Skip to main content

Full text of "Landscapes of memory : a history of social identity in the Western Serengeti, Tanzania"

See other formats








Copyright 1998 


Jan Bender Shetler 


If it takes a village to raise a child,' it takes a world community to produce a dissertation. 

I am grateful for the institutional support I received during the course of my graduate 
studies. This research was assisted by a grant from the Joint Committee on African Studies of the 
Social Science Research Council and the American Council of Learned Societies with funds 
provided by the Ford, Mellon, and Rockefeller Foundations. I also received a research grant from 
the Institute of International Education under the U.S. Fulbright Student Program (1995-96). I 
would like to thank the History Department at the University of Florida for the support received 
over the years, including a write-up grant from the College of Liberal Arts and Sciences, a research 
assistantship and a teaching assistantship. The African Studies Center also aided my work through 
three years of Title VI, FLAS funding, a study carrel and an atmosphere of creative 
interdisciplinary interaction. 

I would like to express my sincere gratitude to the Government of Tanzania for permission 
to do research in the country under the auspices of the Tanzania Commission for Science and 
Technology and the History Department of the University of Dar es Salaam. At the University of 
Dar es Salaam members of the History Department were always generous with their time and 
support. Special thanks to Dr. Fred J. Kaijage, Dr. Bertram Mapunda, Dr. Rugatiri D. K. 
Mekacha, Dr. B. Itandala, Dr. I. N. Kimambo and Dr. Nestor Luanda. 

My deepest appreciation is extended to my professors who have given so much of their 
time and intellectual inspiration to my work. I extend heartfelt thanks to Steve Feierman who has 
fundamentally formed the way I think about Africa and always pushed me toward excellence. 

Thanks also to David Schoenbrun for doing what it took to get me through this with compassion 
and intelligence; to R. Hunt Davis Jr. for advising, editing and encouraging; and to the rest of my 
committee, Jeffrey D. Needell, Kathryn J. Burns and Allan F. Bums for their continuing support. I 
would also like to thank Richard Waller for his interest and ideas over the years. My study at the 
University of Florida was enriched by a community of scholars who demonstrated a cooperative 
and supportive spirit-including Holly Hanson, Tracy Baton, Marcia Good Maust, James Ellison, 
Edda Fields, Todd Leedy, Kym Morrison, Keri Schmit, Dianne Oyler, Rebecca Gearhardt, 
Kearsley Stewart and Rob Ellers. 1 could not have done it without Holly's encouragement and 
reading of many drafts. While I was in Dar es Salaam I enjoyed the support of a wider community 
of scholars at TYCS including, Kathleen Smythe (who also read and edited), Anne Stacie Canning 
Colwell, Patrick Malloy, Frances Vavrus, Lisa Richey, Stefano Porte, Beth Pratt, Simon Heck and 
Ludo Bok from whom I learned and was enriched. 

In Tanzania numerous people aided my work, gave me hospitality and good conversation. 
Thanks to Mwalimu Nyamaganda Magoto for all the time he spent going over thousands of 
cultural vocabulary words in Nata; to Susana Nyabikwabe Mayani for teaching me Nata and being 
my friend; to Adija Sef for her friendship, quick laugh and care of my house and children; to 
Mayani Magoto for his work in interviews, research on various topics and interest in the whole 
process; to Goko Kimori for coming over nearly every day to find out if 1 had learned Nata yet and 
staying to chat; to Morigo, Magoto Nyawagamba, Magoto, Nyangere and Marehemu Nyambura 
Mayani for bringing joy into the days in Bugerera; to Nyangere Faini for helping me with 
transcription; to Faini Magoto and Joseph Magoto for always having a meal and a place to rest 
when I needed it, for all of their hospitality; to Mzee Mswaga for being our community eyes and 
ears and thatching our roof; to Susan Godshall for typing the Ngoreme Dictionary; to Ausgustino 
Mokwe Kisigiro for lending me his Nata dictionary and tapes; to Padre James Eblin, Maryknoll 

Missioner from the Ishenyi mission, for his kindness and friendship, for making contacts in Ishenyi 
and especially for twice rescuing me when the car broke down; to the Tanzania Mennonite Church 
people in Nyabange and Shirati who always extended the hand of hospitality and support; to 
Bishop Solomon S. Buteng'e, Bishop Joram Mbeba and Marehemu Bishop Naftali Birai who were 
always interested in the work and allowed us to lease a church car; to David and Justine Foxall for 
their kind hospitality in Dar es Salaam; and to Brian Farm and Bethany Woodward for their 
hospitality and friendship at Seronera. 

1 feel a deep debt of gratitude to Nyawagamba Magoto who extended the right hand of 
fellowship, invited us into his home and embraced the research as if it was his own. The entire 
Magoto family, the brothers and sister of Nyawagamba and their children, adopted us and cared 
for us as family. I am also grateful for all the work and dedication of those who acted as my 
colleagues to arrange interviews, translate, make introductions and serve as cultural translators. 
Without their help my work would not have been possible. They include Kinanda Sigara, Mayani 
Magoto, Nyamaganda Magoto, David Mganya Masama, Pastor Wilson Shanyangi Machota, and 
Mnada Joseph Mayonga, all of whom also provided me with hospitality and good friendship. 
Mnada rescued me from a car breakdown on a rainy night with his team of oxen and his wife 
always had a pot of milky tea to cheer me up. I was also aided in arranging interviews by Rhoda 
Koreni, Yohana Wambura, Kennedy Sigira, Philemon Mbota, Thomas John Kazi, D. M. Sattima 
and Ibrahimu Matatiro Kemuhe. Zedekia Oloo Siso is a fellow historian and gave me access to his 
work in the Luo-speaking area of the region. Thanks to Glen and Elin Brubaker for facilitating 
that exchange. 

So many other people facilitated aspects of the research. On the trip to Sonjo Michael 
Wambura Machambire gave up his own business to accompany me, Ndelani Sanaya introduced 
me to his home village, and the chairman of the village Emanuel G. Goroi graciously hosted our 

group without prior warning. Father Ambrose Chacha helped me with the Nyegina records. G. M. 
Kusekwa with the S. D. A. records at the highschool in Ikizu, and the archivist in Musoma, 
Fredrick Semkiwa helped to locate files there. In Mugumu I was always grateful for the hospitality 
of our friends from Missions Moving Mountains (Paul and Stephanie, Jim and Joanne, and Cabot 
and Betsy) and the Mennonite Church (Daniel and Prisca Machota, and Wilson and Esta 
Machota). In Nyabange the door was always open from Harold and Christine Wenger, Arlin and 
Velma Schrock, Rachel lgira, Perusi Kyambiriya, Deus and Salome Tumbago, Marehumu Nyerere 
Itinde and his family, and many, many others. In Nyabasi we received welcome from Elva Landis 
and Elizabeth Birai; in Iramba from Father Charles Mwiguta; in Kisaka from Pastor Zakayo 
Jandwa and his family; in Mwanza from Juliana Matoha Magoto; in Bujora from Joseph 
Sungulile, Jefta Kishosha and Mark Bessire; and in Nyegezi from Bhoke Magoto. 

My appreciation also goes out to my church family at Emmanuel Mennonite Church in 
Gainesville, who provided the community our family needed in Florida, and at First Baptist Church 
in Dove Creek, who did the same for us when we arrived as strangers in Colorado. Mennonite 
Central Committee and particularly our many colleagues in the organization got me started on the 
quest for understanding in Africa. 1 am grateful for my training at Goshen College under Theron 
Schlabach, John Oyer, Alan Kreider, and James Hertzler who prepared me for this work in more 
ways than they might know. Mary Oyer and Robert Kreider remain models of scholarship and 
thoughtful living. 

My extended family cheered me on and gave me what 1 needed to run the full race. I am 
thankful to Dad and Mom for all their love and support in a thousand ways during this time, to 
Linda for helping me to keep my chin up, and to Jon for all his help with photography and video on 
the field. My Shetler family (Luther, Geneva, Jo, Stan, Terry, Bonnie, Lu, Tami, and Mick) has 
always believed 1 could do this and helped me not to doubt, too. However, my biggest debt of 

gratitude is to my husband, Peter, and to my children, Daniel and Paul, who suffered and rejoiced 

with me through it all. Their contentment and joy in living in rural Tanzania allowed me the 

freedom and security to do my own work. Paul knows and enjoys the stories as much as 1 do, he 

baked bread, gave me foot massages after weary days of walking and greeted all the old ladies 

politely. Daniel was my source of good sense and willing help, his cultural sensitivity and 

flexibility were appreciated even as he was growing into a teenager. Peter fixed cars, computers 

and broken hearts, he did my maps and graphics, supported and loved me in every way, and 

provided good humor and perspective on the things that really matter. To them I give back my 

dreams, in which they also share, in loving gratitude: 

Had I the heavens ' embroidered cloths, 
Enwrought with golden and silver light, 
The blue and the dim and the dark cloths 
Of night and light and the half-light, 
I would spread the cloths under your feet: 
But 1, being poor, have only my dreams; 
I have spread my dreams under your feet; 
Tread softly because you tread on my dreams. 
William Butler Yeats, 1899 






ABSTRACT xx j x 




Oral Traditions in Space and Time 5 

Oral Traditions and Social Identity 28 

Research Methodology 34 






The Social Spaces of Historical Knowledge 56 

The Colonial Contexts of Men's Historical Narratives 63 

Women's Intimate Community Knowledge 81 

Secret Knowledge ] 00 

Conclusion ] 07 



Rupture in Social Identity at the Time of Crisis 116 

Ecological Breakdown 1 23 

Indirect Effects of the Caravan Trade 133 

Maasai Raids ..137 

Conclusion 141 





The Development of a Regional Culture from Diverse Interactions in the Distant 

Past 146 

The Nata Emergence Story: The Union of Hunters and Farmers 164 

Social Reproduction and Homestead Space 169 

The Ikizu Emergence Story: The Division of Authority 185 

Conclusion 191 



The Ecological Landscapes of Interaction on the Frontier 194 

Hills as Zones of Interaction 203 

Relations with Other Hill Farmers: The Sonjo Connection 213 

Relations with Hunter-Gatherers: Learning How to Live on the Land 218 

Relations with Plains Herders: The Silent Texts 230 

Conclusion 245 



Clans as Children of the Homestead 249 

Clan Histories as Pathways of Regional Association 268 

Praise Names and Prohibitions 291 

Prophecy as the Specialist Knowledge of Clans 303 

Conclusion 312 




The Spaces of Important Places 317 

The Living Dead 325 

Ritual Possession of the Land 338 

Nineteenth Century Settlement Patterns 347 

Continuity and Relationship to the Land in the Context of Mobility 362 

Conclusion 369 



Generation-Set Ritual and the Middle-Time Frame 376 

Age-Organization 380 

The Rituals of Healing the Land and the Retirement Ceremony 395 

Interpreting the Walk 410 

Territories and Boundaries 422 

Conclusion 438 




Age-Set Reorganization 444 

Settlement Reorganization 453 

Resistance and Accommodation to Maasai Hegemony 459 

Age-set Territories: The Bounded Space of Saiga 480 

Continuities in Strategies for Coping with Disaster 490 

Conclusion 494 



Ethnic Formation 497 

Recovery from the Disasters and the Larger Context of Ethnicity 508 

The Emergence of New Wealth 528 

Colonial Rule and Changing Forms of Social Identity 547 

Conclusion 560 






Oral Sources: Formal Interviews 576 

Unpublished Archival Sources 585 

Published Sources, Dissertations/Theses and Conference Papers 589 



Figure page 

F-l . Topics for the Study of Nata History xxiv 

F-2. Magoto Mossi Magoto, Nata patriarch, c. 1890-1987 xxvii 

1-1 . From Riyara Hill in Nata looking East 1 

2-1 . Regional Setting of the Western Serengeti 58 

2-2. Tribal Map of the Musoma District 71 

2-3. Mwinani by Local Artist Deus Nyahega Tumbago 89 

2-4. Women's Story Telling 91 

3-1 . Ishenyi Dispersal from Nyeberekera 118 

3-2. The Ishenyi Story of Nyeberekera 119 

3-3. White Fathers' Mission. Nyegina, Musoma, Founded 1911 128 

4-1. Linguistic Maps of the Lakes Region Over Time 152 

4-2. Great Lake Bantu Linguistic Tree 1 57 

4-3. Dialect Chaining Chart of the East Nyanza Languages 159 

4-4. Narration of Nata Emergence Stories 166 

4-5. Homestead Layout 1 76 

4-6. Interior House Design 1 77 

5-1. Ecological Map of the Serengeti-Mara Ecosystem 199 

5-2. Human Interaction with the Environment 209 

5-3. Hunting Vocabulary and Tools 223 


5-4. Kuria Women, 1904 225 

5-5. Two Versions of the Story of Masuche's Bao 240 

6-1. Narrators of Clan Histories 251 

6-2. Blacksmiths and Their Tools 287 

6-3. Praise Shouts at the New Moon Dances 293 

6-4. The Regional Distribution of Four Major Clans 295 

6-5. Two Ikizu Rainmakers 306 

6-6. The Ikizu Utemi List 310 

7-1. Nineteenth century Settlement Sites of the Western Serengeti 321 

7-2. Respect for Ancestors 326 

7-3. Emisambwa in Ngoreme 331 

7-4. Nata sites, Gitaraga and Riyara 333 

7-5. Ikizu sites at Chamuriho and Gaka 341 

8-1 . Sites of the Ikizu Walk 398 

8-2. Rikora Leaders 402 

9-1. Chronology of Generations 449 

9-2. Remains of Stone Walls, Ngoreme Fortified Settlements 455 

9-3. Sonjo Fortified Settlements 462 

9-4. Maasai Relations 470 

9-5. Map of Age-Set Cycles 486 

10-1. Narrators of "the Generation of Opportunity" 512 

10-2. Kuria Warriors, 1904 518 

10-3. A New Generation of Wealthy Men 536 

1 0-4. Stories of the Nyangi 543 


Texts that are quoted in the body of the paper are cited in footnotes. In cases where I want 
to give the reader a fuller sense of the context of an interview or various versions of a particular 
narrative without interrupting the flow of the text, I have included these in the appendix.' The 
following conventions are used throughout the text. 

1. Significant omission of the speaker's words is represented with three dots surrounded by 
brackets [. . .] 

2. Slight omission of the speaker's words due to stumbling or repetition is represented by three dots 

3. A pause by the speaker is represented by three dashes — 

4. Explanations, actions or omissions are included in brackets, without italics [ ] 

5. Long narratives are indented and set in italics to set them off from the rest of the text. All texts 
are translated from either a local language or Swahili; where untranslated words appear in the 
indented quotations they are represented without italics and where a local language word appears in 
the body of the dissertation it is represented with italics. 

All of these narratives were collected in interviews and not in performance situations. 
They are dialogic in nature, with frequent interruptions and pause for assent. Different sections of 
one narrative may be told in different settings for different purposes. All of this makes it difficult 

1 I have followed many of the conventions in Isabel Hofmeyer, "We Spend Our Years as a 
Tale that is Told:" Oral Historical Narrative in a South African Chiefdom (Portsmouth: 
Heinemann, 1994), pp. xi-xiii. 

to extract a "standard" version. Therefore, long quoted sections in the text represent only one of 

many variations of the telling and tend to be rather loose and free flowing accounts. Footnotes for 

interviews include the name of the informant, village where the interview took place, the date of the 

interview, and the ethnic group and gender of the informant in parenthesis. 

Standard Orthography 

The linguistic conventions used follow the orthography proposed by Muniko, oMagige and 

Ruel in the Kuria-English Dictionary . Because this is the only published dictionary currently 

available for any of the Mara languages I have chosen to use this as the standard form, following 

Swahili orthography. Mara languages use seven distinct vowels. This means that in using the 

standard orthography two single letters each represent sounds which are pronounced differently. 

Muniko, oMagige and Ruel resolve the problem in this way: 

The open and closed 'e' are both written as '«' and the open and closed 'o' are both 
written as 'o. ' Where, however, the sound difference discriminates between two words 
that are otherwise identical their respective pronunciation is given in the entry in 
parenthesis immediately following the word itself. In these cases: 

open 'o'as in the English 'hot' is written 'o. ' 

closed 'o'as in the English 'open' is written 'ou. ' 

open 'e'as in the English 'end' is written 'e. ' 

closed 'e'as in the English 'rein' is written 'ei. ' 
thus: -kora (prncd. -koura), to scrape away: -kora (prncd -kora), to do, make, -kenga 
(prncd. -keingaj, to ward off, parry; -kenga (prncd. -kenga), to cut, sever. 

Other conventions used in the Kuria dictionary include the following. Long vowels are 

indicated by doubling. All vowels preceding 'mb', 'nch', 'nd', 'ng', 'nk' and 'rr' are pronounced long 

and not always indicated as such in the orthography. The nasalized 'ng' (pronounced as the English 

word 'singing') is written as in Swahili ng'. A trilled or rolled Y is written 'rr.' All vowels are 

sounded, including those at the end of a word. The use of W as a quasi-vowel is avoided and W 

used only as a consonant. Muniko el al maintain that because there are very few words 


distinguished by tone alone, Kuria is not to be considered a tonal language, although intonation is 

important. Therefore no marks of intonation are used. 2 

Names of languages, people and place names are used without the prefixes. Therefore, the 

language Egikuria, the Abakuria people and the place Bukuria all become Kuria, distinguished by 


Glossary of Mara Words Used in the Text 

(In Nata unless otherwise indicated) 

abanyikora — generation-set members in power (Ikizu, Zanaki) 

agecha — famine 

aghaso — ritual of purification for those who kill a lion, leopard or Maasai 

ahumbo — distant fields for crops 

aka — homestead 

akoromo — digging stick (lkoma) 

amachi — house for young unmarried boys 

amang'ana ga kare — matters of the past, history 

ambere — drum 

ambirisi — red oat grass, Themeda triandra 

amuma — oath (Ngoreme) 

amusera — rain medicine (Ishenyi), omoshana (lkizu) 

anchara — famine, path 

ang'ombe umwando — cattle given to the clan at inheritance 

anyumba — house 

arachana — topi skin worn over the shoulder for elderly men 

aring'a — oath 

asaraka — gazelle skin worn by women as an apron 

asaro — circumcision ceremony 

ase or ahase — soil, land 

asi — hunter/gatherers 

asimoka — to sprout up, to wake up, the rise of a river (ensemoka in Kuria), origins 

asimoora — livestock corral 

asire — debt between women who borrow goods as neighbors 

bao — board game common throughout Africa 

bene — the people of, abahiri in Ikizu 

bisa — enemy, Maasai 

budodi ~ wire wrapped bracelets or anklets 

buhoro — whole, healthy 

2 S. M. Muniko, B. Muita oMagige and M. J. Ruel, eds., Kuria-English Dictionary 
(Hamburg: Lit Verlag, for the International African Institute, 1996), pp. v-vi. 

chashabashi — ivory bracelets worn by elders of a certain rank 

chesiri — women's self-help farming groups (Ngoreme) 

chubha - south 

ebereri ~ fall pits for hunting wildebeest 

ebeshona — strips of hide worn ritually on fingers (Ikizu) 

ebimenyo — settlement sites 

egeshoko — log which closes the livestock corral, ekeshoko (Ngoreme), egesaku (Kuria), doorway. 

eghise — black wildebeest or cattle tail used ritually for eldership titles 

egitara — grain storage bin 

ekeburu — temporary huts near to the fields 

ekebuse — sandy upland soils 

ekehita/ebehita ~ doorway or gateway 

ekehwe/ebehwe — ghosts, spirits of the dead 

ekerisho — grazing land 

ekerongori — thin porridge for drinking 

ekimweso - ceremony of purification, sanctification (Ikizu), usually with fire, ikoroso (Ishenyi and 

Ngoreme), shishiga (Ikoma) 
ekinyariri — month of green flush after the burns (August or September) 
ekireri — leader of the generation-set (Ishenyi) 

ekishomba — state of ritual impurity, danger, blood shed between families 
ekitana/ebitana — medicine bundles 

ekyaro/ebyaro - ritually controlled lands, ikiaro (Kuria), clan lands 
emigiro — avoidance 

enchobe — horn of the generation-set, ekombyo (Zanaki) 
erisambwa/emisambwa — spirits of the land, spirits of lineage ancestors, places where spirits of the 

ancestors who guard the land are buried 
eseghero — clay bottom-land soils 
-gaba — to divide the inheritance 
-gonka — to suck at the breast 
-guha — to grow old 

-gutaacha asaiga — ceremony to come into your age-set 
hamate — clan, place, strangers 
-haraga — to prepare soil for planting in the dry season 
hengere — short people 
-ibaka — to praise oneself 
ikwabhe — Maasai 

injama — Kuria secret council of the territory 
irigiha/amagiha — hearthstones 
-itaberi — to bless the land 
kang'ati - leader of the generation-set (Nata) 
kaswende — syphilis 
-kerera — the walk of the generation-set 
kicheneni — of "pure descent" 

kigori — mass circumcision ceremony to close an age-set 
-kerani — to greet, to exchange (difference in stress) 
-kunguha — to grow old 
kwawibancha - front door of the house 

kwibiserani angibo — ritual to remove impurity, to reconcile two families with blood between them 

kyawisiko — emergency back door of the house 

magoro — feet or legs 

mame — maternal uncle 

masabha — north 

masubho — secrets, medicines of eldership ranks or in general 

materego — woodland, wilderness 

-menya — to build 

mkamwana — daughter-in-law, woman married by another woman to produce her heir. 

msororo — beer party in which participants contribute in turn to the preparation 

mtemi — Sukuma chief or Ikizu rainmaker/chief 

-musa — to bless by sprinkling 

mwami — Zanaki rainmaker's title 

mwanangwa — colonial headman 

mwinani — ogre, monster of folk tales 

mwiro/bwiro - non-blacksmith or non-potter 

mwisenge — paternal aunt 

ndezi — leather and cowry shell bracelet used in the investiture of the Ikizu mtemi 

ng'ombe ya baki — the cow of the young woman 

ntemi — ritual scar on right breast (Sonjo) 

nyancha — west or lake, specifically Lake Victoria 

nyangi — celebration of life stages or eldership titles 

nyika -- wilderness 

nyina — mother 

-ny wa — to drink 

obokima — in Swahili "ugali", thick porridge eaten as staple food 

obosongo — arrow poison 

obugabho — prophecy, healing, divination 

obugo — fortified settlement (Ngoreme) 

oburwe — finger millet 

omnibi — wealthy man, cattle 

omoboibororu/abibororu — native born 

omogimba — rainmaker 

omogomba — childless woman, barren 

omogongo wa mwensi - medicine of the generation-set (Ngoreme), etnigongo (Zanaki) 

omogongo — back 

omogore/abagore — a person who is bought for food during the famine 

omokina/abakina — speaker 

omokoro/abakoro — the elders, ancestors 

omokuungu/abakuungu ~ old woman 

omorema/abarema — farmer 

omorokingi - eldership title of the Nata 

omoroti/abaroti — dream prophets 

omoseese/abaseese — slave or dog 

omosimano/abasimano — stranger 

omosimbe/abasimbe - independent woman, managing her own homestead and family, not married 

omosino — widow 

omotangi — leader of the generation-set (Ngoreme) 

omotoro — gift of food 

omotware — male wife, married by an independent woman 

omugabho/abagabho - prophet 

omugambi/abagambi — speaker 

omuhabe — poor person, orphan 

omunase — prophet (Ikizu) 

omunywa ~ mouth 

omuryango — outer room of a woman's house for keeping small stock at night 

omusangura - a kind of tree at emisambwa sites 

omwame — wealthy man, or woman 

omwerechi/abawerechi — speaker (Ikizu) 

omwikarabutu — woman specialized in circumcision ceremonies 

omwisani — friend, bloodbrother (wa saraga, bo maguta) 

omwiti ~ the one who hits the lion first, omunoti, the one who hits second 

orokoba — protection medicine around the land or a hide rope for milking 

orubago — fence 

oruberi — settlement 

orutani — long distance trade, to Sukuma 

orutanya — long staff of the generation-set leader 

-ragera — to eat porridge, to eat 

ribancha — homestead yard 

rihaha — rinderpest 

rikora/amakora — generation-set 

risaga - work party in which labor is rewarded with feasting and beer 

-risi — to herd livestock 

rogoro - the east 

-rota — to dream 

ruaki - stone fort for protection during a raid (Nata) 

-sagari chatugo — cattle clientship 

saiga — age-set, asega (Ngoreme) 

sarara — lungs 

-sengera ~ to propitiate an ancestral spirit, "to beseech" 

sesera — iron bracelets 

-sisimoka — to spring up, as in to wake up or to sprout up from the ground 

sukubi — cattle hump 

-tindeka — to store, to bury the dead 

turi — blacksmith or potter 

utemi — chiefship (Ikizu) 

yowe — alarm call 


When I was choosing a research site it seemed wise to return to the Mara Region where 
my husband and I spent six years working with the Tanzania Mennonite Church in the 
coordination and facilitation of community development projects. This work took us to all parts of 
the region, developing significant personal relationships in many places. I would thus have a 
natural set of contacts and social networks through which to begin my work. But more than that 
my desire to return to graduate school and to write African history was born out of my experiences 
and friendships in the Mara Region-and so I felt a debt there. 

I hoped that the formal study of African history would give depth and understanding to 
patterns that I saw in operation, but about which I was ill-equipped to make sense. At that point, 
as a development worker, I saw history as a tool of empowerment, providing the means for local 
people to understand themselves well enough to take development into their own hands, rather than 
relying on help from the outside. Yet how 1, as an outsider learning local history, would facilitate 
this was never clear. At least 1 hoped to be able to tell the history of a people whose "voice" had 
not been heard, to show the significance of a people and a place which had been seen as an empty 
spot on the map. It was this love and respect, also mixed with some amount of paternalism, that 
allowed me to think 1 was needed to recover this neglected history. Idealism falls hard in graduate 
school and I am now less naive and more skeptical, but no less driven to present a history from the 
perspective of those who live it, even if it is told in and moderated by my own voice. 

When I returned to the Mara Region to do my research 1 was returning to a place that 
seemed a lot like home. This was the place where 1 had given birth to my second child, bumped 

endlessly over the dusty roads and sat with friends over cups of sweet milky tea. 1 learned to speak 
as a child does but had been given responsibilities such as an adult has and struggled for what 
seemed right. I had seen close friends die and had new babies named after me. It was a place that 
1 "felt" as well as intellectually "knew." It was a set of landscapes and people that I had come to 
love and which had deeply changed me. 

When we were searching for a place to live it seemed prudent to avoid living on church 
property or in direct affiliation with the church in order to distinguish old roles from new. It also 
seemed important to live in a rural village that still functioned as a community in order to be part 
of the agricultural cycle. Since this was conceived as a regional study I also wanted to be 
somewhat centrally located with transportation at hand. As much as these criteria were taken 
seriously, in the end the decision to move to Nata was based on personal and subjective choices. 
Nata is in some ways the smallest and least significant historically of all of the ethnic groups in the 
area. Yet by living here, Nata narratives, language and custom took on a centrality they would not 
otherwise have had. 

The decision to move to Nata was based on a personal relationship and invitation from 
Nyawagamba Magoto. We got to know him as a local development coordinator and a lay church 
leader. He is an energetic man, committed to introducing innovation in everything from imported 
dairy cows to wheel-thrown, glazed pottery. We came to learn that he is also part of a large and 
influential Nata family-the eighteen sons and daughters of the local patriarch. Magoto Mossi. 
Magoto's children and grandchildren are influential in the community as politicians, teachers, 
development workers, church leaders, businessmen and village council members. One son has a 
university education and another son was the founder of the first cattle herders' cooperative in the 
region, "Wafugaji wa Mara," in the 1950s, and a former member of parliament. Magoto himself 
went through all of the eldership titles, holding one of the highest ranks until he renounced this 

position later in life. He was also an innovator, introducing the first tractors and grain mills in the 
area. One of his daughters was the first Nata girl to attend school. Magoto was wealthy, his cattle 
numbering in the thousands at one time. In the early 1 930s he opposed the Nata Chief, Rotigenga, 
and left to live in Ikizu-nearly 400 Nata followed him. For better or worse my research in Nata 
was under the protective shield of Magoto's legacy. 1 

Nyawagamba promised that if we would come to live in Nata he would build us a house 
and help to facilitate the research. My husband, two sons (9 and 12 years old at the time) and I 
arrived in Nata at the beginning of February, 1995, after having spent a month in Dar es Salaam 
clearing the formalities and doing an initial survey of the archives. We found the round walls and 
roof framework of the house up, with Nyawagamba rushing to finish thatching before the big rains 
began. We camped near the house for two weeks in the rain while the house was being finished- 
our lives and possessions open to all who walked down the path. When we finally moved into the 
house we had become known locally as "wazungu wa Nyawagamba" (Nyawagamba's foreigners). 

Our house sat in the small village of Bugerera, administratively part of the larger village of 
Mbiso on the main trunk road between Musoma and Arusha. Bugerera is the place where many 
Mbiso families have some of their fields and send second wives or elder sons out to live for the 
farming season. It lies on the boundary between Nata and Ngoreme. Just over the hill one finds 
the Ngoreme village of Mosongo and the no-man's wilderness of cattle raiders and thieves. When 
we first drove out to Bugerera, after having committed ourselves to live there, we drove the ten 
kilometers from Mbiso without seeing more than a few isolated homesteads. I was sure that my 

' One of the ways in which I fulfilled the ties of reciprocal obligation to the Magoto family 
was by helping them to write a personal history of Magoto. I am now working with the Swahili 
manuscript by Mwalimu Nyamaganda Magoto Mossi, assisted by Masoye Faini and Chuba Faini 
on behalf of the Magoto Family, "Historia ya Mzee Magoto Mossi Magoto, Katika Maisha Yake " 
Nata. 1996. 

research would be doomed by living where there were no people. Yet it was the location of 
Bugerera that taught me to look for people in the hills rather than on the flood plain where the car 
could travel and to see the relationship between larger villages and their related farming satellite 

Since we were friends and colleagues of Nyawagamba's before coming to live with him 
nothing between us could be settled on the basis of payment for services. We were forced into 
functioning in an "economy of affection" and learning the unspoken rules of reciprocity. 2 We did 
not pay rent on the house, a salary for Adeja (Nyawagamba's wife) to help with household chores, 
compensation for long days spent on interviews instead of farming, or for the milk, meat and 
vegetables that were frequently provided from the Magoto family surplus. We. in turn, provided 
transportation to Musoma or Mugumu for shopping, hospital or other emergencies. We sometimes 
bought gifts for the family when we went to the town. My husband, Peter, who volunteered in 
Serengeti National Park, made arrangements with Serengeti tourist hotels to buy Nata vegetables, 
fruits, milk and cream and transported them to the hotels when he went to the park. This system of 
reciprocity was also used for the colleagues that assisted me in the other ethnic groups outside of 
Nata. These colleagues spent a lot of their own time and resources to facilitate my research 
without direct compensation. I was told by Nyawagamba when I left that it was better to leave 
myself in the debt of others, rather than the other way around, so that they would feel free to 
continue the relationship when 1 returned. 

The Magoto children incorporated us into their family as brother, sister-in-law and 
children. We were greeted in this way throughout Nata-treated with both the respect and 

2 Reference to the "economy of affection" from Goran Hyden, Beyond Uiamaa in 
Tanzania: Underdevelopment and an Uncaptured Peasantry (Berkeley: University of California 
Press, 1980). 

obligation which followed these titles. This allowed me the freedom to travel alone with my 
brothers-in-law, or ask for help from my sisters-in-law, and allowed the children a protective shield 
against the curiosity of school children. The knowledge 1 have gained is a "situated knowledge" 
which is firmly seated in a particular social network and sense of place. 

From the moment I arrived in Nata it was clear that the Magoto family had taken us, and 
my research, as their collective project. Nyawagamba talked to elders before I came and prepared 
them to share the history that they knew. I told the family that 1 intended to spend the first three 
months studying Nata so that I could do interviews in the local language rather than Swahili, in 
which I was already fluent. This also seemed convenient since it was the rainy season and it would 
be almost impossible to travel for the next three months. 1 asked the wife of one of the Magoto 
brothers who lived nearby to help me with language learning and we worked at it daily with a tape 
recorder, notebook and pen. 

By the beginning of March, soon after we were living in our new house, Nyawagamba and 
his brother Mayani decided that it was time for me to start research. They were anxious that 1 
should start as early as possible and not waste time on language study, since I was sure to "pick it 
up" without any difficulty anyway. Nyawagamba and Mayani sat in our house for a couple of 
days writing a list of the things that I should learn if I was to know Nata history and culture. They 
envisioned this list as a set of topics to be systematically covered in my interviews. It was entitled 
"The History of the Nata People" (in Swahili, Historiaya Wanata). This list is reproduced in full 
on Table F-l. 

Soon after the list was finished Nyawagamba and Mayani arranged an interview with 
Megasa Mokiri in Motokeri. He had taken the top eldership title in Nata and was known as a 
person knowledgeable about the past. He was also in some indirect way related to Magoto and so 
bound to honor the request for an interview. When we arrived he was out drinking a neighbor's 


Figure F-l: Topics for the Study of Nata History 

(Mayani Magoto and Nyawagamba Magoto, Bugerera, 1 March 1995) 

History of the Nata People 

1 . Asimoka (origins) of the Nata 

2. Ukwengera (growth) of the Nata 
l.Obokwiri (marriage) of the Nata 

4. Amborori (divorce) of the Nata 

5. Amasambwa (spirits of the land and the 

ancestors) of the Nata 

6. Emigiro (avoidances) of the Nata 

7. Nyangi (life stage celebrations and 

eldership titles) of the Nata 

8. Ebehita (patrilineages) of the Nata 

9. Chanyumba (matrilineages) of the Nata 

1 0. Chahamale (clans) of the Nata 

1 1 . Chasaiga (age-sets) of the Nata 

1 2. Amakora (generation-sets) of the Nata 

13. Asaro (circumcision) of the Nata 

14. Rihe (war) of the Nata 

15. Charing'a (oaths of peace or blood 

brotherhood) of the Nata 

1 6. Asama (moving) of the Nata 

1 7. Ebimenyo ne mipaka (settlement sites 

and boundaries) of the Nata 

18. Ebilana (medicine bundles for 

protection) of the Nata 

1 9. Imieri (months) of the Nata 

20. Ebisego bya kwerekiani amabaga gase 


2 1 . Emeremo gya Saiga (work of the age- 


22. Emeremo gya Rikora (work of the 


23. Asumo (trade) of the Nata 

24. Anchara (famines) of the Nata 

25. Omobureni (young men) 

26. Omuki (young women) 

27. Okuibora (birth) of the Nata 

28. Orokurya (death) of the Nata 

29. Agabho (inheritance) of the Nata 

30. Ang'ombe ya Baki ("the cattle of the 

young women") 

3 1 . Ang'ombe yu Mwando ("the cattle of 

lineage inheritance") 

32. Ababisa (enemies) of the Nata 

33. Obuluri (blacksmithing) of the Nata 

34. Rirema (farming) of the Nata 

35. Obotugi (herding) of the Nata 

36. Ribiema (hunting) of the Nata 

37. Emeremo egiende (other work) 

38. Abana kwegi emeremo (the work of 


39. Amang'ana ga Kare (things of the past) 

40. Obugeni (hospitality for guests) of the 


41 . Risaga (mutual aid) of the Nata 

42. Chakabari (co-wives) of the Nata 

43. Amarina (names) of the Nata 

44. Okusohe kwa Abalaki (the entrance of 

the colonial force) 

45. Omutemi ombele (the first chiefs) 

46. Kebuno betemiri (how they ruled) 

47. Omutemi wo kabere (the following 


48. Abato Maarufu (famous people) 

49. Risau (alarm call) of the Nata 

50. Kusagari chatugo (cattle clientship) 

5 1 . Obogwani bwa matongo (relations with 

other "tribes") 

52. Obusani (friendship) 

53. Obwiterani (murder) 

54. Aghaso (ritual for killing a lion, leopard 

or Maasai) 

55. Kwiraheri (praise names and oaths) 

56. Imiembo (songs and dances) 

newly brewed beer and was only called away with the promise that some would be brought home 
for him. This interview was conducted strictly in Nata with almost no translation to help me follow 
the conversation. Mayani and Nyawagamba based their questions on the list with only four or five 
points being covered in the next four hours. My job was to hold the tape recorder. 

I take these stories as a metaphor for my research and its final product. Although I 
eventually gained more control over interviews, this was a collaborative project which depended on 
networks of relationship established in an earlier time of my life. The project succeeded only in so 
far as people like the Magoto family in Nata, Pastor Machota in Ikoma. David Maganya in 
Ngoreme, Kinanda Sigara in Ikizu and Mnada Mayonga in Tatoga took an interest in the research 
and made it their own. These men spent uncountable hours, days and weeks driving, bicycling or 
walking to remote homesteads with me, arranging for interviews and dealing with the fall-out after 
I was long gone. The stakes were particularly high for the Magoto family, who would have 
suffered severe community criticism and ridicule if things had gone badly for their guests. 

The elders that 1 interviewed only agreed to talk to me because they were convinced that 
the product of my research would be beneficial to their grandchildren after they were gone and 
their knowledge forgotten. The local men who shared their historical manuscripts with me. the 
archivists who took time to dig through unsorted boxes, the priests and primary school teachers 
who uncovered dictionaries and ethnic histories from dusty shelves and filing cabinets, Mwalimu 
Nyamaganda who walked out to Bugerera numerous times in spite of his illness to go over endless 
lists of cultural vocabulary and all of those women who shared their homes, meals and stories with 
me countless times did this in large part because of their commitment to the preservation of local 

It is this enormous debt, this trust, that I struggle with as I put the words into a narrative 
form which must sustain a linear argument. Nyawagamba and Mayani's list remains a symbolic 


measure of how closely I have fulfilled these promises. One thing that the list tells me is that the 
story is not linear, but that peoples' lives are made up of a vast number of parts, which in their 
multiplicity constitute the pasts of the people we now call Nata, Ikizu, Tatoga, Ikoma, Ishenyi, or 
Ngoreme. The list also tells me that Mayani and Nyawagamba considered history to be 
synonymous with its cultural forms-that marriage, oaths, medicine bundles and naming are all as 
much a part of the definition of what it means to be Nata as are accounts of origin, migration or 
chiefs. There is no obvious organizing structure to this list, except as a simple naming of all of the 
elements, without establishing hierarchies or relationships among the parts. Perhaps Nyawagamba 
and Mayani felt that the connections were obvious, or perhaps they left that task for me-allowed 
me the freedom to unite all of those disparate elements into one story. In this sense the debt 
becomes a burden because it is a difficult task. 

How can I make local history live in narrative form while not denying its multiplicity and 
contextuality? The translation of historical imagination which must be trained into a focused 
thesis will have succeeded if the spirit of this unruly list and the vision of those who made the 
research possible remain. The story that 1 tell is my story and I must take responsibility for its 
outcome no matter how indebted it is to the dedicated people who made it possible. I must speak in 
my own voice while respecting those who gave me that voice. It will have succeeded if western 
Serengeti people do not accept this as a final product but contest it, debate it and write their own 

I dedicate this dissertation to Magoto Mossi Magoto and to his sister Nyabikwabe Mossi, 
whose legacies have made this work possible. The only photo of Magoto is reproduced here out of 
respect for his patronage and in hopes that the hospitality of his sons and daughters might be 
reciprocated in a work about which he would be proud [See Figure F-2]. Nyabikwabhe was a 
woman who managed her own homestead after her husband disappeared on a migrant labor trip to 

Figure F-2: Magoto Mossi Magotto, Nata patriarch, c 1890 - 1987 


Nairobi. She left her legacy with many children and grandchildren in Nata. She is remembered by 
her family as a storyteller, knowledgeable about the past. This is dedicated to her in the hope that 
the tradition of story-telling and remembering that she represents will not be lost. 

Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 



Jan Bender Shetler 

May 1998 

Chairman: Steven Feierman 

Major Department: Department of History 

This dissertation analyzes changes in social identity from the distant past to the early 
colonial years in the western Serengeti, Mara Region, Tanzania, through an interpretation of oral 
tradition. My analysis of the core spatial images of oral tradition demonstrates the link between 
different ways of representing space and historical changes in social identity. Multiple social 
identities grounded in the ecological diversity of the region developed out of long-term social 
processes based on the elaboration of generative principles of social organization. The generative 
principles of gender, economic production, clan, lineage, generation-set, age-set and ethnicity 
produced various kinds of social identities and institutions according to the historical context. 
Social identities changed profoundly at the end of the nineteenth century when an era of ecological 
disasters caused western Serengeti people to undertake major social transformations. They 
responded to the crises of this period of stress by renovating their systems of age-set organization 
and by redefining the criteria for ethnic identity, not only to ensure their survival but to convert the 
difficulties of the early years of colonialism into prosperous ones. 

"The landscapes of memory" refers to the way in which memories about the past are stored 
in spatial form. Oral traditions represent the past through spatial images that correspond to 


particular forms of social identity. An interpretation of the core spatial images of oral tradition 
provides the historian with a culturally grounded representation of social processes in the past. A 
representation of the past reconstructed through other forms of evidence like archaeology, historical 
linguistics, comparative ethnography, ecology and written sources demonstrates an amazing 
congruence with the historical understanding afforded by oral tradition. This dissertation seeks to 
understand the past, as much as possible, from the perspective and categories of local historical 
consciousness. The spatial analysis of oral tradition contributes to the historical reconstruction of 
the precolonial past in Africa and social history elsewhere in the world, particularly in places such 
as the Mara Region where historians must rely almost exclusively on oral tradition as a primary 

Think of the past as space expanding infinitely beyond our vision. It is not a record of 
progress or regress, stasis or change; uncharted, it simply, smugly, vastly is. Then we choose a 
prospect. The higher it is, the wider and hazier our view. Now we map what we see, marking some 
features, ignoring others, altering an unknown territory, absurd in its unity, into a finite collection 
of landmarks made meaningful through their connections. History is not the past, but a map of the 
past drawn from a particular point of view to be useful to the modern traveler. 

Serious study of a community's history does not begin with a raid to snatch scraps to add 
color and flesh or nobility to the history of another community. It begins when the observer adopts 
the local prospect, then brings the local landmarks into visibility, giving the creations of the 
community's people-the artifacts in which their past is entombed, the texts in which their past lives 
—complete presence. ' 





' *-~Jc«-" - 





Figure 1-1 : From Riyara Hill in Nata looking East toward Gitaraga and Mochuri 
Mountains with Bangwesi Mountain in the background and Serengeti Plains beyond. 

1 Henry Glassie, Passing the Time in Ballvmenone: Culture and History of an Ulster 
Community (Philadelphia: University of Pennsylvania Press, 1982), p. 621. 




This is a project, as Glassie defines it in the preceding quotation, of adopting the "local 
prospect" in order to "give presence" to the landscapes and the places through which a people 
imagine their collective pasts and "in which their past lives." This dissertation explores these 
landscapes containing the memories of the past to understand social change from the perspective of 
those who live it. As Glassie writes, concerning the great historical epics of Ireland, "time is 
absorbed into place, and place into mind ... the land becomes history, and history becomes 
thought as people cross space in awareness." 2 

In this dissertation 1 explore the "landscapes of memory" which structure knowledge about 
the past in spatial form through an interpretation of oral traditions from the western Serengeti, 
Tanzania. This project will contribute to the growing body of research among Africanist scholars 
on the ways in which the organization of space functions in oral tradition. 3 My analysis of the core 
spatial images of oral tradition demonstrates the link between different ways of representing space 

2 Ibid, p. 664. 

3 Michele Wagner, "Whose History is History?: A History of the Baragane People of 
Buragane, Southern Burundi, 1850-1932," 2 vols. (Ph.D. Dissertation, University of Wisconsin- 
Madison, 1991); Henrietta L. Moore, Space. Text and Gender: An Anthropological Study of the 
Marakwet of Kenya f Cambridge: Cambridge University Press, 1 986); Karin Barber. I Could 
Speak Until Tomorrow: Oriki. Women and the Past in a Yoruba Town ( Edinburgh- Edinburgh 
University Press,1991); David William Cohen, "The Cultural Topography of a 'Bantu Borderland': 
Busoga 1500-1850," Journal of Africa History 29 (1988): 57-79; David William Cohen and E. S. 
Atieno Odhiambo, Siava: The Historical Anthropology of an African Landscape (London: James 
Currey, 1 989); Tamara Giles- Vernick, "Na lege ti guiriri (On the Road of History): Mapping Out 
the Past and Present in M'Bres Region. Central African Republic," Ethnohistorv 43, 2 (Spring 

and historical changes in social identity. In this dissertation, 1 understand social identity as a 
socially shared definition of self in relation to others that is situational and relational. Multiple 
social identities coexist within the individual, activated according to the situation and the 
relationships involved. Oral traditions produce and reproduce these identities through their 
representation of space. This study encompasses an ecologically -defined region— the western 
Serengeti of Tanzania-a region larger than one ethnic group. Thus, multiple and shifting forms of 
social identity represented in oral tradition over time and space must be investigated. 

These multiple social identities grounded in the ecological diversity of the region developed 
out of long-term social processes based on the elaboration of generative principles of social 
organization. The generative principles of gender, economic production, clan, lineage, generation- 
set, age-set and ethnicity produced various kinds of social identities and institutions according to 
the historical context. Social identities changed profoundly from a time in the distant past when 
people gradually developed a regional system of relationships for achieving prosperity in a 
marginal land to the end of the nineteenth century when an era of ecological disasters caused them 
to undertake major social transformations. For the western Serengeti, as for many other parts of 
East Africa, crisis defined the late nineteenth century. This time of famine, disease, war, and 
dislocation served as the dividing point between the far and recent past in oral tradition. People 
responded to the crises of this period of stress by renovating their systems of age-set organization 
and by redefining the criteria for ethnic identity, not only to ensure their survival but to convert the 
difficulties of the early years of colonialism into prosperous ones. 4 

4 Although German East Africa was established as a colony in 1885, German colonial rule 
was not effective around Lake Victoria until 1891 with the establishment of a military post at 
Mwanza. The British Mandate of 1922 incorporated the Tanganyika Territory into the British 
Empire as a result of the League of Nations settlements following World War I. 


Scholars cannot understand these profound late nineteenth century changes apart from the 
long-term historical processes in which people developed patterns for building strong local 
communities around their relationship to the land and to other communities within regional 
networks of reciprocity. Yet the historian understands these older processes only through oral 
traditions, traditions that the events of the late nineteenth century have significantly altered. 
Western Serengeti people radically transformed their societies in the period of disasters by drawing 
on the generative principles of long-term social process, yet in their oral representation of these 
processes they interpret of them in light of their desire to seek historical continuity with, and 
validation for, new ways of building and maintaining strong communities. The fundamental 
importance of the era of disasters to oral representations of the deeper past requires that 1 interpret 
the oral traditions of the pre-crisis era in light of the experiences of the post-crisis era. 5 The design 
of the dissertation reflects this necessity by weaving reflections on the influence of the era of 
disasters on the character and content of all the oral evidence presented into my analysis. 

The primary oral sources for this study come from the traditions of five agro-pastoral 
ethnic groups that speak Bantu languages (Nata, Ikoma, lshenyi, Ikizu and Ngoreme) and two 
pastoral ethnic groups that speak Dadog languages (Rotigenga and Isimajek) in the Serengeti and 
Bunda districts of the Mara Region. These stories guide my version and compel me to take 
seriously the perspective and categories of local historical consciousness. 6 A culturally sensitive 

5 See Gwyn Prins, "Introduction," The Hidden Hippopotamus: Reappraisal in African 
history: The early colonial experience in western Zambia (Cambridge: Cambridge University 
Press, 1980), pp. 1-16, who analyzes "core concepts" in order to place nineteenth century colonial 
contact in long-term local context. 

6 Richard Price eloquently expressed this sentiment in his book on the history of maroon 
communities in Saramaka, "All history is thus: a radical selection from the immensely rich swirl of 
past human activity. The uniqueness of this book lies in its taking seriously the selection that is 
made by those people who gather together at this shrine. It is about those distant people and those 
long-age events that Saramakas today choose to think about, talk about, and act upon: but it is also 

historical reconstruction seeks to translate the inner logic of oral traditions and the profound impact 
of late nineteenth century crises on their content into the language of academic historical 
discourse. 7 The historian can respect the integrity of oral traditions and build a chronology to 
explain social change by using oral sources together with other kinds of evidence, from historical 
linguistics, archaeology, comparative ethnography, ecological studies and written sources. 

The study of African social history in the precolonial period has recently fallen into 
relative neglect after a prolific and optimistic outpouring of research in the 1 960s and 70s. This is 
in large part due to the difficulties inherent in using oral traditions as primary historical sources. 
We still know too little about the precolonial period and, as a result, historians of the colonial 
period build on questionable foundations. It is my hope that this dissertation will inspire another 
look at oral sources through an investigation of the spatial dimension of oral memory. Although 
the subject is social transformation at a particular place and time, historians could apply the 
approach to the interpretation of oral sources from other places in Africa and to social history 
elsewhere in the world. 

Oral Traditions in Space and Time 

In recent years, academic debate over the utility of oral sources for writing history has 
centered on Jan Vansina's positivist approach to oral traditions. He asserts that, through rigorous 
application of the proper methodological tools, historians could discover the objective past on 

about the ways that Saramakas transform the general past (everything that happened) into the 
significant past, their history. This book is an attempt to communicate something of the 
Saramakas' own special vision of their formative years." Richard Price, First-Time: The 
Historical Vision of an Afro-American People ("Baltimore: Johns Hopkins University Press, 1983), 
p. 5. 

7 Steven Feierman, The Shambaa Kingdom: A History (Madison: University of 
Wisconsin, 1977), p. 3; Thomas Spear, "Oral Traditions: Whose History?," History in Africa 8 
(1981): 8. 


which narrators once based these texts. David W. Cohen and others have engaged this claim by 
subtly undercutting and destabilizing its first premises-that an orderly "chain of transmission" 
would lead the historian back to the original testimony, that there was a single tradition with many 
variations rather than multiple and conflicting histories, and that different sets of oral traditions 
represented independent evidence. Luise White's recent work on rumor and gossip questions the 
criteria used by historians to evaluate their sources. 8 She asks, if people all over East Africa tell 
consistent stories about vampires does that mean they are "true?" This primary emphasis on oral 
tradition as itself a historical product leaves little room for a history from oral traditions that 
discusses issues from earlier periods. Are we to focus only on how people talk about the present 
by reference to an imagined past? This dissertation argues that a middle ground between these 
two approaches exists that incorporates the critique of objectivist methodology and yet still 
supports writing about the distant past, within the standards set by the discipline of academic 
Oral Traditions. Memory and History 

Historians using oral traditions as their principal source of evidence have been confounded 
by the central problem of ascertaining the time depth of oral narratives. It has been demonstrated 
many times that the content of oral tradition is not stable and that it changes from performance to 

8 See the collected papers from the international conference, "Words and Voices: Critical 
Practices of Orality in Africa and in African Studies," Bellagio, Italy, February 24-28, 1997, 
University of Michigan, Ann Arbor. On this issue see David Newbury, "Contradictions at the 
Heart of the Canon: Oral Historiography in Africa, 1960-1980;" Carolyn Hamilton, "Living with 
Fluidity: Oral Histories, Material Custodies and the Policitics of Preservation;" Luise White, 
"True Stories: Narrative, Event, History and Blood in the Lake Victoria Basin;" David William 
Cohen, "In a nation of white cars ... one white car, or "a white care" becomes a truth." This 
destablization is not confined to oral history alone, for an analysis of related issues affecting the 
more common methodologies of the historical profession see, Peter Novick, That Noble Dream: 
The "Objectivity Question " and the American Historical Profession tCamhrirlgp- Cambridge 
University Press, 1 988). 

performance over time and in relation to the historical contexts in which the traditions are told. 
Different social groups tell different stories about the past and in different ways to legitimate a 
particular social order. The present not only influences the narration of the past but knowledge of 
the past most surely influences our experience of the present. 9 The historian's analysis of any one 
tradition must take into consideration the present context in which narrators tell it, as well as all of 
the other historical contexts through which it has passed in transmission. Because of these 
difficulties, many have despaired of finding any verifiable historical content in oral traditions at 
all. 10 

One way to assess the historical content of oral tradition is through an understanding of its 
narrative form. Studies of oral memory have shown that narrators construct (rather than 
reproduce) oral traditions in performance through the use of mnemonic systems, the central 
elements of which scholars of oral tradition call "core images" or "cliches." By recalling these core 
images narrators improvise the entire narrative as they tell it. In the Nata origin story the core 
images are a hunter following his prey from the wilderness and a woman at her cave by the spring. 
Narrators elaborate details of how they met and what they said around these core images to form 

9 Paul Connerton, How Societies Remember (Cambridge: Cambridge University Press, 
1989), pp. 2-3; Elizabeth Tonkin. Narrating Our Pasts: The Social Construction of Oral History 
(Cambridge: Cambridge University Press, 1992). For an analysis of social memory outside of 
African history see, Patrick J. Geary, Phantoms of Remembrance: Memory and Oblivion at the 
End of the First Millenium (Princeton: Princeton University Press, 1994). 

10 David Henige, "Oral Tradition and Chronology," Journal of African History . 12, 3 
(1971): 371-389; Joseph C. Miller, ed.. The African Past Speaks: Essays on Oral Tradition and 
History (Folkestone, Kent: Dawson Archon, 1 980); Paul Irwin, Liptako Speaks: History from Oral 
Traditions in Africa (Princeton. Princeton University Press, 1981). 

episodes or narrative units that they string together to create the larger story anew in each 

Historians of oral tradition have long postulated that it is these core images that hold the 
key to historical interpretation. Vansina proposed guidelines for interpreting the "implicit 
meaning" of these "core images" or "cliches," such as comparison with other traditions and other 
cultural expressions. 12 Joseph Miller later suggested that since core images serve as the mnemonic 
device for recalling the story, people pass on these images from generation to generation, even if 
they no longer understand the original meaning. As time goes on the parts of the story that 
narrators elaborate with each telling tend to lose their detail and become generalized, or are 
replaced with present-day experiences. Miller postulated that the core images held the best 
possibility of bearing "information from and about the past." 13 Steven Feierman's structuralist 
interpretation of the core images in the Shambaa origin myth of Mbegha, in terms of the historical 
development of kingship, remains one of the best examples of this kind of interpretation. '" 

Some of the most important core images found in African traditions are spatial images of 
landscape, place or topography. As Elizabeth Hofmeyr put it, "oral memory has a close mnemonic 
relationship with place and location, and in a variety of societies people often bank information in 
the landscape." She questions whether people can sustain memory if they lose touch with the 

" The theory is first argued in A. B. Lord, The Singer of Tales (Cambridge, Mass: 
Harvard University Press, 1964); see also, Joseph C. Miller, "Introduction: Listening for the 
African Past," in The African Past Speaks , ed. Joseph C. Miller (Folkestone, Kent: Dawson 
Archon, 1 980), pp. 5-9. 

12 Jan Vansina, Oral Tradition as History (Madison: University of Wisconsin Press, 1985), 
pp. 144-6. 

" Miller, "Listening for the African Past," p. 8. 

14 Feierman, The Shambaa Kingdom , pp. 40-69. 

places and landscapes of the core images." However, the recognition that core images often 
appear as spatial images does not solve the problem of their interpretation. The first generation of 
historians to interpret oral traditions in Africa accepted the literal meaning of place names in 
migration or clan origin traditions resulting in untenable reconstructions of the movement of large 
and discrete groups of people over long distances. 16 The internal meaning of spatial images is not 
always explicit and not all spatial images are core images. 

How then does the historian discern which spatial images in oral traditions contain 
information about the past and how she should interpret them? For example, place-names can be 
added later to migration stories or even changed. The Ikoma origin story claims that the original 
ancestor migrated from Sonjo, on the other side of the Serengeti plains, and then lists a number of 
other places that are within the western Serengeti. The Sonjo, however, claim that the British gave 
them this ethnic name in reference to the Sonjo bean that was prevalent in the area (the Sonjo call 
themselves Wantemi). If this is true then this ethnonym-cM/n-place-name could not have been part 
of a precolonial tradition. Nevertheless, through a careful interpretation of the important spatial 
references in oral tradition the historian can analyze the historical connection between Sonjo and 
Ikoma. On closer examination of many Ikoma origin stories one notices that, in place of Sonjo, 
some narrators use the name Regata. A present day village named Rhughata in Sonjo claims its 
origins at Jaleti and Ngrumega (perhaps a transliteration of the Rivers Mbalageti and Grumeti in 

15 Hofmevr. "We Spend our Years ." pp. 106, 125, 132-133, 160. 

16 In his paper for the Bellagio Conference," Words and Voices," 1 997, David Newbury, 
"Contradictions at the Heart of the Canon," called this group the "fundamentalists," including; G. 
Hartwig, "Oral Tradition Concerning the Early Iron Age in Northwestern Tanzania," African 
Historical Studies 4 (1974): 3-115; J. B. Webster, D. H. Okalany, C. P. Emudong, and N. Egimu- 
Okuda, The Iteso D uring the Asonva (Nairobi: East African Publishing House, 1973). 

the western Serengeti). 17 Ngoreme tradition mentions origins in Sonjo but specifically name the 
village of Tinaga, a village that Sonjo people claim the Maasai destroyed in raids, causing the 
people to flee to "Ikoma." 18 The landscape described in these stories is the bush, inhabited by 
hunters following the wildebeest migration. Indeed, if one followed the wildebeest migration from 
Sonjo, he would end up in lkoma or Ngoreme. The historian must conclude that a thick set of 
historical interactions from different time periods existed between Sonjo and Ikoma, expressed by 
oral traditions as the migration of a single ancestor. 

Examples of prestigious places being added onto the beginning of origin stories to claim 
affiliation with powerful kingdoms also exist. The claim of association may be at the root of Jita 
claims to Oanda origins or Kanadi (Sukuma) claims to Hima origins. Both Ganda and Hima 
peoples represent royal clans who formed centralized states on the other side of Lake Victoria, 
states whose power people across the wider Lakes Region admired, feared and resisted. Many 
western Serengeti traditions that claim origins in the east name the specific places of Mount 
Kilimanjaro and Arusha, which are major points of reference for Tanzanian nationalism. In 
western Kenya many ethnic groups claim origins in Misri (Egypt) in a biblical exodus model." In 
each of these cases the place-names following the prestigious name are specific, localized places 
with rich cultural meanings attached to them. Usually, the direction of the place relative to the 
community's present location is important either symbolically or historically. The historian cannot 
simply identify spatial references and accept their unmediated historical veracity but must rather 

17 Interview with Emmanuel Ndenu, Sale, 6 December 1995 (Sonjo &). 

" Interviews with Peter Nabususa, Samonge, 5 December 1995; and Samweli Ginduri, 
Samonge, 6 December 1995 (Sonjo tf). 

19 William R. Ochieng', "Misri Legends in East African History," East Africa Journal 
(October 1972): 27-31. 


pay attention to the cultural meaning of these core spatial images and carefully interpret them 
alongside other traditions and other kinds of evidence. These images are important in any 
historical reconstruction because they are integral to oral memory itself. 

The historian best approaches the interpretation of spatial images by first understanding 
how and why our minds spatialize memory. Studies of memory have shown that people store 
recollections of the past as spatial rather than temporal images. 20 We remember events and people 
by locating them in particular places. Thus memories appear to us as a sequence of places rather 
than as the orderly passage of time. In his exploration of the "poetics of space," Gaston Bachelard 
writes, "Memories are motionless, and the more securely they are fixed in space, the sounder they 
are." 21 This insight has profound implications for the historian using the evidence of memory as a 
primary source. As Bachelard notes, "to localize a memory in time is merely a matter for the 
biographer and only corresponds to a sort of external history, for external use, to be communicated 
to others." 22 The job of the historian, like the biographer, is to fix memories within a chronological 
sequence in order to understand change over time and its possible causes. Yet if Bachelard is right, 
memory cannot provide the historian with precise temporal sequences or duration. 

This insight corresponds with my experience in listening to oral traditions in the western 
Serengeti. Elders took great care to give me sequences of place-names yet without narrative 
explanation. In the early stages of my research 1 was puzzled because people seemed to care more 

20 Frances A. Yates, The Art of Memory (Chicago: University of Chicago Press, 1966); 
Jonathan D. Spence, The Memory Palace of Matteo Ricci (New York: Viking, 1984); Mary J. 
Carruthers, The Book of Memory: A Study of Memory in Medieval Culture (Cambridge: 
Cambridge University Press. 1 990); George Johnson, In the Palaces of Memory: How We Build 
Worlds Inside Our Heads (New York: Alfred A. Knopf, 1991). 

21 Gaston Bachelard, The Poetics of Space , trans. Maria Jolas (New York: Orion Press, 
1964; first published by Presses Universitaires de France, 1958), p. 9. 

22 Ibid. 

about specifying place rather than time frame. When asked when they took place the narrators 
usually said either, "a very long time ago before the grandfather of my grandfather." or, for more 
recent memories, they would name a specific age- or generation-set. In contrast, specific place- 
names and spatial images were the central organizing features of oral traditions. For example, the 
Nata origin story took place at Bwanda, in the ecological space where farming and hunting 
landscapes meet. The last great battle with the Maasai took place at Ndabaka, which means "the 
plain of tears" in Dadog. In its starkest form narrators presented the period of settlement as a 
simple list of place-names or migration points with almost no elaboration. People attached great 
importance to the accurate recitation of lists of place-names and spatial features in oral narratives. 

In his study of "Ilongot headhunting," Renato Rosaldo found that the Illongot people 
presented stories about the past as sequences of the "names of places where they had 'erected their 
house posts' and 'cleared the forest.'" They conceptualized history as a movement through space, or 
as a group of people walking in a single file along a trail and stopping at a sequence of named 
resting places. Narrators named no dates or time periods in these stories because events were 
"mapped onto the landscape, not onto a calendar." Rosaldo called the Ilongot historical idiom a 
"spatialization of time" and convincingly reconstructed Ilongot history by making correspondences 
between place-name sequences and temporal chronologies from references in other sources. 23 
Another example is Bruce Chatwin's description of how Australian Aboriginals conceptualize 
history as "song lines" or "dreaming tracks" that are particular paths on the ground, established 
when the ancestors "sang" the world into existence. Each track is a "song" or a "map" by which 
people remember the past. 24 Similarly, in the American tradition, the mention of Plymouth, 

23 Renato Rosaldo, Ilongot Headhunting 1883-1974; A Study in Society and History 
(Stanford: Stanford University Press, 1980), pp. 42- 58. 

24 Bruce Chatwin, The Songlines (New York: Viking, 1987), pp. 2, 12. 

Jamestown, Bunker Hill, or Gettysburg, for example, evokes an emotion-laden history in the 
popular consciousness. 

Clearly, the historian cannot reconstruct the temporal framework of oral memory without 
an internal knowledge of its spatial component. To return to Bachelard's image, as "biographers," 
we are in the business of putting memory back into chronological sequence to present our "external 
history, for external use, to be communicated to others." To do this we need to pay particular 
attention to the indigenous conceptual frameworks that govern the use of time and space in oral 
narratives. As we have seen, oral traditions use space as a metaphor for time or as a mnemonic for 
social processes that are no longer temporally anchored. Oral traditions often use an epic spatial 
scale as a metaphor for emphasizing the importance or permanence of certain social processes. 
History itself is a cultural construct that determines the structure and use of oral narratives about 
the past. The historical interpretation of oral tradition must, in effect, attempt to reverse the 
spatialization of memory to recover what people originally meant these images to convey. 

If the spatial elements of oral tradition are part of a mnemonic system then the historian 
can use them as "evidence in spite of themselves" that provide tangible information about the 
past. 25 The spatial elements of oral tradition-references to place-names, landscapes, topographical 
features and the social organization of space-are crucial elements in the historical reconstruction 
of this region, rather than geographical background. While historians have often disregarded these 
elements as useless details they provide bits of evidence from the past, transmitted to the present 
because of their function in oral memory. Imagined landscapes, embedded in oral traditions as 
core images, are artifacts from the past that, although people might understand their meaning 

25 Marc Bloch, The Historian's Craft, trans. Peter Putnam (Manchester: Manchester 
University Press, 1954), p. 61; Feierman, The Shambaa Kingdom , p. 4; Miller, "Listening for the 
African Past." pp. 6-8. 

differently or lose their meaning altogether in different time periods, remain tenacious fragments of 
past social worlds transmitted in oral memory. 26 Like the ceramic artifacts that an archaeologist 
unearths, a particular shard may have been used in subsequent generations as a shallow water 
container for chicks in the yard or later picked up off the refuse heap by a child to be made into a 
toy. Still, the archaeologist can sometimes reconstruct its original use and historical context 
through careful comparison with similar shards found in other places, other kinds of artifacts found 
nearby, and contemporary pottery forms and their uses. 

These encoded fragments yield information about the past only as historians interpret them 
within their cultural context and alongside other kinds of evidence. For example, when 1 was 
trying to decipher what the lists of settlement names meant historically, most elders could tell me 
where the places were located but few could tell me anything else about these places. However, 
once I began asking to go visit those places, I soon realized that only people from particular 
lineages could take me there because these are the sites of important ancestral graves. This gave 
me some idea of the social groups that inhabited these settlements, information about residential 
mobility and why some knew the names and not these more precise meanings attached to the place. 
When we arrived at those places, elders told the stories of the rainmaker or prophet buried there. I 
observed how settlements at these sites might have been situated ecologically or in relation to other 
settlements. The same places might be mentioned in other traditions, a redundancy that provided 
valuable alternate meanings for the place. The place-names themselves could sometimes be 
translated literally, suggesting historical association with the place. Mapping these sites in relation 

26 See Glassie's poetic treatment of the Irish landscape as a mnemonic artifact in which the 
past is entombed in Passing the Time, pp. 621-65. See also Barber, 1 Could Speak , pp. 27, 34 on 
oriki praise poems as "fragments of the past." 

to each other showed how settlements were spatially related. The conclusions 1 draw from this 
evidence are not firm but do represent a logical set of possibilities. 

Core spatial images also yield important historical content because they refer to particular 
forms of social relationship or identity. In his study of social memory, Paul Connerton theorizes 
that "our memories are located within the mental and material spaces of the group." Individuals 
preserve memories as members of a group and these memories are situated within the socially- 
specific spatial framework provided by that group. 27 Memories are not only spatially located but 
also socially located within particular groups. We can identify each oral tradition with the history 
of a particular social unit. Different social groups located in one place may preserve radically 
different memories about the same time period because each builds on its own "mental map." 28 

Landscapes are not neutral backdrops to the events of history. They shape and are shaped 
by human action as people imagine and use the landscapes. 29 A classic argument, first articulated 
by Durkheim and Mauss in 1 903, holds that the built spaces we inhabit represent the structure of 
our society. Scholars have demonstrated this mainly in the layout of homestead and village as well 
as the interior design of houses. 30 In this dissertation, I extend this observation to hypothesize that 

21 Connerton, How Societies Remember , p. 37, Connerton draws on the early work of the 
French social theorist, Maurice Halbwachs. 

28 Ibid, pp. 20, 28. 

29 For the social theory of space see for example, Edward W. Soja, Postmodern 
Geographies: The Reassertion of Space in Critical Social Theory (London: Verso, 1989); and 
Allan Pred, Making Histories and Constructing Human Geographies: The Local Transformation of 
Practice. Power Relations, and Consciousness (Boulder: San Franciscan, Oxford: Westview Press, 

30 Emile Durkheim and Marcel Mauss, Primitive Classification , trans, and ed. Rodney 
Needham (Chicago: Chicago University Press, 1963; first published in French, 1903.) Bourdieu 
enlarges this argument with his notion of habitus, Pierre Bourdieu, Outline of a Theory of Practice , 
trans. Richard Nice (London: Cambridge University Press, 1977; first published in French, 1972), 
pp. 1-71. For an application of spatial theory to African ethnography see the collection of essays, 

oral traditions encode social relationships and identities by employing a spatial imagery that 
includes landscapes and topography. Different ways of imagining the landscape in the spatial 
images of oral traditions correspond to different social identities that can be situated in the past by 
understanding their cultural context and by using other kinds of evidence. These spatial images 
both represent forms of social identity and persons adopting these social identities use spatial 
images to represent them. 

For example, oral traditions about the origins of clans describe each clan ancestor as a son 
of the founder of the ethnic group, within one territory. Yet, the spaces invoked by clan histories 
and praise names are dispersed, with each clan's spaces interspersed among those of other clans. 
This supports the inference, based on several kinds of evidence, that in earlier times, as in more 
recent ones, regional networks of trade and the movement of ritual experts, connected dispersed 
clan settlements. The same clan names with the same associations of place and ritual avoidance 
are found throughout the region in nearly every ethnic group. Oral traditions say that dispersed 
Hemba or Gaikwe clan members enjoyed access to trade with the Asi hunter/gatherers (who lived 
on the margins of the western Serengeti) through friendship oaths. Since clan networks like these 
do not function any more and because little evidence exists for them in the colonial period, one can 
reasonably assume that they belong to an earlier period. We cannot date precisely the period 
during which these clan networks functioned, but situating them relatively in time may be possible 
by mapping the regional distribution of clan names, words that refer to clan functions, and 
variations in lineage organization. 

Anita Jacobson-Widding, ed., Body and Space: Symbolic Modes of Unity and Division in African 
Cosmology and Experience (Uppsala: Almqvist and Wiksell International, 1 991 ); and Denise L. 
Lawrence and Setha M. Low, "The Built Environment and Spatial Form," Annual Review of 
Anthropology 19 (1990): 453-505. 


Even relative chronologies for the forms of identity to which the core spatial images refer 
are difficult to identify, because oral traditions consist of layers laid down in different time periods. 
Some have described oral traditions as resembling palimpsests, or tablets that various people have 
written over with the older writing just barely visible beneath. Christopher Wrigley, writing on 
Ganda "mythical traditions," says that oral traditions are like textbooks of law, "which are being 
constantly re-edited to reflect new knowledge or interest but keep the basic shape given them by 
their original authors."" Thus, origin stories may indeed contain both existential reflection on 
enduring realities of the human condition and very old historical content. Both of these contents 
may receive new interpretations when new identities, in relation to new communities, gain greater 
importance. In identifying which layer is older, the spatial images around which people tell the 
stories provide crucial clues. 

For example, in the stories about the origins of western Serengeti ethnic groups, the core 
spatial images are not the spaces of bounded ethnic territories. They are instead the gendered 
spaces of the homestead and the ecological spaces of economic subsistence strategies. In addition, 
people tell the origin story of first man, the hunter, and first woman, the farmer, throughout the 
region, not confined to one ethnic group. Each separate group creates its own elaborations on a 
common story. The historian can date the economic subsistence patterns and homestead layouts 
represented in these spatial images to the distant past through historical linguistics and comparative 
ethnography. Thus, one can reasonably argue that the core images of these origin stories are based 

31 Christopher Wrigley, Kingship and State: The Buganda Dynasty (Cambridge: 
Cambridge University Press, 1996), p. 49. Thanks to James Ellison for his archeological 
explanation of this concept: "Imagine a landscape inhabited, people dropping artifacts that reflect 
relations of exchange over a great distance and to the south. Then people die, the site is covered by 
eolian deposits, and other people move in who drop artifacts that reflect trade in another set of 
directions and at a much closer radius. They die. Winds deflate the sediments leaving these quite 
different artifacts side by side," personal communication, 23 June 1997. 


on the founding myths of much older social groupings that narrators have reconfigured into ethnic 
stories. Because the characters in these ethnic origin stories are sometimes members of particular 
clans, it also seems likely that people have used these stories to understand the relationship between 

In the search for historical meaning, this dissertation analyzes and contextualizes the core 
spatial images of each set of oral traditions. The relative age of these images may be suggested by 
interpreting their cultural meaning alongside of other forms of evidence and in light of commentary 
from other traditions. Through this process one can see how the different elements of a single 
tradition might have been added at different time periods. Origin stories may have begun as 
founding myths explaining the relationship between men and women and the interdependence of 
hunting and farming as economic strategies, or, the importance of farming in "civilizing" human 
families. Then particular variations of these stories about first man and first woman and their 
children became identified with the clans of those who first claimed the land. In this case the 
stories explained the relationships between clans and provided the basis for the priority in 
particular circumstances of some clans over others. Later, as ethnic groups began to define 
themselves, narrators turned the same stories that underwrote the rights of clans to tell of 

Multiple, situational, and relational identities seem to be an ancient pattern in this region, 
perhaps as a result of the interaction of diverse languages, subsistence patterns and cultures from 
earliest times. Even if particular sets of oral traditions tell of particular forms of social identity and 
in relation to a relative time periods in the past, other forms of social identity surely existed too. 
People who lived on the marginal lands of the western Serengeti built their security by forming 
extensive regional relationships based on many different kinds of identities, rather than centralized 
hierarchies. We cannot imagine a time when no lineages or clans, age- or generation-sets existed, 

but we can dimly perceive times when, in specific historical circumstances, regional clan networks 
played a more important role than they now do; when lineage was the idiom for asserting a 
relationship with the land and when the principle of age began to eclipse that of generation. The 
interpretation of spatial imagery thus provides tools for understanding these important shifts in 
social process over time. 

In a society where identities were multiple, situational and relational, many different sets of 
oral traditions, describing the diverse histories of many different social groups, operated at the 
same time. If a particular social group remembers and performs a particular set of oral traditions, 
its people can only preserve the particular histories and landscapes of that group and cannot carry 
the weight of a more integrated social history. One elder, who is a member of many different 
social groups, may relate, for example, ethnic origin history, clan migration history, generation-set 
ritual history, age-set initiation history, and lineage settlement history all in one session, with each 
kind of history situated in a particular temporal frame. Each of these separate histories has become 
one ethnic history in the memory of one man. Yet how can we understand the indigenous 
chronology that orders these different kinds of stories in one man's narration of the past in relation 
to each other and in terms of a linear chronology? 
Periodization of Oral Tradition and Concepts of Time 

The organization of this dissertation follows the indigenous chronology implicit in the 
corpus of oral traditions told today, beginning with the stories of ethnic and clan origins and ending 
with the stories of the late nineteenth century disaster and recovery. I do not accept this 
chronology in any absolute sense but contingent upon an understanding of the kinds of 
temporalities reflected in oral tradition. However, indigenous periodizations of oral traditions bear 
some relationship to the relative age of information about the past carried into the present. They 
are not purely imaginative reconstructions of the past based on the present. 


The historical consciousness of the western Serengeti, in spite of the preference for spatial 
rather than temporal organization of oral traditions, employs a division of time into chronologically 
ordered periods. Oral traditions can be grouped into three types of traditions according to this 
indigenous periodization. Ethnic origin stories of first man, the hunter, and first woman, the farmer 
refer to the oldest time period. They often include clan narratives in which the children of the 
ancestral parents disperse. Oral traditions of settlement sites or migration stories and accounts of 
rituals concerning the land characterize the middle period. The most recent period contains the 
historical accounts of the generations from the last two decades of the nineteenth century onward. 
These traditions tell of the disasters of that period and the ways in which people coped and even 
began to prosper in the early colonial period. We can understand only these last stories as 
"historical" in so far as they employ a linear chronology of past events ordered by cycles of age-set 
or generation-set names. 

For the last or historical period, narratives are grouped according to the memories of a 
particular "generation." Those who were in their youth at the time of the disasters refer to this 
experience as the formative point for their generation. Those in the generation who were in their 
youth during the early colonial years oriented their identity around the cattle wealth acquired from 
new opportunities for trade. Rosaldo used the concept of "cohort analysis" to find structure in the 
individual biographies of age-peers set in a particular historical context. As he followed many of 
these lives through time he saw that age-peers began defining themselves by formative historical 
experiences when their "shared collective identity was formed." This then became the enduring 
characteristic of a self-conscious group. 32 

32 Rosaldo, lloneot Headhunting , pp. 1 10-113. 


The division of time into three periods in oral tradition corresponds closely to what oral 
historians have found elsewhere in Africa. Rich oral traditions typically surround the period of 
origins and that of recent times, separated by a "floating gap" of scarce and cryptic information in 
the middle period. In the western Serengeti, the narratives of this middle period contain lists of 
place-names referring to settlement sites or stops on a migration route. Earlier historians of oral 
tradition in Africa accepted this indigenous periodization as representative of relative chronological 
time. Vansina assumed that these bodies of tradition referred to successive time periods and he 
explained the "floating gap" by the fact that beyond a certain time depth, oral memory reaches it 
limits. As traditions got older, they became more generalized and mythologized. 33 

Anthropologists met this positivist method of ordering oral traditions into a chronological 
framework with an unsettling critique. Structuralists demonstrated that these three types of 
tradition do not refer to time at all but to structures of society in the present. They saw the 
mythical accounts as "founding myths" that justify the existence of the present social system. For 
these scholars the middle period represents a static model or "social charter" of the same system 
and only in the recent period do oral historians offer accounts of change and pose explanations as 
to causality. 34 While this understanding of indigenous periodization challenged the possibility of 
historical reconstruction, many historians went on to demonstrate that even as "mythical charters," 
oral traditions still contain evidence about the past. 35 Some historians, such as Feierman, 

33 Vansina, Oral Tradition as History , pp. 23-24. 

34 For a classic example of the structuralist interpretation of oral traditions see. Claude 
Levi-Strauss, "The Story of Asdiwal," in The Structural Study of Myth and Totemism . ed. 
Edmunch Leach (London: Tavistock Publications, 1967), pp. 1-47. 

35 Jan Vansina, Oral Traditions as History , p. 23; Spear, "Whose History?," pp. 165-181; 
Thomas Spear, Kenya's Past: An Introduction to Historical Method in Africa (Harlow, Essex: 
Longman, 1981); Miller, "Listening for the African Past," p. 4. 

accommodated the critique by incorporating structuralist analysis into their historical 
reconstruction. 36 

Thomas Spear picked up this challenge to the historicity of oral tradition by finding a 
different basis for maintaining that the three kinds of traditions represented time periods (recent, 
middle and early). He showed that each kind of oral tradition is characterized by a different sense 
of time (linear, cyclical and mythical respectively). His own work demonstrated that Kenyan 
Mijikenda oral traditions of the earlier periods "contain accurate historical narrative over four 
centuries and continue to describe accurately institutions that have not existed for 130 years." The 
different time sense in each type results from increasing abstraction, changing in "meaning from 
literal to intended to symbolic as the events they recall recede into the past." Thus oral traditions 
of the early period describe "things-as they became," the middle period describes "things-as-they- 
should-be" and the late period describes "things-as they-are." 37 

Although a rough correspondence exists between these three types of oral traditions and 
three levels of historical time depth it would be a mistake to accept this periodization as 
corresponding to any absolute sense of chronological time. Any one of these types contains 
material from many different time periods. The function of oral traditions locally is not to archive 
the past in any "pure" form. Narrators have transmitted traditions because they provide useful 
information for negotiating the paths of present social relationship based on the experience of the 
past. 38 They represent wisdom from the past distilled into an idealized form. 

36 Feierman, The Shambaa Kingdom , pp. 40-69. 

37 Spear, "Whose History," pp. 167, 171-2. 

38 See Feierman, "Introduction," The Shambaa Kingdom , pp. 10-16. 


However, this indigenous periodization of oral traditions according to a relative chronology 
does give some indication of time depth in that the spatial images of each genre can be correlated 
with a particular temporality (rather than time period). The core spatial images of the origin stories 
refer to ancient patterns of production and reproduction in the region, as they had developed out of 
the gendered interactions of hunters, herders and farmers in the deep past and as they have 
continued to exist in more recent periods. The lists of settlement site names refer to historic 
residence patterns in the not-too-distant past but before the more immediate memories of the late 
nineteenth century famines. 

We can be reasonably sure that oral traditions do contain information about the past 
because of the remarkable congruence between historical reconstructions based on the core spatial 
images of oral tradition and that based on the evidence of historical linguistics, comparative 
ethnography, archaeology or written sources, which we assume to have historical validity. If those 
who tell oral traditions cannot have known about this other evidence how otherwise would they tell 
such similar stories concerning social processes in the distant past? Oral traditions provide a 
culturally grounded expression of historical processes partly available to us from other kinds of 

The historians must also accept the limitations of oral traditions. The genres of oral 
tradition corresponding to "mythical time" and "social process time" cannot by themselves show 
change over time. Only the comparison of various versions of the same types of traditions 
throughout the region or among different social groups can accomplish this result. 39 Even then 

39 See Matthew Schoffeleers, "Oral history and the retrieval of the distant past: On the use 
of legendary chronicles as sources of historical information," in Theoretical explorations in African 
religion , eds. Wim an Binsbergen and Matthew Schoffeleers (KP1: London, 1985), pp. 164-188. 

these lead only to tenuous hypotheses, usually confirmed by written sources (that are not available 
here before the late nineteenth century). 

Traditions of the oldest indigenous time periods supply a description of the spatial 
organization of social relationships over the long period in which they unfolded. 40 The core spatial 
images of these traditions represent the underlying principles, or what Pierre Bourdieu calls the 
"generative schemes," of social practice, a practice elaborated and improvised on in daily life, 
according to the specific context. Generative schemes are not so much "rules" as "strategies" by 
which people choose from a range of options to enact social practice. 41 These are the mental 
dispositions inculcated during the earliest phases of socialization, inscribed in the body and 
cultivated in the routines of life. People do not consciously understand how the system works but 
unconsciously know how to make it work. The basic principles by which one knows how to carry 
out relationships based on common parentage, age, generation, expertise or wealth can generate 
unlimited practical applications depending on the situation. For example in the western Serengeti, 
a stranger could be incorporated as a son to secure additional labor; peers, as age-mates, could be 
organized on a regional scale for raiding or reconciliation; wealthy men could be induced by 
followers to provide large communal feasts to gain legitimate respect; and networks of friendships 
sealed by blood oaths could be fashioned into long-distance trading partnerships. People make and 
maintain relationships when they are useful either materially or symbolically, rather than existing 
as the result of a disembodied social system. Social maps are not reified representations but 
various and changing according to the context. 

40 Randall M. Packard, "The Study of Historical Process in African Traditions of Genesis: 
the Bashu Myth of Muhiyi," in The African Past Speaks , ed. Joseph C. Miller (Folkestone, Kent: 
Dawson Archon, 1980), pp. 157-177. 

41 Bourdieu, Outline of a Theory of Practice , pp. 1-71. 


Oral tradition rationalizes these generative principles that govern the daily elaboration of 
social practice into an "official" version of social organization to preserve the existing social order. 
Oral traditions turn what is a dynamic process into a static model. When we understand lineage as 
a strategy rather than a structure, we see people's everyday actions as significant because they are 
making choices rather than following a script. The core spatial images of early-period traditions 
present social organization in a rigid and timeless, "traditional" form. Although this is the "given" 
spatial text, people have "read" it in countless ways over time.' 12 The exclusive narration of the 
"official version" has now silenced some of these various interpretations. By understanding how 
various people have "read" these spatial "texts" since the time of documented historical sources, 
one can imagine a similarly varied spectrum of possibilities in the past. Another way of 
understanding the possible past readings that are now silenced is to look at the regional variation on 
these basic generative themes of social organization. Different historical contexts influence the 
regional variations. 

An example of contextual "readings" of generative principles, reflected in the oral 
traditions as a static model, concerns gender relations represented in the spatial layout of the 
homestead. The house is represented as a female domain and the courtyard and cattle corral as a 
male domain. The oppositions of inside/outside, enclosed/exposed, passive/active are embedded in 
these spatial configurations. Through historical linguistics and comparative ethnography the 
historian can show that these gendered spatial patterns are very old and still exist today. However, 
from evidence in the colonial record and from my own observation, women have clearly turned 
these dichotomies to their own advantage and found ways to cross and blur the boundaries. 

42 Ibid; Moore, Space. Text and Gender, pp. 79 - 86, who suggests that spatial 
organization is like a "text" that can be "read." 

Embedded in the origin traditions (primarily as told by men) are hints of the ways in which women 
and men have reinterpreted these spaces over time. 

The historian might better understand the three types of oral traditions, not as referring to 
different time periods or concepts of time, but to different temporalities or time scales. Because we 
can connect each set of oral traditions with a particular social group and the material spaces that it 
occupies, each of these identities also has its own temporality. The indigenous chronology places 
oral traditions connected with identities based on gender and subsistence economy in the oldest 
time frame, almost out of time. Comparison to other evidence shows that these identities are very 
old and relatively stable over time. The earliest traditions represent a different temporal frame of 
very slow changes over long time periods and enduring social forms. The origin stories do not 
objectively represent particular events and years but they do adequately represent long-term social 

This insight draws on Braudel's classic history of the Mediterranean in which he uses three 
different temporalities: the longue duree (history of imperceptible changes in the relationships of 
man to his environment), the conjuncture (the slow but perceptible rhythms of social process) and, 
the evenemenl (the short term political time of remembered history). 43 Braudel's model allows for 
each type of analysis to supply a different kind of historical information through its own time 
frame. Each temporal scale implies a corresponding spatial scale and social unit. A narrative 
concerning the relationship of man to the environment demands a deep time scale and an extensive 
ecological field. In Braudel's analysis the geographical spaces of mountain and sea where people 

43 Fernand Braudel, The Mediterranean and the Mediterranean World in the Aee of Philip 
U, trans. Sian Reynolds and ed. Richard Ollard (Harper Collins: London, 1992; first published in 
French, 1949.). 

fished and farmed were set in the history of the long duree, while the courtly spaces of kings were 
the venues of short term diplomatic history.' 14 

However, if we analyze separately the temporality of each set of social identities, rather 
than the total society for each time period, than we cut social history on a "vertical" rather than a 
"horizontal" axis. In effect, each of these temporal slices represents a particular kind of social 
identity, over an ongoing period of time, rather than a picture of an integrated society during one 
time period. This moves the analysis from a simple past tense to a past continuous tense. For 
example, the space/time frame of origin stories refers to the demonstrably ancient strategies of 
subsistence and social reproduction that are still practiced today. These are basic generative 
schemes around which other social organization is elaborated. The space/time frame of the stories 
of settlement sites and land rituals refers to processes in the not-too-distant past by which people 
established the rights to the land that exist today. The space/time frame of stories concerning 
nineteenth century disaster and opportunity are within a chronological framework of historical 
change and refer to newly evolving forms of social organization— age-set territories and ethnic 
identities. Each space/time frame moves forward at its own rate to the present. 

In reality, one individual simultaneously embodies each of these various identities 
represented by oral tradition. These diverse identities cannot be separated and are functionally 
interdependent. One cannot understand a woman outside her role as farmer, mother, sister or 
healing expert. A scholar cannot divide the history of an integrated society for analytical purposes 
into the separate histories of gender, economy, clans, age-sets or generation-sets. Braudel 
recognized that this dissection of history into "geographical time," "social time" and "individual 
time" was only a device that "divided man into his multiple selves" and "cut across living history 

" Prins, The Hidden Hippopotamus , p. 8, uses Braudel's model in his attempt to "study all 
the time scales of change" and their interplay with each other. 

that is fundamentally one." 45 However, by making this "vertical" dissection one better appreciates 
the historical dynamics of social identities that were multiple, situational and relational in nature. 

Using this method, the oral traditions retain their integrity, allowing for a translation that 
takes seriously local historical consciousness. Explaining nineteenth century changes in social 
identity without understanding the previous forms of social identity that operated over the long- 
term is impossible. Yet to understand these long-term processes without understanding the 
disasters of the late nineteenth century, through which oral traditions about the earlier periods are 
filtered is also impossible. Throughout the dissertation, I interpret oral traditions about the earlier 
period in light of the nineteenth century transformations in social identity, and understand 
representations of social organization in the nineteenth century as products of earlier cultural 
formation. In each chapter I present a different genre of oral tradition, explore its core spatial 
images and the references to social identity in its cultural and historical context. 
Oral Traditions and Social Identity 

People in the western Serengeti who tell oral traditions today refer explicitly to the social 
identity of ethnicity. Yet the stories themselves represent a variety of other kinds of social 
boundaries and they operate within a common set of regional assumptions, or generative principles. 
Similar stories for similar periods in each of the ethnic groups with whom I conducted interviews, 
suggest the possibility of a common regional history, with local variations corresponding to 
differing historical and geographical contexts. Telling this story as seven different ethnic 
narratives may have better represented western Serengeti historical consciousness in the colonial 
and postcolonial periods, but it would not fairly reflect precolonial consciousness. 

45 Braudel. "Preface to the First Edition," The Mediterranean, p. xiv. 


These traditions show that social identity, in the past as well as in the present, is multiple, 
situational and relational rather than unitary and fixed. None but the most recent traditions use the 
space of the ethnic group as the core spatial image. Instead, the core spatial images that represent 
the identities of male/female, farmer/hunter/herder, lineage or clan member, age- or generation-set 
member seem to reflect older historical knowledge. The social identities based on locality, rank, 
authority, expertise, or wealth also figure prominently in these traditions. People today take on 
different identities depending on the identity of those to whom they are talking and the situation in 
which they find themselves. 
Personal Names as an Illustration of Relational and Situational Identity 

Practices surrounding personal identity and naming show people deploying multiple and 
situational identities to negotiate and strategize their interests. A quality not unique to, but 
certainly characteristic of, this region is the use of a multiplicity of names for any one person. A 
man may take a personal name, an ancestral name, a teasing name, a father's name, a mother's 
name, a clan name, a kinship term of address, a praise name, a youth name or an elder's name at 
different times in his life. Only certain people may have the right to call him a particular name. 
No single name fixes a person's identity. Young people use different names when they apply for a 
job or sign up to retake an exam. One of the first missionaries in the region reported that when he 
began teaching local children at his home the children, having been instructed at their homes not to 
give their proper names, called themselves "hammer," "saw," "stone," cartridge," and "war." 46 
After a person has a child, people call him or her in reference to the child's name, "Mama Bhoke" 
or "Baba Mwita" (in Swahili). Names position a person within particular kinds of social 

46 Valemar E. Toppenberg, Africa Has My Heart (Mountain View, California: Pacific 
Press Publishing Assoc, 1958), pp. 52-53. "The parents believed that we would have some magic 
power over them if we were told their proper names." 


relationships. Each lineage relationship has a name, age-mates call each other by another name, 

those who have been circumcised together or done one of the eldership titles together call each 

other by yet another name. 

A partial list of these names that individuals with specific relationships would call one 

another was related to me in this way by a Nata elder: 

Bashori; the name for the one with whom you were circumcised, putting oil on together. 

Barogumu: the name for a "friend of oil, "for whom a goat was killed. 

Banchabo: the name for the person who sponsors you in achieving the Titinyo rank. 

Banagera: the name for the person who runs behind you in the Titinyo ceremony. 

Baguruki: the name for the one who sponsors you in achieving the Aguho rank. 

Semung'anta: the name for the one brings you into the Eghise rank. 

Omusani wa kusaragana: the name for your blood brother from another ethnic group. 

Barera: the name for the man who marries your first daughter. 

Basigero: the name for the man who marries your second daughter 

Babusheni: the name for the one who marries your third daughter. 

Bachoro: the name that the wife of your first son uses to greet her father-in-law. 

Bangondu: the name that a daughter-in-law uses to greet her father-in-law. 

Babogusi: the name that elders of the same rank call each other at a feast. 

Bagechoncho: the name for the one who takes off the headdress in the Aguho feast. " 

These names could be mutually used whenever people met, not only in the particular situation to 

which it refers. 

Greetings often refer to the relationship to the person being greeted, rather than to a 

personal identity. The younger or subordinate of the pair calls out the relationship in the greeting, 

"sister-in-law," "mother," or "paternal aunt." The person responding calls back the same name and 

then the greetings go on as to the day, the cattle, the fields and the health of those at home. People 

of the same generation most often greet each other as brothers or sisters. One can characterize 

each person she meets as belonging to her own generation, the generation of her children, parents or 

grandparents. Therefore, if she does not know what her particular family relationship is to this 

person, she can greet her as "mother," "grandmother," or "daughter." That person would return the 

47 Interview with Megasa Mokiri. Motokeri, 6 March 1 995 (Nata cf ). 

same greeting. I would greet a very old woman as "grandmother" and she would return the 
greeting "grandmother," not "granddaughter," as it names the relationship rather than the person. 

Because any one person in the community may have multiple relationships to any other 
person, much latitude exists for "playing" with greetings and names. Depending on an individual's 
dealings with this person, she might want to evoke a relationship of superiority (older generation), 
comradeship (age-mates), obligation (a paternal aunt) or joking respect (brother-in-law). Since 
names are multiple, they are open to negotiation. Someone may refuse to greet a person as they 
were greeted, demanding a greeting of greater respect. A discussion over the basis of a claim to be 
greeted in this manner would then result, involving complicated genealogical or historical accounts. 
Since we were incorporated into a family in Nata, my own children often and twelve years could 
demand the greeting of "father" from youth twice their age according to the principle of generation 
(the father of these youth was a grandson of Magoto in the same way that my children were 
considered "grandsons" of Magoto, and thus their equivalence). The point of all this seems to be to 
acquire as many names, and thus social relationships, as possible, in the same way as other people 
might acquire material goods. 48 

Relationship and situation define personal identity. Who I am depends on whom 1 am 
talking to, my relationship(s) to him or her, and the context in which we find ourselves, including 
what each of us wants out of the interaction. Social identity is relational and situational rather than 
immutable and unitary or defined by an ontological sense of undivided being. Dr. Mekacha. a 
Nata professor of linguistics at the University of Dar es Salaam, told me that he is called a 
different name by his parents, his grandmother and his age-mates when he is home-besides being 

48 Just as Guyer and Belinga have postulated "wealth-in-knowledge" as a refinement of the 
"wealth-in-people" paradigm, this suggests a further elaboration of "wealth-in-relationships." Jane 
1. Guyer and Samuel M. Eno Belinga, "Wealth in People as Wealth in Knowledge: Accumulation 
and Composition in Equatorial Africa," Journal of African History 36 (1995): 91-120. 

named as an educated man from Dar es Salaam. In terms of social identity, when he is in Nata 
people identify him by his lineage or place of residence. When he is in Musoma, people identify 
him as Nata. When he is in Dar es Salaam, people identify him as Kuria. 49 Each of those 
positional identities is dependent on where he is and to whom he is speaking. 

A person can never assert his social identity in a vacuum. He must state it in relation to 
someone else and to a specific space and time. It is only in Dar es Salaam that people name Dr. 
Mekacha as Kuria. If someone would call him Kuria in Musoma, he would be deeply offended 
because of the connotation, for others in the Mara Region and throughout Tanzania, of Kuria cattle 
theft. What it means to be Kuria today is different from what it might have meant, if the term was 
used at all, at the beginning of the century when people admired and imitated the Kuria for their 
courage and skill in cattle raiding. The boundaries of social identity are constantly in flux. 
Social Boundaries 

Each of these identities defines boundaries with people who are different and, at the same 
time, defines reciprocal obligations with those who are the same. Identity only functions in 
reference to what it is not and within an arena in which people define difference. Although farmers 
in the western Serengeti also hunted and farmed, they defined themselves in distinction to hunters 
and herders in order to establish relations of interdependence between communities which practiced 
different subsistence economies. Multiple identities flourished because they were the most 
important social resource in a harsh environment where land was plentiful and people scarce. It 
was only by successfully calling on the bonds of reciprocity within various groups that a family 
could survive a drought or grow wealthy and powerful. The way to authority was not a vertical 
movement through a hierarchy but a set of horizontal movements aimed at creating and 

' Interview with Dr. Rugatiri Mekacha, Dar es Salaam, 24 May 1996 (Nata cf). 

maintaining intricate networks of relationship. 50 Jane Guyer and Samuel Belinga describe this as a 
"compositional," rather than an "accumulative," process for gaining "wealth-in-knowledge" 
through various social relationships." 

By looking at each of these social identities as it is portrayed in oral tradition, each with its 
own organization of space, I hope to illuminate the complexity and creativity of noncentralized 
societies. Without centralized hierarchies of chiefs or kings (but certainly not without hierarchy) 
western Serengeti people responded creatively to the stresses of the late nineteenth century. They 
fashioned new solutions out of the social resources at hand. Defining a multiplicity of identities in 
order to spread out social capital and minimize risk also maximized opportunity. The peoples of 
the western Serengeti not only survived the nineteenth century disasters but they forged an era of 
unprecedented cattle wealth in the early colonial years. 

Because prosperity in this marginal environment rested on these identities and the networks 
of reciprocity that they composed, oral traditions preserved the knowledge necessary to maintain 
these relationships. We can understand oral traditions as "mental maps," or spatial representations 
of social relationships necessary in daily life. 52 "Social maps" are a metaphorical way of 

50 Miller uses a similar model to explain the trade networks of the slave trade, Joseph C. 
Miller . Way of Death: Merchant Capitalism and the Angolan Slave Trade 1730-1830 (Madison: 
University of Wisconsin Press, 1988). 

51 Guyer and Belinga, "Wealth in People," p. 103. 

52 For some of the earliest work on "mental maps" see, Peter Gould and Rodney White, 
Mental Maps (Baltimore: Penguin, 1 974); David Lowenthal and Martyn J. Bowden, Geographies 
of the Mind: Essays in H onor of John Kirtland Wrig ht (New York: Oxford University Press, 
1976). For an application of "mental maps" in African history see, Barrie Sharpe, "Ethnography 
and a Regional System: Mental Maps and the Myth of States and Tribes in North-Central 
Nigeria," Critique of Anthropolop v. 6, 3 (1986):33-65. 

representing social relationships and also a description of the organization of physical space." 
Narrators transmitted and maintained these maps because they distilled into spatial form the 
wisdom of the older generation who had already walked the paths of these relationships. Yet in the 
transmission of oral tradition, narrators described, as well as reinterpreted, the spaces of social 
relationship each time they told the stories and lived them out in daily practice. 

Each kind of oral tradition represents a different mental map to a different set of social 
identities, corresponding to a particular temporality and space in the historical consciousness. It is 
only by overlaying these maps that the shape of regional social history emerges. I define a region 
as the geographical extent of a historical set of significant social relationships that change over 
time. 54 These mental maps are the very framework within which people store group memories and 
without which social memory ceases to exist. The transmission of these maps to a new generation 
reproduces social identity and the new generation in turn represents these identities in its 
interpretation of the past through the present. This study necessarily encompasses a larger 
geographical area than that of a single ethnic unit because only within a larger region is it possible 
to see multiple maps on multiple scales representing multiple identities in action. 

Research Methodology 

Historians have recently been criticized for handling oral tradition as a "container of facts" 
using a "documentary model," rather than as living discourse." As a result historians must now 

" Shirley Ardener, ed., Women and Space: Ground Rules and Social Maps (New York: 
St. Martin's Press, 1981), pp. 13-14; and Hofmeyr, "We Spend Our Years " pp. 160-166. both use 
the concept of a "social map." 

" For an analysis of "regions" in terms of the "mental maps" of oral tradition see Jan 
Bender Shetler, "'Region' as a Historical Product: Mental Maps of Western Serengeti Oral 
Tradition," a paper presented at the American Historical Association, Seattle, 1998 and to be 
included as an essay in a forthcoming book on African Regional History. 

" Newbury, "Contradictions at the Heart," p. 32. 

look carefully at their own subjectivity as researchers and participants in the process. To evaluate 
the historical analysis of this dissertation the reader must understand how the oral texts were 
created in the interview process, the place of the researcher, and how these texts were then used in 
conjunction with other kinds of sources. 
Situated Knowledge 

Rather than taking the vantage point of the all-seeing observer, this dissertation "sees" 
from the position of a guest in the community trying to understand how people conceptualize their 
pasts. This is what Donna Haraway calls "situated knowledge" in which only a "partial 
perspective" is claimed, allowing for other interpretations without the splitting of subject and 
object. The story told here is the result of a particular set of interactions, conversations and 
relationships between myself and people from the western Serengeti who agreed to share their 
knowledge with an outsider. I do not claim to speak for or on behalf of them, but only to bear 
witness to what I experienced and to what I could "see" from this vantage point. 56 Another way to 
understand the knowledge that this dissertation represents is as a dialogue between myself and 
those I encountered in the western Serengeti, Musoma, Mwanza, Dar es Salaam, as well as my 
colleagues and professors at the University of Florida and the University of Dar es Salaam, 
interaction with the literature and those people with whom I live and work everyday." 

However, because this project was conceived and carried out as a collaborative project 
with so many colleagues in Nata, Ikoma, Ishenyi, Ikizu, Ngoreme, Zanaki, Ruri, Kuria and even 

56 Donna J. Haraway, Simians. Cv borgs and Women: The Reinvention of Nature (London: 
Free Association Books, 1991), pp. 183-201. 

57 For the concept of research as dialogue see Dennis Tedlock and Bruce Mannheim. eds„ 
"Introduction," The Dialogi c Emergence of Culture (Urbana and Chicago: University of Illinois 
Press, 1995), pp. 1-20; See also Dennis Tedlock. The Spoken Word and the Work of 
Interpretation (Philadelphia, University of Pennsylvania Press, 1983). pp. 285-338. 

Sonjo, I am obligated to try to present an account that will do justice to their efforts. I have laid 
out some of these personal commitments in the Foreword. The constraint of these other voices on 
my work has been a constant source of inspiration as well as a reminder of the scholar's larger 
social obligations in a world of injustice and inequality. It is my hope that my readers will hear the 
voices of those in the western Serengeti who contributed their knowledge above the din of my own 

Although 1 have attempted to understand the past as much as possible from the perspective 
of those in the western Serengeti, in terms of their categories, I cannot claim to present this past 
knowledge in an unmediated form, in which the oral traditions are left to "speak for themselves." 
The conceptualization of this material into a form contained in an academic historical argument is 
obviously mine. I interrogate the historical knowledge of the western Serengeti with all available 
professional and critical tools while striving to foreground the understandings contained in the oral 
traditions themselves. As the narrator is located in a particular place of knowledge, conversations 
with other forms of knowledge are not only possible but necessary for understanding, since all 
knowledge is ultimately relational. Although the dialogues that occurred with people in the western 
Serengeti are foundational to this project, other dialogues in Dar es Salaam and on this side of the 
ocean both before and after the research experience have significantly shaped the outcome. I also 
have access to information, through the linguistic analysis of words and the ethnographic analysis 
of cultural traits, that local people do not understand as evidence about the past. This "evidence in 
spite of itself allows an outsider to construct bundles of cultural meaning which local people take 
for granted. 

"Situated knowledge" also does not presume that the position from which history is "seen" 
is innocent or unitary. All knowledge about the past is contested and deeply implicated in relations 
of power. Attempts to see from the perspective of local historical consciousness require us to look 

for the relations of power behind the texts and to comprehend how people use these texts in the 
struggles of everyday life. The multiplicity of knowledge implicit in Mayani and Nyawagamba's 
list of historical topics discussed in the Foreword demonstrates that no single perspective can do 
justice to all social positions. Conflicting positions are often evident even within one text. My 
position of power as a white, Swahili-speaking, Christian, American woman (wife and mother) was 
certainly a factor both in eliciting and interpreting these narratives. 

Throughout my research I encountered suspicion about why I wanted to gather this 
information and how I would use it. My assistants usually explained my situation to those 1 
interviewed, often when I was not present. Most of the elders agreed to share their knowledge with 
the hope of a book from which their grandchildren could learn about the history and culture of this 
region. Many were sure that 1 would make lots of money from this knowledge. At the heart of 
their suspicion was the sense that knowledge about other people gives one power over them. As I 
learned more, my questions began to reflect this knowledge, to which the elders often responded 
with a nervous laugh and an aside to my assistant, "how does she know that?" Elders respected 
this knowledge and took me more seriously in later interviews, resulting in more careful and 
detailed answers. On the other hand the elders also became much more wary and certain areas 
were deliberately hidden. By the end of the interview process I understood most of what people 
said in the local languages, since they are all closely related. Nevertheless, I still felt more 
comfortable asking the questions in Swahili. Therefore, some answers in local languages were 
directed toward my assistants, containing information that they would not otherwise have shared 
with me. 

A general suspicion existed that an American would not bother living in the "bush" unless 
she had some ulterior motive. I heard many possibilities suggested. We lived near the park and 
since people were always suspicious about new plans to take their land away for park expansion, 

they wondered if the park had sent me to gather information for this purpose. Since cattle raiding 
and game poaching are large security concerns of the government in this area, people also met 
these topics with some caution. Some thought we were prospecting for gold. A deeper and more 
vague concern was that, with rising U.S. power at the end of the Cold War, Americans would re- 
colonize Africa by first learning the secrets of its people. Just as I was doing my main interviews 
Tanzania was getting geared up for the first multi-party elections, so some people assumed that 
some party or candidate in the elections had paid me to promote their cause. Many communities in 
which 1 did interviews had known me as a Mennonite Church development worker (1985-91). 1 
rented the car I drove from the Tanzania Mennonite Church that had the church name written on 
the door. The only time anyone totally denied me an interview was not because of my church 
connections but because I was white. The elder said: 

These people are like God, they know where the sun goes at the end of the day because 
they can follow it with their airplanes: people like that are too powerful to be messed 
with, you cannot predict what they will do with the information that you give them. 58 

The single most important factor in calming these fears and suspicions seems to have been 

that I am a woman, married with half-grown children. The general assumption is that women are 

straightforward, open and willing to work for altruistic reasons, while men always have an ulterior 

and deceitful motive, for personal benefit. Only women with children are treated as adults. In the 

system of generational relationships people in alternate generations have the most intimate 

relationships, while those in adjacent generations maintain a more distant and formal relationship. 

Some of the men I interviewed considered me as their grandchild but more often I was considered 

in their daughter's generation. Thus the interviews were formal but very serious and 

58 Interview with Simora Nyamotoma, Robanda, 12 July 1995 (Ikoma <r). 

straightforward. Interviews with men who treated me like a granddaughter sometimes produced 
unexpected information but were harder to keep on track." 

Tanzanian historians Buluda Itandala and Isaria Kimambo also reported dealing with 
similar issues of suspicion in their dissertation research. Kimambo stated that old men were loathe 
to discuss ritual practice openly because "such information formed the secrets of their society- 
secrets that they revealed only to initiated members of the society." He slowly broke through this 
resistance by using the little information that he gained from those who trusted him to learn to ask 
the right questions of the elders. Both commented that they only began to get inside information 
when they used "well-known and trusted local people" to introduce them to the informants. 60 The 
personal bond of reciprocal relationship allowed for the possibility of divulging secrets. 

A history told from the standpoint of "situated knowledge" may differ in its interpretation 
from the sources of its knowledge. The Nata, Ikizu, Ikoma and other elders who told me these 
stories of the past may not agree with my rendition. I must take into consideration a historical 
consciousness that the transformations of the late nineteenth century have seriously altered. I must 
in a sense privilege the knowledge of the ancestors and their ways of knowing over that of the 
living, seeking out the specific ways in which the ancestors live in the present and learning from 
them. If the pact of obligation for a fair rendition of the past is with anyone, it is to them, and their 
way of imagining the past, that I am held accountable. 1 strive to bear witness to each of these 
concerns, knowing that I am a product of my own multiple identities situated in time and space. 

Similar generational interactions during his research are described by Rugatiri D. K. 
Mekacha, The Sociolinguisitic Im pact of Kiswahili on Ethnic Community Languages in Tanzania: 
ACaseStudvofEkinata (Bayreuth, Germany: Bayreuth African Studies, 1993), p. 51. 

60 A. Buluda Itandala, "A History of the Babinza of Usukuma, Tanzania, to 1890" (Ph.D. 
Dissertation, Dalhousie University. May 1 983), p. 9. Isaria N. Kimambo, A Political History of 
the Pare of Tanzania, c. 1500-1900 (Nairobi: East African Publishing House, 1969), p. x. 


Oral Sources 

This research, conducted over a period of eighteen months in Tanzania (January 1995-July 
1996), was primarily concerned with the search for the remains of this previous historical 
consciousness in oral traditions told today. Although people reworked oral traditions to fit new 
identities during the late nineteenth century disasters and the subsequent colonial era, they did not 
obliterate earlier understandings. Many versions of each kind of oral tradition remain, without a 
central dynastic account. The narration of oral tradition resembles a conversation rather than a 
recitation and history emerges in a process of dialogue. No formalized genres of oral literature 
exist, except perhaps the recitation of praise names and proverbs. Almost no local popular 
histories appear in print that would influence the content of oral tradition. Many versions of the 
same stories from different social positions and among different ethnic groups were collected for 
comparison and cross reference. 

I collected historical narratives among five different Bantu-speaking peoples (Nata, Ikoma, 
Ishenyi, Ngoreme and Ikizu) and two different Dadog-speaking peoples (Rotigenga and Isimajek) 
of the western Serengeti, as well as a few interviews among neighboring peoples (Sonjo, Kuria, 
Sizaki, Zanaki and Ruri). 61 I asked open-ended questions, trying to solicit all kinds of knowledge 
about the past, without confining it to the expected ethnic narratives. The interviews were often 
four hours in length, following the interests and knowledge of the elder rather than a set list of 
questions. After many interviews among many ethnic groups I found repetition of some stories, 
with variations, as well as references to similar social and cultural institutions of the past. From 
these similarities and variations a wider regional history began to come into focus. When I put 
these accounts side by side, similar experiences and forms of organization emerged in 

1 See References Cited for list of informants, ethnic origins and other personal data. 

corresponding generations across ethnic lines. The elders were telling a common history as seven 
unique histories. 

The Mara Region is an ideal venue for this kind of regional comparative study of oral 
traditions because so many different ethnic groups exist within a small geographical area, each 
claiming its own unique identity and history. It is also a region in which colonial penetration 
happened late and incompletely. 62 A much different process was underway here than in 
neighboring areas, where the emergence of a pan-Sukuma, pan-Luo and pan-Maasai ethnic identity 
became an important political force, with clear implications for the transmission of historical 
knowledge. Without a tradition of chiefs or hierarchical leadership, the Mara Region also differs 
from the kingdoms of the Great Lakes Region where "big" dynastic history often overshadowed 
"little" commoner or clan histories. By the same token, no single, centralized tradition exists to use 
in comparison with marginal traditions. 

A strategy employed in the research was to collect historical narratives from different 
positions in the social network. Elders who had reached the top titles in the nyangi system, leaders 
of the age or generation-sets, lineage or clan leaders, rainmakers and prophets each had a slightly 
different version of the past that highlighted the power and authority of each of these kinds of 
social relationship. Although colleagues usually took me to see people of respected position, I also 
tried to talk to ordinary participants in these social organizations and to people in their roles as 
fanners, hunters, herders, specialists and household members on a more informal basis. 

I also began to collect manuscripts of ethnic histories written by local intellectuals, often 
primary school teachers or government clerks. They usually handwrote these in school notebooks 

62 For an analysis of this process see Kirsten Alsaker Kjerland, "Cattle Breed; Shillings 
Don't: The Belated Incorporation of the abaKuria into Modern Kenya" (Ph.D. Dissertation 
University of Bergen. 1995). 

and stored them away in trunks. One Ngoreme man who worked in the parish church office had 
his manuscripts typed and duplicated but I had trouble finding even one complete copy twenty 
years later. These accounts, similar to "encyclopedic informant" narratives, were compilations of 
histories learned from elders and written in the same style. I am trying to get some of these 
histories published in Tanzania for local use. 63 

Ethnographic participant-observation in a rural community and an informal household 
survey in one village augmented this study. 64 I went with the leaders of the different groups to the 
important historical places on the landscape, hearing the stories connected to those sites and 
observing their location, ecology and situation. All of this information added depth and insight to 
the historical processes described in oral tradition. Most of my cultural knowledge outside formal 
interviews is a result of living as a guest of the Magoto family in Bugerera/Mbiso Nata. My 
husband, two children and I lived there continuously from February 1995 through April 1996. 
My Nata language teacher was a sister-in-law of the family, my cultural vocabulary informant a 
brother and my main resources and colleagues for everything from interviews to logistics were two 
other brothers. Because of this locus of learning, my account is specifically situated with Nata at 
the center. 

Oral traditions were collected in an interview setting because of the need to visit many 
ethnic groups in a large area. Elders do not normally perform these stories in a formal setting but 
tell them in small segments in the natural course of conversation, often among elders at their own 

63 See References Cited for a list of local manuscripts. On encyclopedic informants see 
Patrick Pender-Cudlip, "Enclyclopedic Informants and Early Interlacustrine History." The 
International Journal of African Historical Studies 4. 1 (1971): 198-210. 

The results and methodology of this survey will be discussed in more detail in Chapter 7. 
In brief, it consisted of asking a few key informants to tell me who lived in each household in the 
village, what their relationships to each other and to others in the village were, along with other 
personal information. 

gatherings. Being a young woman I would not normally have had access to these situations. I 
conducted formal taped interviews with 1 64 informants and had at least as many informal 
encounters over fifteen months. Interviews usually took place in the home of the elder, always with 
a colleague to accompany and assist me. When I interviewed men, I went with a man from that 
community, when I interviewed women, I usually went with a woman from that community. Most 
interviews attracted an audience of family and neighbors who were important participants as 
audience. A young Nata woman did the transcription of most of the taped interviews in local 
languages and then went over their meaning with me. Tapes, translations and transcripts of 
interviews will be deposited with the African Studies Association oral data collection housed at the 
Archives of Traditional Music, Indiana University, Bloomington. Translations and transcripts will 
be forwarded to the University of Dar es Salaam, History Department and the Musoma Regional 

I was able to conduct interviews in each of these seven ethnic groups because of 
preexisting relationships in the region. I worked there for six years as a church development 
worker (1985-1991) and so had valuable social networks already in place. 65 A local assistant in 
each ethnic area arranged the interviews, accompanied me to them and helped with translation and 
explanation during the interview. These were not hired assistants but friends, colleagues and 
friends-of-friends. The class differences between informants and assistants was minimal and both 
were residents in the same community. In Nata I went with members of the family with whom we 
lived, in Ikizu a man from our village, in Ikoma an older pastor, and in Ngoreme a young 
development projects' coordinator from the program in which I had previous worked. 

65 1985-1991, Co-Coordinator for the Congregational Development Department of the 
Tanzania Mennonite Church, Lake Diocese and Co-Country Representative for Mennonite Central 
Committee, North America, stationed in Nyabange, near to Musoma, Tanzania. 


These people were not chosen because of their particular knowledge of history but because 
they knew their communities, people there respected and trusted them, and they were available. 
Some of the hardest work they did took place in the introductions, often involving a long period of 
questions and answers about my work. Each assistant was committed to the process because of an 
interest in history and because the bonds of reciprocity connected them to me. People did not see 
them as my dependents but as hosts of a guest to the community who could bring potential 
benefits. Because my assistants were also learners asking their own questions during the 
interviews. 1 learned by watching their interactions. They were a valuable source of cultural 
information and also colleagues for discussing how the interview had gone. I sometimes took an 
assistant from one ethnic group to interviews in a different ethnic area. Their questions often 
probed the places where the two groups differed, or where ethnic stereotypes were most intense. 
These assistants played a critical role in the formation of my own historical consciousness. 
Written Sources 

However, oral sources are not the only problematic sources of historical evidence, and as 
this section demonstrates, each of the other sources (archival, ethnography, historical linguistics, 
archaeology) have their own histories and difficulties as evidence of the past. The Mara Region 
itself also presents some problems for this kind of study. Almost no academic historical research 
has been done in this region, requiring the preliminary establishment of the most fundamental 
historical framework. 66 Early written accounts are scarce to nonexistent: a very few travelers' 

66 Gerald Hartwig, The Art of Survival in East Africa: The Kerebe and Long-Distance 
Irade (New York, 1 970); and A. O. Anacleti, "Pastoralism and Development: Economic Changes 
in Pastoral Industry in Serengeti 1750-1961" (Master's thesis, University of Dar es Salaam. June 
1 975), are the two exceptions. 

accounts; 67 a little colonial ethnography; 68 and colonial archival sources. Both the German and 
British papers in the National Archives for this region have significant sections entirely missing. 69 
The first missionaries in the region were the White Fathers, who only visited the western Serengeti 
from their lakeshore stations on rare occasion in the first decade of the twentieth century. The 
Seventh Day Adventist missionaries also came during this time but seemed more interested in 
obliterating local custom than in recording it. Mennonite Mission arrived in the 1930s, by which 
time other Europeans introduced them to local custom and history. 70 

67 Oscar Baumann, Durch Massailand zur Nilquelle: Reisen und Forschungen der Massai- 
Expedition. des deutschen Antisklaverei-Komite in den Jahren 1891-1893 (Berlin: Dietrich Reimer, 
1894); Paul Kollmann, The Victoria Nvanza: The Land, the Races and their Customs, with 
Specimens of some Dialects , trans. H. A. Nesbitt (London: Swan Sonneschein and Co. Ltd., 1899); 
and Max Weiss, Die Volkerstamme im Norden Deutsch-Ostafrikas (Berlin: Carl Marschner, 1910; 
reprint ed., New York: Johnson Reprint Corporation, 1971) are some of the few available. 

68 Edward Conway Baker, Musoma District Commissioner, his collected papers are 
available on microfilm from Oxford University Press, as well as in the Cory Papers at the East 
Africana Library, University of Dar es Salaam. Ethnographic accounts by other authors are also 
available among the Hans Cory Papers at the University of Dar es Salaam. For the Zanaki see, 
Otto Bischofberger, The Generation Classes of the Zanaki (Tanzania) (Fribourg. Switzerland: The 
University Press, 1972). For the Kwaya see Hugo Huber, Marriage and Family in Rural Bukwava 
(Tanzania) (Fribourg, Switzerland: The University Press, 1973). For the Kuria see Eva Tobisson, 
Family Dynamics among the Kuria: Agro-Pastoralists in Northern Tanzania (Goteborg, Sweden: 
Acta Universitatis Gothoburgensis, 1 986) and various articles by anthropologist Malcolm Ruel. 

69 The National Archives in Dar es Salaam has excellent indexes for locating materials. 
Much of what is listed in the indexes on the Musoma District is unavailable. In comparing the 
amount of material listed in the index for the Tarime sub-district of Musoma to the Musoma office 
itself it appears that much of the Musoma district materials never made it to the National Archives. 
1 do not yet have an explanation for this. I was not able to find any colonial papers in the Musoma 
Archives. Much of my archival materials are consequently from the Lake Province (Mwanza) and 
Secretariat files rather than from the district itself. This is unfortunate because what is largely 
missing is the original letters from the Musoma Chiefs in Swahili. The German files were also 
few. The story of the Germans burning their papers as they fled Musoma to Busegwe where th 
S.D.A. mission was used as their temporary headquarters is told in, Toppenberg, Africa Has My 
Heart , p. 67. 

70 White Fathers' sources were found in the archives of the White Fathers' Regionals' 
House in Nyegezi, Tanzania, a few other Catholic sources were located at the Maryknoll Father's 
Language School library in Makoko, Musoma. The White Fathers' stations closest to the research 


If a full range of early written sources were available more comparison of versions written 
at an earlier time could be done. A more complete colonial record would also facilitate the 
understanding of how historical consciousness has changed in the last century. This work relies 
primarily on oral sources, with scarce written sources supplying confirmation at critical points. 71 

I worked in the Tanzania National Archives and the East Africana Collection at the 
University of Dar es Salaam over a period of three months. The archives of the White Fathers, 
Seventh Day Adventist, Mennonite, Bujora Sukuma and Mara Region were consulted while I lived 
in Nata. Given more time and resources I would consult the European archives as well. The 
Lancaster Mennonite Historical Society Archives and the Seventh Day Adventist Archives were 
consulted in the United States. Northwestern University holds seminar papers from Universities in 
Dar es Salaam, Nairobi and Makerere in Kampala. The Musoma District Books, E.C. Baker 
Papers and Church Missionary Society papers are available on microfilm. 

area was Ukerewe Island and Nyegina station (established 1911). S.D.A. primary sources were 
unavailable in Tanzania. I located some in the General Conference Headquarters in Silver Spring, 
Maryland. For an account of the S.D.A. mission in Tanzania see, K. B. Elineema, Historia va 
Kanisa la Waadventista Wasabato Tanzania 1903-1993 (Dar es Salaam: Dar es Salaam 
University Press, 1 993). Mennonite Mission papers were also unavailable in Tanzania for the 
early years and located at the Lancaster Mennonite Historical Society in Lancaster. PA. For an 
account of the Mennonite Mission in Tanzania see, Mahlon M. Hess, Pilgrimage of Faith: 
Tanzania Mennonite Churc h. 1934-83 (Salunga, PA: Eastern Mennonite Board of Missions and 
Charities, 1985); see also Joseph C. Shenk, Kisare: A Mennonite of Kiseni (Salunga, PA: Eastern 
Mennonite Board of Missions and Charities, 1984); Joseph C. Shenk, Silver Thread: The Uns and 
Downs of a Mennonite Fa mily in Mission (1895-1995) (Intercourse PA: Good Books, 1996): and 
David W. Shenk. Mennonite_Safari (Scottsdale PA: Herald Press, 1974). 

" In addition to my own collection of oral traditions 1 also made use of the collected oral 
traditions from North Mara of Zedekia Oloo Siso, manuscripts of ethnic histories from Ikizu, 
Ngoreme, Ishenyi and Sizaki written by local authors and published oral traditions of the Kuria, 
Maasai, Sukuma, Luo and Sonjo, see References Cited. 

Historical Linguistics 

Historical linguistics represents a form of historical reasoning from evidence gained by 
comparing languages spoken today to identify changes in words and sound patterns over time. The 
method of historical linguistics rests on the assumption that language is a system where sound and 
meaning go together arbitrarily. That condition allows historians to figure out how different 
languages are related to each other by regular changes in languages over time, how local language 
innovations came about and also how contact with other languages influenced the development of a 
language. 72 

The classification of languages into families gives the historian an indication of historical 
events and processes that took place among the speakers of these languages. Genealogical trees of 
language families are constructed by comparing lists of core vocabularies from related languages 
to determine how many cognates they have in common (known as lexicostatistics). By identifying 
consistent sound shifts in cognate words for two closely related languages, historical linguists 
reconstruct words in the proto-language from the cognates in these languages. By comparing many 
related languages over a wide region it is also possible to reconstruct the nature of a proto- 
language itself and the relationship among all of the languages that descended from one proto- 
language. Historical linguists can suggest where speakers of a proto-language might have lived by 
looking at the geographical distribution of its descendant languages. 

Splits in a family tree can then be dated very approximately by counting the differences in 
cognate percentages of core vocabularies between related languages (known as glottochronology). 

For an overview of the methodology of historical linguistics see: Derek Nurse, "The 
Contributions of Linguistics to the Study of History in Africa," Journal of African History 38 
(1997): 359-391; David Lee Schoenbrun, A Green Place, a Good Place (Portsmouth: Heinemann, 
forthcoming), pp. 34-48; Jan Vansina, Paths in the Rainforest (Madison: The University of 
Wisconsin Press, 1990), pp. 9-16. 

This method rests on the observation that random replacement of items in the core vocabulary of 
any language accumulates at a regular pace over a period of centuries. For the Great Lakes Bantu 
family of languages Schoenbrun, following Ehret, assumes that sixteen out of 100 items in the list 
of core vocabularies "will be replaced each 500 years, either by morphological analogy, or by 
borrowing from another language." 73 Thus if the group averages for Suguti and Mara languages 
share 56 percent of their cognates, we know that these two branches of the East Nyanza family of 
languages must have been separated about 600 A.D. or 1400 years ago. Within the western 
Serengeti, languages like Nata and Ikoma share 86 percent of their cognates and thus diverged less 
then 500 years ago. These time scales can then be correlated with the independently derived 
chronologies of archaeology, where they are available, although both systems of dating are 
extremely vague with wide margins of error. 

Historical linguists can also look at individual words in a language that may be different 
from the cognates in related languages, to determine innovations in the past. These innovations 
may be the result of adopting words from other languages (loanwords) or developing new words 
internally (innovations). Loanwords from other language groups provide evidence for cross- 
cultural contacts in the past that can be dated by the methods of glottochronology explained above. 
In western Serengeti languages many loanwords from Southern Nilotic languages, particularly 
words about livestock, suggest that Mara Bantu-speakers moving out into the drier areas of the 
interior learned from their neighbors how to diversify and expand their economic subsistence 
patterns by increased stock raising. From a reconstruction of the Southern Nilotic family tree we 
know that Mara Southern Nilotic-speakers were in the region about the same time that East 
Nyanza Bantu-speakers arrived. 

73 Schoenbrun, A Green Place , p. 36. 


Internal innovations of new words are also possible, providing the historian with some 
evidence of historical processes for which new words had to be developed. These words can be 
investigated by looking at how the root word from which it derives was used in other languages 
throughout the region. For example, in the first chapter I demonstrate the status of a post- 
menopausal woman, omokuungu, by looking at the etymology of the word. This word is derived 
from the old proto-Bantu verb -kung, "to gather, to assemble." In other places around the lake it 
refers to a rich person with followers, or to a chief.' 4 Uniquely in the Mara languages this word 
refers to elderly women with many children. This would indicate that the "wealth in people" that 
women controlled through gathering their children was extremely important on this inter-cultural 

Although correlating the speakers of a language with an ethnic group or even a distinct 
community of people in the past would be convenient for the historian, linguistic evidence alone 
cannot support this. Historical linguists can only talk about people who spoke the same language 
(but who might also be multilingual and practice a culture distinct from other speakers of the 
language). If one language disappears and another becomes dominant in the record of historical 
linguistics this does not necessarily mean that one group conquered or expelled the other. A 
language can spread ahead of its speakers and divisions in the family tree of a language do not 
necessarily mean that people had to move." As Derek Nurse points out, "historians and 
archaeologists are frequently prone to interpret a linguistic tree as a literal historical development, 

74 David Lee Schoenbrun, The Historical Reconstruction of Great Lakes Cultural 
Vocabulary: Etymologie s and Distributions (Cologne: Rhdiger Koppe Veriag, forthcoming), #209 
and #210; Schoenbrun, A Green Place , pp. 185-187. 

75 For a similar critique applied to archaeology see Martin Hall, "Origins: Unwrapping the 
Iron Age Package," Farmers. Kings and Traders: The People of Southern Africa 200-1860 
(Chicago; The University of Chicago Press, 1987), pp. 17-31. 

as a movement of people from one point in space or time to another, or as one component of a 
homogeneous 'cultural tradition'." 76 

In spite of these pitfalls, historians can learn a lot about the ways that people lived in the 
past by reconstructing the cultural vocabularies-for example, what kind of subsistence patterns 
they followed, what their settlements and homesteads looked like and what kinds of authority they 
recognized. These reconstructions are based on the assumption that a word in a language refers to 
a real object. If historians have available a series of linguistic stages, with cultural vocabularies 
for each, then they can postulate a sequence of changes in social institutions and activities over 
time. It is often only through the evidence of historical linguistics that we can discern interactions 
with peoples whose culture and language have long disappeared. 

I have relied heavily on the historical linguistic work already done in this region by 
Christopher Ehret and David Schoenbrun. 77 For example, Schoenbrun classified the Lakes Bantu 
languages and Ehret the Southern Nilotic and Southern Cushitic languages in the region. Both 
have published lists of loanwords and etymologies making it relatively easy to compare my lists of 
cultural and core vocabularies. I collected core vocabularies of 100 words (Ehret and 
Schoenbrun's list) in Ngoreme, Ikizu, Nata, Ikoma, Ishenyi, Sonjo and Dadog. I assembled a 
nearly complete 1563 cultural vocabulary list (University of Dar es Salaam list) in Nata and parts 
of it in some other languages. From these core vocabularies I constructed a dialect chaining chart 

76 Nurse, "The Contributions of Linguistics," p. 370. 

Christopher Ehret, Southern Nilotic History: Linguistic Approaches to the Study of the 
Past (Evanston: Northwestern University Press, 1 97 1 ); David Lee Schoenbrun. "Early History in 
Eastern Africa's Great Lakes Region: Linguistic, Ecological, and Archeological Approaches, ca. 
500 B.C. to ca. A.D. 1000" (Ph.D. dissertation, UCLA, 1985); Schoenbrun, A Green Place : and 
Schoenbrun, Etymologies. I used Ehret and Schoenbrun's 200 word list for core vocabularies and 
a combination of Schoenbrun's and University of Dar es Salaam's 1563 word list for cultural 

of the cognate percentages that appears in Chapter 4. The cultural vocabularies and the 
etymologies of words are used frequently throughout the dissertation. 1 also had access to a few 
unpublished dictionaries of some languages in the region and a recently published Kuria-English 
dictionary. 78 

By learning one Mara language, Nata, I could understand at a basic level the other four 
Bantu languages. I learned no Dadog. Knowledge of the local language was important for a more 
nuanced cultural understanding than would have been possible in Swahili. It allowed the elders the 
freedom to speak in their own languages, rather than translate into Swahili. Extensive previous 
knowledge of Swahili was an asset but was also an impediment to learning Nata more fluently. 
Other Sources of Evidence: Archaeology and Ecology 

Although historical linguistics and comparative ethnography are the primary sources of 
evidence used in conjunction with oral traditions, I also rely on published works in archaeology and 
ecology. Little archaeological work exists for the period of my study and in the precise area where 
I worked, however, a lot has been done on pastoral and hunter/gatherer communities during the 
Neolithic period in the Rift Valley and the Serengeti. 79 In addition, a vast body of scientific 
research exists on the Serengeti-Mara ecosystem that provides valuable information on climatic 
patterns, soil and vegetation types and distribution, wild animal ecologies and the effects of human 

78 Muniko, et. al., Kuria-English Dictionary . 

79 See for example Peter Robertshaw, ed., Early Pastoralists of South-western Kenya 
(Nairobi: British Institute in Eastern Africa, 1990); John Bower, "The Pastoral Neolithic of East 
Africa," Journal of World Prehistory 5, I (1991): 49-82; Peter Robertshaw and David Collett, "A 
New Framework for the Study of Early Pastoral Communities in East Africa," Journal of African 
History. 24 (1983): 289-301; Desmond J. Clark and Steven A. Brandt, eds.. From Hunters to 
Farmers: The Causes and Consequences of Food Production in Africa (Berkeley: University of 
California Press, 1984). Closer to Lake Victoria see R. C. Soper and B. Golden, "An 
Archeological Survey of Mwanza Region, Tanzania," Azania 4 (1969): 48-53. North of the Mara 
Region see, J. E. G. Sutton, The Archaeology of the Western Highlands of Kenya (Nairobi: 
British Institute in Eastern Africa, 1973). 

induced perturbations such as burning or intensive livestock grazing on the ecosystem. 80 This 
research sets the material parameters that patterned subsistence economies and social interaction of 
early settlers in the region. 

Throughout the dissertation I also represent the information provided in oral traditions on 
maps of a geographical grid. These maps are tools for understanding the spatial relationship 
between places on the landscape. 81 The bird's eye view is an artificial device used by people who 
have not grown up walking to these places and provides insights into historical process not 
otherwise available. Because local people do not see the landscape from this perspective, these are 
not adequate representations of the "mental maps" of oral tradition but remain useful to the 
Historical Reconstruction and Chapter Organization 

This dissertation is organized into four sections according to indigenous time frames. The 
first section concerns the present era in which the research was conducted, the colonial period in 
which these narratives were reformulated into ethnic accounts and the disaster years in which 
social identity was significantly transformed. Because the analysis of this dissertation is heavily 
dependent on the narrative form of oral tradition and its transmission, the next chapter looks at 
these issues in terms of the social organization of knowledge about the past. The third chapter 
familiarizes the reader with the events of the late nineteenth century disasters because of their 
significance in the reshaping of historical consciousness that must be "read" into the oral traditions 
analyzed in the rest of the dissertation. 

80 The classic study in this regard is A.R.E. Sinclair and M. Norton-Griffiths, eds. 
Serengeti: Dynamics of an Ecosystem (Chicago: University of Chicago Press, 1979). 

81 Thanks to Peter Shetler for his assistance and expertise in GIS and other forms of 
computer mapping in creating these maps. 


The second section is concerned with the oldest period of indigenous history or time frame, 
looking at the traditions of ethnic and clan origins, analyzing them in terms of the social 
organization of space over the tongue duree. Although they have been overlaid with material from 
other time periods, I concentrate on the core images of these traditions that correspond to the 
spaces of production, reproduction, regional networks of association and settlement. These 
traditions do not yield a sense of change over time, but they do provide a basic framework for 
understanding the social organization of space in the distant past in terms of generative principles. 

The third section, on the middle period of indigenous history, looks at the lists of settlement 
sites names and descriptions of rituals concerning the land that refer to a period of time in which 
people established rights to the land that exist today. Long-term settlement patterns, methods for 
incorporating strangers and rituals for marking boundaries are described. These means by which 
people established a relationship to the land were used in the second half of the nineteenth century 
to create new territorial groupings in response to disaster. 

The fourth and final section represents the climax of the story in which the creative 
changes in the spatial organization of social identity take place in response to the disasters of war, 
famine, disease and later colonial intrusion. It describes how age-set territories emerged in the east 
to reconfigure settlement structure and how a sense of ethnicity developed in the early colonial 
years. It is only by setting the events of the late nineteenth century in the context of long-term 
social process that the full meaning of the nineteenth century changes are understood. 

In each chapter successive layers in the interpretation of oral traditions are analyzed by 
looking at different versions as they shed light on various aspects of the form of social identity 
under consideration. In this way the oral traditions themselves set the terms of the analysis and 
other forms of evidence used to elaborate further these themes. 








To approach the history of the western Serengeti from the perspective of its oral traditions, 
the historian must first consider the performance context in which people tell these narratives. 
When and where are stories of the past told, to whom, by whom, for what purposes? Why are they 
remembered? How is this knowledge used in contests of power? How have shifting 
understandings of the past modified historical narratives over the generations? 

Individual members of particular social groups formulate, transmit, and maintain historical 
knowledge within the physical spaces that they occupy in daily practice, physical spaces that 
structure their interaction with others. In particular, a gendered division of knowledge about the 
past is explored in the lived spaces where men and women transmit separate realms of knowledge. 
The present generation of elders developed their ethnic narratives within the expanded space of the 
colonial territory and its reliance on the written word. Although most people consider historical 
traditions as men's knowledge, women significantly influence the transmission of oral traditions 
through their informal knowledge of current community affairs. In addition, women occupying 
exceptional social positions cross these boundaries of gendered knowledge and gain access to men's 

The classification of some kinds of knowledge as "secret" and available only to insiders 
also confines knowledge about the past to particular social spaces. Inside or specialized 
knowledge represents a key social resource that initiates preserve in things and places as well as in 


people. Cohen's caution against a formalist definition of oral tradition as a set of fixed texts allows 
us to see how historical knowledge is exchanged in "the everyday critical, lively, intelligence which 
surrounds the status, activities, gestures, and speech of individuals." 1 Yet the social spaces that 
people inhabit restrict the ways that they can share and alter that knowledge. Although historical 
knowledge crosses social boundaries, it does so in structured ways. Each kind of oral tradition 
contains the knowledge of a particular social group as they represent themselves in relation to 
others, within a field of multiple and overlapping sets of knowledge, rather than as a discrete subset 
of the collective knowledge of an integrated society. Social groups not only represent the spatial 
organization of social relations in their oral traditions but also maintain and transmit these 
narratives within the spaces that structure their relationships. Understanding the social 
organization of knowledge prepares the way to interpret oral traditions. 
The Social Spaces of Historical Knowledge 

Historical narratives are not disembodied social facts but are inseparable from the 
historical context and geographical space in which people tell them. Discerning the social units 
defined by oral tradition and the social definition of the people who tell oral tradition provides key 
insights into the historical meaning of these narratives. 
Locating the Western Serengeti as a Region 

The largest social space of western Serengeti oral traditions defines a region within which 
the narrators of these traditions locate themselves. The principal narrators of oral tradition 
analyzed in this dissertation include the Bantu-speaking hill farmers of the western Serengeti-the 
Ikoma, Nata, Ishenyi, Ikizu and Ngoreme ethnic groups. They now occupy the Serengeti and 
Bunda Districts in the southeastern portion of the Mara Region of Tanzania. However, hill 

1 D. W. Cohen, "The Undefining of Oral Tradition," Ethnohistorv . 36, 1 (Winter 1 989)- 

farmers have not been the only inhabitants of this region over the long-term. The adaptation of 
farmers to the ecology of this region and their ongoing prosperity depended upon interaction with 
pastoralists and hunter/gatherers. The story of Bantu-speaking hill farmers is integrally connected 
to that of Dadog-speaking pastoralists in the western Serengeti, the Rotigenga and Isimajek 
Tatoga. 2 I cannot do justice to the Tatoga story here, but will use it to shed light on the 
development of the region as a whole. I hope that their story can one day be published in its own 
right. The other set of actors in this story, the Asi hunter-gatherers, do not have a voice here 
because 1 could not identify any Asi descendants in the western Serengeti who knew these 
traditions. Regrettably, they are only represented here in the stories of others. I hope that the 
research will someday be done to gather what fragments remain of this tradition. [See Figure 2-1 : 
Regional Setting of the western Serengeti.] 

No local designation exists for this group of Bantu-speaking farmers as a whole except 
Rogoro, or "the people of the east," yet even the area to which this refers varies relative to the 
location of the speaker. These people feel a diffuse sense of collective identity due to their common 
historical background, shared cultural assumptions, and proximity to each other. Another way to 
define this group would be to use the colonial term, "South Mara." 1 rejected this term because it 
also includes peoples along the shores of Lake Victoria, whose traditions are significantly different 
from those in the interior. No absolute set of boundaries defines the regional unity of western 
Serengeti. My research was concerned with five ethnic groups, given the limitations of field 
research, but logically could have expanded to include Zanaki, Sizaki and, at a larger scale still, 
Kuria or the Mara Region as a whole. Limited interviews among these neighboring groups allow 
for a regional comparison. 

1 1 use Dadog to refer to the language and Tatoga to refer to the people, including both 
Rotigena, Isimajek and the larger Tatoga community in other places in Tanzania. 

QUJQ. ™ 


I use the term "region" to refer to the western Serengeti or to the Mara as a whole in the 
sense that they represent the geographical boundaries of intercommunicating, interacting sets of 
people. "Region" defines neither a homogeneous cultural or social unit nor economic relations of 
exchange or formalized marketing systems, as has been the trend in much of the recent regional 
analysis. Regions are rather historical products constantly negotiated and transformed. 3 Even the 
most rigidly conceived regional boundaries with the Maasai to the east or the Sukuma to the south 
were frequently crossed through trade, marriage, prophecy or refuge. Although the case could be 
made for defining a western Serengeti region based on linguistic and cultural unity, it would then 
only include the Bantu-speaking farmers. This region, both past and present, has functioned based 
on its linguistic, cultural and economic diversity. 

I have chosen to call this group "the peoples of the western Serengeti," rather than of 
eastern South Mara, because of their orientation east toward relationships in the Serengeti during 
the late nineteenth century, when the social transformations this dissertation narrates took place. 
"Serenget" is a Maasai word, referring to a historical Maasai section and meaning "wide-open 

3 On some recent theoretical work still using the concept of hierarchical relations of 
exchange to define regions see for example, Claudio Lomnitz-Adler, "Concepts for the Study of 
Regional Culture." American Ethnolog ist 18,2(May 1991): 195-214; Eric Van Young, 
"Introduction: Are Regions Good to Think?," Mexico's Regions: Comparative History and 
Development (San Diego: Center for U.S.- Mexican Studies, UCSD. 1992), pp. 1-36. For a 
review of the literature see Carol A. Smith, "Regional Economic Systems: Linking Geographical 
Models and Socioeconomic Problems," Regional Analysis. Vol. I : Economic Systems (New York: 
Academic Press, 1976), pp. 3-63; Mary Beth Pudup, "Arguments within Regional Geography," 
Progress in Human Geography 1 2, 3 (September 1 989): 369-39 1 . As applied to a historical study 
in Africa, see Charles H. Ambler. Kenya Communities in the Age of Imperialism: The Central 
Region i n the Late-Nineteenth Century (New Haven: Yale University Press, 1988); and Allen M. 
Howard, "The Relevance of Spatial Analysis for African Economic History: The Sierra Leone - 
Guinea System," Journal of African History 1 7, 3 ( 1 976): 365-388. My use of "region" is closer 
to that of Richard Waller for interior East Africa in "Ecology, Migration and Expansion in East 
Africa." African Affairs 84 (July 1985): 356-7. 


spaces." 1 use this word simply because of its widespread recognition. The peoples of this region 
would not recognize this name as their own but would recognize their common story. 4 

The western Serengeti is bounded on the east by the Great Serengeti Plains and 
Maasailand, on the north by the Mara River and Kurialand, on the west there is a gradual shift, 
without any natural division, toward the peoples of Lake Victoria, and on the south by the 
Mbalageti River and Sukumaland. These boundaries define an ecologically unified area, called the 
Serengeti-Mara ecosystem by ecologists, in which the interdigitation of hills, woodlands and 
grasslands allowed farmers, hunters and herders to develop interdependent specializations. Yet the 
ecological unit is not fixed either, as a gradual change in ecology occurs as one nears the lake with 
no natural boundaries. 

From the colonial period on the peoples of the western Serengeti began to see themselves 
as part of the Mara Region, or Musoma District as it was then known. These were the boundaries 
within which the colonial imagination of place shaped the image of the Musoma "tribes" and 
reoriented historical vision toward the lake and to the west. 
Narrative Forms and Narrators 

Historical narratives in the western Serengeti are, like those of many other non-centralized 
societies in Africa, weak and loosely structured. They appear more in the form of conversation 
than as epic poetry in set verse.' No particular word exists in local language for this genre of oral 
tradition except as amang'ana ga kare (matters of the past). No formal experts control this 

4 1 had many interesting discussions with local people about what to call this group. Many 
voted for Rogoro, the people of the east, but just as many declared it was not an encompassing 
term. Some wanted the name of a mountain to designate their unity but could not agree on whether 
that should be Bangwesi or Chamuriho. 

5 Hofmeyr, "We Spend Our Years," p. 4; and David W. Cohen, Womunafu's Bunafu- A 
Study of Authority in a Nineteenth Centu r y African Community (Princeton: Princeton University 
Press, 1977), pp. 8-9. 


knowledge although some are considered more knowledgeable than others. Those who know more 
about "matters of the past" acquired their knowledge through personal desire or aptitude, rather 
than purely as a function of their position. Some people have a "gift" for it, given by the ancestors. 
Elders attain legitimacy as narrators of "matters of the past" through a combination of ability, 
respect, experience and the sanction of the ancestors, manifested in the effectiveness of the tales. 

In the western Serengeti individuals tended to specialize in particular kinds of knowledge 
depending on their own interests and experience. Each was a historian in his or her own right 
because they not only mastered this knowledge but organized and found new meaning in it. 
Similarly, Guyer and Belinga emphasize that the preservation and transmission of knowledge in 
Equatorial Africa were personalistic: "different people took on different parts of the overall 
expanding corpus, largely by choice rather than the ascription of parentage or assignment within 
secret societies."'' 

A profile of the person most often recommended to me as one who knew about these 
"matters of the past" consists of a man more than sixty years of age who occupied some position of 
authority or respect in the "traditional" structures of society. 7 If that man also had education and 
political office, the community valued him as an able intermediary with an outsider. On the other 
hand, many felt that educated people disparaged "traditional" knowledge. Educated men often 
presented the past simplistically as "warring tribes" ruled by "clan headmen." Material wealth was 
not a particular criterion for recommendation. All were respected elders who people in the 
community consulted for their wisdom. Some highly respected community leaders did not have 
much to say about the past, although people assumed that they must. Almost all had some 

6 Guyer and Belinga, "Wealth in People," pp. 110-111. 

7 By contrast among the Tatoga I was often taken to interviews with men in a younger 

experience in the colonial labor force but the most knowledgeable elders had spent much of their 
lives at home. Those with the most extensive social networks, and thus prestige, in the community 
tended to be the ones most knowledgeable about the past. 
Gendered Knowledge 

The formal restriction of historical knowledge to elderly men stands out most in the social 
organization of knowledge. Women, elderly or not, possessed entirely distinct forms of knowledge 
about the past. When I would ask to speak with those who knew about history, local colleagues 
led me almost exclusively to interviews with older men. Both men and women alike agreed that 
men of this generation were the keepers of historical knowledge. When I insisted on talking to 
women, I found that most women did not know the larger ethnic accounts of origin, migrations, 
clans, ritual and battle, which made up the spontaneous content of interviews with men. At first I 
thought that women were just reluctant to give me their versions of the past, but I later became 
convinced that women possessed not just another version but wholly different kinds of knowledge 
about the past. In South Africa, Hofmeyr describes a similar gendered division of oral literature in 
which people said that men told "true" histories while women told "fictional narratives." She goes 
on to claim that the content and style of these stories are similar, the only difference being the 
spaces in which they are told. 8 

1 argue, however, that because people learn about the past in particular gendered spaces, 
men and women share neither styles of oral narration nor types of knowledge about the past. They 
transmit and maintain knowledge by the ways in which gendered space is represented and 
organized in daily practice. Men and women occupy separate spheres of interaction in their daily 
routines, sharing the same world but participating in different, though intersecting, sets of 

discourses about that world. 9 They keep and transmit historical knowledge by the paths that they 
walk each day and the positions that they occupy in the imagined male and female spaces that 
permeate their world. Women may learn some of men's knowledge about the past but they do not 
transmit those stories in the narrative style of men nor in the formal setting of men's courtyard 

The Colonial Contexts of Men's Historical Narratives 
Men's public knowledge of the past is now almost entirely represented as a body of unique 
ethnic traditions, for example, of the Nata or the Ishenyi. As 1 demonstrate in later chapters, 
western Serengeti people created ethnicity in its present form as a result of the late nineteenth 
century disasters, solidified by the colonial experience. The stories they tell about ethnogenesis, 
however, do not date from this period. Men reworked narratives from many different kinds of 
social units, representing different kinds of social boundaries, into a unified corpus of ethnic 
history. While I analyze the specifics of these changes in oral traditions in later chapters, here I am 
concerned with how the space of "tribes" within the larger colonial territory shaped knowledge 
about the past. 10 The creation of ethnic history was not only an imposition of colonial hegemony 
but also a creative adaptation of old knowledge to new circumstances. It was a product both of 
opportunistic men reaping the benefits of the new system and of common people trying to make 

9 See Janice Boddy, Wombs and Alien Spirits: Women. Men and the Zar Cult in Northern 
Sudan (Madison: University of Wisconsin Press, 1989), pp. 5-7, on the analysis of spirit 
possession cults controlled by women as a counter hegemonic discourse to the dominant male, 
Islamic discourse. 

10 1 use the term "tribe" only as it refers to the colonial creation of "tribal" units for 
administrative purposes and not as a traditional category of social identity. I use quotation marks 
to remind the reader of this interpretation. As will become clear in later chapters the way that 
"tribe" came to be defined locally differed significantly from other places in Tanganyika. 

sense of a new world in which old relationships took on new meaning and new relationships were 
assimilated into old patterns. 
Men's Colonial Experience 

The colonial experience deeply affected men's oral narratives of the past. With the advent 
of indirect rule the social unit of political significance became the "tribe." Nevertheless, because 
colonial officers failed to define the boundaries of "tribe" and no traditional system of chiefs 
existed in the Musoma District, local people enjoyed ample room in the negotiations to set 
boundaries. Certain individuals, families or groups gained significant benefits by defining "tribal" 
boundaries to promote their own interests. Wambura Igina of the Simbiti chiefdom became the 
second German chief by translating the people's questions to the German officer as their call for 
Igina's appointment as chief." Fundi Kenyeka of Busegwe became a German chief because he was 
a blacksmith and provided them with nails. 12 Both made themselves useful to the Germans and 
thereby attained the status of representing their "people," among whom they had but little 
"traditional" authority. 

Within the colonial system of indirect rule, many men had the opportunity to negotiate a 
beneficial reconfiguration of "tribe" and authority in their narrations of the past." When the 
colonial officers of Musoma realized that the system of chiefs was not working, because the chiefs 
lacked traditional authority, they began investigating "pre-European tribal organization" to restore 

11 Zedekia Oloo Siso, Buturi, 'The Oral Traditions of North Mara," unpublished 
manuscript in author's collection. 1995. 

12 E. C. Baker, "Tribal History and Legends," 9 December 1929, microfilm, Musoma 
District Books (MDB). 

13 For the wider application of this observation in the Lakes region see David Schoenbrun. 
"A Past Whose Time Has Come: Historical Context and History in Eastern Africa's Great Lakes 
Region," History and Theory 32, 4 (1993): 32-56. 

the "ancient rights and powers" of the clan elders as the basis for indirect rule.' 4 In 1945, Hans 
Cory led the effort to visit each chiefdom, or to call elders to Musoma, to question them about 
"tribal" practice and history. A major reorganization of "tribes" and a new Constitution for North 
and South Mara resulted, in 1 948. " This colonial way of interpreting the past through the lens of 
"tradition" and "tribe" became widespread and strongly influenced local historical narrative. 

Men formulated ethnic histories, not surprisingly, in this era when they had increasing 
opportunity to travel and meet people from other places. Some among the present generation of 
elders were the first to go to school in Musoma; some had careers as government clerks or mine 
supervisors throughout the Territory; others went as far as Burma with the K.A.R. (King's African 
Rifles) in World War Two or to Nairobi, Tanga and Magadi Soda as migrant laborers. Away 
from home, men began to see themselves as part of larger communities, seeking out people who 
came from areas neighboring their homes and who spoke similar languages. Laborers in Nairobi 
popularized Kuria identity by forming the Kuria Union in 1945, expanded one year later to "The 
South and North Mara Tanganyika Union," based on a common tradition and origin. Their goals 
were to promote modernization, help the sick, arrange for funerals, and return fugitive women from 
Nairobi." Migrant laborers walking to Magadi Soda in Kenya found hospitality among the Sonjo, 

14 District Commissioner, Musoma, "Memorandum on the Revival and Application of the 
Clan Regime in the Musoma District," 4 July 1945, CORY #347, EAF, UDSM. 

15 Hans Cory, "Report on the pre-European Tribal Organization in Musoma (South Mara 
Distict and Proposals for adaptation of the clan system to modern circumstances," 1945, CORY 
# 1 73, EAF, UDSM; Hans Cory, "South Mara Constituent Assembly, a new constitution for the 
South Mara Council... ," 1959-60, CORY #385, EAF, UDSM. Hans Cory, "Report on the 
general situation in Kuria Chiefdoms of North Mara and proposals for its improvement...," 1945- 
49, CORY #171. EAF, UDSM. For a discussion of this process as a territorial model see, C. 
Winnington-Ingram, "Reforming Local Government in a Tanganyika District," Journal of Africa 
Administration 2, 2 (April 1950): 10-12. 

16 Tarime District Office, Native Administration, Kuria Union Meetings 1946-52 83/3/2 

in the midst of Maasailand, based on a shared ntemi scar on the right breast. Subtle shifts in the 
historical imagination took place as narratives began to account for a larger nation of "tribes."" 
The Colonial Concept of "Tribe" 

The peoples of Musoma District reworked existing identities to comply with the need of the 
colonial government for "tribes," but they did it on their own terms and created small units that 
were responsive to local control. However, as these units became the basis for political action they 
took on a life of their own and the "tribal" assumed the status of the "traditional." To understand 
the idiom in which elders cast oral traditions today, we must look at how they adopted and 
transformed European notions of "tribe" in this region. 

The "tribal" model of African society is a nineteenth century European idea developed as 
emerging nations sought their ancient origins in self-conscious "tribes." 18 This model sees "tribes" 
as discrete and bounded entities whose movement and influence the observer can trace, like billiard 
balls rolling across space and time as a unit. The influence of this model on colonial officers is 
apparent as they characterized each "tribe" in the western Serengeti with distinct origins and 
customs, despite their obvious linguistic and cultural similarities. One of the first Musoma District 
Officers stated, based on oral tradition, that the Ikizu and Sizaki "are Sukuma" who "arrived 

17 For similar processes throughout Africa see: Leroy Vail, ed., The Creation of Tribalism 
in Southern Africa (Berkeley and Los Angeles: University of California Press, 1991); in 
Tanganyika see John lliffe, A Modern History of Tanganyika (Cambridge: Cambridge University 
Press, 1979); in other places E. Hobsbawm and T. Ranger, eds. The Invention of Tradition 
(Cambridge: Cambridge University Press, 1983); and Terence O. Ranger, "The Invention of 
Tradition Revisisted: The Case of Colonial Africa," in Legitimacy and the State in Twentieth 
Century Africa, Terence O. Ranger and O. Vaughan, eds. (Oxford: St. Anthony's College, ! 993), 
pp. 62-1 1 1. Benedict Anderson. Imagined Communities: Reflections of the Origin and Spread of 
Nationalism < London: Verso, 1983). 

18 See an analysis of the "tribal model" in Igor Kopytoff, "Introduction: The Internal 
African Frontier," in The African Frontier: The Reproduction of Traditional African Societies 
(Bloomington: Indiana University Pres, 1987), pp. 3-4. 

fourteen generations ago," and that the Nata are "an offshoot of the Ikizu from whom they 
separated owing to floods." The same officer described other western Serengeti peoples (the 
Ikoma, Ishenyi and Ngoreme) as Sonjo, who have "become somewhat mixed owing to the 
settlement of Bakuria in the area."" The Musoma District Book records a single, static precolonial 
"tradition" for each group-discovered by seeking the "true" origins of peoples who had come into 
contact with many other groups in their "tribal" migrations. 

In the "tribal" model colonial officers sought the origins of a "tribe" using the concept of 
biological parentage or "blood," rather than adoption, incorporation or assimilation. The image of 
blood relations underwrites the larger historical scheme of evolution in which not only the plant and 
animal kingdoms, but also humans, follow a unilineal trajectory toward a higher state of being. 
Each "tribe" can thus be located according to its social and cultural development, with western 
culture at the pinnacle. The development of humans is represented as a tree in which many 
branches emerge on the way to the top, with one set of roots. Each "tribe" is similarly imagined to 
have one set of roots with many branches that diverged during migration. 20 

However, colonial observers found the heterogeneous composition of the peoples of 
Musoma District difficult to reconcile with these preconceived notions of "tribe." The 
inconsistencies that they record provide the historian, from another perspective, with a unique 
glimpse of people in the process of becoming "tribes." The fluidity of the situation and the ways in 
which local people were reworking their own histories is obvious. Migration histories of the 
Musoma peoples told of ancestors coming from every direction, rather from a single point of 

" Baker/'Tribal History and Legends," 9 December 1929, MDB. See also Native Affairs 
Census 1926-1929, Chiefdom Census 1926, 246/P.C./3/21, TNA. 

;0 For a comprehensive review of anthropology and the "tribafparadigm in Africa see 
Sally Falk Moore, Anthropolog y and Africa: Changing Perspectives on a Changing Scene 
(Charlottesville and London: The University Press of Virginia, 1994). 

origin. They spoke Bantu languages and yet appeared in dress, circumcision rites and age-set 
organization to be Nilotic-speakers. Kinship systems did not consistently follow matrilineal or 
patrilineal patterns, and did not form a neat segmentary system of clans. A "tribe" was nearly as 
likely to raid one of its own clans as those of another "tribe." 21 Traditional chiefs did not exist and 
the "tribal" units themselves were difficult to identify. The smaller units, sometimes called "sub- 
tribes," warranted their own chiefs, but were too small and numerous for effective administration; 
while the larger "tribal" units had no inherent cohesion or unifying institutions of authority. 22 
Since "tribes" were assumed to be biologically separate, though related, societies and 
cultures, colonial writers most often explained heterogeneity as a result of "mixed stock." For 
example, the District Officer characterized the Zanaki as consisting of "an admixture of Kuria 
blood from the north and a heavy strain of Sukuma from the south." 23 The German colonial 
encyclopedist, Schnee, described the Ruri as a people, "strongly coursing with Massai and Wageia 
[Luo] blood, shot through with Bantu, who have also adopted Massai armaments and military 
tactics." 24 

21 An example of a contemporary scholarly formulation of the "tribal model" is colonial 
anthropologist E. E. Evans-Pritchard, The Nuer: A Description of the Modes of Livelihood and 
Political Institutions of a Nilotic People (Oxford- ClarpnHnn Prec 1940), pp. 120-122. Here he 
identifies the important characteristics of a tribe: 1) a common and distinct name; 2) a common 
sentiment; 3) a common and distinct territory; 4) a moral obligation to unity in war; 5) a moral 
obligation to settle feuds and other disputes by arbitration; 6) a segmented structure with 
opposition between its segments; 7) an important structural relation between the lineage structures 
of the dominant clan and the territorial system; 8) a unity within a system of tribes; and 9) the 
tribal organization of age-sets. 

22 Edward Conway Baker, "North Mara paper," 1 935, Tanganyika Papers, microfilm 
project of Oxford University Press. 

23 E.C. Baker, "Tribal History and Legends." 9 December 1929, MDB. 

24 Heinrich Schnee, ed„ Deutsches Kol onial-Lexikon. 3 Vols. (Leipzig: Quelle and Meyer, 
1920), p. 679. 


A closer look at the social realities behind these observations demonstrates that other 
boundaries were at work besides the ethnic in these ascriptions of various groups to a particular 
parentage. For example, elders did not describe the relationship of Nata to Ikizu as one of a 
"splinter group" from parental stock but rather of a shared system of eldership titles. 1 interpret the 
relationship of the Ikizu and Sizaki to Sukuma in their origin stories not as one defined by "blood," 
but by Ikizu and Sizaki acceptance of the limited authority of a Kwaya rainmaker clan from 
Kanadi, Sukuma, to gain access to power over fertility. The language, society and culture 
remained Mara and not Sukuma. Similarly the Ikoma, Ngoreme and Ishenyi, whom oral traditions 
say came from Sonjo, share very few cultural, linguistic or social elements with Sonjo. What they 
do have in common is a shared experience of extreme Maasai pressure in the second half of the 
nineteenth century that created refugees on both sides of the Serengeti plain. Yet, spokesmen from 
each group were willing to present these diverse connections as "blood" relations in the "tribal" 

The categories and terms of classification for the Musoma "tribes" subtly shifted 
throughout the colonial period as those who implemented indirect rule constantly redefined "tribal" 
identity in their search for "tradition." This confusion is evident in the population statistics of the 
various chiefdoms. In 1909 the German officer in Shirati reported a total of twenty-six Sultans 
(Chiefs) north and twenty-eight Sultans south of the Mara River with a total estimated population 
of 1 1 0,000. 25 A German classification of "tribes" listed more than thirty, with the major 
classifications including the Nata and Ikoma as Maasai peoples and the Sizaki, Ngoreme, Ikizu and 

25 Schultz, Schirati, to Governor, Dar es Salaam, 25 December 1909, Schirati 1 909-1910 
G/45/2, TNA. 

Ishenyi as "Shashi" peoples. 26 The first British census in 1928 listed a population total of 199,520 
with nine major "tribes" (Kuria, Girango. Rangi, Jita, Sizaki, Zanaki, Ngoreme, Simbiti and 
Ikoma). 27 A 1 937 report on governance identified "upwards of forty petty chiefs" and thus 
corresponding "tribes." 28 The 1948 census recorded nineteen chiefdoms south of the Mara River- 
the largest, Majita, at a population of 28,696 and the smallest, Buhemba, at a population of 1,505 
or Nata at 1,519. 29 However, the "Tribal Map and Classification of Tribes of Tanganyika 
Territory," also of 1948, listed twelve "tribes" of the Musoma District (Jita with one sub-tribe, 
Kwaya with one sub-tribe, Ngoreme, Kuria with twelve sub-tribes, Zanaki with eight sub-tribes, 
Ikizu, Ikoma with three sub-tribes, Sizaki, Kerewe, Suba with five sub-tribes, Luo with eight sub- 
tribes and Tatoga now being recognized as Musoma natives rather than aliens.) 30 [See Figure 2-2: 
1948 Tribal Map of the Musoma District.] 

This constant renegotiation of "tribal" identity would not have been possible were social 
identity primarily defined by firm and discrete ethnic units. "Tribal" boundaries were not just the 
imposition of the whims of colonial officers, but the result of local people willing to name 
themselves differently depending on the opportunities. Local chiefs and other notables found 

26 Musoma District, "Notes from the Musoma District Books on Local Tribe and 
Chiefdoms in German," [c. 1912?], CORY #348, EAF, UDSM. 

27 Native Affairs Census 1 926- 1 929, Chiefdom Census 1 926, 246/P.C./3/2 1 , TNA. On 
recount the final census figure was 180,136 with the largest being Kuria with a population of 
50,632 and the smallest being Ikoma with a population of 6,454. 

28 E. C. Baker. "System of Government, Extracts from a Report by R. S. W Malcolm " 
1937, MDB. 

29 East African Statistical Department, Nairobi, East African Population Census, 1948, 
African Population of the Musoma District, Secretariat Files, 40641, TNA. The population of the 
sub-district south of the Mara River was 141,547. 

30 Tanganyika Territory Classification of Tribes and Tribal Map, Population Census 
Secretariat Files, 36816, TNA. 




I S 5 to I - 

_ i» g 1 1 S = » j 

1 = 11 


3 =3 3 


V TO 2 .£ F 


> CO (O -; ID J- 

5 jc jz jz j£ C 

1 3 3 3 J 5 >."D § 


2$ hi O 

b in jc E ra 

3 3 3 "^ >» 

ri ^ g in (j) s c6 £ 

historical precedent for moving boundaries and redefining the boundaries themselves to fit their 
purposes, with the implicit consent of the population. All of this evidence from colonial sources 
presents a much less solid image of ethnicity. The negotiable quality of "tribal" boundaries that 
allowed colonial officers to move from separate chieftaincies to federations, to paramount chiefs 
and back to chieftaincies again, indicates a different kind of system at work. Men reworked 
disparate kinds of knowledge about the past into the ethnic histories that they needed to maintain 
smaller locally controlled units. 

Much of the recent work on ethnicity in Africa has stressed the "flexible" nature of ethnic 
boundaries before colonialism. 3 ' However, here, the social boundaries negotiated within the system 
of indirect rule were not so much flexible as they were multiple. Elders brought different kinds of 
boundaries besides the ethnic into play, depending on the situation. Each set of boundaries was 
specifically inscribed on the landscape and related to a particular set of social relationships. Local 
people did not make up new identities or necessarily change preexisting boundaries as much as they 
called on different sets of boundaries, related to other kinds of social identities that both united and 
divided them in different ways. Ethnic identity of a certain kind may have existed in the 
precolonial past but it was only one of many kinds of relational boundaries in operation and only 
became predominant and fixed in the colonial years. 
Literacy and Oralitv 

A central element in the political cultural of colonial rule, and thus of men's evolving 
concepts of legitimate historical narrative, was the imposition of literacy. More than any other 
single feature, writing or the pen became the symbol of the inherent power of the colonial regime. 
Local people both feared it and sought to harness its power for themselves. The colonial chiefs at 

31 See for example, Thomas Spear and Richard Waller, eds., Being Maasai: Ethnicity and 
Identity in East Africa (Athens: Ohio University Press, 1993); Vail, The Creation of Tribalism . 

first resisted sending their sons to school to learn to read and write, but later prohibited any but 
their own sons from attendance. This ambivalence manifests itself in their attitude toward writing 
oral tradition. 

Today's elders and their juniors demonstrate an increasing willingness to rely on written 
versions of the past, products of the colonial historical imagination. One educated elder brought 
out Moffet's Handbook of Tanganyika to prove his migration account of the Ngoreme. 32 Tatoga 
elders insisted that I read Zamani Mpaka Siku Hizi. a wide-ranging account of precolonial 
Tanganyika, uncritically combining historical tradition and colonial anthropology. 33 Although my 
whiteness and association with education may have provoked these comments, respect for the 
written word is far more pervasive. 

Some local intellectuals among the first school leavers have written their own accounts of 
"tribal" history. These manuscripts often lie buried in trunks, moth-eaten, with pages torn out. In 
Ikizu a committee, formed under the Ikizu Development Association, wrote a book of history that 
remains unpublished. 34 A secretary from the Catholic mission wrote a Ngoreme history, getting it 
typed and stenciled at the mission office. 35 Many people promised me manuscripts and then could 
not locate them, others were still in the process of writing. Many Ishenyi elders told me I should 
find a copy of their manuscript and not waste my time doing interviews. There was rumored to be 

32 J.P. Moffet, Handbook o f Tanganyika (Dar es Salaam: Government of Tanganyika, 
1958). Interview with Judge Frederick Mochogu Munyera, Maji Moto, 28 September 1995 
(Ngoreme cf). 

33 Institute for Swahili Research, Zamani Mpaka Siku Hizi (Dar es Salaam: East African 
Literature Bureau, 1 930, revised edition, 1962). 

34 P. M. Mturi and S. Sasora, "Historia ya Ikizu na Sizaki," 1995, unpublished manuscript 
in author's possession. 

35 P. Haimati and P. Houle, "Mila na Matendo ya Wangoreme," unpublished mimeo, 
Iramba Mission, 1969. 


an Ikizu history commissioned by Chief Makongoro in the 1950s hidden because of the secrets it 

contained. These books are a source of pride because they legitimize the group as a "tribe" among 

"tribes." 36 No famous spokesperson has arisen from among these authors, like Sir Apolo Kagwa 

for the Baganda or Samuel Johnson for the Yoruba, as elsewhere in Africa." Yet the instability of 

"tribal" identity in this area has produced a much more widespread, grassroots effort to record the 

history of "tribes." 38 

Writing itself is an instrument of power clearly recognized by the generation of elders 

today. The Germans gave the first colonial chiefs a book and a pen as signs of their authority. 39 

Many elders told me that my own use of pen and paper was a source of anxiety for them (the tape 

recorder was not usually an issue). One remarked. 

In the time of our grandfathers and fathers they were afraid of people with skin like 
yours: they were afraid of the pen. When they saw someone writing, they said that this 

36 See Jan Bender Shetler, "'A Gift for Generations to Come': A Kiroba Popular History 
from Tanzania and Identity as Social Capital in the 1980s," The International Journal of African 
Historical Studies 7.8. 1 (1995): 73-77. 

37 Sir Apolo Kagwa, Basekabaka be B Uganda [The Kings of Uganda], trans, and ed. M. S. 
M. Kiwanuka (Nairobi: East African Publishing House, 1971, first published 1906). Samuel 
Johnson, The History of the Yorubas from Earliest Times to the Beginning of the British 
Protectorate, ed. 0. Johnson (Lagos: CMS (Nigeria) Bookshops, 1960. first published 1921, 
reprinted 1960). There are few academic histories of the Mara Region on which people might base 
their oral narratives. One popular Kuria history in Swahili is Gabriel Chacha, Historia va 
Abakuria na Sheria Zao (Tim es Salaam- East African Literature Bureau, 1963). 

18 While I was in Tanzania 1 contacted some publishing houses on behalf of the authors to 
investigate the possibility of publishing some of these manuscripts. When I left three of them were 
being reviewed by Ndanda Press, a Catholic publisher in Mtwara, Tanzania. Marwa Kishamuri, 
"Historia ya Abakiroba: Desturi na Mila Zao," unpublished manuscript in the author's collection 

39 Interview with Mohere Mogoye, Bugerera, 25 March 1995 (Nata <f), mentions this in 
connection to the story of how Megasa was made the first Nata chief. It is a clear motif in all of 
the chief-making stories throughout the region. 


person would take us with them in the writing. So that is why they did not want us to go 
to school, they said those who go to school will be taken away by the white people. 40 

In my own research, when we visited the site of a rainmaker ancestor's grave the elders told me to 

put the pen and paper away because the spirit of Gitaraga would not like them. 41 

The power of the colonial officers resided in their ability to capture words and then control 
their use in other contexts. Confronted with written accounts that conflicted with their own 
versions, elders would often accept the veracity of the written version. Commitment to the written 
word automatically legitimizes a particular version of the past. Many manuscripts that I saw 
named no author and often listed their sources as "the elders of X." Colonial officers seldom cited 
sources by name or position for the "tribal legends" they collected. In the tribal paradigm, only 
one true version of the past may exist if the "tribe" is organically integrated. Writing constitutes a 
powerful instrument of control because it fixes knowledge, and at the same time dissociates that 
knowledge from the explicit interests of individuals. 

Local intellectuals learned how to use this power in their letters of complaint to the colonial 
government. Those who wrote these letters used "tribal" history to support or to protest against a 
chief and his claim to power. A 1 949 letter from "the people of Zanaki" to the Honorable Chief 
Secretary in Dar es Salaam affirms that Ihunyo is the legitimate Chief (Mwami) of Zanaki because 
he is 21st in the line. The authors cite the fact that even the District Commissioner, E. C. Baker, in 
1 928. investigated these claims and found them to be true, while they suspect that "Bwana" Cory's 
investigations in 1 945 may have brought on the present movement to have Ihunyo removed. The 
precolonial Mwami title was inherited through a line of important rainmakers in the area, though in 
governance each territory was relatively independent. Lacking any traditional Zanaki-wide 

40 Interview with Tetere Tumbo, Mbiso, 5 April 1995 (Nata <f). 

41 Interview with Keneti Mahembora, Gitaraga and Mochuri, 9 February 1996 (Nata a"). 

authority, the written word gave some the means to convert the Mwami title into a chiefly claim 
over a much wider area. 42 

In Sizaki. a continual barrage of letters from 1927 to the 1950's that railed against Chief 
Ruhaga, claimed he was the son of a Sukuma dream-prophet, welcomed in Sizaki not as a chief, 
but thanks to his medicine to catch fish. His son later convinced the Germans that he was the chief 
and had many of the opposition elders killed. The protesters established their claim as "original" 
inhabitants, now ruled by "alien" Sukuma. A 1944 petition, written by soldiers of the King's 
African Rifles, reminded the Governor of their service and loyalty. They likened the Sukuma 
rulers to "Nazis" and themselves to "slaves," appealing to the League of Nations Charter. The 
same group wrote another letter speaking for the Abanyase, a high ranking group of eldership title 
holders, whom they represented as "the ruling Sizaki clan." The colonial government could not 
entertain these claims because the Sukuma had since become the majority in Sizaki and the Chief 
had been elected by popular vote. In 1 952 the colonial government deported the author of some of 
these letters.* 13 Because of the high potential gain or loss in the colonial system men were 
encouraged to rethink their pasts as "tribal" pasts and to commit them to writing to enter the 
colonial field of historical debate. 
Men's Extensive Geographical Knowledge 

Besides their participation in the colonial political economy, men had other reasons for 
possessing historical knowledge that was more extensive geographically than that of women. Men 

Letter from the Zanaki people to the Chief Secretary, Dar es Salaam, 20 June 1949, and 
letter from D. Dowsett, DC Musoma to PC Lake Province, 21 August 1950, 1949-50, Native 
Chiefs, Musoma, Secretariat Files, 29626, TNA. 

43 From Soldiers of the K.A.R. to the Governor, T. T, 1 December 1944, Petition to the 
Governor from the Secretary of the Abanyase, Sizaki, 16 May 1945, Native Chiefs, Musoma 
Secretariat Files. 29626, TNA. P.C. Lake Province to Chief Secretary, Dar es Salaam, 1 
September 1952, Complaints, 215/P.C./50/5. TNA. 

hunted, raided and traded in the domestic economy, activities that took them far from home 
seasonally and for long periods of time. These activities required an expert knowledge of 
landscape, terrain, and ecology. Men had to know where game could be found at any time of the 
year, how to track them for days and still find their way home. They relied on generational 
knowledge for the location of the best hunting camps, water holes, shooting blinds, hunting pit sites 
and arrow-poison trees. The retired generation was obligated to escort the new generation on their 
first hunting or arrow-poison gathering journeys. The youth carried the elders' packs and protected 
them at night as repayment for this service. 44 As they went, elders named each hill and rise, and 
noted each river, seasonal water source and pool. The names of these places often corresponded to 
people who lived there or incidents that took place there. 45 Walking the trail, or around the fire at 
night in the hunting camp, the younger generation heard the stories behind these place-names. 
Remembering was crucial to their survival in the wilderness and their ability to bring home meat. 

When men followed cattle raiders to recover their loss, they would often cross hundreds of 
kilometers of bush into Maasailand or north to the Kenya border. An important trade route of the 
late nineteenth century lay toward Sukuma, where western Serengeti men traded wildebeest tails 
and other wild animal products for goats, tobacco and salt. An even earlier trade route took them 
to Geita in Sukuma for iron hoes. On the routes of raid and trade, local men found hospitality and 
formed friendships with those who had been strangers. Sometimes they sealed these friendships 
with oaths of blood brotherhood. Men maintained their friendships by visits in both directions, 
providing more opportunities for trade. Through these friendships men learned the stories of other 
peoples and had to modify their own stories to account for historical similarities and differences. 

44 Interview with Megasa Mokiri, Motokeri, 4 March 1 995 (Nata a"). 

45 In an interview with Yohana Kitena Nyitanga, Makondusi, I May 1995 (Nata <f), he 
named 1 13 such places, north, south, east and west of his present home. 

Finding that other peoples used the same praise names, cattle brands or claimed the same origin 
places caused them to rethink their own stories of the past to incorporate lost brothers or joking 
relations. Men's stories ordered and maintained wide-ranging regional relationships by solidifying 
reciprocal obligation and responsibility. 
The Spaces of Men's Historical Knowledge 

When asked where they learned their historical knowledge most men said that a father, 
grandfather or uncle taught them. One man told me that the ability to learn history is a gift from 
God, just like healing, prophecy or wealth. Like these other gifts it. too, runs in the family. 44 
Others said that boys learned from older men while working on grain storage bins, fixing the 
corral, building a house or herding. Girls were not usually around at these times. 47 Fathers had a 
responsibility to teach their sons specific historical information that affected the ongoing survival 
of the lineage. This included information like prohibitions against certain marriage partners or 
food, activities required by the ancestors, the location of grave sites, the histories of important 
ancestors for propitiation, and unresolved blood feuds with other families. Without this kind of 
historical knowledge a son might bring ruin on the household. 

Many elders related that in their own childhood experience, elderly men gathered regularly 
in local courtyards to visit and play bao (a board game common throughout Africa), eating their 
meals together there. They had practically no work responsibilities at their homes. Leisure was 
the reward of age. Most people told me that they learned stories of the past from sitting quietly at 
the feet of elders when they were children. 48 The elders were not teaching the children but carrying 

46 Interview with Samweli M. Kiramanzera, Kurusanda, 3 August 1995 (Ikizu <f). 

" Interview with Nyawagamba Magoto (Nata <f), Kinanda Sigara (Ikizu rf) Dhah 
tmara (Nata 9), Mugeta, 9 March 1996. 

48 There were almost always children in attendance at my interviews. 

on their own conversations. The youngest children of the family often heard the discussions of old 
men, since they were more likely to be the ones left in the compound to run errands for the elders 
when the older children and wives went to work in the fields. Young men could also listen if they 
did not interrupt. There is a saying that, 'a youth who does not sit with the elders is like a wild 
animal, running here and there with no home."" Those who learned history had both the natural 
inclination and proximity to these groups of elders. 

Older men prefer to do their socializing with beer and so try to arrange as many 
opportunities for this as possible. While today this is often done at "clubs" where they sell beer, in 
the past other occasions facilitated elders' storytelling over pots of beer. Within one neighborhood 
they each might agree to contribute a certain portion of the required grain and wives' labor for 
making beer; only those who contributed got an invitation to drink (msororo). A man might also 
call his neighbors to help him weed a field; in return he gave the fathers of the young men who 
worked a beer party (risaga). In the past only the elders were allowed to drink beer and thus to 
engage in the protracted discussion of social relations and the past. 50 

Other opportunities for men to gather at leisure to talk were, and still are, the formal 
celebrations of weddings, funerals, circumcisions or eldership titles, which often last for many days 
at a time. At these gatherings, as I observed them, smaller groups eat or drink separately, 
according to age and gender-old men, young men, old women, young women, and children. Most 
of the women are involved with preparing food, hauling water or gathering firewood. Younger 
women are constantly shuttling back to their own homes to make sure that the children they left 

49 Interview with Tetere Tumbo, Mbiso, 5 April 1995 (Nata tf). 

50 Interview with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi a"). 

behind are being cared for and have food. Even the oldest women hold babies or sort beans. 
Women do not have as much leisure time at these gatherings. 

Testimonies of elders and my own observations concur on the format and content of 
discussions about the past at these gatherings of elders. 51 At a beer party or an afternoon bao 
game the conversation usually begins with the news of the day, leading in turn to the necessity for 
action by the community: what should be done about the young woman who ran away from her 
husband, the approach of cattle thieves, a suspected case of witchcraft or the lack of rain? In the 
discussion of these events elders use stories of the past to explain behavior and its consequences, to 
figure out relationships between those involved and their obligations to each other, or to cite 
precedent for the required action. 

People highly value the gift of speech and these discussions often become intense 
arguments as each man elaborates his point by invoking stories from the past. Only after the elders 
chew over these problems can more formally constituted authority take action. Men do not learn 
their historical knowledge from a formal cycle of fixed narratives but from the bits and pieces of 
the past that they deploy to debate community problems. Because these are cases of judgement, 
history is negotiable and used to support a particular agenda. The frequency with which elders tell 
and remember any particular story about the past depends on its power to explain current 
relationships. Few occasions exist when men narrate the historical corpus for its own sake. 

Many rituals that people rarely perform anymore, such as eldership initiation ceremonies, 
circumcision and the propitiation of spirits at sacred places included the recitation of historical 
traditions. At dances (formerly, but no longer, held every full moon) youth sang songs and shouted 

51 As a young woman I did not have forma! access to men's beer parties but did overhear 
enough of their conversation or was told enough about what goes on in them to make these 

praise-names that told about the past. Historical information was woven in and out of all these 
rituals of everyday life. The stories that grandmothers told children after dark included historical 
characters and events. Learning history was not the memorization of a corpus of oral tradition but 
was, rather, a familiarity with "matters of the past" from all these everyday encounters. Many of 
these occasions no longer exist. Elders still recount historical knowledge as a matter of course in 
preparation for weddings (learning about the background of the intended) and at funerals (hearing 
about the life of the deceased and dividing the inheritance). 

Elders fear that they no longer have the opportunity to pass on historical knowledge to 
young people who do not care about these "matters of the past." Old men still have beer parties 
but the young men are at school, in the cities or at their own disco parties. One of the few ways 
that a youth can still learn history is by asking his father, very much in the way of an interview, 
just as colonial school boys did when their teachers sent them home to collect historical information 
from their fathers for composition class. In each research area I found young men with a deep and 
intimate knowledge of the past, committed to its survival. Still, they are the exceptions. As this 
generation of elders fails to pass on its knowledge in a general way through discourse, historical 
knowledge is increasingly becoming the property of ritual experts. The common perception that 
historical knowledge will be forgotten in the next generation is more a result of its changing form, 
than its absolute loss. 

Women's Intimate Community Knowledge 

During the colonial era, when belonging to a "tribe" with a specific history became 
imperative, men significantly reshaped their historical narratives. It was in walking the paths 
outside their home communities in hunting, cattle raiding, trading and migrant labor that they 
formulated their account of a "tribe" among a nation of "tribes." Women did not participate 
directly in the construction of these narratives of ethnic origin and migration, settlement and 

conflict, in part because of the position imposed on them during the colonial era. While women 
had some latitude for formal influence in the community before colonialism, as rainmakers, 
prophets or titled elders, the colonial administration systematically denied them a voice in formal 
politics. 52 Women marrying outside their homes often functioned in the past as intermediaries 
between clans and ethnic groups. When the emphasis shifted to ethnic unity and exclusivity in the 
colonial era women became outsiders and strangers, rather than valued links to others. This 
structural position of women, as outsiders in the homes of their husbands, denied them a formal 
role in the creation of ethnic histories. 

My experience living as a woman in the company of other women in Nata during the year 
and a half of research allowed me to observe women's daily activities and exchange of information 
in a rural setting. 1 describe the female spaces that structure the kinds of information that women 
share in the present, as I observed them. However, these observations are also consistent with the 
testimonies of elderly women about their youth and with the few ethnographies for the area that do 
exist. This leads me to believe that what I describe here would also apply to the colonial period 
and perhaps into the late nineteenth century. Shifts in gender relations that affected the 
transmission of knowledge may have taken place during the disasters when the dangers of 
unsettled times increasingly restricted women's movements and sphere of influence. 

My experience interviewing women demonstrates the differences in narrative style and 
setting of narration between men and women. Older women found the interview process much 

52 This is another story to be told elsewhere but concerns the 1 ) the colonial refusal to 
acknowledge the leadership role of healers/prophets/rainmakers, 2) the denial of a woman's right to 
have an independent household without a man or to obtain offspring through what is known 
elsewhere as "woman-to-woman" marriage, giving woman access to independent wealth, and 3) the 
shifting of "traditional" inheritance and marriage laws to favor men. For an analysis of the 
changing role of women in colonial society see Steven Feierman, Peasant Intellectuals- 
Anthropology and History in Tanzania IMaHknn- University of Wisconsin Press 1990) on 181- 

more intimidating then men, particularly when many men were listening. I had the best interviews 
with women when I used a woman assistant and no men were present. When men were present, the 
woman would defer to the men, or the men themselves would correct the woman. Areas considered 
within a woman's sphere, such as women's circumcision, were not acceptable topics to discuss in 
mixed company. Older women were less likely to understand Swahili than were their male 
counterparts. While men easily launched into long uninterrupted historical narratives, I had to ply 
most women with a continual barrage of questions to solicit more than brief answers. Women 
preferred a dialogue among their peers rather than the monologues demanded by formal interviews. 
Women often loosened up when they sang songs, told a story about one of their grandparents, a 
folktale, or reminisced about the past with their friends." 
The Paths of Women's Daily Interactions 

Women have not been privy to ethnic histories, at least in recent times, because of their 
interior position within the gendered construction of space. Women's knowledge of the past 
consists of the details of family genealogies, family histories (both natal and marital) and 
community stories-all of which concern how everyone is related to everyone else inside the 
community. 54 When I questioned one male elder about the relationships between members of 
households in his village, he frequently went out to consult his wife, who would not budge from the 
kitchen to join us. Women command expertise in this kind of knowledge because this is their 

53 For an analysis of the interview process as part of the "male sociocommunication 
subculture" and thus intimidating to women see, Kristina Minister, "A Feminist Frame for the Oral 
History Interview," in Women's Words: The Feminist Practice of Oral History, eds. Sherna Berger 
Gluck and Daphne Patai (New York: Routledge, 1 991 ), p. 3 1 . 

54 This has been observed by many observers in Africa, including early anthropologist, 
Lloyd A. Fallers, Bantu Bureaucracy: A Cent ury of Political Evolution among the Basoga of ' 
Uganda (Chicago: University of Chicago Press, 1965), p. 90, who states that "lineage males must 
often draw upon the geneological knowledge of wives and mothers.... women often remember 
genealogical complexities better than men." 

source of power and prestige. Women are responsible for feeding their households and raising their 
children. The household is still the basic unit of production and exchange, and women carry out 
their tasks by relying on relations formed by reciprocal obligation rather than cash. In this 
situation the better a woman understands relationships within the community, the more adept she 
can be at establishing mutual obligations for all of the large and small matters of daily life. 

Women's knowledge grows out of the community networks of reciprocal exchange that 
they construct to ensure survival and prestige for themselves and their children. A new bride 
begins this process as a stranger in her mother-in-law's house. Marriage is fundamentally the 
transfer of a woman from her natal home to that of her husband's family, under whom the new 
couple must live, at least until the first child is born." The new bride will have less daily 
interaction with her husband than with her mother-in-law, on whom the bride's successful 
assimilation into a new set of relationships depends. A mother-in-law looks forward to the day 
when her daughter-in-law comes to relieve her of all of the difficult household chores-farming, 
grinding, hauling water, gathering firewood and some cooking. 

A good mother-in-law, who is also a stranger in the patrilineage of her husband, gives her 
daughter-in-law a thorough knowledge of everyone in the family, the genealogical tree as well as 
individual character and reliability. This often happens when guests come and news of the 
community begins to circulate, or when she tells stories about ancestors. A woman gradually 
begins to form her own relationships in the community and among her husband's kin, while still 

55 See Huber, Marriage and Family , pp. 69-91; and Tobisson. Family Dynamics , pp. 134- 
137; for the relationship of a woman to husband's family in this region. 

under the tutelage of her mother-in-law. Families entrust continuity of the most intimate 
knowledge about insiders to those in the family who are structurally outsiders. 56 

A woman does not, however, give up knowledge of her natal community, which is crucial 
to the survival of herself and her children. A married woman maintains her relationship to her 
parent's home and kin by regular, if not frequent, visits and the exchange of gifts. Her mother is 
also required to visit on specific ritual occasions." These exchanges allow a woman to pass on the 
knowledge of these relationships to her children. A woman depends on her father and brothers, or 
her mother's brothers, to be her advocates if something goes wrong in her marriage. Her husband 
also strives to maintain a good relationship with his father-in-law, to gain sympathy in case of a 
conflict (direct contact with his mother-in-law is prohibited). A woman continues to play a critical 
ritual role in her natal home as the paternal aunt (mwisenge) to her brother's children. Her brother 
(mame) has special responsibilities to her children and in some Mara societies they inherit from 
him rather than from their father. Many people described their maternal kin as a place of 
acceptance and refuge in times of trouble. A man could turn to his mother's family to pay the 
blood compensation fine if he were charged with murder. 58 Families value women as wives 

56 Interviews with Susana Nyibikwabe Mayani, Bugerera, 1 February 1 996 (Nata 9 ); 
Kimori Gamare, Bugerera, 28 February 1996 (Nata/Ikoma 9); and Sumwa Nyamutwe, Muget'a, 9 
March 1 996 (Nata ¥ ). In the case of a second or third wife the role of the mother-in-law would be 
taken over by the first wife and is more tenuous. Part of the power of the first wife is this position 
in and knowledge of her husband's family. 

57 Interview with Weigoro Mincha, Kemegesi, 29 March 1996 (Ngoreme 9). 

58 Many of the Lakes peoples as well as Zanaki and lkizu. Interviews with Samweli M. 
Kiramanzera, Kurusanda, 3 August 1 995 (lkizu cf ); Zabron Kisubundo Nyamamera and 
Makang'a Magigi, Bisarye, 9 November 1995 (Zanaki <?). For the role of the maternal uncle in 
Tatoga Rotegenga society see, Interview with Ghamarhizisiji (Uyayehi) Nuaasi (Tatoga 9 ) and 
Gesura Mwatagu (Tatoga a"), Issenye, 8 May 1995. 


mothers, sisters and daughters both for the links that they create between families and the spheres 
of knowledge they manage. 59 

Women also specialize in knowledge of community relations. A woman's daily 
relationships take place in the space of home and the community, revolving around the exchange of 
food, household implements and services. Women are responsible for feeding themselves and their 
children through control over their own fields, storage bins and tools. To do this they constantly 
manage an intricate set of reciprocal relationships, involving an ongoing exchange of goods 
between neighboring homesteads. When a woman runs out of millet flour for the staple porridge 
she goes to her neighbor to ask for some, who in turn would ask her when she ran out of something 
else. Women understand these exchanges as "gifts" and do not repay them one for one. A woman 
does not return borrowed flour until the neighbor that she borrowed it from runs out and comes to 
ask for it. If she borrows a tool not constantly used by her neighbor in everyday work, that item 
stays at her house until the owner needs it and comes to get it. Ideally, a woman wants to have 
both credits and debits outstanding at any one time. It is not to her advantage to balance all 
relationships. People who owe a debt represent potential sources for future needs; those who have 
credits will come to visit when they collect. Women move around the community daily as they 
carry these things from house to house. 

Women visit each other most often as a result of maintaining these relationships of 
exchange (asire), but they also make purely social visits on the occasion of marriage, death, 
illness, ritual, celebration, or friendship. The exchange of gifts, most often food (omotoro), marks 
these social visits. They say that, 'a house without visitors is a house without blessing.' The most 
obvious sign of a prestigious woman is a homestead constantly tilled with people. Women praise 

59 Interview with Machota Sabuni, Issenye, 14 March 1996 (Ikoma cf). 

and seek out a woman known for her generosity, not only for the exchange of goods but also for 
advice and good conversation. They insult the greedy woman, saying that she "does not like 
people," and will eventually isolate her. However, the prestigious woman who gives away a lot 
must also ask for things in return. Otherwise, the circulation of goods and visits ends. A generous 
woman calls in these obligations when she puts on a large feast or a work party. Such events last 
for days and require a lot of women's labor, both before and after the feast, in cooking, hauling 
water, gathering firewood, making beer and grinding grain. This is the ultimate test of how 
skillfully a woman has maintained her community networks. All these situations bring women into 
daily contact with each other. 

A woman's main source of information about the community and respite from work comes 
from these social interactions of daily exchange. Women acquire certain kinds of knowledge in 
these interactions in which men do not participate. Women know the news of the community. 
Information spreads along the paths of women going to each other's houses to exchange favors. 
They must greet everyone they meet on the path and exchange pleasantries. 

The spatial structure of the community facilitates the exchange of news. Paths usually 
pass directly through a household courtyard or alongside it, so that someone walking by must 
naturally greet the people working there. Houses are small, mostly used for sleeping and storage, 
and the daily work takes place outside in the courtyard. A woman on her way to get some flour 
from her friend learns who is going to town that day, who was sick, who has guests, and who is 
harvesting millet. She hears news from the next village and political discussions from the national 
radio station. Men's paths run in direct, rather than circuitous routes, and men seldom visit in 
courtyards during the day except on specific business or social occasions. 

Yet men seek out the informal community knowledge of women before they engage in their 
formal negotiations and exchange of information. They value the wife who attracts many visitors, 

even if it means a bigger outlay in food. The networks of women and their exchange of flour and 
beans inform the big decision-making in the community that men do over pots of beer. 

One of the few formal ways in which women pass on knowledge to future generations is by 
storytelling. Grandmothers or elderly widowed aunts tell stories to their grandchildren before they 
sleep at night or as they do some time-consuming, repetitive work like sorting cotton or beans [See 
Figure 2-4: Women's Story Telling, p. 91]. The stories that women tell to children include the 
typical African animal and monster (mwinani) stories as well as the adventure stories of past 
ancestors. [See Figure 2-3: Mwinani by local artist Deus Nyahega Tumbago, 1995.] Only girls 
hear the stories that grandmothers must tell at night. When girls approach puberty, they move out 
of their mother's house and into another house with other girls their age from the neighborhood, 
supervised by an elderly single woman of the family. Young boys sleep with age mates, without 
supervision, until they marry. These older women must make sure that the girls do not get 
pregnant before circumcision. Because these women are in an alternate, rather than an adjacent, 
generation to the girls they can talk to them frankly and openly about sexual matters. They teach 
the girls about sex, the duties of a good wife, and use stories from the past to illustrate these 
lessons of morality. Because of their position in the community, these single women often know a 
lot about the past. 60 Folktales intersect with oral traditions of a historical nature by providing the 
moral themes that recur in each. As more families live in urban areas, in nuclear family units, 
within a common house, these storytelling spaces are disappearing. Grandmothers now live back 
in the village, where the younger generation rarely visits. 

60 Folktales on videotape, Bugerera, 17 August 1995. Interviews with Raheli Wanchota 
Nyanchiwa, Morotonga, 16 March 1996 (Ikoma ?); Weigoro Mincha, Kemegesi, 29 March 1996 
(Ngoreme ¥). See also David William Cohen, "Doing social history from pirn's doorway," in 
Olivier Zunz, ed., Reliving the Past: The Worlds of Social History (Chapel Hill: University of 
North Carolina Pres, 1985), pp. 191-235. 

Figure 2-3: Mwinani by local artist Deus Nyahega Tumbago ©1995 


Men fear women's knowledge because women know too many intimate details about them. 
Men often ridicule women's knowledge as "foolish gossip" because women can expose their 
misadventures. One elderly woman said that her husband restricted her friendships with other 
women so that she would not gossip. Women are often accused of witchcraft because they hold so 
much power over men with their knowledge. 61 In the past women openly used this power over men 
in their grinding and hoeing songs. A Ngoreme woman told me that women used to form self-help 
groups (chesiri) to weed each other's fields. As they worked, they would sing songs to make fun of 
someone who was lazy or to expose another fault. Other songs would tease men about their lovers 
and chide them for leaving their wives. By singling out a particular man for ridicule, the songs 
served to discipline him and cause change in his actions. Women, singly or in pairs, ground grain 
within a homestead where many people passed by or worked within earshot. A woman sang a 
grinding song to praise or ridicule her lover, or that of her friend. She sang out the attributes, or 
abuses, of a lover in explicit detail, much to the embarrassment of the man. Thus women used the 
power of their intimate community knowledge to change the decisions or actions of men whom they 
could not influence directly. 62 Women brought to the attention of elders the men who needed to be 
disciplined and shame forced the deviant to seek correction. [See Figure 2-4: Women's Story 

Women also have some knowledge that they explicitly keep secret from men. In 
interviews, some women claimed that only men had secrets and that women's eldership titles 
involved no masubho or secret initiation knowledge. Yet women do have secrets, that preserve an 

61 Interview with Sumwa Nyamutwe, Dhahabu Gambamara (Nata ?), Nyawagamba 
Magoto(Nata cf), Kinanda Sigara (Ikizu cf), Mugeta, 9 March 1996. Boddy, Wombs and Alien 
Spirits, p. 45, discusses the relationship between women's knowledge and "gossip." 

62 Interview with Weigoro Mincha, Kemegesi, 29 March 1996 (Ngoreme ¥). 

Dhahabu Gambamara with daughter Kimori, Bugerera, 
12 August 1995, Pounding Cassava 

Weigoro Mincha and friends, Kemgesi, 29 March 1996, women's songs. 
Figure 2-4: Women's Story Telling 

autonomous sphere of female authority over aspects of fertility and reproduction. A woman 
specialized and initiated into this knowledge (omwikarabutu) directed Nata women's circumcision 
ceremonies, the proper disposal of flesh after the ceremony, and the critical days of healing." 
Other women with specialized knowledge supervised the rituals for a young woman's first 
pregnancy and birth. Women maintained this ritual knowledge through eldership titles that ran 
parallel to and separate from men's eldership titles. During the ceremonies of first pregnancy men 
had to leave the homestead. An Ikizu woman could rise through the ranks of eldership titles, 
which involved powerful initiation secrets, despite the rank of her husband." Elderly women 
expressed the most concern about keeping this knowledge secret by questioning how much others 
had told me. They said that men would sell this information for their own benefit rather than look 
out for the good of the community. Women maintain autonomous control over the precious 
community resource of fertility and reproduction with their secrets. 
Extraordinary Women Privileged to Men's Knowledp e 

Beyond influencing men's historical knowledge and actions, women also found ways of 
crossing the gendered boundaries inscribed in the social spaces of the community. As scholars 
have argued for other parts of Africa, biological sex and gender are not necessarily correlated. 65 In 
the western Serengeti categories of women exist with access to men's knowledge. Because of their 
special position in the community these women often exceed men in their knowledge of the past as 
defined by men. I identified these categories by noting what kind of women could tell me the 

63 Interview with Sumwa Nyamutwe, Mugeta, 9 March 1996 (Nata ¥). 

64 Interview with Baginyi Mutani and Mayenye Nyabunga, Sanzate, 8 September 1995 
(Ikizu ¥). 

" Bodd y. Wombs and Alien Spirits , p. 56; Ifi Amadiume, Male Daughters. Female 
Husbands: Gender and Sex in an Afri can Society (London: Zed Books Ltd, 1987), p. 15. 

commonly recognized historical narratives. Because of the situational nature of gender identity, 
these women crossed gender boundaries in particular contexts. The words to describe these women 
predate the colonial period and are used across the region in various ways, making it reasonable to 
assume that these institutions have been in existence at least since the last half of the nineteenth 
century, changing to meet new circumstances over time. 

Perhaps the most obvious but least accessible category of extraordinary women in the 
western Serengeti is that of dream-prophets, healers and diviners, who used historical knowledge to 
practice their art. The general word for prophecy, including healing, obugabho, does not seem to 
have been gendered. Narratives about the past often mention women as rainmakers, healers and 
prophets, including the eponymous ancestresses in the line of Ikizu and Zanaki "chiefly" 
rainmakers. A woman became a prophet, not from her own volition, but because an ancestral spirit 
chose her. When clients came to the prophet for help she used her knowledge of past relationships 
and networks to interpret the problem. The prophet might advise her client to perform a ritual for 
the Maasai or Tatoga ancestors of his grandmother or to sacrifice a goat at the grave of another 
ancestor. These women spoke only reluctantly, if at all, about this knowledge in interviews because 
they understood it as part of the secret of their craft. They said that their historical knowledge 
came from the ancestral spirit in dreams. 

My colleagues often recommended to me elderly childless women as those most 
knowledgeable about history. A childless woman (omogomba) has very little status in a society 
that emphasizes growth and fertility. Her dependency on the goodwill of kin rather than the 
obligation of her children puts her in a particularly vulnerable position. 66 I heard stories of 

66 On the position of these women as widows more widely in Africa see, Betty Potash, 
Widows in African Socie ties: Choices and Constraints (Stanford, California: Stanford University 
Press, 1 986) and Michael C. Kirwen, African Widows (Maryknoll, New York: Orbis Books 

husbands who drove their childless wives out of the marriage (although bridewealth is not returned 
because of infertility). Others allowed them to stay but in a subordinate position to their co-wives 
who gave birth. People assume that childlessness results from a woman's infertility, because if a 
man suffers from infertility he quietly allows his wives to become pregnant through relations with 
other men in order to extend his own lineage. 

An omogomba does not have the labor of children on the farm or in the house. She builds 
and repairs her own house unless someone offers to help (as a son is obligated to do). She has no 
children to care for her when she gets too old to farm. Women do not inherit directly from their 
husbands but only indirectly through their sons, who are bound to support their widowed mothers. 
A widowed omogomba formally stands at the mercy of her brother-in-law, appointed by the family 
to care for her. If she does well, she can pay bridewealth for a mkamwana (a "daughter-in-law" 
under the institution of "woman/woman" marriage) who will provide children to take care of her. 
Most elderly women without children live out their lives moving from relation to relation, without a 
home or security beyond tomorrow. 67 

On the other hand, childless women possess a paradoxical kind of power in the community. 
They are female and yet anomalous because they are not mothers. They lack fertility and yet have 
access to the source of community fertility through the tasks often assigned to childless women as 
midwives and caretakers for young unmarried women. Because they do not care for their own 
children these women often find themselves in a position to help others raise children, deliver 
babies and dispense wisdom. Nata elders propitiate the ancestral spirits at a place called Nyichoka 
(female snake). They ask for rain and fertility at the grave of a childless woman who despaired 
because of the ostracism she felt from her community and threw herself into the pool. She appears 

67 Interview with Mwenge Magoto, Bugerera, 5 November 1995 (Nata ¥). 

as a snake and her hearthstones lie at the bottom of the pool. A menstruating woman cannot draw 
water from the pool or the water becomes churned up and misfortune comes to the community. 
Women dance the eghise at the pool to make her happy. 68 Throughout the Lakes Region people 
have solved the problem of infertility by initiating these women as spirit mediums, who then beget 
spirit children. 69 

A childless woman often has access to men's knowledge, in part, because of her social 
position: without children to care for and feed she can move beyond her household. She has more 
time to go to the fields to farm, to visit friends and family, and to circulate around the village. 
Because of her increased availability, a man often chooses the childless wife to carry his beer 
straws and stool to the beer parties of the elders as a sign of his prestige. When he gets to the beer 
party his wife sits between his legs, drinking beer with them, and hears men's stories about the 
past.™ When 1 insisted on interviewing women, colleagues often took me to meet an omogomba, 
though they never consciously recognized this as a reason for her knowledge. These women said 
that they learned so much more history than other women because of their interest and the 

An independent woman without a husband, managing her own family and property, called 
an omosimbe, also has no man to tell her to stay in the homestead, and so walks about the village 
at her own time and inclination. In the past a woman would not have preferred this status; it 

68 Interviews with Mokuru Nyang'aka, Nyichoka, 9 February 1996 (Nata <?); and Mahiti 
Kwiro, Mchang'oro, 19 January 1996 (Nata <f). 

69 Renee Louise Tantala, "The early history of Kitara in Western Uganda: Process Models 
of Religious and Political Change" (Ph.D. Dissertation, University of Wisconsin-Madison, 1989); 
Schoenbrun, A Green Place , p. 438. 

70 Interviews with Nyangere Faini, Bugerera, 22 November 1995 (Nata ?); 
Nyawagambwa Magoto (Nata <f), Dhahabu Gambamara (Nata 9), Kinanda Sigara (Ikizu <f), 
Mugeta9March 1996. 


would indicate she was the victim of some unnatural circumstances. Nevertheless, the community 
respected independent women, a position found throughout the Lakes Region." A woman most 
often becomes an omosimbe if her father either has no sons or does not think any of them capable 
of handling his property. He then designates a daughter to take over for him when he dies. She 
takes on a man's role as head of the homestead, having children fathered by a casual lover. 72 A 
woman specialized as a healer or a rainmaker sometimes assumes the role of an omosimbe because 
of restrictions against marriage imposed on her by her erisambwa, the spirit who directs her 
work. 73 

One woman that I met had run away from her home in Kuria when her marital family 
accused her of witchcraft because of the deaths of her five infants. She walked to Nata and found 
refuge with a man for a time. Her sister, who also left Kuria because she could bear no more 
children, later joined her. Together they set up their own homestead in another village and two 
more single women, one from Buhemba and another from Ikoma, came to live with them. Each 
sister paid bridewealth for a mkamwana ("daughter-in-law") who also took up residence with them 
and bore them children. 74 Many young Nata women I talked to preferred the life of an omosimbe 
to that of a wife, setting up independent households, often with small businesses in the towns. 
Betty Potash supplies many case studies of East African widows who found ways of avoiding 

71 Schoenbrun, A Green Place , pp. 238-247. 

72 Interviews with Nyamaganda Magoto, Bugerera, 2 October 1995 (Nata cf); 
Nyawagamba Magoto, 16 July 1995 (Nata <?). 

73 "In North Forest society, patriarchal and lineal idioms of inheritance and descent politics 
gave rise to an institution unique in Kivu Rift society, the spirit wife (kehanga, creator's wife). 
This woman was consecrated by her lineage to a spirit who resided either on one of the many 
nearby volcanoes or who was the shade of a departed member of her patriline. As such she 
remained in her natal lineage." Schoenbrun, A Green Place , p. 243. 

74 Interview with Paulina Wambura, Bugerera, 16 April 1995 (Kuria ?). 

levirate marriages to remain independent." In Ngoreme, omosimbe has taken the connotation of a 
prostitute, and they call an independent woman a "widow" (omosino). 16 

An omosimbe assumes many of the characteristics and roles of a man in her relationships. 
She goes independently to men's beer parties and sits with them, exchanging stories. In the past, 
men did not allow women to attend these parties unless specifically brought by their husbands. An 
omosimbe confers as a man, often with the assistance of her brother, in negotiations over 
bridewealth and at other points where her property is in question. One man told me that men often 
hold their secret meetings to plan for age-set rituals or other important community matters at the 
home of an omosimbe. She cooks for them and is allowed to sit in on their council. The role of an 
independent woman obligates her to make beer for her lover and his friends, who help her to farm 
or herd, but also allows her to participate in their conversation. The same man added that married 
men prefer to drink beer at the home of an omosimbe, confiding in her many things they would 
never think to tell their wives. 77 Thus abasimbe and abagomba, acting in men's roles or outside 
women's roles, participate in discussions that most women do not hear. 

Elderly, post-menopausal women have considerable power and respect in their 
communities, particularly if they have a lot children, and also mediate the boundaries of gender in 
particular circumstances. The word for "old woman" in Nata is omokuungu, derived from a Lakes 
Bantu root meaning a "rich person," itself derived from the proto-Bantu verb -kung, "to gather, to 
assemble." In other places around the lake the word refers to rich men with land and followers or 

75 Potash, Widows in Africa , pp. 1-43. 

76 Interviews with Baginyi Mutani, Mayenye Nyabunga, Stella D. Katani, Sanzate, 8 
September 1995 (Ikizu ¥); Alphaxad Magocha Matokore, Kemegesi, 29 September 1995 
(Ngoreme cc). Ngoreme-F.nelish Dictionary . Parish Office. Iramba, n.d. 

77 Interview with Nyawagambwa Magoto (Nata <r), Dhahabu Gambamara (Nata ?), 
Kinanda Sigara, Mugeta 9 March 1996 (Ikizu <?). 

the clients of the king. Uniquely in Mara languages it refers to elderly women and their wealth in 
children. The verb "to grow old" in Mara languages, -kunguha, is also derived from this root but 
the noun form refers only to older women. 78 Sarah Le Vine's study of Gusii women (also Mara 
speakers) notes that elderly women enjoyed the freedom to talk in public, even appearing "raucous 
and openly aggressive" and drinking beer at parties. Women achieved status as elders with others 
to help them, some leisure and a say in family and community affairs.™ 
Concepts of Gender Identity: 

The discussion of extraordinary women and their access to men's knowledge brings us 
finally to a consideration of concepts of gender. Are these extraordinary women then male? Janice 
Boddy reminds us in Wombs and Alien Spirits , that gender is a "symbolic construct," varying over 
time and from one society to the next. As women researchers interviewing women in other places 
we may share the same biology but we cannot assume that we share the same gender. 80 In her 
book, Male Daughters. Female Husbands , [fi Amadiume notes how Nigerian Igbo daughters could 
become sons (male) and women could become husbands (male). 

While the Igbo institution of daughters becoming sons appears similar to the omosimbe of 
the western Serengeti described above, the practice of "female husbands" (often known in the 
literature as "woman/woman marriage") also occurs as a Mara institution which sheds further light 
on local concepts of gender. Amadiume postulates that this separation of gender from sex is 
necessary in a society where women's worlds operate separately from men's. The ability for 

78 Schoenbrun, Etymologies . # 209 and #2 1 0. For a discussion of these terms see, 
Schoenbrun, A Green Place , pp. 185-187. 

79 Sarah LeVine and Robert LeVine, Mothers and Wives: Gusii Women of East Africa 
(Chicago: University of Chicago Press, 1979), pp. 12-14 and Conclusion. 

women to take on men's gendered roles and authority in certain situations mediates these gendered 
barriers. 81 Regina Smith Oboler's study of "woman/woman marriage" shows that theNandi, in 
western Kenya, classify a childless woman who pays bridewealth for another woman to bear her 
children as male, in specified socio-cultural domains. She becomes the social and legal father of 
the children and plays a male role in all negotiations over the property rights involving the children. 
In other areas of life not involving property, people continue to recognize her in a female role. 82 

All languages in the Mara Region use the term ukamwana (the institution of "daughter-in- 
law") to designate "woman/woman marriage." Here they conceptualize the "female husband" as 
the "mother-in-law" who pays bridewealth for the marriage of her dead or fictional son. Because 
of the functioning of the "house-property" system, sons inherit through the "house" of their mother. 
The "mother-in-law" or "female husband" has control over the dispersal of the property from her 
"house," although her husband has overall control of the joint property of the homestead. Thus 
"woman/woman marriage" constituted one way of assuring that 'the house would not die.' People 
said that the bride in a ukamwana arrangement 'married the house.' 83 

The "female husband" represents another category of extraordinary women who occupy a 
liminal position, mediating between male and female roles in different domains of life. Recent 

81 Amadiume, Male Daughters. Female Husbands , p. 89. 

82 Regina Smith Oboler, "Is the Female Husband a Man? Woman/Woman Marriage 
among the Nandi of Kenya," Ethnology 19(1) (January 1 980): 69-88. 

83 Hugo Huber, "'Woman-Marriage' in Some East African Societies," Anthropos 63/64 
(1968/69): 745-752. Fieldwork for this article was done among the Simbete, Iregi and Kenye of 
the northwest portion of the Mara Region. With little variation the practices are similar to what I 
found operating in the Serengeti district in the southeast of the region. See also Denise O'Brien, 
"Female Husbands in Southern Bantu Societies,"in Sexual Stratification: A Cross-Cultural View 
ed. Alice Schlegel (New York: Columbia University Press, 1977), pp. 109-126. Eileen Jensen 
Krige, "Woman-Marriage, with Special Reference to the Lovedu — Its Significance for the 
Definition of Marriage." Africa 44. 1 (1974): 11-36. 

literature on ethnicity and social identity in Africa has shown that ethnic, clan, lineage, or age-set 
identities are situational and relational rather than immutable biological fact. 84 In the same way, 
gender identity in these societies is variable in particular contexts and for particular people. 8 ' The 
structured mediation of gender boundaries allows certain women access to men's knowledge. 
While I found little evidence that these women shared this knowledge with other women in their 
female roles, they did seem to bring their female knowledge to bear on men's discussions of the 
past. They failed to share this knowledge with women, in large part, because the spaces that 
women occupied and the forms of narrative in which they engaged, did not present the opportunity 
for telling ethnic histories among women. These extraordinary women thus acted as the tributaries 
that flowed in one direction between these streams of knowledge. 

Secret Knowledge 

Besides the public historical narratives told by men in their gatherings over beer, the 
family histories passed on to children and daughters-in-law, the place-specific knowledge given to 
young men on hunting and raiding trips, and the bits of historical knowledge shared at funerals or 
in the preparation for weddings, a realm of crucial historical knowledge remains that people cannot 
share publicly. This knowledge, specific to the functioning of particular social groups, must be 
kept within those channels or bring risks to the entire group. 

For example, in the system of eldership titles, initiates learn masubho or "medicines" 
during the ceremony to join the rank. 86 They must not tell these secrets to anyone outside those 

84 For Africa for example see: Vail, The Creation of Tribalism : David Newbury, Kings 
and Clans: l iwi Island and the Lake Kivu Rift. 1780-1840 (Madison: University of Wisconsin 
Press, 1991); and Spear and Waller, Being Maasai . 

85 Oboler, "Is the Female Husband a Man?," pp. 69-88. 

86 See Chapter 10 for a discussion of the nyangi system of eldership ranks. One of the top 
ranking elders in Nata decided that the heart of Nata was contained in those secrets and that unless 

initiated on threat of death, through an oath. Women go through initiation separately from men 
and learn different masubho. Although I did not learn these secrets, elders told me that they 
include a recitation of historical information. Whether in the form of historical chronicles or not, 
the secrets of these ranks represent important sources for understanding the historical development 
of eldership titles. 87 

As young people leave the villages for school and jobs in the city fewer pursue initiation 
into these titles. Young people also complain that the feasts required for taking a rank cost too 
much. The highest ranking elders face a serious crisis in which they cannot pass on their 
knowledge to a new generation because no has the means to be initiated. This information may die 
with the present generation of elders. Concern about this has resulted in attempts to lower the 
feasting requirements to encourage young people to join. No movement has arisen to write down 
these secrets to preserve this information for future generations. 

Elders cannot write down the secrets of the past or they would lose their efficacy. The 
written histories that educated elders have produced do not contain this kind of information. Elders 
keep these "books" as personal property, rather than get them published or make them available for 
public consumption. 88 The books most often contain fairly superficial and basic narratives, 
including almost none of the crucial historical information concerning specific social institutions 

1 learned those 1 would not understand Nata. He wanted to initiate me, which would involve me 
giving a feast for the elders. We had already planned a feast at which time the Nata elders would 
tell Nata history together to preserve on videotape. At that meeting his ranking counterparts 
decided that 1 could not be initiated because I would have to swear an oath never to leave Nata. I 
suspect that a reason for initiating me would have been the fear that I already knew too many 
secrets and that the oath, and my lack of mobility, would keep me silent. 

87 Interview Gabuso Shoka, Mbiso, 30 April 1995 (Nata <?) on the nyangi system. 

88 The elders in the village where I lived requested help in building a museum for the 
preservation of history at my farewell feast. Yet it seemed that most were more interested in the 
material infrastructure and access to tourist dollars than in preserving their own past. 

such as eldership titles. I pursued a highly recommended Ishenyi written history for months before 
those involved finally gave me access. It turned out to be five short pages of elementary Ishenyi 
historical knowledge. Clearly, having the "book" itself was more important than what it actually 
said. The content of these written histories usually consisted of a predictable and pleasing "tribal" 
narrative, calculated to secure status among other "tribes" rather than to pass on the secrets of 

One can only interpret this situation within a wider understanding of local concepts of 
knowledge. Much of what I observed in the western Serengeti corresponds to Guyer and Belinga's 
descriptions of Equatorial Africa where knowledge is "particularly highly valued and complexly 
organized." They noted that, "knowledge was a primary resource that was elaborated, 
differentiated and cultivated far beyond levels that can be explained by the mundane adaptive need 
to exploit land, labor, capital or any other material or social resource." 8 ' Knowledge was multiple, 
diverse and ever expanding, not contained within one coherent body of truth. Individual specialists 
controlled their own personalistic knowledge acquired by birth, sale, capture or initiation. 

A person's knowledge can be embodied in medicines that take on the identity of the person, 
as well as a power of their own. People acquire these medicines of personal knowledge at a cost. 
In the western Serengeti, after someone joins an eldership rank, he or she may add to his or her 
secret knowledge by going to a more experienced member and giving "tobacco" in return for 
knowledge. In the same way, elders often asked me for "tobacco" in return for an interview." 1 

89 Guyer and Belinga, "Wealth in People," pp. 93, 117. 

" I had no clear policy on offering gifts or money to informants. I followed the lead of my 
assistant. Sometimes when we left I gave the man the equivalent of about 1 US dollar for 
"tobacco." Cues on whether to give or not from my assistants seemed to depend on a number of 
factors-how well they knew the man, what their relationship to him was, whether he was poor, 
what his position was, how long we had stayed, how helpful he had been, or whether he asked for 

Guyer and Belinga report that knowledge or medicines that were not "purchased" lost their 

Western Serengeti people also commonly condense powerful knowledge about the past in 
objects such as medicine bundles that they pass on to the next generation. The lineage of the 
person who made the medicine bundle chooses an individual whom they entrust with the secret of 
its composition. The entire lineage faces severe consequences if they neglect the medicine bundle. 
Guyer and Belinga note the equivalence of people and things as repositories of knowledge as an 
important aspect of Equatorial society. 92 Similarly, in the western Serengeti, the ancestors guard 
and ensure continuity of the most important knowledge about the past, whether embodied in 
medicine bundles or at the sites of their graves, through which the living communicate with the 

Local history books, analogous to medicine bundles, represent knowledge embodied in 
things. Elders passed on and preserved the books within the lineage like medicine bundles. People 
assumed that these history books contained more knowledge than what appeared in writing. Elders 
often asked me to pay for the privilege of reading locally produced manuscripts. Although many of 
these groups of elders wanted to have their books published, they did little to pursue this objective. 
If they published the books, the authors would lose personal control over these "medicines." The 
mass production of medicine bundles or books of knowledge is antithetical to the concept of the 
singularity of knowledge situated within particular social channels. 

Written knowledge bypasses the necessity for a personal relationship and reciprocal 
exchange. In village primary schools today teachers complain about the lack of textbooks, but 

1 Guyer and Belinga, "Wealth in People," pp. 111-1 12. 
! Ibid. 


many prefer to be the only one with a book to dispense the knowledge at their discretion. Some say 
that storerooms are full of unused textbooks. Secrecy and orality render knowledge a scarce 
commodity that is accessible only through personalized reciprocity. Secrecy is a general principle 
used in many social interactions to gain power in that relationship. The concept of knowledge as a 
freely accessible resource open to all is foreign and threatening to this system of exchange. 

Even if groups could control written secrets, this knowledge might then lose its most 
important attribute, its flexibility. 93 In the case of the eldership titles, writing this knowledge would 
ossify ritual and prevent its elaboration and development by individual adepts. It would cease to be 
a living tradition. 94 Written knowledge does away with the need for elders to imagine and shape 
the past as they discuss the case at hand over beer. It destroys the context in which oral tradition is 
preserved. Written histories demand conformity to one "true" history rather than a host of 
different kinds of histories told differently in different contexts, especially here, where publication 
is too expensive and rare for democratic use. 

Another example of secret knowledge is the rituals of the "ruling" generation-set for 
healing and protecting, or "cooling" the land and its people. Every eight years the generation-set 
distributed medicine around the community's boundaries in a way prescribed by a prophet. 
Knowledge of the generation-set secrets by outsiders would render the group vulnerable to enemy 
attack, disease and witchcraft. Elders were reluctant to divulge details about the work of the 

93 This is a classic anthropological insight, see for example, Laura Bohannan, "A 
Geneological Charter," Africa 22, 4 (October 1 952): 30 1 -3 1 5, discusses the Tiv genealogies as a 
validation of present relationships, the genealogies must change over time and in different 
situations to be consistent with present social relationships. She argues that writing down these 
genealogies would make them rigid and this incompatible with their usefulness as social charter. 

"This was the case in the 1993 coronation of the Kabaka Ronald Mutebi II in Buganda 
where some of those officiating referred to the missionary/anthropologist John Roscoe's account of 
the ceremony. John Roscoe, The Baganda: An Account of their Native Customs and Beliefs. 2nd 
ed. (London: Cass, 1965). 

generation-set in protecting the land and often avoided my questions or generalized their answers. 
In Nata, elders told me repeatedly that the ruling generation-set's only task was to walk together 
across the land every eight years and feast at different people's houses-like a solidarity walk." 
Just understanding how this ritual functioned in each ethnic group, apart from the particular 
medicines, required patient piecing together of this cryptic information. Although generation-sets 
figure prominently in colonial ethnographies of the region, I have found in them no discussion of 
the generation-set function in protecting the land." 

Elders consider it crucially important that the knowledge of medicines to protect the land 
and the people not fall into the hands of those with malevolent intent toward the community. 
Those who possess knowledge give it to those who deserve to know. If anyone can gain knowledge 
in a book, the elders cannot regulate the integrity of those who have access to this knowledge. In 
my own research, much of the talk and visiting before interviews involved an appraisal by the 
elders of whether 1 was worthy to deal with the knowledge they held. They considered some kinds 
of information better lost than to be made available to those who would misuse it. Greet 
Kershaw's study of Kikuyu oral traditions about land demonstrates that the elders intentionally 
allow some kinds of knowledge to be forgotten. Only the most senior Kikuyu elders know about 
events of the past which brought evil. They decide whether this knowledge should be passed on or 
not. 96 

People understand the giving and receiving of historical information not as a quaint 
antiquarian hobby but rather as a tremendous source of power. This was made abundantly clear to 
me during my research. The people in Robanda, only a few kilometers from the Serengeti National 

" Malcolm Ruel, "Kuria Generation Classes," Africa 32 (1962): 14-36, Bischofberger, 
The Generation Classes . 

96 Greet Kershaw, Mau Mau From Below (Athens: Ohio University Press, 1997), p. 16. 

Park, Ikoma Gate, feared to tell me about their origins in Sonjo, because general knowledge of this 
might lessen their claim to land that the Park wanted for expansion and the Park might send them 
back to Sonjo." In Ikizu a noticeable rift exists between two factions of the ethnic group, each 
with its roots in the stories of origin. The Ikizu fight out this conflict in the political and economic 
arena, with each side appealing to history for its legitimacy. Both sides insisted that I write the 
"true" story of Ikizu, the "full" story of Ikizu, not the lies of the other side. In each interview a 
different set of political issues subtly intruded as elders spoke about the past. 

Knowledge vested in places such as springs, hills or rocks, called emisambwa, comprises 
another kind of knowledge about the past outside the recognized ethnic narratives. These sacred 
places for ritual and sacrifice, appear most often in connection with the grave of an ancestor. The 
stories of these ancestors as great prophets or rainmakers provide important information about past 
events as well as social organization. The landscape can be "read" like a "book" by walking over it 
and hearing the stories connected with each site. Only those in the lineage of the ancestor buried 
there have access to the secrets of the emisambwa, the rituals used to propitiate the ancestral 

The ancestors themselves, whose spirits reside at those places, preserve the historical 
knowledge embodied in these sites. Elders often related stories about serious misfortune visited on 
a community when its people forgot the grave sites of ancestors, the rituals of propitiation, or the 
meaning of past events. The elders of the community would then consult a dream-prophet 
(omugabho) who got in touch with the ancestor involved and told the people what was needed to 
make things right again. A few generations back, a prophet told the Nata lineage (who went to ask 
for rain at the grave-site of Gitaraga) that Gitaraga wanted them to make an offering first at the 

" They had good reasons for these fears from past interactions with the Park and it did not 
help that my husband was doing his research with Serengeti National Park at the same time. 

grave of his wife before they visited him. Through this experience, elders learned the story of 
Nyaheri and her grave site. 98 People gain crucial historical knowledge through the dead, even if the 
living forget. 

Elders continue to keep a certain kind of knowledge secret precisely because they highly 
value it and treat it with caution, either to maintain its efficacy or to prevent disaster. Knowledge 
about the past cannot be divorced from its connection to relationships that have ongoing meaning. 
People remember and pass on stories about the past because this knowledge is part of the social 
landscape of particular groups of people. Secret knowledge, like gendered knowledge, is 
maintained by the organization of social space. Only titled elders can enter the house where the 
secrets are told to a new initiate taking the rank. Only those in a certain lineage can approach the 
sacred places with ritual sacrifices. Only those of the generation-set in power can walk the 
boundaries of the land to spread the medicines of protection. Knowledge about the past flows in 
these particular channels, kept within its banks by the social organization of space. 


This chapter has demonstrated that different kinds of historical knowledge are confined to 
the social spaces of those who transmit them, to the extent that women and men sharing the same 
household do not share the same stories about the past. Over the past two decades, much of the 
debate among historians of oral tradition has centered on a critique of Jan Vansina's early work, 
where he postulated "chains of transmission" for oral traditions, making it possible for the historian 
to compare separate versions. David W. Cohen, and others, made a case for the flow of historical 
information in all kinds of informal channels, an argument that destroyed any hope of establishing 

98 Interview with Keneti Mahembora, Sangang'a, 17 February, 1996 (Nata d 1 ). 


separate and "uncontaminated" sources." Yet this literature has failed to note that most historical 
knowledge is not randomly shared. 100 This chapter has demonstrated the social spaces of 
knowledge in the case of gendered knowledge and the secret knowledge of particular social groups 
such as the medicines of eldership ranks, the rituals of generation-sets and the sacred sites of 
lineages. This evidence returns the flow of historical tradition, as opposed to more ubiquitous 
stories about the past in popular culture, to particular channels, given the inevitable floods and 
tributaries that establish linkages with other streams. Historical knowledge follows the channels 
and direction established by social practice. 

On the other hand, 1 have also shown that institutionalized means for bridging these 
channels also exist. Women, acting in men's roles as prophets, childless women, independent 
women and post-menopausal women learn men's knowledge and participate in both men's and 
women's storytelling experiences. Whether these women tell men's ethnic narratives or not, they 
must use this knowledge in their telling of women's narrative genres. Everyone knows the basic 
outline of the ethnic origin story, although few have opportunity or the expertise to tell it. Men 
know the fundamentals of family and genealogical history but do not have the personal stake in 
memorizing its details. The same men who take eldership titles are also members of the generation- 
set and of a lineage with responsibility to maintain ancestral sites or medicine bundles. The secrets 

99 This insight thanks to personal communication with Steven Feierman. The debate over 
chains of transmission is summarized by Hamilton, '"Living with Fluidity,'" in papers from the 
International Conference, "Words and Voices." See also David Newbury's paper in the same 
collection. Cohen, Womunafu's Bunafu . p. 8-9. J. Vansina, "Some Perceptions on the Writing of 
African History: 1948-1992." Itinerario (1995): 77-91. Cohen's critique is most fully developed in 
David W. Cohen, The Combing of History (Chicago: Chicago University Press, 1994) and Cohen, 
"The Undefining.". 

100 Even though Cohen argued this in the case of Luo women taking care of young women, 
see Cohen, "Pirn's doorway," pp. 191-235. 

of historical knowledge controlled by each of these social groups influence the secrets of the others 
because they are stored as one man's memories. 

I have explored the "lived circumstances" of men and women in the western Serengeti to 
understand how different realms of historical knowledge develop and are transmitted. People 
transmit and maintain this knowledge by the ways in which they imagine and organize social space 
in daily practice. Similarly, in Purity and Exile Malkki demonstrates how "historical 
consciousness is embedded in and emerges from particular, local, lived circumstances." She traces 
the construction of a Hutu identity and national history in the refugee camps as opposed to the 
cosmopolitan identity and denial of history among town refugees. 101 The lived circumstances of 
different social groups confine particular domains of knowledge to the memories of individual 
members of that group. Yet their knowledge is influenced by other realms of knowledge because 
individuals frequently cross social boundaries and are simultaneously members of various social 

The channeling of historical knowledge in the spaces of social practice does not mean that 
oral historians can return to the easy assumption of discrete "chains of transmission." The banks 
of those streams of knowledge are often breeched and numerous tributaries connect one with 
another. Nevertheless, because knowledge crosses the boundaries of social space in structured, 
rather than random ways, the historian can develop tools for interpreting these various 
representations of the past in relation to each other. 

The understanding of historical narrators and their narrations within particular social 
spaces built in this chapter must be kept in mind when specific sets of oral tradition are explored in 
the remainder of the dissertation. The social organization of knowledge provides key insights for 

101 Liisa Malkki, Purity and Exile: Violence. Memory, and National Cosmology Among 
Hutu Refugees in Tanzania (Chicago: University of Chicago Press, 1995), p. 241. 

the interpretation of oral tradition. Nevertheless, before the oral traditions of origin are presented, 
in Chapter 4, it is necessary to look more specifically at the form of historical consciousness in the 
western Serengeti that divides history into two major periods, before and after the disasters of the 
late nineteenth century. 




The previous chapter demonstrated how men reworked oral traditions into unique "tribal" 
histories, during the colonial period. The lens of the "tribe" filters much of what we know about 
past forms of social identity, while the social spaces of gendered discourse and secret knowledges 
structure the transmission and maintenance of knowledge about the past. 

This chapter argues, however, that events of the period immediately preceding colonial rule 
in this region, 1870-1900, also influenced oral traditions in a fundamental way. The disasters of 
the late nineteenth century precipitated radical social transformations that fragmented continuity 
with the past and altered historical consciousness. Although elders remember these disasters as a 
horrible experience, the transformations that resulted constitute a tribute to the creative and 
indomitable spirit of the peoples of the western Serengeti, who not only survived the disasters but 
found in those experiences new ways to prosper. ' 

History, from the perspective of oral traditions, in some sense begins with the disasters. 
The oral traditions concerning earlier times take the form of either mythical and abbreviated stories 
of origin and migration or cryptic lists of settlement sites and clan names. They provide no 
coherent account of events or personalities of the earlier period. By contrast, elders narrate the 

1 For others accounts of social transformation during period of disaster see, James L. 
Giblin, "Famine, Political Authority and Foreign Capital in Northeastern Tanzania, 1840-1940" 
(Ph.D. Dissertation, University of Wisconsin-Madison, 1991); Ambler, Kenyan Communities , and 
Gregory H. Maddox, "Leave Wagogo, You have no Food: Famine and Survival in Ugogo, 
Tanzania, 1916-1961" (Ph.D. Dissertation, Northwestern University, 1988). 


stories of battles with Maasai, suffering during famines, and the walk to Sukuma with elaborate 

Although a middle period of vague and timeless traditions appears as a structural feature 
of oral narrative throughout Africa, the moment at which traditions become historically grounded 
in verifiable events seems to represent a point of transition in social identity. Among societies with 
centralized states, the traditions of historical time often begin with the consolidation of the kingdom 
under a known king, even if the antiquity of the kingdom is extended by reconfigured genealogies. 2 
In the oral traditions of the Maasai historical time begins with the leadership of prophets in the late 
eighteenth century. 3 In comparison to these examples, historical time in the western Serengeti 
begins relatively late and, significantly, with the events of the disasters. From the perspective of 
oral tradition this period represents a rupture in social time. 4 

In the western Serengeti this era initiated not only a break in time but the introduction of 
new concepts of time altogether. A new way of calculating social time, embodied in the 
succession of cycling age-sets, emerged during this period of disasters. Because the age-set names 
cycled after every three generations, this way of calculating time could be, anachronistically, 

2 For kingdoms in Eastern African see Feierman, The Shambaa Kingdom : Newbury, 
Kings and Clans: Randall Packard, Chiefship and Cosmology: An Historical Study of Political 
Competition (Bloomington: Indiana University Press, 1 981 ); Wrigley, Kingship and State . 

3 John Lawrence Berntsen, "Pastoralism, Raiding and Prophets: Maasailand in the 
Nineteenth Century" (Ph.D. Dissertation, University of Wisconsin-Madison, 1979), p. 112. 

Rosaldo describes a similar phenomenon in his study of Illongot society in the 
Philippines. He found that history was divided into two major periods-before and after the 
Japanese invasion of 1 945. Rosaldo wrote: "The stories of 1 945 were so numerous, so vivid, so 
detailed, so often told that it took me over a year to realize that they represented but a narrow strip 
in time." This amplified moment in the historical imagination became "the great divide that 
separated a bygone past from one that merged into the present." Rosaldo's interpretation of 
various oral narratives had to take into consideration that this brief period had been generalized to 
represent the whole period before 1945. Rosaldo, Ilongot Headhunting , pp. 38-54. 

projected back to represent the passage of time before the disasters. With this device people could 
maintain continuity with the past in extremely uncertain times. The process of reformulating age- 
sets is discussed in Chapter 9. Here, 1 use these transformations in concepts of time as evidence 
for discontinuity in social time and the radical reformulation of identity during the period of 

The rupture in historical consciousness indicates that fundamental changes in social 
identity took place during the period of disasters. As has already been shown, particular social 
groups maintain and transmit knowledge that has no independent existence outside that group. 
People cannot integrally maintain and transmit historical knowledge if the social group in which 
that knowledge is based radically reconfigures itself. The emerging stories of new groups, which 
take their heritage from a variety of former identities, incorporate bits of knowledge from older 

Scholars have made a similar case to explain why Maasai oral tradition begins at the end 
of the eighteenth century, when it is clear from other evidence that Maa-speakers were present in 
the Rift Valley beginning around 1600. Lamprey and Waller propose that, "traditions cannot 
survive the communities that produce them and for which they have meaning." They suggest that 
if the identity and composition of a community changes dramatically during a period of stress, "it is 
unlikely that the emergent community will assimilate the old corpus of tradition, except in 
fragmented form, since these traditions again refer to events and processes in which the community 
now has no collective part." s 

' Richard Lamprey and Richard Waller, "The Loita-Mara Region in Historical Times: 
Patterns of Subsistence, Settlement and Ecological Change," in Early Pastoralists of South-western 
Kenya, ed. Peter Robertshaw (Nairobi: British Institute in East Africa, 1990), p. 19. 


For the historian it becomes problematic to interpret the meaning of these dramatic social 
transformations without access to evidence that is not itself a product of these changes. The way in 
which western Serengeti people reconfigured social identity during the period of disasters is the 
product of much longer-term social processes. Yet knowledge of times before the disasters is only 
accessible to the historian through oral traditions, which the new communities that survived the 
disasters and established new identities to deal with the changing situation have fragmented and 
reworked. The fragments of oral tradition that survived have been taken out of context and given 
new meaning. 

The historian must interpret the traditions of the earlier period with knowledge of how 
people transformed social identities during the disasters. Yet the historian cannot understand these 
transformations without knowledge of social process in the earlier period. To understand the 
historical development of this region from earliest times on, through the interpretation of oral 
traditions, a basic knowledge of the events of the late nineteenth century disasters and the 
subsequent changes in social identity is necessary. The changes of this period provide the context 
in which I interpret oral traditions throughout the dissertation. 

The experiences of ecological breakdown, indirect contact with the Swahili caravan trade 
and Maasai raids each caused specific social transformations that were in turn reflected in oral 
traditions about earlier periods. The historical interpretation of pre-disaster traditions depends 
upon an awareness of these changes in the oral traditions. The ecological breakdown, which 
resulted from famine and disease, demanded the reformulation of the networks of security 
previously provided, in part, by the interdependent regional economies of hunters, herders and 
farmers. Western Serengeti farmers did this by developing their own interdependent relations 
between kin living in different ecological niches and practicing slightly different economic 
strategies. Because of this change oral traditions now obscure this earlier reliance on hunters and 

herders, making it extremely difficult to reconstruct the earlier regional system. The ecological 
disasters also required changes in settlement structures. People now understand the new patterns 
as "traditional" and forget what came before. 

The caravan trade, which introduced new diseases and increased trade of wild animal 
products, resulted primarily in ties of dependency to Sukuma and the commercialization of hunting. 
The origin traditions of Ikizu, Nata and Sizaki claim a connection to Sukuma, which we can date to 
this period. People have also accepted forms of eldership titles borrowed from Sukuma during the 
time of disasters and grafted onto the older system as "traditional" ranks. Oral traditions may also 
distort the past significance of hunting, due to the identification of forest products as a source of 
cattle wealth during this period. 

Finally, the experience of intensive Maasai raiding had far reaching effects on historical 
consciousness. Although contact with the Maasai probably began gradually and peacefully around 
mid-nineteenth century, oral tradition tells us that this enmity dates from the beginning of time. 
Although other evidence tells us that the period of most intense raiding took place after the 
intrusion of colonial rule, oral tradition dates the most severe raiding to an unspecified earlier 
period and attributes all of the disasters that later took place to the effects of Maasai raiding. In 
this period, western Serengeti communities drew on Maasai culture to reorder their age-sets. In 
doing so, they revised their concept of time. 

All these changes dating to the period of disasters have fundamentally affected the way 
that people understand their past and what they consider to be "traditional." The historian cannot 
interpret oral traditions of earlier periods without understanding this period. The story of how 
people coped with these disasters and the social transformations that resulted are the subject of the 
last two chapters in this dissertation. This chapter acquaints the reader with the major events and 
material conditions based on oral and written sources from this period. It describes the cycle of 

disease, famine, raids, ecological disaster, and eventual social collapse in which many people 
moved to Sukuma or to other places of refuge. The effects of the caravan trade and the increased 
pressure of Maasai raiding contributed to the cycle of crisis. Oral traditions represent the disasters 
as the turning point between distant and recent history. 

Rupture in Social Identity at the Time of Crisis 

Oral traditions about the period of disasters provide the core images for understanding 
what this rupture meant for historical consciousness. The Ishenyi story of disaster is perhaps the 
most extreme case of rupture in social identity. Its content reveals local interpretations of many 
aspects of the disastrous events of the late nineteenth century. 

The following story comes from the Ishenyi people, who now live between Nata and the 

Ikizu. Nyeberekera, where the story begins, is located on the western edge of the Serengeti 

National Park today [For map see Figure 3-1 : Ishenyi Dispersal from Nyeberekera.]. The Ishenyi 

need special permission from the park to revisit this site for propitiation of ancestral spirits who 

still reside there. The landmark for Nyeberekera is a tall rock outcropping, Bwinamoki, which one 

can see for miles away over the plains and which functioned as a lookout for Maasai raids [ See 

Figure 3-2: The Ishenyi Story of Nyeberekera.]. Nyeberekera refers to a general area of settlement 

and to a pool on the Grumeti River that is an ancient site of ancestral spirits. 6 Mikael Magessa 

Sarota, the son of one colonial Ishenyi chief, told this version of the story: 

Long ago the Ishenyi lived at Nyeberekera, over to the east ofMugumu, inside the park. 
There are hills there, a fertile land that cries buubuubuu . . . when you walk on it. The 
land was called Nyeberekera. This is where we came from. When we left there, we came 
to Nyigoti. The Maasai drove us out in the time before my grandfather. The Maasai 
raided us. We were farmers and they were herders. The Maasai came to steal the few 
cattle that the Ishenyi kept. The Ishenyi had a dream prophet at that time named 

6 Interview with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi d"). He says this is 
where people went to fish and got swallowed up, the clan of Abang'ohe from Bene Okinyonyi, but 
all Ishenyi go there to propitiate the spirit. This is not the site where Shanyangi is propitiated. 


Shanyangi. When the Maasai would come, just enter their land, Shanyangi would make 
biting ants appear which would/ill the river and prevent them from crossing. Then he 
would make bees that would swarm all over and drive them back. They would be unable 
to raid the lshenyi. Then one day the Maasai sal together to decide how to defeat this 
Ishenyi prophet who sent ants and bees against them. They went to their own prophet 
who could stop the rain from coming. This prophet stopped the rain from falling on 
Nyeberekera. At this time the Ishenyi were farming with wooden hoes because the soil 
was so fertile and loose. They had ample food and there was no hunger. This Maasai 
prophet stopped the rain from falling for eight years. All of the food stores were 
exhausted. When this happened, the Ishenyi went to their prophet Shanyangi and asked 
him to send rain. He said that he was not a rainmaker and only knew the medicine of 
war against the Maasai. Nevertheless, they would not listen and sat in their meetings 
over and over again, asking him to make rain. Finally they decided that Shanyangi 
must be lying to them — how could he be such a powerful prophet and not know how to 
make rain? Surely it was not true! When the drought continued, they decided to kill 
Shanyangi. They tied him up on a tree and chopped down the tree, which fell and killed 
him. Yet of course once he died there was no one to protect them and they were driven 
out, dispersed here and there by the Maasai raiders. Some came to Nyigoti, near to 
where the Nata were living. Others refused to come to Nata because the Nata were sick 
with kaswende (syphilis). Many Ishenyi warriors impaled themselves on their own 
spears rather than go to Nata and suffer the slow death o/kaswende. Others went to 
Kuria where they became the Iregi of today, those of the clan ofSarega. The Iregi are 
really Ishenyi people. Those who moved near to the Nata settled and began living there. 
Then another famine came and they were forced to go to Sukuma to beg for food. After 
the drought lifted, they returned from Sukuma and came to live where they are today at 
Nagusi. They did not return to Nyigoti. This terrible famine was called the "Hunger of 
the Feet. " It was called that because of the sores people got in their feet as they walked 
to Sukuma in the dust. 7 

All of the elements of the disaster appear in this story-pressure from the Maasai, drought, famine 

and new epidemic diseases. These events result in dispersal and regrouping in a new place with a 

new identity. Famines bracket this story, beginning with the "Hunger that Finished the Cattle" and 

ending with the "Hunger of the Feet." 

Some hint of former identities is evident in this story. Although elders tell this account as 

an "Ishenyi" story, many elders agreed that those who lived at Nyeberekera called themselves 

"Regata" and those who went to Kuria at the dispersal were known as "Iregi." In Sonjo today the 

inhabitants of a village named "Regata" tell the same story of dispersal. The Maasai Loitai of the 

7 Interview with Mikael Magessa Sarota, Issenye, 25 August 1995 (Ishenyi <?). 


Ishenyi Dispersal from Nyeberekera 


The Generation of Disasters 


9 Ikoma 

O Ishenyi 

• Ngoreme 

Seregeti N.P. 

Ikorongo • !tununu 

to Kcjria 
Iregi , 



I -QSite 




Bwa^° R S -? a 





r\ •Mangwes'i 

O Sanganga 






Peter Shetler 

Dove Creek Information Services 

IDRISI, Adobe Photoshop and Macromedia Freehand 


30 km 

Figure 3-1 : Ishenyi Dispersal from Nyeberekera 

Bwinamoki Rocks at Nyeberekera, 16 February 1996 

Mikael Magessa Sarota, Issenye, 25 August 
1995, narrator of the Nyeberekera Story. 

Figure 3-2: The Ishenyi Story of Nyeberekera 

Loliondo highlands, whose territory extended north into Kenya and east to Lake Natron, reported 
to colonial anthropologist Henry Fosbrooke that the original inhabitants of this area were the 
"Ilmarau," the Maasai name for the Ikoma or Bantu-speaking peoples of the western Serengeti in 
general. 8 Collectively these narratives point to a previous community that straddled what is now 
the Serengeti National Park, dispersed and divided by this series of disasters.' This older 
community may have called themselves the Regata or in Ngoreme traditions they are referred to as 
Masabha or "the people of the north." 10 

During this period of stress, people determined what it meant to be "Ishenyi." They 
connected their identity as a people to their relationship with the land. Therefore, moving 
demanded a reformulation of identity. The murder of the Ishenyi prophet Shanyangi by his own 
people symbolizes this change of identity. Although some Ishenyi now return to Nyeberekera to 
propitiate his spirit, this is an innovation of recent years. Why would Shanyangi choose to help 
those who killed him? Shanyangi's murder was the final break with Nyeberekera and their identity 
as a people under his leadership. The Ikoma also have a story in which a prophet forbade them to 
return to Sonjo, forcing them out to new lands and the formation of new communities." 

8 Henry A. Fosbrooke, "Sections of the Masai in Loliondo Area," 1 953, (typescript) CORY 
#259, EAF, UDSM. 

9 This community was perhaps linked to the deserted irrigation villages of Engaruka. J. E. 
G. Sutton, "Becoming Maasailand," in Being Maasai: Ethnicity and Identity in East Africa , eds. 
Thomas Spear and Richard Waller (Athens: Ohio University Press, 1993), p. 54; and J. E. G. 
Sutton, "Engaruka etc...," Tanzania Zamani 10 (January 1972): 7-10; Louis Leakey, "Preliminary 
Report on Examination of the Engaruka Ruins," Tanganyika Notes and Records 1 (March 1936V 

10 P. Haimati and P. Houle, "Mila na Matendo ya Wangoreme," unpublished mimeo, 
Iramba Mission, 1969. 

" Interview with Machota Sabuni, Issenye, 14 March 1996 (Ikoma if). 


So powerful was this period of disaster for identity formation that the core images of these 
stories are of new beginning and decisive break with the past. The present generation of Ishenyi 
elders seems to have forgotten the entire origin story except the names of the ancestral parents and 
the place where they lived. 12 Mikael's description of Nyeberekera as the place "where we came 
from" reconfigures it as an origin story. The versions of this story that emphasize the pre-dispersal 
paradise of Nyeberekera tell that the people produced so much milk there that the Ishenyi dumped it 
into the river and so attracted the Maasai. The consequences for the Ishenyi were not only 
dispersal and famine but the curse of the prophet Shanyangi that denied them another Ishenyi 
prophet. 13 The Ikoma also experienced a prophetic curse during this time that kept the hunters 
from returning to Sonjo. 14 Elders point to some rocks near Ikoma Robanda where witches were 
pushed off to their death during this generation. That elders would remember this period as a time 
of witchcraft is one indication of extreme societal stress." 

Dating the Nyeberekera dispersal by oral accounts is difficult because of the present value 
placed on histories that are very old. Elders tend to exaggerate the time depth to make ancient 
territorial claims. Many elders, however, agree that the prophet Shanyangi belonged to the 
generation of Maina (C. 1 840) and that the Ishenyi left Nyeberekera during the time of the Saai 
and Chuma (C. 1 860-1880). This matches Mikael's assertion that the dispersal from Nyeberekera 

12 Guka, on the eastern side of what is now the northern extension of the Park. Mugunyi 
and Iyancha are the ancestral parents. 

13 Anacleti, "Pastoralism and Development," pp. 182-184. 

14 Interview with Machota Sabuni, Issenye, 14 March 1996 (Ikoma <f). 

15 Interview with Moremi Mwikicho, Sagochi Nyekipegete and Kenyatta Mosoka, 
Robanda, 12 July 1995 (Ikoma cc). See Hartwig's analysis of this period on the Lake Victoria 
Island of Kerewe and the witchcraft accusations as a result of social stress, Hartwig, The Art of 
Survival , p. 121. 

happened in the time before his grandfather. The Ishenyi lived at Nyigoti during the generation of 
the Saai and the Nyambureti (C. 1 870-90). Ishenyi elders further agree that the "Hunger of the 
Feet" and the eventual move to Nagusi took place during the Kihocha age-set, just before the turn 
of the century. 16 Traditions of other ethnic groups in this area relate similar tales of disaster during 
these age-sets. 

This correlation of generation and age-sets with events of the time further shows a break in 
social time with the reformulation of identity. Note that in the above testimonies the narrators 
marked earlier events by association with generation-sets, and the later events by association with 
age-sets. One Nata elder said that they formed the first saiga or age-sets during the generation of 
the Maina, living at Site, where they divided that generation into the three age-set cycles of 
Bongirate, Busaai and Borumarancha. 17 An Ishenyi elder confirmed that they divided the cycling 
age-sets or saiga when they left Nyeberekera, during the generation of the Maina. 18 These cycling 
age-sets replaced the generation-set as the way of reckoning time. 

Age-sets were not new to the western Serengeti. What was new was the switch from a 
system in which linear age-sets functioned parallel to the more dominant generation-sets, to a 
system in which territorially-based cycling age-sets were incorporated into the patterns of 
generational succession. These complex transformations are explored in more detail in Chapter 9. 
Here it is only important to note how elders reconciled these two systems in historical memory. 
Since each of the cycling age-sets would ideally rule for eight years, elders often produced age-set 

16 Interviews with Rugayonga Nyamohega, Mugeta, 27 October 1995; Mashauri Ng'ana, 
Issenye, 2 November 1995; Morigo Mchombocho Nyarobi, Issenye, 28 September 1995 (Ishenyi 

17 Interview with Kirigiti Ng'orita, Mbiso, 8 June 1 995 (Nata cf), Kirigiti is the last 
surviving Nata generation-set leader of his section. 

18 Interview with Morigo Mchombocho Nyarobi, Issenye, 28 September 1995 (Ishenyi <?). 

lists with cycling names that extended back two hundred years." Although the memory of the 
change to cycling age-sets still exists, the projection of cycling age-set time onto earlier history is 
increasingly erasing that memory. 

For the historian, this makes dating by age-set chronology particularly difficult. 
Historians of East Africa have long recognized that age-sets provide the possibility for constructing 
relative chronologies that correspond to calendar dates. 20 Scholars have rightly criticized these 
methods for presuming to establish precise dates by projecting average age-set intervals back in 
time and assuming that the age-set rituals that promoted the new set would remain constant through 
time. 21 For the western Serengeti, the introduction of a whole new system of dating past events in 
the last quarter of the nineteenth century destroys the historian's hope of establishing accurate 
dating before this point, except through the relative succession of named generations. Clearly, we 
should read these age-set lists primarily as ideological statements rather than chronological records. 
They establish a continuity with the past and the orderly succession of time through the cycling 
death and rebirth of the generations. 

" Anacleti reproduced this pattern back to the beginning of the eighteenth century, 
sometimes using age-set praise names which referred specifically to the events of the 19th century 
as cycling names, for example, abaSanduka, those who carried boxes as porters. Anacleti, 
"Pastoralism and Development," pp. 14-15. 

20 See Alan H. Jacobs, "A Chronology of the Pastoral Maasai," in Hadith 1 : Proceedings 
of the Annual Conference of the Historical Association of Kenya , ed. Bethwell Ogot (Nairobi: East 
African Publishing House, 1968), pp. 10-31; H. A. Fosbrooke, "An Administrative Survey of the 
Maasai Social System," Tanganyika Notes and Records 26 (December 1948): 1 1 ; and H. A. 
Fosbrooke, "The Maasai Age-group System as a Guide to Tribal Chronology," African Studies 15 
4(1956): 194-195. 

21 SeeBerntsen, "Pastoralism, Raiding and Prophets," pp. 83-93. 

Ecological Breakdown 

By all accounts "the generation of disasters," across the western Serengeti, experienced a 
series of devastating epidemics combined with drought and subsequent famine that eventually led 
to ecological collapse. 22 One Ishenyi elder said that they called the famine at Nyeberekera the 
"Hunger that Finished the Cattle" in which the Sonjo or Regata peoples also left Nyeberekera and 
went to their present homes in Sonjo. 23 This may have been the rinderpest panzootic of 1 880- 1 890 
that killed up to 90% of the cattle in East Africa. After this, pastoralist peoples such as the Maasai 
were reduced to "walking skeletons," as the German traveler, Baumann, described the starving 
Serenget Maasai who had taken refuge in Ngorongoro Crater highlands in 1892. He also found 
Maasai sick and dying on the Lake Victoria coast. 24 

More likely the Nyeberekera story refers to a series of cattle diseases before the rinderpest. 
Ikoma and Nata elders also refer to the "Hunger that Finished the Cattle" (rimara n'gombe) as 
rihaha (rinderpest), while others confirm them to be different diseases at different times. 25 Other 

22 For an account of the environmental disasters in Tanzania see: lliffe, A Modern History . 
Chapter 5; Helge Kjekshus, Ecology Control and Economic Development in East Africa History. 
The Case of Tanganyika 1850-1950 (Athens: Ohio University Press, 1 977); and James Leonard 
Giblin, The Politics of Environmental Control in Northeastern Tanzania. 1 840- 1 940 (Philadelphia, 
University of Pennsylvania Press, 1992). For a critique of the "degradation narrative" see James 
McCann, introduction, An Environmental History of Africa. 1800-1996 , forthcoming; and Melissa 
Leach and Robin Mearns, The Lie of the Land: Challenging Received Wisdom on the African 
Environment (Portsmouth: Heinemann, 1 996). 

23 Interview with Mang'ombe Morimi, Issenye Iharara, 26 August 1995 (Ishenyi <?). 

24 The section on the Maasai at the lake in Baumann, Durch Massailand . pp. 44-46; the 
section on Ngorongoro crater is available in English translation, H. A. Fosbrooke, Ngorongoro's 
First Visitor, by Dr. O. Baumann's Journal of 1892 . trans. G. E. Organ (Dar es Salaam: East 
African Literature Bureau, 1963), pp. 12-14. 

25 Interview with Tirani Wankunyi, Issenye, 7 July 1995 (Nata o"). Kjerland cites evidence 
from the Kenya District Books that rihaha was a different cattle disease which preceded rinderpest 
by a considerable period. Kjerland, "Cattle Breed," pp. 134-5. Among my informants rihaha was 
also used to describe the cattle diseases of the colonial period including rinderpest and East Coast 

sources in East Africa provide evidence for an epidemic of cattle lung disease (C. 1880) which 
swept through before the rinderpest. 26 The Maasai in the Nyeberekera story are clearly ascendant 
and not suffering from the rinderpest or cattle lung disease. Because age-set chronologies also date 
the Nyeberekera story to the years between 1 860 and 1 880, this famine was most likely a localized 
hunger (C. 1 870), occurring before the rinderpest panzootic. 

The Nyeberekera famine illustrates the interdependence of pastoralists, farmers and 
hunter/gatherers in a regional economy that eventually broke down because of the disasters. 27 
Drought and disease affected each community differently because of their specialized micro- 
ecologies and economic strategies. The Ishenyi describe themselves as farmers and thus often in 
the position of giving out food when pastoralists or hunter/gatherers suffered, as the description of 
bountiful food at Nyeberekera attests. The Ishenyi avoided a localized drought, as severe as it 
was, by moving a day's walk away, to Nata for example, to find food. Maasai and Tatoga herders 
settled near the western Serengeti farmers after the rinderpest panzootic and traded their children or 
worked as herders for food. Baumann went through Ikoma in 1 892, just before the "Hunger of the 
Feet," and described the surplus of grain brought for trade by the "peaceful inhabitants," enough 

fever, incidence of these diseases confirmed in a report by District Veterinary Officer, Musoma, to 
the D.O. Musoma, Annual Report 1927, January 1928, Mwanza Province 1927-28, Provincial 
Administration, Annual Report, 246 P.C./1/30, TNA. Cattle lung disease said to have been 
introduced only after 1916 by Maasai stock crossing the Kenya border. F.G. (?) Mc Call, Chief 
Veterinary Officer, Annual Report of the Department of Veterinary Science and Animal 
Husbandry, Tanganyika, 1921, Veterinary Department Annual Reports, 1921, 3046/22, TNA. 

26 Richard Waller, "The Maasai and the British, 1 895- 1 905 : The Origins of an Alliance," 
Journal of African History 17 (1976): 530-32; Berntsen, "Pastoralism, Raiding and Prophets," pp. 
276-79, 83; and A. H. Jacobs, "The Traditional Political Organization of the Pastoral Maasai" 
(Ph.D. dissertation, Oxford University, 1965), pp. 95-99; show that the rinderpest followed nearly 
three decades of livestock diseases and just previous to it an epidemic of livestock pleuro- 

27 For a description of this regional economy see: Spear and Waller, Being Maasai . p. 2. 
Waller, "Ecology, Migration and Expansion," pp. 347-370. 


"to pass through Masailand again if we wished." Kollmann, some years later, also reported full 
granaries and fat cattle. 28 This period began with a series of smaller localized famines, which 
could be overcome by dependence on connections in neighboring ecologies/economies, developing 
into a generalized disaster by the 1 890s. 

The "Hunger of the Feet" (agechaya maghoro) which closes this generation, was a new 
kind of famine because of its regional scope, a result of the cumulation of disasters rather than a 
simple lack of rain. Confirmation of the extent of this famine comes from the White Fathers who 
established themselves by 1893 on Kerewe Island. They date the "Great Famine" (presumably the 
"Hunger of the Feet") on the mainland to 1 894. 2 ' During this period many people from the 
mainland came to Kerewe Island in search of food. The Kerewe station reported that a small 
village of Christians had grown up around the mission station, most of whom were former slaves 
and famine victims from Maasai raiding in the interior, particularly Ngoreme. 30 As the famine 
abated these converts returned home and the White Fathers made journeys in 1902 and 1904 to 
maintain contact with them and also to establish the mainland station of Nyegina, with the specific 
purpose of reaching their converts in Ikizu, Ngoreme, Zanaki, Majita and Ruri. 31 By 1919 most of 
the famine victims had gone home, leaving the mission practically deserted. The White Fathers 

28 Baumann, Durch Massailand. pp. 38-42; Kollmann, The Victoria Nvanza . p. 176. 

25 Visitations Book, Nyegina, Mwanza I, 1931-1932, pp. 67-69, White Fathers Regionals' 
House, Nyegezi. 

30 Societe des Missionnaires d'Afrique (Peres-Blancs), "Ukerewe," Chronique Trimestrielle 
de la Societ e de Missionnaires d'Afrique (Peres Blancs) 27me Annee, 1905, p. 133. 

31 Societe des Missionnaires d'Afrique (Peres-Blancs), Rapports Annuels . Sixieme Annee 
(1910-191 1), p. 383; and L. Bourget, Trip Diary, 1904, White Fathers Regionals' House, 

referred to this far-flung group of converts as, "our Christians of the diaspora." 32 [See Figure 3-3: 
White Fathers' Mission, Nyegina, Musoma, Founded 191 1.] 

Ngoreme elders confirm the exodus to Kerewe while Ikoma, Ishenyi and Nata drought 
victims were more likely to go to Sukuma seeking food. All tell tales of being forced to sell their 
children for food to stay alive. Elders remembered their grandparents, who would have been 
children during the famine, telling the stories of lost siblings. Parents found it more advantageous 
to "sell" girls or to dress their boys as girls, with the understanding that the price was an early 
bridewealth payment. 33 The White Fathers also describe refugees selling their children into slavery 
on Kerewe to get food. 34 Gerald Hartwig's reconstruction of Kerewe history shows that from 
1850-1870 the mainland "Ruri" people brought children, probably kidnaped from neighboring 
peoples, to Kerewe to sell for food. Then from about 1 875 on the Kerewe themselves actively 
searched for children to buy for slaves during the famines by taking boats along the lakeshore and 
up the Mara River. 35 Although these children were not heard of again, it seems likely that some of 
them would have made it into the wider slave trade, ending up on the clove plantations of Zanzibar. 

32 Societe des Missionnaires d'Afrique (Peres-Blancs), "Nyegina," Rapports Annuels . No. 
1 1, 1915-16, p. 328-330; and in Rapports Annuels see "Nyegina," No. 13 (1919-20), p. 354; 
"Nyegina," No. 17, 1921-1922, p. 520. For background on the White Fathers Mission see, J. 
Bouniol. The White Fathers and their Missions (London: Sands and Co., 1929). 

33 Interview with Mashauri Ng'ana, lssenye, 2 November 1995 (Ishenyi <?). 

34 Visitations Book, Nyegina, Mwanza I, 1931-1932, White Fathers Regionals' House, 
Nyegezi, pp. 67-69. 

35 Kjerland, "Cattle Breed," p. 135, cites the Mwanza District Books, and Hartwig, The 
Art of Survival, p. 127-128. Hartwig also states that there were a lot of Luo, "Gaya" slaves on 
Kerewe, p. 125-6, for an assessement of Ukerewe slavery, 1 14-128; Confirmed by Kuria 
informants in, Tobisson, Family Dynamics, pp. 12-13. 


Ruins of Original White Fathers' Mission Chapel, Nyegina 

White Fathers' Mission House, Nyegina 

Figure 3-3: White Fathers' Mission, Nyegina, Musoma, Founded 1911 


Most would have been incorporated into Sukuma families as sons and daughters, and into Kerewe 

families as slaves. 

Elders described the stay in Sukuma lasting for a number of years, during which time they 
found Sukuma patrons who provided support and protection. Western Serengeti people established 
these ties of friendship before the drought through a trade of wildebeest tails, wild animal skins and 
arrow poison brought to Sukuma in exchange for tobacco, salt, iron hoes and livestock. 36 Sukuma 
hosts provided the refugees with a plot of land to farm and a place to build their house in return for 
clearing the land and labor on the host's farm. 3 ' When the drought was over western Serengeti 
peoples, began moving back, settling for a growing season or two at a time along the way. One of 
the best remembered of these way-posts was the settlement of Hantachega, now in the western 
corridor of the Serengeti National Park. There, Nata, Ishenyi, Ikoma and Sukuma all built 
together, according to age-set organization. 

The Ishenyi tradition told above mentions kaswende or syphilis in Nata. That young men 
would consider suicide preferable to exposure means that they had some knowledge of the disease. 
Scholars have assumed that the nineteenth century caravan trade introduced sexually transmitted 
diseases (S.T.D.s) such as syphilis. 38 Nata people must have been in at least indirect contact with 

36 Interview with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi tf). 

37 See description of a similar process during a famine in central Kenya, Ambler, Kenya 
Communities , pp. 134-135. 

38 The epidemiology on syphilis is not well understood. By the 1870s syphilis is assumed 
to be rapidly rising along trade routes. The problem is that this may also have been yaws, which 
appears with similar symptoms. The vast majority of childhood complaints were yaws, not 
syphilis while, lesions developing in adults after the turn of the century were probably syphilis. 
Yet with the 800 years or more of precolonial contact on the coast with Arabs and 300 years of 
contact with Portuguese, it is difficult to say when and where it was introduced. It is improbable 
that yaws mutated into syphilis so we can assume that it was introduced. Personal communication 
with Anne Stacie Canning Colwell, M.D., 5 February 1998. 

the caravan trade through their friendships in Sukuma. Another indication of the prevalence of 
S.T.D.s at this time is an Ishenyi and Ikoma story about going to a Tatoga prophet in the 
Ngorongoro crater because of general fertility problems. 39 Many elders said that infertility, 
probably related to S.T.D.s, was a significant problem during the early colonial period. In one 
area, intercourse with a barren woman became taboo for fear of the transmission of disease.™ 

Other "foreign" diseases mentioned in narratives about this period may have been small 
pox, cholera and measles, collectively associated with dysentery and dehydration. 41 One Ikoma 
elder described a disease called nyekekundi, where everyone in one homestead would all die 
suddenly, during the age-set of the Rumarancha (C. 1 890). They went to see a Tatoga prophet, 
Gamurayi, for a cure but he was afraid of exposure and would not open the door. Gifts of cattle 
finally induced his wife to open the door. 42 The Ishenyi story mentions the foot sores that gave the 
famine its name, most likely chiggers from the dusty ground on the path to Sukuma. 

When the Germans began to administer this area from their base at Schirati on the lake, 
near the Kenya border, they quickly identified an epidemic of sleeping sickness in 1 902, that had 
killed 2,000 people by 1 905. This was one of the first areas in East Africa where they identified 
sleeping sickness and the site where Robert Koch and F. K. Kleine did much of the pioneering work 
on the disease. They postulated that sleeping sickness had spread from the west lake, ultimately 

39 Interview with Morigo Mchombocho Nyarobi, Issenye, 28 September 1995 (Ishenyi <f). 
Dating of the Tatoga leaving the Crater due to Maasai pressure around mid-century according to 
Maasai age-set chronology. 

40 Interview with Masosota Igonga, Ring'wani, 6 October 1995 (Ngoreme J); Sira 
Masiyora, Nyerero, 17 November 1995 (Kuria J). 

41 Interviews with Mariko Romara Kisigiro, Burunga, 31 March 1995 (Nata cf); Maarimo 
Nyamakena and Katani Magori Nyabunga, Sanzate, 10 June 1995 (Ikizu cf). These diseases are 
known locally as kyamunda in Nata or nyamugwa in Ikizu, also oborondo, egesaho, etc. 

42 Interviews with Machota Sabuni, Issenye, 14 March 1996 (Ikoma <f). 

from West Africa and the Congo, by the canoe trade and then to the eastern mainland. That the 
disease vector, the tsetses fly, was indigenous to the area was clear, but deciding whether sleeping 
sickness was also indigenous proved harder for early researchers. Researchers gathered local 
traditions as evidence that sleeping sickness existed before the colonial period. 43 

More recent analyses of sleeping sickness in East Africa link the outbreak of the disease 
with the environmental catastrophes of the 1 890s, rather than with the introduction of the disease 
by human hosts from the Congo. They propose that sleeping sickness had been endemic in East 
Africa, kept at bay rather than eradicated, by local patterns of bush clearing, cattle grazing and 
farming. 4,1 Colonial officers in this district correlated the spread of sleeping sickness to times of 
drought when men increasingly sought scarce grazing for livestock in the tsetse infected bush 
areas. Though the government resented the "illegal burning" of bush, this was probably one local 
measure that kept the disease in check. 45 Just as the colonial officers feared, men may have spread 
sleeping sickness during this period of disasters because of their increasing mobility across the 
region in search of food, trade or hunting grounds to cope with the drought. 

43 David F. Clyde, History of the Medical Services of Tanganyika (Government Press: 
Dar es Salaam, 1962), pp. 28-29. Clyde cited traditions from Kerewe Island and Ikoma describing 
a disease which resembled the symptoms of sleeping sickness as evidence of sleeping sickness as an 
ancient disease. In Ikizu and Ikoma this disease was said to have almost depopulated the province 
over the last one hundred years. Local informants said that the disease was contracted by the bite 
of the fly, beginning when the Ruwana and Mbalangeti rivers were in flood. There was a great 
deal of confusion as to whether this was sleeping sickness or severe hookworm disease in man 
coincident with animal trypanosomiasis. 

44 Juhani Koponen, Development for Exploitation: German Colonial Policies in Mainland 
Tanzania. 1884-1914 (Finnish Historical Society, Studia Historica 49: Helsinki/Hamburg, 1994), 
pp. 475-84; John Ford, The Role of Trypanosomiases in African Ecology: A Study of the Tsetse 
Fly Problem (Oxford: Clarendon Press, 1 97 1 ); Kjekshus, Ecology Control. 

45 District Veterinary Officer, Musoma to the D.O. Musoma, 19 January 1928, Annual 
Report 1927, 19 January 1928, Mwanza Province 1927-28, Provincial Administration, 246/ 
P.C/1/30, TNA. 


The outbreak of sleeping sickness in the early colonial years indicates that the cycle of 
disasters had, by the turn of the century, resulted in loss of control over the environment. 
Depopulation had reduced the number of settlements and farms, as well as the ability to burn the 
old grass, thus allowing for the encroachment of bush as a habitat for tsetse flies. The landscapes 
of the western Serengeti have changed significantly in the past one hundred years. Many areas that 
were once open plains have been replaced by bush within the lifetime of elders today. Though 
people did not recognize the relationship between tsetse fly, sleeping sickness and bush 
encroachment, they saw the replacement of open plains areas with impenetrable bush as alarming 
and unhealthy. Western Serengeti people began to rework important social relations during this 
period in order regain their own health and the health of the land. 

Ecological collapse brought an end to the interdependent regional economy of hunters, 
herders and farmers as it had existed before the disasters. Tatoga herders, defeated by the Maasai, 
moved south, following their prophet Saigilo. Bush encroachment also squeezed them out as it 
rendered formerly productive pastures unusable and dangerous for cattle. 46 Asi hunters 
increasingly moved east, as they accepted the patronage of the ascendant Maasai, and Bantu- 
speaking farmers moved west, further into the hills, to avoid raids. Little evidence of the former 
relationships between these three groups remains in oral tradition, which now focuses on the 
opposition between Bantu-speaking ethnic groups, rather than relations between peoples practicing 
other subsistence economies. Settlement and subsistence patterns of Bantu-speaking farmers 
before the disasters are also difficult to reconstruct because ecological collapse demanded new 
coping strategies, later understood as "traditional." 

44 H. A. Fosbrooke, Senior Sociologist, Tanganyika, "Masai History in Relation to Tsetse 
Encroachment," Arusha, 1954, CORY #254, EAF, UDSM. 


Indirect Effects of the Caravan Trade 

The influence of the caravan trade brought about many of these disasters and yet little oral 
evidence exists for the direct presence of caravans in the western Serengeti. Few elders had heard 
anything about caravan traders in the region except in Ngoreme, just south of the Mara River. 
Wakefield's publication of "routes of native caravans from the coast to the interior of Eastern 
Africa," based on Arab testimony, attests to a route from Sonjo, through Ngoreme, to the coast of 
"Ukara," north of what is now Musoma. Otherwise, the western Serengeti remained a blank space 
on the map until almost the turn of the century. The "native" routes across the plains from 
Maasailand usually ended in Kavirondo among the Luo, in what is now western Kenya, near 
Kisumu. 47 However when coming into Ikoma in 1 892 the German explorer, Baumann, noted that 
local people immediately recognized his party as a coastal caravan and greeted him in a 
"Kinyamwezi dialect." Some of Baumann's porters deserted in Ikoma, hoping to stay "as slaves to 
the natives until another caravan passes." Baumann thought this foolish since many years could 
pass between caravans in Ikoma. 48 "Native" caravans, though scarce, were not unknown in the 

Swahili caravans were afraid of following the Maasai route from Kilimanjaro to the Lake 
and consequently did not often attempt it. Europeans did not find a way through until Thomson's 
expedition in 1883-4, which went considerably north of the Mara Region. 49 The peoples of the 

47 T. Wakefield, "Wakefield's Notes on the Geography of Eastern Africa, Routes of Native 
Caravans from the Coast ...." Journal of the Royal Geographical Society 40 fl 870V 303-339. T. 
Wakefield, "Native Routes through the Masai Country," Proceedings of the Roval Geographical 
Society , n.s„ 4(1882): 742-747. Hartwig. The Art of Survival , p. 78 

48 Baumann, Durch Massailand. pp. 38 - 4. The coast was in a state of upheaval in 1892 
and Ikoma may have looked good by comparison to coastal porters. 

49 Joseph Thomson, Through M asai Land: A Journey of Exploration Among the SnowclaH 
Volcanic Mountains and S trange Tribes of Eastern Equatorial Africa... (London: Sampson Low, 

eastern lake had a reputation for being "warlike" and "inhospitable to travellers." Kallmann 
reported that they "massacred whole caravans that were merely crossing the country to purchase 
ivory in Ugaya" (Luo). 50 Jacobs, however, cites evidence that the Arab and Swahili traders had 
exaggerated the dangers of entering Maasailand to "keep the door to the interior closed to European 
exploration as long as possible."" 

The main caravan route went to the south in Sukuma and then across the Lake to 
Buganda. Hartwig claims that the Lake Victoria island of Kerewe was involved in long-distance 
trade from the rule of Chief (Omukama) Mihigo 11 (1780-1840). Migrants fromKanadi in 
Sukuma formed the nucleus of elephant hunting associations, who sent ivory south to the caravan 
trade by way of the chief. As early as the 1850s, coastal traders dealt directly with Kerewe, whose 
chief, in turn, obtained ivory from the mainland. In the next decades Kerewe fell increasingly 
under the influence of the Ganda kingdom across the lake, as its intermediary to the coast. 
Western Serengeti peoples would have had contact with Kerewe as a source of famine food 
because of its higher and more reliable rainfall patterns.' 2 

Marston, Searle and Rivington, 1 887; reprint ed., Frank Cass and Co., 1 968). Thomson states that 
whatever is known about the land beyond Kilimanjaro is from the Wakefield accounts, either 
because the risks were too great or the cost too high. He was commissioned specifically to find a 
way through Maasailand to the Lake by the Royal Geographical Society. 

50 Hartwig, The Art of Survival , p. 541 ; Kollmann, The Victoria Nvanza . p. 177; See also 
Von Hauptmann Schlobach, "Die Volksstamme der deutschen Ostkuste des Victoria-Nyansa," 
Mittteiluneen aus den deutschen Schutzgebieten (Berlin: U. Usher, 1901): 183. 

51 Jacobs, "A Chronology," p. 28. 

52 Hartwig, The Art of Survival , pp. 66-71, 80-81, 116; Itandala, "A History of the 
Babinza," pp. 213-218; E. A. Chacker, "Early Arab and European Contacts with Ukerewe," 
Tanzania N otes and Records 68 (1968): 75-86; C. F. Holmes, "Zanzibar Influence at the Southern 
End of Lake Victoria: The Lake Route," African Historical Studies 4, 3 ( 1 97 1 ): 479-503. 


The main items sought by caravan traders were ivory and slaves. Western Serengeti 
people used ivory locally only for bracelets, which were an emblem in the eldership title system. 
However, evidence exists that elephant hunting societies appeared as a new phenomenon during this 
period, in spite of a traditional taboo against killing elephants. 53 Ikoma elders said that their fathers 
obtained ivory through friendships with local Asi hunter/gatherers. No local institution of slavery 
existed in the western Serengeti. Families adopted and incorporated strangers (abasimano) who 
came on their own or were sold for food during times of famine. They treated these strangers as 
members of the family, whose children would have the same rights as anyone native born. It is 
only toward the lake that the word for strangers translates as "slaves" or "dogs" (seese) and a 
stranger's children were denied the rights of the native born. 54 I found no oral evidence that 
western Serengeti peoples engaged in slave raiding to supply the caravans. 

The devastating effects of the caravan trade on local society are well documented 
elsewhere. 55 The case of the western Serengeti is important because it demonstrates that even with 
the most insubstantial contact the caravan trade had long-term effects. One of the first 
missionaries in the region soon after the turn of the century reported that his student told him that 
the porters,"whispered around the campfire in the evening" that "there is a famine in their country 
and they are going to take our cattle and children to salt them down for shipment to Europe." 56 
Given the nature of the evidence, knowing the more ephemeral consequences of these contacts on 
the ways in which people imagined or re-imagined social relations during this time is impossible. 

See Chapter 10 on Sarabarando hunting associations; for elephant associations related to 
Ukerewe see Hartwig, The Art of Survival , pp. 66-67. 

54 See Chapter 7 on the incorporation of strangers, abasimano. 

55 For an overview see lliffe, A Modern History , pp. 40-77. 

56 Toppenberg, Africa Has My Heart , p. 45. 

Observing the disasters around them and knowing of the new power arising through traders on the 
coast, did local people begin to imagine themselves as part of larger associations of communities, 
opposed to a much more distant "other?" 

One obvious effect of the caravan trade was the increasing trade of forest products to 
Sukuma. Sukuma people lived at the terminus of the overland route at Lake Victoria. They were 
directly involved in trade and worked as porters for the caravans." The Sukuma used their new- 
found wealth from the caravans, in part, for expanded ritual and prestige activities. The demand 
for ivory bracelets, wildebeest tail fly whisks and bracelets, ostrich feathers and eggs, wild animal 
skins and lions' manes increased rapidly in the second half of the nineteenth century. Western 
Serengeti peoples supplied these products in return for iron, salt or livestock. Hunting became 
much more commercialized during this period, far beyond the immediate need for meat. 

Because of their connection to the caravans Sukuma people held the advantage in trade 
relations and western Serengeti people sought them as patrons. This may be the material 
explanation for western Serengeti oral traditions, dating from this period, which establish ritual and 
kin relations with the Sukuma. The disorder of this period seems to have provided the opportunity 
for a Sukuma rainmaker to unite diverse clan territories into the Ikizu chiefdom of western 
Serengeti. Oral traditions then projected these connections to Sukuma rainmakers back to the time 
of origins. The oral traditions concerning hunting and Sukuma origins in earlier times may, in fact, 
reflect more closely this later period of commercialized hunting. Because the caravans themselves 
were rarely encountered in the western Serengeti, the changes precipitated by contact through trade 
in Sukuma were not often recognized as such in oral tradition. 

" Baumann, Durch Massailand. p. 67, when Baumann reached the area of Magu, just 
south of the Mara Region in Sukuma, he observed that, "the natives are great travelers, almost all 
of them were young people who had been to the coast." 

Maasai Raids 

Oral testimonies of the western Serengeti express most directly and keenly the devastating 
effects of the disasters through the experience of Maasai raids. That Maasai raids took place 
during this era is certain, but the relationship western Serengeti peoples had with the Maasai is 
much more complex. Research on the Maasai demonstrates with a fair degree of certainty that the 
Maasai did not move into the Serengeti region until the first half of the nineteenth century, and 
probably not before mid-century. 58 The Maasai entered this region gradually, in search of new 
grazing areas." They did not automatically displace others who preceded them, nor were they at 
first the dominant "Lords of East Africa," which they became by the last half of the century. The 
Maasai pastoral lifestyle depended upon symbiotic interaction with farmers and hunter/gatherers 
within a regional system. At first, western Serengeti peoples seem to have accommodated and 
admired the Maasai. taking on many Maasai cultural innovations as their own. 60 

The Maasai gained dominance by developing a highly specialized form of pastoralism that 
swept the plains in the nineteenth century, rapidly replacing the older regional system where 
farmers also hunted and herded and the preexisting Tatoga pastoralists also farmed a little. 

58 In the age of Merishari (c. 1806-1826) they took the Lake Manyara area from the Tatoga 
and in subsequent ages, perhaps as late as the 1 840's the Maasai forced the Tatoga to withdraw 
from the Ngorongoro Crater and Engaruka area. John G. Galaty, "Maasai Expansion and the 
New East African Pastoralism," in Being Maasai: Ethnicity and Identity in East Africa , eds. 
Thomas Spear and Richard Waller (Athens: Ohio University Press, 1993), p. 74. Berntsen, 
"Pastoralism, Raiding and Prophets," p. 31. 

59 See Berntsen, "Pastoralism, Raiding and Prophets," p. 40, on the process of migrational 

60 "The Lords of the East Africa" is the expression used by Richard Waller in his thesis, 
"The Lords of East Africa: the Maasai in the mid-nineteenth century (c. 1840-c. 1885)" (Ph.D. 
dissertation, University of Cambridge, 1979), to indicate the dominance and status of the Maasai at 
this time period. John Lawrence Berntsen, "Pastoralism, Raiding and Prophets," p. 32. Also Spear 
and Waller, Being Maasai . on economic symbiosis. "Introduction," p. 2-4. 

Through control of access to cattle as a store of wealth and limited pastoral resources, they began 
to impose economic specialization on everyone else in the region, creating the non-cattle owning 
categories of "farmer" and "hunter" from peoples who had practiced both. The pre-Ishenyi 
community at Nyeberekera was one victim of this increasing competition for dry-season grazing 
grounds and water points, located as it was on the edge of the Serengeti plains. The Maasai forced 
the agro-pastoral community that once straddled the Serengeti plains back to the hills on its 
margins, both east (Sonjo) and west (Ikoma). A regional economy developed with the Maasai at 
the center, as the main beneficiaries of the system. As is clear from the experience of the western 
Serengeti, Maasai power lay in both military and cultural domination. We cannot understand 
western Serengeti creativity during this period outside its subordinate and peripheral position 
within this hegemonic system, both in terms of its acquiescence and its resistance. 61 

It seems likely that the most intense Maasai raiding in the western Serengeti actually took 
place after the rinderpest panzootic, as a strategy to recover stock. Before the rinderpest western 
Serengeti peoples kept very little livestock and only began to build up large cattle herds after the 
"Hunger of the Feet," which accelerated the trade of forest products to Sukuma for livestock. 62 
During the generation of disasters a wealthy man owned four head of cattle and most were lucky to 
have one or two, mainly counting their livestock in sheep and goats. The Nata paid bridewealth in 
wild animals skins until the turn of the century. This later period of intense raiding went on up to 
the beginning of British rule. In 191 1 the White Fathers reported that many refugees that had 
come to them were still victims of Maasai raids. 63 The Maasai may have raided the lakeshore even 

61 Spear and Waller, Being Maasai . "Introduction." 

62 This is more completely developed in Chapter 10. 

63 Societe de Missionnaires d'Afrique, "Nyegina (Notre-Dame de Consolation)," Rapports 
Annuels 191 1-1912, p. 392; Societe de Missionnaires d'Afrique, Chronique Trimestrielle de la 

more intensively than the western Serengeti. During the World War One Kenyan Maasai took 
advantage of chaos on the border and made raids into the German colony. 64 This later period of 
raiding was used as a template in the historical imagination to understand the dispersal of the 
previous generation. Oral traditions blame raiding for the most grievous effects of the disasters. 65 

Raids were only the most obvious and resented symbol of Maasai domination that reached 
into all aspects of life. Raids took place mainly during the dry season. They often came in the 
predawn hours, surprising the village, stampeding the cattle, burning the houses and sometimes 
killing men or taking captive the women and children. The Maasai fought with spears and shields, 
while the western Serengeti peoples mainly used bows and arrows. If the Maasai came by 
surprise, they had the advantage in close combat. Sutton attributed Maasai military superiority to 
the use of larger, socketed, spear blades along with novel forms of military organization and tactics 
(which western Serengeti people later adopted). 66 The White Fathers reported in 1904, when they 
took a trip to the interior of the country, that all along the lake people lived in fear of Maasai raids 
from the plains. They would not only raid cattle but burn houses and fields, leaving devastation 
behind them. 67 In 1 902 the Germans built Fort Ikoma in the western Serengeti specifically to 
control Maasai raiding. 

Societe de Missionnaires d'Afrique (Peres Blancs) 24me Annee, No. 94, Avril 1 902, p. 94. 

64 Toppenberg, Africa Has Mv Heart , p. 63 

65 This insight thanks to on-going conversations with Richard Waller. 

66 J. E. G. Sutton, "Becoming Maasailand," in Being Maasai . p. 42. 

67 L. Bourget, Trip Diary, 1904, "Report of a Trip in 1904 from Bukumbi to Mwanza, 
Kome? Ukerewe. Kibara, Ikoma-Mara Region, together with some stories," N.p. n.d. M-SRC54, 
Sukuma Archives, Bujora, Mwanza; See also Societe de Missionnaires d'Afrique, Chronique 
Trimestrielle de la Societe de Missionnaires d'Afrique (Peres Blancs) 24me Annee. No. 94, Avril 
1902, p. 94. 


People throughout the Mara Region know all Maasai (and Maasai-like peoples) as 
"Kwabhe," but more specifically those raiding in Ikoma and Nata were the Serenget of the present 
Loliondo area while those raiding in Kuria, and perhaps Ngoreme, were the Siria of the present 
Narok area. Fosbrooke reported that the Serenget Maasai section was strong throughout what is 
now Serengeti National Park until 1890 when the Loitai Maasai absorbed them. In the intense 
debates surrounding establishment of the Serengeti National Park boundaries in the 1930s 
Fosbrooke and others claimed that the Serenget Maasai had for the past century used the western 
Serengeti as fall-back grazing in times of drought. 68 The Loitai expansion was also responsible for 
pushing the Siria back from the Loita Plains to the Mara River and into competition with the Kuria 
for grazing land. 69 

Because of the highly emotional history of relationships with the Maasai, the historian 
must treat stories about the interaction of western Serengeti peoples and the Maasai carefully. We 
can also recover relations of cooperation, blood-brotherhood, marriage and kinship from this 
period. The interaction of western Serengeti peoples with the Maasai during this era is explored in 
more detail in Chapter 9. This brief summary alerts the reader to the context surrounding 
references to Maasai and cattle raiding that appear throughout the dissertation. This is a further 
part of the puzzle in understanding relations between herders, hunters and farmers before the 
disasters. Many innovations of the late nineteenth century bear the mark of Maasai influence, if 

68 Fosbrooke, "Sections of the Masai in Loliondo Area." For the entire debate see National 
Game Parks files, 21 5/350/vols. I-IV, TNA. 

65 Galaty, "Maasai Expansion," p. 72. The Loitai confederation included the Siria, 
Laitayok, Salei, Serengeti and Loitai. 


not direct borrowing, to the point where many observers of the time concluded that western 
Serengeti peoples must be "of Maasai blood."' 


This basic outline of the late nineteenth century events demonstrates the scope and 
intensity of the disasters. Societal stress does not necessarily result in the transformation of 
identity. However, in this case, the movement of large numbers of people out of the region to 
Sukuma or Kerewe and into the region, from Maasai, seriously disrupted the cohesion of 
territorially based communities. To gain protection from raids and access to a wider range of 
resources, local communities changed the social basis for organizing themselves. They emphasized 
the importance of age-sets in creating social unity where, formerly, clans had been made to do that 
work. Beginning with the next chapter, the traditions containing evidence of social process over 
the longue duree are analyzed. However, we cannot build a picture of this earlier time frame 
without reference to the ways in which historical consciousness was transformed during the 

The historian must constantly keep in mind the possible effects on oral tradition of these 
radical changes in social identity. For example, 1 interpret the oral traditions about early relations 
between herder, hunters and farmers in light of the later breakdown of a regional economy of 
interdependence and the enmity toward Maasai herders. The later commercialization of hunting to 
gain wild animal products to sell in Sukuma must inform one's reading of oral traditions about the 
early hunting economy. Finally, knowledge of the influential position of the Sukuma in the 
caravan trade networks must be brought to bear on traditions of origin in Sukuma. Above all, the 

70 Schnee, Deutsches Kolonial Lexikon. . pp. 121, 679-81; Weiss, Die Volkerstamm e nn 


new identities that emerged during this period of age-set cycle territorial membership and ethnicity 
cannot be projected back onto a "traditional" and timeless past. 






This chapter explores the socially occupied landscapes of the distant past through the 
interpretation of oral traditions of origin or emergence. I investigate the different layers of meaning 
in the emergence stories to present a culturally nuanced understanding of long-term regional 
developments. Although elders tell these narratives as the origin stories of present day ethnic 
groups, the core spatial images are those of the male and female spaces of the homestead and the 
ecological landscapes of hunters, farmers and herders. They refer to the long-term ongoing 
generative principles relating to production and reproduction. The emergence stories suggest the 
frontier processes by which settlers, forging new economic strategies among neighbors of diverse 
cultural backgrounds, developed a new cultural synthesis as the basis for economic prosperity over 
the next millennium. 

Different versions of the emergence stories enrich historical analysis based on other 
sources of evidence such as historical linguistics, archaeology or comparative ethnography. An 
amazing congruence exists between the seemingly "mythical" narratives and historical 
reconstructions based on other kinds of evidence, about which the narrators of these tales could not 
have known. Historical linguistics tells us that Bantu-speaking farmers entered and eventually 
came to dominate a region where herders and hunter/gatherers of other language groups already 
lived. The Nata emergence story suggests how new social identities might have developed as 
Bantu-speakers incorporated hunters and hunting knowledge into their farming settlements on the 
frontier. The lkizu emergence story illustrates a pattern of the gendered division of labor within 


autonomous but interdependent spheres of authority, which settlers on the frontier may have 
deployed to structure relations within the homestead and to others beyond. The next chapter 
continues the investigation of emergence stories by looking at the Ikoma and Ngoreme accounts 
that primarily concern the ecological spaces of interaction between hill farmers, woodland hunters 
and grassland herders on an inter-cultural frontier. 

In my collection of oral traditions, elders most often wanted to begin with the asimoka 
stories-the origin or emergence narratives of first man and first woman. They considered these 
stories the foundational historical narratives, necessary in any account of the past. Local educated 
men translated asimoka as "origins," and more specifically as ethnic origins. The Musoma 
District officers used the asimoka stories that they collected to define "tribes" by reconstructing 
their origins and migrations in the "tribal" model. 1 The way these reports used the term then, and 
the way elders still use it today, seems to imply that each ethnic group can be traced back to one 
root, with many branchings along the way. 

Asimoka, however, has a linguistic derivation whose meaning may provide alternate, and 
less unitary, models for understanding the past. In Nata, the verb -sisimoka means "to spring up," 
as in waking up from a sleep or small sprouts popping up out of the ground. 2 In the related 
language of Kuria the root word is -semoka, meaning "to originate from or the rise (of a river)." 3 
Social theorists have suggested a similar image of a rhizome, instead of a tree, for understanding 
the historical development of different groups of people. A rhizome describes a kind of 
subterranean plant stem that grows horizontally, producing shoots above and roots below. Many 

1 See a discussion of the "tribal model" in Chapter 2. 

2 Augustino Mokwe Kisigiro, "Nata-Swahili Dictionary," unpublished, n.d. 
1 Muniko et al., Kuria-English Dictionary , p. 32, 1 1 5. 

plants that appear distinct on the surface develop from the same network of rhizomes, which can 
cover a large area without a distinct beginning or end. A rhizome is a metaphor of alliance rather 
than filiation, of coming and going rather than starting and finishing. It is a way of understanding 
multiplicity without the need for reduction to a single beginning. 4 

Rhizomatous grasses in the Serengeti have networks of connected underground stems that 
may sprout up in many places at once. Because the rhizomes lie beneath the ground, all the 
surface growth may be burned off, only to sprout up again in different places. This model 
recognizes the interconnected network of rhizomes that recombine and sprout up in new ways 
rather than deriving all things from a single original stock. I will translate asimoka, in this sense, 
as "emergence," the emergence of new identities out of the old tangled underground network of 
rhizomes, without simple and primordial origins. 5 Tracing cultural origins back to a single source 
is not possible, but like the rhizomatous network of Serengeti grasses or the underground 
connection of water courses, cultural change appears as new growth, or new springs, and presents 
creative new options. 

The Development of a Regional Culture from Diverse Interactions in the Distant Past 

Through the evidence of historical linguistics and archaeology, we know that the peoples of 
the western Serengeti, have developed in a rhizomatous pattern, from multiple and interconnected 
stems in the distant past. As early as 500 A.D. peoples speaking languages from four major 
language groups, with diverse economic and cultural practices, each occupying a separate 
community corresponding to a different ecological niche, inhabited what is now the Mara Region. 

4 Gilles Deleuze and Felix Guattari, "Rhizome: Introduction," A Thousand Plateaus 
(Minneapolis: University of Minnesota Press, 1986), pp. 1-19. Thanks to Patrick Malloy for this 
reference and insight. 

5 For the idea of "emergence" instead of "origin" see, Tedlock and Mannheim, The 
Dialogic Emergence , pp. 8-15. 

Western Serengeti culture, as a regional set of common cultural assumptions, emerged out of the 
interactions of these diverse peoples in the distant past. Archaeologists and historical linguists 
working with a broad brush, have produced an overview of the Neolithic and Early Iron-age 
periods the Rift Valley region of East Africa. 
The Evidence of Archaeology 

Archaeologists investigating the origins of food production in East Africa date the 
development of the Pastoral Neolithic period to between 1000 B.C. and 700 A.D. During this time 
local hunter/gatherers in East Africa adopted livestock that had been domesticated in the Sahara. 
The continuity in lithic and ceramic wares between Neolithic and pre-Neolithic sites supports this 
supposition. Some archaeologists hypothesize that small groups of herders, escaping drought in the 
north, brought the livestock. Pastoralism gradually came to occupy a more important place in local 
economies due to the late Holocene drying trend, the introduction of better adapted cattle breeds 
and the establishment of annual wet and dry seasons. During the next stage of development 
discontinuities in the material artifacts do occur, suggesting increased immigration of, or influence 
from, herders originating in the north. Archaeologists find no evidence for large-scale immigration 
in the later stages of the Pastoral Neolithic and economic strategies seem to have shifted slightly 
toward more reliance on hunting and gathering. 6 

Stanley Ambrose further refines this reconstruction of Neolithic history by identifying the 
coexistence of three distinct sets of material artifacts in close proximity, corresponding to the 
diverse ecological habitats of the Rift Valley-Eburran Industry representing a hunter/gatherer 
lifestyle eventually confined to the montane forests, the Elementaitan Industry representing an 
agro-pastoral lifestyle in the forest/savanna ecotone, Highland Savanna Neolithic Industry 

6 Bower, "The Pastoral Neolithic of East Africa," pp. 74-76 


representing a pastoral lifestyle in the open lightly wooded savanna grasslands. No material 
evidence for agriculture at this time yet exists. Ambrose's research shows that about 1 ,300 B.C. 
the material remains of Highland Savanna Pastoral Neolithic communities appear abruptly in the 
archaeological record and continue to coexist with evidence for small groups of Eburran 
hunter/gatherers (who also took up pottery and small stock-raising). Further evidence shows that 
distinctive Elmentaitan material remains appear in the region about 500 B.C., presumably brought 
by peoples from the northwest. Competition among these three groups must have been minimal 
because each had adapted to a different ecological zone. 7 

Ambrose further proposes a continuity between the Neolithic Eburran hunter/gatherers in 
the highlands and Rift Valley of East Africa, identified in the archaeological record, and present 
day hunter/gatherer populations throughout East Africa. The continuity in space and time of the 
hunter/gatherer way of life is remarkable. Both from the archaeological record and from 
ethnographic description in the last century Ambrose shows that over this long period 
hunter/gatherers have shown an unusually high dependence on meat, making up 80% of their diet, 
in contrast to plant foods. They supplemented this high protein diet with honey and fat. 
Hunter/gatherer populations of the past seem to have been fairly sedentary, occupying the eco-tone 
between savanna and forest to exploit both regions. The archaeological remains show that Eburran 
hunter/gatherers mainly hunted resident ungulates of the woodlands, with traps and snares, rather 
than the larger species of the open plains, and also kept some small stock.* 

7 Stanley H. Ambrose, "The Introduction of Pastoral Adaptations to the Highlands of East 
Africa," in From Hunt ers to Farmers: The Causes and Consequences of Food Production in 
Africa, eds. J. Desmond Clark and Steven A. ( Berkeley: University of California Press 1984) 
pp. 222- 33. 

8 Stanley H. Ambrose, "Hunter-Gatherer Adaptations to Non-Marginal Environments: An 
Ecological and Archaeological Assessment of the Dorobo Model," Sprache und Geschichte in 
Afrika 7.2 (1986): 11-42. 


In historic times hunter/gatherers in East Africa nearly always lived in close interdependent 
relationship with agriculturalists and pastoralists. The evidence from archaeology shows that 
incoming Savanna Pastoral Neolithic communities, and later Elmenteitan communities, severely 
restricted the ecological niche occupied by Eburran hunter/gatherers. This may have forced the 
hunter/gatherers to compensate for their loss of resources by developing relations of 
interdependence with the incoming populations. In the western Serengeti context they may have 
also moved farther into the woodlands. 9 Present-day hunter/gatherers have established 
interdependent relations with many different groups, speaking Maa, Kalenjin, or Cushitic 
languages, which may retain traces of a Khoisan language. 10 
The Evidence of Historical Linguistics 

Linguistically inclined historians reconstruct the history of food systems in the Neolithic 
and Early Iron-Age periods from different kinds of evidence. They use the concept of the speech 
community rather than the concept of material cultural traditions to think about the history of these 
early periods. They postulate that the two earliest peoples in what is now the Mara Region of 
Tanzania spoke either a Southern Cushitic language and practiced a pastoral economy, or a Rub 
Eastern Sahelian language and practiced a mixed pastoral and farming economy. Linguistic 
evidence also exists for the presence of hunter/gatherers who become associated with pastoralist 
and agro-pastoralist communities in the greater Rift Valley." 

9 Ibid, p. 30; Ambrose, "The Introduction of Pastoral Adaptations," p. 238. 

10 Michael G. Kenny, "Mirror in the Forest: The Dorobo Hunter-Gatherers as an Image of 
the Other," Africa 51, 1 (1981): 479; Ehret, Southern Nilotic History p 73- Magnir,. „ cobmel 
officer in Maasailand during the 1920s, named eight different "Dorobo" groups, each speaking 
different languages with different levels of integration with other peoples. R. A. J. Maguire, "11- 
Torobo." Tanganyika Notes and Records 25 (June 1948): 1-26. 

" Christopher Ehret, The Classica l Age of Eastern and Southern Africa: A History. 1000 
B.C. to A.D. 300 (Charlottesville: University Press of Virginia, forthcoming). 


Bantu-speakers began moving into this area from their older settlements around the Lake 
Victoria shoreland by or before 300-400 A.D. As these immigrants became separated from their 
Great Lakes Bantu ancestral speech communities, they began speaking distinct languages, that 
historical linguists now call the East Nyanza group. At about the same time, but probably a little 
later, agro-pastoralists speaking a language known as Mara Southern Nilotic began moving from 
the north into the interior of what is now the Mara Region, 12 

These newcomers moved into the ecological zones formerly occupied by the previous 
inhabitants of the region and broke down the agro-ecological boundaries that had confined different 
speech communities to the ecological niches best suited to their agricultural expertises. East 
Nyanza Bantu-speakers who had practiced a root-crop and fishing economy with small stock and 
some reliance on hunting, along the lakeshore, now began to expand their food producing 
possibilities by adopting grain-crop (eleusine millet and sorghum) farming, and increasing their 
hunting and herding expertise as they moved into the drier interior. They learned these new skills 
from their neighbors who had adapted to this environment. Evidence for this is found in the loan 
words relating to herding, grain farming and hunting adopted during this time by East Nyanza 
speakers. ' 3 

Gradually the Southern Cushitic- and Eastern Sahelian-speakers disappeared from the 
historical record. Based on historical linguistics we cannot tell whether they left, died off or 
whether they integrated themselves into the growing community of Mara-speakers. In the western 
Serengeti, Mara Southern Nilotic-speakers, who would have occupied approximately the same 

12 Schoenbrun, "Early History," pp. 156-7, 182-204 

13 Ibid; David L. Schoenbrun, "We are what we eat: Ancient agriculture between the 
Great Lakes," Journal of A frican History 34 (1993): 1-31 ; Ehret, Southern Nilotic Historv. pp 40- 


ecological niche as these two former pastoral groups, seem to have succeeded them. 14 [See Figure 
4-1 : Linguistic Maps of the Lakes Region Over Time.] 

Both historical linguists and archaeologists have noted that the geographic distribution, 
time frame and sequence of events suggested for these different communities of East African 
peoples in these two sets of evidence roughly correspond. To make inferences from this data one 
would have to assume that a correlation does exist between a linguistic group and its material 
culture. On this basis, some have suggested that the past distribution of Southern Cushitic- 
speakers corresponds with Savanna Pastoral Neolithic Industry sites and that the past distribution 
of Southern Nilotic-speakers corresponds with Elmenteitan Industry sites. The relative sequence 
and dating for Southern Cushitic-speakers entering the region before Southern Nilotic-speakers 
also correspond. 15 Historical linguists and archaeologists thus may be looking at the same 
historical processes using different kinds of evidence and historical reasoning. After 1000 A.D. 
historians must rely on the evidence of historical linguistics, since little archaeological work in the 
Mara Region covers this period. 

Bantu languages became dominant throughout what is now the Mara Region by or before 
1000 A.D., but not without considerable influence from the earlier period of diverse interactions. 
Ehret argues that during the second half of the first millennium A.D., when Bantu-speaking 
agriculturalists settled among the more mobile Mara Southern Nilotic-speaking agro-pastoralists, 
they developed an interdependent relationship. East Nyanza speakers adopted, as loan words from 
Mara Southern Nilotic-speakers, the cycling age-set names used today by western Serengeti 
people, which dates this innovation to the period between 500 and 1000 A.D. Other Southern 

14 Ibid. 

15 Ambrose, "The Introduction of Pastoral Adaptations," pp. 233-234. 


•j V 5- 

ai It 


o ** 

. J* 

■ - in o 

quo -< c 
co w * 

n— M » 
J » EC 

Ij Q 

j> < « B 

all 3 


3 • hH 
00O * ~ 
CO u jj „ 

s 2 1.E3 j 

a c -5 " 

■ I ■ M -^ z 

■rtW « c 

•-' (l: .- 

OJ ■( V) ~i v 

tw eg id ■ 

O S 

O © 

<fc < 

<■ -J 










■ — 

£ Q 

= Q 


£ a 

•2 & 
oo 5- 

OS - 


S « 

w i_ 


Nilotic loan words in East Nyanza languages dating from this period of early contact include 
vocabulary connected to livestock (sheepskin, Iamb, he-goat), stages of the life cycle and non-kin 
relations (young man, young woman, friend, oath, age-set), and a new word for the homestead or 
cattle corral, aka."' This evidence may suggest that East Nyanza-speakers, moving into the 
unfamiliar environment of the interior, used common age-sets and the comradeship of peers to gain 
access to livestock expertise and to develop new kinds of homesteads built around the livestock 

Whatever the historical process, today no speakers of the Mara Southern Nilotic language 
remain in the Mara Region. One possibility is that as East Nyanza Bantu-speakers adapted to new 
ecological zones they gradually absorbed Mara Southern Nilotic-speakers into their own 
communities. 1 '' Another possibility is that the descendants of the Mara Southern Nilotic speakers 
today are the Southern Nilotic Dadog-speakers, the Tatoga Rotigenga and lsimajek, who now 
occupy the plains of western Serengeti best suited for pastoralism. I favor the latter explanation, 
although individual Mara Southern Nilotic-speakers surely crossed linguistic boundaries to become 
East Nyanza Bantu-speakers. It seems more likely that most of these pastoralists adopted the 
language and lifestyle of another incoming pastoralist group, rather than that of their agriculturalist 
Bantu-speaking neighbors. Is 

" Ehret, Southern Nilotic History , pp. 130-132, Tables D.l and D.2. 

" Ibid, p. 40-42. Their presence in the past is deduced by Ehret in his reconstruction of 
Southern Nilotic loanwords in East Nyanza languages, containing sounds which were not part of 
Kalenjin or Dadog languages, and which pre-date the split of East Nyanza languages into Suguti 
and Mara branches. 

18 More linguistic investigation would be necessary to determine whether there are 
phonological and lexical transfers from Mara Southern Nilotic to Dadog. There are numerous loan 
words from Mara Southern Nilotic in East Nyanza languages. 


It is not clear from historical linguistics when the Dadog-speakers came to the western 
Serengeti. What we do know from the evidence of loan words in present-day languages of the 
region is that Dadog-speakers were in northern Tanzania, what is now Maasailand, perhaps as far 
west as the Mara, from the first millennium A.D. They spread south into what is now the Maasai 
Steppe and southwest into parts of Kondoa, Mbulu and Singida after 1000 A.D. About the same 
time, incoming South Kalenjin- speaking peoples assimilated the northern Dadog-speaking 
settlements." If the lifestyle practiced by Dadog-speakers presented clear advantages for other 
pastoralists such as the Mara Southern Nilotic-speakers, Dadog-speakers (rather than East 
Nyanza- speakers) could have completely absorbed Mara Southern Nilotic-speaking communities 
by around 1000 A.D. 

The name from which "Tatoga" (for Dadog-speaking people) is derived seems to date from 
the first millennium A.D. in northern Tanzania, suggesting long-term continuity in their sense of 
identity as a people. The characteristic features of Dadog- speaking culture have also remained 
constant over the millennium-as herders of cattle, sheep and goats, who drink milk, bleed cattle 
and pay bridewealth in livestock, cultivate some grain and hunt. They did, however, drop the 
Southern Nilotic cycling age-set system of eight names and adopt a non-cycling linear generation- 
set system, in distinction to their Bantu-speaking neighbors who kept the old Southern Nilotic 
cycling names. 20 

This dramatic change in age-set organization may indicate a period of increased 
differentiation and separation between agriculturalist and pastoralist communities after 1000 A.D. 
By this time, East Nyanza Bantu-speaking communities had already gained herding expertise from 

19 Ibid, pp. 55-62. 

20 Ibid. Loanwords from Dadog appear in Sonjo, Iraqw and Aramanik. The impact of 
Dadog on the ancestors of the Sonjo was particularly significant. 

their Mara Southern Nilotic-speaking neighbors and had developed a lifestyle that exploited each of 
the other economic subsistence patterns (farming, herding, and hunting). Through this strategy 
East Nyanza-speakers gained dominance in the region, perhaps requiring Dadog-speaking 
pastoralists to move farther out on the plains and away from the eco-tones where a combined 
herding, hunting and farming economy was possible. The identity of Dadog-speakers who 
absorbed Mara Southern Nilotic-speakers may also have been more exclusively pastoralist, 
demanding clearer distinctions (within a system of economic interdependence) from their 
agriculturalist neighbors. 

Bantu-speakers and Dadog-speakers also continued to coexist with hunter/gatherers. The 
hunter/gatherer people who appear in oral traditions of the western Serengeti are called the Asi. 1 
could identify no Asi communities in the region during my fieldwork, although western Serengeti 
people living today can remember interacting with Asi hunter/gatherers in their youth. I can only 
speculate that either they moved into Maasailand, where Maasai call them the Ndorobo or Okiek, 
or Bantu-speaking communities absorbed them, leaving little memory of Asi tradition. 21 Because I 
identified no descendants of this tradition, 1 had no way of finally knowing what language they 
spoke, whether Southern Nilotic, Cushitic or Khoisan. 22 Asi seems to be a categorical rather than 

21 This speculation is based on the evidence of oral traditions of the late nineteenth century 
and early colonial reports, see next chapter on the relation of more recent Asi to western Serengeti 

22 Numerous academic arguments exist for the ancestors of most Okiek hunter/gatherer 
communities of East Africa being Kalenjin; see John Distefano,"Pre-Colonial History of the 
Kalenjin" (Ph.D. Dissertation, U. C. L. A., 1985); and Corinne A. Kratz, "Are the Okiek really 
Masai? or Kipsigis? or Kikuyu?," Cahiers d'Etudes africaines 20, 3 (1980): 360, who draws on 
the evidence of historical linguistics and oral history to postulate that all of the Okiek related 
hunter/gatherers once lived together before 1000 A.D., probably in Northwestern Kenya, and 
speaking a Southern Nilotic language related to present day Kalenjin. Ehret demonstrates that the 
distinction of hunter/gatherers as a separate conceptual category was a proto-Southern Nilotic 
innovation, see Ehret, Southern Nil otic History , pp. 79-80. Cushitic speaking hunter/gatherers in 
the greater Rift valley are known by names close to "Asi," which may simply be the word for "bush 

an ethnic name, referring to people who live in the bush, or people of the land, taken from the root - 
Si, meaning "dirt, soil, land, ground, place." 23 Kratz points out that at the heart of Okiek 
(Ndorobo, Asi) identity is their ability to cross boundaries; "a sense of being mediator, code- 
switcher, interstitial." The space of the hunter/gatherer is that of "inhabiting cultural boundary 
areas as their own." 24 They have lived on the boundaries of many different communities (Kalenjin, 
Maasai and Kikuyu), accepting much of the language and culture of others while maintaining a 
sense of their own identity. 

People speaking East Nyanza Bantu languages, and living among people speaking other 
languages, diversified over time as they became separated from each other. The diagram of the 
family tree of Great Lakes Bantu languages illustrates the relationships among these languages 
over time [See Figure 4-2: Great Lakes Bantu Linguistic Tree.]. Those who stayed near the 
lakeshore came to speak Suguti languages (Jita, Ruri, Regi, Kwaya) and those who went inland 
came to speak the Mara languages. Through the method of glottochronology described in the 
Chapter 1 we know that these two communities grew distinct from each other about 1 500 years 
ago. As the Mara-speaking communities spread into new lands, those who crossed the Mara river 

dwellers" in general. See Kenny, "Mirror in the Forest," p. 481 ; Itandala, "A History of the 
Babinza," pp. 24-27; Pare oral traditions also name the original hunter/gatherer populations as 
Asa or Asi, Kimambo, A Political History of the Pare , pp. 14,27; D. F. Bleek, "The Hadzapi or 
Watindega of Tanganyika Territory," Africa 4, 3 (July 1 93 1 ); 273-286. Colonial anthropologist 
Henry Fosbrooke reported that the Lake Eyasi hunter/gathers called themselves "Hesabet" and 
were called "Ngabobwo" (rather than Ndorobo) by the Maasai, and Batandigo or Bahe by the 
Sukuma. Henry A. Fosbrooke, "Sections of the Masai in Loliondo Area" (typescript) 1953, CORY 
#259, EAF UDSM; Sutton, "Becoming Maasailand," pp. 50-5 1 . Depicting a key element in the 
hunter/gatherer economy, the word for "beehive," omutana (in Nata), is a Southern Cushitic 
loanword in East Nyanza languages (-tana), see Christopher Ehret, Ethiopians and East Africans: 
The Problem of Contacts (Nairobi: East African Publishing House, 1974), p. 82, Table 5-2, 
Nyanza Southern Cushitic loanwords in East Victoria and Southeast Victoria Bantu. 

23 Schoenbrun, personal communication. 

24 Kratz, "Are the Okiek really Masai?," pp. 359-360. 


Tree Diagram!. 2. Croup Average - Great Lakes Bantu. 

Percent congation out of a 100-word list. 

90 85 80 75 70 65 60 55 50 45 

5 S.Luhyia 




tral 1 



VortA— i 

10 1 Nort 

11 1-Ruta 














2 8- 

West Highlands—' 



"~~ 1 

3 3 \-Suguti- 


3 5 1 










-South Mara- 






— r—North— 1 
— ' Mara 
75 70 

65 60 55 50 45 40 

1500BP 2000BP 2 500BP 

Glottochronology at shared retention rate of 73-74X per 
thousand years. Note: The nuabers at the left margin 
correspond to nuabered languages in the Outline 
Classification above. 

Figure 4-2: Great Lakes Bantu Linguistic Tree. [David Lee Schoenbrun, "Early History in 
Eastern Africa's Great Lakes Region: Linguistic, Ecological, and Archaeological Approaches, ca. 
500 B.C. to ca A.D. 1000," (Ph.D. dissertation, UCLA, 1990), p 362, with permission.] 

formed the language communities of North Mara-Kuria and Gusii. In South Mara they 
differentiated themselves into three groups, probably becoming distinct about 500-300 years ago- 
Ngoreme, eastern South Mara (Nata, Ikoma, Ishenyi) and western South Mara (Ikizu, Zanaki, 
Shashi or Sizaki). 

Although local convention recognizes each of the western Serengeti languages (South 
Mara) as a separate language today, they are all closely related and, thus linguistically, represent 
one group of people with a common heritage in the past. The dialect chaining chart on the next 
page [See Figure 4-3: Dialect Chaining Chart of the East Nyanza Languages] shows the close 
relationship of the Mara languages to each other. The numbers on the chart show the percentage of 
shared vocabulary from a list of 100 core vocabulary words. 25 South Mara languages are today 
mutually intelligible with differences in pronunciation and vocabulary used locally as an indication 
of home community. 26 Mekacha argues that the South Mara languages (Nata, Ikoma, Ishenyi and 
Ngoreme) should be considered dialects of one language. 27 

A number of innovations in vocabulary took place among South Mara or western 
Serengeti speakers after their languages differentiated from Suguti speakers along the lake shores. 
Many of these words are also different from the Kuria words for the same objects, some of which 
retain the Great Lakes Bantu terms. Based on glottochronology these innovations would date to 

See Chapter 1 for an explanation of core vocabulary words and their use in historical 

26 Schoenbrun, "Early History," p. 157. 

27 Mekacha, The Socioli nguistic Impact , p. 56. South Mara languages in this light would 
be considered a dialect cluster. Among the South Mara languages, Ikizu and Zanaki tend to be 
more conservative in maintaining the old Bantu sound patterns, while the eastern South Mara 
group, toward the Serengeti plains, adopted more of the of non-Bantu words and sound patterns. 
This may be an indication of the necessity of those further to the east to adopt to a pastoral and 
hunting way of life due to the demands of the environment. 
















approximately the last 600 years, since Kuria and the closest South Mara language, Ngoreme, 
share seventy-eight cognates, out of a list of one hundred glosses. The new words in South Mara 
are specifically within the realm of livestock vocabulary. 28 Some of these words are loan words 
from Dadog-speakers, while others are of unknown origin. 29 Another example of these innovations 
is the changing designations for cattle colors. By the time that the Suguti, North and South Mara 
languages had separated around 1 500 A.D., each was using a set of differently innovated words 
for cattle colors. The western Serengeti innovations, referring to the basic colors of black 
(anadaburu), white (iyeru) and red (ambereretu or beriri), are derived from pre- and proto- 
Southern Nilotic roots, which appear in Dadog. The origin of other cattle colors is unknown. 30 

The cattle-word innovations in South Mara demonstrate the ongoing frontier dynamics in 
which western Serengeti peoples continued to interact with Dadog-speaking pastoralists and to 
develop their own economic adaptations to the environment. Of these South Mara innovations, 

28 Examples include: in Nata the word for bull is aheri or satima. in Ngoreme eheeri 
(derived from the proto-Southern Nilotic root eeRi for male cattle, or hirri in Dadog), while the 
East Nyanza term is -galni or -geeni (in Simbete, Kuria and Shashi); Ox or steer is ritwe, while 
the East Nyanza term in taang'ana; the word for cow in addition to the East Nyanza term - 
ha(a)BirI, Nata ahabheri, the term anyaburi is sometimes used, which can refer to mature female 
goats, sheep or wild ungulates; a young she-goat is amwali while the term in East Nyanza is - 
subiini or subeeni. A he-goat is andome while in East Nyanza the term is -gorohe, in Ngoreme it 
is egorohe. Schoenbrun, "Early History," Table 4.26; Nata, Ngoreme word lists and unpublished 
dictionaries; Muniko et al., Kuria-English Dictionary : Ehret, Southern Nilotic History . 
Appendixes D. 1 -D.4. 

29 Other Southern Nilotic loan words date to an earlier period of interaction with Mara 
Southern Nilotic speakers, for example: eesono (barren cow in Nata), risero (hide in Nata), 
risakwa, risako (sheepskin in Nata and Ngoreme), ekimano (kid, lamb in Ngoreme), and iguruki 
(ram in Kuria), entikere (donkey in Ngoreme [from Dadog, from a pre-Southern Nilotic form]), 
egorohe (he-goat in Ngoreme [from Dadog]) Ehret, Southern Nilotic History , pp. 1 30- 1 37; Nata 
culture vocabulary from Nyamaganda Magoto; Ngoreme-English Dictionary, Iramba Parish, n.d.; 
English Kikuria Dictionary, Maryknoll Language School, n.d.; Schoenbrun,"Early History " 
Table 4.26. 

30 Schoenbrun, "Early History," p. 497, Table 4.4. Comparison of cattle color words in 
Kwaya, Kuria and Nata. 

anyabori is a word used for "cow," but also for other female animals and especially for a mature 
female wild ungulate, while the word for "a young she goat," amwati, also refers to an immature 
female wild ungulate. This blurring of semantic domains, between words related to herding to 
words related to hunting, characterizes western Serengeti people's strategy of integrating and 
adapting both hunting and herding practices from different ecological zones. 

By or before 1500 A.D. the present linguistic/cultural foundations of this region were well 
established. Each of the Bantu languages that is now distinct had differentiated itself. Bantu- 
speakers coexisted on the land with Tatoga pastoralists and Asi hunter/gatherers. Although each of 
the Bantu-speaking ethnic groups today insists on its own unique culture and tradition, these 
groups share a considerable heritage, which developed out of interactions with peoples from diverse 
backgrounds in the distant past. Historical linguistics provides evidence for the interactions among 
farmers, herders and hunters that reaches back more than two millennia. Yet, as the latest 
innovations in vocabulary attest, this frontier process of interaction lasted right up through the last 
six hundred years. At that point Bantu-speakers had become dominant, not by economic 
specialization, but by diversification through adopting the expertise of their neighbors. 
The Mechanics of Incorporation in the Distant Past 

Given this evidence from the distant past the historian must ask just how the linguistic 
landscape changed so dramatically over the past millennium-from one in which speakers of many 
different languages occupied separate ecological zones and practiced different but interdependent 
subsistence economies to one in which Bantu-speaking farmers came to dominate over the smaller 
remaining groups of hunter/gatherers and herders by adopting much of the expertise of their 
neighbors. What kinds of social mechanisms were in place at the time that would have made this 
possible? We can only speculate on these processes, finding hints among the mechanisms for 
incorporation that have functioned in historical times. After providing what little evidence exists 

through historical linguistics and comparative ethnography, I will demonstrate how oral tradition 
tells a parallel story that provides insight into these processes. 

Evidence remains to show that East Nyanza-speakers inherited a bilateral descent system 
from their Great Lakes Bantu-speaking ancestors. In the environment of the east lake strong 
matrilineal tendencies developed among East Nyanza-speakers in the distant past. 3 ' Many ethnic 
groups in the Mara Region today remain matrilineal, while others have adopted patrilineal descent 
systems during the colonial years or before. Even those who emphasize the patrilineage use the 
prefix bene plus the name of an ancestor four to five generations back to refer to the level of 
segmentation in the patrilineage known as the ekehita or "door." Christine Choi Ahmed argues that 
this common Bantu lineage indicator is derived from the root word (ny)ina, meaning "a person's 
mother," making these lineages unmistakably matrilineal in origin. 32 Another interpretation of this 
root word is simply that it indicates possession that could be gendered either way.' 3 In either case, 
the evidence seems to suggest that East Nyanza-speakers had at their disposal the tools of lineality 
on either side that they could deploy as it suited their needs. 

Women in the western Serengeti today have control over their own fields and grain 
reserves, even though they live in homesteads controlled by their husbands. If women also 
controlled many of the agricultural resources when Bantu-speakers entered the unfamiliar 
environment of the western Serengeti 1 500 years ago, one might speculate that it would have been 
to the advantage of their lineages to take husbands from hunter/gatherer or pastoral communities 

31 Schoenbrun, A Green Place , pp. 177-179. 

32 Christine Choi Ahmed, "Before Eve was Eve: 2200 Years of Gendered History in East- 
Central Africa" (Ph.D. Dissertation, U.C.L.A., 1996), p. 124. 

33 Schoenbrun, personal communication-the underlying root is a simple possessive particle 
"of," -ny- (the feminine form) cannot be automatically equated with -ne. 

already in the area to expand their food producing range. If these communities emphasized the 
matrifocal household, these stranger men would have been incorporated into the homes of their 
wives, while their children remained in the community of their mother. 34 

On this inter-cultural frontier, land was plentiful and labor the key scarce resource. 
Successful communities were those able to attract new members to exploit these resources 
extensively over a large land area rather than intensively on smaller but more productive plots of 
land. Karla Poewe and others have theorized that matrilineal societies are best adapted for 
incorporating strangers and for expansion on the frontier. In a matrilineal system production is 
individual while distribution is communal. A man's sister's children inherit his wealth, rather than 
the children of his wives, whose production he controls. Those who inherit his wealth most often 
live in distant settlements. This disjuncture between the locality of production and distribution 
creates widespread networks of security through the distribution of wealth, rather than accumulates 
wealth within self-contained family units. The matrilineal system of production and distribution 
tends to be associated with abundant and unrestricted access to resources and situations of 
economic expansion. This situation demands strong networks of security because of the risky 
nature of frontier expansion in a marginal environment. A situation of scarce resources 
concentrated on productive land that people must exploit intensively favors the patrilineage with its 
ability to concentrate wealth. 35 Schoenbrun, also, argues that nondifferentiated descent ideologies 
"accompanied dispersed and mixed farming systems whose main concern lay in opening more land 

34 Also suggested as the mechanism for incorporation in other areas by Choi, "Before Eve," 
p. 1 1 ; and Schoenbrun, A Green Place , p. 1 78. 

35 Karla O. Poewe, Matrilineal Ideology: Male-Female Dynamics in Luapula. Zambia 
(London: Academic Press, 1981), pp. 3, 21, 25-6. 46-7 ; see also Choi, "Before Eve," p. 143; and 
Cynthia Brantley, " Through Ngoni Eyes: Margaret Read's Matrilineal Interpretations from 
Nyasaland," Critique of Anthropology 17, 2 (June 1997): 147-169. 

rather than in protecting specific bodies of wealth from depletion." 36 The ability to reckon descent 
through either line suggests that a flexible strategy was necessary for negotiating various kinds of 
relationships on the frontier. 

In this scenario Bantu-speaking communities incorporated Southern Nilotic or Southern 
Cushitic-speaking men who brought the skills and environmental knowledge necessary for survival 
in the western Serengeti. These newcomers would have allowed Bantu-speaking communities to 
exploit not only their own ecological niche suitable for farming but also take part in hunting and 
increased herding activities. The establishment of affinal connections would also have ensured 
consistent interaction between the two communities. As already mentioned, the evidence that East 
Nyanza-speakers adopted the cycling age-set names and other related words from Mara Southern 
Nilotic-speakers in the early phase of settlement in the region, suggests that the equivalence of age- 
peers across linguistic and economic boundaries may have allowed young men to gain acceptance 
in the community as "brothers" of their age-mates. This speculation, based on evidence from 
historical linguistics, is, in fact, echoed in the asimoka narratives, which tell the story of how 
farmers and hunters met. 

The Nata Emergence Story; The Union of Hunters and Farmers 

This is one narration of the Nata asimoka story. [For other versions of the story see 

Appendix 1 . See Figure 4-4: Narration of Nata Emergence Stories.] 

Our parents, of Nata are — Nyamunywa, he was a man — and Nyasigonko, was a 
woman. Nyamunywa was a hunter — Nyasigonko was a farmer, the woman. They met — 
- this man, Nyamunywa, shot an animal, which fell near to the field of the woman, 
Nyasigonko. The man, Nyamunywa, was thirsty. When he got to where the animal had 
fallen he saw some green grass, which is a sign of water, so he went there to look for 
water. When he got near, he saw there was a person coming out from that place. It was 
a human, like him, and the woman saw him loo. He went to her house in the cave. They 
could only speak in signs because they did not know the same language. The man asked 

36 Schoenbrun, A Green Place , p. 178. 


for water to drink. She got him some from the spring in a gourd (ekebucho). She then 
took some millet from her field and brought it to him in an elongated gourd (akena ya 
oburwe). She put it in his hand and he chewed it. It was mixed with sesame. She asked 
him, "and what do you eat? " He showed her the animal and skinned it. The man went 
outside in the bush and made afire by twirling a stick into a board using an ekengeita 
and ororende . . . shweeeeee. She got wood and they roasted the meat. They ate it. 
They took the meal home and lived in the cave of the woman. Basi (so finally), it became 
their home. The man followed the woman. They gave birth to the Nata, Nyamunywa and 
Nyasigonko. 37 

At one level the Nata emergence story provides in a concrete, locally grounded form, an 
account parallel to that based on the evidence of historical linguistics in which the union of peoples 
from diverse economic and linguistic backgrounds forms a new community. Perhaps a woman 
called Nyasigonko (from the root word, -gonka, "to suck at the breast") or a man called 
Nyamunywa (from the root word, omunywa, "mouth") never existed but, over many generations, 
countless farming women may have met hunting men, whose lineages decided that their 
cooperation would be mutually beneficial. The Nata story is the only one of the western Serengeti 
emergence stories to describe its origins in situ, without reference to ancestral migrations from 
anywhere else. It tells of the springing up, awakening, of a new people right where they are today, 
in Nata territory. 

In this story the hunter comes to live with the farmer, just as he might have if his farming 
wife was part of a Bantu-speaking society based on matrifocal residence patterns. In most versions 
of this story that I heard, elders identified the hunter as Asi and the woman as vaguely related to 
various Bantu-speaking groups such as Gusii or Sonjo. This interpretation of the Nata emergence 
story follows the model exemplified in Packard's analysis of the Bashu origin myth of Muhiyi in 

37 Interview with Jackson Benedicto Mang'oha Maginga, Mbiso, 18 March 1995 (Nata cf). 

Sochoro Khabati telling the Nata Emergence Story at Bwanda, 1 6 February 1 996 







Jackson Benedicto Mang'oha Maginga, 14 August 
1995, Narrator of Nata Emergence Story in Text 

Figure 4-4: Narration ot'Nata Emergence Stories 

eastern Zaire. He argues that an origin myth, "while not necessarily describing a specific series of 
historical events, symbolizes long-term historical processes." 38 

What is amazing about these asimoka stories is that they are locally-created 
representations of a past which is so old that it ought to have dropped out of the historical 
consciousness of the region's peoples. This past comes to the present through the imaginations of 
people who have lived in these landscapes, whose lives have been shaped by its constraints, and 
shaped also by the region's cultural resources. Because of this locally grounded quality, the 
asimoka stories account for the early interactions of hunters and fanners in a way that has rich 
cultural resonances. The stories match the shape of the best historian's hypotheses, when the 
historians reason from historical linguistics and archaeology. The stories work well as historical 
representations because they grow out of a historical imagination more richly informed of ecology 
and social possibility than any outsider's imaginations could ever be. 

This story is from Nata. It was the first story that knowledgeable elders told me when I 
began conducting formal interviews. They told this as a unique Nata story, concerning the essence 
of what it means to "be Nata." It was thus surprising to move onto formal interviews with Ikizu 
elders and hear the Ikizu asimoka stories of emergence about first woman the farmer, Nyakinywa 
(from the word kunywa, to drink), meeting first man the hunter, Isamongo. The stories were 
different with different characters but retained the same core images of a female farmer at her cave 
meeting a male hunter from the wilderness. As I conducted interviews with other ethnic groups, I 
continued to hear echoes of the same story. 

38 Packard, "The Bashu Myth of Muhiyi," p. 174. See also Packard's longer analysis i 
the book . Chiefship and Cosmology and Feierman,"The Myth of Mbegha" in The Shambaa 
Kingdom , pp. 40-64. 


I thus began to speculate on the historical reasons for the widespread occurrence of this 
story with many local variations throughout the region. One possibility would be that these are old 
core images, dating from the time before each of these groups (Nata, Ikizu, Ikoma) differentiated 
itself linguistically, sometime during the past 500 years. This line of reasoning is based on the 
distribution of these core images only among western Serengeti people, while the emergence stories 
of the Kuria to the north or Jita or Kwaya toward the lake contain a different set of core images. 
The variations elaborated around these core images among western Serengeti peoples seem to refer 
to specific events that occurred after each group differentiated itself linguistically, in many cases 
up to the period of the late nineteenth century disasters. 

On the other hand, the historian cannot treat the geographical distribution of oral traditions 
in the same way as the distribution of words or ethnographic material. 39 An arbitrary relationship 
between sound and meaning forms the basis for recognizing changes in one or the other of those 
features of language as an historical change. Oral traditions, on the other hand, are laden with 
heavy ideological freight that does not change unconsciously or arbitrarily over time. Another 
logical explanation for the regional distribution of the emergence stories is that these communities 
were continuously interacting and so shared stories that might only date to the nineteenth century. 
We cannot date these stories, in any case, because isolated communities, providing separate chains 
of evidence, did not exist. 

39 Robert Harms, Games Against Nature: An Eco-cultural History of the Nunu of 
Equatorial Africa (Cambridge: Cambridge University Press, 1 987), p. 6, dates his ethnographic 
material in this way: "The Nunu area can be subdivided into flooded forest, flooded grassland, 
riverbank, and dryland. These zones correspond with cultural subdivisions. By comparing the 
ethnography of the different micro-environments, one can distinguish the cultural traits shared by 
all of the Nunu from those that are distinct to a single environment. Traits shared by all are 
assumed to be old unless it can be demonstrated that they are recent innovations. Therefore, they 
define the more enduring and general features of the Nunu culture. In contrast, if an institution or 
practice is distinct to a certain micro-environment, we can assume that its existence or persistence 
has something to do with conditions unique to that area. It therefore represents innovation." 


However, the remarkable congruence of these stories with the processes recoverable 
through the evidence of historical linguistics and archaeology also seems to place them in a time 
frame much older than the nineteenth century. They echo the ongoing frontier processes in which 
farmers and hunters merged to form new kinds of communities and developed new kinds of 
adaptations to a harsh environment. Rather than each ethnic group entering the region separately, 
with its own history and identity already formed, these emergence stories seem to refer to the sense 
of group identity that developed locally as new kinds of communities sprouted up from the 
rhizomatous networks that preceded them. 

The core image of a hunter from the wilderness coming to found a new community is 
ubiquitous across Africa. 40 This is not surprising since emerging food-producing communities all 
across Africa faced the process of coming to terms with, or differentiating themselves from, 
preexisting hunter communities in the distant past. Nevertheless, it is significant that this variation 
of the hunter myth is the one found consistently throughout the western Serengeti. Bantu-speaking 
immigrants may have brought the hunter myth to the area but it seems to have taken on a particular 
local form as they forged new kinds of identities that owed as much to their hunting past as to their 
farming past. Although Bantu-speakers were well established in the western Serengeti by 1 000 
A.D., the interaction between hunting and farming communities would have been an ongoing 
process, not rendered altogether irrelevant to oral memory. 

Social Reproduction and Homestead Space 
At another level, an interpretation of the core spatial images of the Nata emergence story- 
that of the woman sitting at her cave and the man coming from the wilderness on a hunt-provides 

40 For the representation of hunter founders in African art see, Fritz W. Kramer, Red Fez: 
Art and Spirit Possession in Africa, trans. Malcolm Green, (London, Verso, 1993 first published 
1987), p. 16. 

insight into gender relations within the homestead as a crucial aspect of the frontier process. In the 
Nata story the woman welcomes the man into her home, while the man has no home, only a 
hunting camp in the wilderness. These core spatial images make use of the opposition between 
home and wilderness, inside and outside, female and male. A new society emerges at the contact 
point between these dynamically opposed forces. Although the house belongs to the woman in the 
story, man's presence from the wilderness domesticates the house and makes it a civilized place. 
First man brings fire to the house. Today the hearthstones are still considered the very symbol of 
home and family. No home exists without a fire burning in the hearth. 41 First man knew the secret 
of fire and, in many versions of the emergence story told the woman that he excreted the fire to 
conceal the secret of its origin. In other versions he taught the woman how to cook. [See 
Appendix 1 ] Meat from the wilderness and grain from the homestead are both necessary for 
building this new community. 

These images of an interdependent mutuality between genders, which present women in 
control of an autonomous sphere of authority and men as dependents in women's houses, conflict 
with present gender relations in which men are in absolute control of the homestead and all of its 
productive and reproductive resources. We can date the increasing emphasis on male control of the 
homestead through the patrilineage to the late precolonial and early colonial periods. In the early 
period of settlement when livestock were few and people grouped their homesteads in proximity to 
other lineage members, a woman's home was the primary unit of production and links to her family 
were a way of establishing regional networks of security. During the disasters of the late 

41 See Brad Weiss, The Making and Unmaking of the Hava Lived World: Consumption. 
Commoditi zation and Everyday Practice (Durham- Duke University Press, 1996), pp. 29-31. 51- 
52, who describes the Haya ceremony for blessing a new house which involves lighting the fire for 
the first time by the father or a senior agnate; For the Kuria see Tobisson, Family Dynamics pp 

nineteenth century the dangers of raiding, with men leaving the home to search for food in 
neighboring areas, for raids or to hunt, restricted women's movements. Without the labor of men 
and with the specter of famine the house became a restrictive site. During the early colonial period 
men recovered from famine by selling hunting products to Sukuma in exchange for livestock. As 
cattle wealth increased the man's spatial domain of the cattle corral became increasingly important. 
Nata elders can specifically remember when inheritance (mainly livestock) became patrilineal 
(through a man's sons) rather than matrilineal (through a man's sister's sons). In the new economy 
of commodities, men controlled the cash and women could not keep the proceeds from selling their 
grain. Men accumulated cattle wealth by inheritance through their sons rather than their nephews. 
The colonial record preserves the ongoing gender struggle. In 1 928 the Musoma District Officer 
reported that the Ikoma, Ishenyi and Nata "women have the men completely under their thumbs" 
and "divorce is more frequent than with most native tribes." 42 

One would expect that if the emergence stories dated from the late precolonial and early 
colonial period they would assert the increasing dominance of men over dependent women. A 
rarely told Ngoreme emergence story collected by Odhiambo Anacleti explicitly explains this 
change in gender relations. The story says that men and women once lived in separate camps. The 
women possessed the horn as their symbol of authority while the men had the drum. A child was 
born after sexual relations between youth who were herding. At the meeting to solve the problem 
of authority over the child the men offered the women a fat barren she-goat for slaughter. The men 
picked up the horn of authority and blocked all its apertures with wax while the women were 
chasing the goat. Without their authority, represented in the horn, the women went to live with the 

42 Acting D.O. Musoma to P.C. Mwanza, 10 October 1928, Monthly Report for 
September 1928, 10 March 1928, Monthly Report for February 1928, and 13 September 1927, 
Monthly Report for August 1927, 1926-29 Provincial Administration Monthly Reports, Musoma 
District, 215/P.C./1/7, TNA. 

men. 43 This was the beginning of bridewealth. Women traded their authority for meat. This story 
symbolizes a change in gender relations from that in which men and women controlled separate 
spheres of authority to that in which women were subordinate to men's authority. 

Yet the more prevalent Nata emergence story, and others like it in the other ethnic groups 
of the region, represents a set of gender relations clearly at odds with present patterns that seems to 
refer to relations of an earlier period. If, as I have argued, the emergence stories encode the 
frontier process in which settlers within diverse environments forged new kinds of communities, 
then I must take seriously (and listen closely to) its way of representing earlier forms of gender 
relations and their place in the frontier process. One way to place these core images historically is 
to investigate the gendered homestead spaces to which they refer. 

The gendered spaces of the homestead also represent the same gender relations of the 
autonomous, yet interdependent, authority of men and women, as reflected in the core images of the 
emergence stories. Within the "traditional" homestead as elders remember and, in part, practice it 
today, male space is outside, in the courtyard, while female space is inside, in the house. In East 
Nyanza languages "house" (anyumba) refers not only to the physical house but to a woman, her 
offspring, and the property and dependents attached to her. 44 Men have no houses of their own and 
sleep in the houses of their wives. Men's conversation takes place in the courtyard that surrounds 
the central cattle corral, as the symbol of male property. 

43 Anacleti, "Pastoralism and Development," pp. 189-193. I did not collect any stories 
even remotely like this one in Ngoreme or elsewhere. 

44 Robert A. LeVine, "The Gusii Family," in The Family Estate in Africa: Studies in the 
Role of Property in Fami ly Structure and Lineage Continuity , eds. Robert F. Gray and P. H. 
Gulliver (London: Routledge and Kegan Paul, 1 964), p. 70; Schoenbrun, A Green Place, p. 1 74; 
Tobisson, Family Dynamics , pp. 128-137. 

Homestead Space Over the Long-term 

Although no archaeological evidence exists for homestead layout within the western 
Serengeti itself, one can infer larger regional patterns from the work undertaken in neighboring 
areas. In western Kenya, John Sutton surveyed 1 33 archaeological sites on hillsides, known in the 
literature as "Sirikwa holes," because local Kalenjin people identified them as the cattle-enclosures 
made by former inhabitants whom they called the "Sirikwa." The "holes" consist of saucer like 
depressions, 7-25 meters wide, at a depth of up to 4 meters, surrounded by earthen or stonework 
walls. Two low banks flanked a single entrance on the downhill side, with a mound below it. 
Sutton identified circular houses around the perimeter of the hollow, and these were entered 
through the hollow. The hollow itself was a livestock corral, surrounded by a fence with a narrow 
guarded entrance. Cleaning out mud and dung daily seems to have formed the mounds outside the 
entrance, and perhaps the depression itself. One house that Sutton excavated in Chemage, Kenya, 
was divided into two sections, one for a bedroom and the other for young livestock. 
Agriculturalists who supplemented their diet with livestock seem to have inhabited the "Sirikwa 
holes." Sutton dated these remains to a period between 1600 and 1800 A.D. 45 

Sutton noted that the most striking present day analogies for the "Sirikwa holes" are the 
homestead complexes of the Kuria and Gusii in western Kenya, which consist of a central cattle 
corral surrounded by houses, set into a thorn fence. We could say the same for homestead patterns 
in the western Serengeti, which also match the interior design of the house described by Sutton. 
However, the Kuria houses include two doors and the Kuria cattle corral is not sunken. Sutton, 
therefore concluded that Kuria and Gusii homesteads cannot "be regarded as latter-day "Sirikwa 

45 Sutton, The Archae ology of the Western Highlands , pp. 50-58. 

holes." 46 He believed them to be the work of the ancestors of present-day Kalenjin peoples in the 
area, who built these homesteads to defend against wild animals and limited raids. People stopped 
building in this manner with the advent of large-scale Maasai raiding, which rendered the 
homesteads vulnerable to attack. 47 

Whether the ancestors of present-day Kalenjin or Kuria populations occupied these 
"Sirikwa holes," they do demonstrate the existence of a spatial homestead organization that is 
similar to that found throughout the western Serengeti today. Scholarly debate involves the ethnic 
identity of those who built the "Sirikwa holes" (Kuria, Kalenjin or Maasai), 48 which seems 
anachronistic, given the complex and heterogeneous history of this region. From the distribution of 
Southern Nilotic-speaking peoples reconstructed through the methods of historical linguistics we 
know that Mara Southern Nilotic-speakers occupied the whole region from what is now western 
Kenya through the Mara Region of what is now Tanzania before 1000 A.D. 4 ' The same Southern 
Nilotic cycling age-set names used in the western Serengeti are found among the Kalenjin who live 
in the "Sirikwa hole" area today. In a later re-excavation at Hyrax Hill in Kenya, Sutton looked 
more closely at the house sites to determine the mode of production. He concluded that the Sirikwa 
chronology should be pushed back to the beginning of this millennium because of considerable 
cultural continuity with the Elmenteitan Industry, which has also been correlated with Southern 
Nilotic-speakers. 50 

46 Ibid, p. 62. 

47 Ibid, pp. 60-63 

48 For a discussion of this debate see Distefano, "Precolonial History," p. 127. 

49 Seethe analysis of Kalenjin and Luyia languages in terms of the underlying Southern 
Nilotic component, Distefano, "Precolonial History," p. 141. 

50 Sutton, The Archaeology of the Western Highlands , p. 22. 


A comparison of homestead and interior house diagrams from the Nata, Ikoma and 
Ngoreme within the western Serengeti, with other Mara speakers, the Kuria and Gusii, 
demonstrates cultural continuity with these archaeological remains. The chart on the following 
page illustrates this pattern in contrast to the homestead in neighboring Sukuma, which they also 
build around the central cattle corral. [See Figure 4-5 and 4-6: Homestead Layout and Interior 
House Designs.]. The Nata, Kuria and Ikoma homestead pattern consists of houses built right into 
the corral fence. In all cases they divide the house into at least two rooms, with the outer room 
being used for small stock. Building around a central cattle corral is also a common pattern 
recognized among southern Bantu-speakers." 

Historical linguistics also provides some evidence about early homestead patterns. Great 
Lakes Bantu-speakers (C. 500 B.C.) apparently began building very different kinds of homesteads 
than the ones built by their linguistic predecessors in the forests of the west, which were square, 
with paneled and gabled roofs. Distinctive features among Great Lakes Bantu-speakers were 
round houses with thatched roofs surrounded by tall fences with a main gateway." 

The innovation of building around a central cattle corral seems to date to the time when 
East Nyanza-speakers adopted a new term for a unilineal, dispersed, exogamous group, the eka, 
between 500 and 1000 A.D. This word is derived from the older, Great Lakes Bantu word for 
head of cattle (nka, itself a Sudanic loan word meaning "homestead."). Western Serengeti people 
use this term not for the lineage but for the homestead itself built around the cattle corral, aka. 

The designation of the house as female space, leaving the man without a house, may also 
date to this period when East Nyanza-speakers began moving into the interior. The "house" 

51 Adam Kuper, "Symbolic Dimensions of the Southern Bantu Homestead," Africa 50 1 
(1980): 8-22. 

52 Schoenbrun, A Green Place , pp. 159-160. 



(anyumba), as a term that designates a woman and her dependents, is common among all East 
Nyanza languages." Ethnographers of the Kuria and the Kwaya report that men do not have a 
house and that the courtyard is man's space. Unmarried men sleep together in a house without a 
hearth. 54 This dating, however, remains speculative, since I have no way of knowing, without 
more linguistic evidence, whether the term "house" was gendered in the same way that it is in more 
recent times. Overwhelming ethnographic evidence of gendered homestead space, finally, cannot 
prove how western Serengeti people gendered space in the distant past. It does, however, suggest a 
culturally nuanced way of interpreting the core images of emergence stories that refer to gendered 
homestead space. 
The Ethnography of Homestead Space 

Most of the words used to describe the homestead are common among speakers of all East 
Nyanza languages. The words for cattle pen and yard, recorded here in Nata, seem to be western 
Serengeti innovations. The homestead pattern remembered by western Serengeti elders today and 
still observable in some homesteads is one in which each wife in a plural marriage has her own 
house (anyumba) and grain storage bin (egitara). The man has his own grain storage bin, which is 
kept for emergencies and his own needs for feasting and exchange. The homestead itself (aka) 
consists of a yard (ribancha) enclosed by a brush fence (orubago)." In the center of the yard is 

53 Ibid, p. 174. 

54 Tobisson, Family Dynamics, pp. 128-132, says that Kuria men have no house and that 
the youth's hut has no sections or hearth as a woman's would. Huber, Marriage and Family , pp. 
62-68, reports that there was no hut for the man and that the yard fire was the man's space in the 
homestead. A ritual was necessary for the construction of the homestead gate-way. 

" Orubago is derived from the Great Lakes Bantu root -go(o) "enclosure" which marked 
"the regional appearances of the homestead layout so common throughout Bantu-speaking eastern 
and southern Africa." Schoenbrun, A Green Place . Cultural Vocabulary #26. 

the livestock corral (asimoora) ie and around the perimeter of the yard, the houses of wives and 
sons. The first wife's house is located nearest the gate on the right going in (or in the center back), 
the second wife to the left of the gate and so on. The house for young unmarried boys (amachi) is 
found opposite the gate, or guarding the gate. Only one gate or entrancefefe/Hto) breaches the 
homestead fence and it is closed with a log (egeshoko). If a son marries and has his own 
circumcised children in the homestead they cut another gate for him." 

The gate, the courtyard and the cattle corral are male space and symbolize male (external) 
authority while the house symbolizes female (internal) authority. Narratives about the past 
described girls' circumcision inside the homestead at the site of the grain storage bins, to symbolize 
their future work as married women who would provide food to their families. Boys' circumcision 
takes place outside the homestead under a tree to symbolize their work in conquering the bush as 
hunters. 58 Tobisson argues that for the Kuria the log of wood to shut the livestock corral was the 
most important symbolic marker of male resource control used in rituals of all kinds." Kinship 
terms also refer to these gendered spaces of the homestead. Nata call a person's maternal line the 
anyumba (house) and Ikizu call it the rigiha (hearthstones), while Nata and Ikoma call the paternal 
line the ekehita (homestead gateway) and Ikizu call it the ekeshoka (gatepost). 

The house itself is divided into two areas, the omuryango, or the outer room for sheep, 
goats and young calves to sleep in, or for a sitting room and the inner room or sleeping room where 

56 Asimoora is a Southern Nilotic loanword, Christopher Ehret, personal communication. 

57 Nata homesteads from an interview with Mayani Magoto, Bugerera, 1 8 February 1 995 
(Nata cf). Ikoma and Ngoreme homesteads in Edward Conway Baker, Tanganyika Papers . See 
Tobisson, Family Dynamics , pp. 129-133 for the Kuria homestead. 

58 Interview with Bhoke Rotegenga (Nata ¥ ) and Mgoye Rotegenga Megasa (Nata <?), 
Motokeri, 13 March 1995. 

59 Tobisson, Family Dynamics , pp. 147-148. 


a cooking fire is kept. Kuria call the outer room for livestock the eheero, while they call the 
domestic section of the house with the cooking fire a variation of the word used in Nata for the 
outer room, the omorengo. 60 In Gusii the outer room, the eero, is specifically designated as the 
"husband's room" by LeVine and LeVine. 61 It seems significant that among the western Serengeti 
peoples the eheero does not exist and they instead interpret the woman's domestic section of the 
omorengo as the small stock room, making the whole house female space. The inner room is 
closed to casual visitors, with its outside door (kyawisiko) used only for emergency escape or to 
remove a corpse. Guests always enter the house through the front door (hvawibancha), which 
leads from the courtyard into the omuryango where the small stock are kept. 

Because livestock are the most important form of convertible wealth, the corral in the 
center of the homestead as male space is an important symbol of male authority over livestock 
production and reproduction in the homestead. People equate cattle with reproduction because of 
their use as bridewealth and as markers of other kinds of social exchange. The cattle corral is the 
link between the individual homestead and the larger descent group. 62 Similarly, Rigby shows for 
the agriculturally-based Gogo that " effective kinship and affinal relations are primarily viewed in 
terms of the carrying out of certain jurally defined rights and obligations concerning the most 
valued form of property, viz., livestock."" The corral is located in the center of the homestead in 

60 Miroslava Prazak, "Cultural Expressions of Socioeconomic Differentiation among the 
Kuria of Kenya" (Ph.D. Dissertation, Yale University, 1992), p. 123. Muniko et al., Kuria-English 
Dictionary , this dictionary also defines another word, igiiume as "a man's private house in a 
homestead, hut." 

61 LeVine and LeVine, "House Design," p. 159. 

62 Tobisson, Family Dynamics pp. 150-151. 

63 Peter Rigby, Cattle and Kinship among the Goeo: A Seminastoral Society of Central 
Tanzania flthaca- Cornell University Press, 1969), pp. 1-2. 

contrast to the position of wives' houses on the periphery. Some important men choose burial in 
their livestock corral rather than in front of the house. 

Although a woman is incorporated into a man's homestead and lineage through bridewealth 
in the present patrilineally dominated system, a woman can still exert power through control over 
the house as her locus of authority. Women as farmers of millet and producers of children strongly 
influence a man's prosperity and prestige in the community. Children identify themselves primarily 
by the "house" of their mother, using the inside possessive prefix mwa- (mwaKimori, "of or inside 
Kimori's house"). People often carry these "house" designations to the second or third generation. 
Whatever the formal lineage organization, households have a strong matrifocal orientation. A 
woman has primary responsibility for feeding herself and her family and has fairly autonomous 
control over her own resources to do so. 64 

The independence of a woman's house alludes to another institution of long duration in 
East Africa, known in the anthropological literature as the "house-property complex" or the 
division of family wealth and inheritance according to the "house" of each wife in a polygynous 
family. Each "house," consisting of a woman and her children, controls certain livestock used for 
milking and meat, and paying the bridewealth for its sons to marry. Men allot these livestock to 
each wife at marriage. A woman's "house" gains more cattle by trade, gift or, especially, as 
daughters of the "house" marry and "the house" receives the bridewealth. In the house-property 
system of the western Serengeti men, as heads of homesteads, retained more flexibility to move 
property between "houses" than did their Kuria and Gusii neighbors to the north. Hakansson 
hypothesizes that this more centralized system is found in "high risk environments," like the 
western Serengeti, where periodic disaster demands more interdependence among the houses of one 

64 Tobisson, Family Dynamics, p. 1 47. argues for the pivotal position of the maternal 
house among the Kuria. 


homestead. 65 In either case the system of independent "houses," controlling their own property, 
gave women an extremely strong position in the negotiation of family affairs. In spite of these 
autonomous spheres of control, men and women must ultimately work in unison for the family to 
function. 66 

Both men and women use the powerful metaphor of the "house" today, just as we have 
reason to believe they would have used it in the past, in the struggles between men and women on a 
daily basis. The ultimate sanction that a woman could exert over a man was to deny him access to 
her house or leave her house to find sanctuary in another man's homestead. 67 One elderly Nata 
woman complained that even if a woman had her own house and grain stores her husband, as head 
of the homestead, would give orders for work each day and make decisions about resource use 
without his wives' opinion. She related the story of her father's first wife who butchered a goat to 
eat (trespassing on the male space) during the hunger while he was away getting grain in Zanaki. 
Someone who saw signs of the goat in her inner room attic (trespassing on female space) 
discovered her act and so she left for her natal kin Zanaki. When her husband came to get her the 
youth of the homestead defended her right to sanctuary with their spears but were finally persuaded 
to let her husband take her home. When they got home, he cut off her ears and never entered her 
house again. Nevertheless, as the last word in the story, when he died, his body had to be taken out 
of the back door, from the inner room of her house, as his first wife, for burial. 68 

65 T. Hakansson, "Family Structure, Bridewealth, and Environment in Eastern Africa: A 
Comparative Study of the House-Property Systems/Ethnology 28 (1989): 119. 

66 Discussed by Regina Smith Oboler, "The House Property Complex and African Social 
Organisation," Africa 64, 3 (1994): 344, 351. 

" Interview with Tetere Tumbo, Mbiso, 5 April 1995 (Nata cC). 

68 Interview with Bhosa Rugatiri, Mbiso, 17 June 1995 (Nata ?). 

Oral Traditions and Gendered Space 

The emergence story also represents the gendered spatial organization of the homestead by 
first man leaving his nomadic camp and coming to live in the enclosed space of the cave, which 
remained the woman's house. A Nata elder told me that the man, Nyamunywa, was an omolware, 
a term for a man who goes to live in an independent woman's (omosimbe) homestead and is 
married by her (passive), often called a "male wife." 69 He could never be the head of the 
homestead and was beholden to her goodwill as she could ask him to leave at anytime. In the Ikizu 
story Nyakinywa refuses to marry Isamongo but he stays in her house.™ 

Evidence for homestead space in the distant past from archaeology and linguistics, 
evidence for homestead space in the recent past from ethnography, and the representation of 
homestead space in oral traditions of emergence all present parallel accounts of gender relations in 
which men and women control autonomous yet interdependent spheres of authority in the 
homestead. In the preceding section 1 presented the arguments for assigning these emergence 
stories to the time frame of the tongue duree, reflecting the strategies of settlers on the frontier who 
created new kinds of communities from a diverse social environment. In this section I have shown 
that the core images of these traditions represent a spatial homestead organization that was in place 
by or before 1000 A.D. and continuing, at least in memory, to the present. Although I cannot 
conclusively date the gendering of this homestead space without more linguistic evidence, the 
overwhelming ethnographic evidence suggests that it is also a pattern of long duration. All these 
forms of evidence present a picture of gender relations that is clearly at odds with the increasing 

69 Interview with Sochora Kabati, Nyichoka, 2 June 1 995 (Nata <f). See Chapter 2 for a 
discussion of the omosimbe position. 

70 Note that in the story told by Megasa Mokiri, Motokeri, 4 March 1995 (Nata <r) in 
Appendix 1 , he asserts, in contradiction to all other accounts, that first woman went to live in first 
man's hunting shelter. 

emphasis on the subordination of women to male control of homestead resources that can be dated 
to the late-nineteenth century. Although the argument is not conclusive, the evidence points to a 
pattern of gender relations in the distant past that is different from that which exists today. 

If the gender relations symbolized in the emergence stories form part of the frontier 
process, what crucial strategies might they represent? The spatial metaphor of the male hunter 
moving into the female farmer's home might literally symbolize the historical process by which 
stranger husbands moved into the matrifocal houses of their wives and were incorporated into the 
expanding community. Of course not all farming wives married hunter husbands but the 
emergence traditions might encode this process because it was of crucial importance on the 
frontier. Matrilineages could have enhanced their widespread networks of security by 
incorporating hunter men, and their networks of security, into farming communities. Incorporation 
of hunting men into a matrifocal house ensured that the children of that union stayed in the farming 
community. This strategy would help to explain the success of Bantu-speakers to expand and 
eventually dominate in a region once shared by peoples of many different linguistic groups. 
Another possibility is that the assignment of autonomous spheres of authority to men and women 
within the household was simply a more efficient way of mastering the range of skills necessary to 
diversify the domestic economy on the frontier. 

However gender relations contributed to the frontier process, these strategies were not 
necessarily consensual or automatically assumed. Poewe argues that in a matrilineal descent 
system a form of "sexual parallelism" exists in which men and women control entirely separate and 
distinct resources. However, in a bilateral (nondifferentiated) descent system, gender relations of 
"reciprocal-dependence" exist in which the separate male and female spheres of resource control 
are mutually dependent on each other for ongoing social reproduction. This produces a precarious 
situation in which men and women constantly negotiate and contest the political and economic 


affairs of each gender." Another variation of the emergence story illustrates the contested nature 

of male and female domains of authority employed in social reproduction. 

The Ikizu Emergence Story: The Division of Authority 

The emergence stories use fire and water to symbolize the different kinds of authority 

controlled by men and women and the interdependence of these spheres. Man's secret was fire. 

while woman lived near a spring and supplied water. Many rituals and narratives also use the 

symbols of water and fire as transformative substances of power. 72 A look at another regional 

variation of the emergence story provides a clearer understanding of the spheres of authority 

represented by these gendered domains. In the Ikizu version first woman's secret was water, or 

rain, while man's was fire. The woman won in a contest between water and fire and took authority 

as rainmaker "chief over Ikizu. [See Appendix 2 for other versions of the story.] 

Nyakinywa went by herself (from Kanadi in Sukuma) to the cave ofGaka. When she 
entered the cave ofGaka and stepped on the rock it cried like a drum. That is how she 
knew that she had come to the end of her journey. She laid down her bundles in her new 
house prophesied by her father. The next day she went outside to look at her 
surroundings when she saw some smoke. She found out that the smoke was coming from 
the hill ofSombayo. She went to find it, over the River Kibangi, and to the cave at 
Sombayo. There she met a man named Isamongo who lived in the cave and was of the 
clan ofMuriho. Isamongo asked his guest where she had come from. She said she came 
from her house to see where the smoke was coming from. He asked to see her house, so 
they went to Gaka, to Nyakinywa's place, where he asked her to be his wife. She refused 
but they lived there together as lovers. Those two people each had their own area of 
expertise. Isamongo had the secret of making fire and Nyakinywa had the secret of 
water, that is she could bring rain. Each asked the other to show them their expertise, 
but they could not agree. Things went on like this until one day Isamongo went out to 
hunt. Nyakinywa brought a big rain which completely soaked Isamongo out in the bush. 
Back at home she had put out the fire. When he returned, cold and wet, he found the fire 
out in the cave at Gaka. So Isamongo had to show Nyakinywa how to make fire. He took 

71 Poewe, Matrlineal Ideology , pp. 21-26. 

72 For a discussion of the symbols of fire and water see Anita Jacobson-Widding, 
"Encounter in the Water Mirror," in Body and Space: Symbolic Models of Unity and Division in 
African Cosmology and Experience, ed. Anita Jacobson-Widding (Uppsala: Almqvist and Wiksell 
International, 1990), pp. 177-216. 


out a board and a stick that he twirled until he made fire. So you see, he had already 
shown her the secret of fire. He then asked Nyakinywa to show him the secret of rain. 
She did not refuse but asked him to first bring her sister Wang'ombe from Hunyari. 
Then he should go to the bush and kill a bushbuck, skin it and bring the skin to her. 
Isamongo did all that she asked. The next day she asked him to make the pegs to stretch 
out the hide. When this was done the three of them left with the skin and the pegs and 
went to the pool at Nyambogo. When they arrived. Isamongo was asked to peg out the 
skin on top of the pool. He tried and failed. Nyakinywa tried it and succeeded. 
Isamongo asked her to do it again but she refused. Wang'ombe asked to try and she also 
succeeded in pegging out the skin on top of the pool, like her sister Nyakinywa. So 
Isamongo was told that he had failed the test and would not be shown the secret of rain. 
They returned to their house at Gaka. n 

Although both the Nata and Ikizu emergence stories are basic to the corpus of men's 
historical knowledge, embedded in them is the contested nature of men's and women's autonomous 
but reciprocal spheres of authority. In this narrative the interaction between first man and first 
woman explains the establishment of ritual authority in Ikizu over rainmaking and prophecy. Ikizu 
recognize Nyakiny wa's line as the chiefly line of rainmakers while Isamongo's is the line of 
prophets. The female power of water and fertility triumphed in this story but only with concession 
and compromise. To become the mtemi or "chief of Ikizu she had to agree to certain provisions 
laid down by Isamongo-that Ikizu would retain its basic social institutions of circumcised age-sets, 
generation-sets and eldership ranks. An alternate version of the Ikizu emergence story is told in 
the next chapter. In this story Muriho, the hunter, conquers the malignant spirits of the land and 
establishes Ikizu by marrying a local woman and settling at Chamuriho Mountain. Those who put 
the two emergence stories together say that Muriho was the ancestor of Isamongo and represented 
the original Ikizu people. 

In the story of Nyakinywa and Isamongo the interactions of first man and first woman, 
based on the founding myth, became the vehicle for telling a new story about changing political 
authority, while the spatial arrangement of the homestead became the site of these gendered 

73 P.M. Mturi and S. Sasora, "Historia ya Ikizu na Sizaki," unpublished mss, 1995. 

negotiations. Far from representing women in a subordinate position, in this version woman's 
authority wins the contest, while recognizing the need for mutuality. 

This version of the emergence story reminds us that the space of women's authority is not 
only their marital house but also their natal house, as discussed in Chapter 2. Western Serengeti 
emergence stories emphasize the role of women as wives. Yet in the Ikizu version Nyakinywa's 
sister, Wang'ombe, joins the struggle of first man and first woman. In other versions Nyakinywa 
leaves Kanadi with two sisters. This version introduces Nyakinywa as the daughter of a chief. 
The characters of founding myths in many places throughout Africa represent the genealogical 
prototypes of parts of the kinship system. 74 The sisters in the Ikizu story would seem to indicate 
the strength of the matrilineage, while the Ikoma and Ngoreme stories that are considered in the 
next chapter emphasize the brothers of a patrilineage. 
The Historical Basis of the Ikizu Emergence Story 

Although rainmakers from the Kwaya clan may have been practicing in Ikizu for a long 
time, it was only in the late nineteenth century, in response to the disasters, that Ikizu was 
consolidated as a political entity under the centralized leadership of one rainmaker. In the mlemi 
(chief) list of Nyakinywa's descendants, discussed in the Chapter 6, the first mlemi in the list that 
can be historically dated ruled in the last quarter of the nineteenth century. A similar pattern dates 
the Sizaki chiefship to this period from Sukuma sources. 75 The Ikizu developed this new group 
identity not through assimilation into Sukuma (Kanadi) society but rather through popular 
acceptance of limited ritual authority for the descendants of Nyakinywa as rainmakers. Individual 

74 See Karen Sacks, Sisters and Wives: The Past and Future of Sexual Equality (Westport, 
Conn.: Greenwood Press, 1979). 

75 Gregory Bugomora, Lumuli . 5 August 1949. Lumuli was a White Fathers' Sukuma 
language newsletter of the church. 


Sukuma immigrants seem to have played an important role in the political realignment of the time 
but their numbers were not significant enough to bring lasting cultural or linguistic changes. 
Although colonial officer and anthropologist, E. C. Baker, used this story as evidence to prove the 
Sukuma origins of Ikizu, no other evidence of mass migration from or assimilation into Sukuma 

Because Ikizu retained its western Serengeti culture and language, narrators might have 
adapted the earlier emergence story, similar to that of the Nata, to take into consideration the new 
authority from Sukuma. The Ikizu story about political authority unfolds in the gendered 
homestead spaces because these are the core spatial images of the older founding myth on which it 
was based. By coopting the older story as their story, the Kwaya clan of Nyakinywa gained local 
legitimacy. Many scholars have documented a similar process by which outsiders gained access to 
inside authority by appropriating cultural symbols in the west Lakes region of Buhaya and 
Bunyoro. 76 In the Ikizu example, the outsiders gained authority by inserting themselves into the 
founding myth as the woman who owns the house but does not marry the man. In this society, 
women's power over water, rain and fertility (and thus reproduction) is something that can be used 
as a tool to gain political power. The first four rainmakers of Ikizu in Nyakinywa's line were 
women, after which men usurped the title. This subordination of women's authority to men's 
suggests general trends in the late nineteenth century, which I explore in later chapters. 

76 A good example of this is in the historical analysis of Bunyoro see, Iris Berger, Religion 
and Resistance: East African Kingdoms in the Precolonial Period (Tervuren, Belgium: Musee 
Royal de L'Afrique Centrale, Annales Series, 1 98 1 ); Iris Berger and Carole Buchanan, "The 
Cwezi Cult and the History of Western Uganda," in East African Cultural History , ed. Jospeh T. 
Gallagher, (Syracuse, New York: Maxwell School of Citizenship and Public Affairs, Syracuse 
University, 1976); Peter Schmidt, Historical Archaeology: A Structural Approach in an African 
Culture (Westport, Conn.: Greenwood Press, 1978); Tantala, "The Early History of Kitara"; 
Newbury, King and Clans : Randall Packard, "Debating in a Common Idiom: Variant Traditions of 
Genesis among the BaShu of Eastern Zaire," in The African Frontier , ed. Igor Kopytoff 
(Bloomington: Indian University Press, 1987), pp. 148-161 . 


The Gendered Division of Labor 

This reading of the emergence stories in terms of gendered and autonomous spheres of 
labor and authority provides a further possible parallel to the processes of community formation 
on the frontier that incorporated the knowledge and economic strategies of both farmers and 
hunters. In their representation of the gendered division of labor, both the Nata and the Ikizu 
emergence stories demonstrate that men's and women's autonomous spheres of authority depend on 
mutuality. First woman grows millet. She is a farmer. First man provides meat. He is a hunter. 
In the daily routine of work, women of one homestead do most of the farm work; they control their 
own fields and jointly farm their husband's field. Although colonial rumors existed of women 
hunters, elders characterized hunting as exclusively men's work." While women sometimes took 
their turn at tending livestock, men also controlled this domain. 78 

In spite of these clear gendered divisions of labor, all those I talked with agreed that, in 
practice, both men and women farm. Ikizu elders declared that both men and women share all 
aspects of farming equally and that farming was never considered as women's work alone. 7 ' Ikoma 
elders in the east placed more emphasis on hunting as men's work and farming as women's work, 
although men farmed in the wet season and hunted in the dry season. A young man who wanted to 
marry, first had to harvest from his own fields and fill his grain storage bin. 80 An Ishenyi elder 
said that a man checking out the character of a boy who wanted to marry his daughter would look 

77 From the Game Warden, Kilossa, to the Honorable Secretary in Chief. Dar es Salaam, 
26 February, 1924, p. 62, 2I5/P.C7 14/1, vol. 1 ,TNA, "the craze has gone so far that there are 
even women who spend their entire time hunting, two of these dusky Dianas being particularly 

78 See Tobisson, Family Dynamics , p. 5 1 , on gendered division of labor among the Kuria. 

79 Interview with Kiyarata Mzumari, Mariwanda, 8 July 1 995 (Ikizu <f ). 

80 Interview with Pastor Wilson Shanyangi Machota, Morotonga, 12 July 1995 (Ikoma tf). 

first at what kind of farmer he was and, secondly, how well he looked after his own parents. 81 
Daily practice in which men and women cooperate and participate in the same tasks mediates the 
separate economic realms of men and women asserted in the founding myth. 

The life-cycle and the seasonal cycle determine in key ways the division of economic tasks. 
Young men and women farm together without differentiation, except during the dry season when 
young men frequently engage in hunting or chase cattle raiders for long periods at a time. A 
mature man is less likely to spend time farming as he grows older and elderly women retire from 
farming if economically feasible. Gendered domains are also dependent on age and elderly women 
more easily enter male spaces. Young men and young women maintain an equality in work 
relations that is subordinate to elders of either gender. 82 Nevertheless, while both men and women 
farm, women are more likely than men to farm on a daily basis from their youth to old age. 

If first woman and first man are in some way symbolic of the larger processes in which 
farmers and hunters collectively formed new societies from the contributions of each, then the 
assertion of their mutuality rather than one's dominance over the other is significant. Women 
manage and control the agricultural production of their own fields, fields that provide most of the 
family's food. People placed even more importance and value on agricultural production in the 
past. Men acquired the large livestock herds that dominate the western Serengeti today during the 
early colonial period, as I discuss in Chapter 10. Many elders said that even in the memory of their 
grandparents, livestock were few, counted mainly in goats and sheep. They exchanged bridewealth 
in wild animal skins, hoes, and salt rather than in cattle, as is the practice today. In matrilocal 

81 Interview with Morigo Mchombocho Nyarobi, Issenye, 28 September 1995 (Ishenyi <?). 

For an analysis of age and gender in labor patterns and community organization see 
Elias C. Mandala, Work and Control in a Peasant Economy: A History of the Lower Tchiri Valley 
in Malawi. 1859-1960 tMadisnn- University of Wisconsin Press, 1990). 


communities, bride-service was the norm, until recently. The prestige value of meat also seems to 
have increased since the late nineteenth century with the introduction of new eldership titles gained 
by feasting. 

In the past the community seems to have recognized a gendered division of equally valued 
labor, with women controlling agricultural production and men pastoral and hunting resources. 
Yet in relation to the communities that surround them Bantu-speaking peoples ascribe to 
themselves the corporate identity of "farmers," Rema. They distinguish themselves from the Nyika 
(people of the wilderness) or the Bisa (enemies), whom they identify as those people who do not 
farm-the Asi hunter/gatherers, Tatoga and Maasai herders. In the emergence stories the woman's 
home of cultivated millet fields is the space of civilization. Economic subsistence patterns are an 
important way in which people differentiate themselves from others. 

Oral narratives metaphorically extend the gendered tasks of the homestead (farming, 
hunting and herding) to conceptualize relationships beyond the community with other farmers, 
hunters and herders. This identification of the entire community with the female domain of farming 
runs parallel to the evidence of historical linguistics, comparative ethnography and oral traditions 
that seem to indicate a process by which matrifocally organized farming communities retained their 
own identity while assimilating preexisting hunting communities. In the same way as men and 
women each had their own tasks and realms of authority within the homestead, so farmers, hunters 
and herders each occupied their own place within a regional economic system of interdependence. 


Thus, on one level, the core spatial images of the emergence stories refer to gendered 
homestead space and tell us something about the interactions of men and women in the past that 
made production and reproduction on the frontier possible. In this regard the symbols of female 
control over the house and water and male control over the courtyard and fire represent separate 

spheres of gendered authority, each powerful in its own way. The evidence of historical linguistics 
and comparative ethnography shows that these are long-standing generative principles in the 
region. Gendered household space represents the generative principles through which people 
organized household production and social reproduction on the frontier. 

Bantu-speakers adapted to the new environment they experienced in the western Serengeti 
by learning from and incorporating peoples that they found already in the area. The new kinds of 
homesteads built by these settlers established the productive and reproductive basis for their 
eventual dominance of the region. In an environment of abundant and extensive resources non 
differentiated lineages provided wide social networks and the means for incorporating strangers. 

This chapter provides a view of the emergence stories from the internal perspective of 
homestead space, gendered relations and the organization of production within the community. In 
the next chapter I examine other versions of the emergence story in terms of the external relations 
of Bantu-speaking hill farmers with Asi hunters and Tatoga herders. The core images of these 
stories move from internal homestead space to external ecological space. Oral traditions of 
emergence contain both sets of core images, but some seem to emphasize one aspect over the other. 
The asimoka stories continue to be so effective because we can interpret them in various ways 
without exhausting their rich cultural meaning. 




The western Serengeti emergence traditions also evoke the interdependent ecological 
spaces of woodlands, hills and grasslands of this region. Within these ecological spaces early 
settlers fashioned economic subsistence patterns that took advantage of each niche from the 
position of the hills, in close proximity to the woodlands and grasslands. The congruence of 
regional ecologies with the spaces of farmers and hunters in the Ikoma, Ishenyi and Ngoreme 
emergence stories provides insight into the possible subsistence patterns of different communities 
on the hills, in the woodlands and on the grasslands and ways in which farmers used the hill 
position to gain rights to the land and develop a new combination of economic strategies. Tatoga 
stories adopted by the Ikoma and Ikizu shed light on the changing relationship between farmers and 
herders in the past. Yet these ecological/economic patterns represented in oral tradition do not 
represent an environmental determinism. ' 

Over time people have used and imagined the same ecological spaces in different ways. 
Before Bantu-speakers or Southern Nilotic-speakers ever entered the region, Eastern Sahelian- 
speaking agro-pastoral ists and Southern Cushitic-speaking pastoralists, along with Khoisan- 
speaking hunter/gatherers each occupied a different ecological niche within an interdependent 
economic system. Bantu-speakers broke down these economic barriers that confined each group 
to a distinct ecology by practicing farming, hunting and herding. They settled in the hills, best 

' s ^ Harms. Games Aga inst Nature , for an analysis of the interaction of people with their 
environment, especially "Conclusion: Nature and Culture," pp. 243-256. 


suited for grain farming and, at first, learned from and developed interdependent economies with 
woodland hunters and grasslands herders. Yet as time went on they increasingly became dominant 
in the region by encroaching on the ecological spaces of both hunters and herders, pushing them 
back into more marginal areas. The last remnants of this interdependent economic system 
collapsed as a result of the late nineteenth century diasters when Maasai raiders forced farmers out 
of the eastern hills, Asi hunters moved east to become Maasai clients, and Tatoga herders relocated 
south as Maasai raiders came to dominate the Serengeti plains. As Maasai strength declined after 
the rinderpest panzootic the hill farmers became commercial hunters and wealthy livestock owners, 
dominating all three ecological zones in the western Serengeti. 

In spite of these vast changes in the regional economy the core spatial images of the 
emergence traditions still seem to refer to the earlier patterns of interaction between hill farmers, 
woodland hunters and grassland herders. In its exploration of the ecological spaces of the 
emergence traditions, this chapter shows how western Serengeti people shaped and were shaped by 
the landscapes in which they lived. Historical linguistics, archaeology, and ecological studies 
provide parallel evidence for understanding the historical development of relationships between 
peoples practicing different economic subsistence patterns in the region. People created the social 
identities of farmer, hunter and herder in the distant past to organize a regional economy. Although 
early Bantu-speakers also hunted and herded they saw themselves as farmers within this regional 
system based on difference. 

The Ecological Landscapes of Interaction on the Frontier 

Ikoma, Ishenyi and Ngoreme stories add another dimension to the analysis of the 
emergence of new communities on the frontier in the distant past. The core images still concern the 
interaction of farmers, hunters and herders but they seem to refer to the external relationships 
between communities of different economic subsistence patterns, rather than the internal division of 

homestead labor. The spatial images are more broadly the ecological spaces of the known 
landscape rather than gendered homestead space. This aspect of the core spatial images of the 
emergence stories is also present, though not as apparent, in theNata and lkizu stories considered 
in the last chapter. 

The asimoka traditions represent the interactions of farmers, hunters and herders in a 
regional economic system of interdependence that is amazingly similar to that reconstructed 
through historical linguistics and ecological evidence. Oral traditions suggest possible 
explanations, expressed through local historical consciousness, of how and why this regional 
system of interdependence worked in the past and how Bantu-speaking farmers eventually became 
the dominant players as they reconfigured the entire system. A reading of the core spatial images 
of these traditions, however, must proceed alongside other kinds of sources that provide evidence 
for the ways in which these processes may have operated in the past. All these forms of evidence 
seem to indicate that Bantu-speaking farmers founded successful communities on the frontier and 
out-competed those who taught them how to do it by situating themselves at a favorable position to 
take advantage of other kinds of subsistence patterns and interaction with peoples in other 
ecological niches. 
Ngoreme and Ikoma Emergence Stories 

The Ngoreme and Ikoma versions of the asimoka story are overtly concerned with their 

ancestral migration from Sonjo to settle in their present homes. One version of the Ikoma story 

goes like this: [See Appendix 3 for other versions of the Ikoma origin story.] 

A Msonjo came from Sonjo to hunt. He got lost and went farther to the west and 
rested under an omokoma tree. His name became Mwikoma. He came with his bow and 
arrows and when he got lost he slept under the huge tree that was in the bush. The limbs 
spread out like a house, providing shelter inside. This was at the place called Chengero. 
He killed an animal, skinned it, made afire and ate it under the tree. This then became 
his house and his camp. He would go out to hunt and return here at night. After a while 
he became aware that other people lived in the area. He went to their camps to talk with 


them but they could not understand each other because they spoke different languages. 
He invited a woman to his camp. She only ate grains or porridge and he gave her meat 
to eat. She was amazed and thought how she had only had porridge by itself and how 
good this was. Thus, they began to get to know each other. He said, "I am Mwikoma. " 
They began to live together and then got married, settling among those people who were 
already there. 

He went on and married a second wife, they had children. Then he married a 
third wife. Soon his children had grown up and were adults themselves. Each went off 
in a different direction but Mzee Mwikoma stayed back in Chengero with his first wife 
Nyabaikoma. His second wife's name was Nyabangoreme and they lived around 
Pangwesi Mountain. The third wife's name was Nyabaishenyi and they lived around 
Paori. They separated from each other and multiplied. Thus, today the Ikoma, Ngoreme 
and Ishenyi are one group, one thing, they are from one family. The Nata on the other 
hand come from the Ikizu and the Ikizu come from Sukuma 2 

The core images remain consistent with those explored in the last chapter, hunter meets 
farmer and goes to live with her. Here, first woman comes from a preexisting community that 
incorporates first man. His wives, as founders of "houses," become the ancestresses of the three 
related ethnic groups. Ikoma, Ishenyi and Ngoreme. These images are again remarkably congruent 
with the linguistic evidence for matrilineal or bilateral descent-based communities incorporating 
hunters as husbands. The "houses" of these women are the points from which oral tradition 
reckons differentiation into ethnic groups. The Ishenyi emergence story is an abbreviated one that 
describes the first couple of Iyancha and Mugunyi living in the hills of Guka. 

One version of the Ngoreme story uses the basic emergence images to explain the 
relationship and differences between Ikoma and Ngoreme: they were brothers who migrated from 
Sonjo together. When they got to the mountain Bangwesi (or Mangwesi) the Ikoma brother 
favored hunting and went off to hunt. The Ngoreme brother finally refused to carry the meat of his 
brother and moved away to find better farming land in Ikorongo. 3 In this version two brothers take 
on the role of farmer and hunter rather than the original couple in the Nata. Ikizu and Ikoma 

'- Interview with Machota Nyantitu, Morotonga, 28 May 1995 (Ikoma <f). 
' Interview with Bhoke Wambura, Maburi, 7 October 1995 (Ngoreme ?). 

emergence stories. The separation of brothers according to economic specialization is a familiar 
narrative theme around the lakes, used to explain the common ancestry of peoples who are now 
divided in space and time. 4 The appearance of brothers (rather than first man and first woman) as 
the characters in the Ngoreme emergence story may be interpreted as genealogical prototypes of the 
kinship system and thus indicate a stronger patrilineal emphasis in the east. Some have argued that 
the Ngoreme and their neighbors across the Mara River, the Kuria, became more strongly 
patrilineal because of their greater emphasis on the inheritance of cattle wealth. 5 This emphasis is 
probably the result of late nineteenth century changes as western Serengeti peoples in the east 
gained cattle wealth through trade and moved down onto the plains for the first time. 
The Importance of Place 

What makes the emphasis in these stories different from those discussed in the last chapter 
is that they mention specific places known on the landscape today that are central to the meaning of 
the story. The Ikoma, Ngoreme, and Ishenyi all begin their emergence stories with hunters leaving 
Sonjo. Narrators note particular places along the way to their present home, in the Ikoma story 
these are the home areas of each of Mwikoma's sons. The Nata emergence story takes place in 
Bwanda and the Ikizu story near Chamuriho Mountain. The mention of the Ikorongo hills in the 
Ngoreme emergence story resonates today as homeland, though Kuria immigrants from North 
Mara, who also have a tradition of Ikorongo as homeland, now occupy this area. 

Because these places play such a critical role in the imagery of the stories, I decided that 
they were worth investigating in their own right. Among each of these ethnic groups individual 
elders who took an interest in my work insisted that I go and visit these important sites if I wanted 

4 For Suba see Michael Kenny, "A Stranger from the Lake: A Theme in the History of the 
Lake Victoria Shorelands, Azania 17(1 982): 1 5. 

5 For an argument about Kuria and pastoralism see Kjerland, "Cattle Breed.", 


to understand their history. When I proposed to do some videotaping of my research, the Ikizu 
elders decided that we must go and tape the stories at Isamongo's camp, Nyakinywa's cave, and the 
pond where they stretched the skin. On two different occasions and in two different places, I went 
with elders looking for the Nata origin site of Bwanda. One man told me that someone had painted 
the name "Bwanda" on a rock to mark the spot for future generations. 6 The physical places 
themselves held meaning for these elders. Yet how do these sites speak to a historian, what do they 
reveal about the past and the people who lived there? 
Hill Farmers 

All these sites from the emergences stories of Nata, Ishenyi, Ikoma and Ngoreme were 
located on hills or rises to the east of where they now live and on the western edge of the Serengeti 
plains and woodlands. [See Figure 5-1 : Ecological map of the Serengeti-Mara Ecosystem- 
woodlands, hills, grasslands, with emergence sites.] Mangwesi, 7 one of the highest mountains in the 
area, is a major point of reference in the Ikoma, Ngoreme and Ishenyi stories. Although narrators 
locate the Ikizu emergence story around another mountain, Chamuriho, to the west, an alternate 
Ikizu origin story describes the wanderings of the hunter, Muriho, who takes refuge with an elder 
at Mangwesi before reaching Ikizu. Sonjo itself is found in the hills on the eastern side of the 
Serengeti plains. The Guka hills of Ishenyi emergence are closer to Sonjo than to present day 

6 Interviews with Sochoro Kabati and Makuru Nyang'aka, trip to "Bwanda," 16 February 
1996; the site near Tabora B. We looked for the site near Matare, Mugumu, where Anacleti also 
confirms as the site of Bwanda. with Mariko Romara Kisigiro and the Chairman of Burunga 
village, they had been given instructions by Hassan Irende, the elder who write "Bwanda" on the 
rock, who was then bedridden. This story was confirmed by Charles Nyamaganda Burunga 3 
June 1995. 

7 Sometimes rendered Bangwesi or Pangwesi. 


The Serengeti - Mara Ecosystem 

Graphics by Peter Shetler, 
Dove Creek Information Services 
Arclnfo, Adobe Photoshop Macromedia Freehand 
Based on Map of Sereno.eti National Park and the Surrounding Area 
T.M. Caro, 1970, Serengeti Research Institute 

Figure 5- 1 : Ecological Map of the Serengeti- Mara Ecosystem 

Ishenyi. Linguistic evidence shows that these settlers were following a very old pattern of their 
Great Lakes Bantu predecessors who commonly inhabited the hill ridges. 8 

Because of local soil and climatic patterns, hills constitute the ecological niche that farmers 
would have, of necessity, inhabited. The fact that all these places in the emergence stories 
representing early farming communities are located on hills is fascinating considering the 
ecological evidence about subsistence patterns. The congruence of these two different bodies of 
evidence seems to indicate that these early communities were agriculturally based. Some of these 
hills, like Robanda, are the only elevation rise for miles and miles on a vast grassland, while most 
belong to a chain of hills bordering on the plains or woodlands. Western Serengeti farmers still 
live on the hills and rises to exploit the best soils. Farming practices today provide a means for 
hypothesizing longer term interactions of farmers with their environment. 

The emergence stories describe Nyasigonko, the first woman of Nata, as a farmer of 
eleusine millet, oburwe. Today a thin liquid millet porridge (ekerongori) and a thicker porridge 
eaten with the fingers (obokima) are the staple food of the area. Millet is also used in making beer, 
eaten raw on trips, and is necessary for most rituals of blessing. Millet obokima (porridge) is 
"food"-without it one has not eaten. The word "to eat" (kuragera) means specifically to eat 
obokima (porridge), with another word (kura) for eating other things. Even today elders regard 
cassava or maize obokima as second-rate food. African or finger millet (Eleusine coracana or E. 
africana), a crop domesticated in greater eastern Africa, was adopted by Great Lakes Bantu- 
speakers from Central Sudanic-speaking peoples in the distant past. 9 Grain crops were necessary 
for adapting to life in the drier grasslands of the western Serengeti. Kjerland identified sixteen 

8 Schoenbrun, A Green Place, pp. 160-162. See also Wagner, "Whose History," pp 26- 

9 Schoenbrun, "We Are What We Eat," pp. 10-12. 

varieties of finger millet used among the Kuria, differentiated according to soil tolerances, color 
and end use. 10 This diversity and adaptability of native varieties indicates a long history of 
cultivation in the area. 

Finger millet requires a fertile and free draining sandy loam soil, since it cannot tolerate 
water logging." In the western Serengeti these soils are found only on hills and rises. The plains, 
interspersed between the hills, of the western Serengeti consist of the "mbuga" type soil that is a 
dark, heavy, clay-like "cotton soil," becoming waterlogged and swampy in the wet season. 
Western Serengeti people have only fully exploited the "mbuga" soils with the arrival of ox-plows 
during the colonial period. Many elders described a wooden digging stick (akoromo in Ikoma) as 
the original farm implement, which farmers gradually replaced with wooden and metal hoes. 12 
Although western Serengeti people have traded metal hoes from Geita for centuries before the 
colonial period, these hoes seem to have been a prestige item rather than a common tool. 
Cultivating "mbuga" soils with the digging stick or wooden hoe would be nearly impossible and is 
difficult even with the metal hand hoe manufactured today. 

Local classification of soil types includes ekebuse (sandy upland soils) and eseghero (clay 
bottom land soils). The best soil is a mixture of both, found on the low elevation rises. In the past 
farmers had to work up the clay eseghero soil in the dry season to get it ready before the rains. 
This practice, called kuharaga, has fallen out of use since the arrival of the ox-plow. If the year is 
good, eseghero soils are incredibly productive. Nevertheless, they either get too hard in a dry year 
or too swampy in a wet year to produce a reliable crop. Thus, farmers seek out the ekebuse sandy 

10 Kjerland, "Cattle Breed," p. 37. 

11 J.W. Purseglove. Tropical Crops: Monocotyledons I (London: Longman, 1972), pp. 

12 Interview with Pastor Wilson Shanyangi Machota, Morotonga, 12 July 1995 (Ikoma d - ). 

loam soils for their consistency. If a family controls enough labor, they will exploit both kinds of 
soil in one year. I3 

These soils exist in various mixtures and within close proximity, usually depending on the 
elevation of the slope, so that farmers become adept at staggering the placement of their fields to 
find the right combinations. It is reasonable to assume that people would have followed a similar 
strategy for exploiting these same hill ecologies in the past. Today one family rarely locates all 
their fields in one contiguous space. In the annual preparations for farming, a woman commonly 
asks those who have obligations to her in the economy of reciprocity to plant part of their field for 
her. Depending on the level of obligation, she may come for the weeding or simply return when the 
field is ready for harvest. Often a man sends his second wife or older sons to distant fields for the 
growing season. Land is in plentiful supply but the right soil combinations and the labor strategies 
to farm them are critical issues. People spread out the risks by maximizing the diversity of the 
environment. '" 

Western Serengeti people, now as in the past, build their settlements on the hillsides to 
escape the swampy lowlands in the rainy season. The low areas are considered bad for the health 
of people and livestock in the rains." Elders testify that in the past only the "kitchen gardens" 
were found near the homestead, with the area close by being reserved for herding. Fields were 
located away from the settlement area and temporary houses built there for the growing season. 

13 Interview with Nyamaganda Magoto, Nyawagamba Magoto, Mahiti Gamba, Bugerera 
(Nata cf). ■ 

14 Observations through two farming seasons in Nata and travel throughout the 


15 Interview with Mahiti Gamba, Bugerera, 4 February 1996 (Nata d"). Presumably low- 
land areas are connected with malaria or "fevers." 


Farmers had to protect their fields from wild animals so they farmed in contiguous blocks, 
surrounding the whole area with a thorn fence. 

Hills as Zones of Interaction 

In the emergence stories western Serengeti people also located their hill settlements at the 
interstices of the ecological zones of the area, allowing farmers to exploit herding and hunting 
ecological zones contiguous to each other. This position provided frequent opportunity for 
interaction with hunter/gatherers who lived in the adjoining woodlands and herders who lived on 
the plains. It is this pattern of interaction, afforded by the position of the hills, that gave western 
Serengeti farmers the means for prosperity in a marginal land. These landscapes of hunters, 
herders and farmers, living in close proximity form the ecological basis for the meeting of hunter 
and farmer, represented in the spatial images of the emergences stories of first man and first 
Ecological Zones 

Ecologists refer to the whole area, from Ikizu in the west to Sonjo in the east, as the 
Serengeti-Mara ecosystem, covering some 25,000 square kilometers. This area can be broken into 
three major ecological zones of hills, woodlands and grasslands that have provided the diverse 
environment for inter-cultural exchange. Wildlife populations have also thrived in this varied 
ecosystem, producing the largest herds of grazing mammals in the world with thirty species of 
ungulates (some 2.4 million total) and thirteen species of large carnivores. 16 [See Figure 5-1: 
Ecological map of the Serengeti-Mara Ecosystem, p. 199.] 

16 Sinclair and Norton-Griffiths, eds. Serengeti: Dynamics of an Ecosystem . Sinclair, 
"The Serengeti Environment," p. 41. See Chapter 1, "Dynamics of the Serengeti Ecosystem: 
Process and Pattern," and Chapter 2, "The Serengeti Environment," both by Sinclair for an overall 
view. The Serengeti-Mara ecosystem is defined as that area influenced by the migratory 
wildebeest population, p. 3 1 . 


The extensive grassland plains are found in the southern part of the Serengeti-Mara 
ecosystem, with projections reaching up into what is now Loliondo, to the west of Sonjo, and out to 
Lake Victoria in what is now the western Corridor of the National Park. The Serengeti plain is a 
vast short-grass savanna, with trees along the rivers and among the rocky out-croppings known as 
kopjes. An impenetrable hardpan has been formed under much of the volcanic soils of the plains 
making it possible for only shallow-rooted grasses to survive. In spite of this, the Serengeti is one 
of the most productive grasslands in the world, in terms of biomass. " During the colonial years 
Maasai herders moved into the western Serengeti, at least as far as Moru Kopjes in the western 
hills, during unseasonably dry periods. To avoid disease they grazed the plains during the dry 
season when the large wildlife herds were not there. 18 The Tatoga now graze the plains of the 
western Corridor but they once dominated the Serengeti plains, at least seasonally, as far as the 
Ngorongoro Crater. 

To the northwest of the plains the volcanic soils become finer grained, the rainfall 
increases and taller grasses of different species along with bush begin to thrive, merging into 
woodland. The alkaline ash soils demarcate the limits of the plain. The soils of the woodlands, to 
the north and west of the great plains, are formed from granite or quartzite parent rock, creating an 
acacia thorn-tree scrub and woodland." The small whistling thorn trees (Acacia drepanolobium) 
dominate in the woodlands with poorly drained soils. 20 

17 Sinclair, "The Serengeti Environment." pp. 41-42. 

" H. St. J. Grant, District Officer, "Report on Human Habitation of the Serengeti National 
Park," May 1 954 and from the District Office. Masai, Monduli, 28 May 1 954 to the P.C. 
Northern Province, Arusha, National Game Parks, 215/350/vol. III,TNA. 

19 Sinclair, "The Serengeti Environment," pp. 38-39. 

20 Ibid. 


Oral narratives identify these woodlands as the space of the Asi hunter/gatherers. 
Ambrose demonstrates that the areas best suited for gathering edible plants is on the more acidic 
soils of the woodland bush, not suited for either grazing of livestock, the larger ungulates or 
farming. The low soil fertility and low rainfall of this bush area encourages larger tubers and 
woody plants from which edible plant foods are found. The migrating herds of larger animals from 
the plains move through here in the dry season in search of the permanent water pools along the 
tributaries of the Grumeti and Mara Rivers. Archeological evidence suggests that hunter/gatherers 
in East Africa have always depended on the resident populations of smaller game that live in the 
woodlands, while pastoralists made use of the larger ungulate herds from the plains for meat. 21 

The hunter/gatherers also preferred the eco-tones between woodlands and grasslands and 
probably inhabited the hill areas before the Bantu-speaking farmers arrived. They hunted the 
resident populations of smaller game in the woodlands and grazed small herds of livestock, 
probably sheep and goats, on the grasslands. However, as farmers came to dominate the hill 
ecologies, they increasingly relegated the hunters to the woodland ecologies alone. The woodlands, 
with their supply of edible plants, resident game population and seasonal arrival of larger game, 
sustained a hunter lifestyle. Yet their confinement to the marginal woodlands areas meant that they 
came to depend on hill farmers and grassland herders for livestock and grain in return for 
woodlands products. 22 

21 Ambrose. "Hunter-Gatherer Adaptations," pp. 11-42; Schoenbrun, A Green Place , pp. 
1 04-106. Curtis Marean, "Hunter to herder: Large Mammal Remains from the Hunter-Gatherer 
Occupation at Enkapune Ya Muto Rock-Shelter, Central Rift Kenya," The African Archaeological 
Review . 10 (1992): 65-127. 

22 Ambrose, "Hunter-Gatherer Adaptations," pp. 1 1-42; Schoenbrun, A Green Place dp 


The woodlands and plains finally break up among the ranges of hills in the northwest, 
creating a unique mosaic of many distinct ecotones in close proximity. Rivers form valleys 
between the hills, creating fertile but swampy low-lying areas of woodland. Directly to the south 
of the western hills the grassland plains again dominate, as far west as the lake. Here the Tatoga 
herders pasture their vast herds as close neighbors to the Bantu-speaking farmers who live in the 
hills of Ishenyi or Ikizu. This combination of hills, woodlands and plains, in close proximity, is the 
landscape that lies at the heart of the spatial images of the emergence stories, creating the 
ecological basis for the meeting of hunter and farmer, first man and first woman. 

Sustained farming is not possible either on the short-grass plains or in the acacia 
woodlands. The grasslands have never supported permanent farming communities because of the 
hardpan soils and lack of permanent water sources. An early colonial report states that, "on the 
nine days' track Ikoma to Ngorongoro through the Zerengeti [sic] there are only two perennial 
watering-places." 23 Although the average rainfall would allow for farming in the woodlands, the 
poorly drained soils prohibit it. A colonial resource survey of the region describes these areas as 
the "famine" land of heavy black clay soils, covered by acacia thorn bush. 24 

Those areas that are now woodland or grasslands may not necessarily have appeared this 
way two hundred years ago. The savanna ecosystem oscillates between phases where more 
grasslands or more woodlands exist depending on such large-scale perturbations in the system as 
widespread disease, drought, hunting or fire. For example, after the 1 890 rinderpest panzootic in 
which 95% of the wildebeest and buffalo died, the grasses grew taller, providing more dry fuel for 

3 Geographical Section of the Naval Intelligence Division, Naval Staff, Admiralty, A 
Handbook of German East Africa iLnnrlnn- His Majesty's Stationery Office, 1920; reprint ed. 
New York: Negro University Press, 1969), p. 159. 

24 V. C. R. Ford, The Trade o f Lake Victoria (Kampala: The East African Institute of 
Social Research, 1955), p. 16. 

hotter and larger fires, which destroyed trees and led to the spread of grasslands. The increase in 
the wildebeest population and more controlled burning since then has resulted in a trend toward 
bush encroachment. Ecologists now conclude that these extreme disturbances of the Serengeti 
ecosystem are, in fact, responsible for maintaining the diversity, productivity and resilience of the 
system. 25 Those who try to preserve the Serengeti in its "natural" state by not allowing these large- 
scale disturbances may destroy the feature that created its diversity, productivity and resilience in 
the first place. 
Human Interaction with the Environment 

Patterns of human habitation have done much to create and maintain this unique ecosystem 
over the past millennium, particularly in the use of fire (described in oral narratives as the secret of 
first man). The woodlands of the northwest zone of what is now the Serengeti National Park 
(closest to the area historically inhabited by western Serengeti peoples) contain species of trees that 
serve as markers for what ecologists call a "fire-maintained successional stage." 26 The dominant 
grasses of the western Serengeti, Themeda triandra and associated grasses, are also fire tolerant, 
indicating the evolution of these ecologies in conjunction with annual burning. 

This ecological evidence supports oral narratives that describe subsistence strategies in 
relation to burning. After farmers harvest the crops in July and August the dry grass is set on fire, 
burning extensive areas before it stops. Ecological evidence suggests that this is a very old way of 
maintaining the optimal balance between grassland and woodland. Burning eliminates all of the 

25 Norton-Griffiths,"lnfluence of Grazing, Browsing, and Fire," pp. 332-333, 341-348; 
See also Jane Guyer and Paul Richards, "The Invention of Biodiversity: Social Perspectives on the 
Management of Biological Diversity in Africa," pp. 1-13; and Peter D. Little, "Pastoralism, 
Biodiversity and the Shaping of Savanna Landscapes in East Africa," pp. 37-51; both in Africa 66 
1 (1996). 

26 Sinclair, "The Serengeti Environment, p. 39. 


tall dead grass that has almost no nutritional value for either wild animals or cattle and encourages 
the growth of the more nutritious grasses such as red oat grass, Themeda triandra, at the expense 
of the coarser grasses. 27 Themeda triandra, called ambirisi (from the verb "to herd" korisi), is 
acknowledged in local languages as one of the best pasture grasses, also valuable for thatch. 28 The 
month of ekinyariri ("green lands") refers to the greenflush of grass, especially appealing to cattle 
and wild animals, that emerges right after a burn. [See Figure 5-2: Human Interaction with the 

The landscape created by these burns has an orchard-like appearance with scattered acacia 
trees over a low grass pasture. The tree and grass species of the western Serengeti are particularly 
adapted to fire and regular burns do not destroy them. 25 One of the most important effects of 
burning is to control dense areas of bush which might harbor the tsetse fly. Western Serengeti 
people value and encourage open plains, as the sign of a healthy environment, by burning. From 
ancient times people not only adapted their practices to the existing environment but took an active 
role in creating an environment that was conducive to their economic prosperity. 

Evidence shows that both the practice of burning and agro-pastoral patterns in the western 
Serengeti were also responsible for the control of endemic trypanosomiases. Bantu-speakers who 
farmed the hills kept mainly goats and sheep until the colonial period. Goats and sheep develop 
more resistance than cattle to trypanosomiasis because they browse into the bush, which brings 
them into limited but regular contact with tsetse fly habitats. Because people kept their limited 

27 Norton-Griffiths, "Influence of Grazing, Browsing, and Fire on Vegetation Dynamics," 

28 Grasses identified in Nata by Nyawagamba Magoto and keyed to scientific name in 
D.M. Napper, Grasses of Tanganyika (Dar es Salaam: Ministry of Agriculture. Forests and 
Wildlife, Tanzania, Bulletin, No. 18, 1965), p. 132. 

29 Sinclair, "The Serengeti Environment," pp. 37-40. 

Making Arrow Poison (obosongo), 
Bugerera, 5 January 1996 

Green Flush After a Burn in the Serengeti 
Figure 5-2: Human Interaction with the Environment 

livestock herds close to concentrated settlements, they avoided intensive contact with the tsetse fly. 
They farmed, in turn, on the boundaries of the bush, while visiting woodland areas, where tsetse 
thrived, for hunting, wood gathering or travel. A German report from early in this century states 
that, "the fields in some cases are several hours' journey from the houses, mostly lying in the low 
grounds amongst the rivers and brooks." 30 The tsetse fly officers in the 1930s were concerned with 
the practice of farming away from the homesteads, into the bush. 31 Yet this periodic contact of 
humans with the tsetse vector served to maintain sufficient trypanosomiasis resistance levels. 
Together these practices allowed the farmers gradually to push back the cleared areas and maintain 
a controlled zone of regular contact with the tsetse fly. 32 

Evidence for the successful control of contact with tsetse habitats comes from the colonial 
record. A Veterinary Officer reported in 1928 that, "the fly belt is vaguely demarcated by the 
natives who seem to know where they can safely graze but keep dangerous close . . . goats seem to 
thrive there." 33 In the first reports of sleeping sickness in the Ikoma area the Medical Officer was 
surprised to find that livestock were healthy, although he found cases of trypanosomiasis among 
them and vast tracts of tsetse infested bush surrounded the settlements. The same report also 
found, in the human case, that the incidence of sleeping sickness was only one percent, with no 

30 Geographical Section, A Handbook of German East Africa , pp. 97, in the section 
describing the "Washashi and Wangorimi." 

31 H. G. Caldwell, "Report on Sleeping Sickness in the Musoma District, July and August 
1932, Sleeping Sickness: Musoma District, 215/463, TNA 

Giblin, The Politics of Environmental Control . 

33 District Veterinary Officer, Musoma, to the D. O. Musoma, 19 January 1928, Annual 
Report 1927, p. 4, 1927-28 Provincial Administration, Annual Reports 1927, Mwanza Province 
246/P.C./1/30, TNA. 

tendency to epidemic spread. 34 Western Serengeti peoples had apparently learned how to coexist 
with the tsetse fly. 35 
The Wildebeest Migration 

Wildlife ecologies also affected historical developments in the western Serengeti. In the 
Ikoma and Ngoreme emergence stories the hunter comes to the western Serengeti from Sonjo by 
hunting wildebeest. The route of the annual migration begins in the southeast, in December, when 
the short-grass plain is teeming with vast herds of wildebeest (about 1.3 million at last count), 
mixed with zebra and gazelle. The herds come to the plains in the rainy season to give birth to 
their calves on pastures rich in minerals, finding water in pockets left by rain. The long view 
available on the short grass plains also provides better protection from predators when the calves 
are vulnerable. 36 As the plains dry up from May to June, the herds move north and west into the 
woodlands looking for permanent water sources and fresh grass. 37 The acacia woodland in the 
western Serengeti is the termination point of the migration, just where the Sonjo hunters ended their 

34 J. F. Corson, M. 0., Ikoma, 1 5 April 1 927, "Third Note on Sleeping Sickness," Extracts 
of Report by District Veterinary Officer, 1926-29, Provincial Administration Monthly Reports, 
Musoma District, 215/P.C./1/7, TNA. 

35 On trypanosomiasis see Ford, The Role of the Trypanosomiases - Richard Waller, 
"Tsetse Fly in Western Narok, Kenya," Journal of African History 31 (1990): 81-101; and James 
Giblin, "Trypanosomiasis Control in African History: An Evaded Issue?," Journal of African 
History. 31 (1990): 59-80; for a similar case in Sukuma and interaction with tick ecologies see 
Martin H. Birley, "Resource Control," Africa 52, 2 (1982): 1-29. 

36 These dynamics described in Sinclair and Norton-Griffiths but also for a popular 
audience in James Scott, The Great Mig ration (London: Elm Tree Books: 1 988). 

37 Dennis Herlocker, Woody Vegetation of the Serengeti National Park (The Caesar 
Kleberg Research Program in Wildlife Ecology and Texas A&M Univerity, 1973), p. 9; Sinclair, 
"The Serengeti Environment," p. 33 


The emergence sites are located in this termination zone of the wildebeest migration and at 
a perfect place for hill farmers to take advantage of the sudden seasonal arrival of meat. Even up 
through the colonial period, western Serengeti peoples most commonly hunted during the dry 
season as individuals or in small groups using bows and arrows tipped with poison, traps or snares. 
However, when the wildebeest herds arrived, they engaged in communal hunting using fall-pits, 
ebereri, dug between one hill and another or between another natural means of constricting 
movement through a narrow gap. Hunters would dig many pits in a row and carefully conceal 
them with grass and sticks. 38 Everyone, men, women and children, turned out to stampede the 
wildebeest herds into the pits. Some said that the hunters might kill 100, 100 or 500 wildebeest in 
a day. In 1 899, the German traveler Kollman counted as many as 200 hunting pits in a half-hour 
walk through the Ruwana plain. 39 The massive slaughter of wildebeest reported in the late 
nineteenth century and early colonial period was a result of the commercialization of hunting 
during that era. However, the use of hunting pits for obtaining large quantities of meat to dry 
seasonally seems to be a much older practice. 

In the western hills, farming communities could take advantage not only of the varied 
ecotones in the region, each with its own economic potential, but could also exploit the annual 
migrations of wildlife. They could not live permanently on the plains or in the bush but could use 
these areas by situating themselves in close proximity. Given these parameters, one could predict 
the emergence sites on an ecological map of the region. These ecologies also help to explain the 
connection to other hill farmers in the region, the Sonjo. 

38 Interviews with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata <f); Nyambeho 
Marangini, Issenye, 7 September 1995 (Ishenyi <?). 

39 Kollmann, The Victoria Nvanza. p. 199. 

Relations with Other Hill Farmers: The Sonio Connection 

One of the most perplexing problems in the interpretation of oral traditions of emergence is 
making sense of the purported connection of western Serengeti hill farmers to the Bantu-speaking 
Sonjo hill farmers in what is now the Loliondo district of Maasailand. 40 Western Serengeti peoples 
today bear little resemblance to the Sonjo, either linguistically, culturally or in terms of social 
organization. Yet Ikoma and Ngoreme asimoka stories, told above, assert that their ancestors 
came to the western Serengeti from Sonjo as hunters. It is not clear whether these ancestors were 
supposed to have been Sonjo farmers on a seasonal hunt or hunter/gatherers living in close 
relationship to Sonjo farmers. Because these traditions often attribute the migration of Sonjo 
hunters to the impact of Maasai raids, it is also unclear whether they refer to the period of 
nineteenth century disasters or to much earlier patterns of interaction. 

When I went to Sonjo to find out if similar traditions of interaction existed there, some 
amazing congruencies emerged in spite of the separation of these communities since the early 
colonial period. Sonjo traditions also tell of first man as a hunter who brought fire. They call the 
hunter clan in Sonjo the Sagati, also found among the western Serengeti Ishenyi and Ikoma 
people. 41 In Sonjo, members of the Sagati clan are responsible for blessing the bows and arrows 
before a hunt and when they are made. This clan also kept the secret for making arrow poison (an 
early trade item) and wild herbs for medicines. Sonjo elders said that these hunters maintained 

40 One possibility that was not investigated in this project is the connection between Mara 
languages and Uplands Bantu languages. Fairly solid bodies of culture vocabulary link the two. 
This interaction in the distant past may be represented by the Sonjo cliche. See Schoenbrun. 
"Early History"; and Ehret, Southern Nilotic History . 

41 Among the Ishenyi the Sagati or Sageti clan was reported by numerous informants, 
among the, interviews with Mang'ombe Morimi, Issenye Iharara, 26 August 1995 (Ishenyi <f)- 
Mikael Magessa Sarota, Issenye, 25 August 1995 (Ishenyi a). The Sagari clan among the Ikoma 
was a hunting clan which has since disappeared as an independent clan. Interview with Mabenga 
Nyahega and Machaba Nyahega, Mbiso, 1 September 1995 (Ikoma <f). 

relations of trade between [koma and Sonjo, meeting in the wilderness where Sagati gathered honey 
and hunted. The Sonjo village of Rhughata claims origins from a Sagati hunter and his wife who 
left "Ikoma" because of Maasai raids. 42 Traditions say that they came from Jaleti and Ngrumega 
(perhaps a transliteration of the Rivers Mbalageti and Grumeti in the western Serengeti). 43 
Samonge village elders claim origins from a hunter father who brought fire and a farmer mother 
who controlled water. 44 These traditions from both sides of the Serengeti bear too much similarity 
in specific clan and place names as well as the core images of emergence stories to be coincidental. 

Although a large portion of the migration tradition from Sonjo seems to refer to the late 
nineteenth century period of disasters, much evidence suggests that farming settlements in Sonjo 
and the western Serengeti were in contact from the distant past, largely through the interaction of 
hunters from both places who traveled across the woodlands and plains in search of game. The 
hills of Sonjo are only 80 kilometers from Ikoma straight across the Serengeti where the hills again 
break up the woodlands and plains. Because of the ecological patterns described above concerning 
the wildebeest migrations, hunters from Sonjo who followed the herds seasonally would have had 
occasion to ask for hospitality in Ikoma, while Ikoma hunters would have found themselves near 
Sonjo at the end of the dry season. Ikoma migrant laborers walking to Magadi Soda in Kenya 
during the colonial years in search of work slept the third night in Sonjo and were welcomed as 
brothers by local people who remembered these common traditions. 

Because the hill ecologies of Sonjo and the western Serengeti are similar, we might 
hypothesize that migrants looking for new areas in which to settle would have sought out 

42 Interviews with Peter Nabususa, Samonge, 5 December 1 995; Emmanuel Ndenu Sale 6 
December 1 996 (Sonjo <f). 

43 Interview with Emmanuel Ndenu, Sale, 6 December 1995 (Sonjo cf). 

44 Interview with Marindaya Sanaya, Samonge, 5 December 1995 (Sonjo cf). 

environments most closely resembling their home areas. 45 Once Sonjo hunters established contact 
in the western Serengeti, small groups of people may have gone there to build their homes or to 
marry. Some Sonjo elders say that the great prophet Khambageu came from a large mountain near 
Ikoma in the west, where until recently Sonjo people returned annually to propitiate his spirit. 46 If 
both Ikoma and Sonjo formed their identity as hill fanners within an inter-cultural environment of 
plains herders and woodland hunter/gatherers in the distant past, then they may have felt an affinity 
that they explained by common origins. They were "brothers" within the regional understandings 
of economically-based identities. 

Western Serengeti and Sonjo elders stated to me that the most conclusive proof of their 
common parentage as children of "one womb" was that both have the ntemi scar on the right 
breast. For the Sonjo this is a sign of belonging and identity. 4 ' The Ishenyi, Ngoreme and Ikoma 
also use this mark but understand it as a health precaution for children. When I visited Sonjo, an 
Ikoma and a Nata man accompanied me. Our Sonjo hosts greeted the Ikoma man who had the 

45 Philip Curtin, Steven Feierman, Leonard Thompson and Jan Vansina, African History 
(New York: Longman, 1978), p. 125, Feierman describes the pattern in the migrations of Bantu- 
speaking peoples who, when forced to move, chose those places where they could apply their 
environmental knowledge. David W. Cohen describes the same pattern in, "The face of contact: a 
model of a cultural and linguistic frontier in early eastern Uganda," in Nilotic Studies. Part two. 
Proceedings of the international sy mposium on languages and history of the Nilotic peoples. 
Cologne. January 4-6. 1987.. eds. Rainer Vossen and Marianne Bechhaus-Gerst (Berlin: Dietrich 
Reimer Verlag, 1983), pp. 339-356. 

46 Interview with Emmanuel Ndenu, Sale, 6 December 1 995 (Sonjo <f); stated 
"Khambageu was a prophet and a god, he came from over toward Ikoma. His wife was Nankoni. 
They used to visit back and forth with Ikoma especially in the tenth through the twelfth month. 
They went to worship there and the ones that followed him went there to worship too." See also 
Robert F. Gray, The Sonio of Tanganyik a: An Anthropological Study of an Irrigation Based 
Society. (London: Oxford University Press, 1963), pp. 1 1-12, who relates the tradition of 
Khambageu coming from the Sonjo village of Tinaga, and then cursing the village, leading to its 

See Gray, The Sonio of Tanganyika, p. 1 5, on the ntemi scar. 

ntemi scar as a long lost brother and friend, giving him gifts and special treatment as an honored 
guest. They treated the Nata man, who had no ntemi scar, indifferently. In the past individual 
immigrants from Sonjo may have brought the practice of ntemi scarification but its meaning 
changed outside the context of Sonjo identity. It may have remained in practice for identifying 
"brothers" in the wilderness. 

While the evidence seems to support the hypothesis of a zone of interaction in the distant 
past between Sonjo and the western Serengeti, much remains to be explained. For example, 
although Ikoma, Ngoreme and Ishenyi traditions claim origins in Sonjo, little linguistic or 
ethnographic evidence exists for this. If Ikoma ancestors came from Sonjo, they retained almost 
nothing of Sonjo culture or language. Language shifts take at least three generations and the 
incorporation of loanwords several generations. No linguistic trace of a large migration from 
Sonjo remains, either in the distant or recent past. 48 If small groups of immigrants did come from 
Sonjo, they were completely incorporated into western Serengeti society and culture. The only 
linguistic evidence for interaction in the distant past is shared loan words from a preexisting 
Southern Nilotic-speaking community with which both had contact. 

Evidence exists that today's Sonjo are related to the farmers who worked the ancient 
irrigation agricultural settlements in the Rift Valley, known in the archaeological literature as 
Engaruka, dating to at least 300 years ago. The Sonjo still practice a complicated system of 
irrigated agriculture that forms the basis of their socio-political system. 49 The most influential 
Sonjo leaders are those who control the allotment of water for their sections. Ngoreme praise 

1 Ehret, Southern Nilotic History , pp. 26-27. 
' Gra y> The Sonio of Tanganyika, pp. 53-56. 

names for Sonjo ancestors describe them as "those who irrigate." 50 Yet little evidence exists that 
people in the Mara Region ever practiced this kind of agriculture or settled in the patterns evident 
in these archeological sites." 

Memory of the connection to Sonjo draws more directly on the common experience of 
Maasai pressure during the period of disasters in the late nineteenth century that resulted in small 
groups of refugees moving in both directions. Sonjo sources date the dispersal to Ikoma to two 
generations ago. 52 This is discussed more thoroughly in later chapters, along with the ideological 
reasons why the Ikoma, Ishenyi and Ngoreme would have preferred origins in Sonjo to local 
origins. Similar to the Ikizu case already discussed, the Ikoma and Ngoreme seem to have used the 
older founding myth to legitimate later changes in identity that resulted from the nineteenth century 

Whatever the particular connection of western Serengeti peoples to Sonjo in the early 
period, the Serengeti plain was clearly a zone of interaction rather than a barrier. 53 Ecological 
patterns provided the environment in which these two sets of peoples would have met each other 
during their everyday subsistence activities. According to narratives on both sides, hunters, 
traders, settlers and pilgrims frequently crossed the Serengeti plain. The location of emergence 

50 Interview with Silas King' are Magori, Kemgesi, 21 September 1995 (Ngoreme <f). 

51 For a Sonjo ethnography see Gray, The Sonio of Tanganyika . For archaeological 
investigation of Engaruka see Leakey, "Preliminary Report;" Sutton, "Engaruka etc.," pp. 7-10; 
John Sutton, A Thousand Years of East Africa (Nairobi: British Institute in Eastern Africa 1 990) 
pp. 33-40. ' 

53 This thesis was first expounded by A. O. Anacleti, "Serengeti: It's People and their 
Environment," Tanganyika Notes and Records 81/82 (1 977): 23-34; and by the same author, 
"Pastoralism and Development." 

sites shows that western Serengeti people once lived much farther east, while Sonjo tradition 
described settlements located much farther west. 

This evidence would suggest a set of settlements in close proximity, identifying themselves 
as hill farmers, who later differentiated into Sonjo and Mara peoples as they moved apart in the 
nineteenth century. 54 Whether the common culture they shared was closer to present-day Sonjo or 
Mara, or entirely different, is now difficult to tell. The long-term pattern of interaction and the 
deep connections built between communities over the centuries forms the context in which new 
identities developed in the nineteenth century. During the Maasai raids refugees from Sonjo may 
have come to Ikoma because they had relatives, friends or trading partners there. These ecologies 
were known landscapes that people on both sides of the plains frequently crossed. 

Relations with Hu nter-Gatherers: Learning How to Live on the Land 

Hill settlements, located to take advantage of other ecologies and of the annual wildebeest 
migration, provided a space in which farmers consistently interacted with hunter/gatherer peoples 
in the woodlands. The genius of western Serengeti Bantu-speaking peoples' adaptation to this 
environment lay in their maintaining a strong identity as farmers, separate from but interdependent 
with hunter/gatherer and herder neighbors, while doing some seasonal hunting and herding 
themselves. The hunting skills of Bantu-speakers would have expanded as they learned from 
hunter/gatherers whom they incorporated into their communities. An investigation of the 
relationship between farmers and hunters suggests further ways in which the Asi hunters were 
"fathers" to the farmers in relation to the land. 

54 The Maasai Loitai of the Loliondo highlands, whose territory extended north into Kenya 
and east to Lake Natron, reported to colonial anthropologist Henry Fosbrooke that the original 
inhabitants of this area were the "Ilmarau," the Maasai name for the Ikoma or Bantu-speaking 
peoples of the western Serengeti in general. Henry A. Fosbrooke, "Sections of the Masai in 
Loliondo Area,"1953, (typescript) CORY #259, EAF, UDSM. 

First Man as Asi Hunter 

If the emergence stories are in some way symbolic of long-term social processes then 
another level of interpretation is possible by comparing these traditions with what we know about 
the historical relationships between hill farmers and hunters. Nyamuny wa, remembered as the first 
father of the Nata people, was an Asi hunter. The Asi, as a hunter/gatherer community in the 
western Serengeti who were there before the Bantu-speakers arrived, have a long history of 
interaction with farmers and herders. 

One might argue that the Asi hunter in the emergence story is only a symbol of the 
wilderness in a purely mythical founding story of how civilization came to be. The structuralist 
symbols of wilderness/home, male/female, cooked/raw are all apparent in the story." The Asi are 
liminal boundary shifters, an elusive symbol of the wilderness, whom farmers sometimes call "wild 
animals." Some versions of the Ikizu and Ngoreme emergence traditions describe how their 
ancestors won the land by driving out the original short people with big heads, the Hengere. 56 It is 
these people, rather than the Asi, who represent the forces of nature conquered by civilization in the 
emergence traditions. Because the Asi hunter represents a historical community of known people 
with whom fanners had an ongoing relationship, this myth also has a historical reality behind it. 
An investigation of the relationship of hill farmers to Asi hunters may provide some insight into the 
meaning of the emergence stories. 

Other kinds of evidence show that Asi hunter/gatherers played a central role in the early 
establishment of Bantu-speaking settlers in the western Serengeti. Oral traditions that represent the 

* Edmund Leach, ed., The Structural Study of Myth and Totemism (London: Tavistock 
Publications, 1967). 

56 For an analysis of the common myth of first peoples and "short people" see, Wyatt 
MacGaffey, Religion and Society in Central Africa (Chicago: University of Chicago Press. 1986); 
Kenny, "Mirror in the Forest," 482-484. 

hunter as first father also reflect the importance of the Asi. Western Serengeti peoples give 
precedence as "first-comers" to Asi descendants in farming communities, identified by clan 
designation. For example, the Nata clan, Gaikwe, and the Ikizu clan, Hemba (from to light a fire) 
is associated with and uses Asi praise names. The Asi founded one Kuria section called Nyabasi in 
North Mara where Asi ancestors came to farming communities to trade arrow poison for cattle." 
Some have speculated that the name "Asi" is derived from the Bantu root for "earth" or "soil," in 
Nata ase or ahase." 

In the early years when Bantu-speaking farmers first entered the region they depended 
upon the Asi hunters to teach them the skills necessary to survive in an unfamiliar environment. It 
also seems likely from ecological and archaeological evidence that the Asi hunters would have been 
living in the hill ecotones where they could exploit both the grasslands and the woodlands. Yet 
because this was also the ecological niche best suited for farming they came into competition with 
incoming Bantu-speaking settlers. The emergence stories, at one level, describe the 
accommodations and conflicts between these two groups on the frontier. As the hill farmers 
gained familiarity with the environment, they increasingly pushed those Asi who refused to marry 
into hill farmer communities into the more marginal areas of the woodlands as they took over the 
hill ecologies. The hill farmers still had a close and interdependent relation with the Asi but came 
to dominate them rather then depend on them as their population expanded and many Asi 
assimilated as farmers and herders. Those Asi who lived in the marginal woodland areas bordering 

" Interview with Sira Masiyora, Nyerero, 17 November 1995 (Kuria cf). 

58 Kenny, "Mirror in the Forest," p. 482. Unfortunately, I was unable to find any 
descendants of the Asi who could recount their oral traditions. The Asi have either totally 
assimilated into farming communities or have gone to live in Loliondo under Maasai patronage. 
Thus, until more research is done the view of Asi history presented here is entirely from the 
perspective of the farmers. But because western Serengeti farmers consider the Asi one of the 
original parents and first-comers to the land, farmers respect their knowledge and history. 

the hills gradually began to rely on the farmers for grain and livestock in exchange for products of 
the wilderness. The symbiotic relationship between them developed from one of farmers' reliance 
on hunters to the farmers' dominance and control of the best land in the hill environments. 

In the second half of the nineteenth century the Asi increasingly became the clients of the 
Maasai who came to dominate the greater Rift Valley. However, this was a slow process and in 
the early part of this century the Germans still distinguished between Ndorobo in the Serengeti area 
who spoke a Maasai language and the "pure Wandorobbo in the Zerengeti [sic] steppe on the 
Syonera [sic] living as nomads" who spoke a different language, for which the "Washashi in 
Ikoma" acted as interpreters. This testimony indicates a close relationship between hunters and hill 
farmers, only recently altered by the presence of the Maasai. The same report said that "during the 
great migration of the Masai, the Wandorobbo were either driven out or forced to submit." 59 

Oral narratives of the western Serengeti acknowledge their debt to the Asi in the lore of the 
woodlands and hunting. Nata people learned the secret of arrow poison (obosongo) from the Asi. 
The Nata kept the secret to themselves, using it as a trade good, until recent times when clan 
brothers gave the recipe to the Ikoma. 60 Hunters make the poison by boiling the woody portions of 
the obosongo tree (Acocanthera fiersiorum) and making a dark concentrate by evaporation. The 
active ingredient in the poison was the glycoside, ouabain, which western Serengeti peoples say 
"freezes" the blood of the animal. This arrow poison was extremely valuable, one small container 
selling for the equivalent of a goat. Many people considered the arrow poison from western 
Serengeti one of the best and during the colonial years, at least, they traded it as far as Shinyanga, 

' Geographical Section, A Handbook of German East Africa , pp. 98-99. 
1 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata J). 


Mbulu and across the Kenya border. 61 There is a hill near Nyichoka that elders still remember as 
the place, protected by medicines, where they stored the obosongo when they brought it home, 
while the men purified themselves with sacrifices so they would not become sick. 62 [See Figure 5-2: 
Human Interaction with the Environment, for a photo of preparing arrow poison, p. 209.] 

Inconclusive evidence also exists that western Serengeti peoples borrowed their common 
style of bows and arrows from the Asi. Western Serengeti peoples use a long bow, more 
commonly found among hunter/gatherers in this part of East Africa. Fosbrooke's 1956 
investigation of Hadzapi material culture on Lake Eyasi demonstrates the contrast between the long 
bows (180 cm.) and arrows of these hunter-gatherers in contrast to those of the neighboring Bantu- 
speaking peoples that were little more than half as long. 63 When the German traveler Kollmann 
studied the material culture of the people of Lake Victoria in 1899 he noted that the "Ushashi," or 
western Serengeti, bows and arrows were also "strikingly large and beautifully worked," in 
contrast to the other bows he had seen on his journeys. [See Figure 5-3: Hunting Vocabulary and 
Tools.] The long bow pictured by Kollmann is 1 70 cms. in length, putting it within the range of 
bow lengths described by Fosbrooke for Eyasi hunter/gatherers. 64 The shapes of arrow heads 
described by Kollmann are also similar to that of the Eyasi hunter/gatherers, both making common 
use of poisoned wooden arrow-barbs. 

61 W. D. Raymond, "Tanganyika Arrow Poisons," Tanganyika Notes and Records 23 
(June 1947): 49-65. 

62 Interview with Makuru Nyang'aka, Nyichoka, 7 March 1996 (Nata cf). 

63 H. A. Fosbrooke, "A Stone Age Tribe in Tanganyika," South African Archaeological 
Bulletin 9, 4 1 (March 1 956): 5. 

64 Kollmann, The Victoria Nvanza. pp 194-5. When I visited Sonjo accompanied by a 
man from Ikoma and another from Nata they laughed at the small size of the Sonjo bows and 
offered to teach them how to make "real" bows. 




adult male 


male young 

female young 













gazelle — 




















Names for wild animals in Ekinata 

Cultural Vocabularies on wild animals and hunting provided by Nyamaganda Magoto and Tetere 

Tumbo, Mbiso, 23 November 1995. 

"■""'- ^''""•"^^^TiFrrnrTTrrn r-i i, ■ ■ ■ ■- . r - . 

Fi(J. jio.— BOW. (One-ieuth natural sue, III fc., 5<i,,;.| 

Paul Kollmann, The Victoria Nvanza: the Land, the Races, and Their Customs, with Specimens of 
Some Dialects . H. A. Nesbitt, trans. (London: Swan Sonneschein & Co., Ltd., 1 899), p. 1 95 

Figure 5-3: Hunting Vocabulary and Tools 


Hunting played a crucial role in the economy of hill farmers and, until the last century, 
provided the major source of meat. Ritual feasting for taking eldership titles still requires a 
prescribed number of pieces of dried wild animal meat. Before the Germans came, western 
Serengeti people paid bridewealth in wild animal skins. They also made clothing of these skins. 65 
Photos from 1904 and the testimony of elders show that an elder draped the larger skin of a topi 
across his shoulders (arachana) while a woman wore the skin of a gazelle as an apron around her 
waist (asaraka). 66 [See Figure 5-4: Kuria Woman, 1904.] 

The languages of the western Serengeti themselves attest to the importance of hunting. For 
example, each of the wild animals is given a separate name according to sex and age, as 
demonstrated for a few selected animals in the chart on the next page [See Figure 5-3: Hunting 
Vocabulary and Tools, p. 223]. This complex naming of wild animals shows that people highly 
valued hunting knowledge. An awareness of the hunting season permeates the agricultural 
calendar. The topis, one of the resident ungulate populations in the area, give birth to their calves 
in September and October. This is the sign that the rains are coming and that the fields should be 
ready for farming. 67 In Ishenyi, Obutir (September) is the month when the topi give birth and the 
time for preparing the fields, in October the first millet it planted. When the wildebeest return to 
the short grass plains in November, it is time to plant sorghum. 68 

Elders retain memories from the beginning of this century that testify to the symbiotic 
relationship between hunters and farmers that existed before the hunters became Maasai clients. 

65 Interview with Jackson Mang'oha, Mbiso, 13 May 1995 (Nata <f). 

66 Interview with Mahiti Gamba, Mayani Magoto. Bugerera, 3 March 1 996 (Nata <t) 
Nyamaganda Magoto, collection of Culture Vocabulary. 

"' Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata <*"). 

68 Interview with Nyambeho Marangini, Issenye, 7 September 1995 (Ishenyi d% 


Figure 5-4: Kuria Women, 1904 [Max Weiss, Die Volkerstamme im Norden Deutsch- 
Ostafrikas (Berlin: Carl Marschner, 1910), p. 289] 

After the farmers gained dominance over the hill ecologies and relegated the hunters to the 
marginal woodland areas the hunters developed friendships with farmer families to trade for grain 
and livestock. A German report states that the "Serengeti Ndorobo" also kept sheep and goats. 6 ' 
Elders today remember this relationship with Asi hunters from their childhood. Ikoma elders tell 
stories of particular Asi families who would regularly come to their homes. Patron-client 
relationships, particularly between Ikoma and Asi families, apparently lasted over generations. 
Asi brought wild animal products like ivory, wildebeest tails, lion manes and honey to their 
patrons' homes to trade for livestock, grain, iron or salt. Some Ikoma lineages identified Asi 
ancestors. Asi friends, clients or kin spoke Mara languages but few Mara speakers learned the Asi 
language. 70 In a colonial case of the theft of nineteen head of cattle in Ikoma by hunter/gatherer 
"Dorobo" from Loliondo, the suspects were identified by name, along with the "two Ikoma who are 
friendly with these Wandorobo." The District Officer further added that the Native Authorities of 
Ikoma Federation would not "take any action to recover the stolen stock or to arrest the culprits as 
they are afraid of the Wandorobo."" Or perhaps they had ongoing obligations with those families. 
Stories also exist of conflicts between western Serengeti farmers and Asi hunters, such as 
the war of Kangashori during the early colonial years (C. 1920s). A Nata hunting party found an 
animal killed with an Asi arrow, but they took it anyway and began to divide out the meat. When 
the Asi arrived a fight erupted and the Nata, being outnumbered, ran away. The Nata left behind 
an old man named Kangashori whom the Asi killed. They next day a Nata party took revenge, 

69 Geographical Section, A Handbook of German East Africa , p. 99 

70 Kratz, "Are Okiek really," p. 359, argues that the fact that the Okiek speak the language 
of their host does indicate the superiority of the host but rather the Okiek ability in crossing 
cultural boundaries. 

71 McMahon, D.O., to A.D.O Loliondo, 7 February 1932, and D.C. to P.C. Lake Province 
Mwanza, 16 Marcha, 1932, Stock Thefts: Musoma District 1932, 215/351, TNA. 

attacked an Asi camp and killed twelve people. The Asi appealed to the Nata chief, Rotigenga, who 
called the elders to perform an oathing ceremony that joined them as brothers. After that, they 
could neither steal from nor kill each other. 72 Even in this story of conflict the boundaries were 
mediated and enemies became brothers. 

Over the centuries the peoples of western Serengeti have learned from the Asi, 
intermarried, traded, and fought with them but, in spite of this frequent interaction, both farmers 
and hunters have managed to maintain seemingly rigid boundaries. However, the distinction 
between western Serengeti farmers as seasonal hunters and Asi hunter/gatherers may not be as firm 
as it appears. Scholars working among the Maasai argue that although social boundaries in the Rift 
Valley corresponded to economic subsistence patterns of herders, hunters and farmers, individuals 
frequently crossed those boundaries. When Maasai lost cattle they "became" hunting Dorobo and 
when Sonjo farmers gained enough cattle they "became" Maasai. 73 In the western Serengeti 
examples exist of individuals crossing these boundaries as marriage partners, adopted sons or out 
of economic necessity. The boundaries may also have been mediated by shared clan names that 
are discussed in Chapter 6 or shared age-sets that are discussed in Chapter 9. As the hill farmers 
pushed the Asi hunters into more marginal wilderness areas the hunters increasingly had to rely on 
the farmers for some basic elements in their subsistence economy. 

72 Interviews with Mayani Magoto; Yohana Kitena Nyitanga, Makondusi, 1 May 1995 
(Nata <f). Arrows are marked with individual and clan markings so that the ownership of a killed 
animal can be determined in the bush. The ceremony is called kura aring'a and is described in 
Chapter 10. 

73 See Spear and Waller, Being Maasai: and Berntsen, "Pastoralism, Raiding, and 
Prophets."; Kratz, "Are the Okiek really Masai?" 

Asi Territory and the Emergence Sites 

Looking at the geographical relationship of the places reported in oral traditions to be 
where Asi hunters and western Serengeti farmers lived affords further insight into the nature of the 
historical relationship between the two. The emergence sites named in the oral traditions of hill 
farmers are all located in areas now identified as Asi territory. The site of Bwanda (near to where 
Mugumu town now stands) was in an area known traditionally as Materego yabaAsi (the 
wilderness of the Asi). 74 Oral testimony also identifies the area north of Nyichoka in Nata 
territory, contiguous with the Ngoreme emergence site of Ikorongo, as Asi country. The Ishenyi 
emergence site of Guka is well into the woodland territory but also located on the hills. The 
earliest German maps (1910) show these same areas as "Ndorobo" territory, described as 
"undulating country of open thorn bush and grass," or "open bush and thick scrub." 75 The colonial 
Game Warden found most hunting activity around the Moru Kopjes, and the Mbalageti and 
Seronera Rivers, where Ikoma traditions say they first settled. That all the emergence sites are 
located within what is known as Asi territory and within the woodland ecologies of 
hunter/gatherers, tells us that early agricultural settlers came into direct competition with hunters 
for the same land. 76 

74 Interview with Sira Masiyora, Nyerero, 17 November 1995 (Kuria cT). 

75 Karte von Deutsch-Ostafrika, A.4 Ikoma (Berlin: D. Reimer [E. Vohsen] : 1 9 1 0), 
German Maps, GM 30/3, TNA. 

76 One interpretation of this evidence could be that western Serengeti farming communities 
developed out of a preexisting hunter/gatherer society and that the emergence sites were their 
remembered hunting camps. Yet this kind of major shift in identity from a hunter/gatherer society 
to one based on farming would presumably constitute a rupture in historical consciousness similar 
to that described in the last chapter as happening during the period of disasters. Then places 
significant to an earlier way of life would be forgotten when divorced from their social context. 
The old sites of hunter/gatherer communities would not have figured in the historical imagination 
of these new communities. The evidence of historical linguistics and archaeology already 
presented points to two separate communities of hunters and farmers with different histories who 


That farmers inhabited these hill sites within Asi territory meant that they had to have 
accommodated the Asi "first-comers" to take control of the land. Each practiced a different 
economic strategy within overlapping ecological zones. Bantu-speaking farmers were strangers 
and newcomers in an unfamiliar land. Their oral traditions relate symbolically how they may have 
come to terms with those who were there before them and who knew how to live on the land. They 
did this by incorporating the hunter into their own communities and by situating themselves in a 
location that could take advantage of new economic possibilities. By encroaching on the ecological 
niche that the Asi occupied, the hill farmers put increasing pressure on them either to leave or to 

Elders remember the Asi hunters as first-comers who performed rituals to maintain the 
prosperity of the land. This relationship to the land, as farmers now ritually maintain it, is 
discussed in more detail in Chapters 7 and 8. In this ritual tradition western Serengeti people 
propitiate the ancestors of particular lineages as "guardians of the land." Many of those ancestors 
are from the Asi hunter clans of Gaikwe and Hemba. If western Serengeti people have maintained 
a ritual relationship to the land over the long-term, as seems likely, then in order for hill fanners to 
settle and prosper they would have to have been accepted as "children" in the lineage of the Asi 
ancestors who were connected to the land. They may have undergone "ritual adoption" into Asi 
lineages to gain access to the land. This hypothesis for the relationship of hunters and farmers in 
the distant past is congruous with the position of Asi hunter as "father" in oral traditions and oral 
narratives about an "oath" between Asi hunters and hill farmers. 77 A parallel process existed in the 
Kenyan highlands where Kikuyu tradition recounts ritual adoptions of Kikuyu into Ndorobo kin 

met on the frontier. 

77 Mahiti Kwiro, Mchang'oro, 19 January 1996 (Nata cf). 


groups to clear the land for farming. I am not sure why Kikuyu traditions would be more specific 
about this process. 78 Most western Serengeti elders deny a relation of kinship with the Asi. 
However it happened, Bantu-speaking farming communities were able to gain ritual access to the 
land and successfully diversify their own economic practices, leading to their dominance in the 
region. As a result the farmers increasingly marginalized the Asi hunters, who had taught, married 
and adopted them. As they gained ritual control over the land hill farmers usurped the place of the 
Asi as first-comers and guardians of the land. 

Relations with Plains Herders: The Silent Texts 
If the emergence stories are simply about the identity of hill farmers in relationship to 
people who practiced different economic strategies, then the silences in the emergence traditions 
speak as loudly as the texts. None of the emergence stories mentions the relationship to herders on 
the plains. 79 This seems to refute the claim made earlier that the core spatial images of the 
emergence traditions also refer to the ecological spaces of woodlands, hills and plains where the 
long-term patterns of interaction between hunter, farmer and herder were enacted. If herders were 
an important part of this interdependent system of economic subsistence then farmers would surely 
remember the presence of herders in oral tradition. One must consult other traditions and other 
evidence to understand this key omission. 

78 Greet Kershaw. Mau Mau from Below (Athens: Ohio University Press, 1 997), pp. 20- 
21 . Kershaw dates the Kikuyu settlement of Kiambu to the era preceeding the Kiraka famine of 

™ Traditions of the Lakes peoples do specifically recount stories of interaction with the 
Tatoga. See Huber, Marriage and Family , pp. 38-40 for the Kwaya; and Hartwig, The Art of 
Surviva| . PP- 47-48, for Kerewe. Among the Sizaki, interview with Thomas Kubini and Jacob 
Mugaka, Bunda, 10 March 1995 (Sizaki if). For the Sukuma, Itandala, "A History of the 
Babinza,"pp. 174-193. 


Abundant evidence, already presented, exists that agriculturalists did in fact interact with 
pastoralists in the distant past. The evidence of historical linguistics shows that these two 
communities have been in close contact ever since Bantu-speakers arrived in the region. Many 
Southern Nilotic loanwords in Mara languages concerning livestock indicate a high level of 
interaction and learning from Mara Southern Nilotic- and Dadog-speaking pastoralists in the 
distant past. 80 The word for sheepskin, among others, was originally Southern Cushitic, attesting 
to an even earlier set of interactions and learning.*' 

Nor can one interpret the silence in oral tradition concerning pastoralists as the lack of a 
herding component in the farming economy. From the evidence of historical linguistics we know 
that the first Bantu-speakers who entered the region had small stock. Although the importance of 
livestock increased in the post-crisis era, herding was thoroughly integrated into the total economic 
strategy of western Serengeti farmers from the distant past. It was another way of minimizing risk 
by maximizing use of each of the available environmental niches. Livestock were valuable for 
exchange and necessary for most rituals. Goats and sheep were kept in greater abundance because 
people could more easily give them up for a feast or in exchange for grain in times of need. Since 
people grazed their livestock near the homestead they could employ young boys and girls as 
herders while concentrating the labor of adult men and women in hunting and farming. Cattle 
trusteeship (kusagari chatugo) was one option used further to spread out the risk of losing an 
entire herd to disease or raid at one time. Western Serengeti people brand livestock with the sign of 

80 See Chapter 4 for linguistic evidence. 

81 Ehret, Ethiopians , p. 83. 

the patrilineage (ekehila) and of the owner. Looking at the brands in his corral one can tell the 
owner's social relations. 82 

Because oral traditions are transmitted and maintained by specific social networks it may 
be that no one group among the farmers had reason to preserve information about encounters with 
herders. On the other hand we have seen that specific farmer clans did have reason to remember 
the nature of their interaction with hunters to preserve their rights to the land. Tatoga were not 
considered "guardians of the land" in the same way as Asi hunters. Differing concepts of territory 
and relationship to the land between hill farmers and Tatoga herders are discussed in Chapter 8. 
Here it is only important to note that western Serengeti people do not associate Tatoga ancestors 
with ritual control over specific places on the land as they do Asi ancestors, represented by clan 
affiliation. Western Serengeti farmers did not build on the plains or encroach much on the 
grasslands ecology of the herders until the last century, while the farmers came into direct 
competition with the hunters for hill ecologies. Farmers acknowledge Tatoga authority as first- 
comers only in relation to mobile ritual items. Herders may not appear in the emergence stories 
because they did not bequeath the ritual possession of the hills for farming. The differences in 
these relationships may have left herders out of the emergence stories but does not diminish their 
importance in western Serengeti history. 

However, despite the importance of these considerations, I will argue that the 
overwhelming reason that emergence traditions do not include stories about relations with herders 
concerns the historical development of relations between herders and farmers over the last 
millennium. The historical linguistic record shows that the most intense period of linguistic 
innovations in herding vocabulary dates to interactions with Mara Southern Nilotic-speakers before 

82 Observations of branding, Nyawagamba Magoto. Bugerera, 14 April 1995 (Nata <f). 

1 000 A.D. This was the period in which East Nyanza-speaking peoples adopted cycling age-set 
names, along with other words concerning life-cycle stages and non-kin relations, including a new 
name for the homestead itself, built around the livestock corral. During the next millennium they 
adopted a few loan words concerning livestock from Dadog-speakers in the region, but innovated 
many more internally. Dadog-speakers also dropped the Southern Nilotic age-set names at this 
time, perhaps indicating that these two communities were not seeking the same kinds of connections 
that facilitated the adaptation of Bantu-speakers to the drier environment of the interior. The 
encroachment of Bantu-speakers into the herding economies dominated by Dadog-speaking herders 
may have led to the increasing separation between these two communities. Bantu-speakers with 
expertise in livestock management simply may not have needed herders as they had when they first 
entered the region. 

The history of farmer-herder relations thus begins with a period of intense relations with 
herders in the distant past, broken by a period of less intense relations, and finally in the late 
nineteenth century by a period of enmity and raiding by yet another group of herders. The 
nineteenth century history of Maasai raiding and their hegemonic position throughout the greater 
Rift Valley was outlined in the Chapter 3. Hill farmers came into conflict, rather than symbiotic 
interaction, with herders during the early part of this century because they were now in direct 
competition for pastoral resources with the increase in cattle wealth as a result of trade. It is 
possible that during this period in which herders became enemies instead of useful allies stories 
about earlier relations with herders were obliterated. Yet it also seems that the relationship with 
Tatoga herders became closer just as the disasters began when farmers sought the ritual expertise 
of herders to solve these new problems. 

Tatoga Herders 

One may gain insight into the past relationship between herders and farmers in the western 
Serengeti, at least during the period in which western Serengeti peoples have been in contact with 
Tatoga herders, by investigating the nature of their interactions today and oral evidence about these 
interactions in the past. The Tatoga now live on the plains south of Ikoma, Nata, Ishenyi and Ikizu 
and call themselves the Rotigenga section of the "Aratoga," known by Bantu-speaking farmers as 
Tatoga or Tatiro. They live in close but separate relation to the Dadog-speaking hunter/gatherer 
Isimajek. 83 

Bantu-speaking farmers and Dadog-speaking pastoralists each occupy particular 
ecological niches in close proximity to each other. When asked about their relationship to the Ikizu 
farmers, Tatoga elders said, "we are the people of the plains, they are the people of the hills-when 
we go over there to the hills we say, we are going to Ikizu; when they come over here to the plains 
they say, we are going to Tatiro." 84 These differences are reflected in the ways in which they 
"map" and name the same landscapes. The Tatoga have names for the gaps between hills and the 
plains while the western Serengeti farmers name the rises and hills. 

The different economic subsistence patterns practiced by each group, in the past, did not 
put them into competition for the same resources, although the Bantu-speaking farmers also herded 
some and the Dadog-speaking herders also farmed. Rather, their occupation of different ecologies 
opened options for trade that were mutually beneficial. The Tatoga traded their livestock for grain 
and established relations of cattle trusteeship in both directions to protect the herds from disease. 

83 G. McL. Wilson, "The Tatoga of Tanganyika (Part 1)," Tanganyika Notes and Records 
33 (1952): 40-41, describes the Tatoga Iseimajek and Rutageink who live in the Ruwana valley of 
Mara, numbering 1,300 people in 1948. 

84 Interview with Merekwa Masunga and Giruchani Masanja, Mariwanda, 6 July 1995 
(Tatoga (?). 

The two communities cooperated in the chase whenever one was raided. Through farmers with 
their large herds of livestock now encroach on the plains ecology of the Tatoga herders, some 
remnant of the past relationship of interdependence remains in practice and in oral tradition. 

Each group, farmers and herders, claims that it was the first to enter this region: each 
describes the other as a newcomer. Both, in fact, are right. In the period of early settlement the 
Tatoga were the first occupants of the plains. Then they left during the nineteenth century famines 
when the Maasai defeated them to follow their prophet south as far as the region of what is now 
Tabora. During this time of radical reconfiguration of social identity the farmers took over the 
status of first occupants. The Tatoga were not present during the most important period of recent 
identity formation. 

Although we might hope that Tatoga oral traditions would shed light on their relationship 
to western Serengeti farmers in the past, these narratives, too, are strangely silent on this account. 
I heard narratives among the Tatoga concerning the division of the different Tatoga groups from 
each other, the division of the Maasai from the Tatoga and the division of Isimajek hunter- 
gatherers from the Tatoga proper. One tradition about the division of Tatoga from the farmers 
appears as an addition to an otherwise independent story in the cycle of narratives about the major 
Tatoga prophets. In this story the farmers are the ones who were willing to kill their mothers to 
secure rain from the Tatoga prophet. It is a statement about the moral superiority of herders over 
farmers rather than a core image. These stories are reproduced in Appendix 4. 

Tatoga emergence stories are characteristic of a wholly different narrative tradition, with 
longer and more detailed stories, usually involving the lives of past prophets, both real and 
mythical. These are the stories of heroic deeds and miraculous happenings that ordered the world 
as we know it. In fact, Tatoga oral traditions bear a strange resemblance to the Sonjo stories of 
their prophet Khambageu, whom they said performed many miraculous deeds. This is not 

surprising given the long-term interactions of the Sonjo with Dadog-speakers uncovered through 
the evidence of historical linguistics. 85 What seems more surprising is that western Serengeti 
farmers demonstrate so little of this influence. In contrast to the western Serengeti stories of first 
man and first woman as unifying stories of hunter and farmer, the Tatoga stories are of division 
and exclusivity. 

The Tatoga retain a distinct identity as herders (as opposed to farmers and 
hunter/gatherers), but at the same time are integrated and deeply implicated in the most intimate 
aspects of the lives of western Serengeti farmers. The Tatoga learn the Bantu languages of their 
neighbors but almost no farmers speak Dadog. Tatoga distinctiveness has been responsible, in 
part, for their ability to play a highly influential role among their neighbors as gifted prophets. All 
ethnic groups in the area still call on Tatoga prophets for rainmaking, healing and protection 
medicines. In some Bantu-speaking ethnic groups descendants of a particular Tatoga prophetic line 
hold leadership positions of great power. Some of the most commonly told stories of the nineteenth 
century raids and famine revolve around the prophecy of a Tatoga healer who gave advice they did 
not heed, causing disaster. 86 

The Ikoma have a tradition in which a Tatoga prophet from Ngorongoro gave them their 
most sacred object of collective identity, the Machaba, a set of elephant tusks. Because of this, the 
descendants of that prophet must participate in the rituals in which the Machaba play a part and the 
Tatoga are designated as "father" of the Ikoma people. 87 The first choice of Ikoma people for a 
colonial chief was a Tatoga prophet. In one version of the Ikoma emergence story [reproduced in 

85 Ehret, Southern Nilotic History , pp. 55-62 

86 See Chapter 1 0, the war of Ndabaka. 

87 See section on the Machaba in Chapters 7 and 8. 

Appendix 3] Mwikoma left Sonjo because the Tatog prophet told him to seek a new land. It makes 
some sense that the Ikoma would have the strongest relationship with the Tatoga because they are 
the only one of the western Serengeti peoples to live in closer proximity to the grasslands than to 
the woodlands. Both Ishenyi and Ikoma traditions say that they killed their own prophets and were 
thus cursed to rely on Tatoga prophets. 88 

The ritual precedence that hill farmers give to Tatoga in prophecy indicates a historical 
relationship quite different from that between Bantu-speakers and Asi hunters. Because herders 
had access to the most conspicuous form of wealth in livestock, they occupied a strong position in 
trade or marriage negotiations and by that gained prestige. Livestock was one of the few 
convertible forms of wealth in a society in which land was abundant and freely accessible. If 
prophetic efficacy was judged by its material benefits, hill farmers may have sought out Tatoga 
prophecy to gain access to the wealth that it represented. Colonial investigations of local politics 
concluded that the Tatoga once held a dominant position from Lake Victoria, up to the Mara River 
and across the Serengeti plains to Ngorongoro Crater. " In the mid-nineteenth century the Maasai 
defeated the Tatoga in a decisive battle for the Crater. Sometime after this, cattle diseases 
devastated their herds and many followed the prophet south. Ikoma elders said that after the 
livestock deaths of the early colonial period the Tatoga were reduced to work for them as cattle 
herders to regain their livestock." 1 

The stories concerning Tatoga as influential prophets and "fathers" seem to date to mid- 
nineteenth century, just before or into the period of disasters. It may be that when the first signs of 

88 Machota Sabuni, Issenye, in a letter, 23 March 1997, recorded by Nyawagamba 

89 R. S. W. Malcolm, "System of Government," p. 4, 1937, MDB. 

90 Interview with Bokima Giringayi, Mbiso, 26 October 1 995 (Ikoma cf). 


the disasters to come began to be felt, (through localized famine, the introduction of new diseases, 

the appearance of the first caravans in the area, and the encroachment of Maasai into hill farmer 

territory) hill farmers turned to the Tatoga for help. The Tatoga would have been natural allies in 

times of crisis because they lived in a different ecological niche and so famine and disease affected 

them differently. Yet just as with the hill farmers' relationship to the hunters, while learning from 

and getting aid from the Tatoga herders, the farmers appropriated their expertise and gained 

dominance in the region. The prosperity and success of Bantu-speakers resulted from their ability 

to break down the ecological/economic boundaries defining the identities that had allowed the 

regional system to function. 

The Prophetic Stories of Masuche and Zeaera 

Through an interpretation of stories appropriated from Tatoga tradition but told by hill 

farmers as their own, one gains insight into the historical process by which hill farmers came to 

dominate, linguistically, culturally and demographically, over herders and hunters in the region. 

For example, in Ikoma elders took me to see a rock with natural depressions, resembling a bao 

board game of parallel sets of holes in which seeds are placed as counters for the play. They told 

me that this was Masuche's bao, which he used to trick God, the sun. [See Figure 5-5: Two 

Versions of the Story of Masuche's Bao.] The Ikoma story told at the site went like this: 

They call this Masuche's Bao (agoreshi e Masuche). Masuche played bao with God, the 
Sun (lriobaj here. Masuche's cattle went out and grazed themselves and came back at 
night because God was close to Masuche. One day they quarreled and the Sun went 
home in anger. Only some cattle had come home to the corral by then, when it got dark, 
because the Sun left. So the ones who were left out became the wild animals — Masuche 
named them zebra, gazelle, topi, impala, and all the others. That is how the wild 
animals came to be. They are Masuche's cattle. 1 " 

" Interview with Moremi Mwikicho, Sagochi Nyekipegete, Kenyatta Mosoka, Robanda, 
12 July 1995 (Ikoma cf). See the story of the Bao game in Buganda tradition, Wrigley, Kingship 
and State , pp. 101. 


Masuje is one of the most important Tatoga prophets and Tatoga elders told me one part of 

his story in this way [The full story is reproduced in Appendix 4]: 

Giriweshi was born of a woman and was the son of God. His son was Masuje who 
tricked God, the Sun, in a game ofbao because he knew how to make the stones revolve 
endlessly without coming to an empty hole. Because the game never ended the Sun never 
set and Masuje's cattle could graze far from home. The Sun became angry and retreated 
into the sky, taking Giriweshi with him.' 2 

Ikoma appropriation of Tatoga stories illustrates the historical process by which Bantu- 
speaking farmers forged new identities by adopting elements from the diverse people who 
surrounded them. The Ikoma could only know these stories of Tatoga prophets if they had been in 
close interaction with their Tatoga neighbors, participating in events at which they naturally told 
these tales. This indicates a relationship of some degree of mutuality. Yet the Ikoma men who told 
me this story did not suggest that they knew that Masuje was a Tatoga prophet. Ikoma elders had 
completely assimilated Masuche into a "traditional" Ikoma narrative. 

Just as early settlers recognized and learned from the greater expertise of preexisting 
pastoralists to cope with an unfamiliar environment, Ikoma acceptance of these foundational stories 
of how the world was created recognizes the efficacy of their knowledge about the world. Many 
examples exist in East Africa of people appropriating the origins of those with more power to 
increase their own standing. This may be why the Kuria claim to have come from Egypt or the Jita 
from Buganda and Buhaya. Here, Ikoma do not claim Tatoga origins and parentage but 
appropriate the source of Tatoga knowledge, and so their power, in prophecy. This story connects 
between Tatoga prophecy to wealth in livestock. In the Ikoma version of the story Masuche, a 
Tatoga prophet, controls all of the original livestock and therefore also the wild animals. 

92 Interview with Stephen Gojat Gishageta and Girimanda Mwarhisha Gishageta Issenye 
27 July 1995 (Tatoga <f). 

Moremi Mwikicho, Sagochi Nyekipegete, Wilson Machota, Kenyatta 
Mosoka, Robanda, 12 July 1995, at Masuche's Bao in Ikoma. 

~ { 

■ > 




Stephen Gojat Gishageta and Girimanda Mwarhisha Gishageta, Narrators of 
the Tatoga Masuche Story, Issenye, 27 July, 1985 

Figure 5-5: Two Versions of the Story of Masuche's Bao 


Another example of Tatoga narrative style in hill farmer tradition is an Ikizu story of one 

of their prophets, Zegera, whose miraculous birth and actions resemble nothing more then the 

Tatoga prophetic accounts. The story begins with an alternate Ikizu emergence story to the one of 

first man and first woman told in the last chapter. In this version Muriho, a hunter, establishes the 

country by using his prophetic powers. It is significant that the elder tells this story (which is one 

of the very few stories of miraculous prophets that I collected among the hill farmers) in relation to 

the story of driving out the Tatoga from Ikizu. For all of the myriad times I asked people about the 

relationship between Tatoga and Bantu-speaking communities this is the only time that I got a full 

narrative. It was also the only response that suggested a relationship of conflict rather than of 

coexistence and peace. The full story is reproduced here because it is a unique combination of the 

older emergence stories with Tatoga prophetic tales. 

Samweli: Muriho came from the west, the lake, Nyanza, through the north to Kisu, but 
he did not go through Gorogosi. He was a healer and had medicine. When he left 
Kisii, Muriho went first out to Ngoreme, to the mountain ofMangwesi. Mangwesi was a 
hunter and invited Muriho to be his guest. Mangwesi was a Ngoreme. Mangwesi told 
Muriho, "you are my friend, why not stay here in my country and build with me? " 
Muriho said, "I am going over there to the mountain, the big one that I dreamed about, 
Chamuriho. " 

So Muriho left and came to his land and was welcomed by a man named 
Nyamwarati. He learned that the Hengere were bothering those people. After living 
with Nyamwarati for a while Muriho asked him, "who is the big man of this country?" 
Nyamwarati answered, "I am, but lam defeated by the terrible people here called the 
Hengere." Muriho said, "if 1 drive them out what will you give me?" Nyamwarati 
answered, "if you drive these people out of my country I will give you my daughter to 
marry. " Muriho tried first to send wild buffalo to bother them day and night in their 
homes, but they did not leave. He sent snakes to bite them and scare them, but they did 
not leave. Then he used ants who would bite them at night, but they did not feel it. He 
said, "what shall I do with these people, the moon is nearly gone. " Then he used bees on 
them and the very day that they stung them the Hengere moved away. The bees drove 
them off. 

There is a point I forgot, Muriho had spoken to Nyamwarati, saying, "I am a 
prophet and we cannot live in one homestead. " So he went and established his home on 
the mountain at Itongo Moriho. Nyamwarati lived in Sarama. [. . .] The Hengere are 
the people of the Congo, the pygmies. When the Hengere were chased out Muriho 
encircled the mountain and the land with protection medicine to make it safe so they 
would not return. 


The first wife ofMuriho was the daughter ofNymwarati — Wanzita. She gave 
birth to Mughabo. We say, "we are the Ikizu, the people of Wanzita and — . " 
Audience: Mughabo. 

Samweli : Muriho had this many wives — /holds up fingere/ 
Audience: Eight. 

Samweli: Muriho had eight wives from different places, but the first wife, the big house, 
was from Ikizu, Nyamwarati's daughter, from Salama, Kombogere, Zahya, the big house. 
Wanzita gave birth to Mughabo and Mugabho gave birth to Kishoko. The sister of 
Kishoko was Wasatu and they built a house together . . . Then Kishoko gave birth to 
Mabere. Mabere was a prophet. Mabere gave birth to her son, Zegera. 
Zegera was born speaking and holding his medicines ... but wait I must go back. 

Before Muriho came the Taturu were here, and together with the Rangi they 
were also driven out by the Hengere. They left, but when the Taturu heard that the 
Hengere were gone they came to ask Muriho if they could build here again. So Muriho 
agreed and asked what they would give him. They said they would give him a wife and 
this is the house of the Batatiro. They came back from Mbulu where they had gone. The 
first person to come was Gambasarakwa, who built at Sarakwa Hill, this is a Tatiro 
place. The next person was Gambamiri, who gave his daughter to Muriho and built at 
Kirinero ... He gave his daughter to Muriho because in the past when you wanted to 
build somewhere it was important to give something in return. After he gave his 
daughter, he was allowed to build here. Muriho was the son-in-law ofNyamwarali who 
passed on his authority over the land to Muriho, he said, go ahead and live here and 
herd. He gave him all of the hills to keep the enemy out. Do you understand? 
Jan: Yes. 

Samweli: Write it then. Muriho became a soldier of the old man to guard the hills so 
that no enemy would come. So that is how the Taturu came. 
Jan: Where did they settle, in the hills there or where? 

Samweli: Muriho's Tatiro father-in-law built in the hills a little there in Kirinero. The 
others were on the plains at Sarakwa, down there. They were so few anyway. Then 
Muriho disappeared. After the Taturu came in, they spent some years and he 
disappeared and returned to Ngoreme, Maji Moto, to his friend Mangwesi. [. . .] Now 
after they said that Muriho was dead, but there was no grave, the Taturu began to 
agitate. They wanted authority over the land for themselves and they had their own 
medicines. They harassed the Ikizu constantly for many years, maybe ten. Then this one 
was born, Zegera, the one who was born speaking. 

When Zegera was born, he said to his mother, "give me the milk of a white cow 
or a black cow. " The elders were afraid, "how is it that an infant speaks?" So they 
sounded the alarm call and brought everyone together. When everyone came together, 
they asked, "what is the alarm all about? " The child spoke again, "ask my mother and 
father, I asked for milk but instead they gave the alarm, what is this all about? " The 
people were amazed. So they left and went to the big Diviner who said, "this child was 
born speaking and had everything, even his own medicines in his hand in a bag, together 
with millet and sorghum seeds. " "He is coming to save the Ikizu from the oppression of 
the Taturu. " The Diviner said, "take this child to the crossroads and lay him there with 
a black cow hide over him, let him sleep on the crossroads. " "If he is hurt by any animal 
by morning know that you have not yet found your savior. " Yet if you find that he is 
unhurt in the morning know that you have already been saved. " So they laid him there 


and slept uneasily in their homes all night, listening. At dawn they realized they had 
passed the night without hearing the laugh of the hyena and they knew that he was all 
right. They went and found him well, they had laid the hide on top of him but he was 
sitting on it watching them come. They took him to do sacrifices on his behalf Then the 
Diviner said, "build a house for him right where he slept. " Zegera made medicines 
again after he was one year old — a child that was born speaking — he was made like 
Jesus — he came miraculously. So he made medicines that they put in the water where 
the Taturu cattle drank. The cattle in the whole area began to die. A great many died 
and when they died the Taturu moved, they went back to Mbulu, others to Sukuma, others 
to Majita. they dispersed . . . After this the land was of the Ikizu. K 

This story acknowledges the Tatoga as having been in the land first, but ineffective against 
the agents of destruction and thus not in control ritually as guardians of the land. Again the hunter, 
Muriho in this case, marries a local woman to gain access to the land. This version acknowledges 
his wife, Wanzita and her child, as the founders of the "houses" of Ikizu. The genealogical line of 
prophets runs through these women and their sisters. His journey to the mountain of Mangwesi in 
Ngoreme and his final settlement at the mountain of Chamuriho resonate with the emergence 
stories that have already been told. 

The story of Zegera depicts a contest for authority over the land between the Tatoga and 
the Ikizu. Although the Tatoga had been in the land first, they were not able to defend it against 
the Hengere and only returned after the Ikizu hero, Muriho (a hunter), had done the job. They 
established reciprocal in-law relations to live together on the land. Nevertheless, after Muriho was 
gone, the Tatoga tried to regain their former authority through prophecy. The Ikizu produced a 
more powerful prophet and drove them off by killing their cattle with medicines. The Ikizu 
challenged the hegemony of the herders by using prophecy, the source of Tatoga power, against 
them. They defeated the Tatoga by killing their cattle, the material basis of their power. By 
controlling a prophetic narrative of their own the Ikizu at once accept Tatoga power and attempt to 

n Interview with Samweli M. Kirimanzera, Kurusanda, 3 August 1995 (Ikizu cf). 
Samweli is a healer by profession and tells this story to defend the authority of the prophetic line of 
Muriho in Ikizu as having priority over that of Nyakinywa from Kanadi. 

overturn it and assert their own dominance. This story provides further evidence that although the 
Tatoga were among the first-comers, they were not the ritual guardians of the land. Samweli's 
narrative claims that the Tatoga could not protect people against the Hengere. 

Western Serengeti farmers are clear in their narratives that they control the land and 
maintain the relationship with it necessary for prosperity. They contend that the Tatoga are 
interlopers and newcomers on the land. However, their silence concerning the relationship between 
herders and farmers in the emergence traditions may indicate that the Ikizu gained a certain kind of 
power by the incorporation of Tatoga prophecy as their own. The Ikizu took the mobile power of 
the prophet, connected with the wealth and prosperity of livestock, and attached it to a fixed site on 
the land. Western Serengeti people still accept the ritual precedence of the Tatoga without 
acknowledgment of their role in the emergence of farming communities on the frontier. People 
value the Tatoga for their ritual power, in spite of their diminished material and demographic 

The few stories that remain, such as the one told by Samweli, suggest that farmers gained 
ascendency over pastoralists who first inhabited the plains, by their ability to incorporate rather 
than to exclude. Farmers gained herding expertise from their herding neighbors, but they also 
learned the more esoteric knowledge of prophecy that gave them additional tools for maintaining 
successful settlements on the frontier and meeting the challenges of the disasters. The Tatoga and 
their Bantu speaking neighbors shared the land and maintained relations of reciprocity because they 
occupied different ecological niches: the plains and the hills. Yet just as with the Asi, conflict and 
competition were also important elements in the emerging historical relationship. Because the 
farming communities were so adept at incorporating new knowledge and practice they began to 
take over areas once dominated by the Asi and the Tatoga. 


The interpretation of the asimoka traditions through the core spatial images provides a 
glimpse into important aspects of the distant past through the internal lens of local historical 
imagination. It allows us to come to tentative understandings of how Bantu-speaking immigrants 
on the frontiers of an inter-cultural environment forged unique new communities in an unfamiliar 
landscape. They did this by drawing on their own traditions as well as incorporating the 
experience of their neighbors. The image of the female farmer bringing the male hunter into her 
home and together giving birth to a nation is one that aptly symbolizes this development. Through 
the incorporation of farmers into hunter lineages they gained ritual control over the land. 

Farmers located their settlements on the hills and at the interstices of the different 
ecological zones to realize a wider economic range and to facilitate interaction with neighbors from 
whom they could learn. Different kinds of social interactions with hunters and herders produced 
very different concepts of social identity and the oral traditions that represent them. In each case 
the success of Bantu-speaking communities was a result of their ability to incorporate rather than 
to exclude, to widen their options, rather than to specialize. Their rhizomatous growth has drawn 
on many connected underground stems, thriving in an interstitial space. 

These long-term processes of inter-cultural interaction on the frontier must be kept in mind 
as we turn to the investigation of a different kind of social identity— that of clans. The 
ecological/economic identities of farmer, hunter and herder formed a regional system of difference 
that allowed for interdependent relations. Yet as we have already seen, people needed to find ways 
of crossing these boundaries to gain access to the expertise, knowledge and resources of others that 
were crucial to their prosperity. They mediated these boundaries primarily through clans, based on 
the very old idiom of kinship. 




Clan histories, in contrast to the stories of ethnogenesis discussed in the previous chapter, 

are a separate but related kind of emergence story. They constitute another genre of oral tradition. 

Elders presented these stories as some of the oldest traditions, reaching back "before the 

grandfather of our grandfathers." Like ethnic emergence stories, they contain layers of meaning 

related to different time frames-the elaborations of more recent experience and also core spatial 

images that represent some regional generative principles at work in the more distant past. This 

chapter explores clan traditions as they shed light on community organization and dispersed 

regional alliances before the era of disasters. It demonstrates how the meaning of clans changed as 

their historical context in relation to larger scale social identities changed. 

In this chapter I argue that clans are not fixed and timeless social structures but rather 

inherently flexible units for creating cohesive local communities and for providing regional 

"pathways" whereby people gain access to the resources and knowledge of others. The structural 

interpretation of Cans by both anthropologists and historians, as units in a segmentary lineage 

system, and by oral traditions told today, as subunits of ethnic groups, distorts an older meaning of 

clan still recoverable in the core images of oral traditions. Rigid interpretations of ethnically 

bounded Cans do not work in the western Serengeti because the same Can names are found among 

many ethnic groups throughout the wider region, rather than uniquely in one ethnic group. This 

configuration results from a history in which dispersed Cans and Can narratives preceded ethnic 

groups and ethnic narratives in the region. Clans were the social unit which existed over the 


longue duree; within them people transmitted emergence stories, and these later became ethnic 
stories. Although clan narratives have most often been interpreted as "migration stories," I analyze 
them as the formalized recognition of "resource and knowledge pathways." 

I argue that clan stories represent long-term generative principles through the core spatial 
images of the homestead and of dispersed regional pathways. The homestead image allowed people 
in the past to unite diverse peoples into cohesive residential territories. In the context of the late 
nineteenth century disasters when western Serengeti people began forming ethnic territorial 
boundaries they used the genealogical aspect of the homestead image to imagine clans, like 
lineages, as subsets of ethnic groups. In the period before the disasters clans seem to have 
functioned primarily to mediate diverse kinds of regional social boundaries marked by economy, 
bodily marking, geography or expertise. These large-scale social identities (like 
herder/hunter/farmers) controlled access to resources and knowledge, creating a regional system of 
exchange and interdependence. Clans crossed these boundaries, giving their members access to 
resources and knowledge that others controlled. These clan networks were formed both by the 
settlement mobility of lineages and by the metaphorical association through common praise names 
and avoidances of people controlling particular suites of knowledge and resources. Today, 
prophets use clan networks as the basis of their power; one of the few remaining examples of clans 
as dispersed pathways. 

Anthropologists struggle with the problem of defining "clan" because what they describe as 
"clan" in one society may not have the same function or organization as "clan" in another. While 
people use various words for "clan" across the Mara Region, they refer in every case to a social 
identity made up of people who trace their descent to a common ancestor, recognize commonly held 
territory, and have occasion to take communal action. Dispersed throughout the region one finds 
clan groups with the same names, sharing a sense of common origins and avoidances. What 

distinguishes clan from lineage in the Mara Region is that people of the same clan cannot ascertain 
their exact relationships to one another, as can the members of one lineage, nor are precise 
genealogical relationships deemed important. Each clan also claims a set of oral traditions that 
explains its origins and dispersal throughout the region. 

I interpret clan narratives as representing part of the longue duree history of this region in 
terms of generative principles, rather than specific events. In Chapter 4, 1 demonstrated from 
linguistic sources that the lineage principle is very old in the region, brought by Lakes Bantu- 
speakers as they forged new kinds of settlements on the frontier. Yet because lineage provided a 
flexible and adaptable strategy we do not know exactly how people would have used it in the 
distant past. For the same reasons, it is also difficult to tell which specific clan histories are older 
than others. Many scholars have argued that the clans claiming "first-comer" status were the first 
to enter the region. Nevertheless, these clans may have usurped those rights from others before 
them and the clan units themselves likely changed over time. 

Clan histories do, however, seem to be old: the emergence stories presented in Chapter 4 
and 5 were probably clan stories before they became ethnic stories. I have argued that the core 
spatial images of ethnic emergence stories represent the principles of production and reproduction 
used by settlers on the frontier in the distant past. Yet if people carried these stories in oral 
memory as the stories of clans, then clans might have been the units that organized local 
communities within the larger regional context of a hill farmer's identity. The clan stories which 
scholars have often interpreted as the "migration" histories of ethnic groups seem to reflect an 
earlier context where clans provided the means for crossing the boundaries of regional identities 
such as farmer, herder and hunter, among others. "Migration" stories make connections among 
diverse peoples across the region; they describe the pathways that gave people access to knowledge 
and resources beyond their communities. 


In this chapter I analyze two different kinds of clan narratives which reflect the two 
contexts in which clan identity functioned. The first are clan "migration" histories; the second are 
histories of clans as "children" of the ethnic group's founding parents. The core spatial images of 
the first type are movement through diverse regional pathways. Those of the second type are 
images of family relations within a homestead. Each type emphasizes an aspect of clan identity 
that served the purposes of people operating in different historical contexts. 

Yet both kinds of core images represent aspects of the generative principle of clanship that 
has operated over the tongue duree. Although homestead images were effectively used to join 
clans into ethnic groups in the reconfigured narratives of the late nineteenth century, they also 
capture an aspect of clanship that is much older. In Chapter 4, 1 demonstrated that the core spatial 
images of the ethnic emergence stories represented gendered homestead space. These same 
homestead images may also have functioned in the period before the disasters to create the 
organizing principle around which settlers built local clan-based communities-like the relations of 
a homestead but on a larger scale. This is shown by an analysis of the words for lineages and 
clans, used from the time East Nyanza Bantu-speakers first entered the region. All these words 
represent homestead space (house, gate, hearthstones). 

Clans as Children of the Homestead 

The first type of clan histories presented in this chapter are those that have been 
incorporated into ethnic stories. They are analyzed not only as they represent clan identity in 
relation to ethnic identity in the post-disaster period but also as they represent clan identity as a 
means for organizing local communities before the disasters. Elders often integrate the origin of 
clans into their narrations of the ethnic emergence stories. This version of the Nata asimoka story 
ends with the creation of the four Nata clans or hamate by the children of Nyamunywa and 


The woman became pregnant and gave birth to a son. Then she gave birth to a 
daughter, and in total four boys and four girls. When they were grown, they were 
married to each other. This is the reason that Nata inherit through the woman's side. 
The children made the clans of Nata. The place where they lived is called Bwanda. 
When they got to be too many, they divided into the saiga [age-set cycles]. ' 

An Ikoma elder, Mahewa, along with my colleague Wilson, listed all of the clan names and 

their founders as subunits of the greater Ikoma people [See Figure 6-1 : Narrators of Clan 


Jan: What is your clan? [1 used the Nata term for clan, "hamate."] 

Mahewa: What does she want to know? 

Wilson: She wants to know your land (ekyaro). 

Mahewa: Oh, so that is it! Ok. 

Wilson: Your gateway lineage (ekehitaj. 

Mahewa: Abarache of the lineage of Obohimaro. [. . .] 

Young man: Let me explain this to you. All of us who live here as a tribe are the Ikoma. 

The Ikoma are divided again among the gates (ebehita) which are the Hikumari, Rachi 

and others. The gate of this man is the Rachi. The Rachi are also divided among 

smaller groups. [. . .] 

Jan: Among all the Ikoma how many clans are there? 

Wilson: There are eight clans. Who is the eldest of them? 

Mahewa: Mrachi, then Hikumari, then Getiga, then Himurumbe, then Gaikwe, then 

Mwancha, then Serubati, then Sagarari, that is all. 

Wilson: But explain to her so she understands, the name of Mrachi was what? 

Mahewa: Mgosi. 

Wilson: This man was named Mgosi. 

Mahewa: Hikumari's name was Kumari. 

Jan: Was this the name of the one who founded the clan? 

Mahewa: Yes 

Wilson: And for the Getiga was Mago. Murumbefor Himurumbe. 

Mahewa: Gaikwe, I do not know his name, he was an Asi. Mwancha was Marakanyi. 

Serubati was Nyawatika. Sagarari was Mumare. 

Jan: Are these their names or the places where they came from? 

Wilson: They are the names of the eldest. The ones who began each clan. 

Mahewa: They are the children ofMwikoma, the founder of all Ikoma. 2 

1 Interview with Megasa Mokiri, Motokeri, 4 March 1995 (Nata a"). 

2 Interview with Mahewa Timanyi and Nyambureti Morumbe, Robanda, 27 May 1995 
(Ikoma cf). 


Nyambureti Morumbe and Mahewa Timanyi with Mahewa's wife between, 
Robanda, 27 May 1995. 

Samweli M. Kirimanzera, holding staff and black tail, with friends and research 
colleague Kinanda Sigara (standing), Kurusanda, 3 August 1995. 

Figure 6-1 : Narrators of Clan Histories 


These narratives fully incorporate clans into the ethnic account, with the clan founders 
represented as children of the original parents. These accounts order the relationships among 
people of different clans within one ethnic group. The Ikoma clans acknowledge their unity as 
children of the same parents, and also recognize their individuality and differences in seniority. 
Many historians and anthropologists have explained this family idiom as a mechanism for ordering 
internal political relationships. 3 The status of "first child" often gives that clan ritual precedence 
within the ethnic group. David Newbury's account of the royal First Fruits ceremony on Ijwi 
Island in Lake Kivu, shows the ritual role of each clan in relation to its social position within the 

Western Serengeti peoples, too, identify clans with certain ritual and functional roles 
within the ethnic group. Nata give a representative of the clan of the first son, Gaikwe, a token 
inheritance at every funeral before dividing the rest of the inheritance. In the Nata story told by 
Sochoro Kabati in Appendix 1, first woman Nyasigonko was from the Getiga clan of Sonjo and 
first man from the Gaikwe clan, an Asi hunter from the wilderness. Gaikwe, as the hunter, has 
"first-comer" status-since in oral memory it was the Asi who taught farmers how to hunt and 
make fire, and gave them rights to the land. The two Ikoma clans that keep the elephant tusks, the 
Machaba, have this authority as first sons and first-comers. In Ikizu the rainmaker "chiefs" come 
from the clan of first woman, Nyakinywa, while the prophets come from the clan of Isamongo, first 
man. The Ngoreme and Ikizu ethnic groups that are, by their own account, amalgamations of 
many different peoples, nevertheless tell a story of an original nucleus of clans descended from the 
first parents which then drew in other peoples from elsewhere over time. 

3 For an example of this argued on the basis of regional sources see Kenny, "Stranger from 
the Lake," p. 9. 

4 Newbury, Kings and Clans , pp. 200-226. 

From Clan to Ethnic Histories 

While clan stories are now presented as subtypes of ethnic stories, it seems more likely that 
the narrators of the ethnic emergence stories presented in the last chapter created them by joining 
together the narratives of the clans that made up newly forming ethnic groups. Although I have 
hypothesized that hill farmers in the early period of frontier settlement shared a founding myth 
containing the core spatial images of the homestead and the ecological landscapes of economic 
subsistence patterns, these stories had to be transmitted and maintained by a coherent corporately 
based social group, which 1 suspect was the residential clan community. 

Quite a bit of evidence supports the claim that some form of residential organization based 
on the idiom of clan/lineage existed among Bantu-speakers of the Lakes Region, at least from the 
time that they entered the western Serengeti. 5 At the same time, I found no evidence for 
organization into ethnic groups, as we now know them, until the late nineteenth century. Abuso 
demonstrates for the Kuria that, "before the twentieth century they did not know themselves as 
Abakuria but by either their various clans or by the "provinces" from which they came." Kuria 
identity emerged as an amalgamation of clan "provinces" during the colonial era, with the name 
"Kuria" being first applied to the whole group by early chiefs for political reasons. 6 Clans, rather 
than ethnic groups, were the unit that organized communal resistance to colonial measures. 7 All 
available evidence suggests that before the era of disasters the residential clan was the largest scale 
level on which communities organized themselves and took corporate action. Other kinds of large 

5 Schoenbrun, A Green Place , pp. 168-181. 

6 Abuso, A Traditio nal History , p. 7. The same conclusion is reached by Tobisson, 
Family Dynamics , pp. 94, 115-1 16. 

7 For the Kuria see Tobisson, Family Dynamics , p. 95, she also concludes that the Kuria 
"only began to look upon themselves as a distinct group by the beginning of the lO" 1 century." p. 
94. See also Prazak, "Cultural Expressions," p. 45. 

scale social identity did exist, such as those divisions based on ecology and economy discussed in 
the previous chapter. Nevertheless, people did not organize their local communities on the basis of 
subsistence identities. Although the bonds between hill farmers represented a kind of "ethnicity," 
they were much too loose to secure the claims of reciprocal obligation. When the pressures of the 
nineteenth century disasters demanded larger scale identity formation within bounded units, 
western Serengeti peoples formed ethnic groups in each area as amalgamations of residentially 
based clans. (This process is explained in more detail in Chapter 9.) 

Since we know that territorial clans existed before the nineteenth century disasters and that 
ethnic groups developed as a result of those same disasters, then it seems reasonable to presume 
that clan histories, in some form, preceded ethnic histories. Clan histories and clan identity also 
changed as a result of the nineteenth century disasters but the stories retained the core images 
associated with earlier social forms even as they became the stories of new groups. 

The contradictions inherent in the stories themselves testify to these changes in the social 
basis of the narratives. Clan "migration" stories still exist separate from and cross the boundaries 
of ethnic narratives. Clan "migration" stories and ethnic narratives which incorporate clans are in 
many ways incompatible with each other. For example, if the clan founder Gaikwe was the son of 
first man Nyamunywa how can one tell a Gaikwe history without reference to the larger ethnic 
group of which he is a part? How can Nata identify first man, Nyamunywa, as a Gaikwe clan 
member if Nyamunywa's sons founded the clans? Many elders were clearly uncomfortable with 
talking about these two kinds of histories in sequence. In fact, the account of Samweli, reproduced 
below, was one of the few that admitted to a relationship between clans of the same name in 
different ethnic groups. Most elders said that the Gaikwe of Nata and the Gaikwe of Ikoma had no 
connection except in the coincidence of names. In the clan history told below the Ikizu narrator 

Samweli avoided the issue of whether he could claim clanship with the Gaikwe of Ikoma by 
describing the overall unity between the Ikizu and the Nata , 

The process of ethnic formation is recent enough that inconsistencies remain in clan stories 
that narrators have not yet fully assimilated into an ethnic account. Elders still readily identify the 
characters in the ethnic emergence stories with their clan. The stories some elders tell as clan 
history in one ethnic group, elders from another ethnic group may tell as ethnic history. For 
example, a Gaikwe clan elder in Nata told me the story of hunters following the wildebeest 
migration from Sonjo, and Ikoma elders told me the same story as one about ethnic emergence. 
Particular clan histories often include what is now known as the ethnic emergence story as part of 
the clan narrative. When I asked about Nata origins, one knowledgeable elder told how members 
of a particular clan from Gusii came hunting at the mountain Magaka in Nata. Nata people 
incorporated these hunters as friends, and later as family, into an existing lineage. When asked 
how it was that the people who lived there accepted these hunters, this elder began a narration of 
the Nata origin story of first man and first woman. 8 Clan and ethnic stories overlap so much that 
the narrators themselves have trouble distinguishing them. 

The story of Ngoreme ethnic origins, told by a Timbaru clan elder, is a good example of 
this merging of clan stories into ethnic stories. This elder seems to be elaborating an ethnic origin 
history around an Iregi clan praise shout that goes like this: "the Ngoreme of Isabayaya, of 
Wandira, of Mangwesi-Regata-Manyere, waters which trickle down from above, the Ngoreme of 
Nyahaba."' The Iregi clan, among others who claim origins in Sonjo, say that Isabayaya was their 

8 Interview with Maguye Maginga, Nyeketono, 21 June 1995 (Nata tf). 

' "Abangoreme ba Sabayaya re Wandira re Mangwesi-Regata-Manyere-Manchira 
Baigora (Batahera Manche na Migeri);" interview with Apolinari Maro Makore, Mesage, 29 
September 1995, in a handwritten manuscript of Ngoreme history. Praise shouts or a praise names 
are commonly used across Africa to commemorate a famous person, group or event. They are 


clan founder. 10 The Iregi are also those whose story of dispersal from Nyeberekera during the 

disasters was told in Chapter 3. Their name, derived from "Regata," links them to Sonjo. Their 

clan origins have become the origins of all Ngoreme. This is the ethnic story built by Mzee Silas: 

Silas: We Ngoreme, as I learned from the elders who told me, they said that we came 
from the southeast. Maybe the areas to the north or east. [. . .] To say for sure, the 
places themselves are not named. To the north over there, the area of Sonjo, where we 
border south in the hills, that is where we began. From there they came to the area of 
Mangwesi Mountain. Before they got to Mangwesi Mountain they were harassed by 
others, at the place ofManyere. The elders said that Mangwesi was not where Manyere 
was from, Manyere came to Mangwesi. So Manyere must be the name of a person, not a 
place. But the place was named after him. [. . .] The Taturu call us Manyerega. So I 
think that this person Manyerega was the same as this person Manyere. [. . .] After 
coming here Manyere and his people were dispersed, maybe because of war. His son 
was Wandira . . . He was a farmer, his other brothers were hunters. But he liked to 
farm and liked to find good fertile soil to farm. This was Wandira. [. . .] After Wandira 
was Isabayaya. He was the son of Wandira. At this time they had come from the 
southeast, they had come as far as Serengeti in the hills. They came to Mangwesi 
Mountain and then to lkorongo. But before they got to Ikorongo, there at Mangwesi 
Mountain, the lkoma separated off and went to Ikoma. The lkoma liked to hunt. So they 
went off on their hunt ... I do not know, this is what the elders thought. They went to 

But the others went off looking for soft earth to farm. The soil here was good 
because the elephants had dug up the soil over there and it dried up, so they had to 
move. At that time they farmed with sticks. There were not even hoes then. That is what 
they told me. So they came looking for a place to farm and to build their houses safe 
from enemies. The hills would protect them. They did not have anyone to bother them. 
That is how they came to Ikorongo. At Mangwesi Mountain Isabayaya was born to 
Wandira there. Then after Isabayaya moved to Ikorongo, he gave birth to Mongoreme. 
So that is why this is called Ngoreme. Even today we say, "we Ngoreme are the people 
of Isabayaya and Wandira. 
Jan: Do you have kinship with the Sonjo? 

Silas: We have the praise name of the Sonjo, that they have the irrigation channels. We 
pray Bigoro Manche ra Mogera, " the water that trickles down from on high. " This means 

performed in public gatherings like dances or feasts. See Barber. 1 Could Speak , for an analysis of 
Yoruba praise names. 

10 P. Haimati and P. Houle, "Mila na Matendo ya Wangoreme," (unpublished mimeo) 
Iramba Mission, 1 969. Philipo Haimati is from the Taboori clan, which along with the Maare are 
said to be the first to arrive from Sonjo. 


that we came from Sonjo. When we pray, we ask for a blessing from where we came. 
When we do eldership titles, we pray this. We pray this to the ancestors from Sonjo. ' ' 

The process of the incorporation of clan histories into ethnic accounts is ongoing and 
documented in the written sources. Today, elders rarely tell clan histories any more as an 
independent narrative form. When Musoma District Officer E. C. Baker questioned elders in the 
1920s and early 1930s about clan names, avoidances, and place-names, he got quite thorough lists 
throughout the region. 12 In my research, few elders could recite clan history with any detail. Clans 
had political saliency in the early colonial period when the residential clans of Zanaki and Kuria 
got their own chiefs. Yet with the reorganizations of the 1950s, clan territorial chiefs became local 
headmen and all the people elected a Paramount Tribal Chief. In both cases, however, the 
administration understood clans as subsections of "tribes," referred to by Baker as "sub-tribes," 
and not as dispersed networks of affiliation. Thus people ceased to tell publicly, and gradually 
forgot, clan histories that made regional connections." 
The Core Spatial Images of Clan Traditions 

The more inclusive core spatial image of the homestead is overshadowed in clan traditions 
integrated into ethnic narratives by the image of a bounded community, defined by the genealogical 
relationships of children to their parents. Although genealogical relatedness forms a part of the 
homestead image, in these clan narratives it becomes the central image. These narratives 
characterize clans as subsets of a particular ethnic group whose relationships are contained within 
that unit. This image is consistent with the classic anthropological understanding of clans as a 

" Interview with Silas King" are Magori, Kemegesi, 21 September 1995 (Ngoreme cf). 

12 "Avoidances" or "totems" are animals or objects respected or "avoided" by specific 
groups like clans. 

13 Baker, "Tribal History and Legends," MDB; Baker, Tanganyika Papers . 

form of social organization based on lineage or genealogical descent. In a segmentary lineage 
system a number of clans form an ethnic group. These clans are made up, in turn, of a number of 
maximal lineages, that are made up of a number of minimal lineages, approximating a neat set of 
nesting boxes. This model gives the impression that ethnic groups, made up of discrete clans, 
formed isolated and autonomous units. 14 

Beginning more than two decades ago, Africanist scholars mounted an incisive critique of 
segmentary lineage theory. Some have concluded that the concept is a fantasy of the west and that 
blood ties never formed the basis for social organization among aboriginal peoples." Kuper went 
as far as to assert that, "the lineage model, its predecessors and its analogs, have no value for 
anthropological analysis." His reasons for this radical statement were that the lineage model does 
not represent folk models of society and that there are no societies that "a repetitive series of 
descent groups ever organized.'"" 

Western Serengeti ethnography would confirm that the segmentary lineage model in the 
strictest sense does not apply here. If the model functions one would expect to find that lineages 
are unique subsets of particular clans. In Nata, for example, the ekehita or lineage of Abene 

14 See Evans-Pritchard, The Nuer for the classical case and R. Cohen, "Ethnicity: Problem 
and Focus in Anthropology," Annual Review of Anthropolop v. 7 (1978): 379-403, for its more 
recent manifestation. See also Aidan Southall, Alur Society: a study in processes and typ es of 
domination (Cambridge: Heffer, 1956); and Aidan Southall, "The Segmentary State: From the 
Imaginary to the Material Means of Production," in Early State Economics , eds. Henri J. M. 
Claessen and Pieter van de Velde (New Brunswick: Transaction Publishers, 1991), pp. 75-96. 

" See A. Kuper, "Lineage Theory: Critical Retrospect," Annual Review for Anthrop ology 
1 1 (1982): 71-95; D. W. Hammond-Tooke, "In Search of the Lineage: The CapeNguni Case " 
Man, (n.s.) 19, I (March 1984): 77-93; Jane I. Guyer, "Household and Community in African 
Studies," African Studies Review 24, 2/3 (June/September 1981): 87- 137. See also Parker 
Shipton, "Strips and Patches: A Demographic Dimension in Some African Land-Holding and 
Political System," Man (n.s.) 1 9, 4 (December 1 984): 6 1 3-634. 

16 Kuper, "Lineage Theory," p. 92. 

O'Mugabho is found in each of the four hamate or clans. We have already seen that clans are not 
unique subunits of ethnic groups. In segmentary lineage terms this is not a neat set of nesting 
boxes in which each set is a subset of the next higher level of lineage organization. Elders 
sometimes explained these inconsistencies by citing the stories of individuals who moved between 
clans in times of stress. Some life stories that I collected told of a grandfather who was left as an 
orphan and went to his mother's kin to seek help. The clan then adopted him, but he kept the 
lineage name of his father. 

However, one cannot explain this lack of fit within a segmentary lineage system by the 
exceptional circumstances of individuals crossing clan lines. On the contrary, lineage terminology 
itself is flexible at every level and a single lineage designation does not necessarily refer to units of 
the same value. After a great Nata man, like Magoto, who had lots of children dies, his 
grandchildren may start calling themselves "of the Magoto family" or Abhene O'Magoto. A new 
lineage (ekehita) spontaneously forms around an important man instead of continuing to use the 
previous lineage name. At the same time, other lineages of the same level go on using the older 
names. If no great men appear in recent generations the descendants might continue using one 
name for longer than five generations. The ekehita can also divide into smaller groupings to retain 
the original name. The ekehita of Gitaraga refers to the great rainmaker Gitaraga and that name is 
necessary for those who continue to propitiate his spirit for rain at his grave-site. Yet the lineage 
has gotten so big that they now subdivide it into three sections." In time little structural similarity 
exists between lineages of one level of segmentation. The names of clans that suffer extreme 

" Interview with Keneti Mahembora, Gitaraga, 9 February 1996 (Nata cf). 


misfortune are often changed while others spontaneously adopt auspicious names. 18 One clan name 
may refer to different levels of segmentation in different ethnic groups." 

This evidence confirms what Moore succinctly concluded about lineage theory: "descent 
ideology in its most basic form is an ideology of identities ... it is an idea that can be used in 
many different ways and it is enormously adaptable and manipulatable." 20 The same ideology can 
lead to different forms of social organization according to the historical context. As Guyer 
suggests, the social identity translated as "clans" in one place is not necessarily "comparable in 
size, internal structure, legal status, kind of corporate property, or principles of recruitment" to 
what are known as "clans" elsewhere. 21 Although most recent studies have adopted a more flexible 
notion of lineage, just how scholars should describe local social organization remains uncertain. 
Kuper declares the segmentary lineage model bankrupt, but offers no alternative. Guyer worries 
that an emphasis on the negotiability of lineage as ideology will make each study too particularistic 
to allow for comparison. 22 

The core spatial image of the homestead embodied in clan accounts is much broader and 
more flexible than its genealogical aspect. The dominant image of the blood relationships of a 
bounded community in the clan stories noted above seems to have resulted from the reformulation 

18 Siso, "The Oral Traditions of North Mara." 

" Prazak, "Cultural Expressions," p. 49, describes the discrepancies in the Kuria 
segmentary lineage system in this way, "the lack of distinctiveness of descent groups results from 
the circumstantial character of genealogical rationalization and a tendency for descent groups of 
different order to phase into each other." 

20 Quoted in Guyer, "Household and Community," p. 92, from S. F. Moore, "Descent and 
Legal Position," in Law in Cultu re and Society, ed. L. Nader, (Chicago: Aldine, 1969), p. 380. 

21 Guyer, "Household and Community," p. 92. 

22 Ibid, p. 93; Kuper, "Lineage Theory," p. 93. 

of clan stories into ethnic accounts at the end of the nineteenth century. What changed during this 
time were not the core spatial images themselves but the emphasis and interpretation of particular 
aspects of the images. The clan stories transcribed above are part of the ethnic emergence story 
and thus reflect the core spatial images of the homestead, where the ideology of descent plays a 
central part. 

Using the model of the homestead, people imagined larger units of social organization. 
Genealogical relatedness ideally forms the basis of the relationship of those people who live 
together in a homestead. Yet every homestead incorporates strangers and dependents using kinship 
idioms or by finding other obscure lines of descent. The homestead is an indigenous model of 
social organization that creates a sense of a discrete bounded community with corporate property 
but still allows for the flexibility needed to incorporate outsiders. Kuper has shown that people in 
Southern Africa also used this model as the basis for organizing the Zulu state in the late 
eighteenth century. 23 

The ideology of lineages and clans based on the model of the homestead seems to be a very 
old concept in the region. The oldest words for lineage and clan refer to the homestead itself. 
Schoenbrun's study demonstrates that, early in the first millennium A.D., when East Nyanza 
speakers were moving into the drier and more open environments of the east lake, they brought 
with them a concept of residential groupings of people based on dispersed and exogamous lineages, 
expressed by the term, eka. This word was an older Great Lakes Bantu word, derived from a 
Sudanic loan word, for a head of cattle or a cattle camp. Schoenbrun postulates that settlers 
adopted this form of social organization because "accumulations of wealth in animals were stable 

23 Adam Kuper, "The 'House' and Zulu Political Structure in the Nineteenth Century,' 
Journal of African History . 34 (1993): 469-487. 

enough to warrant their conservation by developing unilineal principles for inheriting them." 24 The 
male space of the cattle corral signified broader connections to kin with claims on that wealth. 
Earlier I argued that in the western Serengeti this East Nyanza Bantu word for lineage, eka, was 
used to create the homestead built around a central cattle corral as a formative strategy for 
expansion on the frontier. A comparison of the terms and the ethnography for clans across the 
region, shows the elaboration of the homestead concept. While the segmentary lineage model 
presumes that social organization based on kinship is diametrically opposed to that based on 
territory, the homestead model incorporates them both. 25 
Hamate: Clan as Locality 

An important aspect of the homestead model is its function as a residential, localized form 
of social organization. Although clan organization and the words used to describe it vary across 
the region they hold in common the ideological view that a clan is a group of people who live 
together on clan controlled land. Today the clan unit is the largest grouping of people who speak 
about their relationships to each other in lineage and descent ideology below the level of the ethnic 
group. In the ethnic groups farthest west, Ikizu and Zanaki, each of the clan groups lives in one 
area and claims clan lands. In the ethnic groups farthest to the east (Ikoma, Ishenyi, Nata, 
Ngoreme) clans have been dispersed among the three different age-set territories, with people of the 
same clan tending to settle as neighbors. In Ngoreme, traditions tell explicitly of people moving 
out of clan lands during the disasters and gradually back again during the colonial period. Even in 

24 Schoenbrun, A Green Place , pp. 170-1. 

This is a classic anthropological assumption, Sir Henry Sumner Maine, Ancient Law 
(London: J. Murray. 1861); Lewis Henry Morgan, Ancient Society (New York: World Publishing. 
1877); these ideas were picked up by later anthropologists as the foundation of much ethnography 
in Africa, A. R. Radcliffe-Brown, The Study of Kinship Systems: Structure and Function in 
Primitive Society (New York: Free Press, 1 965), pp. 49-5 1 ; Meyer Fortes, Kinship and the Social 
Order (London: Routledge and Kegan Paul, 1 969), pp. 27ff. 


"Ujamaa" villages organized after the Arusha Declaration in 1977, clan members tended to build 
their houses together. Clans seem to have been the largest scale level of residential organization 
before the disasters of the late nineteenth century. 

One word for clan used throughout the region is hamate, a locative word that specifically 
refers to the politically independent territorial clan grouping. Oral testimony identifies nine Zanaki 
and fifteen Kuria clan territories that anthropologist Malcolm Ruel describes as "provinces." 26 The 
prefix for place designation, bu, is added to the clan name to indicate clan lands, as in Busegwe 
(Zanaki) or Bukiroba (Kuria). In Kuria, as elsewhere in the region, all land belonged to the clan. 27 
The British vacillated between appointing a separate chief in each hamate and appointing a 
paramount chief for the whole ethnic group. Colonial anthropologist, Hans Cory, struggled to 
figure out which was more consistent with traditional practice. Little evidence exists that in 
precolonial times these autonomous clans ever united with other clans for joint action. 28 

The term hamate designates a place or territory, with the particle "ha" referring to an 
exact, particular place, as in ahaase or land. Each of the hamate in its praise names gives a place 
of reference. The word hamate is also used generically throughout the region as "place." For 
example in Nata one can use the term to say, "a beautiful place." Because age-set cycles became 
territorially based in the east at the end of the nineteenth century, people in those areas often refer 

-' Ruel, "Kuria Generation Classes," pp. 14-36; on provinces, Bischofberger, The 
Generation Classes, pp. 15-16. In Zanaki the clan "provinces" are called ekyaro while the "sub- 
clans" are called hamati which are not territorial and move between provinces. Clans are not 

27 E. D. Dobson, "Comparative Land Tenure of Ten Tanganyika Tribes," Tanganyika 
Notes and Records . 38 (March 1955): 31-39. 

28 Ruel, "Kuria Generation Classes," pp. 14-36. Hans Cory, "Report on the pre-European 
Tribal Organization in Musoma (South Mara District and ... Proposals for Adaptation of the Clan 
System to Modern Circumstances," 1945, pp. 1-14, CORY #173, EAF, UDSM. 

to both clans and age-set cycles as hamate. Thus at a basic level hamate means people who live 
together on the same land. The Kuria dictionary defines hamate as "outsiders, people at large." 29 
Although one may find it incongruous to use this word to refer to outsiders, the same territorial 
boundaries that define insiders also designate outsiders. 

People within the residential hamate are responsible to care for the land, celebrate their 
own circumcision ceremonies, and take public action. The hamate corresponds to the boundaries 
of what western Serengeti people call the ekyaro or the ritually maintained lands of a group of 
people (discussed in Chapter 8). In the Ikoma narrative reproduced above Wilson explains the 
term I used for "clan" by referring to the ekyaro or land. Tobisson describes the essence of 
belonging to a Kuria ikiaro as the "close link between agnatic descent (real or fictive) from an 
eponymous clan founder and affiliation to a particular territory associated with him as the 'first- 
clearer of the land.'" 30 There are certain corporate rights and obligations expected from those 
living in the residential hamate. They contribute to compensation in blood suits if another clan 
accuses a member of murder. At a Nata funeral, the hamate of the deceased receive a cow out of 
the inheritance (ang'ombe umwando). 3 ' 
Clan as Metaphorical Homestead 

Those people who live together on clan lands attest to their relatedness based on a common 
ancestor. Yet most of the words for clan or lineage refer not to genealogical relatedness but back 
to the core spatial images of the homestead, analyzed in Chapter 4 as fundamental to the 

29 Muniko et al, Kuria-Enelis h Dictionary , p.2 1 . The Gusii also use the term 
eamate/chiamate to mean clan. Philip Mayer, The Lineage Principle in Cmsii Snrigfy (London- 
Oxford University Press, 1949), p. 9. 

30 Tobisson, Family Dynamics , p. 97. 

31 Interview with Nyamaganda Magoto, Bugerera, 4 October 1995 (Nata d). 


organization of production and reproduction. This suggests that people have made a metaphorical 
connection between the relationship among those who live together in a homestead and that among 
those who share clan membership. People have found it useful to expand the homestead model 
beyond one family unit to describe relationships within a larger territorial community. 

Except for hamate all of the words used for "clan" in the western Serengeti refer to the 
spatial organization of the homestead. In Ikizu the clan is called irigiha (hearthstones), in Ikoma 
ekehila (gateway), in Ishenyi and Ngoreme ekeshoko (gatepost). Among related Mara-speaking 
peoples, the Kuria and Gusii, the clan is also called the egesaku (doorway) and the lineage the 
irigiha (hearthstones). 32 The hearthstones, located in the female space of the house, represent 
Ikizu matrilineal inheritance principles. In Ikizu the clans (hearthstones) are grouped into loose 
associations called zenyumba or "houses." The patrilineal Ikoma and Ishenyi use terms for clan 
that refer to the male domain in the homestead, the gate of the livestock corral. 

A common characteristic of words for clan and lineage among all Lakes Bantu speakers is 
that these words are derived from the spatial imagery of the homestead. This seems to indicate that 
this idiom for expressing relationships is old and probably brought by Bantu-speakers when they 
first entered the region (c. 500 A.D.). 33 These images suggest that local people conceptualize the 
relationship among those within one settlement as an extended homestead rather than as a 
genealogical line. 34 A person says that they come from the same "house" as all those whom they 


32 See Tobisson, Family Dynamics , pp. 98-100; and Mayer, The Lineage Principle , p. 5. 
3 Schoenbrun, Etymologies # 1 07. #121, #135, #1 17; Schoenbrun, A Green Place , pp. 

4 Adam Kuper does this in his analysis of the Zulu state as representing deep continuities 
with the house model rather than in terms of lineage. "The House and Zulu Political Structure." 
See also David L. Schoenbrun,"Gendered Histories Between the Great Lakes: Varieties and 
Limits." International Journal of African Historical Studies 29, 3 (1996): 470-480. Christopher 
Ehret disagrees and points to the etymology of -longo for a line of object in The Classical Age . 


collectively associate with their mother or sister, regardless of the exact relationship. Likewise 
they come out of the same "gateway" as those whom they collectively associate with their father or 
brothers. The relationships among many of these people is by "blood," but because the terms are 
flexible and metaphorical, people calculate these relationships in many different ways, allowing for 
the use of a kinship idiom to unite diverse people, to merge generations, or to adopt people from 
other kin groups. 

People also use the same homestead imagery to describe lineage (as opposed to clan) 
relations but in this case they designate a particular known ancestor of four to five generations 
removed as the founder. For the lineage divisions below the level of clan people usually use the 
prefix abhene in Nata or abhahiri in Ikizu (meaning the people of a particular ancestor), followed 
by the name of a known ancestor four or five generations removed. Nata call this lineage group the 
ekehita (gateway) and the Ikizu call it the gate post (egeshoko). 

Although people use the same homestead imagery to represent both lineage and clan, oral 
traditions about the two forms are quite different and presumably represent different kinds of social 
identity. Western Serengeti elders can usually recite their exact genealogies to four generations on 
both their mothers' and fathers' sides— naming up to their grandfathers' grandfathers." This 
knowledge is immediately accessible since most of them grew up knowing their paternal 
grandfathers (or maternal grandmothers) and therefore would have heard the names of their fathers 
and grandfathers (or their mothers and grandmothers) within the lineage as a matter of course. 36 

35 lona Mayer, "From Kinship to Common Descent: Four Generation Genealogies among 
the Gusii," Africa 35, 4 (October 1965): 306, demonstrates that four generations is "the usual 
maximum limit for exact genealogical tracing in many African kinship systems" because this is the 
"natural limit of historical record in a preliterate culture." 

36 In the naming system today a person uses his personal name, followed by his father's 
name, followed by this grandfather's name. 

Clan members, by contrast, cannot tell their exact genealogical relationship to each other or to their 
founder, nor is this knowledge considered important. 

Clan members relate to each other as equals within a common category while lineage 
members must determine the exact nature of their relationship according to gender, generation and 
distance. This is because lineage members have specific ritual and jural obligations to each other 
concerning, for example, land, livestock, inheritance or funeral mourning. Therefore, knowing the 
exact genealogical relationship with other lineage members is necessary to know the degree of 
personal obligation. Evans-Pritchard defined this difference between clan and lineage as one 
between relations of personal obligation versus structural relations between collectivities. 37 

When a man recites his genealogy by saying "I am the son of x, who was the son of y, who 
was the son of z . . . ," after four or possibly five generations he names the sub-clan or clan 
founder and then the ethnic group founder, rather than the known father of his great great- 
grandfather. This "telescoping" of earlier generations is a common feature of lineage reckoning 
throughout Africa. The number of relationships to other groups that people must explain 
determines the number of generations recited in a genealogy. Based on similar observations in 
Gusii, Iona Mayer argued that lineage and clan or what she called "kin lineage" and "ancestral 
lineage" should not be merged in one model of kinship. Instead, the anthropologist can only 
properly interpret the kin lineage as a kinship group. 38 

This distinction is also useful in the consideration of oral traditions about lineages and 
clans. Lineage traditions are frequently reminiscences about known ancestors. People still 
associate lineage ancestors furthest removed in time (five generations) with the particular places 

" Evans-Pritchard, The Nuer . p. 366. 
38 Mayer, "Four-Generation," p. 383. 

where they lived. They can communicate with these lineage ancestors through rituals performed at 
these sites. Lineage narratives are the subject of the next chapter. On the other hand, people do 
not associate clan ancestors with exact places but with general directions nor do they propitiate the 
spirits of clan ancestors. The name of a clan ancestor often has symbolic meaning, such as 
Mwancha, for the lake or west, rather than the name of a remembered person. Although the 
localized clan could take political action based on the defense of corporate property, dispersed clan 
members had little corporate responsibility to each other. 

The lineage idiom implicit in the recitation of a line of ancestors grows out of the core 
spatial images of the homestead. Ethnic stories create new units by imagining clans as subunits of 
the ethnic group, just as lineage groupings made up the territory of a residential clan. However, 
when elders tell clan stories independently, they do not invoke images of bounded subunits related 
to each other by genealogy. These contrasting images seem to reflect differing social realities 
behind clan in different historical contexts, before and after the period of disasters. In the context 
of emerging ethnic units in the late nineteenth century, western Serengeti people increasingly saw 
clans as units that were like lineages only bigger. The clan stories that remain, not assimilated into 
ethnic stories, but sometimes presented as the "migration" stories of ethnic groups, present an 
understanding of clan that seems to predate the disasters. 

Clan Histories as Pathways of Regional Association 

A few elders told me clan histories unrelated to an ethnic story. The story of the Hemba 
told by Ikizu narrator Samweli Kirimanzera is a particularly full account. He told this story after 
we had been talking about matrilineal inheritance, in his case as a Hemba clan member. He said 


that inheritance changed as people married into other ethnic groups, as for example when people 

fromZanaki married in Ikizu: 39 [See photo Figure 6-1: Narrators of Clan Histories, p. 251.] 

Jan: There are Hemba in Zanaki too, aren't there? 

Samweli: Yes, of course. 

Jan: And are you the same as them? The Hemba of the Zanaki and the Hemba of the 


Samweli: We are all one. They separated and some went there and others came here. 

Jan: Did the Hemba begin in Zanaki and then come here or did they begin here and go 


Samweli: Why do you bring up a new history now? You know that the history of the 

Hemba is different from the history of the Ikizu. 

Jan: So shall we leave it aside for now? 

Samweli: You want to know the history of the Hemba? You mean where they came 


Kihenda: She wants to know where they came from, who they came with, and what — 

Samweli: Is that what you want to know? 

Jan: / want to know it all. 

Samweli: . . . The Hemba came — they were great hunters, fierce people who lived in the 

wilderness. They came from the east, around the area of Kilimanjaro. When they left 

Kilimanjaro they went to Tamoga. Singira and Ragega. 

Jan: Aaah haa 

Samweli: They came and passed Ndamio. Then they came to Ikoma where they left 

behind some Hemba called Gaikwe. Others divided off from Ikoma and went to Nata 


Jan: So the Gaikwe are Hemba? 

Samweli: They are one thing. If you hear Gaikwe in Nata it is we Hemba, we are one. 

When they left there they entered Tarime, while they were hunting These people are 

called Nchage and they are also Hemba, they are our people too, the Timbaru. When 

they — Have you written that? 

Jan: Uh huu. 

Samweli: When they left there they passed here in the land of the Kenye, the Wahiri 

Hemba, they are our people too, in Bukenye. From there they went to Mobasi, the Asi of 

Kiagata. When they left there they multiplied greatly and lived a long time and then 

separated again and went to Buhemba, Saani — the Hemba of Zanaki. While in the 

area of Saani others broke off and went to Butuguri of Zanaki. They are the Hemba of 

Zanaki. Again Zanaki B. Zanaki A is here. They were hunters, going here and there to 

hunt. After that — clan — let's go slowly — 

Jan: Ok, I'm with you. 

Samweli: Of the clans that were left, others divided and came to Ikizu. — When they got 

to Ikizu others dispersed, small groups went to Sizaki, Changuge, the Hemba of 

Changuge, the people ofMbasha Megunga, they were — one. Then other Hemba 

39 Interview with Samweli Kirimanzera and Kihenda Manyorio, Kurasanda 3 August 
1995 (Ikizu o"). 


dispersed and when they got to Bigegu, they went to the area of Mwibaghi, where there 

were lots of animals. After Mwibaghi they came to Rindara, in Majita. When they lived 

in Rindara, one elder named Guta built at the place now called Guta, he died there in 

Guta, this Hemba man. They went to Kirio. near Guta. Then when they were finished, 

other Hemba of Rindara, after some years went to Majita. That is why you hear then 

talking about Wiyemba, Wiyemba — they are in Majita and are the same as we Hemba. 

We came from one place and dispersed. Then other youth of Guta moved to Nyatwara 

and then to Nasa. When the Germans came there was a youth named Kitubaha of Guta 

who was made Mwanangwa (headman) of Nasa. ... All of them are Hemba, coming 

from up there. Those that stayed behind there and then dispersed in Tarime, they 

returned and are called those you hear of. the Ndorobo. They are still hiding out there 

and even the foreigners cannot capture them, they don't pay taxes. Even when Nyerere 

ruled they still didn't pay taxes. But they eat meat from the hunt and honey. 

Jan: So those that we call Asi — are they those? The same? 

Samweli: The same ones. 

Jan: They came from Tarime? 

Samweli: Ehhhee. They returned again, after coming from the east they came here, and 

then they went down again, they went again to the wilderness and stayed there. Don 't 

you see? In a household, one child becomes a farmer, another a trader, another a 

hunter — so the household that turned back was of the hunters, they love meat, they 

don't want to herd or farm. Those who came down were the ones who wanted to farm and 


Kihenda: I have never seen a Ndorobo who was caught by the Game Guards — 

Jan: There where you came from in the east — where is it? Soncho Loliondo or farther? 

Samweli: I told you — It is to the east, on the side of Kilimanjaro, there are a few hills . 

. . much farther east than Soncho. Soncho is our neighbor. 

Jan: So when the Nata tell the story ofNyamunywa and Nyasigonko the first parents who 

gave birth to the Nata — they say that Nyamunywa was Asi. 

Samweli: Yes, he was of these people. This one, Asi, they used the name Asi, but when 

they began to disperse — you see the Hemba were the ones to discover fire. Fire was a 

problem then. They were hunters who ate raw meat until they discovered fire. So in 

some places they are called Bahemba morero, "to light the fire. " They are the ones to 

discover fire. Isamongo oflkizu was a Hemba too. [. . .] They shortened the name as 

language changed. 

Jan: So they came from over there and at each stop left people? 

Samweli: They left groups here and there and went on. 

Jan: Could you still go to those of the Ikoma and claim clanship with them? 

Samweli: The Nata and the Ikizu are one thing. Until now, even our eldership ranks are 

the same. 

Kihenda: Even if you go to my house you will see that the brands for goats and cattle are 

the same as the Nata. 

Samweli: And he is the same as me, he is a Hemba. 

The core spatial image of this story is not the genealogical relatedness of a bounded 

residential unit but the movement of a group of people from one place to another across the 

landscape. Similar images are encountered in the Ikoma emergence story in the last chapter, of 
Mwikoma moving from place to place before settling in Robanda, or the Ngoreme story told above 
of the lineage of Isabayaya moving until they found their home in Ikorongo. One can identify clan 
stories, even those incorporated as ethnic accounts, by these core images of dispersed pathways. 

Where these stories are incorporated into an ethnic account elders present them as 
"migration" stories. Here the sons and grandsons of the ethnic group founder move from place to 
place until they finally settle in their present home. The narrators of these accounts project 
genealogical unity on the diverse origins of present day ethnic groups. These accounts present 
places which have been important in the past in a linear narrative as stops on the migration route. 
However, taken without the ideological assumptions of the ethnic group, one can also see these clan 
"migration" stories as a mental map of a very different kind. 

The listing of all these places and groups of people who live there makes a connection 
between and unifies seemingly unrelated elements. The lineage idiom does not seem to function at 
all. Samweli's list cuts across ethnic lines, economic subsistence patterns and geographical 
distance, reaching from Mt. Kilimanjaro to the shores of Lake Victoria. In this account Asi 
hunters are the clan brothers of Bantu-speaking hill farmers and lake fishermen from all across the 
region. Mapping the points mentioned in the story would produce the image of a crisscrossing 
pathway, doubling back on itself across the region. How were these paths formed by historical 
process and what kind of relationship do they signify? 

I argue that the core spatial images of clan stories, which scholars have often interpreted as 
"migration" stories, represent the regional connections of dispersed clans in the past. The evidence 
of historical linguistics suggests that some of the earliest ways in which Bantu-speakers used 
lineage principles to organize local communities were also used to maintain claims outside their 
local communities. Schoenbrun demonstrates through the reconstruction of words for lineages that 

just as people used lineage as a way to conserve wealth, at the same time the inheritance of wealth 
by lineage members in other settlements created webs of clientship outside their local residences. 40 
Separating these two aspects of lineage or clan is impossible; the internal and the external exist in a 
dialectical tension. The community builds itself by boundary formation and exclusion but cannot 
maintain itself without using those same bonds to establish regional linkages. 
The Material Basis for Clan Networks 

Regional networks were a crucial part of the total economic strategy as settlers moved into 
the drier and more marginal agricultural lands of the interior. The environment of the western 
Serengeti is a productive one in good years but prone to an equal number of years without enough 
rain, too much rain or unseasonable rain. Differences in rainfall patterns are highly localized, 
meaning that one hill settlement might experience drought while their neighbors a day or two walk 
away would harvest in abundance. To survive these fluctuations people needed to develop 
relations of reciprocal obligation with other people in a variety of locations. Herders and 
hunter/gatherers accommodated these climatic fluctuations through transhumance. They could 
pick up and move to the area where resources were more abundant in any one year. Yet hill 
farmers could not be so flexible. Although they moved every 5-10 years, they were anything but 
nomadic and only moved short distances. Farmers were tied to the land. The problem for them in 
times of environmental stress was how to make claims for assistance on people who did not live in 
their immediate locality. 

Using clan, with its implicit reference to the space of the homestead, was an emotionally 
powerful way of asserting these claims. People bound by clan ties felt some of the same sense of 
corporate obligation to those whom they saw infrequently as to those living in the same homestead 

1 Schoenbrun, A Green Place , pp. 170-171. 

or settlement. Many of our Sonjo hosts greeted the Ikoma man on the trip, whom they had never 
met, by saying "we are of one womb." It is clear from the composition of the internal homestead 
that this sense of unity was not necessarily dependent on blood relations. People find it hard to 
avoid these claims of obligation between clan members. The metaphor of the homestead allows for 
many different kinds of relationships (subordination, superordination or equality) based on the 
generational, parental, sibling or marital relationships of the homestead. Kuper also claimed that 
"the patron-client relationships of the house system" provided the framework for much larger scale 
relationships in the Zulu state. 41 

Although theoretically it makes sense that people used clan ties (in which they invested 
social capital in times of plenty) to make claims among distant communities in times of need, how 
were those paths described in Samweli's story created? How were connections made across great 
geographical distances and across ethnic and economic boundaries? Again, it is useful to speculate 
on this process by looking both at patterns that have existed in historical times and at testimonies 
about the past. If people moved a lot in a frontier situation, they needed to develop not only 
structures for organizing new settlements but also the means for maintaining connections to people 
from their home communities. Clan identity served both needs. In his treatment of the "African 
frontier" Kopy toff theorizes that groups who moved out in search of new land maintained contact 
with home communities, eventually establishing communication networks over long distances. 42 
A Tradition of Mobility: 

Hill farmers of the western Serengeti have a long tradition of settlement mobility. They 
might move whole settlements twice or more in a decade. Repeatedly elders would tell me, "we 

41 Kuper, "The House and Zulu Political Structure," p. 486. 

42 Kopytoff, "The Internal African Frontier," pp. 7-22. 

are a moving people, we never stay long in one spot." 43 In their life histories many elders related 
all of the moves they had made in their lifetimes, including stints as migrant laborers. Some moved 
out of their ethnic area as youth but moved back again as elders. Today many Ikoma people live in 
the lshenyi and Nata areas, with no plans to return. Elders stated that in the past an entire 
settlement would move together. 

Oral narratives suggest that, ideally at least, moving was a decision made by the elders of 
the settlement, or in consultation with a prophet. The decision rested on the perception of problems 
in the home area such as infertility and insecurity or the promise of better conditions in another 
place. It was also possible for one or more families to move out without disturbing the critical 
mass of the settlement. One family might follow a relative who found a fertile new area in which 
to settle. People used already established relations of reciprocity to gain a foothold in a new area. 
Elders most often said they moved to find better farming and grazing lands. Some families said 
they moved because of a number of deaths and misfortune in their previous home. 44 

The important thing to note in these testimonies is that people chose their destinations 
based on having a connection, kinship or otherwise, with someone who already lived there. Carole 
Buchanan's study of inter-ethnic clan relations in Uganda suggests that clan affiliation made these 
movements less random and more secure. It "maximized the safe options for people in crisis and 
influenced groups to move less capriciously, particularly if kinsmen were known in comparable 

43 Interview with Bita Makuru, Bugerera, 1 1 February 1995 (Nata <?). For Kuria 
settlement mobility see Prazak, "Cultural Expressions," pp. 65-89, he found that an average 
homestead lasted for 10-20 years, p. 124. 

44 These generalizations are taken from many informal conversations, the narratives of 
traditions about this period, the life stories of elders, including their narration of their parent's life 
histories and especially interviews with Surati Wambura, Morotonga, 13 July 1995 (Ikoma ?); and 
Mariko Romara Kisigiro, Burunga, 3 1 March 1 995 (Nata <f ). 

ecological zones." She maintains that clan identity provided a regional framework within which 
individuals could be more easily adapt to new localities. 4 ' 

Older people told me that before the advent of trucks, buses or bicycles to facilitate moves 
they carried everything on their heads or backs to the next site of settlement. They described 
moves as festive times when people called all their relatives, friends and neighbors to help carry 
loads to the new site, often a full day's walk or more away. If enough people were there to help, 
they could move the entire homestead-wooden house poles, furniture, utensils and livestock— in a 
couple of days. They often sent the youth of the family ahead by a full season to plant and build 
temporary housing. Donkeys or other pack animals were not used in the western Serengeti, not for 
lack of knowledge but because of tsetse fly infestation and a lack of consistent need for them. 
Pastoral peoples, such as the Tatoga and Maasai, used donkeys in this area. Elders simply stated 
that donkeys were not kept because they fed and cared for the animals without getting any use out 
of them. This clearly suggests a pattern of settlement mobility over a period of years rather than a 
transhumant lifestyle following the seasonal grazing patterns. It also means that new settlements 
were located fairly close to the old sites, usually within a day's walk. 

A settlement would be located so that residents could move the outlying fields (ahumbo) 
for several seasons without moving the settlement. Once they exhausted all of the potential fields 
in the area, the settlement itself would have to move to a new center of possible fields and grazing 
areas, while still located near enough to harvest the fields left behind for another season. Thus, in a 
leapfrogging fashion, settlements would gradually spread out to their outlying fields and then skip 
over them to the next settlement site. The Nata village where I lived had been a settlement in the 
past, was abandoned, then became the ahumbo fields for a new village called Mbiso, located on the 

45 Carole A. Buchanan, "Perceptions of Ethnic Interaction in the East African Interior: The 
Kitara Complex," The International Journal of African Historical Studies 11,3 (1978): 425-426. 

main road ten kilometers away. When the government recently stopped enforcing its policy of 
"villagization" people from Mbiso started moving out to take up permanent residence in their 
ahumbo fields. Scholars have described this sort of "migration" as "natural drift," which shows a 
patterned mobility rather than random moving. 46 

Mobility was also possible here because of the relative ease of clearing the land for new 
fields and obtaining materials for building. Because of vegetation patterns in the area, people 
cleared new fields by burning brush and chopping down larger trees in the months before planting. 
This meant that farmers put little long-term labor investment into any one field. Land was not 
"owned" in the past in the sense of exclusive and inheritable rights of land use. Settlers used fields 
for a number of seasons before moving onto other fields within the area ritually maintained by the 
clan. In contrast to this is the system of land "ownership" by Kikuyu lineage groups in Kenya who 
laboriously cleared farmland over a period of years out of the highland forest. 47 

If one extrapolates patterns of mobility in the western Serengeti over the long term, it is 
easy to recognize how networks of affiliation developed. Over time the connections between new, 
old and still older settlements would diversify and extend over long distances. People would be 
drawn into the networks of others within the settlement as they became acquainted with their 
neighbor's visiting relatives or friends. As time went on the pathways between communities would 
become denser and more diverse, no longer linked to specific genealogical lines. The particular 

46 Vansina, Paths in the Rainforest , p. 55, he calculates that spread of 22 km in 10 years is 
quite possible with this type of "drift". Christopher Ehret, Southern Nilotic History , describes 
migration as a slow and gradual process of small groups moving onto the next pasture or field, a 
process of assimilation rather than extermination, pp. 26-27. See also D. P. Collett, "Models in the 
Spread of the Early Iron Age," in The Archaeological and Linguistic Reconstruction of African 
History, eds. Christopher Ehret and Merrick Posnansky (Berkeley: University of California Press 
1982), pp. 182-195. 

" Kershaw, Mau Mau From Below , pp. 22-23. 

connections of a few clansmen might become generalized to represent reciprocal connections 
between specific places, as they appear in Samweli's narration. The connections between places 
resemble paths of communication and interaction that had practical value. 

The most common way that historians have explained this East African pattern of regional 
clan networks, like the one Samweli describes, has been by dispersal from a central place of origin. 
The pioneering academic histories of East African societies based on oral sources interpreted clans 
as descent groups that migrated from original homelands to many places in the region, creating new 
ethnic groups by the amalgamation of many clans. 48 These histories were based on the collection 
of rich and detailed clan traditions, largely accessible to students returning to their home 
communities. They wrote histories in which the place-names were stops on a migration route- 
producing dots on a map connected by arrows and correlated with a chronology of generations. 

This kind of historical reconstruction is problematic because it projects the image of clans 
or ethnic groups as we know them today moving as fixed and solid units across the landscape from 
place to place. It does not take into account the possibility of forming new kinds of units, 
permeating the boundaries of the units themselves, or incorporating strangers. "Tribal" migration 
histories draw on models produced in the nineteenth century by Europeans re-envisioning 
themselves as self-conscious nations of people with a common past as wandering "tribes." 49 
Narrators easily adapted the place-name lists of the western Serengeti to this migration model and 
many elders tell their story in this way. However, a closer look at these places, the imagined 

48 For examples from the surrounding regions see B. A. Ogot, A History of the Southern 
Luo (Nairobi: East African Publishing House, 1967); Paul Asaka Abuso, A Traditional History of 
the Abakuria. C.A.D. 1400-1914 (Nairobi: Kenya Literature Bureau, 1 980); Itandala, "A History 
of the Babinza;" William R. Ochieng', A Pre-Colonial History of the Gusii of Western Kenya 
(Nairobi: East African Literature Bureau, 1974); and Henry Okello Ayot, A History of the l.uo- 
Abasuba of Western Ken ya, from A.D. 1760-1940 (Nairobi: Kenya Literature Bureau, 1979). 

49 This point is made by Kopytoff, "The Internal African Frontier," pp. 3-4. 

landscapes that they evoke and the social relationships represented in them bring the migration 
paradigm into question. However, although historians have criticized these early histories for their 
literal acceptance of clan migrations, scholars have suggested no plausible alternative to explain 
clan dispersal, apart from this literal movement of people away from their homes, whether in mass 
and unified migrations or by migratory drift. 

There is little doubt that migration played a central role in creating these dispersed clan 
associations. 50 The pathways described in the narratives were often literally the paths that people 
walked back to known home communities or to those of distant relatives to obtain aid. Although 
clan histories describe a unilineal movement, the paths seem to have been worn in both directions 
over extended periods of time. Clan histories were social maps that would enable the next 
generation to find the paths that led to people on whom they could depend. 
The Historical Context of Regional Identities 

Clan networks provided the means for crossing social boundaries created by distance and 
time to settle in a different area or gain access to its resources and knowledge. These clan 
networks did not result only from a simple process of settlers going to new areas and maintaining 
their old clan names. Depending on the conditions surrounding their arrival they might introduce 
their clan name as a new one in the area, they might assimilate into an existing clan as a new 
lineage or they might abandon their clan affiliation altogether. For example, the Kobera clan 
among the Ishenyi are in the process of "disappearing." Traditions say that the Kobera people 
lived near the Ishenyi long ago at Nyeberekera. When the famine came they did not go to Sukuma 
with the others but the Nata invited them to move to Nagusi with the Ishenyi. There the Bene 

50 Cohen, "Pirn's Doorway," p. 196; and David William Cohen, "Reconstructing a 
Conflict in Bunafu: Seeking Evidence outside the Narrative Tradition," in The African Past 
Sneaks, ed. Joseph Miller (Kent, England: Folkestone, 1 980) and Womunafu's Bunafu . pp. 48-67. 

Omugenyi clan in all of its lineages took them in, or "swallowed them." 51 The Kobera go back to 
the hill of Kobera where they came from to make offerings to their ancestors but also make 
offerings at the lshenyi ancestral sites. They maintain a dual identity but are increasingly 
identifying themselves as lshenyi. n 

We better understand clans as a strategy for asserting useful relationships across social 
boundaries rather than as fixed and enduring social units of equal value brought by immigrants 
and inserted into a preexisting structure. People used the generative principles of clan association 
represented in the core images of the homestead to improvise their relationships, both near and far. 
By asserting membership in a clan associated with the clan who controlled the land, new settlers 
could claim rights to the land. The bargaining power that immigrants had with their hosts, who 
provided them access to land and protection in return for clearing and farming new land, may have 
determined whether immigrants kept their old clan names or accepted membership in existing clans. 
This may account for the fact that clans constitute themselves in different ways in different places, 
depending on the circumstances. In Ikizu the Moriho clan claims pride of place as "first-comers," 
while in Zanaki the Moriho live as a small lineage within another clan territory, and in Nata the 
Moriho are one clan among four founding clans. Units with similar names did not necessarily have 
similar forms or functions." 

The simple image of clans originating in one place and gradually spreading out by a 
process of settlement mobility across the region does not explain why relatively few clan 

51 Eating is a powerful metaphor for instrumental power throughout the Lakes Region, 
especially of kingship and chiefship. It is significant that it is used here to mark the disappearance 
of a clan group name. David Schoenbrun, personal communication. 

52 Interviews with Mashauri Ng'ana, Issenye, 2 November 1 995 (lshenyi <f); Jackson 
Mang'oha, Mbiso, 13 May 1995 (Nata <?). 

53 Moore, Space Text and Gender , p. 1 7 . 

associations are prominent region-wide out of the unlimited number of possible clan names if each 
father or mother potentially generates a new lineage. Some critical ideological or material criteria 
must determine the privileged position of some clans over others. Neither does the mobility of 
individuals of particular descent groups account for the ideological association of some clans with 
other clans of different names. It seems that the pathways depicted in clan narratives also represent 
a "deeper and more abstract or metaphorical meaning at a level where clans are embodied ideas." 
Waller suggests that clan names might indicate both historical settlements and "social pathways or 
claims between communities." 54 

Although scholars have long acknowledged inter-ethnic clan relationships there are few 
who attempt to explain these dynamics outside a process of migration. Two notable exceptions are 
Gunter Schlee in Identities on the Move and an article by Carole Buchanan on ethnic interaction in 
the "Kitara Complex" of Uganda." However, although their rich data provide evidence for forms 
of association beyond that of migrating kin maintaining old clan alliances, their analysis does not 
go far enough. Buchanan suggests that the patterns observed in clan designations seem to suggest 
that the system of clans constituted a larger "contextual framework" within which people operated. 
"Clan membership was intrinsic to belonging to the social order and an essential feature of 
ethnicity itself." 56 

54 Richard Waller personal communication. 

55 Gunther Schlee, Identities on the Move: Clanship and Pastoralism in Northern Kenya 
(Manchester: Manchester University Press, 1989); Buchanan, "Perceptions of Ethnic Interaction.' 
See also M. d'Hertefelt, Les Clans du Rwanda Ancien. Elements d'Ethnosociologie et 
d'Ethnohistoire (Tervuren: Musee Royal de I'Afrique Central, 1 97 1 ); and Luc de Heusch, Le 
Rwanda et la Civilisation lnterlacustre (Brussels: Universite Libre de Bruxelles, Institut de 
Sociologie, 1966). 

56 Buchanan, "Perceptions of Ethnic Interaction," pp. 417, 425, 427. 


Newbury's study of clanship in relation to kingship on Ijwi Island of Lake Kivu 
demonstrates that clans were not archaic forms of social organization, eventually superseded by the 
formation of centralized states. Rather, clanship and kingship are "seen as different aspects of a 
single larger process, that is, as complementary as well as opposed concepts." One cannot 
understand clans without seeing them within the broader context in which they functioned. 
Newbury's work shows how changes in clan identity not only result from the incorporation of 
strangers and the movement of clan members to other areas but also reflect changes in the structure 
of clan categories themselves." 

These changes in the structures of clan categories in relation to the broader context renders 
the oral traditions surrounding clans in the western Serengeti confusing for the historian. Clans 
have operated differently in at least two identifiable historical contexts, mediated by the disasters of 
the late nineteenth century. The two kinds of clan narratives investigated here characterize each of 
these periods of clan identity. During and after the disasters the ethnic groups as we know them 
today began to take shape, with clans increasingly reduced to isolated subunits. Samweli's Hemba 
history places clans in a very different historical context. What then were the wider social 
identities that formed the context in which clans operated before the disasters? 
Wider Social Identities Before the Disasters 

Within a regional set of large-scale social identities, clan was one of the few identities that 
crossed boundaries and united people otherwise divided. People used clans as a mediating device 
across various kinds of social boundaries to gain access to resources and knowledge controlled by 
others. Before the disasters a number of large-scale identities operated, based on different 
principles of organization and function. Some, like the boundaries between the Mara peoples and 

" Newbury, Kings and Clans , pp. 4, 227-23 1 . See also, D. Newbury, "The Clans of 
Rwanda: An Historical Hypothesis," Africa 50. 4 (1980): 389-403. 

the Sukuma were marked by bodily markings, while others, like the boundaries between North and 
South Mara peoples were marked by geographical barriers, and still others, like the boundaries 
between black-smiths and non-blacksmiths were marked by expertise and technical skill. Each of 
these kinds of boundaries created a regional system of exchange through difference. Although they 
were marked in different ways each kind of boundary controlled access to certain kinds of 
resources and knowledge. 

For example, we have already seen how the identities of farmer, herder and hunter created 
a regional system of economic interdependence. People identified others by the ecological niches 
that they occupied in relation to their subsistence strategies. These categories corresponded to the 
basic divisions in language groups, East Nyanza Bantu-speakers (farmers), Southern Nilotic 
Dadog-speakers (Tatoga herders) and perhaps Southern Cushitic or Southern Nilotic-speakers (Asi 
hunters). Even through farmers also hunted and herded, in their regional interactions they defined 
themselves as farmers to structure their relations with others. Yet they also had to find ways of 
crossing those boundaries to gain ritual control over the land from the Asi hunters or to gain 
prophetic expertise from the Tatoga. They did this by claiming common clanship with the Asi (the 
Gaikwe or Hemba clan in Samweli's story) or kinship as sons of a ritual "father" with the Tatoga. 

Western Serengeti people deployed the kinship idiom of clan to make claims to land, ritual 
knowledge and trade items across all kinds of social boundaries. We cannot understand clans 
isolated from the historical context in which they functioned as mediating devices. Yet the 
mediation of boundaries by clans did not destroy those other social identities of regional difference. 
People could claim common clanship with each other while still recognizing their difference. 
Samweli explained how different people could still be related using the familial idiom: "In a 
household one child becomes a farmer, another a trader, and another a hunter." 


Clans crossed the boundaries of other kinds of social identities besides those of economic 
subsistence patterns. Regionally people distinguished themselves from each other by bodily 
markings— circumcised and uncircumcised peoples. Uncircumcised peoples bordered East Nyanza 
Bantu-speaking farmers to the north, south and west. Western Serengeti peoples today define the 
land south of the Mbalageti River as the land of the Sukuma, in spite of their settlement beyond 
this boundary. They call the Sukuma Kereti (perhaps those who carry "crates"), 58 who in turn call 
the Mara peoples, Shashi. We know that the term Shashi was in use during the period right before 
the disasters because the earliest travelers who had Sukuma guides named the whole region on their 
maps Ushashi. 59 The most salient feature of this identity for the Sukuma is that Shashi people 
practice circumcision. The Western Nilotic-speaking Luo, who live North of the escarpment, and 
the Bantu-speaking peoples around Lake Victoria, of Buganda and Buhaya, do not circumcise. 
Mara peoples know the Luo as Gaya (slaves), who in turn call the Mara peoples Mwa. 

It is difficult to say whether these divisions were part of the historical context in which 
clan identity functioned before the disasters. Western Serengeti people took both the terms Gaya 
and Kereti directly from the period of the disasters in their reference to the slave and the caravan 
trades. However, since elders always described the social divisions between themselves and the 
Kereti or Gaya as related to the practice of circumcision, the boundaries themselves may have 
preceded their present names. In spite of changes in cultural practice encouraged by schools, 
government and churches, circumcision remains a crucial marker of Mara identity and also of the 
transition to adulthood. This became an issue during the colonial period when the administration 
brought teachers from Sukuma and Buganda to staff the primary schools. Circumcised teenage 

58 Interview with Thomas Kubini and Jacob Mugaka, Bunda, 10 March 1995 (Sizaki <?). 

59 R. Kiepert and M. Moisel, "Victoria Nyansa" 10 January 1896, GM 157/1 A3, TNA, is 
one example. 

students would not submit to corporal punishment by their teachers who were categorically still 
children. 60 This also made marriage with uncircumcised peoples more difficult, though not 

The institution of age-sets and initation into age-sets by circumcision seems to have been 
adopted by East Nyanza-speakers from Mara Southern Nilotic-speakers before 1000 A.D. as a 
means for learning from agro-pastoralists already familiar with the dry inland environment. 
Circumcision and common age-sets united hill farmers with Tatoga herders, Maasai herders and 
Asi hunters. Yet at the same time it divided them from other hill farmers like the Sukuma and the 
Luo, with whom they came into contact through trade. Shared clans mediated these boundaries 
created by bodily markings. Creating brothers among people who were outwardly different. 

Today western Serengeti peoples also mark a difference between themselves and the Lakes 
peoples to the west by reference to cardinal directions and geography. Western Serengeti people 
know those who live on the lakeshore as Nyancha (including the ethnic groups of Jita, Ruri, Kwaya 
and Kerewe), using the same word for both "lake" and "west." The emergence traditions told by 
lakes peoples trace migrations to their present homes from across the lake. The lakes people speak 
the Suguti languages of the East Nyanza Bantu family, becoming distinct from the Mara languages 
by about 500 A.D. The congruence of these categories of lake and inland with linguistic and 
historical distinctions suggests that this division has also been one of long duration. 

However, we must remind ourselves that these categories of lakes and inland were 
situational and relational. The Lakes peoples knew the western Serengeti peoples as Rogoro, or 
"the people of the east." Yet who fit that category depended on the context. A Sizaki might refer 
to his Ikizu neighbor to the west as Rogoro, but the Ikizu man's Nata neighbor to the east might 

1 Interview with Nyamaganda Magoto, Musoma, 8 March 1996 (Nata <?). 

refer to him as Nyancha. In a report of the White Fathers from the Nyegina station in 1 9 1 9, 
among the Ruri people, the priests cautioned visitors not to get directions from local people by 
asking, "where is Bururi?" "For the inhabitants of the lake will indicate the mountains of the 
interior, while if you are coming from the east the inhabitants of the interior will indicate the 
Iakeshore." 61 

In addition to the east/west division, the Mara River divides the region north and south. 
The Mara is the largest river in the district. In some places it spreads out into marshy areas many 
miles wide. Only a few places exist where the river can be crossed or forded with any consistency. 
It was only in 1 989 that the government built a bridge at Kirume, after many years of unsuccessful 
construction attempts and the use of an unreliable ferry. The colonial government divided the 
administrative district of North Mara from South Mara because of the difficulty of crossing the 
river. The oral traditions collected by Siso in North Mara describe one Luo group returning back 
across the Mara River (even after their hosts offered them land in the south) because the river 
would cut them off from their kin in what is now Western Kenya. 62 This geographical boundary 
may have marked social difference between western Serengeti peoples and Kuria-speaking people 
to the north of the river. 

However, clans cross both the east/west and the north/south boundaries in the region. The 
same clan names are found among lakes people, inland people and North and South Mara peoples. 
Those boundaries marked geographically represent divisions according to economic specializations. 
The lakes people specialized in fishing and hippo hunting and engaged in trade around the Lake 

61 Societe des Missionaires d'Afrique (Peres-Blancs), "Nyegina," Rapports Annuals No 
13, 1919-1920, p. 353. 

62 Siso, "Oral Traditions of North Mara;" E . C. Baker, "North Mara Paper," July 1935, p. 
4, Tanganyika Papers . 

Victoria shorelands. The North Mara people live on the higher elevations of the escarpment where 
rainfall is consistently high, allowing them to grow bananas. These divisions established a regional 
system of exchange and interdependence. Yet Western Serengeti people needed to establish 
contacts across those boundaries to facilitate trade and to find patrons who would give them food 
in times of localized drought. Common clanship provided the link between otherwise different 

A final social division that seems to be of long duration in the region is that marked by 
technological expertise-blacksmiths/potters (Turif 1 and non-blacksmiths/potters (Bwiro). Today 
blacksmiths (and potters) have their own endogamous lineages but are incorporated into the clan 
structure of many ethnic groups. Turi are kept ritually and socially separate from Bwiro. If 
sexual relations occurred between them (or in the past if a Turi sat on the stool of a Bwiro or if a 
Bwiro picked up a Turi hammer) the offenders would have to perform special rituals to protect the 
entire community from misfortune. In the case of sexual relations the offender would build a small 
grass hut and go inside while the house was set on fire. Then he or she would run out naked before 
it burned to the ground. In Ikizu, Turi cannot accompany Bwiro to their sacred sites for 
propitiation of the ancestral spirits even if they share the same clan. 64 [See Figure 6-2: Blacksmiths 
and their Tools.) 

63 From the root -tuli, to castrate, to hammer, from the Mashariki protolanguage. 
Schoenbrun, Etymologies . #27. 

64 Interviews with Sarya Nyamuhandi and Makanda Magige, Bumangi, 1 November 

1995 (Zanaki cf); Kinanda Sigara, Bugerera, 27 May 1995 (Ikizu 0% Isaya Charo Wambura, 
Buchanchari, 22 September 1995 (Ngoreme <?); Apolinari Maro Makore, Mesaga, 29 September 

1996 (Ngoreme cf); Bhoke Wambura (Ngoreme 9) and Atanasi Kebure Wamburi, Maburi, 7 
October 1995 (Ngoreme <f); Bischofberger, The Generation Classes , p. 51, describes the 
avoidance of Turi by Bwiro in Zanaki; Gray, The Sonio of Tanganyika, on Sonjo Turi 
blacksmiths, p. 78. 


Sarya Nyamuhandi, Bumangi, 10 November 1995, 
Blacksmith, Holding old Trade Hoe 

Blacksmith Implements, Sukuma Museum, Bujora, Mwanza 
Figure 6-2: Blacksmiths and Their Tools 


Although the Bwiro seem to be ostracizing the Turi, elders compared the relationship 
between Bwiro and Turi to that between blood brothers (aring'a or amuma), among whom sexual 
relations and theft are prohibited. This relationship is a result of Bwiro respect for the power of 
those who work with iron. Testimonies put both rainmakers and iron makers in equivalent but 
exclusive categories. Both are people with medicine and secrets. Through a ritual of oath-taking a 
Turi can become Bwiro, and cross the ostensibly rigid boundary. 6 ' 

The word mwiro/bwiro comes from the Great Lakes Bantu root mwiru meaning "farmer" 
in distinction to the balud or original hunter/gatherers. In the Western Lakes era the word mwiro 
took on the additional meanings of client, follower or subject in distinction to pastoralist peoples. 66 
It thus seems that the East Nyanza meaning of non-blacksmith, with the connotation of owners of 
the land, was unique to this area where patron-client relations between farmers and herders did not 

The social division of blacksmiths and non-blacksmiths, too, seems to have a long history 
in the region. The same traditions and practices surround the relationships between Turi and Bwiro 
all over the region, even where no Turi presently live. Turi tell their emergence traditions as clan 
traditions with origins in Geita, in what is now Sukuma. Other traditions describe patterns of trade 
existing in the distant past for iron hoes and salt in Sukuma. Almost no tradition of iron smelting 
in the Mara Region is recoverable from historical sources; blacksmiths got raw iron from Sukuma 

65 Interviews with Riyang'ang'ara Nyang'urara, Sarawe, 20 July 1 995 (Ishenyi <f ); Silas 
King'are Magori, Kemegesi, 21 September 1995 (Ngoreme <?); Makuru Moturi, Maji Moto, 29 
September, 1995 (Ngoreme &); Bhoke Wambura, Maburi, 7 October 1995 (Ngoreme 9); Sarya 
Nyamuhandi and Makanda Magige, Bumangi, 1 November 1 995 (Zanaki tf). 

66 Schoenbrun, A Green Place , p. 157; and Schoenbrun, Etymologies . #196 and #331. 

to the south or Luo areas to the north. 67 A Turi ethnic group, with its own territory called Buturi, 
now exists in North Mara; it adopted Luo speech and custom within the last two generations. 68 

The distinctions between blacksmith and non-blacksmith established exclusive control over 
certain economic resources-blacksmiths guarded access to the secrets of iron-working and iron 
trade while non-blacksmiths guarded access to the secrets of ritual control over the land and 
farming. Yet because Mwiro and Turi lived together as neighbors they needed to cross these 
boundaries frequently to engage in trade and mutual assistance. In the western Serengeti shared 
clans assured that blacksmiths and non-blacksmiths were never exclusive categories. 

Differences-marked by economy, bodily marking, geography and technological skill- 
defined control over certain resources and knowledge. Clans operated in this context to mediate 
these rigidly defined boundaries so that people on both sides might engage in mutually beneficial 
exchange of resources and knowledge. Shared clan membership crossed each of these boundaries. 
Thus in this context clanship appears as a crucial means for reaching beyond locality to find 
security from drought and for reaching across social boundaries to gain access to the expertise and 
knowledge controlled by other groups. As Newbury suggests, clanship existed in dialectical 
tension with wider social identities of difference. 

Yet in these contexts of great geographical and social distance the probability of "blood" 
kinship decreases greatly. How then were these clan networks that mediated many different kinds 
of social boundaries formed apart from kinship ties? How were people able to establish a bond of 
clanship across the space of these various boundaries? Because clanship works through the idiom 

67 Bischofberger, The Generation Classes , p. 51, reports that blacksmiths from Zanaki 
went to Uzinza to get iron heart-shaped hoes. 

68 Siso, "Oral Traditions of North Mara," reports that the people of Buturi in North Mara 
used to smelt iron. 

of kinship, we may presume that sometimes people developed common clanship across social 
boundaries either through intermarriage or adoptive kinship. The last chapter discussed the 
possibility that hill farmers sought adoption into Asi lineages to gain access to the land or that hill 
farmer matrilineages sought husbands among Asi hunters to gain hunting expertise. In both cases 
the Asi became "fathers" to the hill farmers and thus remembered as ancestors of particular clans. 
This could also have been the story of Sonjo and western Serengeti people who met on the hunt. 
They may have sealed hunting alliances and trade agreements through the ntemi mark and through 
adoption into each other's lineages. Blacksmiths coming to a new area may have sought adoption 
into a clan in order to live peacefully. Disparate peoples literally became kin and thus clan 

From this perspective one can derive a new level of meaning from the ethnic emergence 
stories. The similarities between the story that Samweli tells of the Hemba moves and the ethnic 
emergence stories are apparent. Both tell the story of the clan of first man the hunter who was the 
keeper of fire. Yet in the clan version first man leaves his ethnic moorings and crosses cultural, 
economic and geographical boundaries. In the last chapter first man was interpreted as an Asi 
hunter. Nevertheless, first man as a Hemba or Gaikwe clan member could have been an Asi 
hunter/gatherer or from any one of the Bantu-speaking hill farmer localities, from Kilimanjaro to 
the Lake. It was through the link of clan that these communities could share expertise, knowledge 
and rights to the land. 

Kinship, or even Active kinship was not the only way that people in the western Serengeti 
formed clans networks which crossed the boundaries of other regional social identities. Clan 
networks also seem to have been based on metaphorical associations. Those who controlled similar 
resources or knowledge in their home communities associated themselves with others in similar 
positions throughout the region through clan networks. The clan that gained rights to the land or 


the medicines of the prophet controlled those resources within clan networks. Those clans who 

were recognized as first-comers with rights to the land as Asi hunter descendants may have formed 

associations with clans in other areas who claimed similar rights and expressed their position 

symbolically. This symbolic association through clan networks of diverse peoples controlling 

similar resources or knowledge in their home communities is demonstrated in the use of praise 

names and common avoidances. 

Praise Names and Prohibitions 

Clans do not represent themselves as genealogically based groups in their praise names and 

avoidances, but rather as symbolic associations of people who "praise" the same objects. Each 

"clan," is associated with a set of praise names and avoidances (emigiro, from the Bantu root - 

gldo)."' 1 In the past, young people sang or shouted the praise names at public dances. [See Figure 

6-3: Praise Shouts at the New Moon Dances.] For example, Nata has four clans. The Gaikwe 

avoidance is the zebra and the praise names associate the Gaikwe with Asi hunter/gatherers, 

coming from Rakana, Moturi and Buhemba. The Mwancha avoidance is fish and the leopard and 

they name the place of Muganza, to the west and refer to the lake. The Getiga avoidance is the 

kunde bean, and they name places in Soncho. The Moriho avoidance is cattle; if a drop of milk 

spills they touch their finger to it and put it on their forehead as a blessing. They name Bwiregi, a 

place in North Mara known for its love of cattle. One version of the Muriho clan praise names is 

reproduced in full: 

We are the Muriho, people who honor the cattle, the Iregi people ofUhlsacha and 
Tunda: those who store freshly churned butter in the attic, together with the Iregi of 
cattle, coming from Itiyariro. Those who go to the fields are farmers, we took the 

' Schoenbrun, Etymologies. #287 and #288. 


branding iron and branded nine calves and the rams complained that they were not yet 
branded. We are the Iregi who praise cattle and millet. '" 

The symbolic representation of things, places and economies allowed clans to associate 
themselves with other clans using similar praise names throughout the region. I identified the same 
clan names and names associated with them, in conjunction with common avoidances and places of 
reference, in many other ethnic groups throughout the region. For example, 1 found references to 
the Iregi, named here in relation to the Moriho. all over the Mara Region, far beyond the territorial 
boundaries of Nata. Ishenyi tradition says that the Iregi left Ishenyi during the disasters at 
Nyeberekera. The Kuria Iregi clan has a lineage group called the Isenye with a tradition of 
migration from the Range hills near Ikorongo, in western Serengeti. Among the Southern Nilotic 
Kipsigis in western Kenya a clan named Rangi takes its name from these same hills. 71 One could 
extend these chains of association across vast geographic and cultural distances. 

In the Moriho praise shout no indication of lineage-based organization, or descent from one 
ancestor exists. Instead, the clan praises itself with reference to economy (herding), things (cattle, 
butter, millet), peoples (the Iregi) and places (Mbisacha, Tunda, Itiyariro). Around the region, the 
Moriho clan name, and associated names like the Iregi, all have the same identification of 
economy, things, peoples and places. When E. C. Baker collected clan names in the 1930s, he 
made an unambiguous connection between places and people names in the praise shouts and the 
origins of the clans. 72 The eiders with whom I spoke, however, described these as metaphorical 
associations-the Muriho clan extolls cattle and so admires the Iregi of North Mara who have so 

70 Interview with Megasa Mokiri, 4 March 1995, Mayani Magoto, Bugerera, 5 April 1996 
(Nata cf). 

71 Abuso, A Traditional History , pp. 83-86. 

72 Baker, "Notes on Tribes," pp. 36-54; Baker. "Tribal History and Legends," MDB. 


Mabenga Nyahega, Singer, Bugerera, 19 August 1995, with Nata Dancers 

Figure 6-3: Praise Shouts at the New Moon Dances 


many cattle. The word for shouting clan praise names is -ibaaka, meaning "to praise oneself (by 
formal declaration).'" 3 Thus, nothing about these praise names suggests that these are ancestral 
places or people names. 

As an example of the regional distribution of clan associations, their avoidances and place 
references, 1 took the four clans of Nata— Getiga, Gaikwe, Moriho and Mwancha— and noted all 
other clans with which they had association in other ethnic groups around the region in Figure 6-4: 
The Regional Distribution of Four Major Clans. No doubt with more exhaustive research one 
could fill out this table further and extend it more widely within the region and beyond. 

This evidence seems to suggest that at another level clan culture was a metaphorical 
association of peoples through a common set of symbols, rather than a line of descent or a history 
of migration. As Paul Abuso explains, "these people who come to embrace that particular totem 
[avoidance] need not necessarily have the same historical origin.... [they] agree to accept the myth 
of origin, as is indicated by the origin of the totem for themselves." 74 People seem to have formed 
association with other clans, who were not related by kinship but by common symbolic 
Resource Pathways 

The metaphorical association of clans, at one level, seem to represent economic networks 
of regional trade and specialization. The symbols of clan praise names and avoidances often 
represented economic subsistence patterns. The four clan associations shown in the chart represent 
the economies of hunting, farming, herding and fishing. This pattern seems to suggest some kind 
of regional system rather than a random occurrence. 

73 Muniko et al, Kuria-English Dictionary , p. 48. 

74 Abuso, A Traditional History , p. 143. 





-5.1 6 .8 

Z H ^ < 

J5 oj 

N 1C 

2 1 

9 3 

N £ 


^ ■ 


O E 
O J 

« E <? 


.t; ■- T) O '-C 

c^ i-s 

i 1 1 5 a 

fil S I S?S-§ | S3 

1 c J P o -5 

I O Sj !- w M 

! * EI S3 .s 


<= 2"-! 



H E. 

5 <fl 
< 03 






CO ~. 

1 5 § 1 

a °l 

Si c C 


s « 

£ o 

B S •« 

i < — — 

i 2 0= CQ 

fl oa.* c c g 
£ 2 £ * CQ Z 

S 1 


,•■§8 1 
h* if 

O on 

•k S 


5 c c 



do 9 o 

1 3 

s 1 

Iff I 

£• 9 


in t/> 

5 ^ 

H*! 6 

c 9 

< 1 






For example, the association of the Gaikwe with Asi hunter/gatherers may have been a 
way for farmers to secure access to forest products for use in regional trade. Asi hunters came to 
Ikoma homesteads to trade ivory, rhino horn, wildebeest tails and lion manes in exchange for 
livestock. Gaikwe clan elders most often mentioned friendships with Asi hunters. Gaikwe clan 
elders said that the Gaikwe and the Asi had taken an oath of friendship together." If this oath 
refers to the adoption of hill farmers into Asi lineages to gain access to the land, as I proposed in 
the Chapter 5, then some hill farmers and Asi would have shared an ongoing clan identity. Or it 
might refer to the association of peoples harvesting and trading forest products. Clan members 
may have passed on forest products to other Gaikwe (Hemba) related clans in the networks of 
trade. It is possible that Gaikwe clan affiliation, although originally based on the adoption of 
farmers into Asi lineages, later became an important link in the regional economy. 

The association of each clan with a different aspect of the economy (herding, farming, 
fishing, hunting) may have developed in relation to regional specializations. The clan links of 
western Serengeti peoples with Turi blacksmiths may have given them access to trade with Sukuma 
for iron and salt." Membership in the Mwancha clan, associated with the lakes people (Mwancha 
is derived from the word nyancha, for lake or west), may have provided access to the flourishing 

75 Interviews with Bokima Giringayi, Mbiso, 26 October 1995 (Ikoma cf); Mahewa 
Timanyi and Nyambureti Morumbe, Robanda, 27 May 1995 (Ikoma <?); Mahiti Kwiro, 
Mchang'oro, 19 January 1996 (Nata cf). 

76 For Sukuma salt trade see H. S. Senior, "Sukuma Salt Caravans to Lake Eyasi," 
Tanganyika Notes and Records 6 (1938):87-92; Michael Kenny. "Salt Trading in Eastern Lake 
Victoria," Azania 9 (1975): 225-8; Elias Nchoti, "Some Aspects of the Iron Industry of Geita, c. 
1850-1950 A.D." (M.A. Thesis, University of Dar es Salaam, 1975). 


trade on Lake Victoria. 77 Each of these trades was crucial to the total economy of the western 
Serengeti and the means by which men with wide connections achieved prosperity. 

On the other hand, like many aspects of the regional configuration of clans, the pattern of 
four clans and four economies seems more symbolic than experiential. Idealized patterns are 
characteristic of the ways in which elders conceptualize clan associations. For example, in 
reference to the four clans traced in Figure 6-4, each represents a cardinal direction and place of 
origin besides their association with economy. The Mwancha clan is associated with the Lake or 
west, fishing peoples from the shores of Lake Victoria, around Mugango. The Getiga are 
associated with the east, from around Sonjo. The Gaikwe are also associated with the east or the 
south, from the woodland wilderness ecologies rather than from the hills of the farming 
communities. The Muriho are associated with the north, or the cattle keeping areas, in connection 
with Kuria or Gusii. Yet since these directions of origin take on an ideal pattern to represent each 
of the four cardinal directions, one might question if this is only historical coincidence. 78 

Within the ethnic groups as they exist today, clans also take on an idealized numerical 
pattern. Most ethnic groups have four, eight or a multiple number of clans. Four is the ideal 
number and in most rituals and stories, things and events come in multiples of four. In Kuria a 
word exists for the additional cattle that must be added if a three or seven is agreed on as the 
bridewealth. Within ethnic groups clans are usually divided into moieties with two or four clans in 
each half. These perfect structures seem to be the result of conscious manipulation rather than the 

77 Michael Kenny, "Pre-colonial Trade in Eastern Lake Victoria," Azania 14 (1979): 97- 

1 07; Gerald Hartwig, "The Victoria Nyanza as a Trade Route in the Nineteenth Century," Journal 
of African History 11, 4 (1970): 535-52; Margaret Jean Hay, "Local Trade and Ethnicity in 
Western Kenya," African Economic History Review 2. I (1975): 7-12. 

78 Buchanan, "Ethnic Interaction," pp. 418-419, reports the same use of cardinal directions 
for clans in the Kitara complex. 


accident of birth. Ethnic groups could have maintained these ideal structures by forming new clans 
out of existing lineages, adopting stranger clans or eliminating clans that had diminished in number 
or influence. 

These ideal patterns suggest an organizational paradigm for creating regional systems 
rather than the random spread of clans to new areas. It indicates a systematization of elements that 
were already functional in many separate communities, in communication with each other over 
long distances. From this evidence of idealized clan structures and regional patterns that go 
beyond the random movements of descent groups we might conclude that clan histories took shape 
after people had already developed the associations between communities based on various kinds of 
affiliation like kinship, economy or rights to resources. By joining them in a historical account 
narrators gave substance to the reciprocal claims embodied in these informal pathways. Clan 
traditions represent a process of looking back and ordering or reordering these regional 
connections, lineage based or otherwise. 

People would then have elected to associate themselves with the regional associations that 
brought the most benefits. Communities seeking to assert an association with other communities 
may have adopted the names and avoidances deemed most efficacious. Clan names may have 
begun as the names of ancestors but, in time, emigrants took them to unrelated communities while 
at the same time outside communities adopted the names. Historical examples exist of groups who 
changed their clan name or avoidance because of misfortune." They took the names of prosperous 
groups. Some peoples adopted new clan avoidances because of particular experiences. The Sweta 
of North Mara relate a story in which a group of baboons saved them and so they adopted the 

' Baker, "Notes on Tribes," pp. 13-14. 

baboon avoidance. 80 They made this choice with full knowledge of other baboon clans in the 
region with whom they would now be associated. Many Kuria wild animal avoidances are the 
same as those used in the Lakes kingdoms of Busoga, Bunyoro and Buganda-states that came to 
influence what is now western Kenya and the Mara Region in the nineteenth century." Clan 
names and avoidances may have spread ahead of migrating people. It may have been a little like 
joining a club or lodge today where members treat each other like family. 
Clans as Pathways of Knowledge and Resources 

This line of thinking suggests that people formed clan networks by metaphorical 
association in order to cross various kinds of regionally based social boundaries to gain access to 
resources and knowledge controlled by others. Membership in a clan allowed a person to make 
claims on the particular expertise or resources of other members. Although all western Serengeti 
people practiced a similar agro-pastoral-hunting economy, they maintained suites of knowledge 
that made each collectively valuable specialists within a larger regional system. If a person wanted 
to obtain a lion mane or an ivory bracelet, he would go to a Gaikwe clan community where they 
maintained relations with Asi hunters, for knowledge of cattle medicines to the Moriho clan or for 
fish poison to the Mwancha clan. Historical pathways between communities became a means to 
gain access to valuable sources of knowledge and resources. 

80 Siso, "Oral Traditions of North Mara," Sweta. 

81 David William Cohen, ed. Towards a Reconstructed Past: Historical Texts for Busoea 
(Oxford: Oxford University Press, 1 986); Wriglev, Kingship and State : Kagwa. Basekabaka be 
Buganda; Roscoe; The Buganda : Hartwig, The Art of Survival , p. 40, Hartwig shows that many of 
the wild animal avoidances in Ukerewe traced their origins to Bunyoro. For the influence of these 
kingdoms on the other side of the late in the nineteenth century see, David W. Cohen, "Food 
Production and Food Exchange in the Precolonial Lakes Plateau Region," in Imperialism. 
Colonialism and Hunger: East and Central Africa , ed. Robert I. Rotberg (Lexington Mass.: D. C. 
Heath and Co., 1983), pp. 1-18. 


A person of that clan may not have had the desired knowledge himself but would have 
acted as the intermediary to those in his clan who would know. This process is similar to the way 
in which I gained access to knowledge as a researcher. I could not go straight to the person I 
wanted to interview but had to find an intermediary to whom he or she was connected by reciprocal 
obligation. Knowledge bases even today are diverse, specialized and highly localized. A cure for 
hepatitis, the skill to feather arrows correctly or the knowledge of circumcision songs were some 
areas of expertise that I was aware of people seeking out through these clan channels. 

Clans did not confine the knowledge that they controlled to economic resources. People 
also maintained control over ritual knowledge within the paths of clan and lineage. The Gaikwe 
clan controls the major places of spiritual power over the land in Nata, at Gitaraga, Nyichoka and 
Geteku. The Gaikwe have this power as the clan of first-comers who must maintain the ritual 
relationship to the land necessary for the prosperity of the community. Clan members hold this 
knowledge as an asset that guarantees their prestige and authority in relation to others who wish to 
ask for rain or a good crop. 

Wealth-in-knowledge is a concept proposed by Guyer and Belinga seeking to move away 
from the wealth-in-people model, based on accumulation and domination. They suggest, from 
Equatorial ethnography, that knowledge was diverse, multiple and widely distributed among adepts 
on the basis of personal capacity, going beyond what would be necessary for survival. Because 
knowledge, represented in individual people and things, is diverse and multiple, people acquire 
wealth by composition rather than accumulation and concentration. A wealthy man is one who can 
tap the differentiated knowledge of a wide variety of people over a diverse social landscape. If 

"multiplicity and expansive frontiers" were valued then the social dynamic was not one of 
domination and appropriation but of improvisational incorporation. 82 

In this sense Samweli's list is an invocation of the diverse and singular knowledge and 
power represented by each of the places and peoples named in the Hemba clan history. The 
process of naming asserts the individuality and personal qualities of each place. The clan is the 
pathway which connects these peoples and places and brings each into its legitimate social context. 
Although the representation of economic symbols for each clan avoidance may indicate certain 
kinds of knowledge, the nodes on the pathway, represented by the places in Samweli's list, would 
have been far more diverse and specific, as a result of a long-term process of continual change. 
New adepts would surface, new nodes of power become actualized, and accommodated by the 
pathway of clan. 

Gunther Schlee argues that clans are conservative structures in contrast to ethnic groups 
that were much more flexible in the past. He surmised that in Northern Kenya the same clan 
names had been operational for the last 400 years. 83 The duration of many clan names in the 
western Serengeti over a long time period also may be a reasonable assumption. However, if they 
do represent pathways of knowledge, rather than only genealogical relationship, the particular 
places, peoples and things listed in clan histories are likely to change as the nodes of power change. 
Some would stay the same but take on different meaning as new practitioners arose. While the 
name of the clan and even its avoidance might remain the same, the content of its knowledge 
represented by the composition of its place-named nodes would have changed. 

82 Guyer and Belinga, "Wealth in People," pp. 91-120. 

83 Gunther Schlee. "Interethnic Clan Identities Among Cushitic-Speaking Pastoralists,' 
Africa 55. 1 (1985): 20. 

Prophecy as the Specialist Knowledge of Clans 

Confirming this hypothesis of clans as pathways in the flow of compositional knowledge is 
difficult today because these regional clan associations have been considerably neglected. Yet one 
clear instance in which clan pathways remain is that of the specialist knowledge of prophets and 
rainmakers whose power depends upon their regional clan associations. It is the connection of 
ritual specialists to a regional set of networks that not only legitimizes their authority but gives 
them the knowledge necessary to carry out their tasks. 

Prophets, healers, diviners and rainmakers are some of the categories of people possessing 
specialized ritual knowledge in the region. Although one can "buy" specific medicines, the power 
of prophecy is always a result of clan or lineage connections. The spirit of a particular clan or 
lineage ancestor must choose the person wishing to become a ritual specialist and actualize the 
individual gifts and inclination of the initiate by directing him or her through dreams. The 
knowledge base that this person taps into is that of the dispersed and diversified clan. Using this 
kind of power to consolidate hierarchical authority was difficult. Yet these specialists were the 
most influential in a society without chiefs or kings. Colonial officer Baker stated that, "in nearly 
every case where a chief was established in Musoma District before European occupation of the 
country, he obtained his position through the supposed possession of supernatural powers." 84 
Baker found this to be the case in Ikizu, a number of Jita clans, Zanaki and Ukerewe. 

The glimpse we get of rainmakers in the early colonial period demonstrates that their 
authority was regional and dispersed rather than localized and concentrated. Rainmakers could 
demand tribute in kind and in labor and had a powerful voice in the affairs of the people. 
Nevertheless, they were experts whom people called on in times of need rather than pro-active 

84 Baker, "Rain,"( typescript), p. 4, Tanganyika Papers . 


agents of authority. Authority over land and people was locally based, in the ekyaro or landed clan 
territory, while rainmakers operated on a regional scale, between clans and even between ethnic 
groups. Claims of a pan-Zanaki identity among the various residential clans in the precolonial era 
are based on the office of the head rainmaker at Busegwe, who authorities later made the Zanaki 
Paramount Chief. 85 In the emergence story, the Ikizu gave the head rainmaker authority only as 
she agreed to the ultimate authority of local institutions. 

The local community controlled the specialist as much as she or he controlled them. Early 
colonial reports tell of sanctions taken against rainmakers. In Ngoreme, when the rains were 
delayed too long, the highest ranking women would call together other women by beating a drum 
and together march off to see the rainmaker. They would beat and curse the rainmaker until he or 
she promised to send rain. 86 Baker reported that the community would take all the rainmaker's 
cattle until the rain fell. He would be placed on a high rock in the sun until the drought ended or 
his 'death from starvation and exposure proved him to have been an imposter.' 87 Each community 
also had other rainmakers in other areas to turn to, or their own sacred places where they 
propitiated the ancestral spirits for rain. No one person or office could monopolize these powers. 
Obueabho: Mara-stvle Prophecy 

Western Serengeti people, and people throughout the Mara Region, conceptualized 
prophecy and healing as the power of distribution and disbursement, rather than concentration and 
accumulation. The general word for prophecy, healing, divining and rainmaking throughout the 

85 Bischofberger, The Generation Classes , pp. 17-18, writes that the rainmaker had no 
political authority, this was vested in the generation-set. 

86 Interviews with Maro Mugendi (Ngoreme cf) and Maria Maseke (Ngoreme ?), Busawe, 
22 September 1995; Njaga Nyasama, Kemgesi, 14 September 1995 (Ngoreme ?). 

8 ' Baker, "Rain," p. 4. 

region is obugabho. This word is derived from the old Lakes Bantu root, -gabd, or "to divide up, 
distribute," usually in the sense of one who gives big feasts or gives things away generously. 88 
Mara languages use the verb kugaba only in reference to the division of inheritance. Other Lakes 
Bantu-speakers use this root in reference to one of the oldest forms of authority in which "big 
men" divided out land among other resources to their clients in return for protection. This 
developed into ritual roles for protecting the land through the office of the chief or king. 8 ' The 
restriction of this word to prophecy and healing in the Mara Region may indicate the importance of 
the role of prophets and also their value in sharing out power and blessings. 

What abagabho controlled was knowledge, the compositional knowledge of clan/lineage 
pathways and their own idiosyncratic learning from multiple sources. Their power lay in their 
ability to dispense this knowledge in response to the needs of those who asked. Peoplejudged 
prophets by their efficacy rather than by their position. The widespread regional distribution of 
obugabho in its many variations is an indication of its ancient roots. A consistent feature of 
obugabho is that it is always inherited through clan or lineage and is considered the patrimony of 
the clan. These features of prophecy in relation to clan are investigated more thoroughly in the 
caseoflkizu. [See Figure 6-5: Two Ikizu Rainmakers.] 
The Kwava Clan of Ikizu and the Ulemi 

The rainmakers of Ikizu also seem to have been obugabho-styh prophets in spite of their 
purported clan origins in Sukuma. The clan affiliations of first woman, the rainmaker Nyakinywa, 
present another level of meaning in the Ikizu emergence story. Narratives state that Nyakinywa 
came from the Kwaya clan of the ulemi (chiefly) line in Kanadi, Sukuma. Before the investiture of 

88 Schoenbrun, Etymologies . #162. 

89 Schoenbrun, A Green Place , p. 187. 



a new Ikizu mtemi a delegation vists the Kwaya clan in Kanadi, obtains ritual items for the 
ceremony, and also consults with the Kanadi elders. " 

Oral traditions from Kanadi confirm this historical connection through the utemi. These 
narratives relate that the founders of the Kwaya clan came from a place called Usonge and 
wandered many places, meeting many people from Uganda to Ukerewe before settling in Ururi, 
Majita on the lakeshore of what is now the Mara Region. There they ruled as walemi for many 
years and got the name Kwaya because of the spears that they used to prove their strength to the 
local people. Two daughters of the mtemi, Hoka and Magawa ran away with their brothers to 
escape the wrath of their father. The women ended up in Ikizu. Magawa was married to the ruler 
of one clan territory (Hunyari) while Hoka was married to another (Kihumbo). The brothers went 
on to found the utemi of Kanadi with the cooperation of Tatoga herders who already lived there. 
The herders gave fire to the brothers in return for rain, and provided leather for making the bracelet 
of utemi investiture (ndezi), which the Ikizu must get in Kanadi. Later a man from Ikizu, in the 
line of Hoka, named Chamuriho, after the mountain, went to rule Kanadi. 91 

The pathways of Kwaya clan knowledge thus extend from the Lakes kingdoms of Uganda 
and Ukerewe to the dispersed Jita communities of Ururi. The Kwaya clan taps the power of 
Tatoga herders in the ritual of investiture, embodied in the ndezi bracelet. The clan mediates the 
regional boundaries of circumcised/non-circumcised, east/west, herder/hunter/farmer. The Kanadi 
story demonstrates that a one- way. one-time event did not establish the pathways of clan but they 

90 Interview with Ikota Mwisagija, Kihumbo. 5 July 1995 (Ikizu cf), Ikota is from 
Nyakinywa's clan. 

" Mtemi Seni Ngokolo, "Historia ya Utawala wa Nchi ya Kanadi ilivyo andikwa na 
marahemu Mtemi Seni Ngokolo mnamo tarehe 10/6/1928," provided by his son, Mtemi Mgema 
Seni, 20/5/1971 to Buluda Itandala. Thanks to Dr. Itandala for his help on Sukuma traditions 
about the Mara Region. 


were the result of considerable traffic in both directions. Even conceptualizing the Kwaya clan as 
"Sukuma" in the sense that we think of it today is misleading. In this story the two Kwaya women 
who marry Ikizu men come from Ururi, a place that is linguistically and culturally not at all 
"Sukuma," but closely related to Ikizu. 92 A present day ethnic group called the Kwaya lives on the 
lakeshore near Musoma. 93 The brothers of these women only later go on to Kanadi and at least one 
mtemi comes from Ikizu to rule Kanadi. In fact, the sisters settled in Ikizu before the brothers ever 
reached and founded Kanadi. The stories of fire and water suggest further the mutual interaction 
of oral traditions between Kwaya and Ikizu. 

Nyakinywa was the first in the line of rainmakers or walemi chiefs in Ikizu. Ulemi comes 
from the Sukuma word for chiefship and carries the meaning of the "first clearer of the land" or the 
"owner of the land." The Sukuma ntemi comes from outside, from the wilderness and brings peace 
to the land. 94 Just as in Sukuma, the Ikizu inherit the utemi through the maternal line. 
Nevertheless, in Ikizu the mtemi does not have chiefly authority over the land as a whole. Rather, 
he or she is a particularly powerful rainmaker in the tradition of Mara obugabho. Given the 
relationship between the Kwaya clan and the various places in the Mara Region it also seems likely 
that Kanadi itself did not accept a Sukuma-style ulemi until much later. Many Ikizu elders gave 
me a list of the Ikizu rainmakers in the line of Nyakinywa as part of the emergence story. A 

92 The Kwaya, Ruri and Jita speak languages of the Suguti branch of the East Nyanza 

93 See ethnography of the Kwaya ethnic group near Musoma, Huber. Marriage and 
Family . 

94 Per Brandstrom, "Seeds and Soil: The Quest for Life and the Domestication of Fertility 
in Sukuma-Nyamwezi Thought and Reality," in The Creative Communion: African Folk Models of 
Fertility and the Regeneration of Life , ed. Anita Jacobson-Widding (Uppsala: Almqvist and 
Wiksell International, 1990), pp. 167-186; Hans Cory, The Indigenous Political System of the 
Sukuma and Proposals for Political Reform (Nairobi: East African Institute of Social Research, by 
Eagle Press, 1954). 


comparison of various accounts is reproduced on the next page. [See Figure 6-6: The Ikizu Utemi 

The utemi list of the descendants of Nyakinywa includes anywhere from eight to fourteen 
rainmakers. The written Ikizu history represented in the first set of boxes has assigned dates to 
each rule, presumably based on the author's own estimation. For the rainmakers since the 
beginning of this century written accounts exist to corroborate the dates. Among the five versions 
of this list that are reproduced here, no two agree on the names or their order before Gibwege (c. 
1 890)." It is significant that this was the general time frame of the "Hunger of the Feet" and the 
massive reorganization of society outlined in Chapter 3. 

The reasons for the discrepancy in names before Gibwege may be that no unified Ikizu 
utemi existed before the disasters. Rather, many different obugabho-styk rainmakers operated 
throughout what is now Ikizu. The stress of the disasters resulted in the need, or opportunity, for a 
more centralized authority and the creation of Ikizu. Sizaki was also consolidated under Sukuma 
chiefship at this time. 96 The oral traditions of emergence were reconceptualized to account for this. 
If so then it makes sense that former rainmakers, from the Kwaya clan, in various localities would 
be incorporated into the genealogical line of Nyakinywa to legitimize the centralization of 
authority." With the formation of ethnic groups, such as Ikizu, the function of clans would have 
changed from that of diffuse pathways of regional knowledge to a consolidated line of power within 

95 Other traditions collected in the 1930s testify that Muesa was "the first remembered 
chief," (C. 1895) Richard C. Thumwald, Black and White in East Africa: The Fabric of a New 
Civilization, a Study in Social Contact and Adaptation of Life in East Africa (London: George 
Rutledge and Son, Ltd., 1935), pp. 46-47. 

96 Bugomora, Lumuli . 

97 For a recent critique of king lists see Wrigley. Kingship and State , pp. 27-41. 


Informant #1 

Informant #2 

Informant #3 

Informant #4 

Informant #5 

1 . Nyakinywa 

1 . Nyakinywa 

1 . Nyakinywa 

1 . Nyakinywa 

1 . Nyakinywa 

2. Nyakazenzeri 

2. Wakunja 

2. Nyekono 

3. Hoka 

2. Hoka 

2. Wang'ombe 

3. Nyakazenzeri 

4. Kesozora 

3. Nyambube 

3. Kesozora 

3. Kisozura 

4. Hoka 

5. Hoka Nyabusisa 

4. Kirongo 

4. Nyekono 

5. Guya 

6. Wekunza 

5. Kisusura 

5. Kerongo 
(first male) 

4. Wekunza 

6. Kesozora 

7. Nyambobe 

6. Nyakinywa 11 

5. Mayai 

8. Gibwege 

6. Gibwega 

7. Gibwege 

6. Gibwege 

7. Gibwege 

9. Mwesa Gibwege 

7. Mwesa 
(first man) 

8. Mwesa 

7. Mwesa 

8. Mweda 

10. Nyakinywa II 

9. Nyakinywa 

1 1 . Matutu Mawesa 

8. Matutu 

9. Matutu 

8. Matutu 

10. Matutu 

1 2. Makongoro 
Matutu (1926-1958) 

9. Makongoro 

1 0. Makongoro 

9. Makongoro 

1 1 . Makongoro 

13. Matutu Matutu 

1 1 . Matutu 

14. Adamu Matutu 
(1986 -) 

1 0. Adamu 

1 2. Adamu 

1 1 . Adamu 

Informant #1 - P.M. Mturi, "Historia ya Ikizu na Sizaki," unpublished mss., 1995 (Nyamuswa) 
Informant #2 - Ikota Mwisagija and Kiyarata Mzumari, Kihumbo, 5 July 1995 (Kihumbu) 
Informant #3 ~ Maarimo Nyamakena, Sanzate, 10 June 1995 (Kirinero) 
Informant #4 - Zamberi Manyeni, Guti Manyeni Nyabwango, Sanzate, 15 June 1995. 
Informant #5 - E.C. Baker, "Notes on the Waikizu and Wasizaki of Musoma," Tanganyika Notes 
and Records . 23 (June 1947): 66-69. 

Figure 6-6: The Ikizu Utemi List 

Ikizu. Other unlikely sources of obugabho-sty\e prophecy still exist in the Mara region, not 
consolidated into chiefly power. 
Tatoga Prophet Clans 

Among the Tatoga certain clans also carry the knowledge of prophecy. The Ghaoga are 
the rainmaker clan whose first prophet came from Lake Victoria and disappeared into it again 
rather than dying. The Ghaoga go to the lake to propitiate his spirit and their houses all face west. 
One "house" of the Ghaoga clan, the Omoghira, are not circumcised to show their authority as first 
sons. The Relimajega clan is the prophet clan who specialize in protection medicines. Their first 
prophet came from the east and they must return there to the volcanic mountain Gijisem to 
propitiate his spirit. Their houses all face east. These same clans with the same prophetic stories, 
indication of places and expertise can be found among all Tatoga groups in East Africa, the 
Barabaig of Mbulu, the Burerega of Sukuma and the Rotegenga of Mara. They also appear 
among the Isimajek fishermen and hunters whom the Rotegenga otherwise despise and ostracize. 

Although I found no evidence that clanship crosses the boundary between Dadog- and 
Bantu-speakers, the concept of obugabho, if not the name itself, seems to be the same.' 8 Tatoga 
elders said that the most important and powerful people in the community were the prophets. 
When I asked Tatoga elders who were their leaders, they would immediately reply, 'the prophets.' 
Yet the power of these prophets was the same diffuse and distributive kind of power that 
characterized obugabho. Tatoga elders told me that they moved according to their prophet. But 
when questioned further, it seemed that the prophet did not "tell" them to go as much as he 
"foretold" that they would go. Another elder said that the Ghaoga and Relimajega prophets would 
"rule" the Tatoga together. Yet on further questioning it seemed that each clan had more than one 

98 The Ikoma use the Relimajega Tatoga prophet clan and the Ishenyi the Ghaoga Tatoga 
prophet clan. 

prophet at any one time, in fact all Relimajega were potential prophets. Their authority was 
ephemeral in nature based on their efficacy and the particular situation that confronted the people.™ 

Because of these shared concepts of obugabho power through dispersed rather than 
concentrated knowledge, hill farmers could incorporate Tatoga prophecy. Traditions from both the 
Ikoma and Ishenyi relate how they killed their own prophets when they failed to make rain or keep 
away the Maasai. As a result the prophets cursed them, never again could they have another 
prophet of their own. 100 Western Serengeti people did not seem to understand this as turning over 
local authority to outsiders but rather the practical need to cross social boundaries and tap into 
other networks of knowledge that were efficacious. Characteristic of this compositional social 
process, western Serengeti peoples were broadening and diversifying, rather than constricting and 
making exclusive, their sources of knowledge. Western Serengeti people used the idiom of kinship 
and clanship to gain access to Tatoga prophetic knowledge by incorporating them as "fathers." 
They crossed the boundaries that divided herders and farmers in the realm of prophetic power but 
otherwise maintained their difference. 


This chapter has explored the various levels of meaning in oral traditions about clans. It 
has gone from an understanding of clan as a bounded, residential unit based on the core spatial 
image of the homestead to that of the clan as a dispersed regional association based on the core 
spatial image of pathways of knowledge and resources. Both aspects of clan identity must be kept 

99 Interviews with Ginanani Chokora and Gejera Ginanani, Kyandege, 26 July 1995 
(Tatoga o"), of the Relimajega Prophet Clan and Stephen Gojat Gishageta and Girimanda Marisha 
Gishageta, Issenye, 28 March 1996 and 27 July 1995 (Tatoga <f), of the Gaogha Prophet Clan. 
Gilumughera Gwiyeya, Girihoida Masaona, Gorobani Gesura, Issenye, 28 July 1995 (Tatoga &). 

m Machota Sabuni, Issenye, in a letter, 23 March 1997, recorded by Nyawagamba 
Magoto, ties the Ikoma and Ishenyi curses together. 

in creative tension to come to some understanding of the local meaning of clan narratives. The 
long-term generative principle of clans was inherently flexible and adaptable, allowing people to 
form cohesive communities based on the obligations of clan members as well as regional networks 
based on metaphorical associations. These metaphorical associations operated side by side with 
kinship-based associations so that people did not distinguish who were "real" clan brothers and 
who were "fictional" clan brothers. 

The functions of clans, using the same generative principles inherent in the core images of 
clan tradition, changed according to the historical context. In the era before the disasters clans 
functioned as a mediating device for crossing regional boundaries. In the era after the disasters 
clans became subunits of ethnic groups. As I show in the last chapter, during the late nineteenth 
century, when clans functioned less as mediating devices, other social identities filled that role in an 
inter-ethnic regional context. In a recent book on ethnicity and gender, Sandra Greene 
demonstrated for the Anlo of southeast Ghana that the way of classifying clans and their function 
within society changed considerably over time."" In both cases, the seemingly rigid "text" of clan 
identity was "read" differently according to need. 

The picture formed by clan narratives that refer to the pre-disaster era is one of settlements 
grouped within a territory controlled by a clan. The clan that held ritual control over the land 
encouraged settlers to come and incorporated them in various ways into the clan structure. Yet 
those settlers also maintained contact with their former communities (with whom they also had clan 
connections) and with clan members in distant communities on whom they could depend in times of 
hunger. In addition people with similar economic or ritual specializations augmented regional clan 
networks through metaphorical associations. All of these kinds of connections known as 

101 Sandra E. Greene, Gender, Ethnicity, and Social Change on the Upper Slave Coast : A 
History of the Anlo-Ewe (Portsmouth: Heinemann, 1996). 

"clanship" operated together and formed complicated regional pathways used to mediate the 
boundaries of other social identities and to gain access to the resources and knowledge of others. 

In the next section we will turn to look at traditions of a more recent era, not precisely 
dateable by age- or generation- set but within the time period when people established particular 
rights to the land before the disasters of the late nineteenth century. Lineage traditions, in contrast 
to clan traditions, describe particular settlements within these larger clan territories. 






The oral traditions that represent the "middle time period" of indigenous chronology are 
characterized by narratives that establish claims to the land for certain lineages who represent 
themselves as "guardians of the land.'" These traditions differ from the clan narratives discussed 
in the last chapter because they refer to individual lineage settlements rather than to clan territories. 
The traditions from this period are usually little more than a list of place-names occupied by those 
lineages. I interpret these traditions concerning claims to the land in the "generation of settlement" 
alongside other evidence to form tentative hypotheses about settlement patterns in the nineteenth 
century: on what social basis settlements were formed, how new settlements were established, and 
how new settlers were incorporated. 

In this chapter I provide a reinterpretation of place-name lists in oral traditions which are 
commonly understood as stops on a migration route or as mythical places. 1 argue that these 
places represent a long time period collapsed into memories about one generation of ancestors just 
before the period of disasters who are responsible for maintaining a relationship to the land by 
mediating the dangerous forces of the wilderness. Individual lineages who have connections to 
specific places that they no longer occupy, preserve oral traditions about the places and perform 

' See this concept in J. M. Schoffeleers, ed„ Guardians of the Land: Essays on Central 
African Territorial Cults (Gwelo, Zimbabwe: Mambo Press, 1 978); and Gregory Maddox, James 
Giblin and Isaria N. Kimambo, eds., Custodians of the Land: Ecology and Culture in the History 
of Tanzania (London: James Currey, 1996). Also on territorial cults see J. Matthew Schoffeleers, 
River of Bl ood: The Genesis of a Martyr Cult in Southern Malawi, c. A.D. 1600 (Madison: 
University of Wisconsin Press, 1992. 


rituals there as representatives of "first-comers," those with ritual authority over the land. Western 
Serengeti people claim a ritual relationship to the land by peopling it with their ancestors, rather 
than occupying or "owning" the land. 

The places-name lists of oral tradition represent former settlements of people organized by 
the idiom of kinship, yet attracted to these settlements by the patronage of prosperous men. People 
who were connected to each other by their relationship to the land, through ancestral spirits who 
guarded the land, used the homestead model to represent the relationship between members of one 
settlement-as coming out of the same "gateway" or living in the same "house." Mechanisms were 
in place to incorporate strangers as "native born" where inclusiveness was necessary to make use of 
extensive land resources. The settlements positioned themselves within a wider constellation of 
multiple and overlapping social networks that radiated out from these fixed and knowable points on 
the landscape. 

The Spaces of Important Places 

The narrative process has reduced oral traditions about the middle period of time in 
indigenous chronology to their core spatial images, representing social processes as place-names. 
Historians interested in a locally grounded interpretation of the past must therefore investigate 
these named places because they are the idioms through which narratives convey knowledge about 
the past. The task of translating local representations of the past into academic historical 
categories requires the translator to emphasize the cultural meaning of place and space. 
Settlement Site Lists 

After telling the emergence story, elders often immediately proceeded to recite a list of 
place-names, representing the time after first man and first woman and their children. For ethnic 
groups with "migration" stories place-names from this time period formed the later part of the list 
of migration stops, including known and local places. In Nata, where migration histories are not 


told, the place-names refer to abandoned settlement sites. In Ikizu, where many in-migrating 

groups have come together, the list of place-names refers to the places where different immigrants 

settled or originated. This is an example from the Nata: 

We are the people ofGilaraga and Mochuri, Rakana and Moteri, Sang'anga and 
Kyasigela, Torogoro and Site, Magita and Wamboye. 2 

Although similar lists of place-names occupy an essential place in the corpus of oral 
tradition throughout Africa, scholars have not often taken the cultural meaning of these places 
seriously in their own representations of the past. Many anthropologists have assumed that these 
are mythical places and so treat them symbolically. Others recognize that they are known places, 
but argue that naming them serves primarily to validate present claims to land or to convey a sense 
of immediacy and validity to the text. 3 Yet elders in the western Serengeti do not understand the 
place-name lists in the same way that they understand the places of mythical origin or emergence. 
Each are the product of the history and self-understanding of a particular social group. The 
ecological and gendered spaces of frontier hill farmers, the places marking the dispersed pathways 
of clan knowledge, and the ancestral places of lineages all represent histories about very different 
subjects. However, a literal interpretation of place-name lists as migration routes has also led to 
historical anachronisms, as were discussed in the last chapter. 

The settlement site or place-name lists that represent the middle period of indigenous 
chronology have a different provenance from the clan "migration" stories to which narrators often 
append them. The social basis for the "migration" traditions presented in the last chapter are clans, 
while the social basis for the settlement site lists are lineages. As 1 argued in the last chapter, 

2 Interview with Megasa Mokiri, Motokeri, 4 March 1994 (Nata a"). 

3 See for example the analysis of T.O. Beidelman, Moral Imagination in Kaguru Modes of 
Thought (Bloomington: Indiana University Press, 1986; reprinted., Washington: Smithsonian 
Institution Press, 1993), pp. 67-83,. 

lineages and clans are very different kinds of social identities, although narrators present them both 
within an idiom of kinship ideology. Both are flexible ideologies that can be used in different ways 
in different contexts. Lineage, like clans, do not fall into a neat order of segmentation. The last 
chapter presented evidence for the hypothesis that clan traditions represent long-term generative 
principles in the region. However, elders represent the settlement site lists preserved by lineages as 
dating to three to five generation ago, conveying the histories of known ancestors who are still 
actively involved in the life of the community. 
Walking the Places 

Many of the elders who recited lists of place-names were anxious that 1 visit those sites. I 
gained additional insight into the cultural meaning of place-name lists by going to these places and 
walking over them with elders who knew their histories. While visiting past settlemet places elders 
spontaneously told other related stories and identified the socially significant elements in the 
landscape-each rock outcropping, hill and stream with its own history. Places serve as mnemonic 
devices to remind men of the stories behind them. The land is a "text" of history and walking over 
it with the elders who tell their stories is an act of "reading" the past. As long as people remember 
these places they will also remember the ancestors and their histories. 4 

The ancestors claimed by specific lineages dwell at these places. While elders identify and 
tell stories about many hundreds of places only a few appear in the list of place-names in the 
historical narrative. Those places are most often either emisambwa sites, places where spirits of 

4 In much the same way Ranger describes pilgrimage places in Zimbabwe, Terence 
Ranger, "Taking Hold of the Land: Holy Places and Pilgrimages in Twentieth-Century 
Zimbabwe," Past and Present 1 1 7 (November 1 987): 1 58-1 94. For a similar approach outside of 
Africa see Keith H. Basso, Western Apache Language and Culture: Essays in Linguistic 
Anthropology (Tucson: University of Arizona Press, 1990). 

ancestors who have power over the land reside, or they may also be ebimenyo sites, places where 
important settlement sites of the past were located. 
Mapping the Places 

The exercise of walking the places showed me that place-name lists refer to known and 
identifiable places, rather than to the mythical or forgotten places of the emergence stories of ethnic 
groups or clans. Men who have a thorough knowledge of local geography from hunting, trading or 
raiding trips, can locate these now-uninhabited places in a wilderness without roads or maps. 
Western Serengeti people remember these places because they mark the graves of ancestors, 
ancestors whose spirits still reside there and who are still intimately involved with the workings of 
present day society. 

All these place-names, even those from groups with migration stories from Sonjo (Ikoma, 
Ishenyi, Ngoreme), can be located within the western Serengeti. Ishenyi tradition provides some 
evidence that Regata, the place often referred to as the Sonjo origin place for many western 
Serengeti groups, was found within what is now the Mara Region. The place-names refer 
primarily to important social processes within the western Serengeti rather than to migration 
stories, although people were surely moving. This chapter discusses the "generation of settlement" 
without positing massive migrations from anywhere else. [See Figure 7-1 : Nineteenth Century 
Settlement Sites.] 

Although elders list place-names of the "generation of settlement" as discrete sites 
associated with their own ethnic group, mapping the sites on a geographical grid reveals a pattern 
of settlement that invites a different explanation. All these sites are hill settlements, demonstrating 
similar subsistence patterns to those described in Chapter 5. They are located farther to the west 
than the emergence sites. The sites named by each ethnic group situate themselves somewhat 
territorially. Yet the sites of each ethnic group are also located side by side with, and often closer 


19th Century Settlement Sites 
of the Western Serengeti 

O Nata 

• Ikoma 

O Isenye 

• Ngoreme 

® Ikizu 

Peter Shetler 

Dove Creek Information Services 

IDRISI and Macromedia Freehand 

30 km 

Figure 7-1 : Nineteenth Century Settlement Sites of the Western Serengeti 

to, the sites of other ethnic groups than to other sites of their own group. This patterning could be 
read to suggest that if ethnic groups existed at all they did not control a bounded and exclusive 
territory. Their proximity to each other as neighbors suggests a pattern of interaction, rather than 
isolation, between people who now claim and maintain different ethnic identities. Different ethnic 
groups claim some of the same settlement sites, such as Mangwesi Mountain, indicating that the 
distinction of ethnic groups did not exist at all or that the site was contested. 

Although Nata elders often claimed that they were reciting the place-names in 
chronological sequence of settlement by one group moving from place to place, this ordering may 
refer to concepts of space rather than of time. When asked to name the places in a certain area, 
elders ordered place-names linearly, as they would appear sequentially on a journey. Thus when 
Nata elders list place-names sequentially from east to west one must question whether these are 
successive settlements in time or in space. While some elders list these names in a sequential order 
of settlement, others simply list all of the places without order. I suggest that the ordering of place- 
names is spatial rather than temporal and that narrators conflate all of the settlements into the time 
of one generation just before the disasters. Elders order the settlement site names in this way not 
because they are sequential settlements in time but because that is how people remember their 
sequence on a journey. Memories are attached to a place rather than to a time sequence. 
Dating the Places 

If the ordering of place-name lists is not a temporal sequence but rather a spatial sequence 
how might the historian locate places from lists of place-names in time? Narratives link the place- 
names and their landscapes to a specific period of time, known locally as the "generation of 
settlement" (in Nata the word for settlement, ebimenyo, comes from the verb, komenya, "to build"). 
This is the period of time that Vansina refers to as the "floating gap." It has been a particularly 
difficult period for historians to analyze because of the cryptic nature of these traditions, 

"telescoping" a long period of time into one generation. Vansina explains this "gap" by the 
inability to keep chronology after a certain time depth, depending on the nature of the social 
structure and its way of reckoning time. 5 Although this has validity, what is more important is 
understanding why this particular set of place-names and the "generation" to which it refers 
represents the period of settlement in the western Serengeti. 

Place-names from the lists of oral tradition represent the period during which the land was 
settled by telescoping a long period of time into the memories about the generation just before the 
late nineteenth century disasters. Although many elders could not put a relative date to the 
settlements (they assured me it all had happened a very long time ago) others identified them as 
having happened in the time of the generation of the Abamaina, Amatara, or Amasura. These are 
among the earliest generation-set names that people remember in connection with specific ancestors 
and the last generation before the massive social transformations of the late nineteenth century. 
They probably date to between 1 850 and 1 870. In this area cycling names for both age- and 
generation-set are in use today. Yet narratives usually refer to these settlement sites in relation to a 
generation-set, rather than an age-set. This leads me to think that they refer to settlements before 
the late nineteenth century disasters when the way of calculating time shifted, particularly in the 
east, from one based on generation-sets to one based on age-sets. 6 

The social processes referred to in these narratives about one generation just before the 
disasters operated over a long period of time. We know from the evidence cited in Chapter 4 that 
hill farmers had settled in this area before 1000 A.D. However, these oral traditions tell of the 
generation of ancestors in living memory who established claims to the land that everyone still 

5 Vansina, Oral Traditions as History , p. 23. 

6 This shift in the reckoning of time from generation-set to age-set is argued more fully in 
Chapter 9. 

recognizes. Although settlers had sparsely populated the land before, people preserve the memory 
of "the generation of settlement" because it represents direct historical continuity with those living 
on the land now. The particular place-names in the lists recited by elders today seem to refer to 
remembered people and places just before the period of disasters, while representing social 
processes of much longer duration. 

Because social identity changed so drastically during the late nineteenth century disasters, 
people preserve only unconnected bits of knowledge from the period immediately before the 
disasters. Still, the survivors of the disasters did pass on the knowledge preserved by lineages who 
had obligations to their ancestors at specific sites on the land. The history of the period before the 
disasters, like these points on the landscape, appears as unconnected images of life in the nineteenth 
century, with no master narrative. 

The generation of settlement represents the first settlements of people whom western 
Serengeti people identify as lineage ancestors and who thus possess the power in present day life 
and historical imagination to provide protection and security for the living. People do not propitiate 
the spirits of first man and first woman in the emergence stories, nor do they remember their 
graves.' Clan founders, too, have no known graves nor do they have power to influence events in 
the present. The generation of settlement is the first generation that narrators remember by place 
and thus by name. Those ancestors are a living presence, with their own demands and obligations, 
among western Serengeti peoples today. This generation is responsible for preserving the health of 
the land and its people. 

7 The exception to this is the grave of Nyakinywa as first woman of the Ikizu. The 
reformulation of this origin story in the late nineteenth century was explained in Chapters 4 and 6. 

The Living Dead 

The dead are a part of the community and their descendants must maintain relationships to 
them with as much care as they give to relationships with living people and with the same 
possibilities for benefit and obligation. The word for the burial in Nata is kutindeka or "to store." 
A deceased person has not gone away but simply taken on a different form. The ancestors buried 
at the sites listed in these traditions occupy a special place as guardians of the land and the living 
must appeal to them for rain, fertility and protection. 
Obligations to the Dead 

Western Serengeti people bury their dead in the homestead and abandon the graves when 
they leave the settlement. Families or lineages do not maintain common burial plots through the 
generations. People are, however, expected to remember the grave-sites of their ancestors for at 
least two to three generations, to clean their graves and offer gifts there at least once a year 
(kusengera, meaning "to beseech" in Nata), often on the anniversary of death. They must return 
often to old settlement sites to meet their obligations. 

Numerous conversations over the course of my research convinced me that should a 
mother or father on their deathbed administer a curse on their children, this curse could have 
consequences in the children's lives. I heard many stories of people who refused to go to their 
father's death bed because they were too busy or did not do what their father asked before he died. 
Later their children began to get sick and die, they lost their cattle and wives, or their business went 
bankrupt. A father's blessing brings prosperity, good crops, many children and cattle, a thriving 
business and a large home full of people. The foolish son does not take this power seriously. 

People also remember their ancestors by naming children after a deceased grandparent. 
[See Figure 7-2: Respect for the Ancestors.] Until very recently parents would not give their child 
the name of a living person, because the ancestor lives through the child. The child acquires the 


Granddaughters named after Nyangere Bukaya, at her grave, Mbiso 

Sons of Magoto Mossi Magoto: Faini, Nyamaganda, Joseph Sillery, Mayani, Mossi, 
Nyawagamba, Manyika, at their father's grave, Mbiso on the anniversary of his death. 

Figure 7-2: Respect for Ancestors 

characteristics and personality of the person he or she was named after, and is treated as that 
person reborn. When a child is born parents or grandparents on either side often have dreams in 
which an ancestor appears to them and expresses the desire to have the child named after them or 
they see that their child resembles a deceased grandmother. 

Western Serengeti people experience the presence of the ancestors in the community by 
their frequent communication with the living in dreams. If the family no longer uses an ancestor's 
name or forgets his or her grave site, the spirit passes into a more dangerous realm of "loose" 
spirits, without community moorings." When problems occur in the homestead, such as illness or 
death, the head of the homestead consults a diviner, who often diagnoses the misfortune as the 
result of forgetting the ancestors. 

The word for the ancestral spirits of a family in the western Serengeti is simply 
omokoro/abakoro (derived from the word for "big" or "elder") or ekehwe/ebehwe ("ghosts" or 
"shadows"). Other Lakes Bantu speakers refer to the spirits of the ancestors by using the root - 
zimu (from -dpn,"be lost, extinguished"). These are the homestead ancestral spirits for whom a 
small spirit hut is constructed and whose propitiation is handled by the head of the homestead. 
Peoples of the western Serengeti do not normally build spirit huts in the homestead (exceptions 
may be healers and diviners).' 
Emisambwa: Guardians of the Land 

Yet the spirits connected to many of the important places named in oral traditions of the 
generation of settlement are different from the abakoro or ancestors in general. They are the 

8 See M. J. Ruel, "Religion and Society Among the Kuria of East Africa." Africa (London) 
35,3 (July 1965): 295-306, for outline of different kinds of spirits among the Kuria. 

This information about spirits comes from innumerable discussions but in particular an 
interview with Kinanda Sigara (Ikizu <?), Nyawagamba Magoto (Nata d - ) (who were my 
assistants), Chamuriho (Ikizu cf) and Mahiti Kwiro (Ikizu cr), Mchang'oro, 19 January 1996. 


erisambwa/ emisambwa spirits, spirits of important ancestors, often rainmakers or prophets, who 
are buried at the named sites where their lineages still make offerings to them. These spirits are 
always connected to the land and people propitiate them for help in resolving community problems 
such as rain, protection and fertility. Although ancestral graves (abakoro) are usually located in 
identifiable places and the spirits dwell in those spots, they have no efficacy over the land itself 
unless they are emisambwa. People consider the spirits of ancestors who were efficacious in life as 
rainmakers or prophets as emisambwa: those to whom the living will appeal for assistance. 
Nevertheless, not all important prophets or rainmakers become emisambwa. 

Erisambwa/ emisambwa in Mara languages derives from a Great Lakes Bantu root, samb- 
(ua), meaning "territorial or nature spirit, which protects first comers (often represented as an 
agnatic group)." 10 In other Lakes Bantu languages it means variously: "nature spirits of rivers," 
"spirits attached to larger lineage groups and to areas associated with these groups," "woods- 
dwelling spirit," (North Nyanza), "clan spirits and habitat for them-wild animals, rivers, etc.," and 
the "protective spirit of a settlement" (Rutara). Both Kirwen, who interviewed many diviners in 
Luo and Zanaki areas of the Mara Region, and Ruel, who worked with the Kuria, interpret 
emisambwa (abasambwa in Kuria) as forgotten and potentially dangerous, malevolent spirits." 
The western Serengeti elders with whom I spoke consistently referred to the emisambwa as 
powerful but none-the-less beneficent spirits, while the few Kuria interviews that I did confirmed 

Ruel's assessment. In other parts of the Lakes Region emisambwa can also be malevolent spirits. 12 

10 Schoenbrun, Etymologies . #347; See discussion of spirits in Chapter 5 and 6 of 
Schoenbrun, A Green Place: This substantive meaning was itself derived from a verb meaning "to 
judge" or, possibly, "to grant blessing," found more widely in Savannah Bantu. 

" Ruel, "Religion and Society," pp. 296. Michael Kirwen, The Missionary and the 
Diviner (Maryknoll, NY: Orbis Books, 1987), p. 6 

12 Schoenbrun, Etymologies . #347. 


The shift in meaning among peoples of the western Serengeti, from territorial, nature or 
malevolent spirits to ancestral spirits which guard the land, seems to suggest a particular kind of 
relationship to the land, dating to as early as five hundred years ago when North and South Mara 
speakers separated from each other. Western Serengeti people have collapsed the spirits of 
particular important dead ancestors (abakoro) with the spirits of particular places (emisambwa) to 
the point where they are now indistinguishable as emisambwa. This explains why western 
Serengeti people perceive emisambwa as beneficent-because they are known ancestors, not 
forgotten and lost spirits. The elders themselves were not clear whether erisambwa mean "the 
spirit of the ancestor" or "the spirit of the place," the two meanings have become synonymous. The 
meaning of emisambwa is implicit in the use of the emisambwa places in the place-name lists of 
oral tradition referring to the "generation of settlement." 

The Ngoreme, who are much closer to the Kuria geographically as well as culturally than 
other western Serengeti peoples, join these different conceptions of emisambwa. Many of their 
emisambwa do not demand ancestral sacrifices but are places remembered for important events. 
At these locations elders gather mud, particular tree branches, honey or water for other rituals. 
Animals or monsters inhabit some of these sites that are not connected to stories of ancestors but 
which a particular lineage is still responsible for propitiating. They also have emisambwa that are 
both ancestors and spirits of the land. One of the most famous Ngoreme emisambwa sites is the 
hot springs at Maji Moto. An elder of the Kombo lineage, who are ritually responsible at this site, 
said that the people of a whole village live under the water. Sometimes people hear a child cry 
there. When the colonial government came to explore volcanic activity at Maji Moto, the earth 
swallowed up all of their large machinery and those who did the work died once they got home. 

When the machines drilled into the rock, they brought up blood. 13 [Figure 7-3: Emisambwa in 

By collapsing the meaning of specific dead ancestors into the concept of a territorial spirit 
of a place, peoples of the western Serengeti made claims to the land that they occupied. Yet more 
than that, they made a profound identification of themselves, in the form of their ancestors, with the 
land. The spirit of the land is the spirit of the ancestors, one and the same. Kirwen's informants 
said that ancestors "are the ones who put us into the land, they are the elders and owners of the 
land, they are the ones with power to prevent or ward off sickness, famine and death, evil comes 
when we break our ties with the ancestors, they ensure that the moral order is kept." 14 

Although Tatoga herders deny that they have emisambwa sites, as do their Bantu-speaking 
neighbors in the western Serengeti, much evidence exists to suggest shared understandings of 
ancestral spirits. In Ngoreme a local farmer showed me a pair of large rocks worn smooth, 
presumably by the touch of hands, that he described as a Tatoga erisambwa for which Tatoga 
periodically returned to do the sacrifices [Figure 7-3: Emisambwa in Ngoreme.]. As the deeds of 
great prophets of the past were remembered so the landscape was appropriated with their relics. 
Tatoga elders said that the stone axe of the Relimajega prophet Gwataye is still embedded in a tree 
near the Mara River." The hunter/fishermen Tatoga, Isimajega, have a profound attachment to the 
mountain that they call Somega, in the western corridor of Serengeti National Park, now known as 

13 Interview with Maro Mchari Maricha, Maji Moto, 28 September 1995 (Ngoreme <?). 

14 Michael Kirwen, The Missionary and the Diviner , pp. 8. 

15 Interviews with Stephen Gojat Gishageta and Girimanda Marisha Gishageta, Issenye, 28 
March 1996; Merekwa Masunga and Giruchani Masanja, Mariwanda, 6 July 1995; Gilumughera 
Gwiyeya, Girihoida Masaona and Gorobani Gesura, Issenye, 28 July 1995 (Tatoga cf); Wambura 
Nyikisokoro, Sang'anga Buchanchari, 23 September 1995 (Ngoreme <3% Marunde Godi, Juana 
Masanja. Mayera Magondora, Manawa, 24 February 1 996 (Isimajega <f). 

David Maganya Masama, Wambura Nyikisokoro and Mayani Magoto, 
Sang'ang'a Buchanchari Ngoreme, at the Tatoga Erisambwa Rocks, 23 
September 1995. 

Maji Moto Hot Springs 
Figure 7-3: Emisambwa in Ngoreme 


Simiti hill across the Grumeti River from the Girawera game post. The Isimajega call the spring 
there Yiwanda, after the rainmaker prophet Ghamilay who is buried there. People went there to ask 
for rain, fertility, health or prosperity." 

The ancestors as emisambwa are just as real a presence in community life as those more 
recently deceased. An example from my field work shows the feeling of their personal presence. 
When I went with the elders to visit the grave of Gitaraga, a rainmaker, they did not address the 
spirit with formal ritualized speech, they spoke to him as one would speak to a living person. One 
elder from the lineage of Gitaraga brought along a gourd of water. We sat for awhile at the grave 
talking and then went to the place where the rainpots were buried in the ground under an over- 
hanging rock. After discussing whether 1 should take photos here and disallowing the use of pen 
and paper," he poured the water from the gourd out on the ground and spoke to Gitaraga: 

Mzee (elder) Gitaraga, we have come to greet you, we are your children, do not be angry 
with us but send us blessings, do not be astonished that some others of your children 
have come to greet you. They have not come for a long time, but they are never-the-less 
your children. They are from across the ocean. I8 

One the way home that day it poured down rain. [See Figure 7-4: Nata Sites, Gitaraga and 


16 The Isimajega elders requested that I ask the Park for permission for them to return there 
to propitiate the erisambwa. Tatoga ethnographies report large funeral mounds built for important 
elders in their cattle kraals. Poles which were planted in the mound grew into trees and a "sacred 
grove" was established, which would be visited by the man's ancestors for propitiation of his spirit. 
George J. Klima, The Barabaig: East African Cattle-Herders (New York: Holt, Rinehart and 
Winston, 1970), pp. 102-107. He reports that only rarely are these mound built for women. G. 
McL. Wilson, "The Tatoga of Tanganyika, Part I," Tanganyika Notes and Records 33 (1952)- 34- 

17 They finally agreed that 1 could take a picture but the photo did not turn out on a roll of 
otherwise good pictures. This was the last picture on the roll and when 1 went to change film I 
could not find the roll that I was sure I put in my bag that morning. 

18 Interview with Keneti Mahembora, Gitaraga, 9 February 1996 (Nata tf). 

Mokuru Nyang'aka, Barichera Machage 
Barichera, Nyawagamba Magoto, and Sochoro 
Kabhati, Riyara, 7 March 1996, the bee cave. 

Keneti Mahembora, Mokuru Nyang'aka and author, 
Gitaraga, 9 February 1996, Gitaraga's grave. 

Figure 7-4: Nata sites, Gitaraga and Riyara 


The association of emisambwa with forces of the wilderness, as opposed to the civilized 
spaces of the homestead, provides further insight into the identification of the spirits of the land 
with the spirits of the ancestors. People may not cut the groves of trees that grow up around these 
sites and they foster the untamed growth of these groves. Emisambwa sites are always located 
away from present settlement sites, in the bush. Even those in the more densely populated areas of 
Zanaki are found outside the village. One elder described these as places inhabited by leopards, 
snakes and where lions give birth." Traditions associate these places with the ritual symbols of 
water, fertility, women and growth. Many other anthropologists and historians of Africa have 
noted the recurrent ritual theme of mediation between the forces of the bush and the forces of the 
home. Feierman and Packard demonstrate the role of the king or chief as intermediary between 
wilderness and culture. 20 The peoples of the western Serengeti assign this role to a variety of 
important ancestors, located at specific places of power to mediate between wilderness and culture. 
Other Kinds of Emisambwa 

Emisambwa belong to a polysemous category used in a variety of other circumstances. 
Emisambwa spirits can also reside in particular objects or animals. The erisambwa of a place may 
appear as a snake or a hyena. Testimonies often call these animals the messengers of the 
erisambwa, but just as often say they are the erisambwa itself. One elder differentiated the 
emisambwa as "big" and "little" emisambwa. 21 According to this elder, the "big" emisambwa are 
those ancestors at certain places propitiated by particular lineages for rain or fertility. The "little" 
emisambwa are animals associated with lineages or clans, the emigiro or avoidances discussed in 

" Interview with Sochoro Kabati, Nyichoka, 2 June 1995 (Nata <f). 

20 Feierman, Peasant Intellectuals , pp. 69-93; and Packard, Chiefship and Cosmolog y , pp. 

21 Interview with Yohana Kitena Nyitanga, Makondusi, 1 May 1 995 (Nata cf). 

Chapter 6, that people must feed and be careful not to harm. The most common animals are a 
particular kind of snake (often pythons), the hyena, or a tortoise. Some Ikoma lineages with the 
hyena erisambwa have a special gateway cut in the homestead fence for the hyena to enter. Some 
people told me that the erisambwa at Nyichoka is a snake while others related the story of a barren 
woman, told in Chapter 2. 22 

Scholars have interpreted this understanding of emisambwa as "clan totems" or avoidances 
(emigiro). Yet seeing avoidances as emisambwa, with the nuanced meanings enumerated above, 
subtly shifts the traditional understanding of this phenomenon. For western Serengeti peoples, clan 
avoidances are more than a symbolic representation of the spirit of the collective. They are the 
located spirit of an ancestor who provides for the welfare of that clan and the health of the land 
they live on and use. This may also explain the references to "clan" as territorially based. If the 
clan has an erisambwa then that spirit must, by definition, have a dwelling place related to a 
people. This is another example in which narrators blur the distinction between lineage and clan by 
using the same idioms in reference to both. 

Ritual specialists such as healers, prophets, diviners and rainmakers each have their own 
erisambwa that directs their work. They are the spirits of known ancestors or perhaps the spirits 
that their ancestors used to do their own work. The emisambwa communicate with ritual 
specialists in dreams and tell them what to do. While people can learn or buy some medicines, 
ritual specialists do not choose to do this work. The emisambwa choose them and may make them 
ill or appear crazy until they agree to become a prophet. Yet this is not a possession cult as such 
and the ritual specialists do not take on the person of the erisambwa as much as receive help from 

22 The word, Nyichoka, means "the place of a snake." 

23 Claude Levi-Strauss, Totemism . trans. Rodney Needham (Boston: Beacon Press, 1962). 

it. 24 Each erisambwa has its own rules and prohibitions that the ritual specialist must follow to do 
their work effectively. Ritual specialists often have a special ornament or implement used by the 
ancestor that they call the erisambwa. The erisambwa relates to people of a particular lineage and 
ancestry. An erisambwa cannot direct a person outside the lineage, on either side. 

Dreaming is the most important way that the living communicate with the dead. This is 
the source of power for all ritual specialists: prophets, rainmakers and healers. Without the 
knowledge communicated by ancestors in dreams, prophets cannot perform their task. A general 
word for a prophet who dreams (and they all must dream to have power) is an omoroti/abaroli 
(from the verb, -rota, "to dream"). Common people also communicate with the dead in dreams and 
receive instruction, warning or encouragement for their daily activities. 
Emisambwa Sites as the Spaces of Lineage 

Specific lineages always control the emisambwa sites. Not just anyone can guide a 
stranger to these places. A representative of the particular lineage, ekehila, whose ancestors are 
buried at that place, must be present to approach those places and do the required rituals. They tell 
stories about the lives of these ancestors during a particular period of crisis, for example when 
there was famine and the rainmaker brought rain and prosperity. These sites are points on the map 
which represent particular lineages and the histories of great men and women in the past. 
Emisambwa may represent a matriline or a patriline. In Nata, ebehila (patrilines or "gates") 
control all of the communal emisambwa of the land. Those of the matrilines, (nyumba or 
"houses") are individual or "small" emisambwa. The Nata recognize only three important 

24 Differentiated from Cwezi cults, see Renee Louis Tantala, "The Early History of Kitara 
in Western Uganda: Process Models of Religion and Political Change" (Ph.D. Dissertation, 
University of Wisconsin, Madison, 1 989); and John Beattie, "Group Aspects of the Nyoro Spirit 
Mediumship Cult," Rhodes-Livingston Institute Journal (1961): 1 1-35; or Ngoma in John Janzen, 
Ngoma: Discourses of H ealing in Central Africa (Berkeley: University of California Press, 1992). 

emisambwa, while in Ngoreme and Ikizu each community or clan controlled territory, ekyaro, has 
its own. 

The Ikizu describe the emisambwa as both prophets and as spirits whom their founder, 
Muriho, put there when he conquered the land. They say that each prophet came to give direction 
to a particular lineage and place, so that they would have recourse in times of trouble. A particular 
lineage propitiates each of the emisambwa at their grave site. The Ikizu elders who wrote a book 
on their history listed twenty-one ritual places inhabited by the emisambwa. Out of that list more 
than half of the ancestors came from outside of Ikizu, for example, Nyambobe was a Luo woman 
who came in a boat with potatoes and bananas. 25 

Although not stated explicitly, those lineages who propitiate the spirits of their ancestors at 
the emisambwa sites may have authority there because of their status as those who came first. The 
ancestor buried at that place has a special connection to the land. As an erisambwa, the ancestor 
becomes one with the spirit of the land and thus is responsible for the health of the land. 26 Those 
who settled in the area first made the accommodation with the land (perhaps through association 
with first-comers like the Asi hunters) and have ritual authority over it. Because of the association 
of these places with symbols, like the python, that are ancient throughout the Bantu-speaking world 
and commonly used at sacred spots like shrines, one is tempted to hypothesize that these are 
ancient sites of power that have changed hands many times as new groups and new identities took 
control. 27 They could also represent critical sites of productive value located at the ecotone 

25 Mturi, "Historia ya Ikizu," pp. 25-27. 

6 For an example elsewhere in Tanzania of ritual precedence given to "first-comers" in 
ceremonies concerning the land see, H. A. Fosbrooke, "A Rangi Circumcision Ceremony: Blessing 
a New Grove," Tanganyika Notes and Records ( 1 958): 30-8. 

27 See Schmidt's analysis of the "tree of iron" in Historical Archaeology , p. 105. 

between mixed farmers and herders or hunters. In either case they are now represented by fairly 
recent lineage ancestors. 

Ritual Possession of the Land 

In relation to the generation of settlement, the meaning of these lists of place-names is more 
than simply remembering people and events from the past by places on the landscape. By naming 
places on the landscape western Serengeti people took possession of these sites. One cannot 
become a people without a land and the land must be ritually possessed rather than simply 
occupied. Ritual possession makes the land available for proper political use and ecological 
management; it "humanizes" the land. 28 It becomes "our land" through the practice of naming it 
and peopling it. The peoples of the western Serengeti were extremely mobile. They moved their 
homes and fields every 5-10 years, as many as ten times in a life time. 

Many historians have confused mobility with a lack of territoriality or feeling for the land. 
Paul Abuso, in his history of the Kuria states. 

The Abakuria did not have that sacred attitude toward land as a land of their ancestors. 
They had no such claim to the land they lived on. The owner of a home, when dead, was 
buried inside his cattle kraal and when the people migrated from the place the grave was 
completely forgotten. The people had no more attachment to it. This detached attitude 
of the Abakuria was seen in the way they welcomed many strangers in their midst. 29 

A similar sentiment is expressed by B. A. Ogot in his history of the Southern Luo. 30 Both seem to 

confuse attachment to the land with the exclusive occupation of one kingroup in the same plots of 

land for generations. This bias is perhaps the result of comparison with the dominant studies of 

kingdoms and centralized states in Lakes Region. 

David Schoenbrun, personal communication, with reference to ritual possession of the 
land throughout the lakes region. 

29 Abuso, A Traditional History , p. 35. 

30 Ogot, Southern Luo . pp. 38-39. 


People possessed the land, not in terms of ownership, but rather identified with it by ritual 
occupation, or by peopling it with the spirits of the ancestors who continue to respond to 
propitiation. That is why oral traditions represent the land as empty before first man and first 
woman came-because one's own ancestors did not people it. The ancestors of others must either 
be expelled or coopted in order to live peacefully there." The land is empty not when no people 
occupy it, but when no emisambwa dwell there. The Nata claim a much larger territory than that 
in which Nata people now live. However, the graves of their ancestors inhabit that larger territory 
and thus embody its extent. 
Muriho Possesses Ikizu bv Planting the Emisambwa 

The story of Muriho taking possession of Ikizu is the most vivid example of the process of 

peopling the land with the ancestors among the western Serengeti narratives: 

Muriho himself was a healer and a prophet, in his prophecy he was promised authority 
in a land of tall mountains and so it is at this high mountain that he established his 
settlement, ltongo Muriho. His goal was to possess the mountain now called 
Chamuriho, but he was not at first successful because people were already living there, 
called the Abahengere (short people). Muriho came up with a plan for overcoming these 
people. The Abahengere also became aware of the presence of Muriho and his people at 
Rosambisambi. So they came to attack Muriho, but when they arrived they found 
nothing there because Muriho had surrounded the settlement with protective medicine 
called orokoba. Muriho then passed the protective circle of medicine all the way around 
the mountain of Chamuriho and installed his own ancestral spirits in those places. There 
were five spirits in all, each had a name, specific powers, appeared in the form of a 
different snake and were placed in each stream where they are still appealed to today. 
The water in those streams became bitter so that the Abahengere were unable to drink it 
and had to leave the mountain of Chamuriho. . . . Muriho went on chasing them out by 
putting medicine in the water to make it unfit to drink and after they left each successive 
place he would pass the orokoba, the circle of protective medicine, so that they would not 
return to live in those caves. He chased them all the way to Lake Victoria. When 
Muriho was sure that his enemies would not return he went home and made a plan to 
complete the authority that he had gotten for himself. He went to the mountain 
Chamuriho and did the ritual purification called ikimweso, in order to bless these acts of 
courage and to protect his new settlement established on the mountain Chamuriho. By 

31 For a discussion of firstness as a principle of legitimacy see Schoenbrun, A Green Place . 
Chapters 4 and 5; and Kopytoff, "The Internal African Frontier," pp. 52-61. 


this time the prophet Muriho had many followers and had married eight wives, who were 
given to him as gifts of thanks for his actions. After this, once more, he had to use his 
powers and his medicine to drive out spirits and encircle the land with protective 
medicines, so that they would not return. 32 

The mountain where oral tradition says that Muriho lived is today called Chamuriho. the 
place of Muriho. Chamuriho is the tallest mountain in the area. The Germans used it for one of 
their heliograph stations. Approaching Chamuriho from the west, the small hills and tightly packed 
settlements of Zanaki suddenly spread out to the great Serengeti plains at the base of Chamuriho. 
The mountain marks a boundary between the lakes and hill peoples of the west and those who live 
on the interstices between hill and vast plain, to the east. Ikizu still make offerings on Chamuriho 
and know it as the origin spot and most powerful erisambwa of the Ikizu. Most of the other 
important Ikizu rituals either begin or end here. The Ikizu claim the land because Muriho planted 
their emisambwa at specific places to guard the land. [See Figure 7-5: Ikizu Sites at Chamuriho 
and Gaka.] 
Nata and Ngoreme Ritual: Mediating the Forces of the Wilderness 

The Nata ritual at the emisambwa sites illustrates the ongoing role of these spirits in 
mediating the dangerous but fertile boundary between wilderness and culture which makes 
habitation possible. When a problem arises the whole community comes to ask the lineage elders 
to do the sacrifices at Gitaraga so that they might have rain. Anyone can come along but only the 
lineage of Abene O'Gitaraga will do the sacrifices. When they go to Gitaraga only men and one 
young woman go. taking along a black sheep. The young woman carries a gourd full of water and 
dresses in traditional skins and beads. When they get to Gitaraga they clean out and refill the rain 
pots. Elders kill the black sheep and cut it in half from head to tail. The half with the head is for 
Gitaraga and the other half for his wife Nyaheri, at another nearby site. The group roasts and eats 

! Mturi, "Historia ya Ikizu.' 

Mtemi Adamu Matutu, Warioba Mabusi, Ntabusogesi Nying'asa, Wilson 
Wanusu, Godfrey Mayai Matutu, Mturi Wesaka, Joseph M. Nyaganza, Ikota W. 
Mwisagita, Kibiriti Kekang'a, Makongoro Wambura, P.M. Mturi, Ikizu elders at 
the grave of Nyakinywa, Gaka, 31 August 1995 

Figure 7-5: Ikizu sites at Chamuriho and Gaka 

the meat at the site. Then a young man climbs the tree (omusangura) near to the grave of 
Gitaraga and pours water over the head of the young woman who is bent over below the tree. The 
others cut branches from the tree and wave them while the youth pours water. They also sing 
songs. As this is happening, a lineage elder asks Gitaraga to send rain. The same thing happens 
at the other Nata site, Geteku: the elders propitiate both male and female emisambwa, pour water 
out to imitate the rain, sacrifice, roast and eat an animal. At this site women are specifically there 
to sing and dance the eghise. If the erisambwa is happy with the ritual, the beat of a drum sounds 
(ambere). 33 

In Ngoreme, the emisambwa of Kimeri and Nsoro are two springs up on the hill behind 
Maji Moto. The Gitare lineage cleans the springs periodically. They take tobacco, milk and honey 
for the prayer, the women take flour and the men a white tasseled goat. The women draw lines on 
the ground with flour and the men spit the mixture of honey and milk to the four corners and onto a 
stone that they walk around as they invoke the spirit. Elders said that the rain would start before 
they reached home. They do not kill the goat there but take it back home with them. They also 
bring leaves from certain trees at the site home for other rituals. At circumcision time the initiates 
come there to get water and white mud for the ceremonies. 34 The Gitare perform these rituals in 
times of trouble, such as lack of rain, ill health, infertility or threat of enemies, rather than at 
regular intervals. They visit the emisambwa regularly at the initiation of a new age- or generation- 
set. The new set goes there to receive the blessing of the spirits for a prosperous period when their 
age-set is in power. 

13 The most important informants on the Nata emisambwa were from Gabuso Shoka, 
Mbiso, 30 May 1995; Mokuru Nyang'aka and Keneti Mahembora, Gitaraga and Nyichoka,9 
February 1996; Makuru Magambo, Geteku, 9 March 1996; Mahiti Kwiro, Mchang'oro, 19 
January 1 996 (Nata o"). 

34 Interview with Reterenge Nyigena, Maji Moto, 23 September 1995 (Ishenyi if). 


These rituals reenact the relationship of people and land. The young fertile woman 
standing beneath the tree represents the land receiving from above the male spirit rain. The women 
dance the eghise to please the spirit who blesses the land with rain. The people bring the products 
of their labor on the land as farmers (flour), herders (goat and milk), and hunter/gatherers (honey) 
and offer it back to the spirit that has made prosperity possible. In turn they take the powerful 
things of the wilderness (leaves, clay, water), now made safe for use in the civilized world by the 
spirit of the land, back to perform community rituals. 

People recognize prohibitions commonly associated with emisambwa sites. Where there 
are pools or springs, women can only draw the water using traditional vessels like gourds. They 
may not use metal or plastic buckets, nor anything that is red. The forest surrounding these sites 
never burns nor can it be cut. 35 Each of these items marks the emisambwa as spirits associated 
with the wilderness, whose power is mediated by the ancestors. 
The Ikoma Machaba E risambwa: Domesticating the Ancestral Spirits of Others 

The Ikoma erisambwa of mobile elephant tusks, rather than a fixed place, demonstrates the 
adaptability and flexibility of these concepts through which people gained possession of the land 
through the spirits of the ancestors. Here a relationship to the land could only be claimed by 
appropriating the ancestral spirits of the Tatoga herders. Although the Ikoma lineages have their 
own emisambwa sites, these have become subordinate to the collective Ikoma erisambwa-^ large 
set of elephant tusks known as the Machaba. Relative dating by generation-set places the story of 

35 Interview with Sochoro Kabati, Nyichoka, 2 June 1 995 (Nata <f). 


how the Ikoma got these tusks in the middle time period and around the mid-nineteenth century. 36 

[See Appendix for other versions]: 

Then later the rain slopped again and they called everyone together to take action, along 
with the Ishenyi people. They met to decide where to go to get a prophet who would help 
them. They decided to go south to the Tatoga. the Bachuta, at Ngorongoro Crater, in 
Mbulu. They went to beg him for help and he gave them their erisambwa, the Machaba 
(which is named after the Tatoga prophet himself). It was to be for both but the Ishenyi 
were unable to carry them. The prophet gave one cow to the Ikoma and one to the 
Ishenyi. The Ishenyi were not happy with their cow because it was so thin and they 
wanted the Ikoma cow that was fat. The Ishenyi had more people and so thought they 
should have the fatter cow. They took the Ikoma cow and the prophet let them go ahead 
and do it. When they butchered it. they found it was thinner than the other. When they 
got the erisambwa of the Machaba the Ishenyi could not carry the tusks. Mwishenyi said, 
"let them go ahead and carry it home and we will take it from them there. " They came 
first to Ikoma and then went onto Nyeberekera. They said they were tired and would 
come back for the Machaba later, but they never did. It stayed in Ikoma. The Machaba 
is their elder. Anything you want to pray for he can grant. It is a Tatoga erisambwa. 37 

Another version varies only slightly. 

They went to the Tatoga prophet, east in a crater but not Ngorongoro. another one near 
Mbulu called Mwigo wa Machaba. There was a lake in the crater. They went there 
because they had a problem with fertility. The Ishenyi, who were more numerous than 
they, came along too. The Ishenyi slept at the first place inside the gate, the Ikoma slept 
outside the gate. The prophet said they should grab a sheep as they jumped over the 
gate. Ikoma got a skinny one and Ishenyi a fat one but when they butchered them, the 
sheep looked the same and when they were cooked the Ikoma one was fatter. The prophet 
tried each of their bows and shot the Ikoma arrow far off and said they should follow it. 
He prepared the things that they should take along with them (mbanoraj and showed 
them the path to take when they saw vultures up ahead. The youth ran ahead to get the 
prize. The first to get there was Mayani (a Gaikwe clan member of Ikoma) who took the 
top (right) tusk of the elephant and second was a youth of the Ikoma Himurumbe clan 
who took the lower (left) tusk. The Ishenyi wanted to take it from them but the prophet 

36 While most informants would not date the Machaba story one elder said that the Ishenyi 
were at Nyigoti (Mang'ombe Morimi, Issenye Iharara, 26 August 1 995) which would put it during 
the period of late nineteeth century disasters, others dated it to the time when the Ishenyi were still 
at Nyeberekera, just before the disasters (Morigo Mchombocho Nyarobi, Issenye, 28 October 

1 995 and Machota Sabuni, Issenye, 1 4 March 1 996). Tatoga informants dated it to the time of the 
prophet Saigilo's father which would also date it to the mid-nineteenth century period just before 
the disasters (c. 1 850- 1 870). The fact that they went to the Tatoga prophet because of infertility 
problems would suggest that the disasters had already begun. 

37 Interview with Machota Sabuni, Issenye, 14 March 1996 (Ikoma tf). 


had said not to fight. They tried to take it away but could not lift it or move it. The 
Ikoma were at Tonyo at that time™ 

At all important communal rituals the Ikoma bring out the Machaba tusks. The people 
receive a blessing by touching the tusks. The Ikoma clans are divided into two moieties, Rogoro 
(east) and Ng'orisa (west). Each moiety guards one tusk. Elders tell many stories of times in 
which the Machaba were hurt or taken and the bad consequences of that action. The colonial 
government tried to take them as well, but failed and brought them back. 39 

The Machaba story also defines the relationship between the Ikoma and Ishenyi, both 
going together to find prophecy concerning their lack of fertility. This story explains why the 
Ishenyi are such a small group today and the Ikoma relatively larger. The Nata have a version of 
this story in which they go along too and the prophet gives them a set of buffalo horns that were 
later lost when one group failed to pass them on at the proper time. The Machaba story stands 
alongside the origin stories as a way of explaining the relationship between western Serengeti 
peoples. Yet given the time period of this story (c. 1870) it may also commemorate the formation 
of Ishenyi and Ikoma ethnic identities as ritual communities related to particular emisambwa. 

In Ikoma the clans, rather than the lineages, control the emisambwa, the Machaba. The 
Gaikwe (Ng'orisa) have the right hand, upper or male tusk which confirms this clan as first- 
comers, similar to their status in the origin stories as the clan of first man. The Himurumbe 
(Rogoro or "east") control the female side or left tusk, often related to first woman from Sonjo (to 
the east). One elder confirmed that the Himurumbe clan was also Asi, or hunter/gatherer in origin. 

(Ikoma <f) 

38 Interview with Bokima Giringayi, Mbiso, 26 October 1995 (Ikoma <f). 
Interview with Mabenga Nyahega and Machaba Nyahega, Mbiso, 1 September 1995 

These clans have special ritual functions when they bring out the Machaba and they also keep the 
tusks, in separate places, one located in the east and one located in the west. 

On the other hand, in contradiction to all other western Serengeti emisambwa theory, the 
Machaba are not ancestral spirits of the Ikoma but of Tatoga origin. This brings the Ikoma into a 
very special relationship with the Tatoga. One Ikoma elder said that the Tatoga were "people of 
the oath (ring'a)," or "our parents." 40 The Ikoma commemorate the Tatoga role as spiritual parents 
in ritual practice, as described in the next chapter. In one version of the origin story the first Ikoma 
man from Sonjo came because a Tatoga prophet told him to follow the animals until he found a 
place where lions lived, he should then stay at that place. 41 The Machaba story seems to mark 
another era of negotiation and interaction between Bantu-speakers and Dadog-speakers long after 
the original period of settlement. 

Because the Machaba is a Tatoga spirit, it is a mobile, rather than a located erisambwa. 
The Ikoma have domesticated this spirit and appropriated its power by fixing it to the land. 42 The 
Machaba erisambwa dwells in the tusks rather than a physical feature of the land. The spirit 
represented by the tusks is a Tatoga prophet from Ngorongoro. The next chapter will describe the 
rituals to encircle and protect the land carried out by the generation-set. The Ikoma generation-set 
fixes the tusks to the land by carrying the Machaba in their ritual walk to seal the boundaries of the 
land. In addition, they cannot take the Machaba across the Grumeti River, which Ikoma 
acknowledge as their "traditional" territorial boundary. When the British took the Machaba across 

40 Interview with Sabuni Machota, Issenye, 14 March 1996 (Ikoma <f\ he was an 
important informant on all aspects of the Machaba. 

41 Interviews with Mabenga Nyahega, Bugerera, 5 September 1995; Moremi Mwikicho. 
Sagochi Nykipegete, Kenyatta Mosoka, Robanda, 12 July 1995 (Ikoma <f). 

42 For this process in other parts of the Lakes Region see Schoenbrun, A Green Place . 
Chapter 5. 

the Grumeti their car broke down and the Machaba were out of the tin house (where the officer 
stored them) in the morning. When one of the keepers of the Machaba moved them across the river 
his whole family began to sicken and die until he moved them back. 45 

Thus, the Machaba represent a further cultural elaboration of the polysemous category of 
emisambwa spirits and their guardianship over the land and the health of the people. The Ikoma 
are the people farthest east and farthest out on the plains. The village of a Robanda clusters 
around a hill which rises out of an otherwise flat and featureless plain. The ecological setting 
suggests that the only way to prosper on this kind of land is to appeal to the spirits of those who 
own the grasslands, the Tatoga herders. Nevertheless, the Ikoma fix and domesticate mobile 
Tatoga power by an Ikoma understandings of the relationship between land and people. 
Nineteenth Century Settlement Patterns 

The cultural understanding of the places listed in oral traditions about the pre-disaster 
period not only provides insight into the sequences of relationships between early and later settlers 
to the land, but also sheds light on the kinds of settlements occupied during the nineteenth century. 
Because these traditions only provide glimpses into individual settlements, reconstructing patterns 
or forming generalities, without a master narrative, is difficult. Nevertheless, I use evidence about 
specific settlement sites in conjunction with ethnographic and linguistic evidence to reconstruct 
some idea of what settlements were like in the nineteenth century. 
Abandon ed Settlement Sites and the Patronage of "Big Men " 

Some places in the list of place-names are not emisambwa but primarily old settlement 
sites or what the Nata call ebimenyo (literally "built places"). Traditions distinguish these sites 

Interviews with Kimori Gamare, Bugerera, 15 July 1995 (Nata/Ikoma ?); Nyaruberi 
Kisigiro, Morotonga, 12 July 1995 (Ikoma <f); Sabuni Machota, Issenye, 14 March 1996 (Ikoma 

from the emisambwa sites by arguing that, although people remember specific ancestors and events 
of the past at those sites, they do not propitiate the spirits there by sacrifice nor do these spirits 
have power over the health of the land and its people. Elders from different lineages may recite 
these settlement site names differently, given that they would remember those of their own lineage. 
While all elders fairly consistently name the emisambwa sites, they disagree on the ebimenyo sites. 
Elders speak in most detail about the sites more commonly known that seem to have been the most 
prosperous settlements of important or wealthy individuals. 

People often name places after the well-known individuals who lived there, without formal 
title but with charismatic ability to attract people as a "speaker" (omukinalabakina, 
omwerechi/abawerechi, omugambi/abagambi). A speaker in western Serengeti tradition is a man 
whom people respect for his ability to speak the mind of the community with wisdom and fluency. 44 
These men may have had wealth to back up these claims but they measured their wealth in people 
whom they could attract through extensive relations of reciprocity rather than in things. They were 
men with many "children" and large homesteads. The place-names Magita and Wamboyi in the 
Nata list refer to such wealthy men and their settlements. 

Local languages use the term omwame (or omonibi particularly for cattle wealth) to refer 
to a wealthy man. Many conversations over the research period concerning the definition of wealth 
convinced me that people value wealth in crops or livestock only in so far as it generates people as 
wives, children or dependents. Men aspire to the respect that wealth brings when they are able to 
feed a large crowd of people at a feast, a community ritual or a dance. 45 All wealth is not equal 

44 Schoenbrun, A Green Place , pp. 1 99-200, demonstrates the connection between the 
power of speech and healing or divination, or more generally speech as creative power. 

45 Interviews with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi cf); Sarya 
Nyamuhandi and Makanda Magige, Bumangi, 10 November 1995 (Zanaki d% dealt with these 

and only the wealth generated by the sweat of farming and herding is called umwame, used to 
produce a large lineage. The wealth from the mines or from hunting elephants is illegitimate wealth 
with which one can never build a homestead. 46 

The wealthy men of settlements remembered in oral tradition are lineage elders who 
represented a period of prosperity. People still remember these men and the places where they lived 
because they forged networks of reciprocity over a large region. A wealthy man was one who 
"fed" his people in times of trouble and, in turn, commanded respect, labor and support. In the 
uncertain environment of the western Serengeti these lateral links to wealthy men provided security 
and resources for building prestige throughout the region. A wealthy man was an omwame 
because of his informal ability to control widespread and diverse networks of security through his 

The word mwame is an old Lakes Bantu word from the root -yaami, meaning "chief." 47 In 
the early period of Great Lakes Bantu (500 B.C. to 500 A.D.) settlement these leaders held their 
position, in part, because of their ability to distribute wealth. 48 Vansina describes similar "big 
men" (named mukijni/) in the Equatorial tradition whose authority lay in their ability to attract 
followers through their wealth. Historical linguists have established the ancient "link between 
leadership and the exchange of goods" through political terminology generated from the words for 
"gift," "to give away," or "to divide" throughout Bantu-speaking Africa. 4 ' The difference between 

topics in particular. 

46 Interview with Philemon Mbota, Mugumu, 17 November 1995 (Kuria o"). 

47 Schoenbrun, Etymologies . #261. 

48 Schoenbrun, A Green Place , p. 183; See also Ehret, Classical Age . Chapter 5. Miller 
uses a similar model to explain the slave trade, Miller, Way of Death . 

49 Vansina, Paths in the Rainforest , p. 74. 

Vansina's "big men" in Equatorial tradition and the Lakes Bantu use of the word mwaami was that 
in the Lakes region these leaders achieved their role, in part, as lineage elders. 50 

The use of the term mwame in the western Serengeti represents a variation on an ancient 
bundle of political culture and practice in response to an environment where resources were 
extensively available in plentiful but marginal land rather than intensively controlled for exclusive 
use. Here the mwame was not a chief but a wealthy man, who was the leader of a (at least 
purportedly) lineage-based settlement. His extensive and informal links throughout the area forged 
by his wealth were more important than a concentration of followers at home. The word for a poor 
person (omuhabe) is the same as the word used for an orphan or a person without family, affirming 
the central role of descent idioms in elaborating theories of "wealth-in-people." Elders were at loss 
to give me the words for "patron" or "client" because the concepts did not exist, except in reference 
to the late nineteenth century practice of capturing slaves, in which case they used the Suguti term 
for slave (omuseese).*' 

The Zanaki called the powerful rainmaker from the clan territory of Busegwe, the mwami, 
which is also derived from the word for a wealthy man. In the colonial years the Zanaki 
successfully argued that this was a chiefly title. Indeed many aspects of the Zanaki mwami's 
power were chiefly-other clan territories brought him or her tribute in goods or labor and clans 

50 Schoenbrun. A Green Place, p. 183. Ehret, Classical Age . For the ethnography of "big 
men" elsewhere see, J. P. Singh, Politics of the Kula Ring: An Analysis of the Findings of 
Bronislaw Malinowski (Manchester: Manchester University Press, 1971, first published 1962); 
Douglas L. Oliver, A Solomon Island Society: Kinship and Leadership among the Siuai of 
Bougainville (Boston: Beacon Press, 1955). 

51 Various sessions with Nyamaganda Magoto over the course of 1995 to fill out the 
Cultural Vocabulary list. Seese-"a wild unruly animal, usually a dog, sometimes a jackal," 
Schoenbrun, personal communication. 

throughout Zanaki respected his or her word. 52 However, 1 suspect, given the use of the term 
throughout the Mara Region, that mwami, as a title of chiefly office rather than a generic term for 
a rich man, was reintroduced in the nineteenth century from across Lake Victoria. Busegwe 
tradition says that the first mwami came from the Mugango peninsula on the lakeshores, as an 
immigrant in the line of the great female rainmaker Muse from Buhaya. The rainmaker lineage 
claims kinship with the mukama (chief) from Kerewe Island, also with origins across the lake. 53 
Ruri elders on the Mugango peninsula said that the Busegwe must return here to make the powerful 
raindrum used by the mwami?* The mwami lineage of Busegwe included powerfully respected 
rainmakers who began assuming the authority of chiefship only in the late nineteenth century and 
in the context of colonialism. Immigrants from across Lake Victoria seem to have brought the 
meaning of the term mwami as a chief, rather than a wealthy man, as a tool for asserting the 
authority of their lineage. On the other hand, the shift of mwami from a wealthy man to a chief 
may just as easily have occurred as an internal innovation in the context of nineteenth century 
societal stress and increasing hierarchical accumulations of wealth from the caravan trade. The 
connections to the coast and across the lake may in that case have underpinned the access to wealth 
that the Zanaki used to transform the rainmaker into a chief. 55 

52 Benjamin Mkirya, Historia. Mila na Desturi za Wazanaki (Ndanda, Tanzania: 
Benedictine Publications Ndanda-Peramiho, 1991), pp. 45-55. 

53 Hans Cory, "Report on the pre-European Tribal Organization in Musoma (South Mara 
District and ... Proposals for adaptation of the clan system to modern circumstances)," 1945, 
CORY# 173, EAF, UDSM. See also an evaluation of the Zanaki mwamiship, Kemal Mustafa, 
"The Concept of Authority and the Study of African Colonial History," Kenya Historical Review 
Journal 3.1 (1975): 55-83. 

54 Interview with Daudi Katama Maseme and Samueli Buguna Katama, Bwai, 1 1 
November 1995 (Ruri <f). 

Thanks to David Schoenbrun for this alternate interpretation, personal communication. 

It also makes sense that where chiefly authority began to develop in the Mara Region 
(Zanaki, Ikizu and among the Lakes people) rainmakers assumed this power through the obugabho 
tradition (discussed in the last chapter). This word is derived from the old Lakes Bantu root, 
-gabira, or "to divide up, distribute," usually in the sense of one who gives big feasts or gives 
things away generously. 56 Mara languages use the verb kugaba only in reference to the division of 
inheritance and the noun form only in reference to ritual specialists such as prophets and 
rainmakers. It may be then that those most closely connected to the "big man" tradition described 
by Vansina were the prophets and rainmakers, abagabho, who distribute a different kind of wealth, 
in knowledge. However, in the Mara Region obugabho was not a gendered term and some of the 
most famous rainmakers, like Muse (Zanaki) and Nyakinywa (Ikizu) were women. Vansina's 
characterization of big men would thus have to be revised. In other parts of the Lakes region 
people conceived of political and spiritual or healing power as separate categories in the two offices 
of the chief and the prophet. In the western Serengeti, without political centralization, the 
categories were not distinct. 57 

The association of "big men" with the distributive power of prophecy is evident in some 
settlement sites that are not emisambwa sites and yet represent wealthy men or women who 
controlled powerful "medicines." In Nata one of the important settlement sites is at a hill, named 
after a man called Riyara. Elders said that Riyara was a prophet who had control over bees that he 
kept in a cave here. When a conflict emerged, the Nata warriors came to Riyara, who gave them a 
small wooden box with bees and other medicines inside. On the battlefield they released the bees 
against their enemies. The bees are still at the place of Riyara and someone in his lineage inherits 

56 Schoenbrun, Etymologies . #164. 

57 Schoenbrun, A Green Place, pp. 194-201. 

the power of his medicine bundle. [See Figure 7-4: Nata Sites, Gitaraga and Riyara. p. 333.] 
Medicine bundles for protection in war, ekitana, are commonly used and inherited through the 
lineage. The prophet Riyara links this medicine bundle to a bee hive in a cave at Riyara and so 
places it spatially. Important Nata generation and age-set rituals must use honey from the bees at 
Riyara even today. People do not make sacrifices there because this is not an erisambwa, instead 
they ask Riyara, whose spirit dwells there, to calm the bees so that they can take the honey. Only 
Nata people can take honey there, because the others do not know how to call on Riyara to calm 
the bees. Other ebitana of the Nata are place-fixed in that the person in the lineage who is chosen 
to keep the ekitana must not cross the river boundaries of Nata. 58 

These ebimenyo or settlement-sites recounted in oral tradition may also be associated with 
the age-set or generation- set which lived there, and some important events of the time. The 
remembered sites were usually prosperous with lots of food and people. Elders remember 
Torogoro and Site because these settlements produced so much food that they had lots of leisure 
time to dance. The Abamaina generation danced so much that the youth pounded the dance field at 
Torogoro into a depression that one can still see today. 

Elders say that some settlements, such as the ones as Sang'anga and Kyasigeta for the 
Nata, were settled long, long ago, abandoned and resettled around the time that the Germans began 
to build Fort Ikoma in that area. At that time a woman at Sang'ang'a used the ekitana from Riyara 
to protect against the Maasai with bees. She had a drum in her loft that would sound when 
enemies were near. She took the drum outside and stripped off her own clothes as she released the 

58 The main information on Riyara from a trip to Riyara with Makuru Nyang'aka, Sochoro 
Kabati, Barichera Machage, Nyichoka to Riyara, 7 March 1996 (Nata &). 

bees. 59 Her medicine was so powerful that when the Germans first began to build their fort at 
Sang'ang'a every morning they would awake to find their foundations torn down. They finally 
gave up at that spot and built at Nyabuta, Fort Ikoma. The history of re-occupation of settlements, 
of settlement names used in other settlements and the telescoping of time makes it almost 
impossible to differentiate the settlements chronologically. 

The spaces of these settlement sites represent the spaces of small scale communities as 
points on a landscape. Elders can identify almost every settlement site according to the lineage or 
clan that lived there, without reference to ethnic group. Since wealthy men and ritually powerful 
women, held authority, in part, because they were lineage elders, traditions identify their 
settlements according to lineage. These settlements seem to be separate communities, inhabited 
during approximately the same time period, who were linked to each other by networks of 
redistribution, across what are now ethnic boundaries. By naming and remembering settlements of 
prosperous men and times of plenty, people identify themselves with the same processes of 
reciprocity between land and people that made prosperity possible. 
Settlement Patterns: the Oruberi 

These stories about particular settlements are consistent with the testimony of elders 
concerning settlement patterns before the disasters. The people I talked to described settlements in 
the pre-crisis era as a relatively concentrated collection of patrilineally or matrilineally related 
homesteads called an oruberi. Further confirmation of this pattern comes in the identification of 
nearly all past settlements with particular lineages. One person said that the homesteads were 

59 Interview on the ekitana of Mantarera with Sochoro Kabati, Nyichoka, 2 June 1995 
(Nata cf). 

close enough to each other that people could hear a shout from the adjacent homestead. 60 People of 
the same clan grouped their settlements within one area, or on one set of hillsides, called the ekyaro 
or territory. A brush fence often surrounded the oruberi that contained four to ten or more 
homesteads. Elders contrasted this pattern with the much more highly concentrated settlements and 
stone forts at the height of the Maasai raids. Without archeological evidence, to tell how far before 
the disasters this pattern might have extended is impossible. 

A common grazing area called the ekerisho, always lay near the oruberi." It is not clear 
whether farmers always grazed their livestock near to the oruberi or whether this was another 
change that resulted from the time of raids. In any case elders remember no other pattern and 
grazing near to the village is consistent with the strategy for maintaining trypanosomiasis immunity 
discussed in Chapter 5 and is a wide-spread regional pattern. Members of the oruberi grazed their 
livestock together and farmed in areas often many hours walk from the oruberi called the ahumbo 
fields. 62 Each family, and each wife, had their own ahumbo fields that were adjacent to each other 
and surrounded by a brush fence. Young people were left at the fields during the night to guard 
against wild animals. They stayed in temporary houses called ekeburu during the farming season. 
The old and the very young stayed back at the oruberi, along with enough youth to guard the cattle 

60 Interviews with Surati Wambura. Morotonga. 13 July 1995 (Ikoma S); Jackson 
(Bcncdicto) Mang'oha Maginga, Mbiso, 18 March 1995 (Nata <f); Mariko Romara Kisigiro, 
Burunga, 31 March 1995 (Nata <?). Much of this information is pieced together from 
conversations which contrasted the fort settlements of the disasters with this early pattern and the 
contrast to present village structure. 

61 See F.hret. Classical Age , for a sense of the great time depth here. 

° 2 An early German report states of the "Washashi and Wangorimi," "the fields in some 
cases are several hours' journey from the houses." Geographical Section, A Handbook of German 
East Africa , pp. 97. 

in case of a raid. Economic cooperation and mutuality within the oruberi were as important as 
within the homestead. 

People discussed the ideal oruberi of the past as a homestead on a larger scale. The 
oruberi fence had one gate for livestock and a secret backdoor for emergency escape, just like the 
homestead. Elders compared the brush fence surrounding the oruberi to the homestead fence, 
using the same word (orubago). By definition everyone in the oruberi was part of the same 
patrilineage. as they would be in the ideal homestead. Yet just like a homestead, many people 
were incorporated that did not share a genealogical connection. The ways that communities 
incorporated strangers in the past demonstrate how the homestead model makes allowances for 
people not genealogically related. Lineal descent was the idiom through which western Serengeti 
peoples conceptualized the relationships within one settlement. 

The settlement structure of villages today, even with the vast changes of the past century, 
represents the ideal of lineage-based settlement. These changes include the reformation into 
concentrated age-set settlements at the end of the nineteenth century; the movement back to 
dispersed settlements oriented around wealthy men in the colonial period; and finally the imposition 
by the independent government of Tanzania of concentrated "Ujamaa" villages in the 1970s. 63 In 
spite of this history, organization of most villages in western Serengeti, whether "Ujamaa" or not, 
tends to revolve around the relationships of a couple of key families. The spatial organization of 
the village often situates lineage-related families in the center, with those related by friendship or 
patronage to those families on the periphery. 

B "Ujamaa" was a nationwide scheme instituted after the " Arusha Declaration" in 1 967 to 
deploy "African Socialism" by resettling everyone in planned villages rather than dispersed 


The present day village of Mbiso in Nata is a good example of these ideal patterns in 
practice." Magoto Mosi, the father of my host in Nata. founded this village. His original 
rectangular, tin-roofed house, sits at the crossroads of the village, on the truck route to Arusha. 
Musoma and Mugumu. The rest of the village grew out from that point with his sons' houses 
closest to his and his daughters', who married within the community, farther out. Many people in 
the village came there because they respected Magoto and because his prosperity had generated 
many relationships of reciprocity, relationships he considered friendships. Almost everyone in the 
village, whether of his clan/lineage or not, can trace a relationship of some kind back to Magoto 
himself. The village grew up around the personal patronage and influence of a "big man." 

The internal spatial structure of the village differentiates itself into lineage and age-set 
based categories. Almost everyone in the village is of the age-set cycle of Bongirate, reflecting 
changes in the residence patterns at the end of the nineteenth century. Bongirate of Magoto's clan, 
Moriho, live at the east end of the village, while Bongirate of the Getiga clan inhabit the west end 
of the village. A few families of the Saai age-set cycle, who came in because of association with 
Getiga clan members of the Bongirate age cycle, live on the west end of the village. A number of 
Getiga clan members from ikoma also joined fellow clan members at the west end of the village. 
Quite a mix of people came in at the time of "Ujamaa" from the settlement of Sibora that they 
abandoned because of its designation as a game-controlled area by the government. Sibora people 
were all Bongirate and so moved into the east end of the village, of whatever ethnic group or clan. 
The exception to this pattern is the enclave of school, court, police and other government workers 
living in the government housing section on the west side of the village. 

This information is based on an informal village survey that I did with Nyamaganda 
Magoto in late 1995 in which he and others named the inhabitants of each house in the village, 
their clan, age-set. relation to Magoto. occupation and other relevant information. 


These diverse people who moved to Mbiso because of personal ties to Magoto found a 
place for themselves by settling near to those with whom they shared age-set, clan or lineage 
membership. People related both through Magoto's mother's family and his father's family claimed 
kinship as the basis for moving to the community. Since most Nata people are related to each other 
in multiple ways, these settlers used kinship to assert their strongest claim to a reciprocal 
connection with Magoto. Structurally, then, the village is a lineage-based settlement, but the logic 
of those who decided to move was based on personal patronage. 

Considering these more flexible ideas about how people understand their relationships to 
each other and how they come to congregate in one area, we must also question the assumption of 
strict lineage-based settlement in the past. Because local languages derive the names for lineage 
from the model of the homestead— gateway (ekehita), house (anyumba), hearth (rigiha), we might 
more usefully think of the understood relationship between peoples within one settlement as an 
extension of the homestead model. 65 People may have congregated together in one settlement 
because the presence of an influential man assured them that they could find prosperity and 
security there. Yet the formal ways in which they explained their choices and related to others 
within the settlement were based on the idiom of lineage and later age-set. 
Abasimano: The Incorporation of Strangers 

While elders said that everyone in an oruberi claimed one lineage, the underlying 
philosophy of people and land in the western Serengeti was not exclusive. Because land was 
plentiful, and people scarce, communities gave inclusiveness high priority. According to local 
testimony specific mechanisms existed to incorporate individual strangers (abasimano) and even 

65 Kuper does this in his analysis of the Zulu state as representing deep continuities with 
the house model rather than in terms of lineage. Kuper, "The House and Zulu Political Structure;" 
Schoenbrun, "Gendered Histories," pp. 470-480. 


stranger clans into the group. Their incorporation rendered them natives (abibororu. meaning 
"those who were native born"). Wealthy elders incorporated hard working strangers into their own 
lineages or grafted the stranger lineage onto their own. These strangers accepted initiation into the 
local system of titles and took an oath not to leave the land or betray their adopted people. 

Lineage and genealogical relatedness was the idiom through which people understood their 
rights and obligations to each other but it was their common residence that united them and made 
them one people. The children of an omosimano, or stranger, are ommbororu (abibororu). or 
"native born," and not differentiated from their peers of native born parents. At issue is not blood 
or biological inheritance but where a person was born. These devices quickly erase the origins of 
abasimano and few signs of it remain for their descendants. The family cannot discuss their 
stranger origins until a couple of generations have passed. While Nata genealogies disguise this 
diversity, almost everyone can identify some abasimano ancestors. Many people declared that they 
were "pure" Nata but when I questioned them more closely they would tell me stories of a 
grandmother or a great-grandfather coming from another place. The structure of genealogies 
completely erases stranger origins by incorporating them into existing lineages. 

People not only tolerate and incorporate strangers but also value them highly. The life 
histories of elders today provide evidence for inferences about strangers in the past." Although 
most elders contracted their marriages within the immediately surrounding localities, some took 
stranger wives because of friendships between their fathers and men of other localities. Other 
women fled their homes, sometimes with young children, exiled because of pregnancy before 
circumcision or witchcraft accusations. Nata men sought stranger wives because their children 

66 Each informant that I interviewed was also asked about their own life history. This 
information generalizes from many of those histories and from specific conversations particularly 
with the Magoto family on strangers. Informal discussions in Ikoma on stranger wives confirmed 
these ideas. 

would then inherit from their father rather than their maternal uncle. 67 In Nata the children of an 
omosimano wife inherit equally with their paternal uncles at the death of their grandfather. A 
stranger wife carries on the homestead of her husband, which his brothers do not inherit. Neither 
do they inherit his widow, as is normally the case. Stranger wives also represent important in-law 
connections outside the community. These are useful on trips, in trade and to gain support in 
political conflicts. Nata respect and fear an omosimano wife for her outside connections and 
strong internal power at inheritance. The liabilities of marrying an outsider are that she may be 
culturally and linguistically inept and cause embarrassment to the family. Witchcraft accusations 
most often fall to the stranger wife. 

Another common way in which the community incorporated abasimano. particularly 
during the period of disasters was as abagore or "people who were bought." This was an 
important mechanism for coping with famine in the past. Droughts were often local and when a 
family ran out of food their only option might be to take a child to a neighboring group where they 
had connections and leave the child in exchange for food. If the child was a girl, the food would be 
considered as bridewealth, if a boy, as sale. These children were not treated as slaves but as 
members of the family and incorporated as other abasimano children. Chief Megasa bought 
Rotegenga, who later succeeded him as chief, from Simbete parents during a famine, yet few 
questioned his ability to represent Nata because of his origins. 

The mechanisms for the incorporation of strangers worked a bit differently in Ngoreme. 
reflecting the diverse histories of the region. The Ngoreme do not use the term abasimano at all or, 
if used, it refers to slaves. The term for people of "pure blood" (whose parents were both 
"Ngoreme"). was kicheneni. in contrast to the Nata and Ikoma emphasis on birth place. Certain 

' Also discussed by Huber, Marriage and Family , pp. 95-96. 

Ngoreme rituals of the lineage require a native kicheneni only, whereas elsewhere in western 
Serengeti children of strangers may participate in these rituals. The Zanaki, Kuria and Lakes 
peoples have a specific word for slave (ommeese/abaseese) that is also used for these strangers. 
The word literally means "dog" and suggests a very different treatment of strangers than in Ikoma 
and Nata. They also commonly used the term for "someone who is bought."*" The reasons for this 
pronounced difference in attitude may be a result of higher population densities among the Lakes 
and highland peoples and thus a greater need to control wealth within the lineage. It may also be a 
result of closer interaction with the coastal caravan slave trade that operated around the lake from 
Buganda through Ukerewe to the ports in Sukuma. 

Many lineage or clan histories base their narratives on the arrival of a stranger and his 
incorporation to form a new section of the territory. Western Serengeti people valued strangers in 
the homestead as wives and sons and also honored them as great and powerful ancestors. This was 
part of the strategy of wealthy men to incorporate many people as his dependents. Among the 
important prophets of the past, Gitaraga of the Nata was a stranger, who arrived as a child with the 
implements of rainmaking in his hands. A man without children adopted him so that the lineage 
would not die. 69 Elders say that the woman who makes the medicine bundle of the bees at Riyara 
(Materera) must remain an omosimbe (an independent woman) but take a stranger omotware (male 
wife). The spirit propitiated at Nyichoka was a stranger wife. When the community needed to 
consult a prophet they often went far away to find one who was efficacious. The relationship of the 

68 Ngoreme Dictionary, Iramba Parish, n.d. Interviews with Zabron Kisubundo 
Nyamamera and Makang'a Magigi, Bisarye, 9 November 1995 (Zanaki cf); Sarya Nyamuhandi 
and Makanda Magige, Bumangi, 10 November 1995 (Zanaki tf); Daudi Katama Maseme and 
Samueli Buguna Katama, Bwai, 1 1 November 1995 (Ruri cf); Elfaresi Wambura Nyetonga. 
Kemgesi, 20 September 1995 (Ngoreme cf); Bhoke Wambura (Ngoreme ?) and Atanasi Kebure 
Wambura (Ngoreme <f), Maburi, 7 October 1995. 

69 Interview with Mahiti Kwiro, Mchang'oro, 19 January 1996 (Nata <f). 


Tatoga prophets to Ikoma is one in which strangers have become "parents" with ritual authority in 
some of the most important Ikoma ceremonies to maintain the health of the land. In each of these 
situations, the incorporation of the power of strangers was considered efficacious to the health of 
the local community. 

This ethnographic and linguistic evidence, although not conclusive, suggests that there is 
continuity, at least from the nineteenth century to the present, in the organization of settlements on 
the basis of a lineage idiom. Yet what brought people together in these diverse settlements was the 
patronage networks of "big men" and mechanisms for incorporating strangers. Some of the key 
terms from which I reconstruct these patterns such as oruberi (settlement), omwame (wealthy 
man), omosimano (stranger) can be argued on the basis of the comparative method to have been 
innovations by Mara speakers in the last 500 years. Recent ethnography provides models for how 
these institutions might have functioned in the past. The evidence from both these sources is 
consistent with the oral traditions concerning individual settlement sites dating to the period prior to 
the disasters. 

Continuity and Relationship to the Land in the Context of Mobility 

The last piece of the puzzle concerning nineteenth century settlement and the relationship 
of people to the land is how the patterns just described fit into the context of settlement mobility. 
In the last chapter I argued that western Serengeti people adapted to a marginal environment by 
moving their farming settlements fairly frequently over distances travelled in a day or two. If, as I 
argue in this chapter, people maintained their relationship to the land through the ancestors located 
at particular places, then was it possible for people to move into lands not controlled by the spirits 
of their ancestors? If people moved as individuals or in small family groups the mechanisms of 
incorporation described above could easily accomodate their assimilation into a new settlement. 
However, there are also examples, both in oral tradition and in recent times, of larger groups of 

people moving into new lands to start their own settlements. The Ikoma story of the Machaba 
erisambwa demonstrates how mobile spirits solved this problem. Yet many other cases exist where 
people resolved the problem by accommodating the new spirits of the land, establishing their own 
spirits or moving back to older sites. 

People may have moved frequently but their relationship to the land constrained their 
movements. Those whose ancestors inhabited the land as emisambwa ritually controlled it. A 
family who was not living in an area in which their lineage was responsible for propitiation of the 
spirits had to establish reciprocal relationships with those who "owned the land." Good reasons 
existed for doing this but people were ever mindful of returning to the places where they had a 
connection to the land. Philip Mayer argued that among the Gusii, "the lineage attracts the return 
of its own members because of its association with patrimony, protection, and the influence of 
ancestor spirits."™ People did not take lightly permanent migration to new areas, which meant the 
establishment of new emisambwa to protect the land and its people. It was only under spiritual 
direction by ritual means that people were willing to undertake migration to the land of others. 

Two recent example of larger groups of people moving and establishing themselves in new 
land provide possible models for these patterns in the nineteenth century. These stories also 
illustrate continuity in ideas about leadership and the relationship to the land that I have discussed 
in relation to the mid-nineteenth century. Some have argued that because local societies lost 
political control with colonialism they also lost the ability to generate new innovations on these old 
principles of social action. 71 Yet these examples suggest that incredible continuity remains despite 
an utterly changed historical context. 

70 Mayer, The Lineage Principle , p. 3 1 

71 Vansina, Paths in the Rainforest, pp. 245-248. 


The first case is of the Nata patriarch Magoto who moved to Ikizu in 1932 and back to 
Nata in 1964. The second is the move of Kuria Nyabasi from North Mara into the Mugumu area 
of the western Serengeti in the 1950s. Although these are both cases taken from the colonial period 
they seem to reflect the same concerns about land and settlement discussed in relation to oral 
traditions about the nineteenth century, before the disasters. These cases illustrate cultural 
continuity in the ways that people have settled and found prosperity on the land. 
Magoto Mosi and the Nata Moves to Mugeta (1932-64) 

During the late 1 920s and early 1 930s sleeping sickness became epidemic in the western 
Serengeti and the colonial officers were concerned about the dramatic decrease in the Nata 
population. They sent out a Tsetse Fly officer to investigate the causes. Chief Rotegenga was 
adamant that the cause of population decline was death from sleeping sickness but the officer 
began to suspect otherwise. He concluded that something else had driven the people to move out of 
Nata, which had in turn encouraged the return of bush and attracted the tsetse fly. He left without 
ever solving the problem of why so many Nata decided to leave. 72 

Back in Nata 1 heard the other side of the story. During this time (1 920's- early 30's) the 
Bongirate age-set cycle, of whom Magoto Mosi was a part, felt oppressed by Chief Rotegenga (of 
the Busaai cycle) and had many conflicts with him. Rotegenga had already forced Magoto to leave 
Nata earlier, after having openly defied the authority of the Chief. In addition this was a time of 
famine. Magoto Mosi. known as an omukina, or a speaker, called together the Bongirate of Nata 
and any others not happy with the political situation. He convinced them that it was time to move. 
They secured land and permission from the neighboring Ikizu chief, Makongoro, and moved to 

72 H.G. Caldwell, "Report on Sleeping Sickness in Musoma District, July and August 
1932,"pp. 1-7, 215/463, TNA. 


A friend and fellow labor migrant in Nairobi introduced Magoto and the Nata delegation to 
Chief Makongoro. Makongoro heard their request and promised to make Magoto mwanangwa 
(headman) of the new area at Mugeta if Magoto could get lots of Nata to move to Ikizu. Chief 
Rotigenga prevented the first people to move in 1932 from taking out their livestock until Magoto 
and his men went to Musoma and got the District Commissioner to intervene. Both chiefs were 
competing for the right to claim these people as their own. Most people came to Mugeta between 
1 933 and 1 934, when Magoto was mwanangwa (headman). When the Nata heard of his position, 
even more moved to Mugeta, including some Ishenyi and Ngoreme, who came seeking sanctuary. 
Estimates of the number of people to move to Mugeta in the 1930s vary from 200 to 400 people, 
along with their livestock. By 1938 Magoto had too large a herd of livestock to stay in Mugeta. 
He decided to move again, out on the plains, near to an Ishenyi friend. Magoto moved two more 
times, following his growing herd and trade with the Tatoga. 

Finally, in the 1 956 Magoto decided it was time to move back to Nata. Magoto's son, 

Nyamaganda, in a biography of his father, recreated the speech his father gave in Mugeta to the 

gathered Nata men: 

The time has come to return our youth to their home in Nata, we elders are getting old, 
some of our sons have children and even grandchildren. If we die first who will show 
them the place where they were born and the names of the places where we have built 
and the places where our fathers lived? The thing that we came here to get. God has 
helped us to get in abundance, that is cattle, goals and sheep, and further. He has 
blessed us with people. All of this wealth we must return to our homes in Nata" 

The move happened slowly and it was not until 1 963 that Magoto established the village of Mbiso 

on the crossroads of the Musoma/Arusha/Mugumu road. Mbiso was a good place to live because 

73 Mwalimu Nyamaganda Magoto Mosi, "Historia ya Mzee Magoto Mossi Magoto Katika 
Maisha Yake," unpublished manuscript, Natta, 1996. Other information about Magoto was 
collected by talking to and living with his family in Nata. Formal interview with Nyamaganda 
Magoto, Bugerera, 3 March 1995, and Faini Magoto, Mbiso, 6 March 1995 (Nata d-). 


of the respect that Magoto carried and people began to move there. When the moves of "Ujamaa" 
came in 1977 people from outlying areas congregated at Mbiso. 

Magoto only had formal authority during the few years he served as mwancmgwa in 
Mugeta. Yet in his informal capacity as a "speaker" he was responsible for a large migration of 
people to Mugeta and back, following him because of his charismatic leadership. People 
considered him a man of wisdom and when he spoke they listened. His power was based on an 
extensive network of reciprocal relationships beyond that of lineage and clan. Magoto had been an 
orphan who went to live with his mother's people as a young man. This entailed changing age-set 
cycle as well as clan. He had friendships developed during his youth when he traded in Sukuma, 
collected arrow poison in Kuria and did migrant labor in Kenya. As he grew wealthy in cattle, he 
used these cattle to help the sons of his friends to go to school or to begin their own herds. These 
informal ties of patronage made Magoto a trusted person. Issues of politics and power conditioned 
mobility. Nevertheless, in the end Magoto had to bring his children back where they had 
connections to the ancestors and to the land. 

Magoto represents a combination of old and new ideas about authority and leadership. His 
achievement of wealth through livestock is a phenomenon of the early colonial period. The formal 
title of headman is also a colonial element. However, the methods of historical linguistics date the 
informal leadership of the "speaker" back to the time of early Lakes Bantu speakers. 74 He later 
gained the power of "medicines" when he was initiated into the highest eldership title of the Nata as 
an omorokingi. His relationship to the land within a settlement pattern of mobility is congruent 
with the evidence presented in this chapter for the nineteenth century. Magoto eventually moved 

74 Schoenbrun, A Green Place , pp. 1 99-200, demonstrates the connection between the 
power of speech and healing or divination. As is evident in the example of Magoto the power of 
speech was linked to both political and prophetic roles. 

back to Nata, but many other leaders did not move back and had to establish relations to the land 
in a new place. The next story provides a present day model for those who did not return to land 
they once knew that is congruent with evidence about nineteenth century settlement patterns. 
Establishing a New Land: The Kuria Move to Mugumu (1956-61) 

Some Kuria families from the Nyabasi clan territory in North Mara moved south over the 
escarpment across the Mara River beginning in 1 956 to establish themselves close to what is now 
the town of Mugumu and capital of Serengeti District. During this time people in the fertile 
highlands of the escarpment were beginning to suffer from land shortage and younger sons had to 
find new land for their cattle and fields. Yet Nyabasi elders do not cite these material factors as the 
reason for the move. 

The testimonies of Nyabasi elders agree that they came to Mugumu because their prophet 
Gesogwe prophesied the move near the end of the German period. They were not able to go at that 
time because Maasai raids made the area too dangerous. The prophecy said that they would keep 
moving until they reach the mountain Gaoga, in the northern extension of the Serengeti National 
Park. There they would encounter the Maasai and end their expansion. Although the park has 
precluded this goal, the Kuria today keep pressure along the whole northwestern boundary of the 

The origin story of the Nyabasi says that they once lived at Ikorongo in South Mara, now 
Ngoreme. Their name relates them to the Asi, hunter/gatherers who figured in the Nata and other 
emergence stories told in Chapter 4 and 5. One Nyabasi elder said that their ancestor was an Asi 
hunter, who came to North Mara and traded arrow poison for cattle until he became rich and 

s Interviews with Kisenda Mwita and Hezekia Sarya, Matare, 15 March 1996 (Kuria <?). 

founded the Nyabasi clan. 76 The Ishenyi emergence story says that when the Ishenyi dispersed 
throughout the region some moved to Kuria and became the Iregi, now known as the Kuria 
territory of Bwiregi which is a neighbor to Nyabasi. Clearly, important links existed between 
South and North Mara before the formation of Kuria ethnicity. 

Nyabasi elders say that the Mugumu area was then open space, the hunting area of the 
Ikoma, Nata and Ngoreme. All three chiefs gave them permission to settle and divided the area so 
that each could benefit from tax revenue. The Kuria immigrants suspected that the western 
Serengeti peoples welcomed them as an eastern shield against Maasai raids. People did not live in 
the Mugumu area then, in part, because it was a corridor for raids from Loliondo. The colonial 
government had already begun to settle Maragoli and Luo immigrants from Kenya in this area as 
well. The Kuria intermarried with the Maragoli until they returned to Kenya at "Ujamaa." 

The dream prophet (omoroli) Mbota decided it was time to fulfill the prophecy of Gesogwe 
and led the Nyabasi immigrants to Mugumu. 77 He had no power to order people to go, but as with 
Magoto, where he went, people followed. It took one or two days to move everything from 
Nyabasi to Mugumu, sleeping one night in Ngoreme. The head of the homestead and some young 
men preceded the others by at least a year to prepare a place and harvest a crop. It took two years 
to complete the move, with the cattle coming last. The advance group found friends already in the 
area to help them choose a place to settle. After they had chosen a spot, the immigrants marked the 
trees around the perimeter of the area as an indication of possession. After one family settled, their 

76 Interview with Sira Masiyora, Nyerero, 17 November 1995 (Kuria cf). 

77 Kjerland, "Cattle Breed,"p. 140, says, "the seers instructed people how to move and told 
them where to go and when." In her questionnaire on moves back to Kebaroti-from the Zebra 
people-70 out of 74 in survey claimed their parents moved according to the words of the seers. 
The four who didn't mention it were youngest respondents. Twenty-four named seers were listed, 
according to each lineage. 


lineage, clan or dependents tended to settle around them on the basis of personal patronage. Today 
people use clan or lineage designations to identify all of the Kuria villages around Mugumu. 78 

The Nyabasi settled in Mugumu do not intend to go back to Kuria country in North Mara. 
No land is available for settlement there and they have become ritually independent from North 
Mara. Kuria elders said they did not have emisambwa located in a place but depended on their 
prophets. When they arrived in Mugumu, the prophet gave them medicine to spread around the 
boundaries to make the land good. They also sacrificed an animal. Among the Kuria it is the 
secret council of the injama who are responsible for the land. The new injama formed in Mugumu 
carried some of the old secrets from Nyabasi but now has new secrets for the new land. The 
boundaries that the new injama established with the medicines of the prophet are secret and known 
only to them, they do not correspond with tax boundaries." The Kuria have established possession 
of a new land through the moral prescription of the prophet and ritual control of the land. 80 


The Kuria case demonstrates the ongoing potency of ideas about settlement and possession 
of the land. The relationship between particular places, their spirits and certain peoples undergoes 
a constant reconfiguration in a context of mobility. People know that their relationship to the land 
is good when they prosper and grow. People discard relationships that lose their efficacy for ones 
that are successful. People gather around "big men" who bring prosperity by their wisdom and 

78 Interview with Kisenda Mwita, Hezekia Sarya, Matare, 15 March 1996 (Kuria <f). 

79 Interview with Philemon Mbota, Matare, 27 January 1995 (Kuria <f). Philemon, now a 
Mennonite Church pastor, is the son of the prophet Mbota who brought the people from Nyaba; 

80 This is a highly controversial and politicized topic in Serengeti district today, the 
Ngoreme, Ikoma and Nata have a very different interpretation of the Kuria migrations. 


leadership. Good crops, expanding herds and many people are signs that the land has blessed the 

The place-names that people remember in oral traditions, either as emisambwa or 
settlement sites, are a result of the last reconfiguration of those relationships to the land, perhaps in 
the mid-nineteenth century. Yet they represent social processes, if not the particular settlements, of 
longer duration. As part of the tribal paradigm introduced in the colonial period it became 
important to establish an ancient history in order to claim legitimacy. One Ngoreme historian, 
who had his account mimeographed by the Catholic mission in Iramba, dated these settlement sites 
to the fourth century A.D. When asked how he arrived at this date he said that it seemed "far 
enough back." 81 New chronologies push back the dates of oral traditions that were intended for 
other purposes. These places are important not because they establish the most ancient claims to 
the land but because they establish connections to the ancestors who still control the land and thus 
the health of its people. 

To prosper on the land people must maintain a right relationship with it, through ritual and 

memory. The peoples of the western Serengeti continue to recite these place-names because 

without the identification of people and land they cannot prosper there. They possess the land by 

the propitiation of ancestors who have become synonymous with the land as emisambwa. They 

remember the "big men" with their extensive networks of people and the places where they lived to 

connect themselves with that prosperity. It is a form of patronage in which the patron is dead but 

his spirit keeps "feeding" the people. Because of the importance of this identification of land and 

people, oral traditions pass on these place-names to the next generation. As Magoto said: 

If we die first, who will show them (our sons) the place where they were born and the 
names of the places where we have built and the places where our fathers lived? 

" Interview with Philipo Haimati, Iramba, 15 September 1995 (Ngoreme <?). 


These sites, each with their own history "read" on the landscape, represent all of the 
various groups of people that the emerging ethnic identities of the late nineteenth century. The 
historian better understands these lists of place-names as historically important places in the local 
politics of ritual and power rather than sequential settlements in a "tribal" migration history. 
Narrators created a unified ethnic history by joining the histories of unlike units of lineages, clans, 
generation and age-sets and identification with powerful prophets or "big men." In the early 
colonial period the pointillist history of settlements became a territorial history of the "tribe." The 
land of many smaller ebyaro (territories) became the administrative boundaries of colonial tribes. 
By coopting the identification of places and thus the histories that they represent, oral traditions 
helped to create ethnic identities out of an amalgam of other kinds of identities 

The pointillism of settlement sites does not necessarily mean that they were isolated from 
each other. Social identity was multiple and situational. One person may have lived in a 
settlement of his patrilineage, but visited the prophet of another settlement for advice, asked for 
help from a "big man" of another and shared a generation or age-set name with people in 
settlements all across the region. Clan groupings of settlements into the ekyaro (territory), 
responsible for maintenance of the land, participated in networks of alliance with peoples of the 
same clan name all across the region. From the traditions of the "floating gap" period nothing 
indicates that these larger regional connections necessarily corresponded with ethnic identity. One 
can imagine a map not of fixed ethnic blocks or territories but of smaller communities, dependent 
upon an identification with the land, whose connections radiated out in complex networks of 
affiliation and identity all across the region and beyond. 

The next chapter adds further to the understanding of the relationship between land and 
people by exploring the territorial unit which linked individual settlements into one ekyaro. The 
ekyaro was formally clan-based but it was the generation-set that performed the rituals necessary 

to "cool the land" and provide for its internal security. While the form of social identity explored in 
this chapter was that of the lineage, that of the generation-set and its historical development in the 
region is explored in the next chapter. 




This chapter looks at testimonies concerning rituals rarely performed anymore by the 
generation-set to protect the land and its people. An analysis of these rituals demonstrates that 
people defined themselves territorially even where the land encompassed by that territory was not a 
fixed unit. Each time the generation-set performed these rituals it defined anew the territorial 
identity of a people united by their ritual relationship to the land. These findings challenge 
historians and anthropologists to rethink their analyses of precolonial African society based on 
either the assumption of a lack of territorial identity or the definition of territory in terms of a 
kinship group. 

While the last chapter explored oral traditions about particular lineage-based settlements in 
the mid-nineteenth century, this chapter examines the space of the larger clan territory, or the 
ekyaro, that encompassed a number of settlements in one area during the same period. As with 
emergence stories and migration traditions, these rituals in Ikizu, Nata, Ishenyi and Ikoma have 
become the rituals of the ethnic group and legitimate present claims to a territory. However, they 
also represent much older concepts about the relationship of land and people that suggest further 
how communities formed the territorial identity that allowed for internal cohesion in a context of 
expansion and mobility in the mid-nineteenth century before the disasters. 

Nineteenth century western Serengeti peoples used the homestead images of houses and 
gateways to organize themselves into settlements and territories but they united these disparate 
peoples through an egalitarian and inclusive relationship to the land. Although individual lineages 


carried out the sacrifices to ancestral spirits of the land, the generation-set representing all mature 
men regardless of lineage or clan membership performed the most important rituals to "cool the 
land." Because people across the Mara Region used the same system of generation-set names they 
could move into new clan territories and participate as equals with their own generation-set. The 
generation-set fostered the community consensus necessary for the health of the land and its people. 

The generation-set, or rikora in power maintained the relationship to the land, or the 
ekyaro, by the ritual of walking over the land (kukerera) every eight years and spreading the 
medicine of protection and rain, assuring fertility and security, or "cooling the land." As 1 
discussed in the last chapter, a right relationship to the land and the ancestors who inhabited it was 
necessary for prosperity and growth. The rikora was responsible for the health of the land and 
identified with the land. Many elders made this connection clear by asserting that because the 
rikora no longer walked, the land was ruined. The well being of the land was synonymous with the 
collective well being of its people. 

This chapter probes the social identity of the generation-set and its ritual relationship to the 
land over time. It postulates that the words for the unique type of generation-set organization 
found here dates back to 100 B.C. - 400 A.D. when East Nyanza Bantu-speakers adapted to the 
drier lands of the interior by learning from agro-pastoral Southern Nilotic-speakers. The institution 
of the generation-set provided an ideal device for creating community consensus around the 
inclusive territorial principle of a relationship to the land rather than around the exclusive principle 
of lineage. 

Although the ritual practices of the generation-set seem to be fairly stable, the territory 
defined by this ritual walk has varied over time. For example, the territorial unit of communal 
identity enclosed by the ritual walk shifted from that of the clan to the age-set during the period of 
the disasters and then to the ethnic group during the colonial period. The core images of these 

rituals, including enclosure, binding, covering and mediation of outside forces remain the same 
despite incredible geographical variation in ritual practices. 1 trace the shifting and contested 
meaning of ekyaro territory established by the seemingly timeless ritual practice of "walking the 
land" (kukerera). The rituals described in this chapter represent the unification of clans into one 
age-set cycle or one ethnic group. 1 argue that generation-set rituals used the same symbols before 
the disasters to unify diverse lineages into one clan territory. 

Although social hierarchy did exist-in the form of inequalities in wealth, gender, seniority, 
expertise, lineage or clan membership-the generation-set leveled these differences, at least between 
men of one generation. The rituals to "cool the land" reinforced the communal authority of the 
generation-set over other forms of emergent political authority such as "big men," rainmakers, 
prophets or lineage elders. By promoting the "youth" as the "generation-set in power" the elders 
masked their own authority in controlling the actions of the youth. The "generation-set in power" 
was the visible hand to carry out the will of the elders who directed them from beyond the public 
gaze. By this device, the elders fostered community consensus by silencing discussion on the basic 
inequalities of seniority and gender. These rituals represented and reinforced the authority of 
elders, based on the principle of generational seniority, as the form of authority responsible for 
protecting and healing the land and its people. 

Finally, I use the generation-set rituals representing the core images of enclosure and the 
mediation of outside forces as the basis for looking at the concept of territory and territorial 
identity over time. Although 1 have already shown that western Serengeti people had various ways 
of conceptualizing the spatial relations of their various social identities, they employed the core 
spatial images of enclosure to define an identity with the land independent of lineage ideologies. 
Historical evidence suggests that people used these concepts of enclosure to define territories 
containing different social units in different time periods. 1 conclude this chapter by looking at how 

western Serengeti narrators employ these concepts of territory to define newer kinds of ethnic and 
national boundaries and the identities that they enclose. The generation-set rituals best express 
these common understandings about boundaries and an identity linked to the land that operated 
across the Mara Region. 

Generation-Set Ritual and the Middle-Time Frame 

1 discuss the generation-set rituals of "cooling the land" within the middle time frame of 
indigenous history. One might argue that we should treat the generation-set rituals as 
representative of the longue duree, because the generational principles of growth and healing seem 
to be quite ancient. However, although the generation-set system itself is very old in the region, the 
rituals themselves cannot be dated. Oral testimony recounts the transfer of many generation-set 
functions and ritual to the age-set during the period of disasters. These testimonies indicate that 
before the disasters the generation-set alone was responsible for the rituals to protect the land and 
its people. The unit of land around which the generation-set walks is the ekyaro, but the definition 
of ekyaro changed in the late nineteenth century from the territory of a clan to that of an age-set. 
Perhaps before the nineteenth century the generation-set walk encompassed a different social unit. 
Thus, without more evidence, I can only convincingly discuss these rituals in relation to the period 
immediately preceeding the late nineteenth century disasters. 

1 also treat the generation-set rituals in the middle time frame (or "floating gap") of 
indigenous history because that is where elders themselves place these narratives. Western 
Serengeti elders say that the generation-set rituals are very old but that they began after the 
descendants of the first parents had multiplied to the point that they needed separate territories and 
before the period of disasters. Few local traditions describe the origin or development of 
generation-sets. One informant said that the ancestors formed the generation-sets to unite all the 

clans in one land. 1 Kuria testimonies stated that the Saai and Chuuma generation-set cycles were 
named after the wives of Mukuria, their ethnic group founder. 2 Kuria in the province of 
Nyamongo call two twin hills the "hill of Chuuma" and the "hill of Saai," suggesting a territorial 
origin for the two groups. 3 These cryptic stories, like the place-name lists in the last chapter, 
indicate that elders understand them as belonging to the middle period between the emergence 
stories and the historical recollections of the disasters. 

In elder's testimonies, the generation-set and its ritual represent the concern for "repetitive 
social processes." as Spear described traditions of the middle period. 4 In Braudel's terms the 
middle period concentrates on the slow but perceptible rhythms of social time. 5 By setting the 
discourse on territory and boundaries squarely in this middle time frame of history, western 
Serengeti narrators are putting it out of the reach of overt political debate. Claims to land, 
territorial identity and the social authority of elders presented in the context of the middle period 
become unquestioned "tradition." 

However, I argue that these rituals are more than simply an "invention of tradition" to 
legitimate the present social order, or, in this case, specific rights to a territory. Scholars from 
many disciplines understand ritual as a symbolic text whose meaning they decode. Social 
historians and historical anthropologists have analyzed rituals by looking at their changing meaning 

1 Interview with Kirigiti Ng'orogita 8 June 1995 (Nata <f). This man is the last surviving 
leader from the Saai generation cycle who knows the rituals. 

2 Abuso, A Traditional History pp. 1 6- 1 7. This is because the two cycles of the 
generation-set (Saai and Chuuma) are often refered to using the prefix Mwanya- indicating the 
houses of two wives. 

3 Ruel, "Kuria Generation Classes," p. 20. 

4 Spear, "Whose History?," pp. 165-181; and Spear, Kenya's Past , p. 47. 
' Braudel, The Mediterranean , p. xiv. 

within particular historical contexts." Sociologists and anthropologists have seen ritual as a means 
of communicating shared values and dealing with internal conflict. 7 All try to interpret the 
meaning of rituals beyond their alleged purpose. While each of these approaches provides insight 
that contributes to the following analysis, they all deny the claim by participants that these rituals 
commemorate past events rather than present structures. 

Yet Connerton argues that "if there is such a thing as social memory, we are likely to find 
it in commemorative ceremonies," which he defines as rituals that ostensibly reenact the past. 8 
Through an analysis of ritual language and gesture in Europe he shows that the structure of ritual 
in "commemorative ceremonies" builds in a certain invariance because of the performative, 
formalized, and stylized language on which the reenactment depends. Those who perform rituals 
do so as members of a group that habituates them to certain bodily practices reserved for ritual, 
passed down with little variance from the past. While anyone can narrate oral traditions, only 
members of the group can perform the rituals. The positions and gestures of the body in 
performance form the mnemonic system of core spatial images around which rituals are elaborated. 
Performers understand these actions as reenactments of past prototypical actions. Ritual suspends 
linear time and reconnects people with their past by reenacting the past. The ritual performance of 

The literature on ritual is vast and sophisticated. See M. Bloch, From Blessing to 
Violence: History and Ideology in the C ircumcision Ritual of the Merina of Madagascar 
(Cambridge: Cambridge University Press, 1986); and Newbury, Kings and Clans . 

7 E. Durkheim, Elementary Forms of Relipinns life (London: Allen and Unwin, 1 954); 
Victor Turner, The Forest of Symbols: Asp ects of Ndemhu Ritual (Ithaca: Cornell University 
Press, 1967); Max Gluckman, Rituals of Rebe llion in South-F.ast Afri™ (Manchester: Manchester 
University Press, 1954). 

8 Connerton, How Societies Remember p. 71 . 

the "commemorative ceremony" conveys and sustains "an image of the past" through which a 
community understands its identity.' 

In western Serengeti ritual, the definition of the land encircled by the generation-set. the 
ekyaro, changed according to the historical context. However, the relationship between people and 
land embodied in ritual remained stable as the core spatial images, analogous to that of oral 
tradition, but inscribed in bodily practice. These core spatial images include those of the enclosure, 
binding, covering and mediation of outside forces, all of which serve to define the boundaries of 
social identity and group cohesiveness within a territory. 

We know that the core images of enclosure ritual are old because they are found in 
generation-set rituals throughout the Mara Region and cannot be traced to recent innovations. The 
language of generation-sets reconstructed through historical linguistics is demonstrably old. The 
variations in each ethnic group suggest that each group elaborated a given set of rituals from an 
older pattern. If western Serengeti people adopted these rituals in the last century we would expect 
them to reflect relations between generations at that time. As 1 demonstrate in Chapter 10, during 
the early colonial years young men gained autonomy from the authority of their elders by 
accumulating their own cattle wealth through hunting, trading, and raiding. The new generation of 
wealthy men established their own networks of patronage outside of the channels controlled by 
elders. The historical context of class differentiation and the autonomous authority of young men 
is not represented in the principles of egalitarian responsibility to the land and the authority of 
elders over juniors embodied in the rituals of the generation-set. Although these images seem to be 

9 Ibid, p. 70, 41-71. See also Renee L. Tantala, "Verbal and Visual Imagery in Kitara 
(Western Uganda): Interpreting 'the Story of Isimbwa and Nyinamwiru,"' in Paths Toward the 
Past: African Historical Essays in Ho nor of Jan Vansina . eds. W. Robert Harms, Joseph C. Miller, 
David S. Newbury, and Michelle D. Wagner (Atlanta: African Studies Association Press 1994) 
pp. 223-243. 

quite old, in this chapter I reconstruct the meaning of generation-set rituals in the period right 
before the disasters because narratives about the rituals refer to this period and because it is 
impossible to know how these rituals might have functioned before this period. 


Because the generation-set carries the memory of concepts concerning territorial identity in 
the bodily practice of ritual, we must first understand the ways in which people have used the 
social logic of "age" and "generation" to organize social relationships before turning to their ritual 
function. In the anthropological literature the generic terms "age-organization" or "age-system" 
describe social organization based on age, generation or both. The term "age-grade" refers to the 
nearly universal tendency toward peer grouping, while "age-set" or "generation-set" is only used 
where persons are grouped into hierarchically ordered sets with specific social responsibilities as a 
unit. Some anthropologists use the term "class" interchangeably with "set." "Age-set" recruitment 
is based on age at initiation, while "generation-set" recruitment is determined at birth by the father's 
set. In an age-set system, a boy and his uncle could be in the same set if they were the same age, 
while this is impossible in a generation-set system. "Age or generation-set cycles" are systems in 
which a cycle of successive names is assigned to each group as it is formed over time. 10 Peoples in 
the western Serengeti have used the logic of organizing social relationships on the basis of both age 
and generation, although historically the relative importance and function of each have varied. 

The western Serengeti rikora (generation-set) and its rituals belong mainly in the male 
domain. Although women seldom undergo circumcision anymore, in the past women joined an 
age-set at circumcision and had their own initiation names, different from the boys. When they 
married, they became part of the age-set of their husband. Before they married, young women 

10 P. T. W. Baxter and Uri Almagor, "Introduction," Age. Generation and Time: Some 
Features of East African Age Organisation (New York: St. Martin's Press, 1978), pp. 1-2. 

enjoyed public dances with their age-set peers. Women acquired a generation-set name from their 
fathers at birth but did not participate in most of the generation-set rituals or leadership, nor did 
they maintain a parallel age structure. A woman's generation-set determined legitimate marriage 
partners-with someone of her own or an alternate generation, never with someone of her father's or 
children's generation." Despite the absence of women in these rituals, the generational principles 
expressed in the rituals draw on the female symbols of inside space, the enclosed womb, water and 
fertility.' 2 
The History of Age- and Generation-Sets in the Mara Region 

The Mara Region is somewhat unique in East Africa since social relations based both on 
age and generation can be identified and the region does not conform to the stereotypical 
parameters of societies that emphasize age-organization. Anthropologists in Africa have most 
often studied age-organization in relation to Nilotic- or Cushitic- speaking pastoral peoples. Mara 
peoples are not only Bantu-speaking but agriculturalists, surrounded by peoples who have no 
generation-set systems, cyclical or otherwise. Using a model of diffusion, Bischofberger was at a 
loss to explain how the Mara peoples acquired this system. I3 Mara age organization also 
confounded the typological classification of early anthropologists. P. H. Gulliver attempted to 
classify age organization into three types, 1 ) those like the Maasai and Sonjo with a linear age- 
based system, 2) those like the Nandi, Kipsigis and Luyia with a cycling age-based system, and 3) 
those like the Jie, Pokot, Tatoga, and Kikuyu with a non-cycling generation-based system. The 
Mara Region had a linear age-based system, like the Maasai ( I ), that is subordinate to a 

" Interview with Mwenge Elizabeth Magoto, Mbiso, 6 May 1995 (Nata 9). 

12 For an analysis of the symbols of wombs and enclosures see Boddy, Wombs and Alien 
Spirits , pp. 47-8 1 . 

11 Bischofberger, The Generation Classes , pp. 99-102. 

generation-based system, like the Jie (3), but using many of the same eight cycling age-set names 
oftheNandi(2). H 

These earlier typologies and characterizations treat age-sets and generation-sets as if they 
were discrete models that were uniformly adopted in different times and places. The examples 
from Mara show that age and generation are principles of social logic that people applied in 
different ways in different historical contexts according to their needs. The system of age- and 
generation-sets in the Mara Region developed not by a process of diffusion from outside sources 
but by drawing on a common substratum of inherited generative principles from both Great Lakes 
Bantu and Southern Nilotic societies. Mara peoples used those principles in a variety of historical 
contexts to create institutions that met their needs. In the context of the late nineteenth century 
disasters western Serengeti people used these old principles of age-organization to create unity and 
cohesion in the midst of chaos. 

The cycling generation-set names used throughout the Mara Region have a long history. 
As I noted in Chapter 4, the evidence of loan words in East Nyanza Bantu languages concerning 
livestock, non-kin relations and the homestead from Mara Southern Nilotic languages suggests that 
East Nyanza-speakers, moving into the unfamiliar environment of the interior, used common age- 
sets and the comradeship of peers to improve their livestock expertise and to develop new kinds of 

14 P. H. Gulliver, "The Age-Set Organization of the Jie Tribe," Journal of the Roval 
Anthropological Institute o f Great Britain and Ireland. 83 (1953), pp. 147-168 and Bischofberger, 
The Generation Classes, pp. 75-81. For other analyses of age-set organization in East Africa see,' 
P. H. Gulliver, "Turkana Age Organization," American Anthropologist 60 ( 1 958): 900-922; J. g' 
Peristiany, 'The Age-set System of the Pastoral Pokot: The Sapana Initiation Ceremony." Africa 
21 (1951): 188-206; Robert A. LeVine and Walter H. Sangree, "The Diffusion of Age-Group 
Organization in East Africa: A Controlled Comparison," Africa 32, 2 (April 1 962): 97- 1 1 0; J. J. 
de Wolf, "The Diffusion of Age-Group Organization in East Africa: A Reconsideration," Africa 
50, 3 (1980): 305-310. For age-groups in the Mara Region see, E. C. Baker, "Age-Grades in 
Musoma District, Tanganyika Territory," Man 27 (1927): 221-224; and E. C. Baker, "Age- 
Grades in Musoma District, Tanganyika Territory." Man (April 1953): 64. 

homesteads built around the livestock corral. Sometime before approximately 400 A.D., East 
Nyanza-speakers adopted the cycling age-set names used today as generation-set names, perhaps to 
facilitate their relationships with pastoralists already familiar with the environment.' 5 

The generation-set system that is found all over the Mara Region, including the Lakes 
people who speak Suguti languages and the Kuria who speak North Mara languages, divides the 
eight cycling Southern Nilotic names into two cycles, (Mwanya)Chuuma and (Mwanya)Saai. 
Each of these divisions has a cycle of four names used for each generation in succession. The 
ceremonies for passing on authority to a new generation take place first among the Saai cycle of 
the Saai generation followed by the Chuuma cycle of the Mairabe generation and then back to the 
Saai cycle of the Nyambureti generation and so on. Elders say that it takes 100 years to complete 
the cycle. 

Saai Cycle Chuuma Cycle 

Saa ' Mairabe (Ngorongoro among Kuria and Ghibasa among lkizu) 

Nyambureti Gini 

Gamunyere Nyangi 

Maina Chuuma 

Ehret reconstructed six of these names and their standard order in pre-Southern Nilotic 
speech communities: l)Sae 2)Gorongoro 3) Unknown 4) Gini 5) Unknown 6) Nyangi 7) 
Maina 8) Cuma. Because these six common names are not derivable from Southern Nilotic or 
Bantu words, he postulates that pre-Southern Nilotic speakers adopted the names from Eastern 
Cushitic-speakers with whom they had contact on the Ethiopian borderlands as early as the first 

" Enret > Southern Nilotic History, p. 46, shows that the Kuria generation-set names could 
not have come from later Kalenjin sources because of the sound changes and thus attests them to 
interactions with Mara (Victoria) Southern Nilotic-speakers in the late pre-Southern Nilotic times. 
All East Nyanza languages share the same generation-set names. 

millennium B.C. There, Pre-southern Nilotic-speakers who had age-set systems adopted a cycling 
system along with these names, and also circumcision and clitoridectomy. East Nyanza speakers 
adopted the names of Nyambureti, Mairabe (replacing Ngorongoro) and Gamunyere at a later time. 
Ehret argues that the (Curia did not adopt the cycling names from their Southern Nilotic Kalenjin 
speaking neighbors to the north but more likely from the earlier Mara Southern Nilotic-speakers in 
the region.' 6 A number of groups in East Africa use these cycling names, including Nandi and 

Although East Nyanza Bantu-speakers adopted cycling age-set names from Southern 
Nilotic-speakers, they adapted the system to fit their own needs. The evidence of historical 
linguistics shows that the earliest Great Lakes Bantu speakers (C. 500 B.C.) practiced 
circumcision and initiation. East Nyanza speakers, already practicing some type of age- 
organization would have been receptive to elaborations that linked them to existing pastoral 
communities in the region. In the western Lakes region, age-set institutions were lost with the 
centralization of political and religious authority and the emergence of clientship. Schoenbrun 
postulates that the continued importance of age-organization occurred in those places of the Great 
Lakes Region, such as Mara, where political and religious hierarchies did not develop as a means 
for cutting across lineage-based identities to create alliances." From earliest times, age- 
organization formed the basis for unifying local communities of diverse origins. 

Although Mara peoples now use the eight cycling Southern Nilotic names as generation-set 
names, one cannot tell whether they adopted the names as part of a generation- or an age-set 
system. The word used for the generation-set throughout the Mara Region today is 

16 Ibid, p. 45. 

" Schoenbrun, A Good Place, pp. 1 76-77. "Mara, Luhyia, Forest and Rwenzori branches 
of Great Lake Bantu have maintained these institutions," p. 176. 


rikora/amakora derived from the Bantu root meaning to grow, gokora, implying that (although the 
names of each generation were adopted from Southern Nilotic-speakers) the generation-set system 
itself was a local innovation around the generational principles of growth, fertility and successional 
development." There is no evidence to suggest that the cycling Southern Nilotic names were ever 
used in the Mara Region except in reference to relationships based on generation. 

The word used in the Mara Region for age-set, saiga, is derived from the Dadog word for 
age-set, saigeida, presumably adopted from Tatoga neighbors who practice a non-cycling 
generation-set system." Although western Serengeti people may have adopted this word at any 
time, right up to the nineteenth century, it seems to date to an earlier period since the Gusii, who 
were not in direct contact with the Tatoga in the nineteenth century, also use this word (esaiga) for 
the house of young unmarried men. 20 As I noted in Chapter 4, the fact that the Tatoga dropped the 
Southern Nilotic cycling age-set names, sometime after 1000 A.D., while their Bantu-speaking 
neighbors kept the names seems to indicate a period of differentiation and separation between 
herding and farming communities, where institutions that had unified these communities no longer 
took priority. However, this also seems to be the same time when Bantu-speaking hill farmers 
adopted a linear age-set system, called saiga, in addition to the older cycling generation-set system 
to maintain these useful relationships with herders across ecological zones. 

18 In the Suguti language of Kwaya -kukura, but not in Kuria. Augustino Mokwe Kisigiro, 
"Nata-Swahili Dictionary," n.d., author's collection, The word kukerera, for the walk of the 
generation-set is likely a prepositional form of kukila a proto-Great Lakes Bantu word meaning, 
"to overcome, surpass, heal, unify." The prepositional form implies an object, in this case the land 
and all that it stands for. David L. Schoenbrun, personal communication. 

19 Each circumcision set within the saiga are called siriti. 

20 LeVine and LeVine, "House Design," p. 162. 


From this scarce evidence. I have tentatively concluded that the generation-set system in its 
present form predated the age-set system in its present form and was based on generational 
principles of growth brought by Lakes Bantu-speakers as they moved into the Mara Region. On 
the basis of the words associated with the institutions alone, the generation-set names were adopted 
before the name for the age-set, saiga. Obviously both ways of organizing relationships are quite 
old in the region and both have been used in various ways over the past millennium. 

Western Serengeti peoples rarely practice the rituals of the generation-set any longer, in 
large part, because of events during the colonial period. During the years after the Germans fled 
the region and before the British took administrative control, generation-sets in territories 
throughout the region rose up and overthrew the chiefs appointed by the Germans. When the 
British came in, they reinstated the German chiefs and outlawed the organization of elders known 
to them as the "baGini" (which is a generation-set name) on accusations of witchcraft. 2 ' The 
Musoma District Books, compiled between 1916 and 1927, record under the Native Court at 
Nagusi (Ishenyi) that saiga (or age-set) "is forbidden." 22 The colonial officers were not clear about 
the distinction between age- and generation-sets and may have feared the military connotations of 
age-sets. Government persecution forced age-organization to function underground since its rituals 
were still considered necessary for prosperity on the land. 

1 E. C. Baker, "System of Government, Extracts from a report by R. S. W. Malcolm," 
Sheet 51 , followed by page 2, "Major Coote took the part of the Chiefs and in an attempt to 
improve the administration in 1919, the Bagini were suppressed." General Meeting at Musoma 
2/6/1919. MDB. 

22 Musoma Sub-District 1916-1927, p. 59, MDB, TNA. Note this is a different and 
earlier version of the Musoma District Books than is on micro-film and available in the United 
States through CAMP. This earlier version is only available at the TNA as far as I know. 


Few people living today have participated in generation-set rituals. The Ikoma have not 
performed the generation-set walk since 1 978 and the Ishenyi since 1 950. 23 The next Nata 
generation, Chuuma, has lost all of the elders who know the rituals and so must learn the rituals 
from the Chuuma of Ikizu if they decide to renew the ritual practice. Saai rituals and secrets are 
kept separately from the Chuuma, so that the same generation-set cycle in different ethnic groups 
shares the same rituals. 24 No rikora leaders remain in either cycle today. In Nata, Ikoma and 
Ishenyi the generational names under each cycle have fallen out of use and only a few of the oldest 
men can remember any of them at all. One Nata informant told me that Gamunyere and 
Nyambureti were simply used as names for greeting rather than being generation-set names. 

In Ikizu the Gini, of the Chuuma cycle, was the last rikora to walk, around 1940. Ikizu 
elders say that the rikora does not walk anymore because the last colonial Ikizu chief, Makongoro, 
would not allow it. He gained the chiefship from his uncle Chief Matutu after a controversial 
succession struggle. However, he was not invested with the primary symbol of office, the ndezi, a 
cowry and leather bracelet. Some say that since Makongoro usurped the ulemi by force and was 
not properly installed with the ndezi, he could not allow the rikora to walk. The rikora walk is 
part of blessing the land at the installation of a new mtemiP In the Ikizu emergence story the first 
woman, Nyakinywa, as mtemi rainmaker, had to make compromises to her authority by allowing 
the generation-set to continue. Makongoro would not let the rikora walk because they represented 
a powerful rival to his authority and a threat to his legitimacy. By telling me these stories the Ikizu 

21 Interviews with Mang'ombe Morimi, Issenye Iharara, 26 August 1995 (Ishenyi <f); and 
Moremi Mwikicho, Sagochi Nykipegete, Kenyatta Mosoka, Robanda, 12 July 1995 (Ikoma d% 

24 Interview with Kirigiti Ng'orogita, Mbiso, 8 June 1995 (Nata <f). 

25 Interview with Ikota Mwisagija and Kiyarata Mzumari, Kihumbo, 5 July 1995 (Ikizu 


elders asserted an older communal authority embedded in the generation-set and the control of 
elders over it, standing against the authority of the chief. 
The Historical Priority of Generational Principles 

The variation in age-organization throughout the Mara Region presents a complicated and 
intricate puzzle to fit together. Yet behind this complexity lies a simple logic of generational 
authority and communal unity in relation to the land. Out of these underlying principles the 
variations and combinations of age- and generation-sets developed. 

One of the most difficult problems in understanding Mara age-organization is to untangle 
age- and generation-sets that have a complicated and intertwined history. Overall a regional 
pattern exists of two parallel institutions, age- and generation-sets, in operation simultaneously. 
Young people joined the linear age-set system when they were circumcised at the time of puberty. 
Circumcision ceremonies took place every couple of years in separate ebyaro (territories) and 
several groups of initiates were combined to form one age-set. Each age-set took a unique name 
that referred to an important event or characteristic of the time when they were circumcised. Since 
circumcision ceremonies were highly visible and festive events, news of the chosen age-set names 
spread rapidly and initiates all across the region adopted the same age-set names. 

On the other hand, a person acquired his generation-set at birth according to the 
generation-set of his father. A person's generation-set name followed the sequence in the cycle, 
according to whether his father was from the Chuuma or Saai cycle. A ceremony took place when 
the generation-set that held power was ready to retire and pass on authority to the next generation. 
These took place in approximately the same year(s) throughout the region, although the last 
generation has been decades late to retire in many areas. 

The question then is how these two parallel systems related to each other historically. R. 
G. Abrahams argues that, because age and generation systems operate on different principles, 

societies that use both, must logically give one priority over the other. 26 Gulliver notes that 
generation systems are only possible where the primary function of age organization is not 
military, since a generation-based system does not ensure that all young men will be recruited at 
their prime. 27 The age-set is a useful tool for societies that promote the ethos of a warrior class 
while the generation-set is concerned with the orderly succession of generational growth and 
authority. The warrior class of an age-organized society often gains a fair amount of autonomy 
from their elders and prestige while a generational system is usually more firmly under the control 
of the elders. 28 

Although western Serengeti peoples practiced either age- or generation-sets or both since 
the distant past, oral testimonies and the evidence of comparative ethnography seem to indicate 
that, at least by the nineteenth century, the system of cycling generation-sets predominated. With 
the advent of the disasters age-set structure took on a new and more dominant position among the 
easternmost peoples, exposed to Maasai raids and to the influence of Maasai culture. The demands 
for military expertise made age-sets more valuable at that time. Yet even as age-sets became more 
dominant in the nineteenth century, they were incorporated into a cycling system that retained the 
organizing principles of generation with the outward ethos of the age-set. 29 

26 R. G. Abrahams, "Aspects of Labwor Age and Generation Grouping and Related 
Systems," in P. T. W. Baxter and Uri Almagor, editors, Age. Generation and Time: Some Features 
of East African Age Organisation (New York: St. Martin's Press, 1978), pp. 37-68. 

27 Gulliver, "Turkana Age-Organization," p. 921. 

28 See Berntsen's discussion of the increasing autonomy and power of the warriors over the 
elders in Maasailand, "Pastoralism, Raiding and Prophets," pp. 225-227, 310-317; Galaty, 
"Maasai Expansion," pp. 75-86 

29 See Chapter 9 for more details on the new age-set system. 

People of the Mara Region seem to have derived the great variation in age-organization 
that exists today from a common set of generational principles identifying people with the health of 
the land. Oral testimonies associate these principles with generation- rather than age-sets in the 
past. Among the Lakes peoples, or the Suguti language group, the memory of these generation-set 
names survives but has entirely fallen out of use with more extensive contact and influence from 
around the lake in the nineteenth century. Their surviving age-organization is more like age- 
grades, without specific institutional responsibility, although they do practice initiation and 
circumcision. Ruri elders along the lake, however, still remember the eight named generation-sets 
that once "ruled the land" and functioned in protecting the crops, providing rain and keeping out 
disease. 30 Among the Kuria and Zanaki peoples, who are farther inland, generation-sets remained 
the dominant form of age-organization. Elders from among the peoples living farthest east, 
Ngoreme, Nata, Ikoma and Ishenyi, all acknowledged that generation-sets are "senior" to age-sets, 
although age-sets began to perform the generation-set rituals and take over many of their functions 
during the nineteenth century disasters. During the disasters, the Ngoreme divided each generation 
into age-sets and the Nata, lkoma, Ishenyi derived the age-cycles of the late nineteenth century 
from generational principles. 3 ' All these variations have in common the ritual precedence given to 
generational principles. 

The generational principle also takes precedence in everyday interactions. Greetings are 
structured on the generational relationships of the pair greeting each other (i.e., grandmother, 
mother, daughter, granddaughter). At any gathering of people, even today, participants arrange 

30 Interviews with Daudi Katama Maseme and Samueli Buguna Katama, Bwai 1 1 
November 1995 (Ruri <f). 

31 Interviews with Sochoro Kabati, Nyichoka, 2 June 1995 (Nata <f); Kirigiti Ng'orogita 
Mbiso, 8 June, 1 995 (Nata cf). Both of these informants are Mwanyasaai. which was the only 
generation to "walk" in Nata memory. 

themselves in groups for eating and socializing according to generation and gender. 32 A ritualized 
apportionment of meat corresponds to gender and generation. A common verbal game is to argue 
over whether one should be greeted according to someone's generation or someone's chronological 
age. Because of the lack of total correspondence between generation and age a young person may 
demand a greeting of respect from an old man, whose father was equivalent in generation to the 
younger man. 

The generation-set integrates clans within the ethnic group or the age-set cycle and also 
serves an integrative function on a regional scale. Ethnic groups across the entire region use the 
same generation-set names, while they base the particular ceremonies on the individual ekyaro or 
clan territory. The generation-set unites the larger region with one set of names and generations. 
Because people at approximately the same stage in life use the same generation names across the 
region one can travel and receive hospitality over a wide geographical area based on norms of 
relationship between generation-sets." Age ceremonies were also once regionally coordinated. 
Elders say that both age- and generation-set ceremonies take place in succession, moving from east 
to west. After the Maasai and Sonjo perform their age-set ceremonies, the Ikoma follow, then the 
Ishenyi, the Nata, and the Ikizu. The generation-set ceremonies to maintain the health of the land 
take place after the age-set ceremonies but coordinated with them in the same year, assuring the 
success of the new age-set. 34 

Generation-sets function differently in each ethnic group today and take on slightly 
different forms. In Nata and Ikizu the two generation-set cycles crosscut clans and lineages so that 

32 Discussed for the Kuria by Ruel, "Kuria Generation Classes," pp. 30-3 1 . 

33 Ibid, p. 32. 

34 Interviews with Sochoro Kabati, Nyichoka, 2 June 1 995 and Kirigiti Ng'orogita, Mbiso 
6 June 1995 (Nata <f). 

each cycle would be found in each clan. Here the generation-set functions as a unifying device 
between the clans. In Ngoreme and Ishenyi each clan territory or ekyaro has only one rikora and 
Ngoreme is divided territorially into Saai and Chuuma generation-set cycles. The age-cycle 
territory unifies a number of clans but corresponds to only one of the generation-set cycles. In this 
configuration one generation-set cycle unifies clans within the age-set cycle territory, rather than 
the ethnic territory. 

Peoples in the Mara Region and throughout East Africa have used age-organization in 
countless ways and in countless forms. Rather than understanding them as a specifically defined 
institution we should see them as tools by which individuals accomplish the important tasks of 
society. Generation-sets survive only because they continue to perform important ritual functions 
that protect the people and the land. 35 Although Mara people, as well as anthropologists, enjoy a 
debate on the political or military functions of the rikora, studying the one function of ritual, which 
has survived as secret and vital information for maintaining prosperity, seems more important for 
the historian. By performing rituals to protect the health of the land, the generation-set inscribed 
on the ground a specific territory (ekyaro) and they expressed the identity of the people who lived 
within those boundaries. 

Scholars writing on age- and generation-sets in East Africa have extensively investigated 
the political and military functions of age-organization and have paid little attention to their ritual 
role in relation to land and health. 36 Colonial administrators and early anthropologists established 
this precedent with their concern to identify local structures of political and military control in 

35 Baxter and Almagor, "Introduction," Age. Generation and Time, p. 24. 

36 Ibid, pp. 1-3. 

areas where no chiefs existed. 37 The few academic writings on generation-sets in the Mara Region 
extend their analysis beyond the military and political function of age-organization, but fail to 
make the connection between generation-set ritual and a communal identity related to the land. 
Bischofberger, writing on the Zanaki, concluded that the generation-set had political, social and 
ritual functions, with the overall effect of integrating the kinship-based Zanaki. 38 Bishofberger 
mentions that the Zanaki generation-sets were responsible for rituals that would assure the well- 
being of the ekyaro, but could elaborate no further. 39 However, whether these scholars emphasize 
the political/military/ritual role or the more diffuse organization of daily social interaction, they do 
not situate age-organization in a specific relationship between land and people, critical to prosperity 
and the ongoing continuity of the generations. 40 

Although the elders ultimately controlled the generation-set, the rituals of the generation- 
set in relation to the land created the social cohesion and identity necessary for political authority 

37 Tanganyika government anthropologist, Hans Cory, wrote that the "supreme power" in 
Ikoma was the "warrior age-grade," that was "divided into three military units." After they retired 
from warriorhood, the elders constituted a "chama" under the leadership of a "mukina" who 
"managed the civil affairs of the respective military units only." Although Cory admitted to his 
lack of information on the political functioning of the "age-grade system in pre-European times" he 
assumed that it was "best adapted to meet military emergencies" and was thus now "obsolete." 
Hans Cory, "Report on the pre-European Tribal Organization in Musoma (South Mara District 
and ... Proposals for adaptation of the clan system to modern circumstances." 1945, CORY #173 

38 Bischofberger. The Generation Classes p 100-101. Mustafa, "The Concept of 
Authority," pp. 55-83. Ruel saw generation-sets among the Kuria as "embodying a systematic 
pattern of relationships which serves to determine the status of any individual person vis-a-vis 
others." He preferred to see generation-sets as "social categories" rather than "corporate groups." 
Ruel, "Kuria Generation Classes," p. 17, 33. Here he refers mainly to the relational norms between 
generations: those in adjacent classes have a relationship of respect and deference while in alternate 
or the same class one of joking and intimacy. 

39 Bischofberger, The Generation Classes , pp. 1 1-16. 

40 This is due in large part to the reluctance of Mara peoples to talk about the nature of 
these secret rituals. 

to function. Baxter and Almagor explain that age organization is always subordinate to the 
corporate authority of lineage or kinship-based organization, which controls property relations 
through older family heads. They posit that scholars have overemphasized the political/military 
function of age-organization. For these scholars, age- or generation-sets function as "agents of 
force" rather than "controllers of force," and act as "agents of force" only when community 
consensus occurs. Age-sets perform critical tasks but are not the source of authority. All age- 
organizations exist in coordination with other types of social organization. If elders controlled 
productive resources, including labor, in these societies then age organization was a powerful tool 
in their hands. 41 To say that age-organization does not have political power is not to weaken its 
importance, nor should one conclude that age-organization must have had more political power in 
some pristine "tribal" past. The ritual function of generation-sets was a powerful political force on 
its own terms. 

Scholars in this region have found it difficult to identify the ways in which Mara peoples 
used the principles of generation to express a ritual relationship to the land, in part, because of the 
secret nature of these rituals. Outsiders' knowledge of these rituals would endanger the health of 
the land and its people by opening it to malevolent intervention. In Nata, Ikoma and Ishenyi 1 had 
difficulty learning anything about the rikora except that it existed. The details about the 
generation-set rituals are second only in secrecy to the eldership title secrets about which initiates 
swear an oath of silence. Elders usually answered my questions about generation-set ritual with 
vague and generalized descriptions of the generation-set walk as analogous to the solidarity walks 
of today. Similarly, Zanaki elders told Bischofberger that the Zanaki rikora in power would 
"walk around" to keep in shape and strengthen the solidarity of the group. Each cycle would do 

"" Baxter and Almagor, "Introduction," Age. Generation and Time , pp. 10-19 

this walk separately, wearing traditional skins, carrying sticks and stopping to eat at generation-set 
members' houses. 42 Secrets are another way of putting the discourse on territory and identity into 
an unquestioned category of "tradition" and removed from public debate. The elders who control 
these secrets assert authority over the definition of territory and a place-oriented identity. 
The Rituals of Healing the Land and the Retirement Ceremony 

The main task of the generation-set ritual is to maintain the health of the land. I describe 
these rituals in the "ethnographic present" because that is the way elders told them to me. Most 
elders whose accounts appear here participated in these rituals, some as leaders. Yet they did not 
describe the particular "walk" that they observed. Rather, they gave me a general account of a 
"traditional" walk formed from their own experience and the stories they heard of other walks. 
This was a deliberate narrative strategy to place these stories in the middle time frame of social 
process and outside the flow of historical events. Connerton argues that in order for these 
"commemorative ceremonies" to reenact the past they must move into "ritual time" where an event 
is indefinitely repeatable. 43 Elders enhance the authority of the ritual by portraying their own 
contingent experience of the ritual as a timeless pattern. 1 describe these rituals with commentary 
to point out what 1 believe to be the meaning behind these rituals in their historical context. 

The practice of this ritual among the seven ethnic groups from whom 1 have accounts 
varies extraordinarily around several central themes. Ideally, every eight years in Nata and Ikoma, 
the men who are in the "ruling generation" of the two cycles, Chuuma and Saai, alternately walk 
around the boundaries of the land, kukerera. The walk takes place together with the age-set 
ceremonies as a way of preparing for their "rule." In Ikizu and Zanaki, where no cycling age-sets 

42 Bischofberger, The Generation Classes , pp. 65-67. 

43 Connerton, How Societies Remember, p. 66. 

exist, the walk is part of the retirement ceremony of the generation-set every 20-30 years. 44 In 
Ishenyi and Ngoreme the walk takes place in symbolic terms, around a tree or at a feast. Yet in 
each case the bodily practice of the walk physically inscribes the boundaries of social identity in 
space. The appropriation of territory through ritual control over the land secures the health of the 
land and the people. The ritual creation of enclosed space defines a territory and the identity of 
those who live there as one people.'" 

The walk in all its variation focuses on passing the orokoba or the medicine of protection 
around the land. The leaders of the generation-set literally plant medicine bundles in the soil at 
intervals around the land or in the bark of certain trees that they slash. This medicine protects the 
land from enemies and disease and ensures its fertility. Either a prophet, sometimes from a distant 
land, prepares this medicine or the generation-set itself keeps and passes on the medicines. In either 
case the job of the rikora (generation-set) is to walk around the land and encircle it with protective 
medicines. The walk also preserves the peace of the land, it "cools the land," and makes it 
productive. The "coolness" of the land assures the health and well-being of the people. The 
generation-set achieves this goal by different methods in each of the ethnic groups but the basic 
principles remain the same-the ritual acts of binding and enclosing ensure safety and health. The 
word kukerera is likely a prepositional form of a proto-Great Lakes Bantu word (kukila) meaning 
"to overcome, surpass, heal, unify." The prepositional form implies an object, in this case, the land 
and all that it stands for. 46 I describe the rituals of each ethnic group, as elders narrated them to 

44 Interview with Kirigiti Ng'orogita, Mbiso, 8 June 1995 (Nata <?). For the Zanaki, 
Benjamin Mkirya, Historia, Mila na Desturi za WaranaH (Peramiho: Benedictine Publications 
Ndanda. 1991), p. 33. 

45 See Robert J. Thornton, Space. Time, and Culture among the Iraqw of Tanzania fNev 
York: Academic Press. 1980), p. 19. 

David L. Schoenbrun, personal communication. 

me. I emphasize those aspects of the rituals of each ethnic group that demonstrate how these 
rituals function to create territorial identity and group cohesion under the authority of the elders. 

In these narratives the ritual unifies diverse clans into one ethnic or age-set territory, 
ekyaro. In previous chapters I showed that clan territories, known in the past as the ebyaro, were 
the largest functional unit before the disasters of the late nineteenth century. One can imagine that 
western Serengeti people might have used these same rituals in the pre-disaster era to unite diverse 
lineages, or people of diverse origins, into one ekyaro. 
Ikizu: Uniting Clans Lands Under the Communal Authority of the Elders 

The ritual walk in Ikizu unites the various clan lands that make up Ikizu by asserting the 
communal authority of the generation-set elders. The Ikizu walk begins at a point in the 
westernmost part of the land and moves east, each night the abanyikora (new generation-set 
members) feast and sleep at a different clan center, hosted in a wealthy man's homestead. On the 
third day they pause at the tree called Sarama mo Mogongo and send a small delegation to the 
mlemi rainmaker's house to get the medicines of rain (omoshana) [See Figure 8-1 : Sites of the 
Ikizu Walk, for a photo of the Sarama mo Mogongo tree and the present mlemi at Nyakinywa's 
cave]. The mlemi. as a descendant of first woman who brought rain, sends his delegate to oversee 
the rain medicine but does not participate in the walk. The Ikizu prophet (omunase), as a 
descendant of first man who made fire, provides the medicine of protection and the medicine for the 
new fire. The whole group arrives for the ceremonial climax near Chamuriho Mountain in the 
easternmost part of Ikizu, where Muriho first claimed the land by planting the spirits of his 
ancestors, emisambwa. 

The ceremonies at Chamuriho Mountain feature the ekimweso or the sanctitlcation 
ceremony of the fire to pass on authority from one generation to another. The older generation 
retires and the new generation builds a new fire, with the medicines of the prophet, after all Ikizu 


Mtemi Adamu Matutu, Gaka, 31 August 
1 995, at the cave of Nyakinywa 

Kinanda Sigara, Kihiri Mbiso and son, Sarama, 20 July 1995, standing at the tree of 
Sarama mo Mogondo 

Figure 8-1: Sites of the Ikizu Walk 


extinguish their homestead fires. The elders sacrifice a white goat and cut its hide into strips that 
the new generation-set wear on their middle fingers smeared with butter.'" When the ceremonies 
finish, the new generation-set that has just taken power retraces their steps back east to complete 
the circle. They spread the medicines for healing and rain and distributing the new fire to all the 
homesteads as they go. This act renewed the peace of the land. 

Although the Ikizu walk holds many elements in common with rituals around the region, in 
Ikizu the walk takes on a particularly salient political meaning. The elders use the walk at once to 
assert the unity of clan lands and to claim their own communal authority over the rainmaker and 
the prophet. The walk evokes the emergence story to legitimize the authority of the elders as the 
representatives of Muriho (who first established possession of the land) over first woman 
Nyakinywa (the rainmaker chief) and first man Isamongo (the prophet). The elders spatially 
represent their interpretation of the political make-up of Ikizu through the walk. The generation- 
set pauses in each of the clan lands of Ikizu. The walk does not even approach the seat of the 
ulemi in Kihumbu, except to obtain the rain medicine. The rikora (generation-set) takes 
responsibility for healing the land rather than the "chief." The prophet, too, provides the protection 
and fire medicines but does not participate in the ceremony. The walk of the rikora embodies all 
these issues of contested forms of authority in Ikizu. Through this ritual, the elders reassert their 
control over potentially powerful sources of authority in the prophet and the rainmaker. The 
generation-set is their tool, protected by its association with timeless "tradition" and the high stakes 
of ensuring health and fertility. 

" Interview with Ikota Mwisagija and Kiyarata Mzumari, Kihumbo, 5 July 1995 (Ikizu 
cf); Mturi, "Historia ya Ikizu na Sizaki." 

Ishenvi: Variation on the Theme of Territorial Unity 

Ishenyi elders testify that here the walk of the rikora consists only of a symbolic walk 
around a certain tree (msingisi or msari) n to prepare for the initiation of a new age-set into power. 
The clans (hamate) of Ishenyi are divided into two moieties, with each moiety representing one 
generation-set cycle, either Saai or Chuuma, who carry out the ritual separately. By contrast, in 
Nata, Ikoma and Ikizu each clan claims members of both rikora cycles. In Ishenyi the rikora did 
not crosscut and unify the clans. However, it did unify the age-set territories that developed by the 
end of the nineteenth century to replace clan territories. Age-set territories contained many clans in 
both generation-set cycles, each time the generation-set would walk it covered the entire age-set 
territory. 49 

The Ishenyi ritual holds many elements in common with generation-set ritual across the 
region. The ceremony to install the new age-set takes place at the homes of the newly chosen 
leaders, lasting for eight days of feasting, singing and dancing. As in Ikizu, the new age-set cuts 
strips of hide from the ritually slaughtered animal and wears them on their fingers (ebeshona). 
Generation-set elders bless all the people in the ceremony, who then extinguish all of the homestead 
fires in the ekyaro and light them from the new ceremonial fire started by twirling a stick in a 
board (just as first man lit a fire). The generation-set encircles the land with the medicines of 
protection or rain according to need. The age-set obtain rain medicines (amusera) from the 
rainmaker that they mix with milk and flour. The age-set leaders spread the medicines with a cow's 

48 Interview with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi <f\ Each 
particular kind of tree used in rituals has a symbolic significance. 

49 Main interviews on the Ishenyi kerera were: Mang'ombe Morimi, Issenye Iharara, 26 
August 1995; Rugayonga Nyamohega, Mugeta, 27 October 1995; Morigo Mchombocho 
Nyarobi. Issenye, 28 September 1995; and Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi 

tail as one leader moves east and another west around the boundaries of the ekyaro?" [See Figure 
8-2: Rikora Leaders.] 
Nata: Generational Growth through the Mediation of Outside Forces 

In Nata the age-set ritual, has taken over much of the work of the generation-set. The 
ritual symbolizes the generational principles of fertility and growth and occurs regularly every 
eight years. When it is time for these ceremonies to take place, the elders chose the leaders 
(abachama) of the new age- or generation-set and send these youths on various errands. The 
youths go outside Nata, north, south, east and west, to collect the ritual ingredients that the elders 
mix together and sprinkle as a blessing on the new age-set. These ingredients include water from 
Lake Victoria, honey from Riyara, excrement of an unweaned child, livestock manure, and millet 
or other grains. They also choose a young unblemished black bull for sacrifice. Finally, the new 
age-set builds a ritual fence around the homestead of their leader where the ceremony will take 

At the culmination of the ceremony the head of the homestead (the father of the new age- 
set leader) takes a black cow or wildebeest tail, eghise, as the symbol of his elderhood, and 
sprinkles a mixture of the ingredients brought from outside, along with the stomach contents of the 
slaughtered bull, on the new set as they stand in a circle with their wives, rubbing it on the breasts 
of the women. Since the new age-set entered their age-set, saiga, eight years after circumcision 
they were often married with children. As the ritual father sprinkles he prays that the youth might 
'have children, abundant livestock, good harvests and rain' during the "rule" of their saiga. 

50 Interview with Rugayonga Nyamohega, Mugeta, 27 October 1995 (Ishenyi if). Two 
children who wet their beds accompanied the two leaders and actually did the work of spreading 
the medicine. 



This ritual unifies the clans of Nata— with one clan moiety represented at the ceremony by 
the symbol of a bell on the neck of a young black bull and the other by an old heart-shaped trade 
hoe from the Sukuma Rongo clan. The bell and the hoe are symbols of productivity that the elders 
pass on to the next generation. The elders cut the hide of the sacrificed bull into strips that the 
eight (aba)chama leaders wear on the middle finger. At the ceremony the two leaders, one from 
each clan moiety, stand together on another hide, with the longest strip of the slaughtered bull hide 
stretched between them. An elder cuts the strip in half and declares them blood-brothers, 
baragumu. After feasting and dancing all night, they move on the next morning to feast at the 
homestead of the other clan moiety leader. 51 

Although the above paragraphs describe a Nata age-set ceremony, the important exchanges 
embodied in it are generational. The fathers (rather than the retiring age-set) choose the new 
leaders, perform the blessing, cut the hide strip between the leaders, and pass on the clan symbols. 
In fact, the retiring age-set has no part in this ceremony and no ritual marks their passing. In the 
restructured Nata age-set system of the nineteenth century each age-set cycle lived in a different 
territory, which meant that they hand over authority to a group physically removed from 
themselves. Elders describe the relationship between adjacent age-sets as one of animosity. A 
mock battle of sticks takes place well before the ceremony so that the new age-set "drives out" the 
old. 52 Conflict must not enter the rituals themselves. Furthermore, although elders characterized 
the age-set as the defense against cattle raids, all of the age-set ceremony symbols concern peace, 

51 The most important interviews for information on Nata saiga ritual were: Mang'oha 
Machunchuriani (Mwekundu, elders who make preparations for the Saiga ceremony), Mbiso, 24 
March 1 995; Sochoro Kabati (Kang'ati of the Gikwe for the Bongirate Saiga), Nyichoka, 2 June 
1995, Kirigiti Ng'orogita (Rikora Mchama), Mbiso, 8 June 1995; Mang'oha Morigo (Kang'ati of 
the Moriho for the Bongirate Saiga), Bugerera, 24 June 1 995 (Nata cf). 

52 This is also the case with Tatoga generation-sets. 

productivity of land and people and prosperity. The ritual blessing of both women and men alike 
also suggests that the concern is primarily with fertility and not war. In short, the age-set 
ceremonies seem to have coopted the symbolic content of the generation-set, leaving it with 
responsibility for the walk, kukerera, alone. Nata elders said that the rikora, or generation-set, is 
"bigger" or "senior" to the age-set, with authority over the land that was fundamental to everyone's 
well-being. They claimed that the saiga or age-set "ruled" the land but also portrayed it as a 
"game" of youth. 53 

The spatial symbols of the ritual represent the mediation of external forces to protect the 
internal community. Youths leave Nata territory to collect the symbols of prosperity outside the 
community, and they return with these things, bringing them inside the ritual homestead fence. 
Just as first man came from the wilderness and civilized the home by bringing fire, the symbolic 
reenactment of this movement assures prosperity. The Nata generation-set walk, kukerera, also 
involves collecting things from many places outside Nata used to make the medicine for the land. 
Elders said that these things brought health to the people, fertility, wealth in livestock and abundant 
harvest of crops. 

The Nata generation-set walk itself is a communal ritual that requires the participation of 
all for the medicine to work. The institution of the generation-set emphasizes the equality of all 
those within one generation and the authority of the elders. Everyone walks with the generation-set 
all around Nata. In the colonial days the walk concluded at the chiefs house to feast on the last 
day. Elders considered it an honor, and also a huge expense, to entertain the rikora during the 
walk. The host provided meat from his own herd and made enough beer for everyone. 54 The 

55 Interview with Sochoro Kabati. Nyichoka, 2 June 1 995 (Nata <?). 

54 Interview with Mang'oha Machunchuriani, Mbiso, 24 March 1995 (Nata <?). 

community held "big men" accountable by requiring them to "feed" the people in return for their 
respect and support. In Nata during the last walk one man refused to have the rikora come to his 
house and his generation subsequently cursed him. 
Ikoma: Appropriating Tatoga Spirits to Protect the Land 

The Ikoma use the Machaba tusks, as their erisambwa obtained from a Tatoga prophet, in 
the generation-set rituals to unify the territory. The Tatoga prophet of the lineage of Gambareku, 
of the Relimajega clan carries the Machaba at the head of the walk. He makes the medicine, the 
orokoba of the Machaba, that the age-set plants around the land, and serves as an Ikoma age-set 
leader." The Ishenyi also use the Tatoga in their age- and generation-set ceremonies. A Tatoga 
prophet of the rairunaking clan of Gaoga serves as the top leader of the age-set, fully functional in 
the most intimate of Ishenyi ritual matters. This prophet also provides the medicines for rain and 
to bless the new-year's seeds. 56 The Ishenyi reliance on the Tatoga is a result of having killed their 
own prophet, Shanyangi, at Nyeberekera. Both Ishenyi and Ikoma acknowledge the first-comer 
status of the Tatoga in their rituals to appease the land. 

The Ikoma ritual also spatially symbolizes the unity of the clans into one land. The 
Ng'orisa clan, which controls the male tusk, begins the walk where the tusk is kept in the west and 
the Rogoro clan begins at the place where the female tusk is kept in the east. They circle the land 
and meet in the center, each going through the homesteads of the members of their clans. Each 

55 Interview with Machota Sabuni, Issenye, 14 March 1996 (Ikoma <f). The last time the 
Machaba appeared in Ikoma ritual was in 1994. 

56 Interview with Stephen Gojat Gishageta and Girimanda Marisha Gishageta, Issenye, 28 
March 1 996 (Tatoga tf). 

night they feast, dance and sleep at a different and chosen compound in the different clan 
settlements. 57 
Ngoreme: Generational Passing and the Authority of Elders 

The Ngoreme generation-set rituals emphasize the generational authority of elders. Here 
no rikora walk exists but the generation-set still functions in the protection of the land. Each 
rikora carries out its function within a particular ekyaro— defined here as clan territory, local 
community or the territory of an age-set cycle. When the sons of most of the rikora mature 
(having their own wives, children and homestead) it is time for the generation in power to retire. 
These men have already completed the eldership ceremonies and carry the black tail as a symbol of 
their eldership. They decide together that it is time to retire and then at another communal 
ceremony of the ekyaro they pass on the symbols of their office. 

The public symbol of this transfer of generational authority is the spatial arrangement of 
generations at a public feast and the division of meat, as is common across the region. When the 
elders are ready to retire, they allow the young men to eat the meat of the back, omugongo, at 
communal gatherings. At any feast people sit in groups according to generation and gender and eat 
the appropriate kind of meat. The oldest retired men sit separately by the grain storage bins and 
eat the softest parts of the cow like the lungs (sarara). The ruling elders sit inside the cattle corral 
and eat the back meat (omugongo) and the head. The young men in their esega or age-set status 
sit together outside the cattle corral and eat the chest meat, legs or the hump (sukubi). 5 ' The spatial 
position of the elders inside the cattle corral and their possession of the best cut of meat is a sign of 
their dominant authority in generational affairs. They are the "back" of society, leaving the youth 

57 Machota Sabuni, Issenye, 14 March 1996 (Ikoma if). The Ikoma were some of the 
most reluctant to talk about the rikora and keep secret much of this information. 

58 Interview with Mwita Maro, Maji Moto, 29 September 1995 (Ngoreme d"). 

to do the physical work. The young men most poignantly sit expectantly outside the corral. The 
oldest men become like women as they sit at the granaries, where young women undergo 
circumcision. They also eat some softer internal organs that are the portion of women, 
corresponding to her inside and enclosed space. A common method of putting a young man in his 
place is for an elder to ask if he has 'tasted the back meat' yet. 

Elders mark the passing of generational authority by the exchange of several objects. At 
retirement the Ngoreme elders give the new generation the horn that calls the rikora together for a 
meeting or sounds the alarm in times of danger. Each clan (hamate) has its own horn of the 
rikora. [See Figure 8-2: Rikora Leaders] They make this instrument from the long horn of the 
oryx or greater kudu with special medicines embedded in the strips of buffalo and lion hide 
wrapped around the horn. Each time the generation-set passes on the horn they slaughter a bull 
and use the hide to renew these wrappings. The tight binding of medicines with hide strips employs 
the same symbolic logic as the encircling of the land with medicines of protection or the wrapping 
of medicine bundles of protection for war. The retiring generation-set passes on the horn to the 
ritual care of the first son born to their members, the omotangiP Ritual horns are common 
symbols of generational authority throughout the region. 60 

The last thing that the retiring generation passes on in Ngoreme, at least among the Saai of 
Bumare, is the generation-set medicines called the omugongo wa mwensi, or the protection 
medicine of the generation. The omotangi, or the firstborn of the generation, receives this along 

The most important Ngoreme interviews on these rituals were, Nsaho Maro, Kenyana, 
14 September 1995; Mwita Maro, Maji Moto, 29 September 1995 (Ngoreme <f) who kept the horn 
for the Iregi of Bumare and was the aba Maina mlangi, having obtained the horn in 1 957 he was 
well overdue to pass it on to the abaSaai. 

60 Each Ishenyi clan also had a horn (enchobe) which was passed on as the older 
generation retired. Rugayonga Nyamohega, Mugeta, 27 October 1995 (Ishenyi if). 

with the horn. Elders consider this medicine more powerful than the medicine of the horn and in 
Bumare the last group of retiring elders refused to pass on these medicines. I assume that the 
Ngoreme rikora uses this medicine to protect the land. In this variation the rikora itself passes on 
the secret medicine, rather than going to a prophet to obtain it. 61 
Kuria. Zanaki and Tatoga: Recurrent Symbols of Generational Growth 

Generation-set rituals draw on a common set of symbols recognized throughout the Mara 
Region, beyond what 1 have defined as the western Serengeti and beyond East Nyanza Bantu- 
speakers. The Kuria, Zanaki and Tatoga also practice generation-set rituals using the same 
symbols. Kuria generation ceremonies resemble Ngoreme ones because they both commemorate 
and normalize rather than confer or effect a collective change in the status of its members. When 
the younger generation is ready to establish their own homesteads separately from their fathers, 
their fathers perform a ceremony called "going to the enclosure" in which they build a symbolic 
homestead and ritually bless their sons and their sons' cattle and wives. At the end of the eight- 
day ceremony, each homestead lights a ritual fire. "Sons" could not take individual eldership titles 
until their "fathers" promoted the entire generation. The generation only acts as a defined social 
group with its own corporate identity in a ritual context. The southern Kuria territories perform a 
similar egekereero or "retirement" ceremony in which a tree, usually a fig species (makereero), is 
identified with the retiring generation. Elders do not allow the presence of a stranger in the ekvaro 
during this ceremony. 62 The recurrent symbols of regeneration and growth appear in the tree, the 
symbolic homestead, the blessing of cows and wives, and the passing on of things from father to 

61 Paulo Maitari Nyigana and Ibrahim Mutatiro Kemuhe, Maji Moto, 29 September 1995 
(Ngoreme tf). 

62 Ruel. "Kuria Generation Classes," p. 21-28. 

son. Both closing the ekyaro boundaries to strangers and enclosing the symbolic homestead with a 
fence defines a territorial identity. 

The Zanaki generation-set retirement ceremony takes place every twenty or more years in 
each ekyaro or clan territory at the emisambwa sites of sacred groves or large rocks. There the 
retiring generation forms a circle, with the new generation behind them. They take off the skins 
that they wear around their waists and tie them over their shoulders as retired elders normally 
dress. Their last act is to cut another opening in the symbolic homestead fence for the youth. The 
retiring elders give the new generation secret medicines (emigongo), including a horn (ekombyo). 
The Zanaki generation-set in power also perform a ritual if the land is in an unhealthy state 
(okutura ekiaro). The elders bless the gathered people with a mixture of a sacrificial goat's 
stomach contents, millet, and water. Each homestead extinguishes their fires and the generation-set 
lights a new fire by twirling a stick from the mirama tree on a piece of ivory and distributes it to 
each homestead. Then they pass the orokoba protection medicines around the land. 63 

Finally, the Tatoga practice of linear generation-sets, called saigeida, share many ritual 
symbols with their neighbors. When the new saigeida comes into power, they build a large bonfire 
over which the new initiates must walk. Elders said that one Tatoga prophet, Gishageta, used spit 
and a "cowhide rope" in his prophecy. He embedded his medicines in certain kinds of trees to 
ensure, fertility, prosperity, rain and protection. Another Tatoga rainmaker killed a black sheep 
and gave everyone the strips of skin to wear on the middle finger of their right hand. The 
Relimajega clan controlled protection medicine against the Maasai. An Isimajek informant said that 

63 Mkirya, Historia. Mil a na Desturi . pp. 39-43, 59-62. Interview with Zabron Kisubundo 
Nyamamera, Bisarye, 9 November 1995 (Zanaki cf). 

during the circumcision ceremonies they put strips of hide on their fingers. 64 All these are 
examples of a shared symbolic world of ritual between farmers and herders. Neither seems to be 
"imitating" the other, rather through generations of living side by side they have developed a 
common ritual language, each adopted to a specific lifestyle and historical circumstances. 

Interpreting the Walk 
Because the rituals of the rikora seem to be based on very old generational principles in 
relation to the land, rather than introduced after the period of disasters, we can then interpret the 
core spatial images of bodily practice as they might have applied to the historical context of the 
nineteenth century, before the disasters. Connerton argues that the rituals of "commemorative 
ceremonies" are relatively invariable over time. Similarly, Maurice Bloch in his analysis of 
circumcision ritual among the Imerina of Madagascar shows that little change has taken place in 
the internal content of this ritual over a period of two hundred years, yet the "functional" role of 
this ritual changed often over that time. 65 The three previous chapters have reconstructed a picture 
of economic subsistence, homestead, settlement and territorial patterns as they might have existed 
in the nineteenth century. 1 analyze the symbolic content of the rituals to illustrate how 
communities might have used these rituals (which we know to be old) in the nineteenth century to 
form a communal territorial identity in relation to the land. 

64 Interview with Stephen Gojat Gishageta and Girimanda Marisha Gishageta, Issenye, 27 
July 1995 (Tatoga cr). MarundeGodi, Juana Masanja, Mayera Magondora, Manawa, 24 
February 1996 (Isimajega <f). 

65 Bloch, From Blessi ng to Violence . See also Feierman's analysis of the concepts of 
"healing the land" and "harming the land" reinterpreted in changing historical contexts of Shambaa 
in Tanzania. Feierman, Peasant Intellectuals . 

The Space of the Walk: The Ekvaro 

The space of the walk is coterminous with the ekyaro or the "land," which defines the 
imagined community, embodied in the act of kukerera. The ekyaro is not a fixed unit. In ordinary 
speech the ekyaro refers to anything from the local community to the nation state. Whereas in 
Ikizu the space of the walk corresponds to the united clan territories, in Ishenyi elders define the 
ekyaro as the age-set territory that developed in the second half of the nineteenth century. 
Ngoreme, Kuria and Zanaki all define the old ebyaro as the autonomous units of clan territory. 66 
The generation-set of each ekyaro performs their rituals separately. The smaller ethnic groups of 
Nata and Ikoma today define ekyaro as the entire ethnic territory. They also call the age-set 
territories ebyaro. In effect people define the ekyaro situationally as the land over which they have 
control by a ritual definition of its boundaries. The heart of the concept is the relationship between 
people and land rather than a fixed unit. Yet, I interpret the ekyaro as a "territory" because it was 
defined as an enclosed and bounded area of land. 

The space of the rikora walk changed according to the shifting definitions of group 
boundaries. Ikizu was one of the few ethnic groups in the Mara Region with an emerging 
institution of chiefship when the Germans arrived. The utemi rainmaker "chief united the 
heterogeneous clans and lineages into one land, ekyaro. Before the utemi each clan territory might 
have performed their own generation-set ceremonies. However, in these rituals the generation-set 
asserts its authority over the mtemi to represent a united Ikizu. In Ikoma, Ishenyi and Nata where 
the generation-set now walks over the ekyaro of the ethnic group, a change in the definition of 
ekyaro occurred, from clan land to age-set cycle land to ethnic group land. 

66 Ruel, "Kuria Generation Classes," p. 15. Ruel estimated that the Kuria ebyaro of the 
past consisted of between 3,000 and 10,000 people, which is about the size of all of Nata or Ikoma 
or Ishenyi today. 


If we accept the evidence presented in Chapters 6 and 7 concerning cian territories and 
lineage settlements, then the ekyaro defined by the generation-set in the nineteenth century before 
the disasters was the clan territory of loosely affiliated lineage settlements within one group of hills 
and separated from each other by empty wilderness bush area. Within a context of settlement 
mobility and scarcity of people rather than land, each ekyaro would be competing to attract 
newcomers to open more land. Wealthy men could provide security and extensive regional clan 
networks with access to a variety of resources that would ensure prosperity. The mechanisms for 
incorporating strangers as native born were also well in place. The problem that remained was 
how to unite these diverse peoples into a territorial unit with a sense of communal identity. 
Western Serengeti peoples used the generation-set to overcome the divisions inherent in lineage- 
based settlement structures to form territories. 

The generation-set walk functioned to unify the ekyaro. Nata elders, while reticent about 
the details of the generation-set walk, emphasized it as a ritual for all of Nata-"it brings Nata 
together." Bischofberger states that the Zanaki generation-sets integrated the autonomous clans. 
Although Ikizu was also divided into smaller territorial ebyaro of the clans, the generation-set walk 
symbolically brought together all of Ikizu in the space of the walk. Nata elders said that the saiga, 
or age-set, divided, while the rikora unified. The main work of the rikora was to bring together the 
clans and the age-sets. 6 ' Ruel states that Kuria generation-set designations in a communal setting 
emphasize solidarity and responsibility to the larger group rather than division and divergent 
interests. 68 

67 Interviews with Sochoro Kabati, Nyichoka, 2 June 1995; Kirigiti Ng'orogita, Mbiso, 
June 1995 (Nata tf). 

68 Ruel, "Generation Classes," p. 33. 


The walk identified the rikora itself with the ekyaro or the land. A Nata elder said that 
rikora leaders could not move outside Nata. 69 Some elders described the walk encircling the 
boundaries of the ekyaro, while others said that it passed through the homesteads. This may have 
varied from place to place and according to the advice of the prophet who prepared the protection 
medicine. Ngoreme call the ceremony for the new Ngoreme age-set kwilaberi asega. Kwitaberi is 
derived from the verb -itaberi, "to bless the land." This term explicitly connects the initiation of a 
new age-set (asega) with the prosperity of the land.™ The ekyaro, the land, defined the identity of 
a people. When I asked elders what the ethnic group names meant, most said that they referred to 
a place. 71 They could not abstract group identity from its grounded context in the land. 

The walk of the generation-set formed and was formed by communal consensus. The 
ritual health of the land was dependent upon this consensus. To carry out this large-scale ritual, 
the elders mobilized the whole community and each did their part-prophets and rainmakers 
provided medicines, wealthy men provided food, women prepared food and beer, elders and youth 
cooperated in the preparations and the huge investment of time and energy in the process. The 
walk itself was symptomatic of the state of relationships in the community and thus the health or 
"coolness" of the community. 
The Symbols of the Walk: the Orokoba and the Fire 

The symbols of the walk, the orokoba and the fire, are central to the core spatial imagery 
that defines ekyaro or territorial identity. The images of encircling to form protective boundaries, 
binding up exterior forces to control their power, and covering the land with the smoke of 

69 Interview with Sochoro Kabati, Nyichoka, 2 June 1995 (Nata tf). 

70 Interview with Nsaho Maro, Kenyana, 14 September 1995 (Ngoreme <f). 

71 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata tf). 

purification all indicate how people have understood the forces necessary for the health of the land 
and its people. The spatial images of enclosure and boundaries are different from those of 
gendered homestead spaces, interstitial ecological niches, clan pathways, or nodes in a network of 
relationships because they represent different kinds of social identity. The territorial identity of 
enclosure creates a sense of cohesiveness for a diverse and mobile population living together on the 
same land. 

The term, orokoba, used for the medicine that the rikora passes around the land, is a 
symbolically charged word evoking the spatial images of enclosure, binding and interiority. An 
orokoba means literally a cowhide thong or rope but its meaning is multivalent and contextual. 
People use the leather thong for tying a cow to milk and so associate it with productive labor. The 
rope is made from hide and thus suggests the hide as the exterior cover, boundary or enclosure of 
the animal. That the hide thong is used to tie things up connects it to medicines for neutralizing 
malevolent power. 

Elders also used the term orokoba to refer to the strips of hide from the sacrificed animals 
that the generation-set members wear on their fingers during the ceremony. This symbol of unity 
and solidarity is widespread in East Africa, not respective of language group or culture. The 
boundary ceremony of the Iraqw reported by Thornton makes extensive use of strips of hide in this 
manner. 72 In the western Serengeti, one kind of blood-brotherhood is formed having each partner 
hold the end of an orokoba between them that is ritually cut by an elder. They are then true 
brothers in that they cannot steal from each other, sleep with the other's wife, or form marriages 
between their children. As described above, the leaders of the Nata age-set representative of each 
clan moiety perform this ceremony. 

72 Thornton, Space. Time, and Culture, pp. 88-97. 


In Nata orokoba also means the matrilineage, your mother's kin, and seems to be 
symbolically compared with the unbroken line (rope) of inheritance through the mother's side. The 
Ngoreme dictionary defines orokoba as "the umbilical cord." 71 Local understandings of 
conception give precedence to genetic inheritance through the mother. When elders consider a 
potential marriage partner, they investigate his or her mother's side for inheritable disease, mental 
illness or character flaws. People describe the closest kind of relationship between people as that of 
"one womb." Men contribute to the growth of the fetus by "feeding it" through intercourse, rather 
than, primarily, by supplying its substance. Men become fathers by the exchange of bridewealth 
rather than by a biological function. 74 In the rare and drastic case where brothers of one womb 
disagree and can no longer be reconciled, lineage elders perform a ritual in which they pass an 
orokoba through the wall of the maternal house, each brother holding one end of it. The thong is 
cut in half. The house, anyumba, now divides them rather than unites them as "children of one 

The rikora was not the only group to use orokoba medicine. Certain lineages possessed an 
orokoba in the form of an ekitana or a medicine bundle for protecting homesteads from theft or 
illness or to protect young men when they went after cattle thieves. 76 The medicine was always 

73 Maryknoll Fathers, Iramba Parish, Ngoreme-English Dictionary, n.d. 

74 For the importance and interpretation of folk models of procreation see, Poewe, 
Matrilineal Ideology , pp. 4-7; and Anita Jacobson-Widding, ed., The Creative Communion: 
African Folk Models of Fertility and the Regeneration of Life I Uppsala: Almqvist and Wiksell 
International, 1990). 

75 Interview with Nyawagamba Magoto, Musoma, 25 November 1995 (Nata cf). 

6 Ekitana from the root -tana, "to overcome, wear out, bewilder," the noun form perhaps 
meaning "medicine providing protection from violence." David L. Schoenbrun, personal 

protective rather than destructive. An individual homestead or an entire village hired the owner of 
the ekitana to come and pass the orokoba for protection against witchcraft, cattle theft or disease. 

When elders spoke of the medicine that the rikora passes around the land for its protection 
and healing as orokoba they evoked these other related meanings. They passed the orokoba 
around and encircled the land— containing it as the womb contains those linked by one maternal 
orokoba. Just as the hide of the goat, it marks the exterior boundary. Another word for clan lands 
(ekyaro) in Nata is orokoba? 1 The meaning here seems to be the land around which the orokoba 
was passed, the ekyaro. It referred not to the lands farmed or "owned" by the clan, but to the land 
over which the clan had ritual control by means of the orokoba. In all of the rituals described 
above the image of encircling and containing was repeatedly used (passing of rain and protection 
medicines on the boundaries of the land, strips of hide, fence enclosures, the medicines wrapped 
around the horn, the walk of encirclement). 

The power of the orokoba is in binding, surrounding or encircling. Brad Weiss discusses 
the use of protection medicines in Haya which bind (okuzinga in Haya). Wedding ceremonies, new 
houses, as well as death itself must undergo "binding rites" to ensure protection and peace.' 8 
Schoenbrun describes old Lakes Bantu ideas about medicines that bind up malevolent power to 
control it, evident in the common practice of the tight binding of various substances together in 
medicine bundles and charms. In many Bantu languages heat is associated with witchcraft while 
coolness is associated with peace and prosperity. Yet the control of heat through binding is also 
necessary to activate medicines of healing. The Nata word used to describe people or land that is 
whole and healthy is buhoro that comes from the Lake Bantu root word -podo, meaning quietness, 

11 Interview with Nyamaganda Magoto, Cultural Vocabulary list, Mbiso and Bugerera, 
1995-1996 (Nata <f). 

78 Weiss, The Making and Unmaking , p. 39. 

cold water, and good health. 79 The orokoba works to "cool the land" by tying up the powers of 
disorder. Thus the orokoba evokes both the idea of the containment of the womb or the homestead 
and the rope that ties up potential danger. Dangerous forces are neutralized by tying up their 
symbolic representations in medicine bundles or horns. The power of the wilderness or outside 
forces is civilized by bringing them inside. The control and containment, rather than the isolation, 
of outside forces make the land and its people productive. 

However, an encirclement or enclosure without an opening is associated with death. The 
prosperity of the homestead depends on its gateway that leads to the outside. The term for the 
extinction of a lineage group in Kuria is to be "stopped up" or "blocked off." Elders perform the 
ritual "piercing" or opening of a cow's stomach and sprinkling the stomach contents as the central 
act in any ceremonial sacrifice. All western Serengeti peoples pierce and elongate the holes in their 
ears, which is also the mark of an adult. Ruel reports that the Kuria derisively call an adult 
without pierced ears "a shut in" or "a blocked thing." Things that are completely closed in are 
potentially dangerous (an unpierced gourd, a calf born with skin covering its openings). 
Circumcision itself is an act of opening or cutting. The community must be enclosed for protection 
but dies without linkages to the outside, for wives, dependents and security. A healthy community 
mediates the dangers of outside forces by controlling its boundaries. 80 

The spatial metaphor of outside/inside is operational in all these rituals by using the 
stomach contents of the slaughtered animal (inside), the hide strips placed on everyone's fingers 
(outside), the building of ceremonial homesteads (inside) and fences (outside). Weiss connects the 

Schoenbrun, Etymologies . #335 and in personal communication. 

80 Malcom Ruel, "Piercing," June 1 958, Makerere Institute of Social Research. Conference 
Papers (1954-1958). For an in depth analysis of "blockage and flow" in Rwanda see, Christopher 
C. Taylor, Milk. Honev and Money: Changing Concepts in Rwandan Healin g (Washington: 
Smithsonian Institution Press, 1992), pp. 11-12. 

ideas of binding and interiority to the physical space of the house. Inferiority is used to define 
close social relations-people of one house or one womb. The construction of the house itself 
serves as a metaphor for "binding" as the circular rings on the roof are bound around the house. 81 
Ritual inscribes boundaries and identity on the land as well as on people's bodies. 

The images of fire to purify and cover with smoke, too, have rich and multivalent meaning 
in ritual. Fire was the civilizing gift of first man, but brought from the outside. The fire or the 
hearth is what constitutes the moral center of a house. Lighting a fire established a household in its 
own right. 82 People fear the purifying, transforming fire of the blacksmith and compare the smithy 
with the womb of the woman. Elders build a fire to see if the time is auspicious for any act, if the 
smoke goes straight up, success will follow. When the new Nata saiga goes to Riyara to gather 
honey for their ceremony they first build a fire to see if the time is right. 83 

Elders use fire ritually in many contexts, apart from the generation-set fire, for purification 
and blessing. Ishenyi elders light the ritual fire, ikoroso, for a single clan (hamate) or a single 
homestead when confronted with the problems of death, sickness or infertility. 84 The prophet 
instructs the elders to make the ikoroso fire in the wilderness from certain kinds of trees found 
there and to add the prescribed medicines. Elders said that the smoke from the fire "covered" the 

81 Weiss, The Making and Unmaking , p. 47. 

Ibid, pp. 29-31, 51-52, describes the Haya ceremony for blessing a new house which 
involves lighting the fire for the first time by the father or a senior agnate. These fires too have 
medicines to protect the house made from specific trees. 

83 Interview with Makuru Nyang'aka, Sochoro Kabati, and Barichera Machage Barichera 
Riyara, 7 March 1 996 (Nata tf). 

84 Interviews with Mang'ombe Morimi, Issenye Iharara, 26 August 1995; Mashauri 
Ng'ana, Issenye, 2 November 1995 (Ishenyi <?). 

whole land for its purification. 85 Rikora leaders perform the ekimweso fire ceremony in Ishenyi to 
bless the land (ekyaro) as a whole. Ikoma call this ritual the shishiga and perform it when 
prescribed by a prophet. 86 In Ngoreme elders make the ikoroso fire from particular trees (esebe, 
omoreto. omorama) that grow at the erisambwa site. 87 

Water from emisambwa springs is also used in the rituals to bless the land. 88 Water was 
the gift of first woman, while fire was the gift of first man. Both are transformative and 
transitional symbols used to mediate the boundaries, the inside/outside dichotomies of the orokoba, 
the womb and the house. A fundamental ritual act in these ceremonies of purification, protection 
and sanctification is the sprinkling of a mixture of contents (water, milk, honey, millet flour, etc.) 
onto the gathered people as a blessing and a prayer for fertility (komusa). The sprinkling covers 
the bodies outside, its boundaries, with the symbolic things of sustenance and life from the inside. 
Elders say that the smoke, like water, covers and protects the participants in these rituals. Weiss 
points out that fire and smoke are often used in binding rites, both to surround a house and to drive 
off malevolent forces. 89 

These symbols operate on all levels from the individual homestead to the territorial ekyaro. 
It is this contextual aspect that makes these symbols so powerful. Testimonies compare the health 

85 Interview with Rugayonga Nyamohega, Mugeta, 27 October 1995 (Ishenyi <f). Each 
homestead extinguishes its fires and a new one is ceremonially lit with medicines and distributed. 
Elders ritually slaughter a goat and sprinkle its stomach contents over the people (komusa), then 
they light the new fire and distribute it to each homestead. 

86 Interviews with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi <f); Bokima 
Giringayi, Mbiso, 26 October 1995 (Ikoma cf). 

87 Interview with Paulo Maitari Nyigana and Ibrahim Mutatiro Kemuhe, Maji Moto, 29 
September 1 995 (Ngoreme if). 

88 Interview with Nyambeho Marangini, Issenye, 7 September 1995 (Ishenyi <f). 

89 Weiss, The Making and Unmaking , pp. 48-50. 

of the homestead to the health of the larger territory. They symbolically equate the enclosure of the 
womb, the homestead fence and the orokoba around the land. These ritual practices create an 
identity embedded in a bounded territory controlled by elders. People use the same symbolic 
language throughout the region to express these concepts of territorial identity. 
The Leaders of the Walk: The Chama 

Peoples throughout the Mara Region call the leaders of the generation- and age-set 
(aba)chama, usually numbering eight members in all, with one top leader whose job it is to "guard 
the land." The secret council of Kuria elders is called the injama. The Ikizu call every member of 
the rikora in power an (omo)chama while the Nata use this term only for the eight chosen leaders. 
An Ikoma elder said that these leaders are responsible for anything that concerns the land—rain, 
disease, peace, war, and hunger. Chama is derived from an old Bantu word (-yama) with wide 
use, usually meaning a group that works together or a council. The Kikuyu elders' lodge and the 
Maasai elders meeting are both called the kiama. Because of the dominant position of Maasai in 
the Rift Valley in the nineteenth century, most obviously manifested in their age-set organization, 
one wonders if the Maasai adopted the word chama/kiama from their Kikuyu Bantu-speaking 
neighbors and then Mara-speakers adopted it from the Maasai in the nineteenth century. 90 

The requirements for and character of the generation- or age-set leader indicate that Mara 
peoples gave the generational principles of growth and peace priority over the warrior ethic most 

90 Interviews with Mashauri Ng'ana, Issenye, 2 November 1995 (Ishenyi tf); and 
Rugayonga Nyamohega, Mugeta, 27 October 1995 (Ishenyi o"). In Kiswahili, kuchunga nchi. 
Among the abachama there were specialists with other names. The Ishenyi ekereri, whose job it 
was to "guard the land," led both the age- and generation-set and had eight abachama from each 
clan to support him." The Nata chose eight abachama for the generation-set in power and eight 
abachama for each age-set, four from each clan moiety and a leader, kangati or omotiro, from 
each. Another Ishenyi rikora official was the omusamu who prepared the meat on sticks over the 
fire for the ceremony. Interviews with Mang'ombe Morimi, Issenye Iharara, 26 August 1995; 
Nyambeho Marangini, Issenye, 7 September 1 995 (Ishenyi <?). 

often associated with age-sets. The generation-set leader's body must be unblemished, without 
scar, sore or disability." His parents, children and wife must be alive and healthy. He must be a 
man of good character and of peace. Once he is chosen, no one was allowed to see him naked, 
even to bathe. He could not touch blood, handle raw meat, drink water other than spring water, or 
take lovers outside his homestead. His wife( wives) observes similar prohibitions. His father must 
be native born, but his mother may be a stranger. 

The age-set leader could never fight and on a raid took up the rear position. Instead of 
carrying a weapon, the kang'ali ("leader" from the verb to walk ahead) carries a long stick, 
orulanya, as the sign of peace. He had only to lay the stick between two people to stop them from 
fighting. Some say that even the Maasai respected the orutanya. 92 The leader of the rikora 
commanded more respect than the leader of the age-set. Far from the warrior hero, he was a man 
of peace, embodying the peace of the land. Elders said that the ongoing work of both the 
generation-set and the age-set, in the east, was to "guard" the land. Although some suggested that 
this task included a military function, in fact, neither the saiga nor the rikora functioned as a 
military regiment. When a raid occurred, every able-bodied man took up the chase. Those in their 
saiga might have gone ahead but did not organize separately from every other man on the chase. 93 

91 The ritual leaders of the Maasai age-set also had to have a "pure" body and come from a 
"pure" family, he was a "man of peace" and stayed in the rear during a raid. Berntsen, 
"Pastoralism, Raiding and Prophets," pp. 79-81 

92 Although this description comes from interviews with Nata, the qualifications for rikora 
leadership were the same in all the groups I interviewed and the long stick a universal symbol of 
leadership, called the ekinara in Ishenyi. Interviews with Mang'oha Machunchuriani, Mbiso, 24 
March 1995; Sochoro Kabati, Nyichoka, 2 June 1995; Kirigiti Ng'orogita, Mbiso, 9 June 1995 
(Nata cf); Mang'ombe Morimi, Issenye Iharara, 26 August 1 995. Rugayonga Nyamohega, 
Mugeta, 27 October 1995 (Ishenyi <r). Berntsen, "Pastoralism, Prophets and Raiding."p. 202, the 
Maasai delegations to see the prophets went unarmed with a long black stick as the sign of peace. 

93 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata <?). 


Besides the (aba)chama of the generation-set and the age-set in each clan territory, the 
Ngoreme have (abajchama for the ekilana or orokoba, sometimes called the (aba)chama of the 
land or ekyaro, and the (abajchama of the rain. These people serve as agents to carry out the 
orders of the elders, prophet or rainmaker, rather than constituting the authority itself. Elders said 
that the most important attribute for the job was the ability to keep the medicines secret. The 
(aba)chama swore an oath of secrecy. Some of these occupational (aba)chama served for one 
year, others for a lifetime. They had to meet the requirements and prohibitions enumerated above 
for other (aba)chama. 94 

The Ngoreme pattern makes explicit the role of the generation-set as "agents of force" 
rather than "controllers of force." When the colonial government began looking for leaders, they 
elevated those in the role of (aba)chama to chiefs. No wonder then that the colonial administrators 
found the Ngoreme chiefs to be powerless and swayed by public opinion. In these cases the most 
visible people carrying out these rituals were not the ones with authority. The community ritually 
endowed the generation and age-set leaders as visible symbols and representatives of community 
consensus. However, it was the authority of the elders, the generation of fathers, that controlled 
their actions. 

Territories and Boundaries: 
The walk of the generation-set ritually created the space of a bounded and enclosed 
territory and the territorial identity of the people contained within those boundaries. To appreciate 
this local concept of territory and territorial identity defined by ritual space, we must find a more 

94 Interviews with Paulo Maitari Nyigana, Maji Moto, 29 September 1995; Isaya Charo 
Wambura, Buchanchari, 22 September 1995; Nsaho Maro, Kenyana, 14 September 1995 
(Ngoreme <f). The abachama of the orokoba passed the protection medicine on behalf of the 
prophet who provided the medicine (omogitana), the abachama of the rain passed the rain 
medicine (omusano) on behalf of the rainmaker (omogemba). 

nuanced understanding of territory than the traditional academic understanding of territory as "the 
ecological locus of use or the political limits of domination and sovereignty." 95 Without chiefs or 
kings, people did not define their territory by centralized political authority. In Chapter 5 we saw 
that the ecological landscapes over which western Serengeti people ranged as hunters, herders and 
farmers covered the whole Serengeti-Mara ecosystem, certainly not defined territorially. One 
could argue that the homestead image also defines a political unit without centralized authority and 
the hills define an interstital ecology. However, these definitions presume the prior assumption of 
how these spaces were constituted in the first place. 

Robert Thornton's definition of territory-as the symbolic differentiation of space and the 
appropriation of that space into a structure of meaning, so that it may be represented as a coherent 
and enduring image-is a more useful place to begin. 96 This definition does away with the false 
dichotomy so often posed by anthropologists and historians between "the social definition of 
territory" (assumed to be the precolonial African model) or the "territorial definition of society" 
(assumed to be the western imperialist model). 97 This dichotomy goes back more than a hundred 
years to the work of early anthropologists, Maine and Morgan, who distinguished between kinship 
and territory as two "mutually inconsistent modes of social organization." The functionalist 
anthropologists used this premise to hypothesize that kinship was logically prior to territory and 
that kinship was the basis for territorial groupings. The point made by Thornton is that people 
must first imagine and create the space of the territory, whether the society is organized on the 

' Thornton, Space. Time and Culture , p. 19. 
'' Ibid. 

97 Christopher Gray, Modernization and Cultural Identity: The Creation of National Space 
in Rural France and Colonial Space i n Rural Gabon . Occasional Paper No. 21 (Bloomington. 
Indiana: Indiana Center on Global Change and World Peace, February 1 994), pp. 37-38. 

basis of proximity or kinship. People imagine and represent boundaries and the territories they 
enclose through the performance of ritual and the practice of everyday routines. 98 

Yet space can be appropriated in a number of ways that do not imply a bounded 
"territory." Chapter 7 demonstrated a conceptualization of space defined by the sites of ancestors, 
emisambwa, who acted as guardians of the land. These sites are often located in places well 
removed from the settlements in which people now live. The core spatial images were those of 
networks radiating out from fixed points on the land. Chapter 6 described the regional pathways of 
knowledge controlled by clans that constituted a radically different appropriation of space reaching 
over vast distances of geographical and social space. Chapter 5 suggested that space was also 
defined in terms of the ecological zones in which people practiced certain economic subsistence 
patterns. Yet these too were not enclosed territories but interstitial, interdependent spaces existing 
within a regional economic system. Ethnic groups, clans, lineages, generation and age-sets each 
appropriated space in a different way and formed their identity in relation to those socially created 

For the western Serengeti the only appropriation of space that can usefully be called a 
"territory" is the ritual creation of the bounded and enclosed space of the ekyaro. Although the 
ekyaro was by definition a clan territory, the social unit used to create a territorial identity was the 
egalitarian generation-set. Kinship was not the basis for a territorial grouping. The concept of an 
enclosed and bounded territory was a different way of appropriating space and used to contain 
different social units that were defined territorially over time. The geographer Robert Sack makes 
a distinction between various social conceptions of space and "territory," which he defines as the 

98 This debate discussed in Thornton, Space. Time and Culture , pp. 8-16. Refers to Sir 
Henry Sumner's work, Ancient Law (1861) and Lewis Henry Morgan's work, Ancient Society 
(1877); See Schoenbrun, A Green Place . Chapters 4-6. 

geographical area over which an individual or group asserts authority and control. 99 I put less 
emphasis on the control of these areas since all appropriations of space involve the authority of a 
particular social group. I prefer to understand territory in the western Serengeti as the 
geographical area appropriated by a social group using the core spatial images of enclosure and 
boundaries. While many kinds of social identities define themselves in relation to the land only a 
territorial identity conceptualizes space in terms of bounded units of land. 
"Tribal" Boundaries and Territory 

Elders today use the core spatial images of an enclosed territory embodied in the 
generation-set rituals of walking the land to define "tribal" boundaries. They employ these older 
concepts of territory to conceive of the newer kinds of ethnic territories that gained prominence in 
the colonial period. These boundaries of ethnic territories enclosed the space of a variety of social 
identities such as clan, lineage, age- and generation-set that now made up the emerging ethnic 

When I asked elders to show me their territorial boundaries, they were quick to respond 
with both the "traditional tribal" and the "colonial" sets of ethnic boundaries. Nata elders said that 
their "traditional tribal" boundaries were the Grumeti River with the Ikoma (east), first the 
Sanchate and then the Tirina River with the lkizu (west), the Morega or the Somoche River with 
the Ngoreme (north) and the Mbalageti River with the Sukuma (south). Elders described the old 
lkoma boundaries as the Orangi River with the Maasai (east), the Grumeti River with the Nata 
(west), Bangwesi mountain with the Ngoreme (north) and the Mbalageti River with the Sukuma 

99 Robert David Sack, Human Territoriality: Its Theory and History (Cambridge: 
Cambridge University Press, 1986), p. 19. 

(south). 100 These boundary recitations follow a formula in which boundaries are defined by 1) the 
ethnic identity of who was on the other side of the boundary, 2) a cardinal direction, and 3) a 
physical feature that marked the boundary such as a river or a mountain. 

The "traditional" boundaries indicated by elders usually followed physical features such as 
rivers, mountains or long stretches of impenetrable bush."" These wilderness tracts of a no-man's- 
land often correspond to the boundaries of tsetse fly bush characterized by John Ford as the 
grenzewilderness. m The colonial files assert that a 35-mile wide belt of tsetse fly bush separated 
the Maasai seasonal grazing areas on the western edge of the Serengeti plains from the farming 
peoples to the west in the Musoma district. During the Ikoma sleeping sickness outbreak, colonial 
officers documented that tsetse infested bush surrounded Ikoma, extending for 30 miles to the 
Ngoreme boundary to the north, and to the south across the Mbalageti River to the Sukuma 
boundary. 103 Given the tradition of bush control discussed in Chapter 5, this evidence may suggest 
that tsetse bush was allowed to remain and demarcate social boundaries. Schoenbrun's work 
shows that from earliest times Lakes Bantu-speakers carved out their farming communities 

100 Interviews with Jackson Mang'oha Maginga, Mbiso, 18 March 1995; Mahiti Kwiro, 
Mchang'oro, 19 January 1996 (Nata tf); Pastor Wilson Shanyangi Machota, Morotonga 12 July 
1995 (Ikoma cf). 

101 See Hans Cory, "Land Tenure in Bukuria, " Tanganyika Notes and Records 23 (1947): 
70-79; Prazak, "Cultural Expressions," p. 97. 

102 Ford, The Role of Trypanosomiases . 

103 From H. J. Lowe, Senior Veterinary Officer, Northern Province to Provincial 
Commissioner in Arusha, 6 March 1931, Land and Mines, Chiefdoms' Boundary Dispute Files, 
North Mara District, 83, 3/127, TNA; J.F. Corson, M.O., Ikoma, 15 April 1927, "Third Note on 
Sleeping Sickness," Extracts of Report by District Veterinary Officer, 1926-29, Provincial 
Administration Monthly Reports, Musoma District, 215/P.C./1/7, TNA. 

between uninhabited buffer zones of tsetse bush wilderness. 104 Because rivers, mountains and 
tsetse bush represented barriers to easy communication, people developed a sense of territorial 
identity within the boundaries of their daily interactions and routines. They perceived the peoples 
on the other side of these wilderness boundaries in stereotypical terms and in relation to idealized 
cardinal directions. 

Oral testimonies show that the concept of the "other" was integrally tied to these 
wilderness boundaries. People create boundaries in relation to what is outside those boundaries 
and their identity in relation to someone else. Oral testimonies most often define communal, now 
ethnic, identity in contrast to the "people of the wilderness" (Nyika) or the Maasai and Asi who live 
outside civilized space. Elders consistently identified Maasai herders and Asi hunters with the 
animals of the wilderness, "who go here and there without a home." The plains and bush areas 
were outside possible homestead space and full of danger, from both wild animals and uncivilized 
people. 105 Civilization meant a ritual relationship with the land that transforms it from wilderness 
to home. Anyone who is outside the boundaries is by definition, the stereotypical "other." 

Oral testimony and ethnography reflect this understanding of boundaries as wilderness 
tracts and portray boundaries as dangerous and liminal places. Elders spoke of boundaries as the 
place where they disposed of polluted things. If an uncircumcised girl became pregnant, she was 
forced to flee over the boundaries or pollute the whole land. In the past the Ikizu generation-set 
leaders took breech babies or those whose top teeth came in first out to the wilderness, over the 

104 Schoenbrun, A Green Place, pp. 126 -128; David L. Schoenbrun, "Social Aspects of 
Agricultural Change between the Great Lakes, AD 500 to 1000," Azania . 29/30 (1994-1995)- 

105 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata tf) says theNdorobo 
walk around like wild animals following the herds. Kiyarata Mzumari, Mariwanda, 8 July 1995, 
says that the plains were a fearsome place because of the danger of Maasai, lions and buffalo. 

boundary with Nata or Zanaki, and left them there. The Zanaki generation-set leaders were 
responsible to care for the boundaries and had to dispose of the polluted Kuria bodies deposited 
across their side of the boundary. Court cases in the colonial files describe dead bodies being 
found on the boundaries, in fact on the colonial boundary stone, between two ethnic groups. I06 

Today villages on the boundaries of two ethnic groups are often the frontier refuge for 
young men in trouble who go there to gain a new identity. These places have bad reputations as 
the home of outlaws and people without respect for traditional authority. While I was in the 
region, the clinic at one of these boundary villages treated many arrow and gunshot wounds from 
encounters in cattle raids and other theft on the boundaries. A German tourist was shot in another 
of these frontier villages. These villages contain people of mixed ethnic background and a 
preponderance of young people, living outside the bounds of normal society and morality. R. E. S. 
Tanner's colonial study of cattle theft in the Musoma District showed that the majority of theft 
cases occurred "along uninhabited district borders." 107 

The boundaries that elders described for the old Nata territory were enormous, enclosing 
approximately 3,600 square kilometers. The colonial census in 1 948 lists the population of Nata at 
only 1,519, giving a density of less than a half person per square kilometer. Even if one would add 
in the Ishenyi (2,428) and lkoma (4,474) populations of the time the density would still only be a 

106 Interviews with Zamberi Masambwe, Gisuge Chabwasi, Mariwanda, 22 June 1995 
(Ikizu (f); Zabron Kisubundo Nyamamera and Makang'a Magigi, Bisarye, 9 November 1 995 
(Zanaki <?). Criminal Case No. 92 of 1 93 1 , Villagers of Watende living in Bunjari, 9 September 
1 93 1 , Native Affairs, Collective Punishments and Prosecution of Chiefs, 1 926-3 1 , 246/P C /3/4 

107 R. E. S. Tanner, "Cattle Theft in Musoma 1958-59," Tanzania Notes and Records . 65 
(March 1966): 31-32. 

little more than two people per square kilometer.' 08 Using these figures to estimate precolonial 
populations is nearly useless of course because of the huge loss of population and migration during 
the late nineteenth century. Yet it does give us the sense of the vast tracts of land and small 
number of people in question. Even today the population density of Serengeti District is only 10.2 
people per square kilometer. "" 

Obviously the Nata definition of its territory depended neither on "the ecological locus of 
use" nor "the political limits of domination and sovereignty."" What made it "Nata" territory was 
that elders used the older concepts of bounded territory to enclose a space that included the 
variously conceived spaces of a number of social groups that now made up Nata, filling the need 
for a defined "traditional tribal territory" in the colonial context. These boundaries had to be 
extensive to include all of the sites of the emisambwa still propitiated by Nata lineages, the 
boundaries of the ekyaro defined by the walk of the rikora and to exclude those who were different 
across the dangerous wilderness boundaries. This Nata territory includes the land used to hunt, to 
gather arrow poison or to trade with the Asi or Tatoga. Although the Nata never occupied all of 
this land at once by habitation, they claim this land as Nata territory by claiming all of their past 
interactions within this landscape. The Nata generation-set never walked around these extensive 
boundaries, it was confined to the area in which people lived. However, Nata elders use the same 

108 East African Population Census, 1948, African Population of Musoma District, East 
Africa Statistical Department, Nairobi, 1 October 1948, African Population by Chiefdom, 1948, 
Secretariat Files, 40641, TNA. In the Native Affairs Census of 1 926 the Ikoma Federation was 
listed with a total population of 1 4,799-Nata 1,923, Ikoma 8,664 and Issenye 4,212. In Native 
Affairs Census 1926-29, 246/P.C 73/21, TNA. 

109 The population density of Serengeti District is the lowest in the Mara Region but its 
growth rate at 8.3 is the highest in the region, reflecting a large in migration for open farmland and 
mining. Mara Regio nal Statistical Abstract 1 993 (Dar es Salaam: President's Office, Planning 
Commission. Bureau of Statistics, June 1995), pp. 2, 12. 

"° Thornton, Space. Time and Culture , p. 19. 

core spatial images of enclosure employed by the walk of the generation-set to appropriate this 
larger landscape with which they have formed a relationship to create a "tribal" territory in the 
colonial context. 

Because these new colonial territories enclosed the space of various social identities that 
had existed previously there was considerable overlap and boundaries were disputed. The same 
lineages and clans were now part of different ethnic groups, but maintained claims to the same 
ancestral places. Age-set territories of what were now different ethnic groups overlapped. A 
Ngoreme elder who lived on the present boundary with Nata said that Ngoreme used to include all 
of what is today Ishenyi and Nata, and much of Ikizu, bordering at the Mara River to the north 
with the Kuria and to the east with the Maasai. A Nata elder however said that Nata used to border 
the Kuria at the Mara River and with the Ngoreme at the Ikorongo mountains. The overlap in 
territory of these two statements is enormous. 1 " Yet because these representations of past 
boundaries refer to a combination of various concepts of space operational at different times that 
were not exclusive, they overlapped in various ways. Both accounts were valid depending on how 
boundaries were defined. Other boundaries have become rigid and highly contested. Some versions 
of the emergence story separate the territory of Ikizu between the territory of first man, Isamongo 
(Ikitang'anyi), and the territory of first woman, Nyakinywa (Buraze). Elders recited the 
boundaries of those two halves of Ikizu in great detail, naming the landmarks (villages, springs, 
trees, rocks, hills) around the perimeter of each." 2 

'" Interviews with Mwita Magige, Mosongo, 9 September 1995 (Ngoreme <f); Mayani 
Magoto, Bugerera, 18 February 1995 (Nata cf). 

112 Interviews with Kiyarata Mzumari, Jackson Witari, Musa Matabarwa, Webiro Zeze 
Mariwanda, 8 July 1995 (Ikizu d-). 


During the colonial period these older concepts of territory easily accommodated and used 
new ideas about exclusive "tribal" territory. The elders who told the story about the Kuria moves 
to Mugumu said that the injama passed an orokoba around the boundaries of the new land to 
ensure possession. These boundaries did not coincide with the colonial tax boundaries and only the 
Kuria elders on the secret council of the injama knew where they lay." 3 Yet the Kuria in Mugumu 
also recognized and defended the administrative boundaries that defined their right to live in this 
territory. Many elders across the region worked with colonial officials to define the "tribal" 
boundaries. Once the new boundaries were set on administrative maps the elders defended and 
contested these boundaries by drawing on the authority of "tradition." A Sizaki protest to the 
Governor said, "there are no exceptions to these boundaries since everyone recognized them since 
the almighty Creator made all things on this earth, before the Germans came to Africa."" 4 

What was different about these two concepts of territory was that western Serengeti 
territories were flexible units that were defined anew each time the land was ritually walked over 
while the colonial concept of territory was a fixed unit of land. In the western Serengeti, territories, 
both the boundaries themselves and the social groups contained within the boundaries, shifted over 
time while colonial territories assumed the occupation of a "tribe." Recent scholarship has taken 
this contrast in concepts of territory to mean that if precolonial African territorial boundaries were 
flexible, situational, multiple and shifting, they were not meaningful or enduring indications of 
identity. They theorized that if people were constantly moving they could not form an attachment 
to the land. Kopy toff describes this African "attitude" as a "relative indifference to rootedness in 

113 Interview with Pastor Philemon Mbota, Mugumu Matare, 27 January 1995 (Kuria cf), 
the son of the Prophet Mzee Mbota who came from Nyabasi to Mugumu. 

" 4 From the Askari wa Jeshi la Kikozi cha 46-36-26TT KAR, South East Asia to Bwana 
Gavana DSM. December 1 1944, Native Chiefs, Musoma, p. 23, Secretariat Files 29626, TNA. 

physical space, together with an indifference to a permanent attachment to a place . . . African 
space is, above all, social space."" 5 Yet the western Serengeti rituals of walking the land show 
that place did matter and that a people's identity was deeply tied to the ways in which they had 
appropriated and created the space around them, whether territorial or not. People understood the 
ekyaro as the land that they ritually controlled, intimately linked to their own health and well 
being. However, this space was not fixed and immutable through time nor did it always enclose 
the same social unit. 

The Germans began transforming local ideas about territory into western concepts of land 
as exclusive, measured and fixed units through the process of selling land to the few private 
entrepreneurs and missions that wanted to obtain land. The Seventh Day Adventists bought land 
from the Germans in many places in the region. Each transaction involved an elaborate series of 
steps, including a visit to the site to ensure that no one held claims against it and that compensation 
had been provided for natives. The boundaries were marked with stones (with bottles buried under 
the stones) and maps were drawn so that the land transfer could be registrated in court." 6 The 
Germans took this process seriously as foundational to the creation of civilization. Rather than 
randomly grabbing land for the European newcomers, the German officers made doubly sure of its 
"ownership" and use before granting it as Crown land. 

1,5 Kopytoff, "The Internal African Frontier," p. 22. Christopher Gray, "The Disappearing 
District: The Decline of Precolonial Space in Southern Gabon 1850-1940," a paper presented at 
the American Historical Association, January 10, 1998, pp. 18, also argues that district identity 
was "almost always overriden by clan and lineage identity" and that "territory was defined socially 
as there was no real ownership tie to the land per se and territoriality was merely one strategy 
among others to enhance the wealth of one's lineage or clan." 

1 " All land was regarded as Crown land unless individuals could claim private ownership 
or current occupation. The idea of a plot of farmland permanently cultivated was a new concept. 
Landkommissions Adventisten Mission Magita, 1909-1912, G 15/499; SDA Busegwe File. 1909- 
1913, G 45/34 LR, TNA. Special thanks to Simon Heck who provided German translation 
assistance in the archives. 

Disputes over Colonial Boundaries and Colonial Maps 

The coexistence of these different ideas of territory and the ability of local people to use 
the definition of boundaries to enclose various kinds of social units is evident in the inordinate 
amount of paperwork in the colonial files concerning disputes over "tribal" territory. To settle 
these disputes colonial officers went to the land in question and asked the elders on each side of the 
dispute to walk over the boundaries, indicating the traditional boundary markers. They drew 
sketch maps and placed them in the district files for future reference. Where no obvious and 
agreed upon rivers or hills marked a boundary, the conflict often revolved around earlier colonial 
definitions of boundaries. The elders argued over which tree "Bwana Baker" set as the boundary 
marker and who had been present at the original marking. 

The determination of territorial boundaries was most important to the colonial chiefs 
because they could collect taxes and demand labor only from the people within their territory. 
Thus, chiefs initiated the boundary disputes and the colonial officers experienced continual 
frustration from these obvious efforts at "land grabbing." One colonial officer reported, "there are 
to my knowledge seventy-four latent boundary disputes in Musoma District, nearly all of them 
fostered and exasperated by the chiefs ... not more than six in which the people evince any 
interest at all." This report claimed that the inhabitants did not care what the country was named 
since they intermarried, lived and farmed next to each other across "sub-tribal" boundaries." 7 

The people within these territories, however, seemed to have had their own ideas about 
boundaries and acted on them to the great displeasure of colonial officers and chiefs alike." 8 

117 Annual Report 1933, Musoma District, Native Affairs Section, p.53, Annual Reports, 
Native Affairs Section, Lake Province, 215/924/2, TNA. 

118 Because of the great loss of South Mara files in the National Archives, unfortunately, 
most of these examples of colonial boundary disputes come from Kuria and Luo (recently 
assimilated from Bantu Suguti-speakers) related peoples in North Mara. I do not believe, however, 

Chiefs often complained that people who moved into a new territory continued to pay taxes to their 
former chiefs, of their own ethnic origin. When questioned in court the defendants said that 
although they recognized that this was legally the territory of another chief, in fact, it was 
uninhabited wilderness area and that no one representing that chief was anywhere around. They 
therefore felt free to give their tax to their former chief." 9 These people acted on the old concept 
that the wilderness boundary areas between groups constituted a zone of a no-man's-land that 
settlers could open, make into habitable and civilized space and claim for their own group. 

Within the older concept of territory it only made sense that these settlers on the wilderness 
frontier would pay their taxes or offerings of patronage to the "big man" with whom they had 
already established ties of reciprocity. In the court case cited above, the migrants from Buturi 
might have had debts of patronage to the Chief of Buturi that they reciprocated by the payment of 
taxes. They would have had no former connections to the Chief of Shirati and thus no reason to 
pay the tax to him. These testimonies seem quite close to the older pattern of expansion in which 
young people moved out into the wilderness frontiers while maintaining ties to their home 
communities. 120 In each of these court cases people refused to pay their taxes to the chief legally in 
control of the territory because they defined the place where they settled as a wilderness boundary 

that the issues would have been significantly different in South Mara, had I the documents to prove 

'" From Omukama of Buturi in Buhacha to the ADO in Tarime 24 June 1935; From 
ADO in North Mara to DO Musoma, 3 November 1932; Native Court Testimony, Girango A 
Court, 3 1 March 1 938; From the ADO Tarime to DO Musoma, 4 April 1 940; From Marwa Igina 
Usimbiti Kumuge 14 March 1948; Land and Mines, Chiefdoms' Boundary Dispute Files, North 
Mara District, 83/3/127, TNA. 

120 See Kopytoff, "The Internal African Frontier," The African Frontier , pp. 7-23, on 
frontier movements. 

area. Colonial reports tell of people deliberately building on the boundaries of their chiefdoms or in 
the wilderness areas of neighboring chiefdoms in order then to make territorial claims. 121 

The colonial boundary dispute files also demonstrate the extent to which people still 
equated their relationship to the land with their identity as a people. In a long letter addressed to 
the Governor, the Kiseru Area Council complained that the District Officer arbitrarily changed the 
boundaries of their chiefdom, leaving 170 Kiseru taxpayers who occupied the "lands of their 
ancestors," under the authority of "another tribe which is different from ours." The Council 
appealed for the return of "the children, wives and elders who were taken away from us." In 
another telling phrase the chief said that his people in the disputed area had been "given" another 
"tribe" which was unlike their own. 122 The council expressed its outrage against the separation of 
territorial residence from its moorings in communal identity. It was impossible to conceive of 
living on your ancestral land but owing allegiance to "others" outside. The council characterized 
these people as having been "taken away" or "given" a new "tribe" because of the change in 
authority over the territory in which they resided. Later court evidence demonstrated that the 
customs of these peoples in question were not so different and that they had been intermarrying for 
quite some time. The colonial officer concluded that the Kiseru Council elders were leading the 
people in this agitation because they would stand to lose most from the deal. I23 However, the 
existence of multiple boundaries defining multiple sets of relationships allowed these people both to 

121 From Wakibara Nyamuhika from Suba Kukabwa 8 May 1947, Land and Mines, 
Chiefdoms' Boundary Dispute Files, North Mara District, 83/3/127, TNA. 

122 From the Kiseru Area Council to the Chief Secretary and Governor of Tanganyika 
Territory with copies to the DC and PC, 24 November 1 949, Land and Mines, Chiefdoms' 
Boundary Dispute Files, North Mara District, 83/3/127, TNA. 

123 Report from DC North Mara to PC Mwanza 1 February 1950, Land and Mines, 
Chiefdoms' Boundary Dispute Files, North Mara District, 83/3/127. TNA. 


express outrage in being transferred to a "foreign tribe" while still living in their ancestral land and 
also intermarry and share a common culture with those across the boundaries. 

All these examples of colonial interaction demonstrate that earlier ideas about territory and 
the need for maintaining the ritual health of the land operated side by side with newer ideas about 
territory. At times, the colonial officers and the chiefs used the same language but with very 
different meanings. At other times, the chiefs and elders appropriated the colonial meaning of 
territory for use in local struggles over prestige and authority. They seemed equally at ease 
walking out the boundaries of the "tribal" territory with the District Officer and walking the 
medicines of the orokoba over the land with the generation-set. The flexible nature of the ekyaro 
adapted itself to colonial definitions without negating its previous and still situationally important 
Nverere's Mwenge: The National Space of the Ekvaro 

People in the Mara Region have extended the concept of the ekyaro as the ritually 
maintained territory of a people to construct a national identity in present day Tanzania. 
Mwalimu Julius K. Nyerere, the first president of independent Tanganyika and later Tanzania, 
comes from the Mara Region, Zanaki, Butiama. He is the son of the second colonial chief of 
Butiama, Nyerere Burito (ruled 1912-1942), and the half-brother of Edward Wanzagi Nyerere, the 
last colonial chief of Butiama and of the reunified Zanaki Federation (ruled 1952-61). I24 In spite of 
his training and later his baptism as a secondary school student in the Catholic Church, Nyerere 
grew up with the concepts of the ritual health of the land as a member of the generation-set. 

124 Biographical accounts of Nyerere include, Judith Listowel, The Making of Tanganyika 
(Chatto and Windus, London, 1965); Kemal Mustafa, "The Development of Ujamaa in Musoma: 
A Case Study of Butiama Ujamaa Village" (M.A. Thesis, University of Dar es Salaam, 1975); 
Mustafa, "The Concept of Authority"; William Edgett Smith, Nyerere of Tanzania. (Victor 
Gollancz: London, 1973). 

During my stay in the region, 1 heard many stories about the auspicious signs surrounding 
Mwalimu Nyerere's birth, his inheritance of prophetic powers through his lineage and his use of 
these in the politics of state. Given his background in generational authority it is no wonder that 
Julius Nyerere was one of the first and only African presidents to retire, which he did using the 
Zanaki word for the retirement of the generation-set, kunyatuka. 

The two most important symbols of national unity instituted by Mwalimu Nyerere are the 
torch on top of Mount Kilimanjaro and annual "walk"of the torch or mwenge around the nation of 
Tanzania. Each year since the Arusha Declaration, the torch, mwenge, is set on the back of a 
truck and escorted throughout the nation by soldiers and political dignitaries of each region through 
which it passes. All other vehicles must clear the road and wait as the long line of new white land 
cruisers and army vehicles passes. Disrespect for this convoy can incur serious consequences as a 
national offense. Different routes are chosen each year so that the torch passes through as many of 
the remote rural areas as possible in a few years time. The mwenge always returns to the place 
that it started at the end of its "circular" tour. Planning the passing of the torch begins months 
prior to its arrival by the collection of local "donations," the preparation of a feast, arrangements 
for songs and dances by local school children, speeches by political officials and the attendance of 
crowds along the roads to cheer it on. The torch is a symbol of national unity and its journey 
around the nation an attempt at building new symbols of national identity. 

The issue of the mwenge figured highly in the political debate surrounding the presidential 
elections in 1995. Opposition candidates called it an example of the misappropriation of funds 
better spent on development. Others denounced it as a scheme by local dignitaries for redirecting 
funds. Observers in other regions told me that many people found the whole institution less than 
convincing. Yet people in the Mara Region, by contrast, seemed to understand and support the 
mwenge institution. 1 heard the mwenge discussed as an orokoba for the nation. Some claimed 

that the mwenge contained medicines that were spread across the land for protection and healing. 
It was Nyerere's medicine through his prophetic line. The torch's fire symbolized the spread of the 
new fire to each homestead throughout the ekyaro. One local interpretation said that just as the 
new generation extinguished the old fire of the previous generation, so the new fire protects the 
authority of those who rule. The medicine of the mwenge "puts out" the opposition to the authority 
of the nation. People said that these things had to be kept secret, just as the rituals of the walk of 
the generation-set were secret, to protect the nation. All of this is highly suggestive and 
unsubstantiated since I was not able to discuss motivation and intent with Nyerere himself. It is 
nevertheless significant that, whether Nyerere understood the mwenge as an orokoba or not, many 
people in the Mara Region do. It is an indication that these older ideas about land, healing and 
territory are still functioning alongside newer ideas about administrative boundaries and 

All these stories, from the walk of the rikora in the nineteenth century to the ride of the 
mwenge in the twentieth century, indicate a 'symbolic differentiation of space and the appropriation 
of that space into a structure of meaning' that creates a communal identity in reference to the land. 
Yet this appropriation of space is different from the others explored in previous chapters because 
its core spatial images are those of enclosure and boundaries, defined here as a "territory." 
Western Serengeti territory has encompassed the clan ekyaro, the age-set cycle and the ethnic 
group depending on the historical context. Lineages, clans, generation-sets and age-sets each 
define their identity in relation to the land in a different way. During the colonial years some of 
these different definitions were combined to enclose the new territorial units of the ethnic group. 
The concept of territory was used to create "tribal" units in the colonial years but in the process 
assimilated the colonial idea of fixed territorial units. 


Each kind of appropriation of space, related to a different set of social identities, also 
implied the locus of social authority used to control it. The "fathers" or elders of the generation-set 
in power had authority over the territory as representatives of communal consensus. Even though 
these rituals are seldom, if ever, performed anymore, the narratives about the walk of the rikora 
told by the elders reinforces their dwindling communal authority and reasserts them as the 
legitimate carriers of "tradition" and the health of the land. 

The stories about the generation-set and its walk, the healing of the land and the definition 
of boundaries all belong to this middle period of history that identifies repetitive social processes 
that are subject to change but at an incremental level. Although the generation no longer walks, the 
ideas about healing the land remain and people have found new ways to express these concerns in 
the new contexts of national life. The idiom of timelessness in which they discuss these issues 
obscures the fact that these are central issues of land and territory that are highly contested. 

By exploring these ideas of territory through the rituals of the generation-set. we can 
imagine how these same concepts were applied to the clan-based ekyaro territory of the nineteenth 
century. In the context of plentiful land but few people, settlements attracted new settlers and 
integrated them into an inclusive territorial identity through the reenactment of these rituals. 
Generation-sets were not corporate, property-owning groups, but they embodied communal 
consensus and an identity with the land that was necessary for the prosperity of those who lived as 

The narrative now turns to the period of historical remembrances of the late nineteenth 
century in which the issues of land and territory remain an important focus for understanding the 
dislocations of that period. The various forms of precolonial social identity discussed up to this 
point form the basis for analysis of late nineteenth century transformations. 






The generation-set maintained a ritual relationship to the land, bringing protection and 
healing to the territories of local lineage-based settlements in the nineteenth century. When this 
protection failed how did people cope in the face of large-scale environmental and human disaster? 
How did western Serengeti people imagine their past to cope with the present in the context of the 
loss of loved ones due to sickness and hunger, the loss of a significant part of the next generation 
due to widespread infertility, the loss of the resources for survival due to the breakdown of a 
regional economy, the loss of control over boundaries due to the encroachment of Maasai 
hegemony, and the loss of control over the environment due to the spread of bush? The disasters 
suffered by people of the western Serengeti were similar to crises across East Africa during this 
period. While describing the disasters in detail because local people felt them so horribly, the 
emphasis here is on the creative response of peoples of the western Serengeti to these problems. ' 

This chapter analyzes the historical memories of the "generation of disasters" (C. 1870- 
1895) and how that generation responded to these crises by reworking existing social relationships 
and patterns of settlement. In particular, they reconfigured the older institution of age- 
organization into territorially based units that provided the means for both the unification and the 
enlargement of scale required for the later emergence of ethnic identities. These larger scale 

' For an account of the environmental disasters in Tanzania see: Iliffe, A Modern History . 
Chapter 5; Kjekshus, Ecology Control : Giblin, "Famine, Political Authority and Foreign Capital." 
For a critique of the "degradation narrative" see McCann, "Introduction," An Environmental 
History : and Leach and Mearns, The Lie of the Land . 


territories combined settlements organized around the lineage idiom and linked them together in 
loose networks of reciprocity. It was during the period of disasters preceding the colonial era that 
western Serengeti people formulated the basic ethnic identities that exist today within the space of 
age-set cycle territories, although the advent of colonial rule solidified ethnic identity into 
territorially based administrative units. 

Transformations in identity came as a response to severe dislocations of population and of 
economic subsistence patterns. Numerous refugees from farming communities fled the region, 
moving as far south as Sukuma. At the same time incoming Maasai, who gained dominance over 
the regional system of economic interaction by controlling pastoral resources and developing units 
of social organization through which to expand that control, drove a large portion of the Tatoga 
herders as far south as Tabora. The Maasai raided western Serengeti farmers to gain more 
livestock, particularly after cattle disease swept through their herds between 1880 and 1890. The 
Asi hunters who had been so important in the previous set of regional relations gradually became 
clients of the dominant Maasai and moved farther east as the farmers moved farther west. These 
events shattered the previously existing regional economic system and left the farmers particularly 
vulnerable to famine and epidemics of disease introduced from outside that swept over the land. 
The reorganization of lineage-based settlements into age-set territories, however, enabled people to 
reformulate the interdependent economic strategies of woodlands, hills, and grasslands. 

Chapter 3 described the main events and transformations of the period of disasters and the 
effects of those changes on oral tradition. This chapter looks more closely at how these 
transformations in social identity came about and how people made creative use of the older 
generative principles of social organization to effect these changes. While Chapter 3 portrays 
western Serengeti people as victims of famine, epidemic, and raid, this chapter highlights their 

agency in forming communities that not only coped with the disasters but forged strategies that led 
to prosperity in the next generation. 

In this chapter I first show how the ethnic groups of the eastern part of this region (Ikoma, 
Ishenyi, Nata and Ngoreme) reformulated generation-sets into age-set cycles more responsive to 
the need for mobilizing young men in raids. Reorganized age-sets resulted in a new way of 
calculating time and in a new way of organizing territorial settlements. These changes, in turn, 
indicate a massive transformation of social identity at the time of disasters. Through an 
exploration of age-set lists, I demonstrate how western Serengeti people maintained continuity with 
the past in the face of radical change. Elders explicitly tell about the generation that was divided 
into territorial age-sets. At the same time, their age-set lists project the age-set cycle names of the 
late nineteenth century back in time before the disasters. I look next at the concentration and 
fortification of settlements as a response to raids and the need for boundary formation in times of 
societal stress. 

The section on western Serengeti response to Maasai hegemony details the effects of 
raiding and also peaceful interaction with the Maasai. As a result of Maasai raiding, Sonjo 
refugees settled in the western Serengeti and brought with them the direct experience of living 
closely with the Maasai. Origin stories emphasized Sonjo ancestors because of their value as 
cultural translators during the period of disasters. Western Serengeti people both resisted and 
accomodated the imposition of a Maasai regional hegemony. They formed new age-sets out of 
admiration for the powerful Maasai warrior ethos but maintained the regional continuity of the 
older generation-sets. The territorialization of age-sets allowed western Serengeti farmers to 
respond to the disasters by spreading out both their risks and their opportunities when the former 
regional system of economic interdependence broke down. The larger-scale identity of cycling age- 
set territories became the basis for ethnic identity in the early colonial years. 

Age-Set Reorganization 

The most important transformations of this period resulted from changes in age-set 
structure, particularly among the most easterly groups of Nata, Ishenyi, Ikoma and Ngoreme. The 
preexisting linear age-set system (subordinate to a cycling generation-set system) was changed into 
a cycling age-set system that largely took over the functions previously assigned to the generation- 
set. I demonstrate these transformations by showing how these changes are represented in the lists 
of age-sets provided by elders today. Although elders explicitly tell stories about the 
reorganization of generation-sets into age-sets, their age-set lists project the continuity of cycling 
age-set names back before the period of disasters. This chapter rests on the basic knowledge of 
age-organization laid out in the last chapter. 
Time: Age Organization and Dating 

As I showed in Chapter 3, the effects of the diasters were so widespread and severe that 
remembered history begins here. Men living today heard the stories of the disasters from their 
grandfathers and grandmothers who lived through them. Although western Serengeti narrators 
now tell all history as ethnic history, it is only beginning in the "generation of disasters" that one 
can confidently speak about the existence of ethnic groups called Nata, Ikoma or Ishenyi. Western 
Serengeti peoples used age-set territories to formulate ethnic group identity and ethnic histories 
during this period of stress. 

Beginning with this generation, the historian can offer a sequence of relative dates based 
on age- and generation-set lists. African historians have long recognized that they can use lists of 
age-set names to establish a relative historical chronology. Elders in societies that use age 
organization can list the names of successive age-groups working back from the present to the past 


and using a consistent interval of years between each group (eight years in this case). 2 However, 
these lists express ideological concerns as much as they narrative relative chronology. Even after a 
careful analysis of the historical development of these lists, they can provide only a relative 
sequence of events.' The lists that I collected among five Bantu-speaking groups in this area were 
fairly consistent, at least back to mid-nineteenth century. Narrators place the stories from mid- 
nineteenth century on by reference to a particular age-set, thus making it possible to order events in 
a relative chronology. However, elders often elide events that happened "a very long time ago," or 
during the "generation of settlement," into the time of the first remembered age- or generation-set, 
around mid-nineteenth century. As discussed in the last two chapters, the telescoping device allows 
narrators to condense a long period of time into the memories of one generation occurring in the 
middle time frame of indigenous history and acts as a bridge between the older and more recent 
time periods. 

When narrating age-set lists, elders seamlessly weave the abrupt changes in social 
organization that clearly occurred during the period of disasters into the ongoing flow of time 
through the birth and death of generations. The encirclement and boundary formation necessary 
for healing the land in a time of stress depended on continuity with the past to ensure that the 
medicines of the land and the propitiation of ancestors as guardians of the land would be effective. 
Elders said that the age-set names recurred in a cycle every third generation, a claim that enabled 

2 Jacobs, "A Chronology," pp. 10-31. For a critique of dating by this method see Berntsen, 
"Pastoralism, Raiding and Prophets," pp. 83-93. See also John Lamphear, The Traditional History 
of the Jie of Uganda (Oxford: Clarendon Press, 1976), pp. 17-60. 

' For a Maasai construction of time, "spiralling upwards with age," see Paul Spencer, 
"Becoming Maasai, Being in Time," in Being Maasai: Ethnicity and Identity in East Africa , eds. 
Thomas Spear and Richard Waller (Athens: Ohio University Press, 1993), pp. 140-156. See also 
the critique of king lists for chronology, Henige, "Oral Tradition and Chronology," pp. 371-389; 
Wrigley. Kingship and State . Chapter 2. 

them to project the names back infinitely into the past. This device allowed people to imagine the 
ongoing continuity of age-sets as an inheritance from the past. The lists of particular age-set 
names which they provided, however, did not usually go any earlier than mid-nineteenth century, 
and elders specifically discussed the reorganization in age-sets necessitated by the disasters. 4 The 
remembered age-sets before the disasters (C. 1 850-1 890) were usually generation-set names or 
linear age-set names assimilated into an age-cycle pattern. Elders never explained how the names 
from age-set cycles that were, by their own account, invented in response to the disasters could be 
applied to age-sets in the eighteenth century or even before. They were willing to overlook these 
inconsistencies in order to establish a necessary link with the past by placing these fundamental 
changes in an understanding of continuous and repetitive time. 

The sense of continuous time and generational authority over its orderly passage permeates 
idealized description of age-set organization. In theory, each age-set or saiga would "rule" for 
eight years before the next group would take over. While the senior cycle "ruled" the next waited 
to "enter their saiga," having already been initiated during the last cycle The third and most junior 
cycle began circumcision of a new set, culminating in the mass circumcision ceremony of the 
kigori. The practical functioning of age-sets was much more flexible than its rules suggest, and 
elders with authority over the ceremonies that promoted new age-sets could delay them at will in 
order to remain in power longer. Other acts could prolong the interval between age-sets for more 

4 Some of the few other works on the Mara Region have taken these names back to the 
beginning of the 18th century. See Anacleti, "Pastoralism and Development," pp. 14-15. 1 would 
argue that there is no evidence to support these longer lists, some of which anachronistically use 
event-oriented praise names (example, abaSanduka) as cycling names. Their dates were also based 
on initiation at circumcision, while local understandings date each saiga, or age-set to eight years 
after circumcision when they are said to "enter, or step on, their age-set" (gutaacha asaiga). 

than eight years. 5 The incoming saiga fought for its position in a mock battle with sticks; if it lost, 
the incoming saiga waited another year to take power. If the prophet whom the elders consulted 
before the ceremony said that the time was not auspicious, elder could delay gutaacha asaiga. 1 ' 
Because western Serengeti people synchronized the ceremonial cycle of age-sets regionally, a delay 
in the ceremonies of one group would delay the others. Elders said that the more easterly Ikoma 
saiga was always ahead of the equivalent Nata saiga by a few years.' The system also depended 
on a cycle of other rituals necessary for the health of the new saiga. The installation of the new 
saiga must take place before the rikora, or generation-set, could begin its walk to cool the land, 
which must itself be completed before the kigori circumcision ceremonies could take place. 
Because of all these contingencies, age-set lists can provide only a relative and idealized, as 
opposed to an absolute, chronology, and only for the period of the late nineteenth century onward.. 
Age-set Lists and Regional Chronology 

A comparison of age- and generation-set lists throughout the region demonstrates how the 
most easterly peoples reformulated generation-sets into age-set cycles and projected these new 
names back in time. They did this by including what were generation-set names and praise names 
of linear-age-sets from an earlier period into the list of the new cycling age-sets. 

Because age- and generation-sets throughout the region often used the same names at 
approximately the same time, correlating the age- and generation-set lists of different ethnic groups 
is possible. From these age- and generation-set lists I have reconstructed a relative chronology of 

5 Eight is the most perfect number in this area and it is not surprising that it is also used for 
the ideal saiga rule. 

6 Interview with Megassa Mokiri, Motokeri, 4 March 1995 (Nata tf). One Nata saiga was 
delayed when a raid was imminent to keep the more experienced men in power. 

7 Comparison to Kjerland's Kuria list shows that they were also "ahead" of Nata age-sets. 
Kjerland, "Cattle Breed," Appendix. 

successive age-sets that covers the entire region. [See Figure 9-1 : Chronology of Generations.] 
Note that the most easterly ethnic groups (Nata, Ikoma and Ishenyi) divide the age-sets into three 
cycles: Busaai, Bongirate, and Borumarancha Each of these has three recurring names (not so 
obvious on the chart because the names mainly recur in the later colonial period), making the age- 
set name of a man the same as his great-grandfather, if they were both of the same cycle. The 
Busaai age-set comes into their saiga (gutaacha asaiga) first, followed by Bongirate, then 
Borumarancha, and finally returning to Busaai to start the cycle again, with each cycle "ruling" for 
eight years. The first age-sets in this new cycling system are those with the same names as the 
larger cycle, the Busaai, Bongirate and Borumarancha, dating to approximately 1 870- 1 895 ("the 
generation of disasters"). 

Groups to the north and west, Ngoreme, Kuria, Ikizu, and Zanaki still follow the older 
generation-set system that consists of two generation cycles, each with four recurring names. The 
age-set system for these groups (groups who continued to give primary importance to the 
generation-set) remained the linear type, with a unique praise-name chosen for each new group. 
The cycling age-set names of Ikoma, Nata and Ishenyi developed from the cycling generation-set 
system and thus can be correlated with generation-set names in other ethnic groups. 

The correlation of age- and generation-set names across ethnic boundaries provides 
evidence for the change from a linear age-set system to a cycling age-set system among the Ikoma, 
Nata and Ishenyi. Although Ikoma, Nata and Ishenyi elders list the age-set names before the 
disasters within the three age-set cycles, these names often correspond with linear age-set names 
among groups that kept the linear age-set system (Maase, Ngirabhe). Even after the Ikoma, 
Ishenyi and Nata adopted the cycling age-set system they continued to use the regional linear age- 
set names as praise names (Sanduka, Romore, Kambuni). Other age-set names, listed before the 
institution of age-set cycles, are generation-set names (Maina, Saai, Nyange) assimilated into the 


Chronology of Generations, Western Serengeti, Tanzania 


Cycling Age-Sets (Saiga) of 
Ikoma, Nata and Ishenyi 

Cycling Generation-Sets 
(Rikora) of Ngoreme, Kuria 

Ikizu and Zanaki 

Bongirate ziorumarancha 



Praise Names 
Kuria Nyabasi 

(c. 1820) 



(c. 1828) 

abaGamunyere :-abaGinj ~- — 

-1 . Gesetwi™ 



(c. 1836) 


(0. 1844) 


2 Kehanga 
3 Gesambiso 

the Generation; 
of Settlement 

(c 1845) 



(c. 1852)- 




ic 186C 

4 Ngibabe 

5 Machare 
(1865, 67 : 69> 

(c 1868) 


The Generation 

of Disasters 


(c. 1876) 



(c 1884) 





6 Getiira 

7. Maase 

8. Nginogo 

TWGeneratibn- ' 

of Opportunity' 

abaKihocha- ' 
(c 1900J' 

- -9.- Komore-" . 

.-_{T893, 9 

10 Nginaro 



-t^M 1907^-09^:: 

Kuria Nyabasi list from Kirsten Alsaker Kjerland, "Cattle Breed, Shillings Don't: the Belated 
Incorporation of the abaKuria into Modern Kenya" (Ph.D Dissertation, University of Bergen, I995), 
Appendix, using dates at the time of circumcision. 
Dates of Cycling Age-Sets at the time of entering their Saiga, ideally eight years after circumcision. 

Figure 9-1 : Chronology of Generations 

three age-set cycles. It seems that elders took the linear age-set and cycling generation-set names 
before the disasters and represented them as part of the three age-set cycles to create a sense of 

Narrators of cycling age-set lists also included regional praise names of linear age-sets 
after the disasters. Both cycling and linear age-sets tended to choose the same praise names 
throughout the region. In the cycling system each age-set has two names. The first name, taken at 
circumcision, is a praise-name referring to events of the time. The second name, taken at the 
assumption of age-set "rule" eight years after circumcision, is a "traditional" cycling name that 
would have been the name of their great grandfather's age-set. For example, the praise name used 
across the region in the nineteen-thirties was Sanduka, which refers to the boxes that migrant 
workers brought home filled with things purchased in the city. The age-set name from the 1950s, 
Hobasi, refers to the "Habasha" or Ethiopians of Second World War who resisted Italian colonial 
domination. Other praise-names included Ngerecha (English) used in 1960 and Chabani 
(Japanese) used in 1 968. The same praise-names can be found across the whole region, at least 
from about 1870 on, as indicated by the inclusion of the Kuria linear age-set list in the final column 
of the chart. This regional coordination of praise names does not mean that formal institutional 
arrangements operated between these groups. Rather, it signifies only that they were aware of each 
other and found practical advantage in defining themselves as age-mates. 

Western Serengeti people understood corresponding age- and generation-sets to be one 
"generation," even if their dates of initiation did not exactly correspond. They worked both 
generation- and age-sets, either cycling or linear, into a regionally-based understanding of time. 
The textured coding on the chart in Figure 9-1 represents this understanding. These shared 
equivalencies took on practical importance in this era when people were redefining boundaries and 
populations were in flux. Equivalent age- and generation-set understandings allowed for the 

incorporation of strangers from different parts of the region and the formation of friendships in 
other areas that people could appeal to for local hospitality in times of hunger, in travel or for 
trade. The chart that I have drawn represents this idealized understanding of time and society, 
broken into discrete blocks, or "rungs of a ladder," across the regional space of the western 
Serengeti. 8 

The equivalencies between generation-sets and age-sets were possible because the newer 
system of cycling age-sets was an innovation on and still corresponded to the generation-set 
system. For example, one of the new age-set cycles used the name Saai, which is also a 
generation-set cycle name. The new system broke up the larger "generation" into three smaller 
territorial age-set units. These smaller units made it easier to mobilize young men for raids within 
a localized area. A system based strictly on the principle of generation brings men of all ages and 
stations of life into one set, with little group cohesion, while an age-set system capitalizes on the 
fraternity and equality inherent in the age-cohort and competition between different cohorts.' 
Whether or not these newly reorganized age-sets were effective militarily, they appealed to young 
men who admired the dominant power of the Maasai warrior sets who seemed to raid with 
impunity. If the new age-cycle resulted from the growing power of young men and offered a 
solution to the crises of the times, youth ultimately lost to elders who maintained the principles of 
generational authority in the unbroken assimilation of the newer age-cycles into a generational 

8 Much of the above thinking about time thanks to Richard Waller, "Making and Taking 
Time" (Paper presented at African Studies Consortium, University of Pennsylvania, 4 October 

1 996), and in personal communication. 

9 See Baxter and Almagor, "Introduction," Age. Generation and Time , pp. 2-7. 

Oral Narratives of Age-Set Reorganization 

The clearest evidence for these changes in the structure of age-sets during the "generation 
of disasters" comes from oral narratives which explicitly tell about the reorganization of age-sets. 
Elders from Ngoreme, lkoma, Nata, and Ishenyi identified the age-set of Maina (Ngirabhe), 
Matara (Megona) and Masura (C. 1850-75) as the time when they redefined age-sets into cycles or 
territorially-based "associations." A Nata elder said that the first saiga or age-set was the Maina, 
living at Site, where they divided into the three cycles of Bongirate, Busaai and Borumarancha. 10 
An Ishenyi elder confirmed that the people divided into cycling age-sets or saiga when they left 
Nyeberekera or after they got to Nyigoti, which would also have been at the time between the 
Maina and Saai generations." This process of division appears clearly in an Ishenyi text that 
reads, "The Amasura gave birth to the Amatara who then gave birth to the abaRumarancha, 
abaSaai and abaNgirati." 12 In the idiom of the fathers and sons, the generation-set "gave birth" to 
the three new age-cycles, maintaining the continuity of time. 

Philipo Haimati uses a similar generational idiom in his written chronicle of Ngoreme 

history to describe the reorganization of age-sets in response to the feeling of insecurity: 

Then they passed a law that each father should not have all his sons living in one 
homestead in one village. If a war came in one village then not all of the brothers would 
be killed at once. So they combined five circumcision sets in all to be one company of 
soldiers, one age-set. They called the first children of the age-set whom they circumcised 
the Saai and gave the Saai land to live on from Maji Moto up to Busawe. They called 
this land Ikorongo. The Saai called themselves by another praise name that they made 
up, the Mar 'osikeera. They gave them the horn and the drum. The age-set made their 
own weapons. These were the first company of soldiers. The second year they 
circumcised other children, they called them the Amatara to whom they gave the land of 

10 Interview with Kirigiti Ng'orita, Mbiso, 8 June 1995 (Nata if). 

1 ' Interview with Morigo Mchombocho Nyarobi, Issenye, 28 September 1 995 (Ishenyi cf ). 

12 "Kikao cha Mila, Desturi na Asili ya Kabila la Waishenyi Kilichokutana Tarehe 
6/6/1990, Nyiberekera, Ishenyi," copy in the possession of the present author. 


Kisaka. They called themselves the Bongirate and were given the horn and the drum and 
made their own weapons. In the third year they circumcised the next children and to 
whom they gave the same names of Amataara and Abangirate, hut who occupied the land 
of Kewantena and Bumara ... In the fourth year they called the children whom they 
circumcised Abagamutenya and they gave them the land of Ring 'wani up to Masinki to 
live . . . The fifth company of soldiers was called the Amasuura. They called 
themselves the Abarumarancha, living in the land oflramba. They gave them, loo. the 
horn and the drum and they made their own weapons. A man who had five sons made 
this division, following the circumcision sets ... He would spread them out among the 
five companies as they circumcised them in successive years. [. . .] [Each of these 
companies would take turns ruling the whole country, when they would become too old 
they would be driven out by the younger company who would then rule in their turn.] [. . 
.] At that time each lineage lived together in one settlement. They built forts to protect 
themselves from the raids. They built these forts with high walls made of rocks. In this 
way each homestead was inside the big wall and inside each homestead were the houses. 
They built these settlements on the mountain sides and they went down toward the plains 
to herd and farm. The lineages lived separately because they despised each other. Yet 
they helped each other when it was necessary and fought their common enemies. They 
made a plan together to strengthen the companies of youth when they became 
circumcised. There were five companies of soldiers and each had more than 2, 000 

The above story depicts a conscious reorganization of social space. Haimati reconfigures 
the pointillism of lineage-based settlements to an image of more concentrated settlements, joined in 
an enclosed and bounded territory of age-sets and linked together by the patrilineage and primarily 
connected to outsiders by the affinity of age-mates. The emphasis is on the warrior ethos of these 
"companies of soldiers" but the logic behind this formation argues that the spreading out of the 
sons of one man in different areas, linked together into a territory for mutual support, would 
preserve the patrilineage. In times of societal stress, the concentration of lineages and clans in one 
territory was a liability rather than an asset. 

Settlement Reorganization 

The reorganization of age-sets, in turn, led to the reorganization of settlements on the basis 
of age-set rather than lineage. In an era of societal stress settlements became concentrated and 

13 Haimati and Houle, "Mila na Matendo." 

boundary formation took on increasing prominence. Philipo's account mentions both the 
reorganization of age-sets into territories and people building fortified settlements to protect 
themselves from raids. Western Serengeti people were victims of raiding which resulted from 
competition among various Maasai groups in the Rift Valley for dominance. Some of the 
peripheral Maasai-related groups, such as the Lumbwa, after experiencing defeat, began raiding 
farmers. These pastoral groups also began encroaching on the territory of hill farmers for dry 
season grazing. Yet oral evidence attests that the threat of disease, general insecurity and the need 
for boundary definition were equally strong motivations for the concentration of settlements. The 
medicine for protection, orokoba, worked against both disease and raids; its power lay in the act of 
encirclement or enclosure of the land against external danger. Thus, fortification was a visually 
symbolic "medicine" for protection against all external forces. At a time when people were 
reformulating identity, boundary maintenance was increasingly emphasized. Western Serengeti 
people were building "walls" and boundaries where none had existed before. 
Fortified Settlements 

The remains of these stone structures all over the Mara Region testify to the movement 
into more concentrated settlements in fortified positions on the hillsides. Either thick rock walls 
higher than a man surrounded the entire village (obugo in Ngoreme), or smaller stone enclosures 
(ruaki in Nata) protected women, old people and children as a temporary shelter during the raids? 
[See Figure 9-2: Remains of Stone Walls, Ngoreme Fortified Settlements.] The German explorer 
Baumann described an Ngoreme fortified settlement with walls two meters tall and almost two 
kilometers around. One entered the settlement through a gate locked from the inside, finding a 
large open space inside. 14 White Fathers missionaries, who traveled inland briefly in 1902 and 

14 Baumann, Durch Massailand . p. 56. 




H « . ^S*d 


"** A;- 

fe •* ^P^-'^ajST 

,\.\V- - 

' • "^<? 





Figure 9-2: Remains of Stone Walls, Ngoreme Fortified Settlements, Nyansurumunti Kisaka, 21 
September 1995 

1904, reported that Zanaki, Ngoreme and Ikizu people lived in fortified settlements up on the hills 
among the rocks." The German traveler Kollmann (1899) described "Ushashi" villages up in the 
rocky hillsides surrounded by high hedges of euphorbia or thorns. Near to Ngoreme he found even 
more strongly fortified villages with stone walls five feet high and three feet wide. Inside the walls 
a virtual labyrinth of euphorbia and thorn hedges divided the individual homesteads.' 6 An 
Ngoreme elder said that each obugo had a front gate that was guarded and a secret back door for 
escape, the walls were 8-10 feet high and had holes to look out and shoot through." 

Another kind of fortification was to build a tembe or low log house, covering the roof with 
dirt, in the style of the Gogo of central Tanzania. This prevented the Maasai practice of burning 
thatch roofs during a raid to drive the inhabitants out into the open. People could fortify the door 
of a tembe from the inside to prevent intrusion. They adopted this style from the Tatoga who 
brought it from Mbulu during the disasters. The Tatoga themselves began building tembe-sty\e 
houses in spite of their preferred mobile lifestyle." 

15 Societe de Missionnaires d'Afrique (Peres Blancs), "Ukerewe," Chronique Trimestrielle 
de la Societe de Missionnaires d'Afrique fPeres Blancs) . 24me Annee, No. 95 (July 1902): 281; L. 
Bourget, "Report of a Trip in 1904 from Bukumbi to Mwanza, Kome? Ukerewe, Kibara, Ikoma- 
Mara Region, together with some stories," N.p. n.d. M-SRC54 Sukuma Archives, Bujora, 
Mwanza, Tanzania. They attributed this to a period of famine to intertribal war and the raids of 
the Maasai and Luo. 

16 Kollmann, The Victoria Nvanza . pp. 1 77-78. Ushashi is the Sukuma name given to all 
Mara peoples, this is still used as a derogatory name by the Sukuma today. 

17 Interview with Elfaristi Wambura Nyetonge, Kemgesi, 20 September 1995 (Ngoreme 

18 Interview with Stephen Gojat Gishageta and Girimanda Mwarhisha Gishageta, Issenye, 
27 July 1995 (Tatoga o - ); Gilumughera Gwiyeya and Girihoida Masaona, Issenye, 28 July 1995 
(Tatoga <f) The tembe style is an ancient form of architecture brough to the larger East African 
region by S. Cushitic-speakers. See Ehret, Classical Age . Chapter 2. 


Fortified and concentrated villages were found from western Kenya all the way down to 
Sukuma and Nyamwezi during this period. The Sonjo to the east also built substantial 
fortifications against raids." Although previous settlement patterns grouped people of related 
lineages in one area, the intense concentration of settlements during this period seems to have been 
a temporary response to stress. During the early colonial period people moved out of these 
concentrated settlements. The lineage idiom still united the inhabitants of one fortified settlement 
but they now grouped these settlements in a territory defined by age-set cycles rather than by clan. 
Concentrated Settlements 

Scholars have interpreted this trend toward fortified and, more generally, concentrated 
settlements throughout East Africa during the late nineteenth century as a response to the 
insecurities of the caravan trade, lliffe emphasizes the effect of firearms in causing the "ribbon-like 
settlements along the trade routes to give way to fortified villages." 20 Yet, as I showed in Chapter 
3, the caravan trade only indirectly affected the western Serengeti. No important trade routes or 
trade centers existed in this region where concentrated settlements might naturally develop. No 
powerful trade lords like Mirambo forced people into concentrated settlements with the threat of 
arms as in Nyamwezi. 21 

" For western Kenya see R. T. K. Skully, "Fort Sites of East Bukusu, Kenya," Azania 4 
(1969): 105-1 14; and R. T. K. Skully, "Nineteenth Century SettlementFort Sites and Related Oral 
Traditions from the Bungoma Area, Western Kenya," Azania 14 (1979): 81-96. For Nyamwezi 
see R. K. Burton, The Lake Regions of Central Africa (London, Longman, Green, Longman and 
Roberts, 1860), pp. 81-96. For Kuria see Cory, "Land Tenure in Bukuria," pp. 70-79. For Sonjo 
see Gray, The Sonio . pp. 33-34. 

20 lliffe, A Modern History , p. 75. In southern Tanzania the unrest caused by Ngoni 
incursions from the south were responsible for concentrated settlements. 

21 For a nationalist biography of Mirambo see, Norman Robert Bennett, Mirambo of 
Tanzania. 1 840?- 1 884 (New York: Oxford University Press, 1971). 


In genera], people distrusted concentrated settlements and described them as unhealthy and 
potentially dangerous places. So powerful reasons must have existed for people to move into 
concentrated settlements. Many colonial reports described the increase in witchcraft accusations 
with the concentration of settlement. The Mwanza Senior Commissioner reported that, "natives 
have been advised to concentrate ... but as they are steeped in superstition and fear of witchcraft 
in larger communities, refuse." 22 During the colonial years western Serengeti people resisted 
sleeping sickness measures to get them into concentrated settlements. Although the colonial 
government carried out these campaigns in Sukuma and elsewhere, schemes within the Mara 
Region ultimately failed. 23 Not only did western Serengeti people fear witchcraft but the fortified 
structures themselves represented an enormous outlay of labor for people who were accustomed to 
building their houses of mud and thatch in a few days. 

The concentration of settlements contributed to ecological collapse. When people lived in 
concentrated settlements, competition for accessible farmland increased, causing farmers to 
overwork the soil. Since people grazed their livestock near the settlement, concentrated settlements 
also resulted in overgrazed pastures. Where farmers and herders were not opening up new land for 
farming and settlement, tsetse bush tended to encroach into once-clear areas, as when refugees left 
for Sukuma. 24 

Scholars elsewhere have postulated that concentrated settlements were associated with the 
increased authority of political leaders and the control of elders over young people. It might be that 

22 Report from A.M.D. Turnbull, Senior Commissioner, Mwanza, to Game Warden 
Kilossa, 28 March 1924, vol. 1: 1923-29, Game Regulations, 215/P.C./14/I, TNA. 

23 Report by District Veterinary Officer, April 1927 and Monthly Reports, 1928, 1926-29 
Provincial Administration, 215/P.C./1/7, TNA. 

24 See Iliffe's analysis for all of Tanganyika, A Modern History , pp. 75-77, 163-167. 

concentrated settlements were a way elders exploited conditions of uncertainty and vulnerability to 
reassert their authority over young men who had gained some autonomy and prestige as warriors 
or as traders to Sukuma. Yet the elders had few means available to force people into concentrated 
settlements against their will. In a situation of abundant land resources, people who disagreed with 
their leaders could simply leave and be assured of a welcome in any of the other neighboring 
settlements (albeit accepting the comparatively weaker position of a newcomer). 

The threat of Maasai raiding and violence, asserted in oral narratives, stands as the most 
plausible explanation for the concentration of settlements. Yet the relationship with the Maasai and 
their threat to farming communities was more complicated than the simple enmity expressed by 
elders. Although the immediate threat of raids was most palpable, the larger threat posed by the 
Maasai was the loss of a way of life based on farming, hunting and herding inherited from the 
distant past. 

Resistance and Accommodation to Maasai Hegemony 

Beyond individual loss of life and property, the threat of Maasai raids represented the 
imposition of an entirely new economic system and ethnic map in the greater Rift Valley. As noted 
in Chapter 3, the Maasai succeeded in dominating this region by developing a specialized form of 
pastoralism that forced others into the specialized niches of farming or hunting. Western Serengeti 
peoples, having based their adaptation to this region on a combination of farming, hunting and 
herding skills, resisted the imposition of this hegemonic system controlled by the Maasai. As the 
Nyeberekera story told in Chapter 3 attests, the ultimate threat posed by raiding was that the 
Maasai would drive western Serengeti farmers out of areas now claimed by Maasai for dry season 
grazing. The immigration of Sonjo refugees to the western Serengeti as a result of Maasai raids 
most clearly demonstrates this process. 


Although disease, famine and subsequent ecological collapse caused more loss of life and 
dislocation of populations, elders say that the reorganization of age-sets and the fortification of 
concentrated settlements was the result of Maasai raiding. I argued in Chapter 3 that the most 
intense period of raiding occurred after the disasters. The experience of these later raids was then 
projected back onto narratives of raids during the disasters to make sense of an unexplainable 
series of calamities. Identifying a known enemy to account for these troubles was much more 
acceptable than combating the intangible forces disease and drought. Many narratives, such as the 
Ishenyi story of leaving Nyeberekera told in Chapter 3, attribute both disease and drought to the 
medicines of the Maasai prophets. 

In this section I show that because of both admiration and fear of the Maasai, western 
Serengeti people welcomed Sonjo refugees from Maasai raids for their valued experience in dealing 
with the Maasai. The impetus for reformulating age-sets in the western Serengeti may have come 
from the desire of young men to imitate the power of the Maasai murran. The final shape given to 
age-set structure, however, was based on older generational principles, firmly under the control of 
Sonio Refugees of the Disasters 

Intense Maasai raiding on both sides of the Serengeti plains (particularly on the Sonjo 
side) drove refugees in both directions and resulted in the separation of these two communities. 
The long- term interaction of hill fanners living in the western Serengeti and Sonjo was detailed in 
Chapter 5. During the period of disasters western Serengeti settlements moved farther to the west 
and the potentially habitable hill sites between Sonjo and Ikoma were abandoned to the Maasai. 
This was the area that the British later designated as Serengeti National Park because it lacked 
permanent inhabitants. 


Today the six Sonjo villages in the Loliondo District exist as islands of Bantu-speaking hill 
farmers surrounded by a sea of Maasai pastoralists. Each village situated on a hillside or 
mountain depends on springs to water irrigated fields on the valley floor below. The village 
leaders, known as mwanamaji ("those with water") control the allocation of water. Each village 
preserves the oral traditions of their origins and settlement separately, connected only by the epic 
cycle of stories about the prophet Khambageu. In the nineteenth century a complicated set of 
fortifications surrounded each village. 25 [See Figure 9-3: Sonjo Fortified Settlements.] The Sonjo 
practice Maasai-style linear age-sets and know nothing of generation-sets or cycling age-set names. 
However, a connection with the western Serengeti is suggested by the fact that the name of the first 
age-set which elders remember is Olnyamburete about eleven age-sets ago (like the western 
Serengeti generation-set name, Nyambureti). 26 Perhaps before the Maasai came the Sonjo did 
practice a generation-set system similar to the western Serengeti version. Today the Sonjo dress 
and outwardly look like Maasai, wearing red blankets over one shoulder and adorning themselves 
with beaded jewelry. Sonjo murran (warriors) always carry a long knife at their side. One elder 
told me that this gear was necessary for safe passage across Maasailand, as they would be 
indistinguishable from the Maasai, who are their enemies. 

The Sonjo who remain in the Loliondo district have become assimilated into the Maasai 
system of hegemony. They raise only goats and sheep, hunt little, and mainly subsist on 
agricultural production of millet and beans. Sonjo is the name given to them by the colonial 
officers after the "sonjo bean." The Sonjo call themselves Bantemi after the nlemi scar. Scholars 
of Maasai history have shown how Sonjo could pass the ethnic boundaries of economy to "become 

25 See Gray, The Sonio for one of the few ethnographies of Sonjo. 

26 Ibid, p. 88. 

Figure 9-3: Sonjo Fortified Settlements, Emmanuel Ndenu, Sale, 6 December 1995 

Maasai" if they gained cattle, while those who lost their livestock became Ndorobo hunters. Both 
the Ndorobo and Sonjo now follow the ceremonial cycle of Maasai age-sets. 2 ' The strongest 
connection between Sonjo and the western Serengeti seems to date to the period of disasters. In my 
own interviews and in those of earlier ethnographers, Sonjo elders consistently asserted that the 
"Ikoma" (used generically for western Serengeti peoples) and the Sonjo once lived as neighbors or 
as one people, from "one womb." 28 Elders from the Sonjo village of Samongo told stories of a 
village called Tinaga to the north, located more on the plains, unlike the Sonjo villages of today. 
The eight villages (Yasi, Tinaga, Meje, Buri, Hajaro, Hume, Horane and Jema) in this area were 
collectively known as Masabha (of the north). 29 When the Maasai entered this area, they raided 
and burned the Masabha villages and destroyed their fields and granaries. Without a means of 
subsistence, the people of Tinaga dispersed. Some went to Ikoma and others moved, as Tinaga 
clan mates, into other Sonjo villages in the south. One elder from the Tinaga clan said that the 
Maasai and Lumbwa fought with the Masabha people over a period of many years until the 
Maasai took all their cattle and goats and destroyed their villages. 30 Elders from the village of 
Samongo claim that they can still see the graves, homestead foundations and grindstones at the site 

27 Alan H. Jacobs, "The Irrigation Agricultural Maasai of Pagasi: A Case of Maasai- 
Sonjo Acculturation," Dar es Salaam, Social Science Conference (January 2-5, 1968): 1-12; John 
L. Berntsen, "The Maasai and their Neighbors: Variables of Interaction," African Economic 
History, 2 (Fall 1976): 1-11. 

28 Henry A. Fosbrooke, "Sections of the Masai in Loliondo Area," typescript, 1953, 
CORY #259, EAF, UDSM; Gray, The Sonjo . pp. 11-15. 

29 Interviews with Peter Nabususa, Samonge, 5 December 1 995 (Sonjo d - ); Marindaya 
Sanaya, Samonge, 5 December 1 995 (Sonjo cf ). 

30 Interview with Samweli Ginduri, Samonge, 6 December 1995 (Sonjo <?). 


of Tinaga. 31 The Ngoreme tell of ancestors who came from Masabha and also from Tinaga. 32 One 
Ikoma version of the emergence story says that the first hunters came from Sonjo Tinaga following 
the wildebeest migration to get meat during a famine. 53 All of these testimonies indicate a western 
Serengeti connection with specific communities in Sonjo, those most directly in competition with 
the Maasai for pastoral resources. 

The stories of the prophet Khambageu also tell of the connection between Sonjo and 
Ikoma. Some versions of the Khambageu story say that he came from the west in Ikoma and that 
people went there to propitiate his spirit until only a generation ago. 34 Some elders say that 
Khambageu came from the village of Tinaga where many of his miracles took place. He 
subsequently cursed Tinaga, leading to its destruction by the Maasai. 35 The Khambageu prophetic 
stories resemble the Tatoga miracle stories of their prophets. Given the more recent history of 

31 Interviews with Peter Nabususa. Samonge, 5 December 1995; Samweli Ginduri, 
Samonge, 6 December 1 995 (Sonjo J). The Tinaga site was visited by Gray, The Sonio . p. 1 3. 

32 Interview with Nsaho Maro, Kenyana, 14 September 1995 (Ngoreme cC). Philipo 
Haimati, handwritten notebook on Ngoreme history, which I saw on 14 September 1995, says that 
the Ngoreme came from Sonjo "Nyahaba." 

33 Interview with Mzee Taranka, Bugerera, 10 May 1995 (Ikoma <?). 

34 Interview with Emmanuel Ndenu, Sale, 6 December 1995 (Sonjo <?). 

35 Gray, The Sonjo, pp. 11-12. Interviews with Peter Nabususa, Samonge, 5 December 
1995; Marindaya Sanaya, Samonge, 5 December 1995 (Sonjo <f); Samweli Ginduri, Samonge, 6 
December 1995 (Sonjo cf). F. G. Finch, "Hambageu, some additional notes on the God of the 
Wasonjo," Tanganyika Notes and Records. 47 and 48 M9S7V 703-708- H. A. Fosbrooke, 
"Hambageu, the god of the Wasonjo," Tanganyika Notes and Records. 35 (1955): 38-43; E. 
Simenauer, "The Miraculous Birth of Hambageu, Hero-god of the Sonjo," Tanganyika Notes and 
Records, 38 (1955): 23-30. 


Sonjo relations with the Maasai, the Sonjo may have developed their own prophetic institution in 
the nineteenth century to combat the power of the Maasai prophets. 36 

Refugees not only moved from Sonjo to the western Serengeti but also in the other 
direction. Chapter 3 presented the story of the dispersal of Ishenyi people to Sonjo from 
Nyeberekera. Ishenyi elders said that at the time they lived at Nyeberekera they called themselves 
the Regata. In Sonjo today the older name for the village of Sale is Rhughata. An elder from 
Rhughata claimed their origins at Jalati and Ngrumega (perhaps a transliteration of the Rivers 
Mbalageti and Grumeti in western Serengeti) and that the praise names of the Rhugata clans names 
the place called Nyankerekera (perhaps a transliteration of the Ishenyi dispersal place 
Nyeberekera). The original ancestors of Rhugata were hunters of the Sagati clan, a clan name also 
found in Ishenyi and Ikoma. 3 ' Oral traditions from Ikoma and Ngoreme also claim origins in the 
Sonjo village of Regata. Migrations seem to have taken place in both directions from communities 
that were found in areas later claimed by the Maasai. 

Dating the destruction of Tinaga and the dispersal to Ikoma is difficult because elders want 
to assert the ancient roots of this connection. One elder from the Tinaga clan said that this 
happened in the time of his grandfather (C. 1 880). Confirmed dates in Maasai history help to set 
the temporal parameters of these events. Although the Maasai may have been present in the 
western Serengeti since the eighteenth century, expansion based on a specialized form of 
pastoralism did not develop until the nineteenth century. With the advent of prophetic leadership at 
the end of the eighteenth century, the Maasai began a period of increased raiding and territorial 
expansion, forcing the victims of these raids to move or abandon pastoralism. The earliest victims 

16 See David M. Anderson and Douglas H. Johnson, eds., Revealing Prophets: Prophecy 
in Eastern African History (Athens: Ohio University Press, 1 995). 

" Interview with Emmanuel Ndenu, Sale, 6 December 1 995 (Sonjo <f ). 

of this expansion, known in Maasai tradition as the pastoral Lumbwa, were forced to recoup their 
losses by raiding Bantu-speaking farmers with few livestock. Sonjo traditions often tell of raids by 
the Lumbwa rather than the Maasai. Between 1 850 and 1 890 the Purko-Kisongo under prophetic 
leadership completed their expansion into what is now south-central Tanzania. It was probably 
during this period that Maasai sections on the periphery of the central Purko-Kisongo cluster near 
Mt. Meru began seeking control over pastoral resources in the western Serengeti and Sonjo. 
Raiding increased dramatically to maintain a pastoral way of life after 1 890 when the rinderpest 
panzootic destroyed Maasai herds. 38 

Although the Sonjo immigrants were too few in number to change western Serengeti 
language or culture, western Serengeti people may have valued Sonjo immigrants for their 
knowledge of Maasai culture. This might account for the attribution of Sonjo as the place of origin 
and the home of first man or first woman. Although the age-set system adopted by western 
Serengeti people at this time was a local innovation, Sonjo refugees may have brought compelling 
experience with the warrior ethos and Maasai-type age-sets. They would have experienced raids 
earlier and more intensely and had closer contacts with the Maasai than western Serengeti people. 
If Sonjo knowledge provided the means for resisting Maasai raids then as western Serengeti people 
formulated new ethnic identities, they would have acknowledged the crucial role of Sonjo in their 
own emergence as a people. We might learn something about the value of these Sonjo immigrants 
by looking more closely at the ambivalent relationship between western Serengeti peoples and the 

! Berntsen, "Pastoralism, Raiding and Prophets," pp. 1 12-143, 172, 224. 

Relations with Maasai 

Narratives of the disasters concerning the Maasai invariably picture them as the arch 
enemy, the "other" with whom no relationship of peace was possible. Western Serengeti people 
divide the peoples of the larger region into two opposed categories: Bisa (enemies) and Rema 
(farmers). The Maasai and sometimes the Asi hunter-gatherers were in the "enemy" category, as 
those who live in the wilderness (Nyika). This way of categorizing regional relationships does not 
agree with the way that Maasai scholars have interpreted mutually exclusive identities in reference 
to "differential access to resources" and economic specialization. 39 Western Serengeti people were 
reluctant to give up their agro-pastoral-hunting economy for a specialization in which they would 
become subordinate to Maasai pastoralists who controlled a system that defined ethnicity by 
economics, largely to the benefit of the Maasai. 

The western Serengeti understanding of Tatoga as "fathers" defies the Maasai hegemonic 
categories of farmers, herders and hunters. Although the Tatoga were not "farmers" they were not 
considered to be "enemy" and cooperated with "farmers" during Maasai raids. If a Tatoga killed a 
Nata, Ikoma or Ishenyi it was just like killing another Tatoga, rituals of purification were 
performed and a fine paid to the family. 40 Many of the western Serengeti peoples considered the 
Tatoga as their spiritual or ritual "fathers." The Ikoma and Ishenyi peoples gave Tatoga prophets a 
prominent place in the most important rituals of "cooling the land," indicating an acknowledgment 

39 Spear and Waller, Being Maasai . p. 6. 

40 Interview with Gilumughera Gwiyeya and Girihoida Masaona, Issenye, 28 July 1995 
(Tatoga <f ). 

of Tatoga as "first-comers" on the land. 41 The western Serengeti farmers allied themselves with 
other pastoralists to resist the incursions of the Maasai. 

The relationship of western Serengeti people to the Maasai manifested itself in ritualized 
form. The "farmers" of the western Serengeti, the Sonjo and the Tatoga pastoralists practiced an 
important ritual called the aghaso, to purify and reward young men who killed a lion, leopard or 
Maasai. 42 The first man (omwiti) to hit the lion with his arrow or spear and the next two men 
following him (omunoti) received honor and became blood brothers. While still at the lion kill they 
cut out the heart of the beast, the small tip of which they fed to the killer who spat it out three 
times, ingesting the fourth bite. [See Figure 9-4: Maasai Relations.] The killers took the lion skin 
and the claws back as their trophy but they burnt the remainder of the animal corpse. They also 
took Maasai weapons and other things as trophies. As the group came into the village, they sang 
the songs of the aghaso. When the village heard these songs, the mother of the killer came out to 
greet them, throwing sand and smearing them with butter. The father of the killer gave his son a 
cow. The next morning the singing, dancing and meat feasting began and lasted for eight days or 
even a month during which time the killers went around and received gifts of livestock in each 
home. 43 

41 Sutton cites linguistic and oral evidence that the Tatog- speaking peoples once occupied 
the Loita-Mara plains and across Serengeti to the Crater Highlands, being pushed out or absorbed 
by Maasai expansion. J.E.G. Sutton, "Becoming Maasailand," in Being Maasai . p. 48. 

42 Interview with Marindaya Sanaya, Samonge, 5 December 1995 (Sonjo cf). 
Extraordinarily similar practice by the Tatog Barabaig reported by G. McL. Wilson, "The Tatoga 
of Tanganyika (Part II)," Tanganyika Notes and Records . 34 (1953): 35-56, where any killing of 
cattle thieves or lions may be used to collect lots of cattle. He speculates the group most likely to 
engage in this activity are youngest sons without other outlets for status. Among the Barabaig the 
anointing of the killer with butter is a propitiation of the ancestors and the cattle given to him 
equivalent to blood compensation offered to a kinsmen. 

43 Among the Barabaig the killer "adorns himself with women's ornaments, which 
symbolize that he is like a woman who has given birth. Killing an enemy of the people and giving 


The killers were dangerous and liminal characters-like a lion of the wilderness. During 
this time they went through rituals of purification: they shaved their heads and smeared them with 
the stomach contents of a sacrificed sheep and they could not eat or sleep with other people. 44 
This ceremony explicitly categorized Maasai with the beasts of the wilderness. The symbolism 
here was not derogatory but of respect and admiration. The killer ingested the heart of the beast to 
internalize the qualities of courage and power. One elder said that the heart is the place of courage, 
the essence of the beast, by eating it one gains that courage. The bits of heart that they spat out 
were an offering to the ancestors, since they have given the strength for this feat. As in so many 
western Serengeti rituals, they brought the things of the wilderness within-the fire brought into the 
home by the hunter, the things of power brought from the wilderness to perform the rituals of 
enclosure. 45 Yet this was also a direct act of resistance to Maasai dominance. As one informant 
said, 'the Maasai was "boss" then and the aghaso proved our triumph over them'. 46 Western 
Serengeti people admired, accommodated and resisted the Maasai. 

birth is symbolically equated. The killer of an enemy must observe a convalescence period (one 
month) for having given "birth," and is restricted from touching food or doing any work." Klima, 
The Barabaig . pp. 58-60. Western Serengeti peoples also allow women who are courageous in 
birth to dance the aghaso with the men. Interviews with Baginyi Mutani and Mayenye Nyabunga, 
Sanzate, 8 September 1995 (Ikizu ¥). 

44 Interviews with Zamberi Masambwe and Gisuge Chabwasi, Mariwanda, 22 June 1995 
(Ikizu if); Mang'oha Morigo, Bugerera, 24 June 1995 (Nata d - ); Merekwa Masunga and Giruchani 
Masanja, Mariwanda, 7 July 1 995 (Tatoga cf ); Elfaresti Wambura Nyetonge, Kemegesi, 20 
September 1 995 (Ngoreme &); Zabron Kisubundo Nyamamera and Makang'a Magigi, Bisarye, 9 
November 1 995 (Zanaki cf); Marindaya Sanaya, Samonge, 5 December 1995 (Sonjo <f). 

45 Related to the analysis of "other" in the Kramer, Red Fez , p. 2, Kramer shows how 
African representation of the European "other" in sculpture was used to define self; and Boddy, 
Wombs and Alien Spirits , p. 342, Boddy demonstrates that the "zar" possession cult in Sudan 
fosters an "alien world at the heart of culture." 

46 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata <?). 


Elia Masiyana Mchanake and Robi Nyekisokoro, descendants of the founders of 
Nguku Ngoreme, Saroti and Nyaboge (Matiti's daughter), Borenga, 21 September 

Zamberi Masambwe and Gisuge 
Chabwasi, reenacting the Aghaso, 
Mariwanda, 22 June 1995 

Figure 9-4: Maasai Relations 


In spite of this history of animosity toward the Maasai, evidence of interaction and 

cooperation also exists. The first raids remembered by elders began in the 1 870s, while interaction 

and emulation of Maasai culture must have been ongoing much before this. An interdependent 

regional economy of pastoralists, farmers and hunter/gatherers was held together not only by force, 

but also by the glue of common cultural understandings and social interaction with frequent 

boundary crossing. 47 From as early as 1800 the Loitai Maasai expanded from the Rift Valley to 

the Loitai hills where they pushed the Siria Maasai up to the Mara River in Kuria and Ngoreme 

territory. 48 Maasai ancestors often appear in Ngoreme and Ikoma genealogies. Western Serengeti 

people still propitiate their Maasai ancestors using prescribed Maasai implements and cattle 

sacrifice. 49 One localized emergence story in Ngoreme capsulizes this interaction: 

Saroti was Maasai and left behind when others moved on, at the hill ofGisema. He built 
his house near the spring ofKiru and lived by himself for many years. Then one day he 
saw smoke coming from Nyibihori and went to see who it was. There he found a man 
named Matiti. his wife and their children. They became friends and built their houses 
near to each other. Matiti was a farmer and grew millet, he taught Saroti to farm. 
Saroti married Matiti's daughter, Nyaboge, for one storage bin of grain during a time of 
hunger. She cut his hair and shaved his head to make him acceptable for marriage. 
They gave birth to Kitang'ita, Gogay, and Wandwe?" 

47 Spear and Waller, Being Maasai . p. 2. 

48 Galaty, "Maasai Expansion," p. 72. 

49 For Kuria Maasai clan ancestors also reported by Cory, "Land Tenure in Bukuria," pp. 
71-72. See Raids by Masai, 1936, vol. 1, Secretariat Files, 23384, TNA, on a scare of Maasai 
raid because a woman was carrying out one of these rituals on behalf of her Maasai ancestor in 
Sukuma, a young herd boy saw the proceedings and ran to alert everyone that a raid was in 
progress. Interviews with Gabuso Shoka, Mbiso, 30 May 1995 (Nata cf); Bokima Giringayi, 
Mbiso, 26 October 1995 (Ikoma cf); TetereTumbo, Mbiso, 5 April 1995 (Nata o"); Mwita 
Magige, Mosongo, 9 September 1 995 (Ngoreme d"). 

50 Interview with Elia Masiyana Mchanake (Ngoreme cf) and Robi Nykisokoro (Ngoreme 
?), Borenga, 21 September, 1995. There are variations of this story, including that they met on a 
hunt and that Saroti ate porridge (ugali). In some versions of this story Matiti is said to be an 
Iregi, the clan which left the Nyeberekera dispersal center in the Ishenyi story. Saroti is sometimes 
said to be Maasai only in that he was a "vagabond, traitor or outcast" his origin was Gosi, from the 


In this narrative the Maasai pastoralist takes refuge with a farming patron, marries his 
daughter and founds a new clan territory in Ngoreme, Nguku. [See Figure 9-4: Maasai Relations, 
p. 469.] This story illustrates the permeability of the ethnic and economic boundaries in the late 
nineteenth century. Nyaboge made Saroti fit to join the community by cutting his hair, presumably 
the long locks of a Maasai murran, destroying the outward marks of his Maasai identity and 
removing him from the warrior grade. 

The symbiotic relationship of the Maasai and the Ikoma, farthest to the east, was also 

significant. During the rihaha famine, or rinderpest of 1 890, the Maasai came to "sell" their 

children in Ikoma for food. Many stayed and settled near Banagi hill, well into Ikoma territory and 

now part of Serengeti National Park. Ikoma clans adopted young Maasai men and married young 

Maasai women, establishing in-law relationships of long duration. When the Maasai began to 

recover and the "Hunger of the Feet" hit the farming peoples, the Ikoma went to the Maasai for 

help. The Maasai were raiding on the lake during this time and used their Ikoma friends as scouts 

who knew the land better. Even today western Serengeti people know the Ikoma as Maasai 

collaborators." A 1933 report from the Musoma District illustrates this close relationships: 

. . . you must remember that the Waikoma are on very friendly terms with the Serengeti 
Masai. For many years the Masai have brought tails of wild animals for exchange with 
the Waikoma who sell them in Usukuma.... a Maasai can always rely on a bed and a 
meal when he visits Ikoma. No doubt many of them act as guides to raiding parties . . . 

Shirati area. All accounts confirm that the spring at Kiru is a powerful erisambwa place. And all 
are both proud and embarrassed of this important Maasai ancestor. Interviews with Isaya Charo 
Wambura, Buchanchari, 22 September 1995 and Charwe Matiti, Nyeboko 22 September 1995 
(Ngoreme cf). 

51 Interviews with Mahewa Timanyi and Nyambureti Morumbe, Robanda, 27 May 1995 
(Ikoma ef); Machota Sabuni, Issenye, 14 March 1996 (Ikoma tf). 


but it is quite useless to expect the Waikoma to abandon friendly relations which have 
survived the raids. 51 

Maasai dominance also resulted in the development of a common regional culture. Elders 
say the Ikoma, Nata and Ishenyi practice resembles the Maasai in dance, ornamentation and 
songs." Although western Serengeti youth did not use red ochre (a characteristic sign of the 
Maasai murran) in everyday ornamentation, young men would put it in their hair in special 
occassions-to dance, after circumcision or for cattle raids. When the White Fathers visited the 
Ngoreme in 1 904, they described young men wearing their hair in butter and ochre-smeared plaits 
as the Maasai. 54 At the dances held at the full-moon young people wore rows of brass or wire- 
wrapped anklets and bracelets, beaded headgear and ear ornaments. Mara peoples also pierce and 
elongate their earlobes, as did the Maasai. Just when they adopted this style cannot be deduced 
from the available evidence. People now think of it as "traditional." 55 

The cultural patterns that seemed to imitate the Maasai were so widespread that the early 
explorers and colonial officers mistook western Serengeti people to be Maasai. German explorers' 

52 Annual Report 1933, Musoma District, Annual Reports, Native Affairs Section, Lake 
Province, 215/924/2, TNA. 

53 Interview with Mang'oha Morigo, Bugerera, 24 June 1995 (Nata <?). 

54 "Report of a Trip in 1904 from Bukumbi to Mwanza, Kome? Ukerewe, Kibara, Ikoma - 
Mara Region, together with some stories," L. Bourget, Trip Diary, N.p. n.d., 1904, M-SRC54, 
Sukuma Archives, Bujora, Mwanza. Interview with Machota Sabuni, Issenye, 14 March 1996 
(Ikoma <?). 

55 E. C. Baker, "Age-Grades in Musoma District, Tanganyika Territory," Man 27, 151 
(1927): 223, reports that the Kuria abaNgibabe age set initiated in 1858-62 in Nyabasi first began 
piercing the tops of their ears for the insertion of small sticks as ornaments. It would be interesting 
to know if western Serengeti use of Maasai ornamentation differed slightly enough, as Klumpp and 
Kratz show for the Okiek, that it is a visual display of both submission and resistance to Maasai 
dominance, Donna Klumpp and Corinne Kratz, "Aesthetics, Expertise, and Ethnicity: Okiek and 
Maasai Perspectives on Personal Ornament," in Being Maasai: Ethnicity and Identity in East 
Africa, eds. Thomas Spear and Richard Waller (Athens: Ohio University Press, 1993), pp. 195- 
222. Unfortunately there are few of these ornaments left as they are not currently in use. 

Baumann and Weiss both noted the similarity between these peoples and the Maasai in dress, 
ornamentation, ear piercing, use of snuff and weapons. 56 Present day Kuria peoples shown the 
photos taken by Weiss of