Hostname: page-component-7c8c6479df-r7xzm Total loading time: 0 Render date: 2024-03-28T06:09:42.280Z Has data issue: false hasContentIssue false

We Are What We Eat: Ancient Agriculture Between the Great Lakes1

Published online by Cambridge University Press:  22 January 2009

David L. Schoenbrun
Affiliation:
University of Georgia

Extract

A history of food systems in Africa's Great Lakes region is presented using mostly historical linguistic sources, with help from archaeology and paleoecology. The paper moves beyond understanding the causes and consequences of iron-working as the most important feature of the period between c. 1000 b.c. and c. a.d. 500. I argue that a history of agriculture both gives context to changes in technology and introduces powerful new explanations for historical processes of settlement and occupational specialization that took place.

Between 1000 b.c. and 500 b.c., in the Great Lakes region, speakers of three of Africa's four major language families practiced distinguishable food-producing systems. Two groups, Central Sudanian and Sog Eastern Sudanian, depended mainly on growing cereals and raising livestock for their sustenance. The third group, the Tale Southern Cushites, gave decidedly greater emphasis to cattle but probably also grew grains. A fourth group, the Great Lakes Bantu, grew root crops, fished and raised cattle and grain. They inherited much of their knowledge of these techniques, other than cattle-raising, from earlier Eastern Highlands Bantu-speakers. But they incorporated cattle and some grains through longstanding contacts with the two Sudanian and the Southern Cushitic communities. The eclectic food system they thus created allowed them to carry their unified, complex food-producing system throughout the wide variety of environments that they encountered in the Lakes region. After c.a.d. 200 descendants of the Great Lakes Bantu refined this synthesis; they emphasized livestock raising inland from Lake Victoria, and mixed farmers spread throughout the Kivu Rift. Technological, demographic, ecological and sociological explanations of the technological evidence are offered.

Type
Archaeology and Linguistics Among the Great Lakes
Copyright
Copyright © Cambridge University Press 1993

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

2 Vansina, Jan, L'évolution du royaume du Rwanda (Brussels, 1960)Google Scholar; Rennie, J. Keith, ‘The pre-colonial kingdom of Rwanda: a reinterpretation’, Transafrican Journal of History, II (1972), 1053.Google Scholar

3 Berger, Iris, Religion and Resistance: East African Kingdoms in the Precolonial Period (Tervuren, 1981)Google Scholar; Tantala, Renee, ‘The early history of Kitara in western Uganda: process models of religious and political change’ (Ph.D. thesis, University of Wisconsin-Madison, 1989)Google Scholar; and Ogot, Bethwell A., ‘The Great Lakes’, in Niane, D. T. (ed.), Africa from the Twelfth to the Sixteenth Century (Paris, 1984), 498524.Google Scholar

4 Sutton, John E. G., ‘East Africa before the seventh century’, in Mokhtar, G. (ed.), Ancient Civilizations of Africa (UNESCO General History of Africa, II) (Paris, 1981), 277–99Google Scholar; Oliver, Roland and Fagan, Brian, ‘The emergence of Bantu Africa’, in Fage, John D. (ed.), The Cambridge History of Africa, (Cambridge, 1978), Vol. 2: c. 500 B.C.–1050 A.D., 342409Google Scholar; Twaddle, Michael, ‘Towards an early history of the East African interior’, History in Africa, II (1975), 147–81CrossRefGoogle Scholar; Ehret, Christopher, ‘The East African interior’, in el-Fasi, M. (ed.) and Hrbek, I. (asst. ed.), Africa from the Seventh to the Eleventh Century (UNESCO General History of Africa, III) (Paris, 1988), 616–42.Google Scholar

5 For the period 500 b.c. to a.d. 500 there is Schoenbrun, David L., ‘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. thesis, University of California at Los Angeles, 1990).Google Scholar Although focused on periods after a.d. 1000 the following works are extremely suggestive of the new directions we seek to take here: Cohen, David William, ‘The face of contact: a model of a cultural and linguistic frontier in early Eastern Uganda’, in Vossen, Rainer and Bechhaus-Gerst, Marianne (eds.), Nilotic Studies: Proceedings of the International Symposium on Languages and History of the Nilotic Peoples, Cologne, January 4–6, 1982 (Berlin, 1983), 341–55Google Scholar; Newbury, David S., Kings and Clans: Ijwi Island and the Lake Kivu Rift, 1780–1840 (Madison, 1991), 4354.Google Scholar

6 Schmidt, Peter R., Historical Archaeology: A Structural Approach in an African Culture (Westport, Conn., 1978)Google Scholar; Noten, Francis Van, ‘The Early Iron Age in the Interlacustrine Region: the diffusion of iron technology’, Azania, XIV (1979), 6180CrossRefGoogle Scholar; Schmidt, Peter R. and Childs, S. Terry, ‘Innovation and industry during the Early Iron Age in East Africa: the KM2 and KM3 sites of northwest Tanzania’, African Archaeological Review, III (1985), 5394CrossRefGoogle Scholar; Grunderbeek, Marie-Claude Van, Roche, Emile and Doutrelepont, Hugues, Le premier âge du fer au Rwanda et au Burundi: archéologie et environnement (Brussels, 1983)Google Scholar; Grunderbeek, Marie-Claude Van, ‘Essai d'étude typologique de céramique urewe de la région des collines au Burundi et Rwanda’, Azania, XXIII (1988), 1156CrossRefGoogle Scholar; Hiernaux, Jean and Maquet, Emma, ‘Cultures préhistoriques de l'âge des métaux au Ruanda-Urundi et au Kivu (Congo beige): Ie partie’, Bulletin des Séances de l'Académie Royale des Sciences Coloniales, VI (1957), 1126–49Google Scholar; Hiernaux, Jean and Maquet, Emma, ‘Cultures préhistoriques de l'âge des métaux au Ruanda-Urundi et au Kivu (Congo beige): IIème partie suivie de: deux sites archéologiques à briques en territoire Walikale (Kivu)’, Académie Royale des Sciences d'Outre-Mer, Mémoire No. 8, Nouveau Série 10, 2 (1960), 188Google Scholar; Clist, Bernard, ‘A critical reappraisal of the chronological framework of the early Urewe Iron Age industry’, Muntu, VI (1987), 3562.Google Scholar

7 Schmidt, , Historical Archaeology, 286–94Google Scholar; Grunderbeek, Van et al. , Le premier âge, 1983.Google Scholar

8 Sutton, John E. G., A Thousand Years of East Africa (Nairobi, 1991), ch. 2.Google Scholar See especially chapters 7, 8 and 11 in Schmidt, Historical Archaeology. The constriction of forests and the development of a complex iron smelting technology in Rwanda, Burundi and Buhaya are beyond doubt. But we may question whether or not the links made in royal traditions between iron smelting sites, ruling dynasties and the powerful AbaCwezi patron spirits represent actual coincidence in the early Iron Age or result from later attempts by new dynasties to bolster legitimacy by claiming connections with sacred places.

9 Van Grunderbeek et al., Le premier âge; Noten, Histoire archéologique. In related work, Childs has located a possible source of friction between the iron workers and the potters of Buhaya in their competition over swamp clays with plastic and refractory qualities desired by both specialists; Childs, S. Terry, ‘Clay resource specialization in ancient Tanzania: implications for cultural process’, in Kolb, Charles (ed.), Ceramic Ecology Revisited (Oxford, 1988), 24.Google Scholar

10 Schmidt, , Historical Archaeology, 280.Google Scholar

11 Grunderbeek, Van et al. , Le premier âge, 42.Google Scholar

12 Ehret, , ‘Between’, 626–42.Google Scholar

13 Ehret, Ibid. 627–9. For a recent and far-reaching reconsideration of his earlier positions, see Ehret, Christopher, ‘The African Great Lakes Region in the Early Iron Age: shifting mosaics of cultural and economic interaction’ (paper presented at the African Studies Association meeting, St Louis, 1991).Google Scholar

14 Wrigley, Christopher C., ‘Cattle and language between the Lakes’, Sprache und Geschichte in Afrika, VIII (1987), 247–80.Google Scholar

15 For full references and discussion, see Schoenbrun, David L., ‘The contours of vegetation change and human agency in eastern Africa's Great Lakes region: ca. 2000 b.c. to ca. a.d. 1000’, Forest and Conservation History, forthcoming (1993).Google Scholar

16 The analytical and dating methods associated with palynology and paleolimnology are well covered in the literature. See Schoenbrun, David L., ‘Treating an interdisciplinary allergy: methodological approaches to pollen studies for the historian of early Africa’, History in Africa, XVIII (1991), 323–48.CrossRefGoogle Scholar

17 A full list of references to the literature on paleoenvironmental change in the Great Lakes Region is presented in Schoenbrun, , ‘Early history’, 618–54.Google Scholar Data that underlie the summary given here are discussed fully in Schoenbrun, ‘The contours’.

18 See Schoenbrun, , ‘Early history’, 7091.Google Scholar

19 Large concentrations of silted material in association with charcoal from Ahakagyezi supports this reconstruction. See Hamilton, Alan C., Taylor, David and Vogel, Joseph O., ‘Early forest clearance and environmental degradation in south-west Uganda’, Nature, CCCXX (1986), 164.CrossRefGoogle Scholar

20 Schoenbrun, , ‘Early history’, 91–4.Google Scholar

21 Schoenbrun, , ‘Early history,’ ch. 5Google Scholar, presents the career of nomenclature for different sub-units of Great Lakes Bantu. I follow the spirit of the conference on ‘the ancient civilisation of the peoples of the Great Lakes’ where ‘Great Lakes’ or ‘Grands Lacs’ was substituted for ‘Interlacustrine’. See La civilisation ancienne des peuples des grands lacs (Colloque de Bujumbura, 41009, 1979) (Paris and Bujumbura, 1981), 7.Google Scholar To the historical linguist, Great Lakes Bantu is a sub-branch of the Eastern Highlands group of the Niger-Congo family of African languages. The use of the term ‘Eastern Highlands Bantu’ refers to Bernd Heine's ‘Ostochland’ group and reflects a genetic classification. The term ‘Eastern Bantu’, where used below, refers to a referential classification. Unfortunately, many of the people who spoke proto-Eastern Highlands dialects probably did not live in highland environments! Heine, Bernd, ‘Zur gentische Gliederung der Bantu-sprachen’, Afrika und Übersee, LVI (1973), 164–85Google Scholar; and Heine, Bernd, Hoffand, H.Vossen, Rainer, ‘Neuere ergebnisse zur Territorialgeschichte der Bantu’, in Möhlig, Wilhelm, Rottland, Franz and Heine, Bernd (eds), Zur Sprachegeschichte und Ethnohistorie in Afrika (Berlin, 1977), 5772.Google Scholar The use of these terms does not imply agreement with the formulation ‘Western Bantu’ as presented by Vansina, Jan, ‘Western Bantu Expansion’, J. Afr. Hist., XXV (1984), 129–45.CrossRefGoogle Scholar The great time depth of the proposed binary split makes suspect any classificatory decisions based solely on lexicostatistical study. But see Bastin, Yvonne, Coupez, André and Halleux, Bernard de, ‘Classification lexicostatistique des langues bantoues (214 relevés)’, Bulletin des Séances de Académie Royale des Sciences d'Outre-Mer, XXVII (1983), 173–99.Google Scholar We await the next effort by the team at Tervuren, in association with Michael Mann, which will include some 345 Bantu languages.

22 Material for these lists comes from my field collections, those of Derek Nurse, Gerard Philippson and their students at UDSM and other published and unpublished sources. See Schoenbrun, , ‘Early history’, 364411, 422–49.Google Scholar

23 Only the Western Lakes and West Nyanza branches of Great Lakes Bantu are discussed in this essay. Greater Luhyia, East Nyanza and Pre-RuGungu are not pursued.

24 For generally favorable views on this vexed question see Swadesh, Morris, ‘Towards greater accuracy in lexicostatistical dating’, International Journal of American Linguistics, XXI (1955). 121–37CrossRefGoogle Scholar; Anttila, Raimo, Historical and Comparative Linguistics (Amsterdam, 1989). 396–8CrossRefGoogle Scholar; and Ehret, Christopher, ‘Language change and the material correlates of language and ethnic shift’, Antiquity, LXI (1988), 566–9.Google Scholar For unfavorable views see Bynon, Theodora, Historical Linguistics (Cambridge, 1977), 266–72CrossRefGoogle Scholar; and Birnbaum, Hendrik, Linguistic Reconstruction, its Potential and Limitations in New Perspective (Washington, 1978), 17.Google Scholar

25 All such ‘replacements’ are innovations and represent natural processes of language change that occur in both literate and non-literate contexts. See Vansina, Jan, Paths in the Rainforests: Toward a History of Political Tradition in Equatorial Africa (Madison, 1990), 916Google Scholar; Ehret, , ‘Language change’, 564–6Google Scholar; Schoenbrun, , ‘Early history’, 103–31.Google Scholar

26 Sapir, Edward [1916], ‘Time perspective in aboriginal American culture: a study in method’, in Mandelbaum, David G. (ed.), Edward Sapir: Selected Writings in Language, Culture, and Personality (Berkeley, 1985), 410–25.Google Scholar

27 See the collection of papers in Cooper, Robert L. (ed.), Language Spread: Studies in Diffusion and Social Change (Bloomington, Indiana, 1982), especially 562.Google Scholar

28 Ehret, , ‘Linguistic inferences about early Bantu history’, in Ehret, Christopher and Posnansky, Merrick (eds.), The Archaeological and Linguistic Reconstruction of African History (Berkeley, 1982), 5961Google Scholar; Vansina, , Paths, 55–6.Google Scholar

29 When mentioned in the text the word's corresponding number will be given in parentheses.

30 Ehret, Christopher, ‘Cattle-keeping and milking in Eastern and Southern African History: the linguistic evidence’, J. Afr. Hist., VIII (1967), 118CrossRefGoogle Scholar; idem, ‘Sheep and Central Sudanic peoples in Southern Africa’, J. Afr. Hist., IX (1968), 213–22Google Scholar; Nurse, Derek, ‘Extinct Southern Cushitic communities in East Africa’, in Bechhaus-Gerst, Marianne and Serzisko, J. (eds.), Cushitic-Omotic: Papers from the International Symposium on Cushitic and Omotic Languages, Cologne, 6–9 January, 1986 (Hamburg, 1988), 93104Google Scholar; Schoenbrun, , ‘Early history’, chs. 5, 6.Google Scholar

31 This and subsequent discussions of the chronology and location of settlement in the Great Lakes region based on historical linguistic criteria are bolstered by correlations with a sometimes uneven archaeological record of ceramic traditions. The methodological focus of this exercise often deflects attention from the history of economic and social process, though precisely economic and social factors are implicated in the development, dispersal and promotion of a ceramic industry. See Schoenbrun, , ‘Early history’, 267–90Google Scholar, for ideas about correlation both of the historical linguistic record and regional pottery traditions.

32 For comparative linguistic evidence see Ehret, Christopher, ‘Patterns of Bantu and Central Sudanic settlement in Central and Southern Africa (ca. 1000 b.c.–500 a.d.)’, Transafrican Journal of History, III (1973), 171;Google Scholaridem, ‘Agricultural history in Central and Southern Africa, c. 1000 b.c. to a.d. 500’, Transafrican Journal of History, iv (1974), 4ffGoogle Scholar; and idem, ‘Proto-Central Sudanic reconstructions’ (n.p., 1986). For more work on Central Sudanic see Demolin, Didier, ‘Some problems of phonological reconstruction in Central Sudanic’, Belgian Journal of Linguistics, III (1988), 5396Google Scholar; Goyvaerts, Didier L. (ed.), Language and History in Central Africa (Antwerp, 1986)Google Scholar; and Bender, M. Lionel, ‘Central Sudanic segmental and lexical reconstruction’, Afrikanistische Arbeitspapiere, XXIX (1992), 561.Google Scholar Sorghum and Pennisetum both thrive in lower rainfall regimes with much less atmospheric moisture than required for the yam or oil palm. Sorghum cannot tolerate frost but can briefly withstand waterlogged soils without suffering significant loss of yield. See Purseglove, J. W., Tropical Crops, Monocotyledons (Harlow, Essex, 1972), 269–71.Google Scholar In eastern Africa, Pennisetum requires a minimum average rainfall of 350–400 mm; Food and Agriculture Organization, Traditional Food Plants: A Resource Book for Promoting the Exploitation and Consumption of Food Plants in Arid, Semi-Arid, and Sub-Humid Lands of Eastern Africa (Rome, 1988), 472.Google ScholarPennisetum sp. needs a more evenly distributed rainfall regime than does Sorghum, although it will grow with as little as 250 mm of rain annually; Purseglove, Monocotyledons, 206. In eastern Africa, Pennisetum requires high temperatures during the day and cool ones at night to mature properly. The mean minima and maxima should be 18°C and 27°C respectively. The crop is rarely found above 1200 m, FAO, Traditional Food Plants, 396–7.Google Scholar

33 The authorities here are Ehret, Demolin and Goyvaerts (see note 32). But work undertaken in Zaire by Mary McMaster of Castleton State University, and soon to be underway by Kairn Klieman of UCLA, should add greatly to our knowledge.

34 Johnson, Douglas H. and Anderson, David M. (eds.), The Ecology of Survival: Case Studies from Northeast African History (London and Boulder, CO, 1988)Google Scholar; Waller, Richard, ‘Ecology, migration, and expansion in East Africa’, African Affairs, LXXXIV (1985), 347–70.CrossRefGoogle Scholar

35 Ford, pointing out the speculative nature of such considerations, concludes that modern distributions of fly fail to indicate ancient patterns of tsetse infestation, but that ‘a vegetational and faunal continuity existed that would have allowed fusca group tsetses and their hosts to have spread along the forest edges associated with the highlands of both the Western and Eastern Rifts…’; Ford, John, The Role of the Trypanosomiases in African Ecology: A Study of the Tsetse Fly Problem (Oxford, 1971), 121.Google Scholar For the sake of simplicity I assume, with Ford, that areas of more open grasslands above about 1500 m were relatively safe for livestock; Ford, Ibid. 5. I also assume that these same areas might have harbored tick-borne ‘east coast fever’. Comparative linguistic data might help to prove the existence of these pests during the early periods of history. A root *-gUpe meaning ‘cattle tick’ in proto-Great Lakes Bantu is the strongest evidence so far. A shift in meaning from the general term for ‘tick’ to that of ‘cattle tick’ seems to have occurred during the proto-Great Lakes period.

36 Dyson-Hudson, Neville, Karimojong Politics (Oxford, 1966), 39Google Scholar; Weatherby, John, ‘Inter-tribal warfare on Mount Elgon in the 19th and 20th centuries’, Uganda Journal, XXVI (1962), 200–12.Google Scholar

37 Ford, , The Role, 8690Google Scholar referenced in Giblin, James, ‘Trypanosomiasis control in African history: an evaded issue’, J. Afr. Hist., XXXI (1990), 63.Google Scholar

38 The name ‘Sog’ is the Kuliak word for ‘mountain’ and is a common loan word in Great Lakes Bantu. See root 23 in Appendix.

39 We are ignoring the question of the presence and role of specialized gatherer-hunters in the Great Lakes region. Since no linguistic evidence yet exists which might be of use in discussing exclusively gathering and hunting peoples the lacuna is justifiable if lamentable. Later efforts at reconstructing vocabulary for hunting, wild food collecting and fishing should reveal the rich contributions these activities made to Great Lakes nutrition.

40 I owe to Christopher Ehret the original observation that several Nilo-Saharan loans in Great Lakes Bantu could not be derived from Nilotic or Central Sudanic sources. Based on linguistic geography and its correlation with archaeological sequences, John Sutton long ago hinted at an ancient Nilo-Saharan presence in the Lakes region; Sutton, John E. G., ‘The Aquatic Civilization of Middle Africa’, J. Afr. Hist., XV (1974), 537.Google Scholar For a collection of reconstructed agricultural terms in Proto-Kuliak see Schoenbrun, , ‘Early history’, 570–1Google Scholar, where all items come from Heine, Bernd, The Kuliak Languages of Eastern Uganda (Nairobi, 1976), 73–9Google Scholar; Ehret, Christopher, ‘The classification of Kuliak’, in Schadeberg, Thilo C. and Bender, M. Lionel (eds.), Nilo-Saharan: Proceedings of the First Nilo-Saharan Linguistics Colloquium (Dordrecht, 1981), 269–89Google Scholar; and idem, ‘Revising Proto-Kuliak’, Afrika und Übersee, lxiv (1981), 81100.Google Scholar

41 The southern Kuliak may have been responsible for the earliest evidence of forest clearance in the Pilkington Bay pollen core. Their apparent reliance on sorghum as the main cereal crop nicely supports the interpretation of a long-fallow, valley-bottom farming system involved in the clearance of lower forests; Schoenbrun, , ‘Early history’, 90–1.Google Scholar Sorghum is resistant to the waterlogging common in swales.

42 Cohen, David William, The Historical Tradition of Busoga: Muhama and Kintu (Oxford, 1972), 77ffGoogle Scholar; Phillipson, David W., The Later Prehistory of Eastern and Southern Africa (London, 1977), 219ff.Google Scholar

43 Ehret's work furnishes the basis for the following review of Southern Cushitic agricultural practice, especially Ehret, Christopher, The Historical Reconstruction of Southern Cushitic Phonology and Vocabulary (Berlin, 1980).Google Scholar For considerations of the role of Southern Cushitic-speakers elsewhere in eastern African history see also Ambrose, Stanley, ‘Archaeology and linguistic reconstructions of history in East Africa’, in Ehret, and Posnansky, (eds.), The Archaeological, 140–2;Google ScholarEhret, Christopher, Southern Nilotic History (Chicago, 1971), 37–9, 4851, 5560, 110–27Google Scholar; idem, Ethiopians and East Africans: The Problem of Contacts (Nairobi, 1974), 744Google Scholar; Robertshaw, Peter T. and Collett, David P., ‘A new framework for the study of early pastoral communities in East Africa’, J. Afr. Hist., XXIV (1983), 289301.CrossRefGoogle Scholar

44 There were several different communities of Southern Cushitic-speaking societies in greater eastern Africa. Ehret, Christopher and Nurse, Derek, ‘The Taita Cushites’, Sprache und Geschichte in Afrika, III (1981), 125–68Google Scholar; Nurse, Derek, ‘Reconstruction of Dahalo history through evidence from loanwords’, Sprache und Geschichte in Afrika, VII (1986), 267305Google Scholar; and Nurse, , ‘Extinct’, 96, 100.Google Scholar

45 The name replaces Ehret's ‘Nyanza Southern Cushites’, Ehret, , Ethiopians, 24–7Google Scholar, in favor of a suitably Cushitic word, but it follows his and Nurse's conception of this linguistic grouping. Tale Southern Cushitic belonged to the West Rift branch of Southern Cushitic (as for example, does present-day Iraqw of north-central Tanzania).

46 A quite detailed history of these contacts emerges from loanword studies; see Schoenbrun, , ‘Early history’, 190251Google Scholar, and Ehret, , ‘The African Great Lakes’, 1726.Google Scholar

47 Ehret, , Historical Reconstruction, 132.Google Scholar

48 E. coracana is a very durable crop. It is extremely drought-resistant and may store seeds on the head for up to ten years, though viability for planting drops 50 per cent after two years. It thrives in free-draining soils with steady moisture levels. It does best between 18 °C and 27 °C, from sea-level to 2400 m. In Uganda it requires 900–1250 mm of rainfall and dislikes waterlogging; FAO, Traditional Plant Foods, 266–7.Google Scholar

49 Proto-Rift Southern Cushitic vocabulary, and presumably Tale Southern Cushitic vocabulary, possessed detailed sets of terms for livestock breeding taxonomies, secondary products and probably a number of terms for specific cattle colors. See Ehret, , Historical Reconstruction, 338–84.Google Scholar

50 See Ehret, , Historical Reconstruction, 347.Google Scholar This noun turns up as a loan in Rutara Bantu, a descendant of proto-West Nyanza Bantu, after a.d. 800 or 900.

51 This sort of observation has been made most recently by Ehret, , ‘The African Great Lakes’, 52Google Scholar, and for pastoralists in the Kenya Rift by Ambrose, , ‘Archaeology’, 140–1.Google Scholar Later work provided an even better ‘fit’ between material cultural traditions, environmental zones and Southern Cushitic sub-groups, though the authors did not intend this as a conclusion. Compare Ehret's historical linguistic discussion of Southern Cushitic in Ethiopians, 7–31, with the distributional dimension of the four different traditions recognized in the new classification of Pastoral Neolithic pottery assemblages by Robertshaw, and Collett, , ‘A new framework’, 290 and 292.Google Scholar

52 Ehret, , ‘Patterns’, 35–9Google Scholar; idem, ‘Agricultural history’, 4ff; Schoenbrun, , ‘Early history’, 165–7.Google Scholar

53 For more on this well-known feature of the ancestral Bantu food system see Vansina, , Paths, 8892Google Scholar; Bennett, Patrick R., ‘On the reconstruction of Bantu technology and its vocabulary’, Muntu, III (1985), 121–35Google Scholar; Ehret, , ‘Lingusitic inferences’, in Ehret and Posnansky (eds.), The Archaeological, 61–2.Google Scholar

54 Yams cannot endure frost or waterlogged soils, and they thrive in warm, humid temperatures around 25–30 °C on deep, well-drained soils; Onweme, I. C., The Tropical Tuber Crops (New York, 1978), 1213Google Scholar; Purseglove, J. W., Tropical Crops: Monocotyledons (London, 1972), 108–10.Google Scholar

55 The oil palm will not tolerate frost either and thrives in more open, secondary forest or forest fringes where it may receive adequate sunlight. It will grow in riverine forests but not deep forests. It prefers a temperature range of 24–30 °C and an evenly distributed annual rainfall regime of about 2000 mm; Purseglove, , Monocotyledons, 485–7Google Scholar; Vansina, JanEsquisse historique de Pagriculture en milieu forestier’, Muntu, II (1985), 13.Google Scholar

56 Vigna subterranea is a synonym for Voandzeia s. and is commonly intercropped in eastern Africa, doing well on dry, inferior soils up to 1500 meters above sea level. It prefers an evenly distributed rainfall of 600–1000 mm and a temperature range of 20–28 °C, FAO; Traditional Food Plants, 510–11.Google Scholar

57 In eastern Africa, Vigna unguiculata tolerates droughts but not waterlogging, loves the sun, but is sensitive to cold, and frost kills it. It will stand a temperature range of 20–35 °C but prefers a range of 22–27 °C. It thrives at altitudes up to 2000 meters in a rainfall range of 280–4100 mm on many different soils but not alkaline. In the higher end of the rainfall range it is grown mainly for its leaves, with the seeds growing in importance to farmers in drier marginal lands. It is commonly intercropped with cereals and root crops; FAO, Traditional Food Plants, 518–22.Google Scholar

58 The most comprehensive presentation of these data is Schoenbrun, , ‘Early history’, 499518.Google Scholar

59 Ehret, Christopher, ‘Historical inferences from transformations in culture vocabu-laries’, Sprache und Geschichte in Afrika, II (1980), 194.Google Scholar

60 Oates, A. V., ‘Goats as a possible weapon in the control of thornbush’, Rhodesian Agricultural Journal, LIII (1956), 6885.Google Scholar

61 Ehret, , ‘Patterns’, 39Google Scholar; Wrigley, , ‘Cattle’, 250–62Google Scholar; Schoenbrun, , ‘Early history’, 505–8.Google Scholar

62 Linguistic data are available in Schoenbrun, Ibid. 516–18.

63 Sherratt, Andrew, ‘The secondary exploitation of animals in the Old World’, World Archaeology, XV (1983), 100.Google Scholar

64 This root has the additional meaning ‘large cow with long horns’ in all members of the Rutara branch of West Nyanza in which there is a reflex. This might refer to the breed we call Zebu, being either Bos indicus or cross-breeds of B. indicus and B. taurus. Both of these breeds tend to be smaller than the small-humped Sanga breeds. Fiona Marshall now feels confident that either of the two former breeds was present in the eastern Rift by 2000 years ago; see Marshall, F., ‘Rethinking the role of Bos indicus in Sub-Saharan Africa’, Current Anthropology, XXX (1989), 236.Google Scholar The biological advantages of B. indicus in times and regions prone to environmental stress are well known. Its apparently late in-corporation in the Great Lakes region lends support to the argument that risk minimization practices in the pastoral sector increased after a.d. 1000. For the latest classification of African cattle, see Grigson, Caroline, ‘An African origin for African cattle ? Some archaeological evidence’, African Archaeological Review, IX (1991), 119–44.CrossRefGoogle Scholar

65 See Schoenbrun, David L., ‘Cattle herds and banana gardens’, African Archaeological Review, forthcoming (1993).Google Scholar

66 This revises views stated in Schoenbrun, ‘Early history’, 196, and is more in line with Schoenbrun, Ibid. 227–34. For livestock in southern and western Uganda during the second millennium a.d. see Reid, Andrew, ‘The role of cattle in the Later Iron Age communities of Southern Uganda’ (Ph.D. thesis, Cambridge University, 1991)Google Scholar; and idem, ‘Ntusi and its hinterland: further investigations of the Later Iron Age and pastoral ecology in Southern Uganda’, Nyame Akuma xxxiii (1990), 26–8.Google Scholar

67 Vansina, , Paths, 86.Google Scholar

68 See Ehret, , ‘Patterns’, 38Google Scholar; Schoenbrun, , ‘Early history’, 490508.Google Scholar

69 Language data are laid out in Schoenbrun, Ibid. 509–13.

70 In the drier grasslands between the Kagera and the Katonga, settlement was probably sparse if present at all. See Andrew Reid, ‘The role of cattle’. But, in the lands between the lakeshore and the Kagera, south of its bend, occupation appears quite ancient. See Nelson, Charles and Posnansky, Merrick, ‘The stone tools from the excavation of Nsongezi rock shelter’, Azania, V (1970), 199–72Google Scholar; Schmidt, Peter R., ‘Early Iron Age settlements and industrial locales in West Lake’, Tanzania Notes and Records, LXXXIV/LXXXV (1980), 7794.Google Scholar

71 This is ‘language shift’. For the emerging debate on this issue, see Nurse, Derek, ‘Language contact, creolization, and genetic linguistics: the case of Mwiini’ (paper presented at the African Studies Annual Meeting, St Louis, MO, 1991)Google Scholar; and Thomason, S. G. and Kaufman, T., Language Contact, Creolization, and Genetic Linguistics (Berkeley, 1988).Google Scholar

72 Archaeological and palynological evidence also exist. Van Grunderbeek et al., provide fascinating though very thin evidence for a cereal-crop component in food systems of central Rwanda during the Early Iron Age; Grunderbeek, Van et al. , Le premier âge 27.Google Scholar And A. Gauthier produced the earliest direct evidence for cattle-keeping in the form of asingle jaw with a terminus ante quern date of a.d. 300. See Gauthier, A., in Noten, Francis van, Histoire archéologique (Butare and Tervuren, 1981), 32.Google Scholar The palynology is covered by Schoenbrun, ‘The contours’.

73 See Schoenbrun, ‘Early history’, ch. 3; Grunderbeek, Van et al. , Le premier âge, 620.Google Scholar

74 Schoenbrun, , ‘Early history’, 500, 509–10, 535.Google Scholar

75 Ehret, , ‘Agricultural history’, 22Google Scholar; Schoenbrun, , ‘Early history’, 500.Google Scholar

76 Schoenbrun, Ibid. 509–10.

77 Schoenbrun, Ibid. 510–11.

78 A development perhaps indirectly reflected in the palynological record from Kigezi. See especially Michael, E. S. Morrison and Hamilton, Alan C., ‘Vegetation and climate in the uplands of south-western Uganda during the later Pleistocene period, II: forest clearance and other vegetational changes in the Rukiga highlands during the last 8000 years’, Journal of Ecology, LVI (1968), 26Google Scholar; Hamilton, et al. , ‘Early forest clearance’, 165Google Scholar, discussed in Schoenbrun, ‘The contours’.

79 Schmidt, Peter R. and Avery, Donald, ‘Complex iron working and prehistoric culture in Tanzania’, Science, CCI (1978), 1085–9.CrossRefGoogle Scholar

80 Schoenbrun, Ibid. 133–6, 235–57.

81 Schoenbrun, , ‘Early history’, 526–7.Google Scholar

82 See Vansina, , Paths, 7383, 152–5Google Scholar; and see the important critique of Vansina in Ahmed, Christine, ‘Not from a rib: the use of gender and gender dynamics to unlock early African history’ (paper presented at the African Studies Association meeting, St Louis, 1991).Google Scholar

83 Ahmed, , ‘Not from a rib’, 78.Google Scholar

84 Schoenbrun, , ‘Early history’, 293.Google Scholar

85 This may well have been the context for the initial integration of the constellation of symbols and ritual process in the umuganuro ‘first fruits’ ceremonies that are shared by enough Western Lakes societies to suggest that umuganuro has very ancient roots. See Newbury, Kings and Clans, 200–26, 290 n. 16, 291 n. 23.

86 The data for these developments will appear in Schoenbrun, ‘Cattle herds’.

87 See Ehret, , ‘The African Great Lakes’, 6Google Scholar, and Schoenbrun, , ‘Early history’, 503.Google Scholar

88 See Ehret, , ‘The African Great Lakes’, 9Google Scholar, and Schoenbrun, , ‘Early history’, 502.Google Scholar

89 See Wrigiey, , ‘Cattle and language’, 250–3 and 256Google Scholar, where he comes close to the position taken here. I employ a different etymology than does Coupez, André, ‘Linguistic taboo concerning cattle among the interlacustrine Bantu’, Acts of the African Languages Congress (Pretoria, 1976), 226–7.Google Scholar