock
of the simple single forms such as the Hydra. The lumen
is usually ciliated and makes possible a transfer of mate-
rial which renders practicable such division of labor as
occurs in this group (Piitter, '11). In the Bryozoa the
different zooids have their body cavities joined in the
simpler forms merely as a branching lumen of the main
wall of the colony ; in others by small openings the more
specialized of which are sieve-like plates (Harmer, '01,
pp. 471 and 496; Delage and Herouard, '97, Vol. 5, p. 62).

The connection between the individuals of the tunicate
colonies is often very complex, due to the fact that in the

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 647

most complex types the stolon (stem) gives rise to new
individuals and possesses all the layers of cells which
take part in forming them. The connection between
different individuals differs in different groups and is
determined by the particular mode of asexual reproduc-
tion. As the individuals are quite independent of one
another in function, these connections do not have the
same significance as in plants. Even where there is a
common blood circulation, as for example in the Clavel-
linidcB (Harmer, '04, p. 71), there is no noteworthy divi-
sion of labor.

iii. Metabolic and Reproductive Relations of Individ-
uals.— The flat worms at certain times consist of chains
of zooids at various stages of development and with
various degrees of independence. Child ( '13) has found
that these chains of zooids present a series of gradients
in rate of metabolic reaction. The rate is highest at the
anterior end of the whole chain and decreases toward the
posterior end, not uniformly, however, for the rate is
lower inamediately in front of each head region than it
is in the head region itself. A gradient is present in the
axis of each zooid. The most anterior head dominates so
long as the chain remains intact. In the corals certain
zooids dominate (Wood-Jones, '11) over the others.
Some types have a single dominant zooid and some more,
while in other cases all are equal.

Among plants whose form is that of a chain or a plate
the individuals are less closely bound together and domi-
nant vegetative regions are probably less well developed.
In the branching types, dominant vegetative regions occur
(Cowles, "11, p. 747; Goebel, '00, Vol. I, p. 206). In the
conifers, for example, there is a leader, a dominant grow-
ing region at the tip of the main stem just as in certain
madrepore corals (Wood-Jones, p. 83). Other plants like
the elm have several vegetative regions which dominate
over others, as they do in the branching madrepores.

Q-rowth form or colony form varies according to cer-
tain laws dependent, in part at least, upon the metabolic

Digitized by

Google

648 THE AMERICAN NATURALIST [VoL.XLVm

relations of individuals. Thus Wood-Jones says of the
corals —

a colony may grow according to five different types of vegetative growth
... it may grow as (1) a spherical mass, (2) an encrusting layer, (3) a
free plate, (4) a branching tree-like growth, or (5) a mere amorphous
lump.

He further notes the division of all the corals into two
groups of normal growth-forms; for all the zooids may
take an eqiud share in the asexual reproduction or, again,
some may be of greater importam^ce than others, and the
asexual reproductive functions may be lodged in a very
few individuals only. Considering the first division
(all zooids taking equal share, the principal types of bud-
ding vary from each other in the actual site of origin of
the daughter zooid from the parent, in the degree of final
separation of the two zooids, and in the thickess of the
intervening partition between the two zooids. The
amount of rising above the general surface by each indi-
vidual zooid is likewise subject to variation.

Turning now to the corals that constitute the second
class (some zooids of greater importance than others)
which in the words of Wood-Jones have some of their
units specialized as active agents of growth,

it is at once seen that the possibilities of variation of normal vegetative
habit are greatly increased. All the elaborate branching forms, plates
and leaf-like growths belong to this class ; and all are evolved by special
peculiarities of the growing point. The zooids that constitute the grow-
ing point may take various forms; they may be arranged as a cluster, as
a creeping edge, or as many varieties of terminal shoots of branches.

In the first instance, it is necessary to draw very sharp distinctions
between two subdivisions of this group. In Group 1 come all those
forms like Montipora, whose distal zooids are the newest formed mem-
bers of the colony; and in Group 2 are included the Madrepora, whose
distal zooid is the most ancient individual in the whole group.

In dealing with Group 1 many forms have to be considered, for when
the youngest are the active zooids their growth cluster may be variously
disposed, and on its disposition the resulting vegetative form entirely
depends.

In Group 2, however, this state of things is entirely altered, for there
one zooid, which is situated at the extremity of the stem, and which I

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 649

shall call throughout the " dominant apical zooid" constitutes the grow-
ing point; and this zooid is the parent of the entire colony.

Various writers make comparable statements or show
comparable principles among hydroids (Motz-Kossowska,
^08) and Bryozoa (Davenport, '91, et al.) and among
plants (Goebel, '00). Of the colony form of the tntiicates
Herdman ('04, p. 82) says:

The marked differences in the appearance of the colonies of conijxmnd
Ascidians is largely due to the methods of budding; even in those of
stolon type where the budding is practically the same in essential nature,
the results may be different in superficial appearance, according as the
buds are formed on a short stolon close to the parent body, or from the
extremity of the post abdomen or from the long epicardiac tube which
may extend for some inches from the ascidiozooid.

Thus we conclude that the innate causes of diflFerent
growth-forms (colony forms) of colonial organisms are
(a) the mode of division of the zooids or vegetative
regions, (fe) the ratio of stem elongation to number of
zooids or buds produced or uniformity or lack of uni-
formity of stem elongation (Wood-Jones, p. 76) closely
related to {c) the presence or absence, number, position
and region of influence of the dominant growing regions
or dominant zooids, and (d), in some cases, the grand
period of growth and the length period of the internodes
(Johnson, '11). The innate tendencies are thus reducible
to a few principles applicable to both plants and animals.

{c) Responses of Motile Organisms

i. Movements, — In motile organisms the most striking
responses are changes in position brought about by
movements usually more or less random, and which bring
the organism into various conditions one of which usually
relieves the disturbance. The organism resumes nonnal
activity in conditions which brought the relief (Jennings,
'06). These conditions are not necessarily advantageous,
but are usually so when the stimuli are those encountered
in nature (Mast, '11). Behavior of motile organisms is
also modified by repetition of action even in animals as
low in the animal series as the Protozoa (Holmes, '11).

Digitized by

Google

660 THE AMERICAN NATURALIST [VoL.XLniI

Jennings ('06) has quoted various botanical workers'
observations on motile plants the behavior of which prob-
ably follows the general laws governing the behavior of
motile animals. As a result of the quick behavior re-
sponses of motile organisms, their distribution at any
given time is a better index of the conditions at that time
than the distribution of sessile organisms, because when
the conditions at a given point become unfavorable the
motile organisms usually move to another situation,
while the sessile forms remain and perhaps die.

ii. Structural Responses. — ^Among motile animals,
structural and color changes occurring as a response to
environmental conditions (stimuli) are usually not of
importance to the organism concerned. The color differ-
ences induced in Lepidoptera by heat and cold (Stanfuss;
Fischer) and the structural differences in Crustacea such
as were brought about in Cladocera by Woltereck, and
other modifications brought forward recently, are usually
of no known advantage or disadvantage to the anhnals
concerned (Bateson, '13, Ch. IX and X). Such re-
sponses in color and general form do not ordinarily take
place in adults subjected to such conditions. The strik-
ing structural responses of motile animals are often
responses to the organism's activity. The use and disuse
phenomena of the Lamarckians, the increase in size and
form of muscles, thickening of skin in man and mammals,
are well-known examples of a type of responses which
have influenced zoological speculation. Child ( '04) con-
trolled the form of Leptoplana by controlling activity.
Holmes ('07) found that the movements of pieces of
Loxophyllum have an important part in shaping the
general outline of the bodies of the resulting forms. The
general forms of motile animals are correlated with their
activities but whether form or structure correlated with
it appeared first in the course of evolution has been the
subject of considerable fruitless speculation.

Digitized by

Google

No. 575] BESPONSES OF PLANTS AND ANIMALS 651

(d) Responses of Sessile Organisms

i. Structural Responses. — The striking phases of re-
sponses among colonial sessile organisms are often
changes in form and structure, or the relative position
of the parts. The changes in structure or position of
parts are not necessarily advantageous or useful, but are
usually so when the stimuli are those commonly encoun-
tered in nature (Cowles, ^11; Loeb, '06, p. 124; Wood-
Jones, '11; Ch. VIII). Indifferent and detrimental re-
sponses are often given under experimental conditions and
no doubt the absence of such variants among sessile ani-
mals collected in a wild state is due in part to the failure
of such organisms to survive. A few sessile colonial
organisms such as ca;cti (Cowles, '11) show little or no
plasticity.

Among sessile animals, the observations of Wood-
Jones form the best examples of response. He found
that the branching type of corals dominated in barrier
pools, tall slender non-branching types in deep water,
and massive boulder types on surf beaten shores. Thus
he figures similar colonies of each of three genera which,
while possessing certain peculiarities of their own, are in
general agreement as to growth form just as sessile
plants usually are ; and this in part for comparable rea-
sons. Thus various conifers occur as Krummholz in the
high mountains, due to severe conditions (Cowles, '11,
p. 732), wind, snow, and in part to the injury of terminal
growth regions of the main stem which gives rise to
lateral branches. The boulder-like corals with the zooid
at the same level occurring on the surf-beaten shores of
coral islands are due, in the case of Madrepora, for ex-
ample, to repeated injury of the terminal dominant zooids.
Conifers in protected situations often grow into tall
slender trees comparable with the (deep) still-water
corah. The barrier pools aflFord conditions where the
terminal buds are less often injured than in the surf and
the tree-like branching corals result from minor injuries
to dominant zooids.

Digitized by

Google

652 THE AMERICAN NATURALIST [Vol.XLVITI

Wood-Jones finds further that still-water corals are
less strongly calcified than those in rongh water, the
strains producing increased secretion analogous to in-
creased tissue production as a result of mechanical
strains in plants (Cowles, p. 669). Corals show diflFerent
kinds of growth under diflFerent environments partic-
ularly when injured. The new part may be diflFerent
from the rest and adjusted to the environment thus
making it appear as though two '* species" occurred in
the same colony. The mode of division of the zooid is
also diflFerent under diflFerent conditions. Plants show
similar variation with changes of conditions, particularly
in the leaves which are divided in submerged portions of
amphibious plants and entire in the emerging portions
(Cowles, '11, p. 595).

As has been noted, there is nothing in sessile animals
that is more than roughly analogous to leaves. Leaves
show marked structural diflFerences on diflFerent parts of
the same tree where the environmental conditions are
diflFerent, as, for example, in the diflFerences which occur
between the upper and lower portions of a forest tree.
While there are, no doubt, diflFerences in similar details
(histology) in the organs of display in diflFerent parts
of the same colony of sessile animals, little or nothing
has been done upon them. As a further indication of the
prevalence of structural response in sessile organisms
of the hydroids Hickson states that there is probably but
one species of Millepora which occurs in a large number
of growth forms. The commercial sponges (Moore, '08)
and common freshwater sponges and polyzoa show many
diflFerent forms under diflFerent environmental conditions.

The major diflFerences in growth form induced by ex-
ternal stimuli in colonial organisms result from modifica-
tions of the rate and character of growth with respect to
the four innate tendencies toward various growth or
colony forms discussed above, and which may be briefly
enumerated as follows: (a) mode of division, (b) amount
of stem elongation, (c) influence of dominant regions and

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 653

(d) grand period of growth and the length of period of
internodes.

The principles are concerned with asexual reproduction
and apply to motile organisms only exceptionally as for
example in the case of colonial pelagic forms. The laws
are applicable to both plants and animals.

ii. Movements. — Movements of sessile animals are
usually contractions or extensions of parts or of the
entire body. Tentacles and comparable organs are capa-
ble of movements for securing prey. Such organs often
tend to wrap about objects which are in motion. Many
sessile animals are capable of opening and closing a
mouth opening and of bending or twisting the entire body.
Plants possess a comparable capacity only occasionally.

(e) Behavior of Sessile Motile Organisms
Most sessile animals are capable of some movement
and r.eact by contraction of parts. The reactions may be
modified by repeated stimulation (Jennings, '06) and
usually by physical factors. Some animals, as Hydra,
Stentor and many others are both sessile and vagile or
free-swimming, and show different types of behavior
when attached and when free. Jennings states that such
protozoa have a more complex behavior than motile
forms. This is due to their combining the types of
behavior of sessile and motile animals.

(/) Response and Taxonomy of Sessile Organisms
Hickson ('98) has stated that there is but one species
of Millepore and believes that sex organs will be found
to be the best taxonomic characters. Wood-Jones states
that there are far fewer species of corals than has
formerly been supposed, and states further that growth
form can not be used to distinguish species. Among
fresh-water sponges and Bryozoa reproductive bodies
{gemmules and statoblasts) have been found to possess
satisfactory taxonomic characters. This is a situation
quite parallel with that in plants where reproductive

Digitized by

Google

654 THE AMERICAN NATURALIST [Vol.XLVIII

organs are used as classification characters. The ideas
of the reproductive organs of plants are now at the
*/ fixity'^ stage which on the animal side is paralleled by
the idea of fixed tropisms and fixed instincts, of a few
years since. Variability of tropisms is now well recog-
nized and reproductive organs in plants are being found
plastic, as those of animals will probably be found also.

HI. PARALLELISM BETWEEN SESSILE AND MOTILE OBGAX-
ISMS WITH REFERENCE TO ECOLOGY

From a summary of the considerations above it will
be seen that for practical comparison the division of
organisms into plants and animals may be abandoned and
only reference to sessile and motile organisms made. We
may now turn to a discussion of a few general principles
making the division into sessile and motile organisms only.
The behavior of motile organisms is plastic. There
are innumerable cases of modification of reaction by
variations of physical factors (Jennings, '06; Loeb, '06;
Mast, '11). If for purposes of discussion we put the
usual ''normal" reactions of motile animals over against
''normal" structure of sessile animals, we note that the
behavior response of the former parallels the structural
response of the latter.

1. Breeding

Motile Organism Fixed (Sessile) Organisms

(a) The breeding activities take (a) Breeding and other acti\'i-

place within narrower limits than ties within same limits, except that
any other activities. Merriam, dispersal may take place over wide
'90; Herrick, '02; Reighard, '08; areas through detachability of
Shelford, 'lla, b, c, '12a, h. seeds and other reproductive bod-

ies.
(h) The selection of breeding (b) Less marked because a se-

place and breeding activities, in- lection of abode by sessile organ-
eluding first activities of the isms takes place through the be-
young, are governed by the same havior of motile young stages or
general laws as other activities. through wide dissemination of non-

motile bodies by wind (etc.) with
growth under favorable conditions
and failure elsewhere.

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 655

probably least modifiable and least and early embryonic stages are
regulatory. less modifiable than the vegetative

parts.

{a,b,c) The maple tree, a sessile organism, is entirely
stationary in its adult stages. The seeds are blown by
the wind. One would not accomplish much in the study of
ecology by studying the distribution of the seeds of the
maple, or, on the other hand, by the study of the distribu-
tion of adult birds, without some further discrimination.

Sessile organisms are not diflScult to associate with
their proper environmental conditions in their adult
stages. As we proceed in our study to forms which can
move readily and rapidly, the diflBculty of associating
them with their definite environmental conditions in-
creases. Sessile organisms have stages which are small
and capable of easy dispersal, as in the case of the maple.
Sessile marine animals and some sessile plants frequently
have motile forms in young stages. In these motile
stages they are governed by the same laws as other motile
organisms. The conditions under which the motile stages
develop into the sessile forms are crucial.

Most fresh-water forms and some marine forms of
sessile organisms are without the free-swimming stage,
and they produce non-motile stages physiologically
comparable to the seeds of higher plants. The winter
bodies (statoblasts) of the Bryozoan {Pectinatella) com-
mon near Chicago, and which is a strictly sessile organ-
ism, are comparable to seeds and probably require
'* ripening '^ by cold, just as do many seeds and the repro-
ductive bodies of some other species of the same group.
Organisms which are highly motile in the adult stages
are not motile in the egg and young stages. The eggs
and young of birds, for example, do not move about, yet
birds are the most motile of all animals.

2. COMPABISON OF THE SeSSILE AND MoTILE ELEMENTS OP

THE Biota
(a) The motile organisms of a (a) The sessile organisms of a

given habitat usually react simi- given habitat (particularly plants)

Digitized by

Google

656

THE AMEBIC AN NATURALIST [VoL-XLVIU

(6) The various structuiraLl de-
vices which meet the conditions of
the environment are ecolog:dcally
equivalent.

larly to two or more stimuli not usually show similar functionai
differing greatly in intensity from rates, such as similar rates ot
their optimum, t. e,, the percent- transpiration among sand dtiU®
age of positive or negative trials is plants,
essentially the same for standard
intensities. There is also probably
similarity in the rates of metabol-
ism, etc.

{h) The specificities of behavior
such as the mode of moving the
organs, e, g., of locomotion, and
in some cases the combined results
of different behavior reactions are
similar and hence are ecologically
equivalent. The size and efficiency
of the organs are also involved.

A testing, for example, of the rheotaxis of a large num-
ber of brook- rapids animals has shown them to be
strongly positive, and when active individuals onl>' are
considered the percentage of positive trials is v^ery
similar for the entire rapids community. Likewise they
are in accord in their avoidance of sand bottom. Id^aiiy
of the animals have special means of attachment \^liicb
may be brought into play with speed.

As has already been pointed out elsewhere, ecological
equivalence is illustrated here. The darters (fish) ^^^
strong swimmers and are able to live in rapids by vjj^tVg
of their swimming powers and positive reaction, x^lj//
snails meet the same general conditions through po^itl^^
rheotaxis and the strong foot which enables them to hol(j
to rocks.

3. Sessile and Motile Organisms in Ecological

Succession
(a) Ecological succession is succession of ecological
(physiological) types over a given area, due to changes of
conditions which both cause migration of physiological
types and transformation of such types as remain (Shel-
ford, '11a, 'lib, 'lid, '12a, '12b and citations). Changes
of conditions are geographic, i. e., physiographic, climatic,

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 657

etc., and biological (due to organisms). Sessile plants
are the chief biological cause of successipn on land and
in fresh water, while sessile animals are the chief biolog-
ical cause in the shallow portions of the sea, especially in
coral reef regions (Wood- Jones, *11). Sessile organisms
are more important causes of succession than motile ones
because they (a) build up the substratum with detritus
and skeletons, (6) interfere with the movement of the
surrounding medium, (c) cut off light from the sub-
stratum where other organisms must reside and their
own young secure foothold, and (d) they usually iiffect
their own environments with excretory products more
than do motile organisms. In general we recognize
ecological succession of motile animals through the differ-
ences of behavior which accompany changes in conditions.
The differences are physiological ; differences in behavior
are the easiest index of the physiological condition. The
character of nests, burrows, etc., are often good indi-
cators also.

IV. INFLUENCE OF RESPONSE PHENOMENA UPON BIOLOGICAL
THEORY AND CONTROVERSY

A glance at some aspects of biological speculation
since before the publication of Darwin's ''Origin of
Species'' is essential to our understanding of the atti-
tude of biologists until recently, toward responses.

1. Teleological View
In the matter of animal behavior response, the earlier
workers interpreted the reactions as intelligent and pur-
poseful, ascribing human sensations, etc., to animals as
low in the scale as protozoa. This teleological tendency
was paralleled on the plant side by the idea of purposeful
adaptive responses. Many common plants respond
(structurally) readily to environmental conditions. As
has been noted, the commonest of the surviving responses
of the wild state are apparently advantageous. This led
some botanists to a Lamarckian teleological conception of
.response, perhaps best represented by Kerner and

Digitized by

Google

658 THE AMEBIC AN NATURALIST [Vol. XLVIH

Oliver's work on the natural history of plants. Accord-
ing to this view, responses are advantageous and for the
purpose of preserving the plant. Thus response and
adaptation become synonymous (Coulter, '08), a usage
quite inapplicable to animal structure. At the beginning
of the recognition of the response phenomena of corals
Wood-Jones takes essentially the view of adaptation
which botanists have tried and rejected.

Lamarck, who was for many years engaged in botanical
work, must have noted many cases of advantageous
structural response in plants. Later he undertook the
study of invertebrates which show great plasticity, and
was naturally much influenced in the development of his
theory of transmutation of species by the response phe-
nomena in the plastic organisms which he studied. Thus
the responses of motile (as well as sessile) organisms
which result from their own activities or the action of
their environments formed an important feature of
Lamarck's (Packard, '01; Cope, '96) theory of transmu-
tation of animal species. His theory is clearly in accord
with the material he studied most. The nature of his
contention and various well-known circumstances caused
his ideas not to be accepted.

2. Natubal Selection View

Characters used in classification of motile animals
before and since the time of Darwin are quite frequently
adaptation characters. Thus the large pectoral fins and
absence of an air bladder are characteristics of an entire
group of fishes, the darters. The divided eyes of the
Gyrinidce, which swim at the surface of the water, are so
adjusted that one half looks downward into the water,
and the other outward into the air. This character com-
bined with the paddle-like hind legs would have served to
distinguish the family. Again larvsB with a head and
thorax modified to fit a circular burrow and with hooks on
the dorsal surface of the fifth abdominal segment, which
is supposed to be an adaptation to prevent the animals

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 659

from being drawn from their cylindrical burrows by
their prey, could serve to distinguish the entire family of
CidndelidcB (tiger beetles). Such cases might be multi-
plied indefinitely.

Following Lamarck came Darwin, who, being more par-
ticularly a zoologist, was probably (proportionately, at
least) less familiar with structural response phenomena.
He was apparently impressed with the *' fixity" of the
so-called adaptation characters in motile animals, and
with the fact that they are often family, generic or specific
characters. With the assumption that they originated in
the environment in which they are now found, Darwin
and his followers on the zoological side credited '* natural
selection" of structural characters with the origin of
species. Though broader than Lamarck, this important
feature of Darwin's theory was quite clearly drawn from
data on motile animals. After the acceptance of Darwin's
theory, biologists were for many years engaged in elabo-
rating the ideas of phylogeny and natural selection by
working out recapitulations and homologies and by point-
ing out cases of adaptation. The investigation was
largely confined to the highly individuated animals. The
morphological method of this period, which indeed has
still continued in use among a minority of zoologists and
which finds a parallel in the recent morphological study
of the sex organs of plants, belongs to descriptive rather
than to analytical science. Since its conclusions are often
based upon the arrangement of species or of stages in
development into series chosen by the investigator, it is a
method which often allows free play of subjective fancy.
Thus unconsciously experimental study of modification by
environment became more and more neglected, and the
dominant type of investigation being such as to show
only the usual course of events in development, the ideas
of fixity grew more and more. Thus the fact that the
external form, structure and color of animals are not
easily, modified without careful experimental methods,
and that the structural responses of sessile animals were

Digitized by

Google

660 THE AMERICAN NATURALIST [VoL.XLVm

SO little known, resulted in structure in animals being fre-
quently regarded as fixed and every resemblance and
peculiarity being too often regarded as significant. The
explanations of supposed adaptations among animals fell
largely to the theory of natural selection which was
strained by some (see, for example, in Komanes, '92, p.
269) to explain origins in great detail, largely on the basis
of the competition of species for food, etc. Explanations
along this line were carried to a reductio ad ahsurdum as
indicated by Livingston ( '13) and have by no means dis-
appeared from the scientific calendar. This tendency
was less important on the plant side. More attention was
given to speculation concerning adaptive response.

From a consideration of the facts just presented, we
note that the characters of the two leading early view
points in evolution were no doubt influenced if not actually
caused to crystallize into their peculiar form by the failure
of workers to recognize the entire series of phenomena
which we have presented above. Thus a review of the
responses of sessile and motile organisms throws much
light on the influences leading to the first conceptions and
later modification of these two leading doctrines. Botan-
ists for many years dwelt mainly on the response of sessile
organisms and crystallized a Lamarckian conception of the
origin of adaptations through the fixing of advantageous
responses as hereditary characters. During the same
period zoologists essentially ignored sessile and other
multiple individualed .animals and their great plasticity
and crystallized the Darwinian idea into Weismannian
germplasm doctrine based on highly specialized single
individualed animals.

3. Supposed Non-Inheritance of Response and the
Germ Plasm Doctrine

The theory of the independence 9£ the germ-plasm from
the soma, and its continuity from generation to genera-
tion, was brought strongly to the attention of zoqjogists
in 1885 by Weismann. It was the natural outgrowth of
the methods and theories of the preceding period and

Digitized by

.Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 661

was largely based upon the non-inheritance of mutila-
tions and the fact that the germ cells of a few organisms
are, morphologically, early differentiated from the soma.
Turning to its influence upon ideas concerning response,
we note that from this viewpoint details of structure
were not of fundamental importance unless traceable to
the germ plasm. Still, structural details were more im-
portant than response, because, with the exception of
instincts, responses were believed to occur independently
of the germ plasm and hence were of interest only on
their own account. Thus the methods used in applying
Darwin's theory led to neglect of experimental study of
response and culminated in the extreme views of Weis-
mann. The germ-plasm theory or the ideas of heredity
which are associated with it has dominated zoological
thought almost if not quite down to the present day.^

4. The Influence of the Study of Response on
Present-Day Biological Theory

One of the most striking developments of recent years
has been the discovery that behavior responses are modi-
fiable to a high degree. Small traces of reagents reverse

2 Unconsciouslj suggestions of the supernatural which come up in connec-
tion with heredity and evolution have stimulated investigators to study and
speculation, though they have often approached the question of heredity wiih
an unscientific attitude. This is indicated by such statements as ''I could
not, however, resist the temptation to endeavor to penetrate the mystery ui
this most marvelous and complex chapter of life" and "the momentous
issues involved" and "no more fundamental problem could well be
stated" bear out this statement. The ardency which appears here and
elsewhere in the discussion of scientific questions, appears to the writer to
be associated with the discussion of problems which can not be referred
to existing facts for solution. Few of the present generation of scientific
men acquired a working knowledge of the methods of science before the
age of twenty-five years, and the early habits of mind were formed in the
atmosphere of the supernatural and dogmatic, which has characterized
hutnan thought for centuries. It is doubtful if the majority of us can
maintain a scientific attitude for more than a short period; we must con-
stantly come back to our tests and principles. This may account for many
of the contradictions regarding scientific principles which one finds in the
conversation of scientific men. When the methods of science have become
the methods of society we may expect a group of scientific men far more
eflfective than we ourFelves can hope to be.

Digitized by

Google

662 THE AMERICAN NATURALIST [VouXLVIlI

reactions. Intelligent behavior occurs in the lower
Arthropods. Even Paramoecium shortens the time re-
quired to turn around in a tube, by repetition. Actions
formerly regarded as instinctive now appear to be mere
innate tendencies perfected by repetition. Thus the ideas
of fixity have essentially disappeared from this field.

The response of organisms to injuries and the general
control of form in the lower groups has done much to
break down the ideas of fixity developed by Weismann
and embryological schools. Thus Child, the leading
American worker in this line, is able to control
size, form, number of eyes in the case of Planarians.
Various writers have found modifications inherited after
several generations of repeated stimulation (see Bateson,
'13). The development of anti-bodies (immunity) has
been shown to be a response occurring in connection with
many normal processes. The discovery of responses of
so many types has led to abandoning ideas of fixity even
among students of embryology and genetics. Thus we
note the recent decline of the doctrine of continuity and
independence of the germ plasm and kindred doctrines
and points of view, which constitute the central ideas of
fixity. It will accordingly be profitable to consider some
further facts which make the germ-plasm doctrine un-
necessary.

5. Aspects of the Untenability of the Gterm
Plasm Doctbine

The presence of primordial germ plasm is assumed
even in sessile colonial organisms such as plants, coelen-
terates, and in flatworms, etc., where under certain con-
ditions any small part of the body may give rise to a
complete organism. Here the theory is not needed to
explain the facts.

Child ('11) said:

The theory of the continuity of the germ plasm as a system, inde-
pendent of the soma, except as regards nutrition, has played an im-
portant part in biological thought during the last two decades, but I
am convinced that it has led in the wrong direction and that it is re-

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 663

sponsible for many pseudo-problems of heredity and development,
which on the basis of a different theory could never have occupied the
attention and wasted the energy of biologists. Briefly my position is,
that the gonad primordium is, at least up to a certain stage of develop-
ment, physiologically a part of the individuality as are other organs,
and that its further history of differentiation into male and female
gametes indicates that it becomes specified in a particular direction, at
least partly in consequence of its correlative environment in the or-
ganism.

The independence of the germ plasm is not well sup-
ported physiologically. Thus Wilson ( '12, p. 163) says
of the effect of prolonged ingestion of alkaline salts by
mice:

No obvious changes were evident in the liver, kidneys, lungs, spleen
and intestines but in the testes some extraordinary alterations were
found. These results are of especial interest because as the cells of the
testes except the basal cells are regarded by many cjrtologists as out of
coordination with the somatic cells. As a result of these experiments
it would seem that they are more susceptible to changes in reactivity
than the surrounding plasma.

Dungay ('13) and authors cited have thrown compara-
ble light on this question.

The facts of embryology themselves are but a pseudo
argument in its favor. The organisms in which continu-
ity is supposedly demonstrable are highly individuated
and their organs highly specialized and many different
organs are early separated from the common mass of
cells. The germ cells thus follow the general law of
development in such animals. The germ plasm is prob-
ably no more independent of other parts of the organism
than is the liver or any other special tissue. *'Germ
plasm'' and ''germinal continuity,'' if such exist, may
thus be merely incidental to the particular type of organi-
zation of the specialized individuals in which they occur.

It should further be noted that on the botanical side
this doctrine of the independence and continuity of the
germ plasm has received little attention and has been
given little credence because ''germ plasm" arises from
different tissues and is neither set aside early from the
soma nor is it in any other sense clearly continuous.

Digitized by

Google

664 THE AMEBIC AN NATURALIST [Vol. XLVU[

Furthermore, the plasticity of plant structures made the
application of the doctrine of natural selection to sup-
posed adaptations untenable, and this type of explanation
has received little more attention with botanists than have
Lamarckian speculations with zoologists. The adaptation
characters of plants can not ordinarily be used as taxo-
nomic criteria (Coulter, '08).

6. The Measure op Values in Biological Science
One hears reference to pure science as something quite
apart from applied science. It is indeed true that inves-
tigators in pure science are to some degree prompted to
push forward in research by interest in the problems for
their own sakes. But the human mind does not work long
isolated from practical affairs or the main channels of
human interest, and it is doubtful if the pure-science
investigator continues long in this way. Observations are
soon connected up in some way, actual or possible, with
some human interest, be it as remote as the improvmg of
human stock in remotely future generations. Thus "fure
science" defined as investigation for investigation's sake
hardly exists so far as the pure-science workers are
concerned, but may be best defined as an indirect method
of attacking problems of general importance. It differs
from applied science in that application to practical
problems is not its aim, though the estimated value of
theories and results in ''pure'' science are often greatly
modified by applicability to practical questions.

Certain problems and groups of facts in biology are
sometimes referred to as fundamental. Some one has
said that a fundamental problem is one the solution of
which biologists have decided will give greatest progress.
It is doubtless true that a few leaders reach such decisions
with regard to particular questions, but the real causes
of their general acceptance as fundamental are social
and imitative. Thus when one investigator or a small
group of investigators arrives at such a decision many
others usually become active along the same lines largely
because it is a popular topic. Thus under the influence

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 666

of a group of investigators among whom Weismann was
a conspicuous leader, problems of the germ cells, the
egg^s early development, and heredity, became *'funda-
nouental problems.'' They evidently argued that since all
comes from the egg and germ cell, all must be discover-
able in the egg. If germ plasm were as independent from
soma, as completely insulated from environment as con-
tinuous from generation to generation as has been
assumed, the study of germ plasm would be the only way
to the solution of the problems of heredity and evolution.
This follows no matter whether the chromosomes or
almost the entire egg are credited with carrying heredi-
tary qualities; only the postulation of continuity and
independence from soma and insulation from environ-
ment are necessary. If the independence of germ plasm
from soma be accepted even in a weakened and modified
form it follows that studies of somatic characters can at
most be of secondary importance from the point of view
of heredity and evolution. Thus in some quarters the
value of various lines of zoological work has been esti-
mated largely, unconsciously, no doubt, in proportion to
the nearness or remoteness of their relation to the ''germ
plasm' ^ question.

Thus it is true that in biology as in all other fields
values are measured consciously or unconsciously by
criteria. In recent years another better criterion of value
has made its appearance among zoologists. The germ
plasm criterion already discussed was primarily morpho-
logical; the second is physiological, borrowed no doubt
from physiologists. It measures values on the basis of
the analysis of the organisms into terms of physics and
chemistry or is concerned with a mechanistic conception
of life in all its manifestations. From this viewpoint the
study of each and every part of the organism is important
because the discovery of laws governing one part is
usually or at least often of general importance. Investi-
gations from this viewpoint have shown that the germ
plasm criterion is clearly illogical in its application to the
study of somatic characters because it is based upon the

Digitized by

Google

666 THE AMERICAN NATURALIST [Vol. XLVIII

tacit assumption that the soma is governed by different
laws from the living matter which makes up the germ
plasm from which it arose. In other words it is assumed
that the germ plasm is so different from the soma that
the discovery of laws governing the soma is a type of
investigation o£ relatively little significance.

Some criterion of values is of course necessary in sci-
ence as well as elsewhere, and for the sake of argument we
would be willing to accept the second when broadly stated
and the first when broadened and modified so as to accord
with the second as appears to be the case among certain
students of genetics. In other words, problems of the
germ cells, the egg, and heredity, are of much importance
when the germ cells themselves are regarded as dynamic
and in their relations to the dynamics of the organism
as a whole.

Granting that these are true and tenable criteria of
values in present-day biological science, what is to be the
method of application? Should biology demand that
results be of direct application to these '* central'' prob-
lems? One has but to look at the history of almost any
branch of science to find that great, if not the greatest,
advances have come through following up results at
points where relations to the central problems of the
period were quite unsuspected, or by the transference
of methods, principles and results from one field to an-
other where relations between the two were not suspected.
Take, for example, immunity and immunization, the his-
tory of which is ably sketched by Adami ( '08, pp. 451-
528). It has been known for ages that one attack of many
infectious diseases yields more or less complete immunity
from subsequent attacks. Thus for centuries in India
and the East individuals, chiefly children, have been pur-
posely inoculated with matter or by contact. The prac-
tise grew out of experience showing that diseases thus
communicated to healthy individuals from weaker ones
are less severe. In 1796 the results of Jenner on vaccina-
tion with cowpox were published. This may have influ-

Digitized by

Google

No. 575] RESPONSES OF PLANTS AND ANIMALS 667

enced Pasteur, who over eighty years later laid the
foundation for the modern epoch of development, by
combating a plague of diarrhoea in poultry (1880).
During the twenty years following, various investigators
added noteworthy contributions, and about 1900 Ehrlich
and Morgenroth evolved the ** side-chain theory" by
which a large number of possible conditions can be pre-
dicted and all the observed facts of immunity explained.
While not expressed in strictly chemical terms, the theory
and the experiments which support it are very important
both practically and theoretically. In recent years the
knowledge of immunity and comparable phenomena have
been greatly extended. Various workers (Pf offer, Vol.
II, p. 262) have shown similar phenomena in the increased
resistance of plants to poisons, thus making the responses
of plants and animals still more generally comparable.
Most recently workers on problems such as fertilization
(LiUie, '13), standing in close relation to the older germ-
plasm doctrine, have discovered facts belonging to this
field and made use of Ehrlich 's theory to explain the ob-
servations. This development has helped to confirm the
conclusion of some investigators that immunity phe*
nomena represent important features of the chemical
mechanism of life. Adami has remarked.

That a plague of diarrhoea in a poultry yard, studied by a professor
of chemistry, should be the seed from which has grown the vast de-
velopment of later years is a strange fact, but a fact nevertheless.

What was the attitude of pure science so called, of
germ-plasm doctrinairies, and biologists generally during
the long period which elapsed before they could make
use of his results? Clearly it was one of indifference, if
not disgust, toward the subject. The probable result of
such attitudes on the progress of the investigation of
immunity phenomena, had it not been for their immense
practical significance, is clear. They could not have
received their proper share of attention. Thus in the
pursuit of the analysis of the chemical mechanism of life
men who sought it directly have failed in this one imp^r-

Digitized by

Google

668 THE AMERICAN NATURALIST [Vol. XLVIII

tant step, and the chief contribution has come from verj''
remote indirect methods. Generally speaking the inves-
tigators who choose a direct method of attack often put
themselves somewhat in the position of the chemist who
would make chemical analysis of living matter when his
first step defeats its own purpose by killing the substance
to be analyzed. The failure of exclusively direct methods
is often evident. Still the ability to obtain results by the
method of direct attack, combined with a far too rare
ability to tie with them indirectly obtained data, some-
times gives noteworthy contributions.

It accordingly remains to be seriously considered
whether or not biology can afford to apply criteria to the
measure of the values of investigation. Their application
is of course largely unconscious, but the effects are not
thereby modified. Noteworthy results of their applica-
tion are (a) concentration of work in certain lines indi-
cated by a given criterion, and (6) an actual abandoning
to a large degree of remote and indirect methods of
attacking the problems which the criterion involves. This
means the partial abandoning of the methods for which
pure science stands.

Criteria can be safely used only in a very broad gen-
eral way, and in application more often to past progress
than to current investigation. They are perhaps most
valuable as a guide to individual investigators working
on problems remote from these more or less central
**pure science" questions. That some guide should be
in the hands of such workers is beyond question. In the
hands of those attacking the problems directly they often
appear detrimental because they soon take on an extreme
form and become regarded as fundamental. At this
stage they are usually in need of extensive revision. If
the investigator is contributing observations and details
only, he is doing a great service, for such information is
needed everywhere. If he is able to combine his own and
others results, he almost invariably draws data from all
sc^rces, direct and indirect, far and near. Granted the

Digitized by

Googl^

No. 575] RESPONSES OF PLANTS AND ANIMALS 669

ability to synthesize, the opportunity to use the ability
sometimes comes to those who attack the so-called cen-
ti^l problems directly. It comes equally often (we believe
more often) to those who have led up to the central prob-
lem from some remote viewpoint, frequently condemned
by the followers of direct method of attack. Granting
the importance of synthesis, if the biologist seeks the
solution of such a problem as the germ-plasm problem,
he should encourage workers to start at points as remote
from the subject as possible, that they may approach it
with new light and from new angles.

In judging the work of another, its value should be
determined more by the (a) strictness of scientific method
used, (6) the thoroughness and completeness of the in-
vestigation, and (c) (and perhaps most important of all)
evidence of ability to synthesize and combine other re-
sults with his own with a view to broader generalization.
It must, however, also be recognized that there are many
biological problems of much human importance, which
must be solved quite independently of the ideal central
problems of pure science.

6. Summary and Conclusions
From the data presented above, we note that the doc-
trine of purposeful, advantageous response (including
anthropomorphic ideas) arose from the uncritical non-
experimental study of the responses (structural) of ses-
sile and (behavior) motile animals. The idea of the all-
sufficiency of natural selection is largely the outcome of
observational study of apparently fixed and yet appar-
ently adaptive characters of motile highly individuated
animals. The doctrine of the continuity of the germ
plasm is likewise the outgrowth of the study of highly
individuated animals in which the various organs are
early differentiated in the dividing egg. No one of the
doctrines is wholly tenable ; no one is more than a partial
truth. Each appears to have arisen from a recognition
of certain more or less unconsciously selected and un-
critically interpreted phenomena by each of several men

Digitized by

Google

670 THE AMERICAN NATURALIST [Vol. XL^TLII

who secured different facte and attempted explanations.

In a few animals the ''germ plasm'' may be morpho-
logically early differentiated and reasonably continuous,
though governed by the same laws as other tissues. In
others, any part of the general tissues may give rise to a
complete organism. The behavior of some organisms is
intelligent and purposeful, while that of others is largely
mechanical. Some structural responses of sessile organ-
isms are advantageous, some indifferent and some harm-
ful. Some of the more fixed structures of the highly indi-
viduated animals are advantageous, some indifferent,
and some disadvantageous (Metcalf, '13). No other type
of general statement appears to be tenable, yet each
extreme of each proposition has at some time or other
been the subject of some all-inclusive doctrine.

Such are the limitations of an individual's knowledge
and the psychic limitations of our race and generation.
In considering the psychology of religion, Ames ('10,
p. 394) points Qut similar well-recognizable tendencies in
that field of human activity and quotes Cooley on social
development as follows :

Much energy has been wasted or nearly wasted, in the exclusive and
intolerant advocacy of special schemes — single tax, prohibition, state
socialism and the like, each of which was imagined by its adherents to
be the key of millennial conditions. Every year makes converts to the
truth that no isolated scheme can be a good scheme, and that real prog-
ress must be advanced all along the line.

Advance all along the line is tvhat biological science
must achiev^e. This I believe means the encouraging of
all lines of indirect attack, whether they at first throw
light on the ideal central question of pure science or
important practical problems or not. .It means the exer-
cising of extreme caution in the application of criteria of
values to scientific results. Such measures tend not only
to stifle the best initiative in good investigators, but also
tend to check the building up of fruitful hypotheses.
The latter danger is greatest in connection with the
mechanistic criterion referred to above. As has already
been stated, criteria of values can be safely applied only

Digitized by

Google

No. 575] BESPONSES OF PLANTS AND ANIMALS 671

as broad general guides, and investigation should be
measured on the basis of its thoroughness, the originality
shown, etc.

In science special schemes of course do not exist recog-
nized as such, but intolerant application of criteria of
values results in essentially the same condition. One
often hears the statement made by so-called scientific
men, that this or that line of investigation has been pur-
sued for several years, but has failed to yield important
advances or generalizations, but they add, we will be very
glad to recognize it as soon as its value is proven. This
seems to us to be a distinctly unscientific attitude, and
but a polite modern statement of a spirit which in former
generations often sent men to the stake or dungeon. This
is true because to these i>olite objectors its value is
rarely or never proven. It*is '* schemes'' (preconceived
theories) thus presented that have in the recent past
stifled the study of responses by discouraging efforts in
Nhat direction and thus contributed materially toward
making zoology the unorganized science which it is
to-day. We must recognize that the various aspects of
zoology pure and applied have never been well corre-
lated, less so we believe than in any other branch of
natural science, clearly less than in botany. In general,
animal physiology has been isolated in medical schools
and genetics, faunistics and morphology have not been
properly influenced by it, while morphologists for many
years held themselves aloof from other workers.

In a discussion dealing mainly with the doctrine of
natural selection in the origination of adaptations,
Mathews ('13) has sounded the keynote of a growing
attitude toward all response questions. Out of the infi-
nite different combinations which may enter into the
proteid molecule and the varying rates at which metabolic
action may go forward, innumerable types of irritability
and correlated structure have been and still are arising
under the influence of environment external and internal.
Of these some are incompatible with life, others indiffer-

Digitized by

Google

672 THE AMERICAN NATURALIST [VouXLVHI

ent, and others advantageous. Upon these physiological
characters natural selection has operated to eliminate,
and with time has perhaps rendered of less frequent
occurrence, those characters that are incompatible with
their conditions of existence. External form, color orna-
mentation, etc., while no doubt often of importance them-
selves are more often the advantageous or indifferent
correlatives of physiological or irritability types which
are compatible with their conditions of existence. The
study of irritability and response may be pursued in
many ways — ^by experiment, by observation in nature
alone or combined with experiment. The mapping of
stimulating conditions in nature, of the distribution of
types of irritability and response, which is one function
of field ecology and modern geography, can hardly fail
to contribute materially to tbe advance of knowledge in
many lines, including that of the physico-chemical
mechanism of life. The student of experimental ecology
has an infinite field of problems and methods thrown #
open to him by the organization of such information
relative to responses. Still in our attempt to make ad-
vances along the line of the study of responses, we must
not forget that it is but one of several lines of advance,
all of which must sooner or later be correlated with a
view to broader generalization.

Hull Zoological Laboeatory,
Univebsity or Chicago,
April 1, 1914

LITERATURE CITED

Adami, J. G. ('08). Principles of Pathology; Vol. I, General Pathology.
Philadelphia.

Adams, Chas. C. ('13). Guide to the Study of Animal Ecology. Chaps,
VIII, IX. New York.

Ames, E. S. ('10). The Psychology of Religious Experience. New Yort

Bateson, W. ('13). Problems of Genetics. New Haven.

Brooks, W. K. ('93). Salpa in its Relation to the Evolution of Life.
Johns Hopkins Univ. Studies. Biol. Lab., Vol. 5, pp. 129-211.

Child, C. M. ( '04). Studies on Regulation, 11. Jour. Expt. Zool, "Vol. II,
pp. 95-133.
('11). A Study of Senescence and Rejuvenescence Based on Experi-
ments with Planaria dorotocephala. Roux. Arch., XXXI,
pp. 537-616.

Digitized by

Google

No. 576] RESPONSES OF PLANTS AND ANIMALS 673

('13). Certain Bjnamic Factors in Experimental Beproduction.

Boux. Arch., Bd. XXXV, pp. 598-641.
Coulter, J. M. ('09). The Theory of Natural Selection from the Stand-
point of Botany. Fifty Years of Darwinism, pp. 56-75.
Cowles, H. C. ('11). A Text-book of Botany. Vol. II, Ecology. New

York.
Cope, E. D. ('96). Primary Factors of Organic Evolution. Chicago.
Davenport, C. B. ('91). Cristatella: The Origin and Development of the

Individual in the Colony. Bull. Mus, Camp, Zool,, Vol. 20,
pp. 101-152.
Delage, Y., and Herouard, E. ('96). Les Vermudiens, Zoologie Concrete

I. V. Paris.
Dimgay, N. S. ('13). A Study of the Effect of Injury on the Fertilizing

Powers of Sperm. Biol, Bull., Vol. XXV, pp. 213-260.
Forbes, S. A. ('80). Some Interactions of Organisms. Bull. Ill, 8t, Lab.

Nat. Hist,, III, pp. 3-18.
Goode, J. P. ('04). Human Besponses to the Physical Environment.

Jour, of Geography, III, pp. 333-343.
Goebel, K. ('00). Organography of Plants. Part I, English ed., Oxford.
Harmer, S. F. ('01). Polyzoa. Cambridge Natural History, Vol. II,

pp. 465-534.
Herdman, W. A. ('04). Ascidians and Amphioxus, 1. c, Vol. VII, pp.

35-138.
Herrick, F. H. ( '05). The Home Life of Wild Birds. New York.
Hickson, S. J. ('98). On the Species of the Genus Millepora. Proc. Zool.

8oc. London, 1898, pp. 241-257.
Holmes, S. J. ('07). The Behavior of Loxophyllum and its Belation to

Begeneration. Jour, Expt. Zool, Vol. IV, pp. 399-418.
(*11). Evolution of Animal Intelligence. New York.
Henderson, L. J. ('13). The Fitness of the Environment. Am. Nat.,

XIVII, pp. 105-115.
Jennings, H. S. ('06). Behavior of the Lower Organisms. New York.
JoKison, Myrtle, E. ('10). A Qualitative Study of the Development of

the Salpa Chain in Sdlpa fusiformxs runcinatcfi. Univ. of Cal.

Pubs, in Zool., Vol. 8, No. 7, pp. 145-176.
Johnstone, J. ('08). Conditions of Life in the Sea. Cambridge.
Klebs, G. ( '03). Willktirliche Entwicklungsfinderungen bei Pflanzen. Jena.
Lillie, Frank B. ('13). The Mechanism of Fertilization. Science, N. S.,

Vol. XXXVIII, pp. 524-528.
Livingstone, B. E. ('13). Adaptation in the Living and Non-living. Am.

Nat., XLVII, pp. 72-82.
Loeb, J. ('06). Dynamics of Living Matter. New York.
MacDougal, D. T. ( '11). Organic Besponse. Am. Nat., Vol. XLV, pp. 5-40.
Mast, 8. O. ('11). Light and the Behavior of Organisms. New York.
Mathews, A. P. ('13). Adaptation from the Point of View of the Physi-
ologist. Am. Nat., XLVII, pp. 90-105.
McGee, W J ('96). The Belations of Institutions to Environment. Smith.

Bep., 1895, pp. 701-711.
Merriam, C. H. ('90). Besults of a Biological Survey of the San Fran-

Digitized by

Google

674 THE AMERICAN NATURALIST [Vol. XLVni

Cisco Mountain Region and the Desert of the Little Colorado.
North American Fauna 3.
Metcalf, M. M. ('13). Adaptation through Natural Selection and Ortho-
genesis. Am. Nat., Vol. XIiVII, pp. 65-72.
Minchin, E. A. ('00). Sponge. Lankester's Treatise on Zoology, Part H.

London.
Moore, H. F. ( '08). The Commercial Sponges and Sponge Fisheries. B«R.

U, S. F, C, 1908, pp. 403-511.
Motz-Kossowska, S. ('08). Quelques considerations ft propoe de Plumularia

lichten^temi Mark Turn et les esp^ces affinee. Arch, Zool, Exp,
et Gen. Notes et Bevue, 4, T. IX, pp. 55-59.
Packard, A. ('01). Lamarck the Founder of Evolution. New York.
Ptttter, A, ('11). Vergleichende Physiologic. Jena.
Beighard, Jacob ('10). Methods of Studying the Habits of Fishes with

an Account of the Breeding Habits of the Horned Dace.

Bull Bur. of Fish,, XXVIH, 1908, pp. 1111-1136.
Bitter, W. E. ('11). The Controversy between Materialism and Vitalisnti.

Can it be ended? Science, N. S., Vol. XXXIII, No. 847, pp.

437-441.
Romanes, G. J. ('92- '97). Darwin and After Darwin. 3 vols. London.
Shelf ord, V. E. ('11a). Ecological Succession, I. Stream Fishes and the

Method of Physiographic Analysis. Bioi, BM,, Vol. XXI,

pp. 9-35.
('lift). Ecological Succession, II. Pond Fishes. Biol. BuU., Vol.

XXI, pp. 127-151.

('lie). Physiographic Animal Geography. Jour, of Morph. (Whit-
man Vol.), Vol. XXII, pp. 551-618.

('lid). Ecological Succession, III. A Beconnaissance of its Causes in
Ponds with Particular Beference to Fish. Biol. Bull., Vol.

XXII, pp. 1-38.

('12a). Ecological Succession, IV. Vegetation and the Control of

Land Communities. Biol Bull, Vol. XXIII, pp. 59-99.
('126). Ecological Succession, V. Aspects of Physiological Classifi-
cation. Biol. Bull, Vol. XXIII, pp. 331-370.
Strassberger,* E., Schenk, Noll F., and Schimper, A. W. F. ('03). A Text-
book of Botany. Translation by Porter and Lang. London.
Stockberger, W. W. ('13). A Literary Note on the Law of Germinal Con-
tinuity. Am. Nat., XLVII, pp. 123-128.
Washburn, M. F. ('09). The Animal Mind. Chap. I. New York.
Weismann, A. ('91). '92, 2v. Essays upon Heredity and Kindred Biolog-
ical Problems. Authorized Translation. E. B. Poulton & A. E.
Shipley. Oxford.
( '93). The Germ Plasm a Theory of Heredity. Translation by N. W.
Parker & H. IWinnfeldt.
Wilson, F. P. ('12). Cell Stimulation by Means of Prolonged Ingestion of

Alkaline Salts. Biochemical Journal, Vol. IV, p. 162.
Wood, Jones F. ('07). On Growth Forms and Supposed Species of Corals.
Prod. Zool Soc. London, 1907, pp. 518-556.
('11). Coral and Atolls. London.

Digitized by

Google

AN APTEROUS DBOSOPHILA AND ITS
GENETIC BEHAVIOR

CHARLES W. METZ
Department op Zoology, Columbia University

Among the various mutants of the fruit-fly, Drosophila
ampelophila, which have arisen from cultures in this
laboratory, is one entirely destitute of wings, and hence
called apterous.^ The study of the heredity of this form
has been difl5cult because of its almost complete (appar-
ent) sterility. In order, therefore, to determine to which
of the three groups of linked characters of Drosophila it
belonged I was obliged, in most crosses, to make use of
heterozygous flies that carried the factor for apterous.
As this process is unique in certain regards, it will be
described in some detail.

Methods of Study

At first it was thought that the apterous mutant was
completely sterile, since none of the first flies, as they
appeared occasionally in certain cultures, could be crossed
even with normal individuals. At last, however, offspring
were obtained from an apterous female by a wild male,
and a permanent line started. But this line could not be
perpetuated by means of apterous individuals, for these
were unable to breed.^ It had, therefore, to be kept up
by means of heterozygous, winged flies. The method was
as follows : The original cross of winged by apterous gave
in F2 approximately 3 winged to 1 apterous. Of the
winged class approximately two thirds were heterozygous
for apterous, and when mated together gave the same

1 This apterous fly is quite distinct from that called wingless in earlier
papers by Morgan, and now known as vestigial.

2 Only twice, aside from the original mating, were apterous individuals
successfully crossed, and then only to winged specimens, never to their own
kind. These two cases axe given in experiments II and III.

675

Digitized by

Google

676 THE AMERICAN NATURALIST [Vol. XLVIII

3 : 1 ratio. Selecting again from the winged flies, the proc-
ess could be repeated indefinitely. The only diflBculty lay
in the fact that no visible character differentiated the
heterozygous from homozygous winged flies, and conse-
quently all matings had to be made in pairs taken at
random, with the result that about 56 per cent, of the
cultures were rendered worthless. In actual practise large
numbers were mated in pairs, and then all discarded save
those producing apterous.^ This was the method used in
keeping up stock.

To obtain the necessary combination of apterous with
other mutant factors, winged offspring from apterous-
throwing parents were mated in pairs to flies of the
desired stock. One third of the normals from apterous
stock were pure for the normal allelomorph of apterous
and rendered worthless all matings in which they were in-
volved; but the other two thirds were heterozygous for
apterous, and when crossed with the desired stock gave
in Fa some apterous offspring. If the Fi flies were bred
671 masse, approximately 15 winged to 1 apterous were ob-
tained, but if bred in pairs, certain pairs (those in which
both members were heterozygous for apterous) gave 3
winged to 1 apterous. The latter method was the one
actually used in most cases. In this manner the same end
result was attained as would have been secured by using
apterous individuals in crosses with other stocks, the
only difference being in the amount of labor involved in
making up a larger number of cultures. Both kinds of
crosses were, in fact, used, as will be seen below.

The use of symbols in this paper follows the system
recently adopted by Morgan and other students of Droso-
phila (Morgan, 1913, a and 6). That is, for any pair of
allelomorphic characters a capital letter is used to indicate
the dominant, and a small letter the recessive factor — ^the
symbol being taken from the name of the mutant Since
the apterous character is recessive, the symbols for the

8 In the fourth experiment a character (black) was introduced which dif-
ferentiated homozygous from heterozygous and thus made it possible to pick
out the heterozygous individuals.

Digitized by

Google

No. 575] AN APTEROUS DROSOPHILA 677

apterous fly become Op-Op, and those for the winged fly
Ap-Ap. In other words, Ap is a factor in the wild fly
necessary for wing production, while Op is its modified
homologue responsible for lack of wings in the mutant.
The apparent contradiction in using Ap, not for the factor
responsible for apterous, but for its normal allelomorph,
may be confusing at first sight, but a little familiarity with
the system obviates this difficulty.

Experiments

Experiment /.—Long-winged, red-eyed c? by apterous,
white-eyed 2 (from miniature wing stock).

Fj All winged. Long-winged, red-eyed females.

Miniature winged, white-eyed males.

Fa Winged and apterous as follows :

Long-winged, red-eyed males and females.
Long- winged, white-eyed males and females.
Miniature-winged, red-eyed males and

females.
Miniature-winged, white-eyed males and

females.

Winged

. [Apterous, red-eyed males and females.

^ [Apterous, white-eyed males and females.

This experiment shows the inheritance of the apterous
character to be Mendelian, giving in Fj all winged, and in
Fa approximately 3 winged to 1 apterous. Table I con-
tains a summary of the offspring from 21 pairs of the F^
and Fg individuals, giving a total of 1,405 winged to 450
apterous,— a ratio of 3.12 to 1.

The absence of apterous flies in Fi indicates at once
that the apterous character is not sex-linked. The pres-
ence of miniature-winged flies in Fj and Fg indicates that
the apterous factor is independent of the miniature-wing
factor, which latter must have been carried by the apterous
female (coming from miniature wing stock), and trans-
mitted to her offspring unaffected by the apterous factor.

Digitized by

Google

678 THE AMEBIC AN NATURALIST [Vol. XLVin

Analysis of the cross:
Apy factor necessary for wing production, (h, its alle-
lomorph, in the apterous fly.
M, factor necessary for the production of long wings

(sex-linked).
m, allelomorph of If responsible for miniature wings.
W, factor necessary for tiie production of eye color

(sex-linked).
w, allelomorph of W responsible for white eyes.

Pi Long, red male ApMWX-Apf

Apterous, white female a^mwX-cipmwX.

Fj Long red females ApMWX-aprmoX, •

Miniature white males ApHipfif^wX.

Fg, leaving out of account the sex-linked factors and
considering only winged vs. apterous :

Gametes of Fi Ap, ch.

Apf Qfp,

Winged

Ap-Op-
Ap—Ap,

Apr-Op.

Apterous Op-Op.

Experiment //. — Long vermilion $ by apterous, white ^.^
. This cross is practically the reciprocal of Exp. I, except
that vermilion replaces red eye color in the winged parent
Like Experiment I it involves two pairs of sex-linked char-
acters, aside from the apterous character. The results are
essentially like those of Experiment I and may be passed
over briefly.

Pi Long, vermilion 2 ApMWX-ApMWX,
Apterous, white c? OpinwX-dp'^

* The white-eyed, apterous c? in this cross is white-vermilion, t. e., the
double recessive, and therefore when crossed with vermilion it gives vermilion
instead of red in F,.

Digitized by

Google

No. 575]

AN APTEROUS DROSOPHILA

679

Long, vermilion $
Long, vermilion c?

AjMWX-opmwX,
AvMWX-^.

Long vermilion $ and c?
Long, white c?
Miniature, vermilion ^
Miniature, white c?

Apterous, vermilion $ and c?
Apterous, white c?

TABLE I

. Winged.

Apterous.

Offspring from Pairs Heterozygous for Apterous in Experiment I.
Parents taken from F„ F, and F,

Mating No.

Winged

Apterous

Ratio

122

3.8 : 1

2.9 :

2.5 .

1.5

1.8

3.04

3.5

4.3

4.4

111

183

2.8

112

117

2.4

118.

2.8

130

4.5

131

3.1

132

3.7

134

36-

2.6

151

4.8

170

7.8

171

4.5

177

129

3.6

1.405

450

Average ratio of winged to apterous, 3.12:1.

An analysis of the Fg is not essential here aad is omitted
for the sake of brevity. It may be derived from the Fj
formulae. Table 11 indicates the expected classes and
ratios in Fg and gives the actual numbers obtained in cul-
ture No. 59, in which each class was recorded separately.
In subsequent cultures of this experiment no attempt was
made to separate any but the winged and apterous classes.
Counts of the latter are given in Table III.

Digitized by

Google

680

THE AMERICAN NATURALIST [VouXLYin

TABLE n
Mating 59

id Classes 1 Expected lUtio

Actaftl Ratio

Actatl No.

12
3
3
3
3

4
2
2

13.8 : 1
4.2:1
2.4:1
2 :1
4 :1

3.2:1
1.2: 1
2 :1

185

Long verm. 9

Long verm, cf

Long white cf

Miniature verm, cf .
Miniature white cf .

Apterous verm. 9 . .
Apterous verm, cf . .
Apterous white cf . .

TABLE in
F, FBOM Mass Cultubss

Culture No.

Winged

Apterous

RaUo

367

4.7 : 1

352

4.2

556

114

4.8

306

4.3

120

597

103

5.08

135

554

104

5.3

137

405

154

298

6.6

157

405

7.6

3.840

699

Average ratio, 5.6

Fa FROM PaIES

Culture No. |

WiDged

Apterous

Ratio

155 1

156 1

141
38

35
11

4.1
3.4

: 1
: 1

179

Average ratio. 3.9

: 1.

F, FROM

Pairs

Culture No.

Winged

Apterous

RaUo

160

119

4.7 : 1

162

123

4.6

172

266

4.3

173

3.4

174

165

4.5 •

175

167

176

171

3.7

177

129

3.6

178

1.319

323

Average ratio, 4.08: 1

179 46

1,319 323

Total from pairs, 1,498 369

Ratio, 4.6 : 1.

Digitized by

Google

No. 575] AN APTEROUS DROSOPHILA 681

It will be noted that the apterous classes fall a little
below the expected numbers in most cases. This is char-
acteristic of all weak races of Drosophila, and is doubtless
due to the inability of some flies to mature. Of the winged
classes the first two and the fifth exceed the Mendelian
ratios, while the other two fall short, due to linkage be-
tween white, vermilion and miniature.^ The distribution
of apterous, however, is entirely independent of the
others, showing that the factor responsible for it is not a
member of the group (Containing those responsible for
vermilion eyes, white eyes, or miniature wings.

The ratio of winged to apterous in this particular cul-
ture is 4.2 : 1. Table III includes a summarj^ of this and
nineteen similar cultures in which the parents were all
descendants of the long-winged, vermilion-eyed female by
the apterous male mentioned above. The first nine are
mass cultures, the next two are pairs, from Fj flies. Below
these are offspring from nine pairs of Fg flies.

It is noticeable that the ratio of apterous to winged is
greater in cultures where pairs are used than in mass cul-
tures, though all parents in the latter are heterozygous.
This, I believe, is unquestionably due to the low viability
of the apterous flies, which prevents some of them from
maturing in cultures where the competition is severe. For
this reason the averages are given separately for pairs
and for mass cultures. The average from pairs is 4.06 : 1,
while that from mass cultures is 5.5 : 1. This low viability
is also shown by pairs, if the food conditions are not good,
or if the culture becomes very dry.

Experiment III.— To determine the relation between
apterous and characters in Group III.

It is obvious from Experiments I and II that apterous
is not a sex-linked character (Group I). The present ex-
periment is for the purpose of determining its relation to
characters of Group III. As a representative of the latter
group pink eye color was chosen. The results of the cross
between this and apterous may be passed over briefly

5 For d'seussion of linkage between these characters Fee Morgan, 1911.

Digitized by

Google

682

THE AMEBIC AN NATURALIST [Vol. XLVHI

since they are similar to previous results in showing no
linkage. A winged, pink-eyed male bred to an apterous,
vermilion-eyed female (from Experiment HI) gave, as
expected, winged, vermilion-eyed males and winged, red-
eyed females in Fi. These inbred gave four classes of
winged and four classes of apterous, i. e., red, vermilion,
pink and orange.® The ratios are such as to show inde-
pendent segregation of apterous and pink. Below is a
summary of the expected and actual results.

Pj Apterous, vermilion $
Winged, pink c?

Fi All winged.

Red $
Verm, c?

OpvPX-OpvPX,
ApVpX-App.

chvPX-ApVpX,
ChvPX-App.

Eight kinds of eggs and four kinds of spermatozoa are
formed by these Fi flies, giving, through random fertiliza-
tion, 32 classes of offspring divided into eight groups, as
shown in Table IV.

TABLE IV

Ft Expected Results

ClasMt

Ratio

Actaal Results in

Experiments

627

6287 1 629

630 { 6S1

^h^

120

fo35

?3>

I}-

Total

Winced, red

Winged, vermilion . .

Winged, pink

Winged, orange. . . .

Apterous, red

Apterous, vermilion.

Apterous, pink

Apterous, orange. . .

402

111

114

In the table red and vermilion have been considered
together as one class, because they both contain P; and
similarly pink and orange have been considered together
because they both contain p. The total numbers for the
four classes give the ratios 402:111:114:34, or 10.8:3.6:
3.35 : 1,— a sufl5ciently close approximation to the expected

« Orange eye is the double recessive pv-pv,

7 In this culture the ratios are seen to diverge widely from the expected,
due, I believe, to the poor cultural conditions in this case which prevented
some of the weaker pink and orange flies from maturing

Digitized by

Google

No. 575]

AN APTEROUS DROSOPHILA

683

9:3:3:1. These results clearly show the independence
(t. e., lack of linkage) of apterous and pink.

Table V includes all matings (giving apterous) in Ex-
periment m, for the purpose of showing the ratio of
winged to apterous.

TABLE V

Mating No.

Winged

Apterous

Ratio

627

178

628

629

168

630

631 •

646

649

650

661

124

662

674

692

1,027

3.66 : 1

2.47 : 1

3.7 : 1

3.53 : 1

3.36 : 1

4.8 : 1

3.1 : 1

3.16 : 1

3.36 : 1

2.7 :1

2.63: 1

27
(16

3.16 . 1

Average ratio winged to apterous^ 3.25:1.

Experiment IV.— To determine the relation between
apterous and characters in Group II.

The mutant called *' black" (having black body and
wings) was used in this experiment as a representative of
Group n. No direct matings with apterous individuals,
such as obtained in the three preceding cases, could be
effected here, and consequently the winged brothers and
sisters of apterous had to be used for crossing with black,
according to the method described in the introduction.
Matings of this kind (in pairs) gave, in Fj, winged flies,
some of which were heterozygous for apterous and black.
These inbred (also in pairs) gave winged and apterous,
and gray and black, as shown below :

Pi Black (? homozygous for wings Apb-Apb,
Gray ? heterozygous for apterous ApB-OpB.

Fi Winged, heterozygous for black Apb-ApB,

Winged, heterozygous for black and apterous Arb-apB.

Only pairs in which both members were of the second
type (^pfe-^pB,— heterozygous for apterous) could pro-

Digitized by

Google

684 THE AMERICAN NATURALIST [Vol. XLVIU

duce apterous. The others, therefore, are ignored. Con-
sidering the second type alone, the analysis becomes :

Fj gametes (expected) Apb-ApB-Oph-OpB,

Aph-ApB-dpb-OvB.

Fg Expected classes.

OpB-OpB apterous, gray.

OpB-Opb apterous (heterozygous for black).

OpB-ApB winged, gray.

dpB-Aph winged, heterozygous for black.

Opb-OpB apterous, heterozygous for black.

Opb-Opb apterous, black.

Opb-ApB winged, heterozygous for black.

Opb-Apb winged, black.

ApB-OpB winged, gray.

ApB-Opb winged, heterozygous for black.

ApB-ApB winged, gray.

ApB-Apb winged, heterozygous for black.

Apb-OpB do.

Apb-<ipb • winged, black.

Apb-ApB winged, heterozygous for black.

Apb-Apb winged, black.

Expected ratios : 3 winged black ; 6 winged heterozygous
for black; 3 winged gray; 1 apterous black; 2 apterous
heterozygous for black ; 1 apterous gray.

Actual results : winged black, winged heterozygous for
black, and apterous gray, as shown in Table VI.

In the last two matings black and heterozygous oflF-
spring were counted as one class.

Total: winged 701; apterous 174 or 4.02:1.

The expectation for the F2 if Ap and B segregate inde-
pendently is equal numbers of black and gray among the
winged and among the apterous offspring. Actually, how-
ever, the apterous flies are all gray, and the winged flies
are all black or heterozygous for black. Furthermore, the

Digitized by

Google

No. 575]

AN APTEROUS DROSOPHILA

685

ratio of heterozygotes to pure blacks in the winged class
shows that the flies which should have been gray accord-
ing to expectation have been added to the heterozygotes.
Likewise the gray flies in the apterous class are about four
times as numerous as anticipated, showing that the ex-
pected heterozygotes and blacks are here gray. From this
it is evident that the factors Op and B, on the one hand,
and Ap and b, on the other, have remained associated in
the combination which they formed in the parents, instead
of independently segregating. Such an explanation ac-
counts for the absence of ApB and apb gametes in the Pi
generation, and consequently for the absence of gray,
winged flies, and of black or heterozygous apterous flies
in Fg. The evidence accords with that obtained for many
other mutant characters in Drosophila, and the explana-
tion is the same as that given for the previous cases {e. g.,
Morgan, 1911, 1912 ; Morgan and Lynch ; Sturtevant, 1913
a and b; Dexter).

TABLE VI
Actual Besults

Mating

Winged

Apteroui

No.*^

Gray

Black

Heterozygous

Gray

Black

Heterozygous

703

724

732

733

741

745

746

166

414

136

726
734

63
79

14
24

The presence of a definite linkage or association be-
tween apterous and black (t. e., between either Op or Ap
and b or B depending upon the nature of the cross) as
shown by this experiment, together with the absence of
any such linkage with characters in Groups I and HE, as
shown by the preceding experiments, indicate that apter-

Digitized by

Google

686 THE AMERICAN NATURALIST [VoL.XLVni

ous is a member of Group 11 and is, presumably, asso-
ciated or linked with all other characters belonging to
that group.

Supposedly this association of the characters results
from an association of the factors responsible for them
in the germ cells. And this latter association has been
explained upon the assumption that factors responsible
for linked characters are located in the same chromosome.
The hypothesis has even been carried so far as to postulate
a linear arrangement of the factors within the chromo-
some— the relative position of the factors being deter-
mined by the degree or amount of linkage existing between
them. This conception and the data upon which it is
based have been amplified by Morgan and by Sturtevant,
and need not be dwelt on here. In the present case no
attempt has been made to ascertain the exact degree of
linkage between apterous and other characters in the
group, except black, because of the difficulty of breeding
the apterous flies. Apparently the linkage between apter-
ous and black is very close, if not complete, since no ease
of ** crossing over'* was observed among the 875 Fa off-
spring in this experiment (Table VI). There is a possi-
bility that the classification of the Fj apterous flies as all
gray is not absolutely correct, because, owing to the diffi-
culty of distinguishing gray from heterozygous black in
apterous specimens, an occasional heterozygous fly might
have passed for pure gray. However, if there had been
any appreciable number of cross-overs in this direction,
there would also have been some in the opposite direction,
which fact would have been indicated by the presence of
winged, gray flies. And since none of these were obs?rved,
it is safe to conclude that few or no cross-overs occurred,
and hence that apterous is very closely, if not completely,
linked to black.

Experiment F.— To determine the relation between the
apterous mutant and the '* vestigial'* mutant.

Among the mutant characters of Group 11 is one called
'* vestigial wing.'' Flies having this character are more

Digitized by

Google

No. 575] AN APTEROUS DROSOPHILA 687

like the apterous individuals than are any of the other
mutants, and since the two characters belong to the same
group the question arose as to whether or not the factor
responsible for one might be simply a modification of that
responsible for the other. Experiment V was performed
to determine this point.

Long-winged flies heterozygous for apterous were
crossed with vestigial winged individuals, and Fi and Fj
generations raised. The Fi flies were all long winged,
which fact in itself indicates the independence of the two
characters, for if they were allelomorphs either apterous
or vestigial should have appeared. In Fg both vestigial
and apterous, as well as long-winged, flies appeared,
showing conclusively, the independence of the two
characters.

Summary of Experiments

Experiments I and II show that the apterous character
is a simple Mendelian recessive, which independently
mendelizes with miniature wings, white eyes and ver-
milion eyes, and hence is not sex-linked (i. e., not a mem-
ber of Group I).

Experiment III shows that the apterous factor is trans-
mitted independently of the factor for pink eye, thus indi-
cating that apterous is not a member of Group HE.

Experiment IV shows a linkage ratio to result from
crosses involving apterous and black, the ratio being such
as to indicate a very close linkage between apterous and
black, and to identify apterous as a member of Group II.

Experiment V shows apterous to be distinct from ves-
tigial wing, to which it bears a considerable degree of
resemblance.

Origin and Characteristics of the Apterous Mutant

The description of the apterous fly has been deferred
up to this point in order that it might be combined with a
discussion of the experimental results.

The mutant has appeared upon several occasions, but

Digitized by

Google

688 THE AMERICAN NATURALIST [VoL.XLVHl

always in the same stock (miniature white), and always
with the same evidence of weakness and low viability.
Fig. 1 is a camera drawing of a typical specimen, made
by Miss E. M. Wallace. In morphological characters the

fly differs from the normal in
being entirely destitute of wings
and in possessing greatly re-
duced balancers. Likewise in
Pig. 1. aptbbous Mutant op physiological characters it de\i-

Drosophila ampelophUa, * "^ ., . i .

ates strikingly trom the normal.
This is best shown by means of a comparison between
apterous flies and normal flies from which the wings
have been removed. The latter are not appreciably
inconvenienced by their loss of wings ; they show char-
acteristic vigor in their active running and jumping
movements, they easily right themselves if overturned,
or extricate themselves if entangled in food or cotton, and
they are long lived and breed as prolifically as do winged
individuals. In fact they show no ill effects except the in-
ability to fly. The true apterous individuals, on the other
hand, show marked abnormalities in all these respects. In-
stead of being vigorous and active they are weak and usually
sluggish ; if overturned they have great diflSculty in right-
ing themselves ; or, if entangled in food or cotton, they are
usually unable to extricate themselves and consequently
perish. Moreover, they are always short lived, even when
kept under the best possible conditions and prevented from
becoming entangled in food or cotton. And lastly they
exhibit a most marked inability to breed, as noted in the
experiments. This characteristic, as has been mentioned
above, is so marked that the apterous flies were at first
thought to be sterile. I am convinced now, however, that
the diflBculty is not one of sterility at all, but is due to a
physical weakness which makes it extremely diflScult for
the flies to copulate, and for the females, even when fertil-
ized, to produce and lay eggs. Cytological examination
has shown that the males produce spermatozoa in an ap-
parently normal manner, yet prolonged observation of

Digitized by

Google

No. 576] AN APTEROUS DROSOPEILA 689

the flies has not revealed a single copulation or attempt at
copulation on the part of an apterous male.® Similarly
the females have been shown to produce rudimentary eggs
in an apparently normal manner, and in two cases females
have produced offspring when fertilized by winged males,
thus indicating their fertility. But many other cases
have been observed in which apterous females were fertil-
ized by winged males (or at least in which copulation took
place), and yet in these observed cases the females invari-
ably died without producing offspring,® because, I believe,
of their physical weakness.

From these facts it appears practically certain that the
apparent sterility is not due to infertility of either sperm
or eggs, but results from a weakness which makes it very
difficult for the apterous flies to perform the reproductive
processes.

This explains why no crosses have been secured between
apterous and apterous, although each sex has been suc-
cessfully crossed to winged. It is also supported by the
fact that from the cross between apterous male and
winged female a large number of offspring were secured,
since the winged female could produce many eggs, —
whereas in the two crosses between apterous females and
winged males only a very few offspring were secured,
because the apterous females could only produce a few
eggs.

When the experiments were first begun it was hoped
that sooner or later one or more inherently vigorous
apterous flies would appear which might give rise to a
vigorous race. But nothing of the sort took place,
although numbers of the apterous flies were given oppor-
tunity to breed all through the course of the experiments.
Obviously, then, the physiological characteristics, as

« Copulation in nonnal flies can be observed with very little difficulty. It
is evident that at least one case of copulation by an apterous male occurred,
namely in Experiment II.

» Judging from the cases observed a large number (probably one hundred
or more) of apterous females must have been fertilized by winged males
during the course of these experiments, yet only three of these gave progeny.

Digitized by

Google

690 THE AMERICAN NATURALIST [Vol. XLVIII

shown by vigor and viability, are directly associated with
morphological characters and are not to be separated from
them by selection. In other words, the ** factor*' respon-
sible for lack of wings is also responsible for physiological
disturbances.

The only suggestion of an inherent difference between
different races, or strains of apterous, is the slight differ-
ence in the percentage of apterous offspring in Experi-
ments I and in as compared with n and IV. In I and III
the ratios of winged to apterous are 3.12 : 1 and 3.24 : 1,
respectively, while in Experiments 11 and IV they are
4.06 : 1 and 4.02 : 1. This deviation is not great, but it is
fairly constant, and is sufficient, I believe, to indicate a
real difference. But whether it is to be explained upon
the assumption that in I and HI the apterous parents were
inherently stronger than in 11 and IV is not so clear. It
might equally well be explained upon the basis of differ-
ences in the winged races to which apterous was crossed
Unfortunately, an experimental analysis of the question is
prohibited by the difficulty of breeding the apterous flies,
and it must, therefore, be left open. One fact, however,
is clear, namely that there is no progressive increase in
viability of the apterous flies, for the apterous parent in
Experiment II, where the viability appears to be low, was
descended directly from that in I where it appears to be
high, and likewise the parent from the apterous side in
IV was obtained directly from HI.

In conclusion it may be profitable to call to mind
briefly the bearing of certain of the above data on the
question of the nature and behavior of Mendelian
''factors.*'

The present case of a definite correlation between lack
of wings, reduction in size of balancers, and weak physical
constitution in the apterous race of Drosophila, shows
clearly that one factor may have far reaching effects, and
not be limited to any particular part or organ,— a fact

Digitized by

Google

No. 575] AN APTEROUS DROSOPHILA 691

which has been long known, and often mentioned,*® bnt
by no means universally recognized.

Correlated with, or resulting from this principle is the
conception that the final result of ontogenetic develop-
ment is not due to the independent action of various
factors and their products, but is due to the combined
action, or the mteraction of these products during devel-
opment. To illustrate by the wing of a fly, — it is probable
that the normal development of such an organ is not
dependent solely upon one factor, but that it is influenced
by many factors. This is strongly suggested by data
derived from the various wing mutations in Drosophila.
These have dealt with a large number of factors, each of
which is responsible for a definite wing modification. For
instance, one factor is responsible for miniature wings,
another for vestigial, another for rudimentary, another
for curved, etc.* * From the fact that these mutant factors
(which may be considered as modifications of factors in
the normal fly) influence the wings, it seems highly prob-
able that their normal allelomorphs also influence wing
production in the wild fly.

Finally I wish to thank Dr. T. H. Morgan for kindly
f^ssistance and advice in connection with this work.

BIBLXOGRAPHY

Dexter^ John S.

1912. On Coupling of Certain Sex-linked Characters in Drosophila.
Biol Bull, Yo\. 23, p. 183.
Morgan, T. H.

1911. An Attempt to Analyze the Constitution of the Chromosomes on
the Basis of Sex-limited Inheritance in Drosophila. Jour, Exp,
Zool, Vol. 2, p. 365.
1912a. Eight Factors that Show Sex-linked Inheritance in Drosophila.
Science, N. S., Vol. 35, p. 472.

10 Most recently, perhaps, by Morgan (1913a, page 9): '*A change in a
factor may have far-reaching consequences. Every part of the organism
capable of reacting to the new change is affected. Though we seize upon the
most conspicuous difference between the old type and its mutant, and make
use of this alone, every student of heredity is familiar with eases where more
than the part taken as the index is affected. Weismann's theory, on the
other hand, seems to identify each character with a special determinant ..."

11 The same is true for various eye colors, and body colors.

Digitized by

Google

692

THE AMERICAN NATURALIST [VoL.XLVin

1912b. A Modification of the Sex Batio, and of Other Batios, :n
Drasophila through Linkage. Zeit, f, ind. Ahst. «. Vererb.,
Bd. 7, p. 323.

19X20. Heredity of Body Color in Drosaphila. Jowr. Exp. Zool., Vol.
13, p. 27.

191 2d. The Explanation of a New Sex-jatio in Drosophila and Com-
plete Linkage in the Second Chromosome of the Male. Science,
N. a, Vol. 36, p. 718.

1913a. Factors and Unit Characters in Mendelian Heredity. Aiceb.
Nat., Vol. 47, p. 5.

191 3&. Simplicity versus Adequacy in Mendelian Formulae. Aic£&.
Nat., Vol. 47, p. 372.
Morgan, T. H., and C. J. Lynch.

1912. The Linkage of Two Factors in Drosophila that Are Not Sex-

linked. Biol, Bull, Vol. 23, p. 174.
Morgan, T. H., and E. Cattell.

1913. Additional Data for the Study of Sex-linked Inheritance in

Drosophila, Jour, Exp, Zool,, Vol. 14, p. 33.
Sturtevant, A. H.

1913a. The Linear Arrangement of Six Sex-linked Factors in

Drosophila, as Shown by their Mode of Association. Jour,

Exp, Zool, Vol. 14, p. 43.
1913&. A Third Group of Linked Genes in Drosophila ampelophxla.

Science, N. S., Vol. 37, p. 990.

Digitized by

Google

SHORTER ARTICLES AND DISCUSSION
FORMULA FOR THE RESULTS OF INBREEDING

In connection with Pearl's recent valuable analyses of the
results of inbreeding (1, 2, 3), a comparison of these results with
those from self-fertilization is of interest. In my note on the
latter (4), I gave a formula for the rate at which organisms
become homozygotic through continued self-fertilization. This
occurs more slowly in the various types of inbreeding, but Pearl
gives no general formula for it. For purposes of comparison I
have worked out from Pearl's data the general formula for the
rate at which organisms become homozygotic through continued
brother by sister mating; as such formulae appear to be of perma-
nent value, it is here given.^ What the formula gives is, pre-
cisely, (1) the proportion of individuals that will be homozygotic
for any given character after any number of unbroken genera-
tions of such inbreeding, (2) the average proportion of the char-
acters of a given individual that will be homozygotic after any
number of unbroken generations of such inbreeding. The nu-
merical value so obtained may conveniently be called the co-
efficient of homozygosis.

The formula turns out to be a combination of the successive
powers of 2, with the successive terms of the Fibonacci series,
which appears in so curious a way in various natural phenomena.
In this series every term is the sum of the two preceding terms,
the series beginning : 0, 1, 1, 2, 3, 5, 8, 13, etc.

Let x = the coefficient of homozygosis.

n = the. number of inbred generations (the number of
times successive brother by sister mating has
occurred).
fu Uy /s» etc., = the successive terms of the Fibonacci series
(thus /i = 0, /, = !, etc.).
Then the formula for the coefficient of homozygosis is:
_2n-i+/^ . 2«-2+72 . 2*'-» . . . etc.

(The terms in the numerator are continued until the exponent
of 2 becomes 0.)

1 In conversation, Dr. Pearl urged the publication of the present note,
otherwise I should not at this time have dealt with a matter which he has
under analysis.

693

Digitized by

Google

694

THE AMERICAN NATURALIST [V0L.XLVIII

Thus, if the number of inbreedings (w) is 1.

a? =21 ==1/2, or 50 per cent.

If n = 4

23 + 0.2« + 1.2^ + 1.2«

— =11/16, or 68.75 per cent.

If n = 9

2« + 0»2^ + 1>2" + 1-2° + 2.2* + 3.2» + 5.2^ + 8.2»

l.^t

If n = 16

_ 63819
^~ 65536

2»
= 457/512, or 89.26 per cent.

or 97.38 per cent.

As these examples show, the formula gives the results *^^^
were obtained by Pearl in the detailed woAing out (so "f ^^"^ ^
this was carried), as given in Pearl's table I (2, p. 62). (X'fc ''^^'l
be noted that Pearl counts as generation 1 the one before int>x"0«d-
ing has occurred, so that his generation 10, for example, is t^l^^-'^ ^T^
which there have been 9 inbreedings (n = 9).)

If one is working out the values of the coefiScient x for a j^ei^^"*
of generations, the above formula may be expressed as a siTX^f^.
rule, applicable after the value for n = l is obtained. ^^
rule is :

The value of the coefficient of homozygosis x for any term (^
the ntK) is obtained by doubling the numerator and denominaio^
of the fraction expressing the value for the previous term, and
adding to the numerator the corresponding (n — 1th) term of
the Fibonacci series.

Or, in view of the peculiar nature of the Fibonacci series, the
rule may be expressed as follows :

Double the numerator and denominator, and add to the nu-
merator the sum of the last two numbers so added.

Thus, since

X for 1 inbreedings 1/2

^ „ „ „ _2Xl + 0

2X2

X2 4

2X4

_2X5 + 1_

= 2/4

2X4 -^/^

2X8

11/16, etc.

Digitized by

Google

No. 575] SHORTER ARTICLES AND DISCUSSION 695

After obtaining x, or the proportion of homozygotes for any
one pair of characters, the proportion y for any number m of
pairs is obtained simply by raising x to the mth power, that is :

Thus, after two generations of brother X sister mating, the
proportion of homozygotes for three pairs of characters is
(l/2)» = l/8, or 12.5 per cent. After 8 generations of such
inbreeding the proportion homozygotic for 10 pairs of char-
acters is :

/222y«_
\256y' ~

24.05 per cent.

The corresponding value in the case of continued self-fertili-
zation is 99.61 per cent. (4, p. 491).

Whether it may be possible to obtain a similar formula for
the coefficient of homozygosis in the cases of mating of cousin X
cousin or of parent X offspring, remains to be discovered.

Pearl's "coefficient of inbreeding" gives the percentage of
lacking ancestors in a given pedigree, as compared with the
number that would be present if all the parents were unrelated.
In order to compare self-fertilization with inbreeding in this
respect, Pearl 's formulae for the coefficient of inbreeding may be
expressed in terms of the number of successive inbreedings (n) ;
for many purposes the formulae appear more convenient so ex«
pressed. The following gives these formulae for self-fertilizatiop
and the three types of inbreeding, together with those, so far as
worked out, for the proportion of individuals homozygotic with
respect to a given character. In all these, n is the number of
successive self-fertilizations or inbreedings.

Coefficient of Inbreeding. Coefficient of Homozygosis.

2* - 1 2* - 1

Self-fertilization —^- ~~2»

i> *u V. Q- * 2" - 2 2»-» +/, ■2"-^+/f2"-« . . -etc.
Brother X Sister ^n ~ " 2*

2*"* — 2
Cousin X Cousin ^ — ?

2»» — n — 1
Parent X Offspring ~ O" ^

It will be observed that in self-fertilization the value of the
coefficient of inbreeding is, curiously, the same as that of the
coefficient of homozygosis, while in the other cases there is no
evident simple relation between the two. Further, the coefficient

Digitized by

Google

696 THE AMEBIC AN NATURALIST [VoL.XLVin

of inbreeding in brother X sister mating is the same as for self-
fertilization, save that it lags one generation behind the latter;
thus the coefScient for the fourth generation of self-fertilization
is the same as that for the fifth of brother X sister mating. Pearl
(1, p. 592) has already pointed out that in cousin mating the
coefficient is one-half that for brother X sister, with a lag of one
generation; as compared with self-fertilization the lag is two
generations. No such simple relation is apparent between the
proportions of homozygotes resulting from the diverse methods
of breeding, though possibly such may yet be discovered.

H. S. Jennings

PAPEES CITED

1. Pearl, E. A contribution toward an analysis of the problem of inbreed-

ing. This JoUENAL, XLVII, October, 1913, pp. 577-614.

2. . On the results of inbreeding a Mendelian population; a correction

and extension of previous conclusions. This Journal, XLVllI, Jan-
uary, 1914, pp. 57-62.

3. , On a general formula for the constitution of the nth generation of

a Mendelian population in which all matings are of brother X sister.
This Journal, XLVIII, August, 1914, pp. 491-494.

4 Jennings, H. S. Production of pure homozygotic organisms from hetero-
zygotes by self-fertilization. This Journal, XLVI, August, 1912, pp.
487-491.

A SHOET-<:!UT IN THE COMPUTATION OP CERTAIN
PROBABLE ERRORS

In his handbook of statistical methods, on p. 38, Dr. C. B.
Davenport^ ^ves a short method for the calculation of the prob-
able errors of some of the commonest statistical constants, in a
table of logarithmic formulae. It would seem that the simple
and obvious short-cut involved has not been given the attention
it deserves in connection with non-logarithmic calculation. The
logarithmic formulae are as follows:^

(1) log Ej, = log .6745 + log 0- - ^ log n since Ej, = .6745 ~ .

log E, = log i;^ - ^ log 2 since E^ = .6745 -J-,

(2) L V2n

or, £. =E^^V2j,

1 Davenport, C. B., ' ' Statistical Methods with Special Belerence to Bio-
logical Variation,'' 2d ed., 1904, New York, John Wiley & Sons.

2 A indicates the weighted arithmetic mean, <r the standard deviation, and
C the coefficient of variability.

Digitized by

Google

No. 576] SHORTER ARTICLES AND DISCUSSION 697

(3)3 log Ec = Jog E, - log A [since E^^E,^ A].

Now, if one is working with a calculating machine, he can
simply carry the value of Ej^ to two or three more decimal places
than are to be retained, and then divide by the square root of
2 to get E^; similarly, the latter value, divided by the mean,
gives Ec*

The writer prefers, however, to calculate the values in the
ordinary way on the machine, using Miss Gibson V table for

-— p— and ' — , and then to use the short method in checking.
Vn ^|2n

The original computations can be indicated and performed with
great confidence and rapidity, since it is hardly possible to make
an error that will not be discovered in the checking.* It is
obviously safer, as well as much quicker, to check in this way
than to repeat the original processes. Howard B. Frost

citeus expbeiment station,
Riverside, Cal.

GALTON AND DISCONTINUITY IN VAEIATION

It seems not to be generally realized that Galton recognized
both continuity and discontinuity, both in variation and inherit-
ance. Of course, all biologists are familiar with ' * Galton 's poly-
gon,'' in which slight oscillations of the polygon on one of its
faces, but without a change of face, are compared with ** small
unstable deviations" (fluctuations), while a larger oscillation, in
which the polygon moves over to a new face, is compared to a
sport ... of such marked peculiarity and stability as to rank as a new
type, capable of becoming the origin of a new race with very little as-
sistance on the part of natural selection.^

Galton 's polygon illustrated for him how the following condi-
tions may co-exist:

(1) Variability within narrow limits without prejudice to the purity
of the breed. (2) Partly stable sub-types. (3) Tendency, when much
disturbed, to revert from a sub-type to an earlier form. (4) Occasional
sports which may give rise to new types.

These four types would seem to correspond rather well to what

8 Formula (3) gives, of course, the approximate or uncorrected value
of EC.

♦ Gibson, Winifred, "Tables for Facilitating the Computation of Prob-
able Errors," Biometnka, 4: 385-393. 3 tables.

s Unless, of course, one misreads the figures from the machine in checking.

1" Natural Inheritance," London, 1889, p. 28.

Digitized by

Google

698 THE AMERICAN NATURALIST [V0L.XLVUI

are now called (1) fluctuations or *' non-inherited" (in reality, I
think, partially inherited) continuous variations; (2) instability
resulting from a heterozygous or partially heterozygous condi-
tion; (3) reversions, now believed to result chiefly from cross-
ing; and (4) mutations.

Galton is equally explicit in other statements on this subject.
Like Darwin, he admitted the facts both of continuity and dis-
continuity in variation; but, unlike Darwin, he also recognized
discontinuity or alternation as well as continuity or blending, in
inheritance. Thus he says, in a paragraph headed "stability of
sports":^

Experience does not show that those wide varieties which are called
" sports " are unstable. On the contrary, they are often transmitted to
successive generations with curious persistence. Neither is there any
reason for expecting otherwise. While we can well understand that a
strained modification of a type would not be so stable as one that ap-
proximates more nearly to the typical center, the variety may be so wide
that it falls into different conditions of stability, and ceases to be a
strained modification of the original type.

In another paragraph,* headed ''Evolution not by minute
steps only,'' he says:

The theory of evolution might dispense with a restriction, for which
it is difficult to see either the need or the justification, namely, that the
course of evolution always proceeds by steps that are severally minute,
and that become effective only through accumulation. That the steps
may be small and that they must be small are very different views; it is
only to the latter that I object. . . . An apparent groxmd for the com-
mon belief is founded on the fact that wherever search is made for in-
termediate forms between widely divergent varieties, whether they be of
plants or of animals, of weapons or utensils, of customs, religion or
langfuage, or of any other product of evolution, a long and orderly series
can usually be made out, each member of which differs in an almost im-
perceptible degree from the adjacent specimens. But it does not at all
follow because these intermediate forms have been found to exist, that
they are the very stages that were passed through in the course of evo-
lution. Counter evidence exists in abundance, not only of the appear-
ance of considerable sports, but of their remarkable stability in hered-
itary transmission.

Again, Galton not only believed in the existence of both
blended and alternative inheritance, but he recognized the im-
2 L. c, p. 30.

8 L. c, p. 32.

Digitized by

Google

No. 575] SHORTER ARTICLES AND DISCUSSION 699

portance of the latter in connection with the survival of new
races. Thus he writes:*

The quadroon child of the mulatto and the white has a quarter tint;
some of the children may be altogether darker or lighter than the rest,
but they are not piebald.^ Skin-color is therefore a good example of
what I call blended inheritance. . . .

Next as regards heritages that come altogether from one progenitor to
the exclusion of the rest. Eye-color is a fairly good illustration of
this. . . .

There are probably no heritages that perfectly blend or that abso-
lutely exclude one another, but all heritages have a tendency in one or
the other direction, and the tendency is often a very strong one.

On the following page Galton remarks that

A peculiar interest attaches itself to mutually exclusive heritages,
owing to the aid they must afford to the establishment of incipient races.

He thus recognizes the invalidity of Darwin's objection to
"single variations" as a factor in evolution, namely, that they
would certainly be swamped by crossing with the general popu-
lation.

It would, therefore, appear that in his recognition of continu-
ity as well as discontinuity both in variation and heredity, Galton
was in advance of his time, and more in accord with some of the
current views. B. Buggles Gates

Univebsity or London

EEPULSION IN MICE

In the February number of the American Naturalist Dr. C.
Little criticizes the results of my mouse-breeding experiments
which I published in the Zeitschrift fur Induktive Abstam-
mungS' und Vererbungs-lehre Bd. VI, Heft 3. The chief point,
on which he disagrees with me, is the interpretation of the results
I obtained in breeding black and albino mice together.

The fact is, that in my paper on mice, I overlooked a serious
error. In three sentences on page 126, relating to test matings
of albinos, the words "black" and ** agouti" changed places. As
printed in the paper these sentences run:

Without exception they have given black or equal numbers of black
and albino young, depending upon the purity of the black used. But
never has one of these albinos produced a single agouti young in a mat-
ing with black. Counting together the colored young of such families I
get 89 black young.

4X. c, p. 12.

B Cases of piebaJdism in such crosses are of course now well-known.

Digitized by

Google

700 THE AMERICAN NATURALIST [Vol. XLVIII

These errors were corrected in an "errata" in Band VI, heft 5,
which Dr. Little unhappily did not find. The sentences should
read:

Without exception they have given agouti, or equal numbers of <igout%
and albino young, depending upon the purity of the black used. But
never has one of these albinos produced a single hlack young in a mating
with black. Counting together the colored young of such families I
get 89 agouti young.

Professor Punnett was so kind as to draw my attention to
these mistakes. They were corrected in the reprints sent out

The facts were simply these : Albinos were bred of two sorts,
with and without 0 (the gene which agoutis have more than
blacks) . These albinos can only be distinguished by test-mating
them to blacks. The albinos with O (aO) give agouti young, if
mated to black (Ag), the ag albinos give black young from such
a mating. In one series, some agoutis were produced, which
were heterozygous for A as well as for 0{AaOg). Ordinarily,
sueh agoutis, when mated inter se, produce 9 agouti (1 AAOO,
2 AAOg, 2 Aa60, 4 AaOg), 3 black (1 aaOO, 2 oaGg) and 4
albinos ( 1 aaOO, 2 aaOg, 1 aagg) in every sixteen. Mated to
albinos without 0{ag) the ordinary AaOg animals give four
kinds of young, agoutis (AaOg), blacks (Aagg) and two kinds
of albinos (aaOg) and (aagg) in equal numbers.

Now these particular AaOg animals did not produce four
kinds of gametes, as expected, namely, AO, Ag, aO and ag, but
only two kinds, Ag and aO. Thirty one agoutis were test-mated
to adgg albinos. These test matings gave 181 young, of which 94
were black (Aagg) and 87 albino (aaOg). No agoutis were
produced.

As a further proof, the result of breeding these agoutis inter se
can be adduced. These matings gave 73 agouti (AaOg), 37 black
(AAgg) and 32 albinos (aaOO) . Of these 32 albinos, thirteen
were tested by inating them to blacks. If one of them should
have lacked 0, it would have given black young. But no black
young were produced. Some young were albino (when the black
parent was heterozygous for A), but all the colored young were
agouti (89 in all).

This, I hope, will make it perfectly clear, that in this series
we have been dealing with a case of repulsion between the genes
A and 0, A. L. Hagedoorn

BussuM, Holland

Digitized by

Google

No. 575] SHORTER ARTICLES AND DISCUSSION 701

THE OSTEOLOGY OP A DOUBLE-HEADED CALF

Through the kindness of Mr. Charles 0. Reed, taxidermist, of
Fairmont, W. Va., the writer received the skulls and anterior
cervical vertebrae of a double-headed calf which seemed of suflR-
cient interest to warrant a brief description.

According to Mr. Reed the calf's mother was a four-year-old,
thoroughbred Herford, living at Grafton, W. Va., owner not
mentioned.

At her first labor this cow gave birth to twins, supposedly
normal, though it was not so stated. The second calf was
''slightly deformed," but in what way Reed did not know. The
third labor produced the double-headed calf in question, which
was of unusual size, and was killed in parturition. According
to Reed "This calf would have lived if it could have been brought
through 0. K." He dissected it and found the "alimentary
canal, blood vessels and trachea normal."

The bones in the occipital region are slightly broken, probably
done in disarticulating the skulls from the neck ; and in the left
skull the left premaxilla was lost and was replaced by a roughly
carved piece of wood for the sake of synunetry.

In macerating the skulls, for the purpose of removing all the
flesh, many of the loose sutures separated, and in gluing the bones
together again it was not always possible to completely close the
sutures.

As may be seen in the figures there is a considerable though not
very great difference in the size of the skulls, the right being the
larger. They were detached from the cervical vertebrae when
received, but the photographs show their approximate position in
relation to the neck and to each other.

Each skull is twisted and bent away from the other, the bend
being most marked just cephalad to the orbits. The left skull is
the more distorted.

The articulation of the skulls with the fused atlas was so crude
that Reed, who had seen the skulls before disarticulation, had to
be appealed to to decide which skull was right and which was left.

Pig. 1 is a photograph of the dorsal aspect of the skulls and
the first three cervical vertebrae. The distortion of the two skulls
is of about the same character but is, as noted above, more marked
in the left skull.

The parietal (p) is normal. The posterior regions of the fron-
tals (/) are normal, but their anterior ends are bent laterally.

Digitized by

Google

702 THE AMERICAN NATURALIST [Vol. XLVIII

which causes a slight curvature in the sagittal suture. It is in the
region of the lachrymals (1) that the distortion is most marked,
so that the lachrymal on the convex side of the bend, especially

Fig. 1. DoBSAi. Views op the Two Skulls and of the First Tbbee Cervical

VBRTEBRiE. MAKDIBLBS IN POSITION.

Oj atlas ; ax, axis ; e, extra bone between maxilla and premaxllla ; /, frontal ;
I, lachrymal ; m, malar ; mx, maxilla ; n, nasal ; o, occipital ; p, parietal ; pm,
premaxllla.

in the left skull, is much longer than that on the opposite side ;
the same is true of the malars (m), of the maxillaries (mx), and,
to a less degree, of the premaxillae (pm). The nasals (ti) are also
unsynmietrical, but do not differ much in size; they are simply,
as a pair, pushed to the side.

Fig. 2. The ventral aspect of the skull shows even greater ab-
normalities than the dorsal. The occipital (o), as noted above,
was somewhat injured by the person who disarticulated the skulls
from the neck, but it is quite unsymmetrical, especially in its
exoccipital region. In the left skull (right in this figure) all the
other bones seen in this aspect are bent, but in the other skull
most of the bones are comparatively straight.

In the right skull a suture in front of the teeth separates off an
extra bone (e) on each side, between the maxilla and the pre-
maxilla, that of the right side being much the larger. In the left
skull these extra bones are not present though a partial suture,

Digitized by

Google

No. 575] SHORTER ARTICLES AND DISCUSSION 703

extending about half way through the left maxilla, is visible in
this view of the skull.

Fig. 2. Vbntbal Views of the Skulls and the Fibst Two Cervical Ver-
TEBRJB. Mandibles Removed.

Fig. 3 shows the curious distortion of the mandibles, which seem
to be bent in more or less the same direction. In the right man-
dible the left half has four incisor teeth, the right half has three.
In the left mandible the right half has four teeth, the left half has
three, though one tooth is missing from each half.

As noted above, the skulls, when received, were disconnected
from the vertebras; but the latter, three in number, were strung
together on a small piece of rope and presumably were the first
three cervicals; they are shown in a dorsal view in Fig. 1, ven-
tral view in Fig. 2, and anterior view in Fig. 3. The first of
these is presumably a compound atlas (a) since it articulates
with each of the skulls, though in a very crude way. It consists
of eight loosely united elements which became completely sepa-
rated in cleaning and had to be glued together again. In the
dorsal view. Fig. 1, is seen a small, irregular bone from which
radiate three somewhat symmetrical bones, the largest lying in
the median plane between the bases of the skulls. This larger

Digitized by

Google

704 TRE AMERICAN NATURALIST [VoL.XLVin

bone is pierced by two large foramina; eadi of the other two
bones shows in this view a foramen which branches and opens
both on the antero-median and the postero-lateral surfaces. The

Fig. 3. Dorsal Views of the Mandibles and Anterior Vibws of ths Fibist
Three Cervical Vertebra.

ventral view, Fig. 2, shows a very irregular group of bones, the
smallest of which is for articulation with the following vertebra.

The second vertebra (oa;), supposedly the axis, exhibits no
indication of an odontoid process and articulates in a very crude
way with the preceding bone. Its dorsal spine is rather elon-
gated in an antero-posterior direction, but otherwise it bears no
closer resemblance to an axis than to any other cervical vertebra.
Its centrum was so loosely fused with the arch on either side
it became detached in cleaning and had to be glued in place.

The third vertebra exhibits no peculiarities that warrant de-
scription.

A. M. Reese.

West Virginia Univeesity,
Morgantown

Digitized by

Google

Two Important Biological Journals

The Journal of Qenetic3

Edited by
W. BATESON and R. C. PUNNETT
A periodioal for the pnblioatioD of reoordB of orieinal research in heredity^ variation, and allied topior,
The JoUKNAi. also contains, from time to time, articles summarizing the existing state of knowledge in the
yarioos branches of genetics. The numbers are follv illostrated, and, when the subject-matter demands it.
colored plates are freely employed. The Journal is issued in parts as material accumulates, and a yolume
(appearing, so far as possible, annually) consists of four such parts. Subscription price, $7.60 a volume ;
nngle oopiea, 12.50.

Biometrikia

Edited by EARL PEARSON
A journal for the statistical study of biological problems, appearing about four times a year. A volume
containii:^ about 500 pages, with plates and tables, is issued annually. Volume X begins in 1914. The
nine jyreoeding volumes form a veritable eni^clopaedia of biological statistics in the fields of anthropometry,
oianiometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, 12.50.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes
mad single copies. ^____^-..««-.__

Published in America for the Cambridge University Press of England by

The University of OHicagro Press
ChicasTO, Illinois

University Control

By J. MoEjqbbn Cattell, Professor of Psychology in Oolambia University

ToostiMr with a series of Two Hundred and Ninety-nine Unsigned Letters by Tieading Men of BelsBes
holding Aoademie Positions and Articles by Josnra Jastbow, Qbobob T. Ladu, John J. BnmraoM, J. B*
Omcuwom, J« MoEsbn Cattbu., Qsobob M. SouTfos, SrawAS? Patqh, John Jat OHAniAS, Jambs P.
MvBBOB and Jaoob Qovld ScsimifAir.

A great Tariety of questions oonoeming geoetal university administration are dealt with In an original and
Mpfui way.— iVocvTi.

These quotations and ezami>les are taken from Professor CatteU's Informed and thoronsh disoussion of the
sahjeet of university eontrol, a subject upon whidi he has had much to say of late, finding oocarion for eanstie
SKftioism of eristing American conditions, and standing as the champion of an academic demoecacy and a teaeb*
Ing profession upon whidi a man may enter without forfeiting his sdf*reqpect.— Tito Dial.

Sentences and paragraphs that betoken the expert, hlghly*trsined mind, the suggestions that come to ve-
fkBsh a d tell OS that a new day is about to dawn in educational writing. — Tk$ BotUm JShmUiao TfOMoripk

SCIHINCB AND EDUCATION

A merles of rolnmes for the promotion of eeientifLo research and ednesiionsl progress
VOLUME L The Foundations of Science. By H. PomoABi, Containing the authorised English translation

by QaoBQB Bbuob Halstbd of "Sdenoe and Hypothesis,'! "The Yahie of Science," and "Scknee

and Method." Price, $3.00 net.
VOLUME n. Medical Research and Education. By Rxokabd M. Pbabob, Wiluam H. Wblob, W. JL

HowBLL^ Fbahxloi P. Mall, LBWB1J.TS F. Babtbb, Cbablbs S. MnroT, W. B. Cahbob, W. T.

OoraoiLMAJr, Tbbobakd Smtth, Q. N. Stbwab? , C. M. Jaoksob, E. P. Ltob, Jaubs B. Hbbbick, Jobs

M. DoDBOB, G. R. BABnBBN, W. OfbOub, S. J. Mbiasbb, Jaubs Ewmo, W. W. Ebbb, Hbbbt H. Doxaso-

■oir, GBBiBnAB A. Hbbtbb, and Hbbbt P. Bowditob. Price, $3.00 net.
VOLUME nL Unireisity Control. N^ Eaa4y. Paget 9+484. Pries, %SM wKL

0ARRI50N* N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The American Naturalist

with Sp«cUI lUf erenc* to the Factors ol Orgwiic Evolutkm mnd H«M<lity

CONTENTS OF THE MAY NUMBER

EetoparMltMofltommAlt. ProfettorV^monLyiiuui
Kellogf.

Regeneration, Variation and Conelation in Thyone.

Professor John W. Scott.
Shorter Artlelefland DitcnsBion: Teimt relating to

Generic Typei. Dr.O. F.Cook.
Notes and Literature: Linkage in the Bilkworm

Moth. A. H. Stnrterant. Naboon't Breeding

BxperimentiwithGraiihoppezB. John a Dexter.

CONTENTS OF THE JUNE NUMBER

liassor John H. Gerould.
Heredity of BrlsUes in the Common Gieenhottle Fly—
A Study of Factors goTeming Distribution. Phinew
W. Whiting.

Physiological Con«]iitions tnd CUmatie iGeaetifloe in

AlfiUfk Breeding. Gea F. Fxeeman.
Taxonomy and Erolution. By X
Shorter Articles and Discussion* Nabonr^a fhass

Misleading Terminologies in GeneticsL Ptotaeor
w. E. Castle.

CONTENTS OF THE JULY NUMBER
Pattern Pereiopment in Mumtnftif gm^ Birde.

Glorer M Allen
Internal Belatione of TMr^f^ri^i AffKTdatfimf.

CONTENTS OF THE AUGUST NUMBER

Molttple Allelomorphs in Mioe. Professor T. H

Morgan.
Thirteen Yean of Wheat Selection. T. B. Hntch-

ItSrHr^^li^l^i^^^: ^ H.L.Babcock.

J. Stomps. Dr. Bradley M. Dayis. Th» Thearet*
ical DisUnction betw^ Multiple AUdomoraS

f<S2o?w*Rc!SlS: ^^'--^'.H.Mo^SSuS^

KotM and literature:

Biometrics. Dr. Raymond Peail. A New Mod-
Hermann J. Muller.

Arthur G. Vestal.
Shorter Articles and Difconion* Another Hy-
pothesis to Aeeonnft for Dt Swingle's Expert*
menti with Cltni. A. C. and A. L. Hagedoom,

CONTENTS OF THE SEPTEMBER NUMBER

Studies on Inhreeding. Dr. Raymond Pearl.

The Chromosome Hypothesis of Linkage ap'pUed to

Cases in Sweet Peas and Primula. Calrin B

Bridges.
The RedupUcation Hypothesis as applied to Droso-

phila. Dr. A. H. Stmrteyant.
Pattern Deyelopment in Mammals and Birds. Dr

GloTer M.Allen.
Shorter Articles and Correspondence :

The Bearing of the Selection Experiments of

Castle and Phillips on the VariaMllty of Genes.

Utrmann J. Muller.

CONTENTS OF THE OCTOBER .NUMBER

Sex-limited and Sex-linked Inheritance. Professor
T.H. Morgan.

Inheritance of Endoepenn Texture in Sweet x Waxy
Hybrids of Mahse. G. N. Ccdllns and J. H. Kemp-
ton.

A Study of Variation in the Apple. W.J.Toung.

Shorter Articles and Discussion :

Variation and Correlation in the Mean Age at
Marriage of Men and Women. Dr. J. Arthur
Harris, Roxana H. Vivian. DupUcate Genes.
Sewall Wright.

Notes and Literature:

A Study of Desert Vegetation. Professor Charles
E.Be88ey.

Single Number 40 Cento Yeariy SabMription, $4.00

THE SCIENCE PRESS

GaniMii, N. Ye Sub-Station 84 : NEW YORK liiiirfitter. Fb.

Digitized by

Google . S-

I I ' i 1

lie 7 ly.4

70L. Xim. NO. 676 DECEMBER, 1914

THE

AMERICAN
NATURALIST

A HOHTHLT JOUSHAL

DeTotad to the Advftnoement of the Biological Scienoei Witt

Special Beference to the Factors of Bvolvtion

OOVTBVTS

L The FaUnro of Btlier to Prodnco Mntattona In DrofophiU.

Profenor T. H. Mobqak 706

n. Tho Analyiif of a Oaio of Contiimoiu YArtatloii In DroiopbiU by a Btody

of lU Linkago Belatloni. ProfeBsor John S. Dbxtbb - - - - 7ia

m. Bliorter Articles and Oorrospondenco : On the Progressiye Increase of Homo-

zygoos Brother-Sister Matings. H. D. Fish ---■.-- 769

IV. Notes andUtoratnre : Mendelian Fluctuations: Q.U.Y. 763

V. Index to YolnmeXLin ----------- 763

THE SOIENOE PRESS

LAHOASTEB, PA. OABBIBON. H. T.

HEW TOBK: SUBSTATION 84

Digitized by

Google

The American Naturalist

MSB Inteided fer piblloatit ■ ud btokt, ete., iiteided for review eboiM be
•eitte the Editor of THE AMERICAN NATURALIST, Gtrrisoo-OB-Hidooo, New York.

Short artiolee ooRtaloiog eoMMmrieo of rotearoli work bearing oe the
probloMO of orgulo ovolotioa are oopoelally woleoae, aid will bo itvee prefisreaeo

One baadrofl reprleto of eontribatloBe are eappliod to autbero f^ee of ebarge.

Ill DO eappiied at eoet
Subobriptlone and aavertleoMento ohoald bo oent to tbe pabliebers. The

Fartber roprlnte will be eappliod at eoet

enbeoription priee ie fear dollare a year.

Canadian poetaie twonty-llvo eonte additional. Tbe ebarfo for elnfle oopieo ie

fbrty eonte. Tbe advertleini ratee are Fear Dollare fbr a page.

Foreiott poetage Ie tiny oonte and

Lancaster, Pa.

THE SCIENCE PRESS
NEW YORK: Sub-Station 84

Garrison, N. Y.

■ttrtd M MOond-elMi nuitttr, A»rtl 2, 1M8, at the Pott Office at Luieuter, Pa., andtr the Act of

Oongnn of lUreli 8, 1879.

FOR SALE

ARCTlCt ICELAND and GREENLAND

BIRDS' SKINS,

Won Proparod Low Priooo

Particulars of

C. DINESEN, Bird Collector

Huoavik, North looland. Via Loidio, Bnffland

JAPAN NATURAL HISTORY SPECIMENS

Perfeot Condition and Lowest Prtoen.
Speoiallj: Bird Skins, Oology, Entomology, Marine
Animals and olhera. Catalogne free. C<»reBpond-
enoe solioited.

T. FUKAI, Naturalist,

KoBOsu, 5altama, Japaa

For Sale

An important collection of Indian Birds'
Eggs, containing approximately 3,300
specimens of about 620 species, and in-
cluding among other varieties the Green-
ish Willow Warbler (Acanthopneuste
viridanus), Red-browed Finch (Calla-
canthis burtoni), Tibetan Sand-Grouse
(Ssrrrhaptes tibetanus), Spotted Sand-
Grouse (Pteroclurus senegallus), Ibis-
bill (Ibidorhynchus struthersi), White-
bellied Minivet (Pericrocotus erythro-
pygius), Bearded Vulture (Gypagtus
barbatus), numerous Cuckoos with their
hosts, etc., etc.

Particulars and full list may be had
from

W. F. H. ROSENBERG

57 Haveratock HiU London, N. W.

Marine Biological Laboratory

Wooda Hole, Maaa.

INVESTIGATION

INSTRUCTION
July— Anfiut

SUPPLY
DEPARTMENT

FftdDtiM fee wear eh bt 2oolosr»
BmbrTolosr, V^»jMULogf and Bo^
any. BevaVfmx privat* labor*>
ionaa, $100 eaoh for not ovw thne
rnoaths. Thirty teUeB am mnA-
able for besUtnen in maaareih who
dedrt to worh under the dlreetioB
of memben of the etaff. The fee
f or raoh a table ie SSaoO.

Coonee of laboratory
with leotoree are offared in lirperte-
brate 2oolo|^* Xnbcyuloajf PlQni*
fliQgj. ICor^olegy and Teinoamy
of me Algae, Phyaielogy aad
Bed^oTSaxine Stand aad Boc
Vecetttioin* Baeh eoane ceouiiee
the fuU time of the atodank I%e,
^~3.00. A lecture eoniae on tbe
" Aapects of Bieloo
tie alee offmSu

Preeenrad material of all Igrpee of
I aad of Algae. Fangi* liver-

Open the Enth« Year anfanals i__

worte aad

oUMWork, or for the mianun.
liring material f uxnlBhed fai aeieoti
aa ordered. Priee liata of ZoO-
loglonl nnd Botanieal ma«c«M
amt on appBeatlon. State whie^
to dariied. Per priee I ata and aS
Infonnatkm wgarding naatann],
addrtaa

GBO. li. GRAY, Cutator, Wooda Hole. Una
The aannal annoonoement will be aent on appBoatioa te
The Dinetor, ICariae Bioiockal Labocatuy, Weeds Bola
Maw-

Digitized by

Google

THE

AMERICAN NATURALIST

Vol. XLVIII Decemb^ 19U No. 676

THE FAILUEE OF ETHER TO PRODUCE MUTA-
TIONS IN DROSOPHILA

PROFESSOR T. H. MORGAN
Columbia Univebsitt

The many mutants of Drosophila ampelophila that have
appeared ** under domestication" have raised the question
as to the cause or causes that have brought about the re-
sult. Since every fly that has passed through our hands
has been etherized once in its life, usually before it begins
to lay its eggs if a female or before mating if a male, it
might appear that this recurring condition was respon-
sible for the mutations. At any rate it seemed worth
while to put this view to a test, if for no other reason
than to remove from one's mind the suspicion that ether
''did it.''

Preliminary trials showed that two drops of ether (on a
piece of cotton) in a quart milk bottle, tightly stoppered
with a cotton plug, would not noticeably ajBfect the flies in
half an hour, three drops made them slightly ''stupid,"
four drops more so, and five drops quieted them. It was
found that they would for tiie most part recover even
after 6, 7 and 8 drops of ether. If etherized twice daily
the flies were so far weakened that they generally died
without laying any eggs. Therefore in the later experi-
ments the flies were etherized only once a day or once in
two days.

The larvae (beginning two days after the eggs were laid,

705

Digitized by

Google

706 THE AMEBIC AN NATURALIST [Vol. XLVm

at which time the eggs have hatched) can stand more
etiier. Twice daily throughout their larval and pupal
lives (approximately 11 days) 6, or 7, or 8, or 9 or even 10
drops of ether were added to the quart bottles. The ether
excited the larvae at first, then quieted them ; later they re-
covered. In such tests the larvae were kept almost con-
tinuously in an atmosphere of ether from birth to emer-
gence of the fly and in a few cases the etherization was con-
tinued with the flies also. By covering a wide range of
stages and conditions I hoped to find the critical point, if
any such existed, when etiier would act. Since, as the
sequel will show, no specific results were obtained it seems
unnecessary to give the details of all these trials.

Double, and in one case triple recessive, stocks were
used for the work, because experience had already shown
that even with great care contamination may occur. One
or two flies that came from escaped mutants would ruin
the value of the data, but the operator can protect himself by
using stocks that have already two or more recessive char-
acters. If such flies mutate in one of the characters in-
volved the presence of the other one will make it certain
that the mutant belonged to this culture, and had not come
in from outside ; if a change appeared in some other part,
the double recessive character would still identify the
stock. Two of the stocks used had sex-linked characters,
i. e., eosin miniature and cherry club vermilion. If a
mutation should appear that involved tiiese characters it
would become evident at once in the male offspring; for,
tiie male gets his single sex chromosome from his mother
and exhibits her sex-linked factors. Of course this
would be equally true for any other sex-linked char-
acter that appeared, but in practise it is impossible to
thoroughly examine each fly in every possible part, so that
I had to confine my attention to certain organs, and in
these cases I concentrated on tiie mutant characters. Con-
spicuous mutations in other parts would, I think, have
been picked up, but minor ones would probably have been
missed. On the other hand, if changes taking place in

Digitized by

Google

No. 576] MUTATIONS IN DROSOPHILA 707

the chromosomal material are the basis for mutation it
would seem perhaps a priori unlikely that the same
changes should occur at the same time in both members of
a pair, and if not the effect would not appear in the next
generation, and not until two flies of the later progeny
each carrying one mutant factor met. Whatever weight
may be attached to this argument— we know really nothing
as to the origin of mutations— it seemed necessary to
carry some stocks to anotiier generation; and this was
done.

The following are the totals of offspring produced by
flies from larvae that had been etiierized twice daily from
the time of hatching to the winged state :

Black vestigial -I ._

Pink black 1,390

Eosin miniature 871

Cherry dub vermilion 364

Pink ebony sepia 1,311

Total 4,802

In the next case fewer drops of ether were used — four
or five throughout larval and pupal life. The parent flies
were changed to new bottles quite often to prevent crowd-
ing and abundant food was supplied :

Black vestigial 2,122

Pink black 6,762

Eosin miniature : . . . 2,603

Pink ebony sepia 953

Total 12,440

The following data are from the offspring of the flies
that had 8 and 10 drops of ether twice daily:

Pink black 3,440

Eosin miniature 2,775

Total 6,215

The next data are the records of the offspring of

Digitized by

Google

708 THE AMERICAN NATURALIST [Vol. XLVm

adult flies that had been etherized several times (usually
twice a day) just after they had hatched:

Black vestigial (6 times) 870

Black vestigial (7 times) H3

Black vestigial (8 times) 694

Pure black (6 times) 81

Eosin miniature (3 times) 206

Eosin miniature (5 times) 428

Cherry club vermillion (8 times) 713

Cherry club vermillion (8 times) 476

Total 3,611

Finally some of the flies that had appeared, in the ex-
periments in which 8 and 10 drops of ether had been used
throughout the larval and pupal life, were bred and gave
in the next generation the following records :

Pink black 2,186

Eosin miniature 666

Cherry club vermillion 709

Pink ebony sepia 539

Total 4,100

In a grand total of 31,168 flies subjected to ether, there
was not a single mutation observed. It seems highly
probable therefore that ether has no specific effect in pro-
ducing mutations in DrosophUa ampelophUa. It might,
of course, still be said that mutations are so rare, that,
although caused by ether, tiiey still are not frequently
enough produced to make 31,000 flies a sufficient guaran-
tee. Granting this, it still remains that since no mutants
appeared under this excessive treatment, ether does not
play the role of a specific agent causing the mutations of
DrosophUa, and one is inclined to look elsewhere for a
solution of the problem.

One of the first mutants that I observed in ampelophUa
appeared in the offspring of flies that had been treated
with radium and although there was no proof that the
radium had had a specific effect I felt obliged to state the
actual case, refraining carefully from any statement of

Digitized by

Google

No. 576] MUTATIONS IN DR080PHILA 709

causal connection.^ Nevertheless, I have been quoted as
having produced the first mutants by the use of radium.
I may add that repetition of the experiment on a large
scale both with the emanations of an X-ray machine and
from radium salts has failed to produce any mutations,
although the flies were made sterile for a time. Loeb and
Bancroft also tried the ejBfect of radium.^ They found a
black mutant tyi>e after treatment with radium but since
the same type appeared in the control they do not believe
that its appearance had any connection with the radium.
They also state that after treatment a white-eyed female
appeared in the first generation, and suggest that a white
eyed male may have existed in a previous generation that
escaped notice, but if it had been found in a previous gen-
eration, the mutation or the contamination must have
been earlier than the one that produced the white-eyed
female ; for, a white-eyed male takes two generations to
reappear again. Pink-eyed flies were also found both in
the control and in the treated flies. In regard to another
mutant type, they state :

We succeeded in producing short winged specimens in two different
cultures by treating them with radium, while thus far we have not
observed this mutation in cultures not treated with radium.

But although *'two hundred different cultures" were
subsequently treated with radium and no short-winged
(miniature) flies appeared, I get the impression that
Bancroft and Loeb must have had stock that was already
contaminated by some recessive mutant factors. All of
these mutants had been obtained and described by us, and
the stock used by Bancroft and Loeb was obtained in part
at least from my friend Dr. Frank E. Lutz, who had at
that time in his possession, as a letter I have from him
states, certainly two of these mutants, black and minia-
ture, that he had received from me. It seems to me not
improbable that the collector, who got the stock from Dr.

1 Science, XXXIII, 1911.

2 Loc, cit.

Digitized by

Google

710 THE AMERICAN NATURALIST [VouXLVm

Lutz for Professor Loeb, included by mistake some flies
heterozygous for these two characters; for in our very
extensive experience with wild stock from Cold Spring
Harbor (the origin of most of Dr. Lutz^s stock) and else-
where" these mutants have never arisen again.

At various times experiments have been made in this
laboratory involving wide ranges of temperature,* salts,
sugars, acids, alkalis without any resulting mutation.
In fact, our experience with DrosophUa has given us the
impression that mutations are rare events, although the
actual number of our mutants is now quite large.

Guyenot* also has treated ampelophila to high tempera-
tures, to radium and to X-rays without result. When the
adult flies were treated with ultra-violet light, however, a
definite type of * ^ black ^^ fly was obtained. The first eggs
tiiat such females lay are normal and give rise to normal
flies. The eggs laid later fail to hatch, although they ap-
pear to begin their development On the third day
amongst the abnormal eggs some were found that gave
rise to flies that were apparently normal. It happened
that they were not examined again until after the flies of
the next generation had appeared (many of them had
died). Both among the living and the dead flies there
was a considerable percentage of black flies. The black
females laid eggs which did not develop, even although
normal males were added. Why the black males were not
also tested by outcrossing is not apparent. The descrip-
tion of the black flies given by Guyenot tallies in some
points with our stock of ebony in which the females were
at first usually infertile but the males fertile. At first,
indeed, we kept the stock by breeding the ebony males to
the heterozygous females. These are intermediate in color.
In fact, Guyenot seems to have had heterozygous flies but
did not, according to his account, obtain any black flies
from them. However, if the ultra-violet light is a specific
agent for these mutations the experiment can easily be
repeated.

8 Science, XXII, 1910.

^Bull Scientifique, XLVm, 1914.

Digitized by

Google

No. 676] MUTATIONS IN DB080PHILA 711

It should be added that only one of Guyenot's two
lines gave dark flies after treatment with ultra-violet light.
This might seem to indicate that the first result was acci-
dental, or due to the presence of a recessive mutation in
the stock prior to treatment were it not that a careful
control is recorded. Guyenot himself speaks with much
caution concerning the interpretation of his results. De-
cision as to their value may be reserved until repetition of
the experiment gives confirmation. Our own experience
with Drosophila shows that mutations appear under con-
ditions where all the other flies in the same culture are
normal and we have become unduly sceptical perhaps
towards evidence which refers a particular mutabt to some
unusual treatment to which the flies have been subjected.
Until we can get definite information as to how mutants
arise, whether through external influences, through acci-
dents of mitosis, through hybridizing, or through changes
in the chromosomes with its consequent dislocations of
the machinery of crossing over, or in some other way, it
seems futile to discuss the question.

Digitized by

Google

THE ANALYSIS OF A CASE OF CONTINUOUS

VARIATION IN DROSOPHILA BY A STUDY

OF ITS LINKAGE RELATIONS

PROFESSOR JOHN S. DEXTER,

Olivet Collbqb
I. Introduction.
II. (The Germinal Constitution of Beaded Flies.

A, Crosses between Beaded and Normal WOd Flies.

1. Behavior in First Generation.

2. Behavior in Second Generation.

3. Behavior in Third and Fourth Generations.

B, Crosses between Beaded Flies and Other Mutants.

1. The Fi Generation.

2. Linkage Belations.

(a) Sex Linkage.

(&) Linkage to Sex-linked Genes.

(d) Linkage to Third Chromosome Genes.

III. The Effect of Environmental Conditions on the Development of Beaded
Wings.

A. General Statement.

B. The Effects of Belative Moisture.

C. The Effects of Covering with Paraffine the Mouth of the Bottle

in which the Flies are Developing.

D. The Effects of Acidity and Alkalinity.

E. The Effect of Belative Temperature.

F. The Effect of Darkness.

rv. The Effect of Selection on the Production of Different Types of Bead-

edness.
v. Mutation in Beaded Stock.

A, General Statement.

B, Perfect Notched Wings.
C Spread Wings.

J). Stumpy Wings.
VI. Summary.

L INTRODUCTION

Haedly had the principles of Mendelism been worked
out in one species of plant than apparent exceptions to
these principles were discovered. MendePs own case of
the breeding true of species hybrids in Hieracium was the
first of these, and since 1900 others have be^n reported.

712

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 713

Further analysis has shown that many of these early
cases are readily interpreted on Mendelian principles,
while for other exceptions, like that of Hieracium, for
instance, the true explanation has been found without in
any way coming into conflict with Mendelism.

The masking of a Mendelian ratio may be effected in
many ways, and some of the most important of the recent
work in genetics has dealt with this problem. Among
the conditions so far brought to light may be mentioned
the following :

(a) Multiple Factors. — ^Recent papers by MacDowell
(1914) and Shull (1914) have discussed at length the
literature and history of this subject. In brief, the work
that has been done shows that in both animals and plants
the production of certain characters is brought about
through the action of two or more independently Men-
delizing pairs of genes that have similar effects on the
developing organism. If the effect of these genes is
cumulative, so that the character is more or less produced
according to the number of dominant genes present, the
type of inheritance known as blended inheritance is
produced. If the effect is not cumulative, the recessive
character does not appear with the frequency of 1 : 3, but
with the frequency of 1 : 15, 1 : 63, etc., according to the
number of pairs of genes concerned.

(&) The Effect of the Environment. — ^A typical case
of this sort is reported by Baur (1912). In crossing a
dark red to a red strain of Antirrhinum, a complete series
between the red and the dark red appeared in the Fg gen-
eration; the effect of light on the plants was such that
plants that had developed in a bright light had a darker
color than those that had developed in a less intense light.
The analysis of the Fg generation, however, proved con-
clusively that one fourth of the Fg plants had been homo-
zygous dark reds, one fourth had been homozygous red,
and two fourths had been heterozygotes. Morgan (1912a)
has described a case in Drosophila in which moisture
conditions in the bottle in which the flies are developing
determine to a certain extent whether or not certain

Digitized by

Google

714

THE AMERICAN NATURALIST [Vol. XLVDI

characters shall appear; and Hoge (1914) has shown
that certain temperatures are necessary for the develop-
ment of reduplicated legs in Drosophila. Other examples
may be found in the literature of genetics.

(c) Lethal Characters. — There have been reported sev-
eral instances in recent years of animals and plants which
are unable to live if homozygous for certain genes. The

Digitized by

Google

No. 576]

VARIATION IN DROSOPHILA

716

case of yellow mice, Baur's ^wrea-strain of Antirrhinum
(Banr, 1912) and the modified sex-ratios in Drosophila
reported by Morgan (1912eZ) are examples of tMs phe-
nomenon.

The object of the present paper is to describe a case of
inheritance in Drosophila that for some years seemed to

Digitized by

Google

7 1 6 THE AMERICAN NATURALIST [Vol. XLVni

defy Mendelian analysis. Though all the details of the
ease have not been worked out, enough has been done to
show that it is brought about by factors which segregate
in the ordinary Mendelian fashion, and that the diffi-
culties which it still presents are not opposed to that
hypothesis.

The case under consideration is that of Beaded wings,
which, according to Morgan (1911a), first appeared in
May, 1910, among flies that had been exposed during part
of their early life to radium rays.

The appearance of these wings can best be understood
from the figures (Figs. 1-12), which represent a few of
the forms that may appear in a stock culture. All grada-
tions may be found between wings perfectly normal and
mere strips, such as shown by Figure 11.

In the early days of its history, according to Morgan,
the Beaded- winged flies did not breed true, but for many
generations produced many normal-winged offspring.
At the time when I took up the experiment, however, the
stock bred almost 100 per cent, pure ; that is, almost every
fly hatched had wings more or less Beaded. I have at
present a strain which breeds true, throwing only Beaded-
winged offspring, and most of the offspring have the
Beading in an extreme form. Most of my work has been
done with this stock.

II. THE GERMINAL CONSTITUTION OP BEADED PLIES

A. Crosses between Beaded and Wild Flies

1. Behavior in First Generation

When a Beaded fly is mated to a normal fly of a normal
Wild stock, a considerable number of flies with Beaded
wings usually appears in the first generation (F^). The
percentage is not constant, but varies between zero and
about fifty per cent. (See Table I.) From Chart 1, it
appears possible that the average percentage of Beaded-
winged offspring per pair is near 10-15 per cent, or else
near 30-35 per cent, of the total offspring. The exact

Digitized by

Google

No. 576]

VARIATION IN DROSOPUILA

717

average is 25.5 per cent. That there is a bimodal curve
produced may perhaps not be significant, as will appear

TABLE I

Crosses op Beaded to Wild Flies, showing Peecentaqes of Beaded-
winged Offspring

Per Cent, of Flies with Beaded Wings

^ «» ^. *. «». «» ^. ^' «i «*.

2;-A.i|i* i s i ^ I I

Number of broods giving this per-
oentage

1 5

fi 7 7 14 a ' 2 2

Average aixe of brood

203 172! 148 120180 113 101 96 130 60

from the following facts, although later evidence will
show that it very possibly is significant.

The per cent, of Beaded-winged offspring given by one
pair (Beaded X Wild) may vary at different times and

o o>

•^^

Chart I

Numbers of broods giving certain percentages of Beaded-winged off-
spring in Fj generation of Beaded X Wild. (See Table I).

under different conditions. For instance, if a pair are
put into a bottle with food and are left there for ten days,
and are then put into another bottle with fresh food and
left another ten days, the percentage of Beaded-winged
offspring will be different in the two broods. Table II
gives the records of such tests. The first two were made

Digitized by

Google

718

THE AMERICAN NATURALIST [Vol. XLVni

with single pairs. In the third case, a Beaded male was
given four virgin females, so that although all the off-

TABLE II

DnTEBENT Percentages of Beaded-winged Oitsp&ing by the Saks

Paeents During Two Separate Tbn-dat Periods

IN DipPBRENT Bottles

First Ten Days

No. of Oir-
spring

Per Cent.
Beaded

Second Ten Days

No. of Off-
spring

Per Cent.
Beaded

First pair 126 I 21

Second pair 117 7

One father X4 mothers.. 389 I 20

179
146
301

41
22

spring have the same father, they come from four
mothers. Inspection of this table shows that it is quite
impossible to assign the parents of any one brood to any
definite class based on the percentage of Beaded-winged
offspring that they give.

Table II shows also that the parents gave a larger per-
centage of Beaded-winged offspring during the second
ten days than during the first period. That this is a
coincidence appears from Table HI. Here it is shown
from the records of fifty broods chosen at random, that

TABLE III

Percentages op Beaded-winged Flies in the First Count op a Brood

Compared with Those op the Last CJount (Interval op

PROM Eight to Ten Days). Based on CJounts prom

FiPTY Broods, Chosen at Bandom

First

Last

First

Count

45 1

37 1

21 1

Last

First

Last

First

Last

First

ILt

Count

Larger percentage of Beaded-winged offspring the first count, 44 broods.
Larger percentage of Beaded-winged offspring the last count, 5 broods.
Equal percentage of Beaded- winged offspring both counts, 1 brood.

Digitized by

Google

No. 576] VARIATION IN BROS OPE ILA 719

the counts, made in the first few days after the flies of
any brood begin to hatch show almost invariably a very
much larger percentage of Beaded-winged offspring than
do the last counts made. This fact will be considered at
some length in the section on environmental effects.

Enough has been said, at least, to show that, whether
the results here described are genetic or environmental
effects, the F^ generation is remarkably inconstant with
reference to the percentage of Beaded-winged offspring
that appear. It is evident that this percentage can be
readily altered by (1) changing the length of the period

Chaet 2.
Bdcf XWad9

daughters, 33% Bd sodb, 13% Bd

Bd cf X Wfld 9

daughters, 16% Bd sons, 3% Bd

Bd cf X Wild 9

daughters, 23% Bd sons. 9% Bd

during which the brood is allowed to run; (2) by chang-
ing the parents from one bottle to another. Extensive
studies of environmental effects have shown other ways
in which the percentages can be altered, but of this we
will treat later.

2. Behavior in the Second Generation

The question at once arises whether the Beaded and
normal Fj flies are alike genetically. To the solution of
this problem two different breeding tests were applied:
viz., matings of Fj normal by normal, normal by Beaded,
and Beaded by Beaded ; and back crosses of both normal
and Beaded to Wild stock. The results of these tests are
given in Tables IV and V. These tables show that when

Digitized by

Google

720

THE AMEBIC AN NATURALIST [Vol. XLVHI

Beaded- winged flies of the Fj generation are used as par-
ents, more Beaded-winged young are produced than when
normal- winged Fj flies are used. This holds true for each

TABLE IV

Matings Betwbbn Fi Plies op the Cross Beaded by Wild, Showing
Percentages or Beaded Offspbing in Individual Broods

Per Cent, of Flies Beaded

0-4.9
5-9.9
10-14.9

(7> 1 o» a»

^ S ^

30-34.9
35-39.9
40-44.9

45-49.9
50-54.9

6.V50.9
60-61.9

Normal X Normal

.-.. 8 2 ! 2
. . . 1 . . .

...... 1

... 3 1

1 ...

Normal XBeaded

... 2
1 2

Beaded XBeaded

i . . .

1 1

TABLE V

Back-crosses to Wild op F^ Flies of the Cross Beaded X Wild, Show-
ing Percentages op Beaded Offspring in Indiyidual Broods

Percentage of Ofl^pring Beaded

t i

I I' I' I

Normal XWild 9

Beaded X Wild 3

6 3

of the five crosses shown in the two tables. Normal-
winged Fi flies do, however, have some Beaded-winged off-
spring, both when mated among themselves, and also,
though more rarely, when back crossed to Wild.

These Fg and back-cross results give little satisfaction
at first sight to the student of Mendelism. If we suppose
that there is one gene on which the Beaded condition
depends, and that it is partially dominant, then Beaded

TABLE VI

Beaded and Normal Offspring by Sexes When One Parent is Beadkd
AND the Other Wild

Beaded
9 9

Normal
9 9

Beaded

Normal

Per Cent.
9 9 Bd.

Per Cent.
cfcf Bd.

Father Beaded..
Mother Beaded .

1.246 .
894 ;

4.488
2.959

948
1,139

4,481
2,684

21.7
23.2

17.6
29.8

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 721

and normal F^ flies should give the same results when
used as parents. Or if we were dealing here with a case
like the * 'yellow mouse '* case, in which homozygous
yellows do not exist: that is, if homozygous ^'Beadeds'^
do not exist, then one quarter of the flies produced by two
Beaded parents from the stock should be normal. But as
was said before, the stock breeds true, every fly produced
having Beaded wings.

It may be noted that a pair of Fj normal flies usuallly
produce less than 10 per cent, of Beaded offspring. If
these normal flies carried a recessive gene for Beaded-
ness, they should produce twenty-five per cent. Beaded
offspring. The Beaded Fj offspring, on the other hand,
though they produced in sdl cases more than twenty-five
per cent., did not produce 75 per cent. Beaded offspring,
as they should have done if a single dominant gene for
Beaded wings were heterozygous in them.

3. Behavior in Third and Fourth Generations

Beaded offspring, that appeared in the Fj generation
of the cross Beaded X Wild, were back crossed to Wild
The process was again repeated with the Beaded off-
spring that appeared, till four generations had been pro-
duced. The results of this test are given in Tables VII
and VIII and in Chart 4.

A striking result is that an Fj Beaded fly or even a fly
of later generations heterozygous for Beaded wings some-

TABLE VII

Bepeated Back-crosses of B£ADEI^WINGED Flies from the Cross Beaded
BY Wild to Wild Stock to Show Percentages of Beaded-
winged Offspring, See Chart IV

Digitized by

Google

722 TBE AMERICAN NATURALIST [Vol. XLVHI

TABLE VIII

Normal Females from Family 2, Generation 2, Back-crossed to Wild
Males, Showing Percentages op Beaded-winged Oitspring

Beaded

Normal

^Bd

Type X

2
50

1.040
342

.02

TypeY

16.3

times has as large a percentage of Beaded-winged oflf-
spring when mated to Wild, as does a fly direct from pure
Beaded stock when mated to Wild, though a comparison
of Tables I and V shows that this is not the usual occur-
rence. This suggests at once the action of a lethal gene
(Morgan, 1912&). Morgan has shown that in a certain
stock of Drosophila there are twice as many females as
males in the offspring of one half the females. No matter
to what male such a female be mated, her daughters are
twice as numerous as her sons, and one half of her
daughters also repeat this phenomenon, and one half of
the daughters of these again. This fact finds its explana-
tion in the assumption that there is in one of the sex-
chromosomes of such females a gene which prevents the
development of any male which gets it.

Now if such a gene had the power of expressing itself
as a dominant in those flies that carried it in the hetero-
zygous condition, if, for example, it caused the wings to be
Beaded, it would be possible to select such flies at sight,
and these flies could then be depended upon to repeat the
phenomenon. (Morgan accomplishes the same end by
mating such flies to mutants carrying a gene with which
the lethal gene shows close linkage, such as that for white-
eyes. He then finds that the red-eyed females carry the
lethal gene, unless, as rarely happens, a ** cross-over'* has
occurred.)

Such a sex-linked lethal gene producing a dominant
wing character has actually been found to occur in the
case of a mutant which arose in the Beaded stock, and
which will be discussed later. For the present we must
note that if the lethal gene were not associated with sex,
its presence could be detected by the absence of certain

Digitized by

Google

No. 576] VABIATION IN DROSOPHILA 723

expected ratios, or classes, or in some other peculiarity of
genetic behavior. In the case before us, we found that
the Fj generation consisted of at least two types; viz.,
Beaded and not-Beaded flies. These were shown to differ
genetically. To obtain such a result must mean that at
least one of the parents was heterozygous in at least one
gene. This result is however a fairly constant one ; and
by virtue of the long-continued inbreeding of the Beaded
stock this heterozygosity must surely have been weeded
out before now if there were no serious hindrance to
homozygosity. The classic example of this sort of effect
is that of the yellow mice.

But the development of Beaded wings can not be
brought about by the action of a single lethal gene, for if
this were true it would be impossible to obtain a stock
of Beaded flies that would breed true, and yet such a
stock, as has already been said, is the one from which
these very crosses derive their Beaded ancestors. There
must therefore be at least one pair of allelomorphs of
which one member is effective in producing Beaded wings,
and can exist in the homozygous condition and possibly
also another pair of allelomorphs of which one member is
a recessive lethal gene. We can explain many of the
facts so far obtained on the supposition, that there are
these two independently Mendelizing pairs of allelo-
morphs concerned in the production of Beaded wings.
The pair containing the lethal gene we will call L (nor-
mal) and 1 (lethal) ; and the other pair B' (Beaded) and
b' (normal). The occurrence of the two genes B' and 1
in one individual usually causes such an individual to
have Beaded wings, though Beaded-winged flies also
occur which do not carry the lethal gene, but are homo-
zygous for B^.

It should be possible then to isolate a stock of Beaded-
winged flies not carrying this lethal factor, 1. Such flies
should give a much smaller percentage of Beaded- winged
offspring in the Fj generation of a cross with Wild stock
(or perhaps none at all, if B' were recessive), than would
those flies carrying IL. Such a stock has not yet been ob-

Digitized by

Google

724 THE AMERICAN NATURALIST [VouXLVm

tained, but occasionally a strain of Beaded flies is met
with that gives only low percentages of Beaded-winged
offspring. See, for instance, Family 4, Table VII. Pos-
sibly such a stock would not be recognized at once, espe-
cially if it were so affected by environmental conditions
that even flies homozygous for the factor B'B^ sometimes
had normal wings. Normal-winged flies, as will be
pointed out in a later section of this paper, do very fre-
quently appear in Beaded stock, but these flies when
mated to each other appear to throw as many Beaded-
winged offspring as do the Beaded-winged flies of the
stock, and often 100 per cent, of their offspring have
Beaded wings.

In this connection it will be of interest to recall that
Chart 1, and Table I gave results that might be inter-
preted as evidence of the bimodal curve that should be
expected if the above hypothesis is correct.

Normal females from the second generation of Family
2 were also back-crossed to Wild males. The results are
given in Table VIII. Most of these normal females gave
very few or no Beaded offspring (Type X) while two of
them gave a considerable number of Beaded offspring
(Type Y). The explanation here is perhaps that the type
Y females were genetically like most of the Beaded
females of an Fj generation (on our hypothesis, B' L b' 1)
while the females of Type X were genetically lacking in
the factors that are usually present in Beaded Fj flies
(i. e., they were B' L b' L). That such an occurrence is
not infrequent in Drosophila is seen in Table IV in which
three broods out of fifteen raised from normal Fi flies
gave 25 per cent, or more of Beaded offspring though the
other twelve broods gave less than fifteen per cent., and
eight broods less than five per cent, of Beaded offspring.
It seems certain therefore that there are two types of
normal-winged offspring in the Fj generation of the cross,
Beaded by Wild; one of these is genetically like the
Beaded flies of the same generation and the other is
genetically different from its Beaded brothers and sisters.

Types X and Y have been found to occur in all of the

Digitized by

Google

No. 576] VARIATION IN DBOSOPHILA 725

tests made of Fi flies whether of matings to Wild stock
or of matings to other mutants such as Black, Pink, Arc,
Ebony, etc. Table XXVI shows these two types as they
appeared in back crosses to normal Pink males of normal
and Beaded females of the cross Pink Beaded by Wild.
Here it was found that more of the normal than. of the
Beaded Fj flies were of Type X, and conversely that more
of the Beaded than of the normals were of Type Y.

It has not been possible to distinguish with certainty
between these two types even by their offspring because
of the large amount of fluctuation that occurs in the per-
centages of Beaded offspring. For example it would be
difficult to say whether a fly giving five per cent, of its
offspring Beaded would belong to Type X or Type Y.

It would be expected that Type Y would be given by
those flies that carried both factors for Beaded, and
Type X by those that lack the lethal factor, and it wiD
be seen later that on the whole the evidence supports
this view.

B. Crosses between Beaded Flies and Other Mutants

1. The Fj Generation

If we examine the Fj generation when Beaded flies are
crossed to other mutants, i. e., to flies of a stock that is
perfectly normal so far as Beadedness is concerned, but
which is unlike the normal Wild flies in some other wing
character, or in eye color or body color, etc., we find an
even greater amount of variability in the percentage of
Beaded-winged offspring than in the Fj generation of
Beaded by Wild. (See Tables I and IX; also Charts
1 and 3.)

The details may be gathered from Table IX, where it
can be seen that there is a certain specificity in the per-
centage of Beaded offspring that appear in any specific
mating.

For instance, it appears that more of the offspring
have Beaded wings if a cross is made with Vermilion-
eyed flies than when Beadeds are mated to Pink-eyed

Digitized by

Google

726

THE AMEBIC AN NATURALIST [Vol. XLVHI

TABLE IX

The Peecentaoes or Beaded-winged Flies in the Fx Generation of
Crosses Between Beaded Flies and Other Mutants

Percentages

Mutant Involved

Oil a*

III

9 0> Ok O)

i?i

t>> io ' 35 1 cii cS

Vermilion Yellow 1

Vermilion

Miniature Eosin

White

Pink Beaded XWhite ... 2

Pink Beaded X Wild 4

Pink I..

Pink Ebony |..

Sepia

Pink Black ' 1

Black ll

Purple Curved ' . .

Truncate

Vestigial

Antlered ' . ; . .

Strap I . . I

1 1 .... 1

.. .. 1 4 ..

.. 1 ......

14 3 11

2 .... 1 ..
1

Total 8 6;5

1 6|6 8672.11.

. 1 ..

..,.. 1 ,..

BeadedX Wild (Table I) . ! 1 5 lo' 6 ! 7

7 14 3 2 2 2 ! . . .

..'..I.. ..

Grand Total 9 11 15

' 1 1

12 13 22 8 9 4 2 11.

. 1 ..

..;..ll ..
1 i ;

Chart III

oo^^Si ^ ^ ^ ^ «;^ &^S5;^^ ^t?^^

Distribution of broods giving certain percentages of Beaded-winged off-
spring in Fj generation of Beaded X Normal (other Mutants or Wild).

(See Table IX.)

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 727

flies, or more in the crosses with White-eyed flies than
in those with Black body color. (In every case, where
the contrary is not stated the flies are normal in other
respects than the one named, e. g., White-eyed flies in
these crosses have Gray bodies and Long normal wings.)

No explanation of this specificity by the assumption
of a segregation of factors in the germ cells appears to
be available here, though such a possibility has not yet
been ruled out, or can be ruled out till certain other phe-
nomena are understood. Tha easiest way of ** explain-
ing" it is that the dominance of the genes for Beadedness
varies in accordance with many other circumstances,
among which are differences in the other genes present,
such as those for Vermilion, White or Pink. Such an
assumption as this, as will appear later, would seem to be
fully in accord with the behavior of the genes for Beaded
wings when in still different relationships.

It is assumed, then, for example, that the percentages
of Beaded-winged flies in the Fj generation of a cross
between Beaded and White are higher than those in the
Fj generation of a cross between Beaded and Black, be-
cause the gene for Black is relatively to the gene for
White eyes an inhibitor of Beadedness. It would appear
as though a series might be made of the mutants of
Drosophila beginning with those in which the genes for
Beaded wings are most dominant and ending with those
in which the Beaded genes are recessive. In order to
construct such a series a large number of pairs would
have to be made for each cross in order to determine the
limits of variability of Beadedness for the cross con-
cerned. The work would probably be greater than the
value of the results obtained, and therefore the attempt
has not been made to carry out this test. From what has
. been done incidentally in carrying out other experiments,
it will be seen that in general the darker eye colors and
body colors are associated with a low percentage of
Beadedness in the Fj generation, and the brighter colors
with a higher percentage. This may, however, only be a
coincidence.

Digitized by

Google

728 THE AMERICAN NATURALIST [Vol. XLVIIl

2. Linkage Relations
(a) Sex Linkage

If in the crosses thus far described the sex of parents
and offspring that show Beaded wings be considered, it
may appear at first as though we may be dealing with a
partially sex-linked gene. For it very frequently happens
that when the mother is Beaded, and the father is normal
(either of Wild stock or of some mutant stock not carry-
ing Beadedness), more of .the sons than of the daughters
are Beaded. For example, in one such brood, there
were 17 Beaded to 128 normal females, and 5 Beaded
to 130 normal males, or 12 per cent, of the females
and 3.5 per cent, of the males. Both of these examples
were deliberately chosen because they were good ex-
amples of the phenomenon described. It would be possi-
ble to select from my records several examples of the
reverse phenomenon, where Beaded females had more
Beaded daughters than Beaded sons, and where Beaded
males had more Beaded sons than Beaded daughters.
Nevertheless, the records of all broods available have
given the numbers shown in Table VI, where it appears
that more sons are Beaded when only the mother is
Beaded and more daughters when only the father is
Beaded.

It may perhaps be significant, on the other hand, that
when the mother is Beaded a slightly larger percentage
of her daughters is Beaded than of the daughters of a
Beaded male, while a very much larger percentage of her
sons is Beaded than the sons of a Beaded male. In other
words, it seems that the daughters are affected to ap-
proximately the same extent, whether they get their
Beadedness from father or mother, while the sons are
affected also by the mother, whether or not she carries
Beadedness. This might mean that there is some gene in
the sex chromosome that does not show except when other
Beaded factors are present. That this is not the case
will appear from Chart 2, which records three generations
of flies in each of which the mother was normal (Wild)

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 729

and the father Beaded. This shows that although the
father transmitted his Beadedness more to his daughters
than to his sons, yet his Beaded sons also had the capac-
ity to affect their daughters more than their sons, and
these sons again repeated the phenomenon. Yet these
males could not have received their X-chromosome from
their father, unless non-disjunction (see Bridges, '13)
had occurred. In fact, to produce the results here given
non-disjunction must occur in one half the females of
the Wild stock. Frequent tests with the Wild stock by
practically all of the students in the laboratory make it
certain that this is not the case. I also tested a consider-
able number of the females by mating them to sex-linked
mutants and found no non-disjunction.

This apparent sex-linkage that does not follow the
** ordinary rules" of sex-linkage must be left for the time
being as one of the still unsolved problems. The only
possibility of explanation that occurs to me is that the
above-described effect would be produced if in the cyto-
plasm of the egg of the Beaded female something were
present which is absent in the egg of the normal female,
and to which the males are more responsive in their
development than are the females. This suggestion has
not a particle of cytological evidence to support it. Mor-
gan (1912(i) has suggested that the influence of cytoplasm
may cause certain peculiar results obtained in crosses
between Miniature- winged and Rudimentary-winged flies.

(6) Linkage to Sex-linked Genes

Matings of Beaded flies to flies with sex-linked char-
caters, including Vermilion and Vermilion-yellow, have
been made and the Fg generation raised. No sign of
linkage was observed. The Fg figures are given in Tables
X and XI. These cases definitely establish that there is
no gene for Beaded wings in the X-chromosome.

Although no sex-linked gene for Beaded wings are
known, there has arisen in the Beaded stock by mutation
a fly with notched wings (Fig. 13) that proved to be

Digitized by

Google

730

TRE AMEBIC AN NATURALIST [Vol. XLVHI

TABLE X
P, Counts peom the Ceoss Vermilion J X Beaded <?

V.B(L9
39

V.Bd.cf
26

V.N. 9
93

V.N.cf
77

RedBd.9
33

R-Bd.cf
22

R.N.9
81

Bd.V:NV.«66 : 170

Bd.Red : N.Red - 66 : 141 . . .
Bd.Total : N.Total - 120 : 311

%Bd.

27.6
28.0
27.8

Expected No. of Bd. if

no coupling ezista

65.4

54.6

(120)

TABLE XI
Pa Counts from the Cross Vermiluon Yellow X Beaded

^eaded

Normal

Beaded

Normal

Beaded

Normal

Beaded

Gray

YeUow

Red

Vermil.

Red

Vermil.

165

227

Normal
YeUow
Venml.

% Beaded

Bd.R. : N.R. - 206 : 261 44.1

Bd.V. : N.V. - 85 : 122 41.0

Bd.G. : N.G. - 200: 261 ' 43.4

Bd.Y. : N.Y. - 91 : 122 42.7

Bd.Total : N.Total - 291 : 383 43.2

Expected No. of Bd. if

no coupUng exists

201.7

89.4

199.1

92.0

(291.0)

caused by a dommant sex-linked
factor lethal when homozygous.
(See page 754.) It will be dis-
cussed under the name ** Perfect
Notched'* and its peculiarities
described in the section on *' Mu-
tation in Beaded Stock.*'

For the reasons given, it seems
certain that there is in the group
of sex-linked genes no gene con-
cerned in the production of Beaded
wings. We shall later bring for-
ward evidence to show that there
is such a gene in the third chromo-
some group. (Sturtevant, 1913.)
The crosses made with flies showing characters whose

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 731

genes are in the second chromosome are still perplexing;
for while the second chromosome exerts an influence on
the statistical results, as will be evident from the figures
to be presented, the nature of this influence is not fully
determined.

The second chromosome characters with which tests
have been made are the wing characters, Arc, Curved,
Vestigial, Antlered, and Strap,^ the body color. Black;
and the eye color. Purple.

Vestigial, Antlered, and Strap stand for wing char-
acters of such a nature that it is not possible to distin-
guish Beaded-winged individuals if any of these other
characters are also present. They are therefore of no
use for determining whether or not there is a second
chromosome gene for Beaded wings. The crosses be-
tween Beaded flies and flies with these characters do not
especially interest us here. It may be said in passing,
however, that in every case in the Fj generation between
these flies and Beaded flies, from 60 to 90 per cent, of the
offspring had non-normal wings, and the author was put
to serious straits to classify the new wing types that
appeared. These were similar in all the crosses, however,
and on the whole resembled Beaded wings.

In the Fg generation, and in back crosses to Beaded
Stock and to Vestigial Stock further complications arose
with more new types of wings, including a new ** muta-
tion" which bred true from the start, and which will be
discussed briefly and described under the name ** Spread'^
in the section that concerns mutation.

We may now return to the crosses between Beaded flies
and Black, or Purple, or Arc, or Curved. These crosses
give results that can be used for the study of linkage,
and they present in common a number of distinguishing

1 strap Wings is a mutant much resembling Extreme Beaded in appear-
ance but its mode of inheritance has not yet been worked out. It may be
that it actually is Beaded plus some at present unknown gene. Beadedness
i0 suspected to occur also in Vestigial and Antlered stock. This may very
likely be true since Strap and Antlered arose in Vestigial, and Vestigial in
Beaded.

Digitized by

Google

732

THE AMERICAN NATURALIST [Vol. XLVHI

characteristics. Tables XII to XX give the results in
systematic form.

TABLE XII
P, Counts op the Cross Beaded ? X Curved <J

Beaded
Curved

Normal
Curved

Beaded
Straight

Normal
Straight

Exp. Bd.
Curved

Kxp. Bd.
Straight

Type 1

61
. 25

16
18

246
106

4
3.6

Type 2

16.4

One of the most striking characteristics of these second
chromosome crosses is that the Fj flies fall into two
classes or possibly into three classes with reference to the
offspring that they produce. These classes I have called
Type 1, Type 2, and Type 3.

In Type 1 there is no linkage between Beaded wings
and the second chromosome character, but Beaded-
winged flies occur with equal frequency in all classes of

TABLE XIII
Pa Counts op the Cross Beaded J X Arc <$

Beaded
Arc

Normal
Arc

Beaded
Straight

Normal
Straight

Exp. Bd.
Arc

i Exp. Bd.
' Straight

Typel

Type 2

39
41

200
152

184
330

970
452

38.+
73

185
298

offspring. In Type 2 there is linkage of Beaded wings
with the second chromosome characters, so that the
Beaded wings appear more frequently in flies showing*
the characters of the Beaded parent. In Type 3, which
occurs only a very few times and is not very marked
except in Table XX, Beaded wings appear to a greater
percentage in the offspring whose other characters are
not those of the Beaded parent. (/. e., * ' repulsion * ' occurs
between the factor for Beadedness and that for the
second chromosome character with which it entered the
cross.) I do not wish to emphasize Type 3, but concern-
ing the other two it is important to note that about one
half of the Fj flies seem to be of Type 1 and one half of
Type 2.
In Type 2 in the cases here adduced the linkage is

Digitized by

Google

No. 576]

VARIATION IN DROSOPHILA

733

strongest with the character Arc and weakest with Black.
If there is a gene in the second chromosome which aids in
producing Beaded wings, it seems probable that it is
located nearer to Arc than to Black and on the side of
Arc away from Black.

TABLE XrV

Baoe-ckosses of Ft Beaded-winged Males or the Cross Beaded J X
Purple Curved J to Purple Curved Females op Normal Stock

Beaded Pr.
Curved

Normal Pr.
Curved

Beaded Red Normal Red
Straight 1 Straight

Exp. Nor.
Bd. Pr. Cv.

Exp. Nor.
Bd. R, Strt.

Type 1

15
2

125
29

15 154
28 27

13.6
10.8

16.4

Jype2

19.2

It remains to consider Type 1, and to find the reason
for the existence in the Fj generation of flies whose off-
spring show no linkage between Beadedness and second
chromosome characters, and in the same brood, flies whose
offspring do show such linkage. The most obvious ** ex-
planation *^ would be, of course, that the factor in the

TABLE XV

Back-cbosses of Fi Beaded-winobd Females or the Cross Beaded J X
Purple Curved J to Purple Curved Males op Normal Stock
Brood 2 is of Type 3 for Purple and of Type 1 for Curved.

Bd. Pr.
Cv.

N. Pr.

cv.

B. Pr.

St.

. N. Pr.
St.

Bd.Red
Cv.

N. R.
Cv.

' Bd. R.

1 St.

N. R.

St.

Brood 1

Brood 2

...... 10

49
47

1
4

22
_6

1
0

17
12

16
1 14

61
71

second chromosome was a *' lethal" such as the factor 1,
described in an earlier section of this paper. If this were
the case, there should be some flies in the Beaded stock
homozygous for L, the normal allelomorph of this gene,
and also for B', i, e., B'LB'L. These flies should have
fewer Beaded offspring than those heterozygous for 1,
and none of these Fj offspring should give linkage with
second chromosome characters. As a matter of fact, in
the Fg results given in Table XVI for Beaded by Black,
no linkage was observed; but this case is not good evi-
dence, for it was made in the first attempts to solve the
problem of Beaded wings, and I had not yet learned the

Digitized by

Google

784

THE AMERICAN NATURALIST [Vol. XLVHI

value of Fj counts, matings in pairs, and back-crosses to
normal. It stands however as the only evidence of ite
sort that I can give at present.

TABLE XVI
P, Counts op the Cross Beaded ? X Black c?

Beaded
BUck

Normal
Black

Beaded
Gray

Normal
Graj

- Expected
No. Bd. Bl.

Expeeted

NoVBd.

Gny

Typel j 19

375

1602

Type 3 is not easy to explain. There are no known
cases of this sort elsewhere in Drosophila and I prefer
not to attempt to answer this question at present.

TABLE XVII

Back-ceosses op Pi Males op the
Females op

Cross Beaded <? X Black J
Normal Stock

TO Black

Beaded
Black

Normal
Black

Beaded
Gray

Normal
Gray

Expected
No. Bd. Bl.

Expected

NoVBd.

Gray

Type 1

7
5

162

110

9
40

187
97

7.3
20.5 •

8.6

Type 2

24.5

In general, it may be noted that technical diflSculties
have disturbed the crosses with second chromosome char-
acters. The wing character Arc is not always easy to
recognize, as it is very often nearly normal in appearance.
On the other hand, the flies with Curved wings, though
always distinct, occasionally get ** stuck up*' with the
food and in their bedraggled condition it can not always
be determined whether or not the wings are Beaded as
well as Curved. I was at first inclined to attribute the
apparent coupling (which was discovered for Arc and for
Curved before it was discovered for Black) to errors
made in the counts.

As for Black, the Fj and later generations give a much
lower percentage of Beaded offspring than do most other
crosses, and this necessitates raising large numbers of
offspring. The results are, however, trustworthy when
obtained.

Digitized by

Google

No. 576]

VARIATION IN DROSOPHILA

735

The crosses with Purple-eyed flies presented no diffi-
culties but ran smoothly aside from the fact that the
purple-eyed flies had Curved wings, and as remarked
above, Curved wings sometimes get bedraggled.

TABLE xvm

Back-ceossbs op Fi Temales op the Cross Beaded <J X Black ? ro
Black Males op Normal Stock

Beaded
Black

Normal
Black

Beaded
Qraj

Normal
Graj

Expected
No. Bd. Bl.

Expected

No. Bd.

Graj

Type 1.
'IVpe2.

78
457

2
28

91
493

3.3
15.5

3.7
17.5

(d) Linkage to Third Chromosome Genes

We have said tentatively that there was perhaps a non-
sex-linked lethal gene for Beaded wings in the second
chromosome, and that possibly the cytoplasm carried by
the egg disposes males toward or away from Beadedness
according to whether the fly that bore the egg was or was
not Beaded. These relations are not securely deter-
mined, and the data are still incomplete. The relation of
Beaded wings to characters whose genes are in the third
chromosome is much clearer. All crosses that bear on
this problem point to one fact, namely, that there is a
gene for the production of Beaded wings in the third
chromosome, and that this gene is very closely linked to
Ebony, and very loosely linked to Pink. Tests have been
made between Beaded and the third chromosome char-
acters. Maroon, Sepia, and Pink eyes and Ebony body
color.

TABLE XIX

F, Counts op the Cross Pink Beaded <J X Black J

Type 1 (?)

Pink
Bd.BI.

Pink
N.Bl.

Pink Bd.
Graj

PlnkN.
Gray

RedBd.
Black

RedN.
Black

RedBd.
Graj

RedN.
Gray

278

Expected No. Bd. if no

1 coupling occurs

Bd.Pink : N.Pink = 22 : 82 j 6

Bd.Red : N.Red - 3 : 323 ' 19

Bd.Black : N.Black - 2 : 70 4

Bd.Gray : N.Gray = 23 : 336 21

Bd.Total : N.Total - 26 : 406 (25)

Digitized by

Google

736

TRE AMERICAN NATURALIST [Vol. XLVm

In the cross of Beaded by Maroon-eyed flies, 1,369 flies
were raised in the F^ generation. Fifty-seven of these
flies had Beaded wings ; only one of the Beaded-winged
flies had Maroon eyes, while fifty-six were red-eyed, ( See
Table XXI.)

TABLE XX.

Back-crosses of F^ Females of the Cboss Pink Beaded ^ X Black $
TO Pink Black Males fbom Normal Stock

PinkBd.

Pink N.

Pink Bd.

Pink N.

RedBd.

RedN.

RedBd.

' RedN.

Black

Gray

Graf

Black

Gray

Typel.....

113

122

Types

Totals . . .

166

171

170 i 4

198

Expected No. Beaded if no

coupling occurs

Bd.Pink : N.Pink - 31 : 337.

Bd.Red : N.Red » 16 : 368. .

...J 24

Bd.Black : N.Black - 29 : 33

..J 23

Bd.Gray : NOmv a 18 - .^69

Bd.Total : N

.Total -

• 47 : 70(

• • • • 1

(47)

Fi males of the cross Sepia by Beaded were back-
crossed to normal Sepia females. Inasmuch as cross-
overs probably do not occur in the male (Morgan, 1912c),
no Beaded Sepia flies should occur in the offspring of this
cross. Table XXII shows that none occurred. The num-
bers are not large, but since they are entirely in accord
with the other third chromosome results, it was not
thought worth while to increase them. That apparent
cross-overs may very rarely occur will appear possible
when we consider the results of crossing Beaded by Pink,
and the probable significance of the phenomenon will be
considered.

TABLE XXI
F, Results op the Cross Beaded J X Maroon (J

Bd. Maroon 1 N. Maroon

Bd.Red
56

N.Red

Exp.Bd.N.

Exp. Bd. R.

1 318

994

In Fg counts of the crosses involving Beaded and the
body color Ebony, totaling 4,417, in which 1,205 Beaded-
winged offspring occurred, not one had the body color
Ebony, and only eleven had Pink eyes. Repeated attempts

Digitized by

Google

No. 576]

VARIATION IN DROSOPHILA

737

to obtain Ebony flies with Beaded wings have failed. The
possibility that for some ^'inherent peculiarity'' an
Ebony fly can not have Beaded wings has suggested itself,

TABLE XXn

Back-ceosses of Fi Beaded Males op the Cross Beaded <f X Sepia ? to
Sepia Females op Normal Stock

Bd. Sepia N. Sepia

Bd.Red ' !

N.Red

0 134

» i

132

and although this would seem very improbable, it may
nevertheless be the fact. At any rate, it appears that
Beadedness either depends on genes which in the presence
of the Ebony body color are completely recessive, or that
the third chromosome gene for Beadedness, B', lies so
close to that for Ebony that cross-overs are extremely
rare even in the female.

TABLE xxin
F, Eesults op the Cross Beaded J X Ebony c?

Beaded Ebony

Normal Ebon j

Beaded Gray

Normal Gray

151

525

Very extensive experiments were carried out with Pink-
eyed flies. The important facts brought out are presented
in condensed form in Tables XIX, XX, and XXIV-XXIX.

In Table XXIV are shown the Fg counts for Beaded by
Pink Ebony. From the eleven Pink Beaded flies obtained
a new stock was derived, which was '* purified^* by a few

TABLE XXIV
Fa Eesults op the Cross Beaded $ X Pink Ebony cJ

^b.^-'N.P.Eb.

Bd. R. Eb.

N. R. Eb.

Bd.P.Gray

0 847

182

N. P. Gray! Bd. R.Gray
157 I 1,177

N. R. Gray
1.360

IBd.Ebony : N.Ebony - 0 : 1029.
'Bd.Gray : N.Gray =- 1188 : 1507.
;Bd.Pink : N.Pink - 11 : 1004. . .
iBd.Red : N.Red - 1177 : 1532.. .

Expected No. Bd. if no coup-
ling occurs
328
860
324
864

Digitized by

Google

738

THE AMERICAN NATURALIST [Vol. XLVHI

Chart IV
Family 1,
Bd9,86; 25M

Bd9,226; 26.7^

Bd9, 110; 27.3^

Bd<f,405;19?f

Bdd*. 185; 8.99(
Family 2,

Bdd<, 460; 28.9<

Bd9 Bd9 Bd9 Bdc^ Bdcf Bdc^ Bd<^ Bdrf" Bdc^ Bdc^ Bdd" Bdc^ Bd<f Bd cf

168 177 162 147 221 220 117 148 284 149 281 167 247 139
46.8^ 38.4^ 8.6^ 42.9^ 29.4^ 15^ 37.6j< 11.9^ 28.2^ 23.5f( 18.4^ 16.2^ 13^ SS.lfC

Bd(^

106
19.8^

Bdc?

27
33,3^

Bdc?
68
27f»

Family 3,

Bdd<, 690; 28.2

.r-.

Bd cf , 26
50^

Bd(^, 193
16.6^

Bd c^, 237
9.7j<

Bdcf" Bdd

Bdd' Bdcf" Bdd" Bdc

Bdd*, 190

18.4)(

Bdd' Bdd" Bdd* Bdd" Bdd' Bdc

Bdd* Bdd' Bdd' Bdd" Bdd" Bdd' Bdd" Bdd' Bdd' Bd

192 41 206 177 190 154 114 90 168 189 215 182 195 80 43

6.7;* 41.5j« 36.9^ 16.45< 6.89^ 13^ 21. H 17.8^ IBM 1% IZ.Si IB.7^ 20.5?t W.ejf 37.2*

Family ^.

Bd9,48; 4.2^^

Bdd", 187; 1.5^<

Bd d", 209 ; 2.4'i
Bd d", 237 ; 4.2;f

Family 5.
Bd 9 , 82 ; 15.9^

d-, 314; 7.65<

Bd 9,282; 1.8^
Bd9,60; 8.3{(

Bd d*, 175 ; 9.7'

Bd d', 144 ; 24.8

Bd d", 50 ; 32'< Bd d", 82 ; 20.7?^

Repeated baok-erosses of Beaded X Wild in successive generations, show-
ing sex of Beaded parent, number of offspring, and percentage of offspring
with Beaded wings.

Digitized by

Google

No. 576]

VARIATION IN DROSOPHILA

739

generations of selection, and now gives approximately 100
per cent. Beaded offspring, though no selection has been
practised for nearly a year. This stock has been used in
one series of crosses to supplement another series in
which Pink and Beaded enter the cross from opposite
parents. The results in each case are essentially similar,
and show that when Beadedness enters with Red it comes

TABLE XXV
Fj Results op the Cross Pink Beaded X Wild

Bd. Pink Normal Pink < Beaded Red Normal Red Exp. No. Bd. P. Exp. No. Bd. R,

171

213

366

964

120

417

out more with Red than with Pink. They show that in
the Fj female crossing over occurs almost independently
of Pink, so that almost the same percentage of Beaded-
winged individuals appears in each class, though usually
the class that is similar to the Beaded parent is consider-
ably the largest. In Table XXVII, however, a record
is given in which a very considerable '* repulsion" oc-
curred, and the high Beaded class is not Pink Beaded,
as is there expected, but Red Beaded. The results from
back-crosses of the brothers of these females to Pink
normal stock show that no mistake was made in record-
ing the cross, which therefore, though somewhat surpris-
ing, must stand.

TABLE XXVI

Back-crosses of Fj Females of the Cross Pink Beaded X Wild to Pink
Males or Normal Stock

Bd. Pink N. Pink Bd.Red | N. Red

Exp. No. Exp. No.
Bd. P. Bd. R.

TypeX

TypeY

333
332

2
58

337
369

2.5
62.6

2.5
66.4

Total

665

706

The tables show also that in the males, crossing over is
of very rare occurrence, if, indeed, it occurs at all. The
records show that out of 566 Beaded flies (Tables XXVII
and XXIX) which occurred as the offspring of an Fj male

Digitized by

Google

740

THE AMERICAN NATURALIST [Vol. XLVIII

back-crossed to Pink normal stock, six flies of the cross-
over class appear. For reasons to be mentioned, it is
improbable that these represent cross-overs, however, but
rather they may be due perhaps either to the presence of
the second chromosome gene, 1, which usually does not
manifest itself in the absence of the third chromosome
gene, or to mutation, or to some. unknown abnormality.
Through carelessness only one of these males was tested

TABLE XXVII

Back-crosses or Fj Males of the Cross Pink Beaded X Wild to Pink
Females op Normal Stock

Bd. Pink

N. Pink

Bd. Ked

N. Red Exp.No. Bd.P.Exp.No.Bd.R.

710

805

or used further in breeding. They were very slightly
Beaded, and had only a very slight **nick^' at the tip of
the wing, even smaller than that shown in Fig. 3. The
single Pink Beaded male mentioned in Table XXTX was
mated to several females but was sterile. Another test
is also possible, and was made as follows. Pink normal
males and females from Table XXIX, which of course
should not carry the third chromosome gene for Beaded

TABLE XXVIII

Repeated Back-crosses of Fi Flies op the Cross Pink Beaded X Wild
to Pink Flies or Noemal Stock

Bd.
Piok

N.
Pink

Bd.
Red

N.
Red

Exp.

No.

Bd.P.

Exp.

No.

Bd. H.

Fi Q XPink c^.

Type X

1
24

228
379

1
79

272
383

1
48

Fi$ XPink d^.

TypeY

! ^

Total

607

655

Ficf XP$

133

135

3.5

i 3.5

wings, were then mated together, and among their 374
offspring three males with slight ** nicks*' at the tip of
their wings, exactly like those of the Pink Beaded male
before mentioned, were produced. One of these males
was sterile. One of the remaining two was fertile, but

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 741

gave no Beaded offspring either in the first generation or
in the Fg generation, although nearly one thousand of his
grandchildren were carefully examined. The remaining
male was abundantly fertile and had one son exactly like
his father in appearance (with a slight nick at the tip of
the wings). The rest of his offspring were normal. This
son was sterile.

TABLE XXIX

Back-cbosses or Fj Plies or the Cross Beaded X Pink to Pink Flies
OF Normal Stock

FicT XPink 9
Fi9 XPink cf

Beaded
Pink

1
70

Normftl Beaded | Normal Exp. Bd.

Pink Red I Red Pink

869 504

114 I 114

580 223

106 84

Exp. No.
Beaded R.

282
100

The results of these tests with five of these supposed
** cross-over*' males show clearly that they were not nor-
mal Beaded flies. As said, they might represent muta-
tions, or the dominance of the gene 1, or some abnormality.
These are mere guesses, but since there are no authentic
cases on record in Drosophila of crossing over in the male
sex in those cases where the mutants dealt with are well
known genetically, i. e., since the only apparent cases
occur in the Beaded wings and some of the other not
well-known and peculiar mutants of Drosophila, we are
not justified in assuming that such crossing over takes
place here.

ni. THE EFFECT OF ENVIRONMENTAL CONDITIONS UPON THE
PRODUCTION OF BEADED WINGS

A. Geneeal Statement

If we have so far interpreted the evidence correctly we
may formulate the following statement as a provisional
hypothesis. A gene B' located in the third chromosome
near that for Ebony is directly responsible for the pro-
duction of Beaded wings. By itself in the homozygous
condition, the fly bearing it may have normal wings,
though it usually will have wings somewhat Beaded. In
the heterozygous condition, it is rarely, though sometimes,

Digitized by

Google

742 THE AMEBIC AN NATURALIST [Vol. XLVHI

dominant. The conditions so far presented which canse
it to be dominant are two. (1) The presence of a gene 1
in the second chromosome which can not exist in the
homozygous condition. (2) The influence, particularly
noticeable in the males, of non-chromosomal constituents
of the egg from which the individual arose, so that if the
mother had been Beaded, the appearance of Beaded wings
in her sons would be increased, and if the mother had been
normal the appearance of Beaded wings in her sons would
be reduced.

Certain facts already brought out (namely, those pre-
sented in Tables II and III) show that the tale is not yet
told. Our hypothesis does not explain the fact that from
definite numbers of eggs laid at different periods in the
life of an individual very different percentages of Beaded-
winged offspring arise, and these differences do not form
a definite series progressing to or from a high percentage
as the individual grows older, but are extremely irregular.
We have not gained control over this phenomenon, but
the evidence we have to present points strongly to the
suggestion that the environmental conditions are the final
determiners of the percentage of the Beaded-winged off-
spring. This environmental control might lie in three
'distinct methods : (1) The destruction of a certain class of
•offspring by their differential viability. (2) In the case
of Table III the results might be explained on the theory
that Beaded flies had a shorter life cycle. This supposi-
tion has, however, been disproved as follows. Five non-
virgin females from Beaded-winged stock and five non-
virgin females from normal-winged stock were put to-
gether without males in the same bottle. When the off-
spring began to hatch they were examined daily. During
the first three days 73 flies hatched, of which 11, or 15 per
cent., had Beaded wings. During the following five days
261 flies hatched, of which 54, or 20 per cent., had Beaded
wings. Since I was particular to take Beaded flies several
days old as the parents of these Beaded offspring, the
experiment shows that if there is any difference in the
length of the larval life, that of normal-winged flies is

Digitized by

Google

No. 576] VARIATION IN DBOSOPHILA 743

the shorter. (3) The determination of whether or not a
fly of a given germinal constitution shall have Beaded
wings. The first of these effects is probably not the
significant one, in view of the following facts.

Although as a rule Fj normal flies give few Beaded off-
spring, and Fi Beaded-winged flies relatively many,
nevertheless, as has been said, at times normal flies give
a high percentage of Beaded offspring and, occasionally,
Beaded flies give a low percentage. This can only mean
that the dominance of the factor B' is variable, and con-
sidering the large number of times that it shows itself as
a recessive, it must be that this varying dominance has a
marked effect on the percentage of Beaded-winged off-
spring that appear.

The possible amount of variation in the environment
surrounding a brood of Drosophila developing under
laboratory conditions is enormous, even when the attempt
is made to keep conditions constant. These variations
depend upon the exact ripeness of the bananas used aa
food, the length of time the food has been fermenting, the
amount of food and filter paper used, the size of the bottle
in which the larvae are developing, the tightness of the
cotton plug, the temperature of the laboratory, etc. Due
to these causes there arise very great differences in the
relative moisture content and carbon dioxide content. If
the food is not properly prepared it may rot instead of
fermenting, or it may mould, or the reaction may be in
one bottle quite alkaline and in another very acid. A
perfect control thus becomes an impossibility, and there-
fore the experiments to be described must be considered
as trials only, and not as decisive tests.

In all the experiments on this subject. Beaded flies of
pure stock were mated to normal flies of Wild stock in
order to learn the effect of particular environments on
the percentage of Beaded offspring in the Fj generation.
On our hypothesis, the pure Beaded flies from stock should
be of two kinds, viz., those with the lethal gene 1 (i. e.,
B'lB'L), and those without 1 {i. e., B'LB'L). Correspond-
ingly there should be two types of offspring in the F,

Digitized by

Google

744

THE AMERICAN NATURALIST [Vol. XLVHI

generation, one of which (B'lb'L) should have a consider-
ably higher percentage of Beaded offspring than the
other (B'Lb'L). If it is possible, however, that B' should
be dominant in the heterozygous condition and in the
absence of 1, then it should also be possible theoretically
to produce an Fj generation every individual of which
should have Beaded wings, while those with 1 as well as
B' (constituting one half the progeny) should have a
more extreme form of Beading. In practice it is not
usual even under the best of conditions to get more than
40 per cent, of Beaded-winged flies, while, as has been
seen, the average amount is about 25 per cent.

B. The Effect of Relative Moisture
Table XXX and Charts 5 and 6 present the data for

table XXX

Percentages of Beaded-wikged Flies in the Fj Geneeation in Rela-
tively Wet and Dey Bottles. Not Done in Pairs, but
Each Bottle Contained Several Pairs

Dry Bottles

Wet Bottles

No. 1 1 No. 2

No, 3

No. 4

No. 5

No.
Flies

Md. ^{i, jSBd.

.^i^ee ^Bd.

No.
Flies

f(Bd.

F^i^esl^Bd..

let count

2d count

38
122
34
33
59
95

31.6
14.7
20.6
27.3
20.3
10.5

32 , 43.7
58 ' 17.2
19 21.1
34 8.8
72 25.0
37 35.1

26 i 26.9
79 32.Q

14
87
16
17
30

28.6
20.7
37.5
41.2
36.7

26 30.8
131 26.7

3d count

41.fi

37 37.9

4th count

6th count

39 35.9
35 31.4

43 41.7
53 18.9

Total

281

17 R 2.'i9 24 fi

203 ' 3.'?.5

164

^fi

2on 9Q a

_... _._ _._

Tot

xvr« 1 , -- --

Total Dry I

>33

3d. 20.5

al Wet, 657

% Bd. 30.3

Counts not made every day.

Bottle No. 2 was very dry and the flies very small during time of last
two counts.

this test. The parents were put into bottles of similar
size with plenty of food. In three of these bottles the
food was very wet and from time to time juice was added
in sufficient amount to Keep the food saturated. The
other two bottles were made relatively dry by putting a

Digitized by

Google

No. 676]

VARIATION IN DROSOPHILA

lib

Chart V

a»

•C 4

'5.

•4

<»

o <

• <

c S

3 :

r*-

r* T

6C1*

ioio

30*

kof

^\^

_ ■

ioio

01.

Wet

Effect of Relative Moisture in Food on Percentages of Beaded-winged
Flies In F^ Generation of Beaded X Wild.

large amount of filter paper into the bottle at night and
removing it the following morning. After two or three
days of this treatment the bottles were so dry that I did
not venture to carry the process farther; the flies from

Chart VI

.-^^^^

^_^

Witi

»*f

nT"^

"Utxv^y^

*•*

/____

Wd>^v^^^

i*i

»i

Drya

Dry*

Effect of Relative Moisture in Food on Production of Beaded Wings, as
shown by Individual Bottles.

Digitized by

Google

746

THE AMERICAN NATURALIST [Vol. XLVHI

these dry bottles were rather small and in bottle No. 2,
they were extremely small in the last two counts.

From Chart 6, where the records are given of the indi-
vidual bottles, it will be seen that there is a good deal of
irregularity from day to day.

Special attention should be called to the curve of pro-
duction of bottle 2, which beginning with a high per-
centage of Beaded offspring gives fewer and fewer for
the first four counts (about six days) and then the per-
centage rapidly mounts again. The offspring given dur-
ing the last two counts were of surprising minuteness and
gave as high a percentage of Beaded individuals as the
average of all the bottles on the first day. It has been sug-
gested that it may not be wetness or dryness or any one
specific thing that brings out the Beadedness, but condi-
tions that are unfavorable to the organism as a whole,
resulting in poor nourishment. It has frequently been

TABLE XXXI

The Influence op Acid, Alkaline, and Fresh Food on the Development

OF Beaded Wings

^ !

« s .

a. *'o

Food Sour
Mother Beaded ; Father normal
Father Beaded; Mother normal

Food Freah
Mother Beaded; Father normal
Father Beaded; Mother normal

Food Alkaline
Mother Beaded; Father normal
Father Beaded; Mother normal

40|151

7 60!

23 25
54*15

41 '
16 1

85
147,

4o:

57l

129
61

92
164

14.7
13.0

21.3
26.9

46.8
32.9

23.8
10.3

21.4
8.4

50.0
21.6

19.1
11.7

21.3
18.2

48.7
27.7

noted that those bottles which gave very tiny flies gave
also a higher percentage of Beaded individuals than the
bottles whose flies were of average size. On the other
hand, the first flies of a brood are almost invariably larger
than the later ones, and yet, as has been seen, they are
more Beaded. This is a paradox, but the behavior of
bottle No. 2 suggests that as a hatch proceeds and the
bottle becomes drier, there may be a certain optimum
point for the production of normal winged offspring, and

Digitized by

Google

No. 576] VARIATION IM DROSOPHILA 747

that this point is so low that the flies are poorly nourished
for lack of water, though they can survive an even
greater water reduction.

It is, perhaps, needless to say that an effort has been
made after these experiments to keep the moisture con-
tent high and fairly uniform in cases where other envi-
ronments were being tested.

C. The Effects of Covering with Paraffine the Mouth
OF THE Bottle in which the Flies are Developing

On observing that the proportion of Beaded to Normal
offspring was lowered as a hatch continued, it seemed
possible that this might be due to one or to both of two
causes: (1) The diminishing water content. This matter
has already been considered. (2) To a changing carbon-
dioxide content. When a brood is first counted the cotton
plug that has been for several days in the mouth of the
bottle is removed, and in removing the flies the air within
the bottle is very apt to be much changed. With this pos-
si]?ility in mind a number of bottles were supplied with
food and flies, and after ten days (when the larvae were
beginning to pupate) the parent flies were removed, a
little new food put into the bottle and a paraffine cap
melted over the cotton so that the bottles were tightly

table xxxii

Counts op Seven Broods which Hatched During Two Periods, the First

OF Which was Spent in a Bottle Sealed with Paraffine,

AND the Second in a Bottle Covered with Cheese

Cloth. Bottle No. 7 was not Sealed with

Paraffine but had been Lightly

Stoppered with Cotton

Bottle 1 Bottle 2 Bottle 3

s 0 s a

Ist count

2d count

36 82 40 54
9 32 3 17

44
6

Total, first count, 426, per cent. Bd., 36.
Total, second count, 228, per cent. Bd., 6.
Total, both counts, 654, per cent. Bd., 25.

Digitized by

Google

748 THE AMERICAN NATURALIST [Vou XLVIH

sealed. At the same time other bottles were very loosely
covered with a light cotton plug. The bottles remained
covered till flies had been hatching for four or five days
and then the plugs were removed and the flies counted.
The paraflBne plugs were not replaced ; after carefully
renewing the air in the bottles, they were covered with
cheese cloth and their brood counted again in four days.
The results of this test are given in Table XXXII. The
results are striking enough at first sight, but I do not
know just what their significance is. They show exactly
the same phenomenon that is described earlier and illus-
trated in Table III. They are more striking than any
case I have yet found of the sort, and yet the first infer-
ence drawn, viz., that the markedly higher percentage of
Beaded flies in the first count is due to these flies having
undergone their late development in a ''close'* atmos-
phere, must be qualified by the statement that ''close*'
does not refer to the carbon-dioxide content.

At first suspecting this to be the case, I made an appa-
ratus by means of which fresh air could be drawn through
a bottle during the entire development of the brood. By
this means the carbon-dioxide content could not become
very high. In order to prevent drying out, a large amount
of food was put into the bottle and the air which was to
enter the bottle was first passed through water. The
hatching period was prolonged in the cool sink. The re-
sults were decisive. One hundred and sixty-nine flies
were hatched in the first four days, of which 32 per cent,
were Beaded. One hundred and eighty-four flies were
hatched in the next four days, of which 10 per cent, were
Beaded.

The same flies that were the parents of this brood were
in the meanwhile transferred to another bottle, which was
covered with paraffine. The first four days of hatching
gave 108 flies, of which 15 per cent, were Beaded.

This case shows conclusively that the carbon-dioxide
content of the bottles is not the feature of the closed
bottles that determines whether or not a fly shall have
Beaded wings. It leaves the question still unsettled as

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 749

to the effect of moisture, but corresponds to the results
obtained in the study of moisture effects.

D. The Effects of Acidity and Alkalinity of the Food

Normally the reaction of food at the time of putting it
in the bottles is acid, the degree of acidity depending upon
the length of time it has been fermenting. This sourness
usually passes gradually away as the larvae grow older,
and by the time a brood begins to hatch the reaction is
frequently quite alkaline, unless fresh food has been put
recently into the bottle.

On the other hand, if the acidity of the food is neutral-
ized at the beginning with sodium hydrate or carbonate,
or if the reaction is made alkaline while yet there remains
a good deal of unfermented banana, the acidity will re-
turn for a time if not carefully guarded against. There-
fore to keep the reaction acid or alkaline is a difficult
matter, and requires occasional stirring of the food to
make the reaction uniform; this operation is likely to
prove disastrous for the developing pupae.

In the tests here recorded I used food that had been
fermenting for one month, so that it had a very acid
reaction that lasted till hatching time. For studies of the
effect of alkalinity I used food that had been fermenting
about one day and mixed with it sodium carbonate,
sodium hydrate or ammonia. The results were unsatis-
factory and the reaction did not remain constant in spite
of my efforts, though on the whole it remained alkaline,
and became strongly alkaline, and also slimy towards the
end of the experiment, and not a great many flies hatched.

I also used food that had not been allowed to ferment
at all, and although I do not know its reaction, it was
certainly not so alkaline as the last mentioned, nor so
acid as the first. It was soon attacked by mold (Bread
mold). I refer to it here as fresh food.

The results are given in Table XXXI, but may be more
briefly summarized here.

Of 483 flies raised on sour food, 17 per cent, were Beaded.
Of 605 flies raised on fresh food, 19.3 per cent, were Beaded.
Of 317 flies raised on alkaline food, 38.1 per cent, were Beaded.

Digitized by

Google

750 THE AMERICAN NATURALIST [Vol. XLVHI

In other words, a high percentage of Beadedness came
from flies raised on alkaline food, a low percentage from
flies raised on acid food, and intermediate amount from
flies raised on fresh food.

A careful study of Table XXXI will reveal the curious
partial sex-linkage of which I spoke on pages 15 et seq.,
and here, too, the explanation suggested there seems to
apply as in other cases of the sort. It is not a littlef
peculiar that in all of these food tests this phenomenon
should have occurred, though I consider this purely a
coincidence. In any case, if we can draw any conclusion
at all from its appearance, it would only be that the re-
action of the food has nothing to do with the occurrence
of the phenomenon rather than the reverse.

E. The Effects of Relative Temperatubes

No evident effect was produced by rearing the Fj gen-
eration in an ice-chest, but ratios were as varying as when
the flies were raised at room temperature. Ratios of 15.4
per cent., 19.2 per cent., 10.3 per cent., 20 per cent, of
Beaded offspring are examples of those given by broods
raised at low temperatures. The cold does, however,
lengthen greatly the larval life and flies were in the case
of the brood last mentioned twenty-eight days in hatching.
The brood consisted of 312 normal and 77 Beaded-winged
flies.

Similar results were obtained in experiments with
heat, except that here the larval life was correspondingly
shortened and was at times reduced to eight days. It
was not found practicable to keep the flies at higher than
30°-33° Centigrade, as they soon died at higher tempera-
tures.

F. The Effects of Dabkness

Flies were raised in complete darkness and sister
broods in full daylight, but no differences appeared in the
offspring. Of 484 flies raised in darkness 30 per cent,
had Beaded wings. Of 360 flies raised in the daylight,
29 per cent, had Beaded wings. This experiment seems

Digitized by

Google

No. 576] VARIATION IN DBOSOPHILA 751

to show conclusively that light and darkness do not influ-
ence the percentages of Beaded- winged flies.

IV. THE EFFECT OF SELECTION ON THE PRODUCTION OF
DIFFERENT TYPES OF BEADEDNESS

Just how much can be accomplished by selection in
Beaded stock was one of the first questions that arose.
Morgan (1911a) describes the origin of pure Beaded stock
as having occurred through the selection of Beaded flies in
the early generations after its first appearance. He says
the first Beaded fly found arose in a culture of Droso-
phila that had been exposed to radium. Mated to his
sisters, 1.6 per cent, of the offspring were Beaded. When
these Beaded flies were inbred 3 per cent, of the offspring
were Beaded. These inbred gave 8.5 per cent. Beaded
offspring.

The same process continued through many generations has finally
produced stock that gives in certain cultures nearly 100 per cent.
Beaded wings.

In continuing these selection experiments, he says more
extreme forms of Beaded wings appeared, and at the
time of publishing (March, 1911) he was attempting ''to
fix some of these extreme variations." While engaged in
this work other wing forms arose, most of which are
among the best -known mutants of Drosophila. Among
these are Truncate, Miniature, Rudimentary, Vestigial
and Balloon wings, and the Black and Yellow body colors.
Most of these forms have been ''purified" now and
Beadedness never appears in them though it can still be
found in Vestigial stock. All of the above-n^med forms,
by the way, with the exception of Truncate and Rudi-
mentary bred true from the start. The Truncate case is
not yet published and Rudimentary has proved (Morgan
and Tice, 1914) to be due to a single Mendelian factor.
The Rudimentary flies were at first self sterile and highly
non- viable, and therefore gave peculiar results in breeding
tests.

When I first began work with Beaded flies (Sept., 1912)
the stock gave 100 per cent. Beaded-winged offspring.

Digitized by

Google

762 THE AMERICAN NATURALIST [Vol. XLVm

So soon that I did not realize it, nor think to count the
generations, I had one stock that gave offspring much
more extremely Beaded than the ordinary stock, and this
stock is the one on which most of this report is based.
About December, 1912, 1 started one stock bottle to form
the basis for a ^'No selection '^ test. The parents of this
brood were '*pure stock Beaded '^ males and females.
The first generation, no normal-winged flies appeared.
The generations following were made up by shaking at
random from the bottle of the generation before a dozen
or two flies into a new bottle.

The second, third and fourth generations gave three
normal-winged flies to 325 Beaded. The sixth, 3 normal
to 100 Beaded. In later generations I occasionally found
normal flies. The stock is in its 27th generation now,
the 25th generation having given rise to a large brood
of which I counted 541 flies (284 $ and 257 c?), all of
which had Beaded wings of a type averaging like those of
Figs. 4-6. It is very apparent that the stock is not under-
going any marked change, though I can not guarantee
that it would give exactly the same results in other
respects as the extreme (selected) Beaded stock that I
have used in the linkage tests.

On the other hand, I have not been able thus far to
increase the Beadedness of the selected stock beyond a
point which it apparently reached many generations ago.
The Figs. 1-12 (excepting 2 and 4), which are here re-
produced, were made under Dr. Morgan's direction long
before I took up the work, and the forms he had drawn
then are as extreme as any that I now have.

If this extreme stock be allowed to go without selection
for two or three generations, it ''reverts" to a less ex-
treme form, from which it can apparently be recovered
by one mass selection. I feel confident that in selecting
the extreme forms one merely selects a large percentage
of individuals that are heterozygous for 1, and of course
when the stock is not selected for a while, LL forms be-
come relatively more numerous. This would account for
all the facts here recorded.

Digitized by

Google

No. 576] VARIATION IN DROSOPHILA 753

On the other hand, selection for less extreme Beading
is also rapidly effective and normal-winged forms appear
soon, but this effect soon reaches its limit apparently,
and a normal strain or even a strain throwing a high
percentage of normals has not yet been obtained. I am
not yet certain that it can not be done. I selected in each
direction for eleven generations without marked success
beyond that here recorded.

V. MUTATION IN BEADED STOCK

A. General Statement

As will be gathered from statements made in the last
section, the Beaded stock has been prolific in giving muta-
tions. There has been no especial attempt made to see
how many different mutants could be obtained from the
stock, and yet a goodly number have appeared. Most of
these have been marked types showing little variation and
coming out regularly and distinctly in Mendelian propor-
tions in crosses with other types. They have in general
bred true from the start without further selection.

A few of these have been of a sort to confuse for a time
the study that I have been making, because of their re-
semblance to certain types of Beaded flies. The criterion
in every case as to whether or not a fly was an ordinary
Beaded fly or a new '* mutant** was its genetic behavior,
and the cases to be here described have, with the excep-
tion of Stumpy, shown themselves to be due to a single
gene conforming in general to those of other well-known
mutants of Drosophila.

B. Perfect Notched Wings

In the beginning of my work on Beaded wings I thought
it might be possible to isolate definite types from the
Beaded stock by crossing out to Wild and extracting the
Fg types that appeared ; or by back-crossing the Fj forms
to Wild again and extracting new types, etc. Several
thousand flies were raised in the hope of accomplishing
this, but the *' types'* found did not breed true, but con-
tinued to behave like ordinary Beaded flies, from whose

Digitized by

Google

754 THE AMERICAN NATURALIST [Vol. XLVIH

many original types none were distinguishable. Finally
a genuine new *'type'' appeared, with both wings alike
and definitely '^notched" (Fig. 13, p. 730). This female
which was at sight named Perfect Notched, was mated
to Wild. Her ancestry was as follows :

The grandmother came from pure Beaded stock, and
the grandfather from Wild stock. Their offspring con-
sisted of 13 Beaded and 69 normal flies.

A Beaded female of this generation was mated to a
normal brother and gave 100 Beaded offspring, male and
female, and one ''perfect notched'* female.

This female and her descendants behaved in a very
different manner, genetically, than the Beaded stock
from which she arose.

She was mated to a Wild male and gave 62 Beaded off-
spring and 112 normal offspring. Of the Beadeds, 50
were notched in a way resembling the parent and of the
50, 49 were females. Several other peculiar wing types
appeared among the remaining 12 Beaded flies of this
generation, but did not breed true and were later dis-
carded. /

The notched male gave ordinary Beaded and normal
offspring and never gave in either the first or later gen-
erations any ''notched'* offspring. He was probably an
extreme variant of a common Beaded type (Fig. 4).

Of the normal offspring of the Perfect Notched female
four pairs were made up. Seven hundred and forty-nine
normal sons and daughters appeared, and no notched.

Of the notched daughters of the perfect notched female,
two were mated to normal brothers and two to Wild
males. Their progeny was :

Notched 9 Notched cf Nonn*l 9 Norm*! cf

By normal brothers 53 0 79 69

By wild males 56 0 47 46

Six of the notched females of this generation were
mated to normal brothers and gave

Notched 9 Notched <^ Normal 9 Normal cf

126 0 144 120

At this time, June, it was necessary to leave New York.
In traveling, the Perfect Notched stock was lost. Enough

Digitized by

Google

No. 576] VARIATION IN DBOSOPHILA 765

had, however, been done to show definitely the nature of
the mutation involved. It is clear that the perfect notched
wings owed their appearance to a dominant sex-linked
gene, lethal for males. This accounts for the fact that
the males are only half as numerous as the females, and
none of them notched, while notched and normal females
occur in nearly equal numbers. It also accounts for the
fact that the normal females of these generations gave no
notched offspring.

Other sex-linked lethal genes have appeared from time
to time in the crosses of Beaded flies with others, but
none of them were dominant, and therefore they made
themselves evident only by preventing the development
of one half of the males. I have not worked out the
inheritance of these cases.

C. Spread Wings

Comment has already been made on the extreme num-
ber of wing types that appeared both in the Fj, Fg, and
back-cross generations of the cross between Beaded and
Vestigial flies. Most of these forms gave results too com-
plex to be analyzed at present. However, among the off-
spring of a considerable number of the Fj females there
were flies with wings perfectly normal in appearance save
that they were held at right angles to the long axis of the
body. In all, 60 flies with Spread wings appeared. One
of the 60 had wings very slightly Beaded. Some of them
were mated together and produced only spread-winged
offspring with no sign of Beadedness. Spread-winged
males were mated to Pink Black females in order to test
the linkage of Spread. (Pink is in the third chromosome
group, and Black in the second.) The Fj generation gave
only flies with red eyes, gray bodies and normal wings
(neither Spread nor Beaded). In the Fo generation were
Black flies. Gray flies, and Red-eyed flies with normal
and with Spread wings, but none of the Pink-eyed flies
had Spread wings, though a large number of Fg Pink
normal flies appeared. The Pink-eyed flies were also
mated inter se, but no Spread-winged flies appeared in
the F3 generation. This definitely places the gene for

Digitized by

Google

756 THE AMERICAN NATURALIST [Vol. XLVHI

Spread wings in the third chromosome gronp. Beaded
wings have not appeared in the stock bottles of Spread
which breeds perfectly tme.

D. Stumpy Wings

Eecently a new non-lethal sex-linked character has ap-
peared in the offspring of the cross of an Fj Beaded male
to a Wild female. Its nature has not yet been worked
out, since only males have thus far appeared. The flies
have wings resembling those of Vestigial, save that they
are not held at right angles to the body, but in the normal
position. Vestigial is not a sex-linked character.

SUMMABY

The character under consideration is that of Beaded
wings in Drosophila ampelophila. All gradations of
form between that of normal wings (Fig. 1) and those
shown in Figs. 2 to 12 occur in the stock bottles, though
certain selected strains of the stock give no normal-
winged offspring.

When a Beaded fly is mated to a fly of a stock not
carrying genes for Beadedness in its germ plasm a vary-
ing percentage of the Fj offspring is Beaded. If the male
parent is Beaded the majority of the Beaded offspring
are usually females ; and if the female parent is Beaded,
the majority of the Beaded offspring are usually males.
A female Beaded fly however gives a larger percentage
of Beaded daughters than does a male Beaded fly. This
phenomenon is repeated from generation to generation,
no matter whether a given Beaded fly has come from a
male or female Beaded parent, and this shows that the
phenomenon is not caused by a sex-linked gene.

This phenomenon is not caused by non-disjunction of
a sex-linked gene, for tests of both the Beaded and "Wild
stocks showed non-disjunction to be a rare phenomenon.
The only explanation suggested was that the male off-
spring were somewhat influenced to or away from Beaded-
ness by the nature of the cytoplasm that was brought in
with the eggj while females were not readily influenced
in this way.

Digitized by

Google

No. 576] VABIATION IN DROSOPHILA 767

A study of the Fg generation shows that the majority
of the normal Fi offspring differ from the majority of
the Beaded Fj offspring genetically in that normals give
fewer Beaded offspring in the Fg generation than do the
Beaded flies.

Beaded wings showed no linkage to any sex-linked
character.

Approximately one half of the flies of the Fj generation
of a cross between Beaded flies and flies with characters
whose genes were in the second chromosome, showed
linkage in the following generation to second chromo-
some characters, while one half of the flies did not show
such linkage. The cases where linkage did not occur
gave a slightly lower percentage of Beaded offspring
than did those where linkage was present. An explana-
tion of these phenomena is sought in the suggestion that
there was in the second chromosome a gene, here called 1,
that was recessive but that in the heterozygous condition
intensified the dominance of another gene, called B',
which was not in the second chromosome. This gene 1
behaves as a lethal factor preventing the development of
any fly that carries it in a homozygous condition.

All of the Ft offspring of the crosses of Beaded flies
by flies with characters caused by genes in the third
chromosome showed linkage in the following generation
between Beaded wings and the third chromosome char-
acters. This was taken to signify that there was in the
third chromosome a non-lethal gene concerned in the
development of Beaded wings. This gene was called B^
This gene was shown to be the essential germinal factor
in the production of Beaded wings. It is sometimes
dominant and sometimes recessive.

The determination as to whether B^ should be dominant
or recessive seems to lie in several possibilities : 1st, the
nature of the egg cytoplasm ; 2d, the presence or absence
of the gene 1; 3d, the nature of the environmental con-
ditions.

With reference to environmental conditions, it was
shown that a larger percentage of the Fi generation had
Beaded wings when the culture was wet than when it was

Digitized by

Google

758 THE AMERICAN NATURALIST [Vol. XLVm

dry; and more when the food was alkaline than when it
was acid. No other environmental factors were discov-
ered which influenced the production of Beaded wings.

Selection of more or less extreme Beaded flies very
quickly moves the average Beadedness of the offspring
in the direction of the selection, but this selection appar-
ently becomes further ineffective in a very few genera-
tions.

Mutation is of very frequent occurrence in the Beaded
stock and the new mutants obtained have in most cases
shown themselves to be produced under the influence of
one normally Mendelizing gene.

I acknowledge with pleasure the kindly interest and
suggestions made from time to time by Dr. A. H. Sturte-
vant and Mr. H. J. MuUer. These have been of much
assistance to me. My thanks are also especially due to
Dr. T. H. Morgan whose advice and criticisms at critical
points have never failed to aid in clearing up the situation,

BiBLIOQEAPHY

Baur, E. 1912. EinfUhrung in die experimentale Vererbungslehre.
Bridges, C. B. 1913. Non-disjunction of the Sex Chromosomes of Vro-

sophila. Jour. Exp, Zooh, Vol. 15.
Hoge, M. A. 1914. The Influence of Temperature on the Development of

a Mendelian Character. Jour. Exp. Zool. In press.
MacDowell, E. C. 1914. Multiple Factors in Mendelian Inheritance.

Jour. Exp. Zool, Vol. 16.
Morgan, T. II. 1911a. The Origin of Nine Wing Mutations in DrosophUa,

Science, N. S., Vol. 33.
Morgan, T. II. 1911&. A Dominant Sex-limited Character. Proc, Soc.

Exp. Biol, and Med., Vol. 9.
Morgan, T. H. 1912a. The Masking of a Mendelian Eesult by the Influ-
ence of the Environment. Proc. Soc, Exp, Biol, and Med,, Vol. 9.
Morgan, T. H. 1912&. The Explanation of a New Sex-ratio in DrosopMla,

Science, N. S., Vol. 36.
Morgan, T. H. 1912c. Complete Linkage in the Second Chromosome of the

Male. Science, N. S., Vol. 36.
Morgan, T. H. 191 2d. A Modification of the Sex-ratio and of other Ratios

in Drosophila through Linkage. Zts. ind. Abst. u. Vererh., Bd. VTI.
Morgan, T. H., and Tice, S. C. 1914. The Influence of the Environment

on the Size of Expected Classes. Biol. Bull., Vol. 26.
Shull, G. H. 1914.
Sturtevant, A. H. 1913. A Third Group of Linked Genes in Drosophila

ampelophila. Science, N. S., Vol. 37.

Digitized by

Google

SHORTER ARTICLES AND CORRESPONDENCE

ON THE PROGRESSIVE INCREASE OP HOMOZYGOSIS
BROTHER-SISTER MATINGS

It has been brought to my attention that the note concerning
inbreeding, written at the request of Mr. Phineas W. Whiting to
add to his paper on ** Heredity of Bristles in the Common Green-
bottle Ply, Lucilia Sericata Meig.,'' which appeared in the Amer-
ican Naturalist for June, 1914, might be taken to mean that my
data had been sent by Dr. E. M. East to Dr. Raymond Pearl by
whom it had been published as his own. I wish to make it clear
by a statement of the facts herewith that no such interpretation
should be placed upon the note. I was seriously ill at the time
and did not submit my manuscript to Dr. Castle or Dr. East for
revision, as I should ordinarily have done. In that case no
doubt, any ambiguity of statement would have been pointed out
to me.

Mendel, in his original paper, showed that if equal fertility
of all plants in all generations is assumed, and, furthermore, if
every plant is always self-fertilized, then in the nth generation
the ratio of any allelomorphic pair (A,a) would be 2" — 1 AA.:
2 Aa : 2*» — 1 aa. This statement was generalized in 1912 by East
and Hayes^ for any nimiber of allelomorphic pairs. **The prob-
able number of homozygotes and any particular class of heterozy-
gotes in any generation r is found by expanding the binomial
[l+(2'' — 1)]" where n represents the number of character
pairs involved. The exponent of the first term gives the number
of heterozygous and the exponent of the second term the number
of homozygous characters.*' A little later Jennings independ-
ently showed how homozygotes are produced from heterozygotes
by self-fertilization.^

East and Hayes^ published no generalized formula for -calcu-
lating the reduction toward homozygosis through any other type
of mating, but that this was thought to be a proper conclusion
deducible from the above is shown by the following quotation
(p. 21) :

1 U. 8. Dept. Agr., Bur. Plant Ind., BuU. No. 243.
2 Ames. Nat., Auijust, 1912.
. iLoc, cit.

769

Digitized by

Google

760 THE AMERICAN NATURALIST [Vol. XLVIH

Close selection, of course, tends toward the same end (homozygosis),
but not with the rapidity or certainty of self-fertilization.

This idea is further shown by their statements under the head-
ing "Extension of Conclusions to the Animal Kingdom" (pp.
39-43).

A little later Mr. Whiting had occasion to work out the results
of random matings of brothers and sisters, in connection with his
work at the Bussey Institution. He found that the amount of
heterozygosis was reduced one eighth in matings of the Pj gen-
eration and from this concluded that the. remaining heterozygosis
was reduced one eighth in each succeeding generation, so that in
the nth generation the number of matings which would produce
at least some heterozygous offspring would be (7/8)"-^. He
showed these figures to Dr. East, who agreed with the general
conclusion (tendency toward homozygosis), but thought that the
ratio would not hold for offspring after the Fg generation. Dr.
East, however, after a casual examination was not able to show
Mr. Whiting the fallacy in his work and did not go into the
matter further.

In the American Naturalist for October, 1913, Dr. Raymond
Pearl criticized the extension of the conclusions for self -fertilized
plants to the animal kingdom.* He applied the figures of Pear-
son, 1904,*^ for random matings, which show that the relative
number of homozygotes and heterozygotes remains constant in
a population where all factors of fertility, virility and environ-
ment have the same effect upon each individual in each genera-
tion. Dr. Pearl's error, as he has since recognized, lies in the
fact that in the Pj generation random mating involves only
brothers and sisters, while in all subsequent generations it also
involves other relationships.

When I read Dr. Pearl's article in October I naturally won-'
dered why there was such a difference of opinion between Dr.
East, Mr. Whiting and Dr. Pearl. Before finishing the article I
computed the amount of homozj'gosis in the Fj generation as 0
per cent. ; in Pj, 50 per cent. ; Pg, 50 per cent. ; P4, 62.5 per cent,
and P5, 68.25 per cent. As soon as possible after that I figured
other generations until the heterozygosis would be reduced to
one half of one per cent, of the maximum of heterozygosis in the

-•East and Hayes, 1912, loc. cit

^PMl. Trans. Boy, Soc. (A), Vol. 203, pp. 59 and 60.

Digitized by

Google

No. 576] SHORTER ARTICLES AND CORRESPONDENCE 761

Fj generation and found that this was accomplished in the F^,
generation, the amount of heterozygosis in each generation being:

Pe, 75.000 per cent. Fj,, 94.312 per cent. r», 98.710 per cent.

Ft, 79.687 per cent. Fj^, 95.398 per cent. Fa, 98.956 per cent.

Fs, 83.594 per cent F„, 96.277 per cent. Fa, 99.155 per cent.

F„ 86.719 per cent. F„, 96.988 per cent. Fa, 99.317 per cent.

Fio, 89.258 per cent. F,t, 97.563 per cent. F,», 99.447 per cent.

F,i, 91.309 per cent. Fi^, 98.029 per cent. F^, 99.553 per cent.

Fu, 92.969 per cent. F^, 98.405 per cent. Fa, 99.638 per cent.

With the approval of Dr. Castle and Dr. East I prepared to pub-
lish these figures.

Shortly after this Dr. Pearl wrote to Dr. East asking for an
opinion upon his article. Dr. East, in the meantime, by a method
diflfering from mine, had worked out the ratios independently.
Before answering Dr. Pearl's letter, however, Dr. East compared
his results with mine. They agreed. Dr. East then wrote to Dr.
Pearl, giving a short rebuttal of Dr. Pearl's arguments, enclosing
some of his own figures and adding that a student of Dr. Castle's
(myself) was thinking of publishing the complete figures. Dr.
Pearl immediately acknowledged his mistake and very gener-
ously asked if he should wait until I had published my article
before he published a correction. Dr. East replied that he could
see no reason for delaying the correction and advised me of this
reply.

Since it seemed proper for Dr. Pearl to correct his previous
article, I decided to withhold my own figures and incorporate
them later in a paper bearing also upon other matters. Dr.
Pearl's second article came out in the American Naturalist for
January, 1914, and this paper together with the third article in
the same journal for June, 1914, shows that his work was en-
tirely independent of Dr. East's or my own.

When Mr. Whiting asked me for a note giving the figures
showing what might be expected in the way of an automatic in-
crease in homozygosity when brothers were mated with sisters in
successive generations, as Mr. Whiting had done with his flies, I
naturally was pleased to have him accept my figures as correct-
ing his own, and at the same time give me an opportunity to ac-
knowledge my indebtedness to those who furnished the idea upon
which my figures were based.

H. D. Pish

BussEY Institution,
Forest Hn.LS, Mass.
August 18, 1914

Digitized by

Google

NOTES AND LITEEATUEE

MENDELIAN FLUCTUATIONS^

When the observed proportions, say of dominants and reces-
sives, in any Mendelian experiment are worked out for small
groups, such as individual litters or the seeds on individual plants
in individual fruits, considerable fluctuations round the expected
proportions may be observed. In the present note the magnitude
of these fluctuations is compared with the magnitude to be ex-
pected if the fluctuations were the result merely of chances of
sampling — corresponding to the fluctuations that would be ob-
served in drawing, say, samples of black balls from a bag con-
taining white and black balls in the proportion of 3 to 1. In so
far as there is good agreement, this is additional confirmation of
the Mendelian process holding good in its simplest form: if
the fluctuation observed is markedly greater than this theory
would indicate, some source of disturbance is certainly present,
but whether this disturbance arises from irregularities in the
distribution of the gametes or merely from extraneous circum-
stances (varying death-rates or diflSculties of sorting) can not,
of course, be determined from the data alone. For albinos in
individual litters of mice (Darbishire's data), and for numbers
of ** green'' or ** wrinkled'' in Mr. Bateson and Miss Killby's
crosses of peas I find exceedingly good agreement, at least if very
small plants are omitted. Lock's data for maize give good agree-
ment for the DR X DR cross, but poor agreement for the
DR X RR cross. Some data given me by Miss E. R. Saunders
for seed characters in the individual fruits of stocks show rather
irregular results. Further comparisons on similar lines would
be of interest, especially for the DR X RR cross, for which very
few data are available. For the case to afford a good test the
sorting should be clear and there should be nothing in the data
to suggest differential death rates obviously.

G. U. Y.

1 '* Fluctuations of Sampling in Mendelian Ratios," G. Udny Yule (Proc.
Cambridge Phil. Soc, XVII, 425).

76^

Digitized by

Google

INDEX

NAMES OF CONTRIBUTORS ARE PRINTED IN SMALL CAPITALS

Age, Mean, at Marriage pf Men
and Women, Variation and Cor-
relation, J. Arthur Harris and
RoxANA H. Vivian, 635

Alfalfa Breeding, Physiological
Correlations and Climatic Reac-
tions in, Geo. F. Freeman, 356

Allelomorphism, Multiole, W. E.
Castle, 383

Allelomorphs, Multiple, in Mice, T.
H. Morgan, 449; and Close Link-
age, The Theoretical Distinction
between, T. H. Morgan, 502; W.
E. Castle, 503

Allen, Glover M., Pattern Devel-
opment in Mammals and Birds,
385, 467, 550

Analysis of a Case of Continue as
Variation in Drosophila by a
Study of its Linkage Relations,
John S. Dexter, 712

Animals, and Plants, Sessile and
Motile, Comparison of Resnonses
of, Victor E. Shelpord, 641

Ants, Gynandromorphous, described
during the Decade 1903-1913,
William Morton Wheeler, 49

Apple, A Study of Variation in the,
W. J. Young, 595

Apterous Drosophila and its Genetic
Behavior, Charles W. Metz, 675

Arachnoidea, The Origin of X Cap-
sella Bursa-pastoris, Henri IIus,
193

Babcock, II. L., The Meadow Jump-
ing Mouse (Zapus Hudsonius)
especially regarding Hibernation,
485

Behavior, Genetic, An Apterous
Drosophila and its, Charles W.
Metz, 675

Bessey, Charles E., A Study of
Desert Vegetation, 639

Biology of the Thysanoptera, A.
Franklin Shull, '161, 236

Biometrics, Raymond Pearl, 505

Birds, and Mammals, Barriers to
Distribution, Joseph Grinnell,
248; Pattern Development in,
Glo\t:r M. Allen, 385, 467, 550

Breeding, Experiments with Grass-
hoppers, Nabours's, John S.
Dexter, 317; Alfalfa, Physiolog-

ical Correlations and Climatic
Reactions in, Geo. P. Freeman,
356

Bridges, Calvin B., The Chromo-
some Hvpothesis of Linkage ap-
" plied to Cases in Sweet Peas and
Primula, 52 i

Bristles, Heredity of, in the Com-
mon Green-bottle Fly, Phineas
W. Whiting, 339

British Islands, The Endemic Mam-
mals of the, T. D. A. Cockerell,
177

Calf, Double-headed, The Osteology
of, A. M. Reese, 701

Capsella Bursa-pastoris arachnoidea.
The Origin of X, Henri Hus, 193

Case, E. C, Restoration of Edaph-
osaurus cruciger Cope, 117

Castle, W. E., Some New Varieties
of Rats and Guinea-pigs and their
relation to Problems of Color In-
heritance, 65; Yellow Varieties of
Rats, 254; Nabours's Grasshop-
pers, Multiple Allelomorphism,
Linkage and Misleading Termin-
ologies in Genetics, 383; The
Theoretical Distinction between
Multiple Allelomorphs and Close
Linkage, 503

Castle and Phillips on the Variabil-
ity of Genes, The Bearing of the
Selection Experiments of, Her-
mann J. Muller, 567

Chandler, Asa C, The Effect of
Extent of Distribution on Specia-
tion, 129

Chromosome Hypothesis of Linkage
applied to Cases in Sweet Peas
and Primula, Calvin B. Bridges,
524

Citrus, Hybrids, Swingle on Varia-
tion in F„ and the Theory of Zy-
gotaxis, Orland E. White, 185;
Dr. Swingle's Experiments with,
another Hypothesis to account for,
A. C. and A. L. Hagedoorn, 446

Climatic Reactions in Alfalfa Breed-
ing, Geo. F. Freeman, 356

Cockerell, T. D. A., The Endemic
Mammals of the British' Islands,
177

763

Digitized by

Google

764

THE AMERICAN NATURALIST [Vol. XLVHI

Collins, G. N. and J. H. Kbmpton,
Inheritance of Endosperm Tex-
ture in Sweet X Waxy Hybrids of
Maize, 584

Color Inheritance, Some New Vari-
eties of Bats and Guinea-pigs,
and their Eelation to Problems of,
W. E. Castle, 65

Cook, O. F., Terms relating to Gen-
eric Types, 308

Correlation, and Variation in the
Mean Age at Marriage of Men
and Women, J. Arthur Habbis
and EoxANA H. Vivian, 635

Correlations, Physiological, and Cli-
matic Eeactions in Alfalfa Breed-
ing, Geo. F. Fbeeman, 356

Correspondence and Shorter Articles,
57, 122, 177, 248, 308, 383, 446,
491, 567, 635, 693, 759

Davis, Bradley Moore, Stomps 's
(Enothera biennis L., 498

Desert Vegetation, A Study of,
Charles E. Besset, 639

Dexter, John S., Nabours's Breed-
ing .Experiments with Grasshop-
pers, 317; The Analysis of a Case
of Continuous Variation in Dro-
sophila by a Study of its Linkage
Relations, 712

Differential Mortality with Bespect
to Seed Weight occurring in Field
Cultures of Pisum sativum, J.
Arthur Harris, 83

Discontinuity in Variation, Galton
and, R. Ruggles Gates, 697

Distribution, The Effect of Extent
of, on Speciation, Asa C. Chand-
ler, 129; Barriers to, as regards
Birds and Mammals, Joseph
Grinnell, 248; A Study of Fact-
tors governing, Phineas E. Whit-
ing, 339

* * Dominant ' ' and ' * Recessive ' '
Spotting in Mice, C. C. Little, 74

Drosophila, The Reduplication Hy-
pothesis as applied to, A. H.
Sturtevant, 535; An Apterous,
Charles W. Metz, 675; The
Failure of Ether to produce Mu-
tations in, T. H. Morgan, 705;
Analysis of a Case of Continuous
Variation in, its Linkage Rela-
tions, John S. Dexter, 712

Duplicate vienes, Sewall Wright,
638

East, E. M., and H. K. Hayes, A
Genetic Analysis of the Changes
produced by Selection in Experi-
ments with Tobacco, 5

Ectoparasites of Mammals, Vernon
LvMAN Kellogg, 257

Edaphosaurus cruciger Cope, Res-
toration of, E. C. Case, 117

Emerson, R. A., The Inheritance of
a Recurring Somatic Variation in
Variegated Ears of Maize, 87

Endemic Mammals of the British
Islands, T. D. A. Cockerell, 177

Endosperm Texture in Sweet X
Waxy Hybrids of Maize, Inherit-
ance of, G. N. Collins and J. H.
Kempton, 584

Environmental Work, Humidity — a
Neglected Factor, Frank E.
LUTZ, 122

Errors, Probable, A Short-cut in the
Computation of Certain, Howard
B. Frost, 696

Ether, The Failure of, to produce
Mutations in Drosophila, T. TL
Morgan, 705

Evolution, and Taxonomy, X, 369

Fish, H. D., On the Progressive In-
crease of Homozygosis Brother X
Sister Matings, 758

Fluctuations, Mendelian, G. Udny
Yule, 764

Fly, the Common Green-bottle, He-
redity of Bristles in, Phineas
W. Whiting, 339

Formulae for the Results of Inbreed-
ing, H. S. Jennings, 693

Freeman, Geo. F., Physiological
Correlations and Climatic Reac-
tions in Alfalfa Breeding, 356

Frost, Howard B., A Short-cut In
the Computation of Certain Prob-
able Errors, 696

Gates, R. Ruggles, Galton and Din-
continuity in Variation, 697

Generic Types, Terms relating to, O.
F. Cook, 308

Genes, The Bearing of the Selection
Experiments of Castle and Phil-
lips on the Variability of, Her-
mann J. MuLLER, 567; Duplicate,
Sewall Wright, 638

Genetic, Analysis of the Chang^
produced by Selection in Experi-
ments with Tobacco, E. M. East
and H. K. Hayes, 5; Behavior of
an Apterous Drosophila, Charles
W. Metz, 675

Genetics, Misleading Terminologies
in, W. E. Castle, 383

Gerould, John H., Species-bmlding^
by Hybridization and Mutation,
321

Digitized by

Google

No. 576]

INDEX

765

Grasshoppers, Nabours 's Breeding
Experiments with, John S. Dex-
TEE, 317

Gregory's Tetraploid Primulas, A
New Mode of Segregation in,

HeEMANN J. MULLEE, 508

GsiNNELL, Joseph, Barriers to Dis-
tribution as regards Birds and
Mammals, 248

Guinea-pigs, and Bats, Some New
Varieties and their Belation to
Problems of Color Inheritance, W.
E. Castle, 65

GuLiCK, John T., Isolation and Se-
lection allied in Principle, 63

Gynandromorphous Ants, described
during the Decade 1903-1913,
William Moeton Whbeleb, 49

Hagedcorn, a. C. and A. L.,
Another Hypothesis to account for
Dr. Swingle's Jiixperiments with
Citrus, 44b

Hagedoobn, a. L., Eepulsion in
Mice, d9y

Harris, J. Arthur, On Differential
Mortality with respect to Seed
Weight occurring in Field Cul-
tures of Pisum sativum, 83; and
EoxANA H. Vivian, Variation and
Con elation in the Mean Age at
Marriage of Men and Women, 635

Ha^es, n. K., and E. M. East, A
Genetic Analysis of the Changes
produced by Selection in Experi-
ments with Tobacco, 5

Heredity, and **The Influence of
Monarchs," V. L. K., 255; of
Bristles in the Common Green-
bottle Fly, Phineas W. Whiting,
339

Hibernation, Notes on the Meadow
Jumping Mouse, L. H, Babcock,
485

Homozygosis Brother-Sister Ma-
tings, The Progressive Increase
of, H. D. Fish, 758

Humidity — a Neglected Factor in
Environmental Work, Frank E.
LUTZ, 122

Hus, Henri, The Origin of X Cap-
sella Bursa-Pastoris Arachnoidea,
193

HuTCHESON, T. B., Thirteen Years
of Wheat Selection, 459

Hybridization and Mutation, Spe-
cies-building by, John H. Ger-
OULD, 321

Inbreeding, a Mendelian Population,
Raymond Pearl, 57; Notes on.

Baymo^^d Pearl, 491; and Bela-
tionship Coefficients, Baymond
Pearl, 513; Formulae for the Be-
suits of, H. 8. Jennings, 693

Inheritance, Color, Some New Va-
rieties of Bats and Guinea-pigs
and their Belation to Problems
of, W. E. Castle, 65; of a Becur-
ring Somatic Variation in Varie-
gated Ears of Maize, B. A. Emer-
son, 87; Sex-limited and Sex-
linked, T. H. Morgan, 577; of
Endosperm Texture in Sweet X
Waxy Hybrids of Maize^ G. N.
Collins and J. H. Kempton, 584

Internal Belations of TerrestriiQ As-
sociations, Arthur G. Vestal,
413

Isolation and Selection allied in
Principle, John T. Gulick, 63

Jennings, H. S., Formulae for the
Besults of Inbreeding, 693

K., V. L., Woods on Heredity and
**The Influence of Monarchs,"
255

Kellogg, Vernon Lyman, Ectopar-
asites of Mammals, 257

Kempton, J. H; and G. N. Collins,
Inheritance of Endosperm Tex-
ture in Sweet X Waxy Hybrids
of Maize, 584

Linkage, in the Silkworm Moth, A..
H. Sturtevant, 315; and Mis-
leading Terminologies in Genetics,
W. E. Castle, 383; Close, and
Multiple Allelomorphs, T. H.
Morgan, 512; W. E. Castle, 503;
Chromosome Hypothesis of, ap-
plied to Cases in Sweet Peas and
Primula, Calvin B. Bridges, 524;
Belations, Analysis of a Case of
Continuous Variation in Dro-
sophila by a Study of its, John
S. Dexter, 711

Literature and Notes, 185, 255, 315,
505, 639, 762

Little, C. C, '* Dominant" and
"Becessive'* Spotting in Mice, 74

Lucilia sericata Meig., the Com-
mon Green-bottle Fly, Heredity
of Bristles in, Phineas W. Whit-
ing, 339

LuTZ, Frank E., Humidity — a Ne-
glected Factor in Environmental
Work, 122

Maize, The Inheritance of a Becur-
ring Somatic Variation in Varie-
gated Ears of, B. A. Emerson,

Digitized by

Google

766

THE AMERICAN NATURALIST [Vol. XLVHI

87; Inheritance of Endosperm
Texture in Sweet X Waxy Hy-
brids of, G. N. Collins and J.
H. Kempton, 584

Mammals, The Endemic, of the Brit-
ish Islands, T. D. A. Cockeeell,
177; and Birds, Barriers to Dis-
tribution, Joseph Grinnell, 248 ;
Ectoparasites of, Vebnon Lyman
Kellogo, 257; and Birds, Pat-
tern Development in, Glovee M.
Allen, 385, 467, 550

Marriage of Men and Women, Vari-
ation , and Correlation in the
Mean Age at, J. Arthur Harris
and RoxANA H. Vivian, 635

Matings, Homozygosis Brother X
Sister, The Progressive Increase
of, H. D. Fish, 759

Meadow Jumoing Mouse (Zapus
Hudsonius) especially regarding
Hibernation, H. L. Babcock, 485

Mendelian, Population, On the Re-
sults of inbreeding a, Raymond
Pearl, 57; Fluctuations, G.
Udny Yule, 762

Metz, Charles W., An Apterous
Drosophila and its Genetic Be-
havior, 675

Mice, ** Dominant" and ''Recess-
ive," Spotting in, C. C. Little,
74; Multiple Allelomorphs in, T.
H. Morgan, 449; Repulsion in, A.
L. Haoedoorn, 699

"Monarchs, The Influence of," and
Heredity, V. L. K, 255

Morgan, T. H., Multiple Allelo-
morphs in Mice, 449; The Theo-
retical Distinction between Mul-
tiple Allelomorphs and Close Link-
age, 502; Sex-limited and Sex-
linked Inheritance, 577 ; The Fail-
ure of Ether to produce Mutations
in, 705

Mortality, Differential, with respect
to Seed Weight occurring in Field
Cultures of Pisum sativum, J. Ar-
thur Harris, 83

Moth, Silkworm, Linkage in the,
A. H. Sturtevant, 315

Motile and Sessile Plants and Ani-
mals, A comoarison of the Re-
sponses of, Victor E. Shelpord,
641

Mouse, Meadow Jumping (Zapus
Hudsonius), especially regarding
Hibernation, H. L. Babcock, 485

Muller, Hermann J., A New Mode
of Segregation in Gregory's Tet-
raploid Primulas, 508; The Bear-
ing of the Selection Experiments

of Castle and Phillips on the
Variability of Genes, 567

Multiple, Allelomorphism, W. E.
Castle, 383; Allelomorphs in
Mice, T. H. Morgan, 449; Allelo-
morphs and Close Linkage, T. H.
Morgan, 502; W. E. Castle, 503

Mutation, and Hybridization, Spe-
cies-building by, John H. Gebt
OULD, 321

Mutations, Parallel, in CEnothera
biennis L., Theo. J. Stomps, 494 ;
Bradley M. Davis, 498; in Dro-
sophila, The Failure of Ether lo
Produce, T. H. Morgan, 705

Nabours 's, Breeding Experiments
with Grasshoppers, John S. Dex-
ter, 317; Grasshoppers, W. E.
Castle, 383

Notes and Literature, 185, 255, 315,
505, 639, 762

CEnothera biennis L., Parallel Muta-
tions in, Theo. J. Stomps, 494;
Bradley M. Davis, 498

Osteology of a Double-headed Calf,
A. M. Reese, 701

Pattern Development in Mammals
and Birds, Glover M. Allen, 385,
467, 550

Pearl, Raymond, On the Results of
Inbreeding a Mendelian Popula-
tion, 57; Studies on Inbreeding,
491; Biometrics. 505: Inbreeding
and Relationship Coefficients, 513

Phillips, and Castle, The Bearing of
Selection Experiments on the
Variability of Genes, Hermann
J. Muller, 567

Physiological Correlations and Cli-
matic Reactions in Alfalfa breed-
ing, Geo. F. Freeman, 356

Pisum sativum. Differential Mortal-
ity with Respect to Seed Weight
occurring in Field Cultures of, J.
Arthur Harris, 83

Plants and Animals, Sessile and Mo-
tile, A Comparison of the Re-
sponses of, Victor E. Shelford,
641

Primulas, Tetraploid, Gregory's, A
New Mode of Segregation in,
Hermann J. Muller, 508

Rats and Guinea-pigs, Some New
Varieties of, and ttieir relation to
Problems of Color Inheritance, W.
E. Castle, 65; Yellow Varieties
of, W. E. Castle, 254

''Recessive,*' and ** Dominant,"

Digitized by

Google

No. 576]

INDEX

767

Spotting in Mice, C. C. Little, 74
Beduplication Hypothesis as applied

to Drosophila, A. H. Stubtevant,

535
Reesk, a. M., The Osteology of a

Double-headed Calf, 701
Regeneration, Variation and Corre-
lation in Thyone, John W, Scott,

280
Repulsion in Mice, A. L. Haoe-

DOORN, 699
Responses of Plants and Animals,

Sessile and Motile, A Comparison

of the, Victor E. Shelpobd, 641
Restoration of Edaphosaurus cru-

ciger Cope, E. C. Case, 117

Scott, John W., Regeneration,
Variation and Correlation in
Thyone, 280

Seed Weight occurring in Field Cul-
tures of Pisum sativum. Differen-
tial Mortality with Respect to, J.
Arthur Harris, 83

Segregation, A New Mode of, in
Gregory's Tetraploid Primulas,
Hermann J. Muller, 508

Selection, in Exneriments with To-
bacco, Genetic Analysis of the
Changes, E. M. East and H. K.
Hayes, 5; and Isolation, allied in
Principle, John T. Gulick, 63;
Wheat, Thirteen Years of, T. B.
Hutcheson, 459; Experiments of
Castle and Phillips, bearing on
the Variability of Genes, Her-
mann J. Muller, 567

Sessile and Motile Plants and Ani-
mals, A Comparison of the Re-
sponses of, Victor E. Shblford,
641

Sex-limited and Sex-linked Inherit-
ance, T. H. Morgan, 577

Shelpord, Victor E., A. Comparison
of the Responses of Sessile and
Motile Plants and Animals, 641

Short-cut in the Computation of Cer-
tain Probable Errors, Howard B.
Frost, 696

Shorter Articles and Correspond-
ence, 57, 122, 177, 248, 308, 383,
446, 491, 567, 635, 693, 769

Shull, a. Franklin, Biology of the
Thysanoptera, 161, 236

Silkworm Moth, Linkage in the, A.
H. Sturtevant, 315

Speciation, The Effect of Extent of
Distribution on, Asa C. Chand-
ler, 129

Species-building by Hybridization

and Mutation, John H. Gerould,
321

Spotting in Mice, "Dominant''
and ** Recessive, " C. C. Little,
74

Stomps, Theo. J., Parallel Muta-
tions in CSnothera biennis L., 494

Stomps 's CSnothera biennis L.,
Bradley M. Davis, 498

Sturtevant, A. H., Linkage in the
Silkworm Moth, 315; The Re-
duplication Hypothesis as applied
to Drosophila, 535

Sweet Peas and Primula, Chromo-
some Hypothesis of Linkage ap-
plied to Cases in, Calvin B.
Bridges, 524

Swingle on Variation in Fi Citrus
Hybrids and the Theory of Zygo-
taxis, ORLANiy E. White, 185

Swingle's Experiments with Citrus,
Another Hypothesis to account
for, A. C. and A. L. Hagedoorn,
446

Taxonomy and Evolution, X, 369
Terrestrial Associations, Internal

Relations of, Arthur G. Vestal,

413
Thyone, Regeneration, Variation

and Correlation in, John W.

Scott, 280
Thysanoptera, Biology of the, A.

Franklin Shull, 161, 236
Tobacco, A Genetic Analysis of the

Changes produced by Selection in

Experiments with, E. M. East

and H. K. Hayes, 5
Types, Generic, Terms relating to,

O. F. Cook, 308

Variability of Genes, The Bearing
of the Selection Experiments of
Castle and Phillips on the, Her-
mann J. Muller, 567

Variation, in Variegated Ears of
Maize, The Inheritance of a Re-
curring Somatic, R. A. Emerson,
87; in F, Citrus Hybrids and the
Theory of Zygotaxis, Swingle on,
Orland E. White, 185; Regen-
eration and Correlation in Thy-
one, John W. Scott, 280; in the
Apple, W. J. Young, 595; and
Correlation in the Mean Age at
Marriage of Men and Women, J.
Arthur Harris and Roxana H.
Vivian, 635; Galton and Discon-
tinuity in, R. RuQGLEs Gates,
697; Continuous, Analysis of a
Case in Drosophila by a Study of

Digitized by

Google

768

THE AMERICAN NATURALIST [Vol. XLVHI

its Linkage Belations, John S.
Dexter, 712

Vegetation, Desert, Charles E.
Besset, 639

Vestal, Arthur G., Internal Eola-
tions of Terrestrial Associations,
413

Vivian, Boxana H., and J. Arthur
Harris, Variation and Correla-
tion in the Mean Age at Mar-
riage of Men and Women, 635

Wheat Selection, Thirteen Years of,
T. B. Hutcheson, 459

Wheeler, William Morton, Gy-
nandromorphons Ants described
during the Decade 1903-1913, 49

White, Orland E., Swingle on
Variation in F, Citrus Hybrids
and the Theory of Zygotaxis, 185

Whiting, Phineas W., Heredity
of Bristles in the Common Green-

bottle Fly, Study of Factora
Governing Distribution, 339

Woods on Heredity and the "In-
fluence of Monarchs," V. L. K,,
255

Wright, Sewall, Duplicate Genes,
638

X, Taxonomy and Evolution, 369
X Capsella Bursa-pastoris arach-

no^dea, The Origin of, Henri

Hus, 193

Yellow Varieties of Rats, W. E.
Castle, 254

Young, w. J., A Study of Varia-
tion in the Apple, 595

Zygotaxis, Swingle on Variation in
F, Citrus Hybrids and the Theory
of, Orland E. White, 185

Digitized by

Google

Two Important Biological Journals

The Journal of Qenetics

Edited by
W. BATESON and R. C. PUNNETT
A periodical /or the pablication of records of original research in heredity^ variation, and allied topics,
The Journal also contains, from time to time, articles summarizing the existing state of knowledge in the
Tarions branches of genetics. The numbers are fullv illustrated, and, when the subject-matter demands it.
oolored plates are freely employed. The Journal is issued in parts as material accumulates, and a volume
(appearing, so far as possible, annually) consists of four snob parts. Subscription price, $7.50 a volume ;
single copies, 12.60.

Biometrik:a

Edited by EARL PEARSON
A journal for the statistical study of biological problems, appearing about four times a year. A volume
ooutaining about 500 pages, with plates and tables, is issued annually. Volume X b^ns in 1914. The
nine preceding volumes form a veritable encyclopaedia of biological stiu^tics in the fields of anthropometry,
orauiometry, medicine, heredity, etc. Subscription price, $7.50 a volume ; single copies, 12.60.

Prices on back numbers of these journals vary, and postage from London is charged on back volumes
and single copies.

Published in America fof the Cambridge University Press of England by

The University of Ohicasfo Press

ChicasTO^ Illinois
University Oontrol

By J. MoEIbbn Cattbll, Professor of Psychology in Colombia Uniyenity

Tofstber with a series of Two Hundred and ^nnety-nine Unsigned Letters by Leading Men ci Bdmum
holding Academic Positions and Articles by Joseph Jastbow, GnoBan T. Lads, Jomr J. SrBimnov, J. B.
Cbkobtoh, J. MoEsnr Cattkll, Qbobob M. Stbattov, Stbwabt PATOVt Joks Jat Ghafmaw, Jaios P.
limrBon and Jacob GouLp Schubmav.

A great yariety of Questions oonoeming geoenJ onivenity administration are dealt with in an original and
beipful way^ — Naiuirt.

These quotations and eftamples are taken from Professor Gattell's Informed and thorough disensrfon ci tbs
■Qbjeot of unirenity oontrol, a subjeet upon which he has had much to say of late, finding occasion for canstie
eritidsm of eiisting American conditions, and standing as the champion of an academic demootaey and a tead»-
Ing profession upon which a man may enter without forfeiting his self-respect. — Th$ DidL

Sentences and paragraphs that betoken the expert, highly-trained mind, the suggestions that come to v^
fresh a d tell us that a new day is about to dawn in educational writing. — Th$ Bot/Um Svmnmo TramteripL

SCIBIVCB AND EDUCATION

A aeries of volumes for Ute promotion ofseienU&e resesreh and educational progress
YOLUMB L The Foundations of Science. By H. PoniaABi. Gontalning the authorised En^ish translatloB

by Gbobob Bbuob Halstbo of "Science and Hypothesis," "The Value of Science," and "Sdenoa

and Method." Price, $3.00 net.
YOLUME IL Medical Research and Bdocatlon. By Riobabd M. Pbabob, William H. Wblob. W. H.

HOWBLL, FbaMKLIH P. MaLL, LbWBLLTS F. BaBKBB, GBABLBa S. MlHOT, W. B. GAHVOVt W. T.

GomioiuiAir, Thbobalb Smith, G. N. Stbwabt, G. M. Jaoxbon, E. P. Ltoh, Jambs B. HaBBioK, Joboi
M. DoDeoN, C. R. Babdbbn, W. OphOls, S. J. Mkaibb, Jambs Ewnio, W. W. Kbbx, Bmamx H. Dobald-
BOB, Ghbibtiab a. Hbbtbb, and Hbnbt P. BowniroH. Price, $3.00 net.
YOLUBiEin. Vnirenity Control. NamBtads. Paget x+48i. Prie$,tSMnd.

aARRi30N, N. Y. THE SCIENCE PRESS Lancaster, pa.

SUB-STATION 84, NEW YORK CITY

Digitized by

Google

The Americsuti Naturalist

A Ifomiilr loonuJ. •stablidiMl in 1M7. Derotod to th* Adnne^muA ci th* Biological SctecM

CONTENTS OF THE JUNE NUMBER

tbmot John H. GerouM.
Heredity of Brisaes In the Oomnum Greenbottto Flj—

A Stadj of Factors gOTemingDistribation. PhlneM

W. Whiting.
FhyslologlGal Conelationfl and GUmatio BeMtiona in

AlfUik Breeding. Geo. F. Freeman.
Taxonomy and Erolation. By X.

Mialeading Terminologlee In Genetica. Proftesor
W.RCaatle.

CONTENTS OF THE JULY NUMBER

Pattern Dereiopment in Maaimahi and Birds.
GloTer M Allen

Arthur G.VestaL
Shorter Articles and Discussion* Another Hy-
pothesis to Account for Dz Swingle's Bxperi-
meatf with Citrus. A. a and A. L. Hagedoora,

CONTENTS OF THE AUGUST NUMBER

Homple AUekmorpha in Mice. ProfcMor T. H

Morgan.
Thirteen Tewni of Wheat Selection. T. B. HoUh-

eaon.
Pattern Deyelopmtnt in Manunala and Birda.

Glover M. Allen.
The Meadow Jamplsff Mooae. Dr. H. L. Babooek.
Shorter Articles and DiacuaBion :

Parallel MntaUona in (Enothera biennis L. Dr.
J. Stomps. Dr. Bradley M. Dayia. The Theoret*
loiU Distinction between Mnltiplt Allelomorphs

Notea and literature: ^
Biometrica. Dr. Raymond Peail. A New Mod-
of Segregation in Gregory 's Tetrapioid Prtmnlasa
Hermann J. Mailer.

CONTENTS OF THE SEPTEMBER NUMBER

Studies on Inbreeding. Dr. Baymond Pearl.

The Chromosome Hypothesis of Linkage applied to

Cases in Sweet Peas and Primula^ Calrin B.

Bridget.
TheRedupUcatlon Hypothesis as applied to Droso-

phila. Dr.A.H.Sturtevant.
Pattern Dereiopment in Mammals and Biida. Dr.

GloTer M.Allen.
Shorter Articles and Correspondcnoe :

The Bearing of the Selection Experiments of

Castle and Phillips on the VftriabUity of Genet.

Hermann J. Muller.

CONTENTS OF THE OCTOBER NUMBER

8ez-limited and Sez-Onked Inherltanoe. Professor

T. H. Morgan.
Inheritance of Endosperm Texture in Sweet x Waxy

Hybrids of Maize. Q. N. Collins and J. H. Kemp.

ton.
A Study of Variation in the Apple. W. J. Tonng.
Shorter Articles and Discussion :

Variation and Correlation in the Mean Age at

Marriage of Men and Women. Dr. J. Arthur

Harris, Boxana H. Vivian. Duplicate Genes.

Sewall Wright.
Notes and Literature :

A Study of Desert Vegetetion. Professor Charles

E.Bes8ey.

CONTENTS OF THE NOVEMBER NUMBER

A Comparison of the Bcsponses of Sessile and Motite
Plants and Animals. PrefessorVictorB.8helfofd.

An Apterous Drofophila and its Genetic BehaTlor.

Charles W.Mets.

Shorter Articles and Discussion :

Formulse for the Results of Inbreeding. Ptofessog

ol Certain Probable Errors. Howard B. Frost.
Galton and Discontinuity in Variation. Dr. R.
Buggies Gates. The Osteology of a Double-
headed Call Professor A. M. Btese.

Single Number 40 Cents Teariy Subicriptiom $4i>0

THE SCIENCE PRESS

<Urri»o, N. Y. Substation 84 : NEW YORK Uiieut.r, P.:

Digitized by

Google

Digitized by

Google

Digitized by

Google

Digitized by

Google

)

Digitized by

Googl^

'7i
■1

Digitized by

Google

Digitized by

Google

Digitized by

Google

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "The American naturalist"

See other formats