Early Agriculture in Africa

REFERENCES Linked references are available on JSTOR for this article:https://www.jstor.org/stable/41856943?seq=1&cid=pdf-reference#references... You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a widerange of content in a trusted digital archive. We use information technology and tools to increase productivity andfacilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available athttps://about.jstor.org/terms

Historical Society of Nigeria is collaborating with JSTOR to digitize, preserve and extend

access to Journal of the Historical Society of Nigeria

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Journal of thh historical society of nigeria vol. vi, no. 2 june 1972 EARLY AGRICULTURE IN AFRICA by Thurstan Shaw any consideration of the beginnings and development of agriculture in Africa the present state of our knowledge must largely be a survey of our ignorance reasoned essay in speculation. What is needed to advance the subject is moie evidence. The debate on the age of tropical 'vegeculture', for example, 'has raged some decades without much in the way of concrete evidence being put forward' (Harlan, 1969, p. 313). This paper will only have been worth while if it asks the right questions and serves to stimulate efforts to win that solid evidence. Reviews of the data have been published fairly recently (Mauny 1967; Clark 1967; Davies et al., 1968) and while there is no point in duplicating these, there need to look critically at the quality of the evidence; but a paper of this length pretend to be exhaustive. Evidence In considering the beginning and development of food production in Africa necessary first to consider the types of evidence which are available to us concerning this, and to realize that different kinds of data have different levels of reliability it comes to assessing from them the balance of probabilities concerning what actually happened. These different kinds of evidence have already been well categorized (Seddon, 1968; 489): (a) direct archaeological evidence - the remains of domesticated plants and animals in context; ( b ) indirect archaeological evidence - all other material discovered in an archaeo- logical context that, by its nature, suggests the presence of agriculture and a food-producing economy ; (c) evidence provided by botanical, stock-breeding, ethnographic and linguistic studies. Under (b) above one may include rock-paintings (which can sometimes be dated) and terracotta models of domesticated animals. Pieces of indirect evidence may be of different weight: an iron hoe or a cattle figurine excavated in a dated archaeo- logical context is of much greater worth and reliability than a grindstone, an undated rock-painting, or a 'digging-stick weight', since such perforated stones can be used for ether purposes. Caution has to be exercised in using the indirect evidence, and this has also been well emphasized (Seddon 1968, p. 489) :This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

The indirect evidence of material artifacts without the proof of cultivated plants and permanent settlement or of domestic stock is often misleading, and 'the short list of technological traits that occur only in association with agricultural economies (prepared by Meighan et al., 1958) demonstrates the difficulty of judging subsistence from indirect sources . . (Gabel 1960, p. 438). A warning against the uncritical use of indirect types of evidence to infer agriculture is provided by a list of traits commonly associated with agriculture but which are also possessed by some societies without domesticated plants or animals (Alexander 1969, p. 124.) As will appear below, Africa lags behind Europe, Western Asia and the New World in relation to archaeological research and in knowledge about the beginnings of food-production. Not only is there less actual evidence bearing specifically upon this particular subject, but in very few areas have there been established those cultural sequences and chronological frameworks through which alone it is possible to trace the patterns of origin and growth of agricultural and stock-raising practices. This was true until recently even of Egypt and the Nile valley, where one might have expected a clear and firm picture; yet the lack of this in the crucial period is one of the things which emerged from all the contributions to a review of predynastic development (Arkell & Ucko 1965). In some areas of Africa, too, radiocarbon datings are from one or two determinations, the reliability of which is less than for series and constellations of dates. The need for adjustment of radiocarbon ages to real ages in the light of the most accurately determined half-life of C 14 instead of the conven- tionally adopted one (Deevey et al ., 1963) and in the light of fluctuations revealed by bristle-cone pine dendrochronology (Stuiver & Suess, 1966; Reed, 1966; Neustupny 1970; Olsson 1971) also needs to be remembered; although for our present purpose, relative ages are more important than absolute ones. Theory This lack of evidence in Africa compared with other parts of the world, where much more is known about agricultural origins, means also that we have open to our choice a number of theoretical models - since different models, rather than a single one, have been shown to give the best fit to the evidence in different parts of the world. The concept of the 'Neolithic Revolution' was introduced by Gordon Childe in terms of simple diffusion; and a fairly straightforward pattern of diffusion still seems to be taken as the dominant mechanism for the beginning of farming in Europe (Grahame Clark 1965) although it is felt that necessary predisposing conditions must have been present for it to find acceptance (Grahame Clark 1969, pp. 62, 67-9). Similarly Balout has described the spread of food-production in the Sahara as a matter of simple diffusion. 'The Neolithic current moved from East to West' (1955). It must be remembered too that diffusion may be by cultural diffusion only or by a movement of peoples - or a mixture of both. Certainly it seems that in the Sahara there were complex movements going on in Neolithic times. Appropriately Hugot concludes a chapter on the Neolithic in the Hoggar by applying to the Sahara Cartailhac's earlier dictum 'Lorsqu'on parle du Néolithique comme d'une seule 144This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

entité on . . . semble oublier qu'il n'y a pas un mais des Néolithiques' (Hugot 1963, p. 168). Lastly, the diffusion idea may only be a preferred hypothesis, 'arising out of a legend rather than out of the data itself' (Higgs & Jarman 1 969, p. 36). On the other hand, the demonstration of independent centres of agricultural invention in the New World (MacNeish, 1965) and probably in eastern Asia (Gorman, 1969; Solheim 1969; Chang 1970, p. 182) has clearly shown that the idea of food production did not have a single origin in the world. If at least three independent centres have been demonstrated, it opens up the possibility of there having been moie. Furthermore, additions to the evidence from south-western Asia have changed the picture of a single, comparatively small nuclear place of origin for agricultural inven- tion, to one of a much bigger area (and a much longer span of time) in which condi- tions were right for innovations, numbers of which may have been made in different places within the general area and which may have interacted with each other (Higgs & Jarman 1969). But it seems that the most fertile advances in our understanding at the moment are coming neither from environmental determinism by itself, nor from cultural determinism - the idea that food-production is initiated when human culture has 'advanced' to a point capable of it. Rather, the concept of an interaction of the two processes, 'cultural ecology', or what has been called the 'Binford-Flannery' model, seems most productive of understanding of the process (Flannery, 1965, 1969; Wright, 1971). Perhaps it is helpful to think of a brush fire model of diffusion, in which there is a main area where the fire spreads and a number of rather haphazard patches of little fire, separated in space from the main fire and set off by flying sparks, i.e., the concept of a rapidly transmitted idea, or more probably a rapidly moving small group of people (or even one individual) the traces of whose movement are too small to show on the archaeological record, and who only reappear to produce a separated occur- rence of the diffused trait when they land up somewhere where the ecology is right or the socio-economic receptivity of the existing society is right, or both. From a very broad look at the distribution of the early food-producing communities, it looks as if they are ecologically, perhaps partly by choice, restricted to hill country on the edges of substantial uplands. Perhaps this relates to very local diversification of the environment in broken country, so that : ( i ) necessary control of water is obtained by choice of suitable site ; (n) the diversified environment means that various different natural resources are comparatively close to hand and do not involve wide-ranging migration to catch up with them at the right time of year ; (//) the country gives a much better chance of defending the immovable crop assets against passing or marauding gatherers. If this sort of ecological restraint (in addition, of course, to the environmental restraint on the crops) did operate strongly, then a brush-fire effect might be very marked. The apparent priority of the Saharan sites over Nile Valley ones would fit such a model. Lastly we have to bear in mind the mechanism of stimulus diffusion, in which culture contact with an idea or a principle serves to trigger off the development of an analogous practice rather than the importation of the practice itself (Chang 1970 p. 183). 145This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

All these mechanisms may have been operative in different parts of Africa at diffe- rent times in relation to food production and we must keep an open mind to see which model fits a particular set of data best. One needs to apply to Africa what has been said about the advent and spread of food-production in other parts of the world: These changes have to be seen in terms of both innovation and historical contact, so that we ask ourselves, on one hand, why certain food-collectors began to experiment with plant and animal domestication in Western Asia or Latin America; and on the other hand, why intrusive agricultural economies did not immediately replace hunting and gathering in Europe and the U.S. once they were introduced (Gabel 1960, p. 438). In the light of the above - our present lack of data and the choice of theoretical models open to us - it is better to proceed, not by trying to pursue the couise of abstracted entities such as 'agriculture', 'vegeculture', 'arboriculture' or 'pastoralism', but first by examining the evidence for domesticated plants and animals crop by crop and animal by animal. After having done this we may be in a position to see what it all adds up to in terms of subsistence patterns and food-winning economies, consi- dered in relation to the ecology of the area. (It has recently been shown that many of the defining characteristics of 'domestication', supposedly clear-cut, are in fact blurred (Higgs & Jarman 1969, pp. 31-2). This is the point at which we should bring in all the botanical, ethnographical, linguistic and geographical data, so that we can now consider food-production practices in terms of response to environment and as a part of a total cultural pattern. Because of the speculative and now discredited nature of some botanical or ethnographical theories about the origins of food pro- duction in Africa (e.g., Murdock 1959; Vavilov 1951) there was a healthy reaction to place more reliability on archaeological types of evidence (Seddon, 1968; Harlan & De Wet, 1971): this in turn produced a counter-reaction supporting 'a total geog- raphical orientation, emphasising ecology' (Yarnell 1968) and an excellent statement of the theoretical problems involved (Seddon 1969). Thus the two types of evidence should not be regarded as conflicting but as complementary, controlling each other. Complementary also to the need to obtain 'hard facts' in the African situation, emphasised in this paper, is the need to interpret the data in the light of generaliza- tions derived from cultural change elsewhere (Harris 1971). A review of our ignorance of this whole question in Africa points up also the need for those ecological studies which have been combined so profitably with archaeolo- gical evidence in south-west Asia to give new insights into the probable mechanics of the process; and the need also for harder evidence on climatic and vegetational changes in different areas of Africa, derived and substantiated from those areas and not merely inferred by extrapolation from elsewhere. PLANT SOURCES OF FOOD Direct evidence for production Direct archaeological evidence for food of botanical origin includes the datable contexts of actual remains of seed, fruit, root or tree crops, their 146This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

impressions of them in such material as pottery. The preservation of actual remains is largely a matter of lucky accident, thanks to aridity, carbonization and charring or water-logging, either of the material itself or in human and animal coprolites; the increasing use of the technique of flotation in excavations (Struever 1968) should help to recover more evidence of this kind than has been hitherto. This applies parti- cularly to seed crops; tubers and root crops are liable to remain much more elusive. African palynology has a long way to go to catch up with that of temperate regions, but some progress is being made (Assemien 1971; Bonnefille 1971(a) and (£); Bronckers 1967; Coatzee 1965; Guers 1970; Guinet 1968; Livingstone 1964; Livingstone and Kendall 1969; Lobreau et. al ., 1969; Maley 1970; Miée 1965; Quetzel 1960; Quetzel and Martinez 1958, 1961, 1962; Smith 1964; Sowunmi 1968; Van Campo 1957, 1958, 1960; Van Campo et al ., 1959, 1960, 1964, 1965; Van Zinderen Bakker 1965-9, 1967 {a) and ( b )). Studying seed impressions on pottery has already paid dividends in the Sahara area of Africa (Munson 1968, 1970; Clark, 1971(a) and could almost certainly be applied with profit to others, although it is not always rewarded (Fagan 1967, p. 62). Direct evidence for the early cultivation of crops in Africa is in total surprisingly small; and it falls into two groups, separated widely in time and space. The first comes from Egypt and the Sahara, extending from the sixth millennium B.C. to the fourth: two pollen grains of a cultivated cereal dated to between 6100B.C. and 4850 B.C. at Amekni in the Hoggar and identified as Pennisetum (pearl millet, bulrush millet), together with other pollen which might be Triticum (wheat) (Camps 1969, p. 188; 1971) - but one has to be cautious concerning pollen identifications of cereals, as the Gramineae are notoriously difficult in this respect; remains of emmer wheat ( Triticum dicoccum ), barley ( Hordeum ) and flax ( Linum ) from Fayum A radiocarbon-dated to the later part of the fifth millennium B.C. (Caton-Thompson and Gardner, 1934, pp. 34, 46-49 ; Seddon 1968, p. 490; Wendorf et al ., 1970, p. 1168); pollen at Meniet in the Hoggar dated to the mid-fourth millennium B.C. has been interpreted on account of its size (40 microns) to be a type of cultivated grass (Hugot, 1968, p. 485) but its domesticated status is now less certain since the finding of wild Sahara grasses with pollen grains up to 50 microns (Clark 1971a) ; emmer wheat and barley (and castor oil seeds, Ricinus communis , but these are probably wild) from the Badarian of Egypt said to date to at least the first half of the fourth millennium B.C. (Arkell & Ucko 1965, p. 150) and for which Hugot (1963) quotes a radiocarbon date of 3160 B.C. (Gro-223) but which is not certainly identified with any horizon and so cannot be an indication of the age of the settlement ; emmer wheat and barley from the unsatisfactory site of Merimda with radiocarbon dates from 41 80 B.C., but in which the excavator may have 'confused a very early culture with later ones' (Baumgartel 1955; 1965), a likelihood perhaps made greater by the presence of club wheat ( Triticum compactum) 'not known till much later anywhere in the Near East except at certain Lower Egyptian sites' (Helbaek 1955); the Predynastic cemetery at El Mahasna produced 'grain' (Ayrton & Loat, 191 1, p. 18) and the cereal from the Predynastic grain kilns of Abydos was identified as wheat ( Triticum vulgare) (Peet & Loat 1913, pp. 1-7). To the foregoing should be added the carbonised fragment of the pericarp of the oil palm, Elaeis guineensis , found in the excavation of Shaheinab in the Sudan with a mean radiocarbon date of 3300 147This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

B.C., 'brought from West Africa 01 the Congo for its food value, although it is con- ceivable that ecological conditions at this time were suitable for wild growth locally' (Arkell 1953, p. 105; Arkell & Ucko 1965, p. 149). Clark (1971(6)) has recently reviewed the evidence for the Nile valley. The second group of evidence all comes, with two exceptions, from the interior of south-eastern Africa and extends forward in time from the middle of the first millen- nium A.D. The exceptions both concern sorghum ( Sorghum bicolor; Guinea corn, Kaffir corn, great millet), first from the late first millennium A.D. (Gif-1292, A.D. 750±100; Kl-293, A.D. 860±65) at Niani, Republic of Guinea (Filipowiak et al., 1968, pp. 617, 645); and second, from a ninth/tenth century A.D. level at Daima, north-eastern Nigeria (Connah 1967, p. 25). In eastern Africa, Engaruka, with radiocarbon dates ranging from the fourth to the nineteenth century A.D., has produced carbonised sorghum at all levels, sufficient- ly well preserved for it to be possible to identify the types represented (Sassoon, 1967, 1971). In the area of south-eastern Africa, we begin with 'what are tentatively identified as a squash seed and a possible bean' from.Chundu (Zambia), believed from its pottery typesto belong to the period from the sixth to the eighth centuries A.D. (Vogel, 1969); Inyanga ruins, belonging to the eighth century A.D. or before, produced car- bonised remains of Sorghum sp., Pennisetum typhoides, Eleusine coracana (finger millet), Voandzeia subterranea (groundnut, peanut), Vigna unguiculata (= Vigna sinensis, cowpea), Citrullus vulgaris (Kaffir melon) and Ricinus communis (Summers, 1958, pp. 175-7); charred seeds of Sorghum sp. and possibly a seed of Pennisetum typhoides come from Mwamasapwa in Malawi, a site considered to belong to the ninth century A.D. on the similarity of its beads to the Leopard's Kopje industry of Rhodesia (Robinson, 1966(a), p. 180); seeds oí Sorghum sp. come from Isamu Pati and Kalundu of the Kalomo tradition of Zambia dated to the period from the ninth to the eleventh centuries A.D. (Fagan 1967, p. 62); the site of Mapungubwe produced carbonised remains of Sorghum sp., Sclerocarpa caffra ('marula', a wild nut), Pseudocadia zambesiaca (an indigenous tree), Citrullus sp. (melon), Vigna unguiculata, and Grewia sp. ('Kruis bessie' - unlikely to have been cultivated) (Fouché, 1937, p. 31); the site has hitherto been regarded as late fourteenth century but it may be up to 300 years earlier on account of the isotopie fractionation of the sorghum whose remains provided the radiocarbon samples (Van der Merwe, 1971). The site of Ingombe Ilede, where the gold burials are now dated to the fourteenth/fifteenth centuries (Phillipson & Fagan, 1969), produced carbonised seeds of Sorghum sp. in levels 4 and 5 and a leaf impression of sorghum with burial II/3, but as date R-908 (A.D. 680 ±40) was actually obtained from a specimen of charred sorghum, this gives a seventh century date for sorghum from the earlier period at the site (Fagan et al., 1969; pp. 81, 85), Klipri- viersberg produced three seeds of Sorghum sp. (Mason 1 967) ; many thousands of seed or grains of Sorghum sp. were recovered during the 1971 excavations at the Olifant- spoort Iron Age site, 60 miles west of Klipriviersberg and dated to the period from the sixteenth to the nineteenth centuries A.D- (Mason 1971); seventecEth/eightccrilh century levels of the Inyanga ruins complex produced maize {Zea mays) (Summers, 1958, pp. 175-77). In connection with the probability that agriculture was in fact, practised at Bambandyanalo at the beginning of the eleventh century A.D., in spite 148This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

of its denial by the excavator, it has been pointed out that on many sites 'abundant domestic animal remains make it easy to overstress the importance of stockbreeding, even to the exclusion of agriculture' (Fagan 1964, p. 343). In addition to the above, there are various pieces of relevant direct evidence, although they do not provide any unambiguous proof of domestication. At the K6 rock shelter near Kintampo in Ghana, seed husks of Ceitis sp. (African nettle tree, African false elm) were found in the Punpun phase dated to before 14C0 B.C. and in the succeeding Kintampo culture levels vegetable remains include husks of Elaeis guineensis and cowpeas, but these are small compared with those now grown in the area and it is not possible to say whether they are wild or cultivated (Flight 1970, p. 72). There have been many finds of Ceitis sp. in archaeological contexts (Hugot, 1968, p. 486) but no evidence of domestication. It is unlikely that the atili seeds ( Canarium schweinfurthii) associated with Nok culture finds indicate its formal cul- tivation (Fagg. 1959, p. 289). At the Njoro River Cave among remains of the Stone Bowl people dated to the beginning of the first millennium B.C. were found the remains of calabashes, Lagenaria vulgaris, (Leakey & Leakey 1950, p. 38), and possibly at another Stone Bowl site at Ilkek (Brown, 1966, p. 66). In Malawi Vigna unguiculata is reported from Nkope, layer 1 (dating therefore to c. A.D. 775±1C0; Robinson 1970, p. 171). What emerges from the above, as so often in an area where archaeological work is uneven, is more a pattern of our present knowledge and ignorance than of what actually happened in the development of food production in the African continent. It is perhaps not so surprising that we are entirely without direct evidence in the whole of the humid area of the Congo basin and the West African rain forest, where archaeo- logical work is difficult, and acid and corrosive soils militate against preservation, but it is very regrettable that we are without direct evidence from Ethiopia, the Horn and East Africa concerning early crops. It is clear that wheat and barley were introduced into Egypt from Asia, probably not later than the sixth millennium B.C. Whether the fact that dates for cultivated cereals in the central Sahara are earlier than in Egypt represents a true state of affairs is open to question. It is widely believed that the sites of the earliest agriculturalists in the Nile delta and in Middle Egypt are buried beneath metres of later silt (Butzer, 1965); the same reason used to be given for the apparent absence of Late Palaeolithic industries in Upper Egypt and Nubia, an absence now shown to be illusory (Wendorf et al , 1970, p. 1161.) As indicated above, present evidence is ambiguous about the early presence of domesticated wheat and Pennisetum in the Hoggar; and at present we are not sure whether during the northern expansion of the tropical monsoon rains, sorghum, Pennisetum and Eleusine were included in the flora of the Hoggar and the central Sahara, predominantly of Mediterranean type, during the Holocene up to the final desiccation, or whether they were confined to the Sahel/Sudan belt as now- adays. This raises a fundamental issue concerning the domestication of the tropical grasses; for another thing which emerges from the foregoing survey of the direct evidence from Africa is that by the later part of the first millennium A.D. sorghum was established as a domesticated crop in northeastern Nigeria and south-eastern Africa, and probably Pennisetum and Eleusine as well. The issue is whether the tropical grasses became domesticated as a result of (a) local interaction with the environment 149This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

without outside stimulus ( b ) experimenting with local resources as the result of the stimulus of the idea of food production received from outside (c) attempts to grow wheat and barley and these being supplanted by their own weeds (Darlington, 1969, pp. 68-9); but tropical weeds in fields of wheat and barley, which must be giown in the cool season, are regarded as unlikely (Harlan 1971). There are now two sources of evidence not yet mentioned which fall into the 'silent millennia' (Darlington, 1969) between our two groups of evidence and which have a bearing on this question. At Adrar Bous in the Air region of the Sahara, pottery impressions have yielded a single grain of Brachiaria dated to c. 4000 B.C. and one of cultivated Sorghum dated to 2000 B.C. (Clark, 1971a). Secondly, in settlements in the Dhar Tichitt region of Mauretania spanning from the mid-second to the mid-first millennium B.C., and divided into seven phases, impressions on pottery revealed the use of the grain of a number of desert grasses. In the first three phases there were many seeds of Cenchrus bifloris (bur grass), still collected to some extent as a famine food, and a single grain of Pennisetum sp. but it was impossible to tell whether it was wild or cultivated. In the fourth phase, Brachiara deflexa and Panicům laetum also appear, with Pennisetum rising to 3 per cent. In the fifth and sixth/seventh phases Pennisetum impressions jump to 60 per cent and 80 per cent respectively and have definite characteristics of cultiva- ted grains, with Panicům turgidum also appearing in the last phase (Munson 1968, 1970). At first sight it looks as if the Tichitt communities were experimenting with the local grass seeds and finally hit on Pennisetum as the best, and as if this was an inde- pendent discovery of how best to exploit the environment, not connected with the use of cultivated Pennisetum only 800 miles away in the Hoggar some four thousand years earlier; but it could have been a delayed introduction from that area or from elsewhere. It has been suggested that impressions on pots excavated at Ntereso in Ghana are of Pennisetum, but the stratigraphy, associations and the claimed dating of the 2nd millennium B.C. are somewhat in a 'suspense account' at the moment (Davies 1968, p. 481; Shaw 1969 (a), p. 228). Returning to our review of the direct evidence in sub-Saharan Africa, we notice that before the end of the first millennium A.D. there are records of cucurbits, beans and groundnuts. As with the millets, their domestication undoubtedly goes back much further than the dates of our earliest evidence. We have already noted wide geographical and chronological gaps in the record provided by direct evidence, and there is one important botanical gap as well - and that is that we have no archaeological evidence for the use of yams and other tubers at all. Nor, because of the very nature of these crops, is the prospect of getting such evidence very hopeful, although perhaps one day some recognizable charred yam may be recovered in a significantly dated archaeological context. Pollen evidence is likely to be either lacking or equivocal since the domesticated varieties tend not to flower (Shaw 1968, p. 501). Nor is the equipment for preparing tuberous crops as likely to provide indirect evidence as that for preparing grain crops, since wooden pestles and mortars replace grinding stones; not only is their evidence ambiguous, since pestles and mortars are also used for grain crop food preparation, but above all chances of preservation are low - although not non-existent (Fagg 1 965, p. 23). Since the indigenous 150This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

domestication of certain yams was probably one of the most important events in the development of agriculture in West Africa, for this we have to rely on our third type of evidence - the botanical, ethnographical and linguistic (see below). Indirect archaeological evidence for production of plant foods Before we can settle the question as to whether Pennisetum was separately domesti- cated in the Tichitt area and in the Hoggar (see above pp. 3, 5) we need more evidence, but there is some interesting indirect evidence relating to this whole question and it takes us back to an earlier period than any we have yet considered. On the evidence of grinding stones and other circumstantial evidence it has been claimed that 'near the end of the Nubian Final Stone Age, techniques were developed to permit the use of the wild grain which grew along the Nile, and thus to make possible the exploitation of a new and rich source of food' (Wendorf 1968, p. 1056). It is suggested that this practice goes back along the Nubian Nile at least to the twelfth millennium B.C. and thus appears to be an entirely indigenous response to the problems of winning food from the local environment. However, this theory is considered by some to rest on very slender evidence (Arkell 1969, p. 488). The question that has to be asked is, if this utilization of grain occurred earlier in the Nile valley than anywhere in south- western Asia, why did it not lead to settled villages, and to domestication ? The answer given lies in those changes of climate which meant that the wild grasses on which those communities were dependent became less and less abundant (Wendorf 1968, p. 1059) - although one school of thought would argue that this itself should have provided the very stimulus for domestication. However, perhaps these Nubian sites are not in fact earlier than the earliest grain-using in south-west Asia, since sites with grinding stones have been found in the southern Negev dated to around 15,000 B.C. (Wendorf 1971). With the slightly moister climate of the sixth millennium there is again evidence of a similar kind in the silt areas of the Dungul region for the specialized gathering and milling of wild grass seeds, especially perhaps Panicům turgidum , which has now retreated to the highlands of Tibesti (Hobler & Hester, 1969). The authors referred to suggest that there may have been an independent evolution in this area from intensive food-gathering to food-production and that it was from this area that ideas of food-production were introduced up and down the Nile valley, reinforcing their argument from the dissimilarity between the food-producing cultures of the Fayum and Khartoum on the one hand and those of southwest Asia on the other. However, two things go against this argument : the fact that the earlier intensive exploitation of wild grains in Nubia did not lead to food-production; and the fact that wheat must have been imported into Africa from Asia. From the 1968-69 season at Jebel Uweinat, no direct evidence was obtained. However 'the ecology (and also the presence of Panicům turgidum in great quantity, close to the neolithic and older sites) would suggest a stage of vegeculture. It appears that Uweinat was more isolated than previously thought, but could be regarded as an area where tropical grasses might have become domesticated as a result of local interaction with the environment without outside stimulus' (Van Noten, 1971). Of course there is a great deal more indii ect evidence, much too much to enumerate in a paper of this length. Increased archaeological knowledge has shown us not only the unreliability of former simple definitions of the 'Neolithic', in which it was implied 151This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

that food-production could be inferred from the presence of certain technological traits such as pottery and ground stone axes, but also that 'agriculture' is not a unity - i.e., that food-production has many forms and that we have to be more precise; and even grinding stones can be used for grinding other things beside food, such as pig- ments, and we know that they were; nevertheless, ethnographic parallels suggest the possibility of dual use and that true quern forms only result from the persistent heavy grinding of grain, not from pigment grinding alone (Wendorf et al ., 1970, p. 1170). In spite of the need for such caution, the finding of tool-kits, admittedly with many regional variants and facies, belonging to what is usually refened to as the 'neolithic' of north Africa and the Sahara, is useful presumptive evidence for filling in many of the gaps left by the exiguous record of direct evidence. In upper Egypt there are sites dated to the eleventh millennium B.C. which yield grindstones and abundant pieces with lustrous edges believed by some to be sickles; and a pollen profile associated with one of these sites has a sudden increase in the frequency of a large grass pollen which the palynologist is '95 per cent sure is barley' (Wendorf 1971). The Terminal Palaeolithic cf the Fayum dated to the sixth millennium B.C. has abundant grind- stones, but no direct data so far on food production (Wendorf 1971). The stone mullers which apper about 4800 B.C. in the Haua Ftea record are suggestive of grain milling (McBurney 1967, p. 298), and the increased number, size and wealth of Naqada I settlements between 3800 B.C. and 3600 B.C. have been interpreted as indicating a growth in the importance of agriculture over Badarian times (Kaiser 1957; Arkell & Ucko 1965, p. 153). Clay models of garlic appear in Predynastic cemeteries in Egypt (Ayrton &Loat 1911, pp. 18-29). Although there are no radiocarbon dates for Early Khartoum, it looks as if pottery is earlier in the area stretching from there through Ennedi right across the Sahara into western Algeria than it is anywhere in the Nile valley from Nubia to the Fayum, since in the latter pottery does not appear before 4000 B.C. (Wendorf 1971) where- as for the Sahara sites there is now a collection of dates from the fifth and sixth millennia B.C. (Camps 1969, pp. 206f.). When we move south of the Sahara we are in a much more difficult position with regard to handling our indirect evidence. One cannot always be sure that stone imple- ments classified as hoes on the basis of their morphology were in fact used for tillage, whether as hoes or as digging stick blades (Hugot 1968, p. 484) and one questions whether it is justifiable on the basis of such tools alone, mostly collected as surface finds, to create a 'culture' and give it a date and a name (Davies 1964, pp. 203-30). Perhaps some edge-damage studies could throw light on the question. The Nachikufan Industry of the woodland savannas of Zambia and Malawi, extending from about 8000 B.C. to 1600 A.D. has bored stones that could have been used as digging stick weights but could equally well have been used in the making of some sort of animal trap; there are also edge-ground stone axes (Clark 19706, pp. 177-8). Stone axes with their cutting edges made by grinding have been invented independently in more than one part of the world, and it is possible that this may have happened in sub-equatorial Africa, although some of the forms found in East Africa bear resemblance to the Egyptian lugged axe (Leakey 1943; Brown, 1969; Van Noten, 1969). It has been suggested that the infilling of the valleys with the deep deposits containing remains of the Nok culture in Nigeria was the result, not of 152This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

climatic change, but of sheet and gulley erosion following extensive clearing of the vegetation cover by agriculturists during the later part of the first millennium B.C. (Shaw 1964, p. 455) and the same thing has been suggested for northeast Angola (Clark 1968, p. 146-8). The presence of grindstones and quern fragments has been taken as "demonst- rating' the practice of agriculture at the northern Malawi site of Phopo Hill, dated to the third to fifth centuries A.D. (Robinson & Sandelowsky, 1968, p. 134), but it can only be classified among the indirect evidence. The stone rubbers at Ngorongoro (Leakey 1966; Sassoon 1968 (a) and ( b )) were probably used for grinding ochre because they have red ochre ingrained in their surfaces; the bowls do not look as if they were used for grinding or pounding. It may be possible to establish that certain types of querns are associated with agriculture; deep dish-querns are only found on agricultural settlements in south-central Africa (Clark 1971a); in southern Zambia, light-weight flat ones, similar to late Stone Age types, continue until late in the Early Iron Age sequence, while deep querns are probably to be associated with maize cultivation (Vogel 1971). An interesting piece of pollen-analytical evidence comes from cores taken from Lake Victoria, which indicate that about three thousand years ago there was a sharp decline of forest species with a corresponding rise in grass pollens, while another genus - Acalypha - which is an early-stage pioneer in the regeneration of forest after gardens have been abandoned, also increases in abundance (Clark 1970(6), p. 206). This is suggestive of agriculture, especially in view of the plausible date, but it cannot be taken as conclusive, since elsewhere a decline in tree-pollen and an increase in the ratio of grasses to trees must be due to natural causes, e.g., in the Hoxne Interglacial of eastern England (West & McBurney 1954, p. 135). The earliest recognizable iron hoes come from Zambian sites: Kapwirimbwe dated to the fifth century A.D. (tang only) (Phillipson 1968), Kumadzulo dated to the sixth/seventh centuries (Vogel 1969), and Chundu dated to the early eighth century (Vogel 1970). The indirect evidence for a specialized form of agriculture at Engaruka (in northern Tanzania) is still rather tantalising, in spite of recent work there; two radiocarbon dates from the terrace sites on the hillside belong to the first millennium A.D., but it has not been possible to link the system of fields and enclosures to the hill-side terrace-platforms; contrary to what was formerly supposed, careful examination has thrown doubt on whether the fields were irrigated except close to the stream (Sassoon, 1967). There may be a possible new line of evidence for millet agriculture, from an elevated C-13 content in skeletal material, due to the constant eating of isotopically frac- tionated material (Van der Merwe 1971). From the archaeological evidence that we have, it is clear that the millets became important in sub-Saharan Africa, but we do not know for how long this had been the case, in different areas, before we first actually meet them at the end of the first millennium A.D. Was some form of domesticated millet in use in sub-Saharan Africa for 500 or for 5,000 years before that ? At this point we have to take our first look at our third type of evidence - the botanical, ethnographic and linguistic. 153This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

Botanical , ethnographic and linguistic evidence for the production of plant foods Botanical evidence has been adduced to show that there is a coincidence of the greatest variability of the wild and cultivated sorghums in the area south of the Sahara, north of 10° S and east of 25° E, and on the basis of Vavilovian theory the implication was that it is in this area that we must look for the origin of the crop (Doggett 1965, p. 58). most probably in the northern part of it (Hemardinquer et al., 1967; Harlan, 1968), and perhaps specifically in Ethiopia (Doggett 1970, p. 2). On the other hand, 'comparative morphological studies indicated that sorghum was probably ennobled independently in at least three African regions, from three morphologically different prototypes' - in the Ethiopian region, tropical West Africa north of the rain forest, and south-east Africa (De Wet & Huckabay, 1967, p. 800). At one time it was thought that linguistic evidence precluded any belief that sorghum reached India earlier than the first millennium B.C., or at most 1500 B.C. (Watt 1893, 6 :291 ; Doggett, 1970, p. 7), and Allchin quotes it as a late-comer in India, not appear- ing until the opening of the Christian era (Allchin 1969, p. 325). However, other evidence has been adduced to claim that it had already reached India by the mid- second millennium B.C. (Marshall 1931, Plate LXXXVII, Photo 5; Vishnu-Mittre, 1 968). That it reached India initially as ships' provisions in early sea-borne traffic seems more likely than that it travelled all the way overland (Doggett 1970, p. 7). (The intro- duction of the cultivated kapok tree {Ceiba pentandra (L.) Gaertn. var . pentandra from West Africa to southern Asia seems to have been a much later development, for which the Arabs were responsible (Baker 1965).) Sorghum reached Mesopotamia and the Middle East from India, and only arrived in Egypt in Roman Byzantine times, so it did not travel from tropical Africa to the Middle East through Egypt (Doggett 1965, p. 61). Thus in the area of north-east Africa specified, we have to think of Late Stone Age people being responsible for the domestication of sorghum. Doggett sees Cushitic people as growing emmer wheat and developing some of the local plants, including Eleusine , teff (Eragrostis abyssinica) and sorghum, for use in the areas less well suited to wheat (1965, p. 59) but it can be argued that crops such as Ensete (African banana, false banana) and teff must have been domesticated before the introduction of wheat and barley to Ethiopia or else they never would have been cultivated. Eleusine coracana was formerly thought to be of Indian origin (De Candolle, 1886; Burkill, 1935; Werth 1937; Porteres 1951) but recent work suggests that it originated as a domesti- cated plant in Africa (Kennedy-O'Byrne, 1957; Mehra, 1962, 1963«, 19636) and was transmitted to India by way of the Sabaean Lane (Anderson, 1960). The whereabouts of the wild ancestors of Pennisetum and its first domestication has long remained something of a mystery (Harlan, 1968), but its story may well be the same as that of Eleusine coracana. Doggett now sees both Eleusine and Pennisetum, as well as Sorghum , as originating in the Ethiopian area and being transmitted thence to India by sea (1970, pp. 2, 7). On the other hand, Dr Rachie's investigations, while they suggest to him that Ethiopia was indeed the centre of origin for cultivated Sorghum and Pennisetum (with the possibility of a second centre for Pennisetum towards the western end of the Sudan zone), they do not provide the same evidence for Eleusine - for which there are many wild varieties in south-west Uganda and adjacent regions (Rachie 1971). Pennisetum has been recorded in Rajasthan and Gujarat in north-west India during the second half of the second millennium B.C., and Eleusine in the middle of that millennium from 154This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

Mysore (Vishnu-Mittre 1968). Most of the millets grown in Bantu Africa today are said to be of eastern varieties, and agronomists stress their difference from the western (Grundeman 1968, pp. 2, 4). Digitaria exilis ('fonio', 'hungry rice') is not known wild but is cultivated as a secondary cereal throughout the savanna zone from northern Nigeria to the Atlantic, as well as, on a much smaller scale and in a much more rest- ricted area, D. iburua (Dalziel 1955, p. 526). Clark (1971(6), pp. 67-70) has recently reviewed the ethnographic evidence on the use for food of the grain of the grasses of the Sahara, the Sudan zone and the Nile valley. Just as it is likely that many tropical grasses were at one time giown for their grain, so it is likely that a number of African tubers were formerly much more widely cultivated than now and have come to be displaced by more successful ones; Coleus dysentericus seems to be a case in point (Busson 1965, p. 405). The yam-bean ( Sphenostylis stenocarpa) is widely cultivated in tropical Africa, both for the seed and for the tuber (Dalziel 1955, p. 262; Busson 1965, pp. 245-8). It seems, then, that Late Stone Age peoples of north-eastern Africa south of Egypt were cultivating sorghum, but we have little evidence about its spread from here to other parts of Africa. Gloss-edged trapezoids, looking very much as if they are parts of a composite sickle, make their appearance (along with pottery and ground stone axes) in the Iwo Eleru sequence of Nigeria after about 3000 B.C. but sickles are not suitable, and not normally used, for cutting sorghum, and the trapezoids may have been used in a composite implement adapted to the purpose or for gathering wild or other cultivated grasses (Shaw 1969 (è), p. 369; and forthcoming). The same can be said of possible obsidian sickle blades found in burial mounds of the Stone Bowl culture in the Ngorongoro Crater of Tanzania dated to the second half of the first millennium B.C. (Sassoon 1968a, 19686). The stone bowls of this culture also may or may not be associated with agricultural practices; at Ngorongoro they were probably used for grinding ochre as they were very smooth and had ied ochre engrained in their surfaces ; one was blackened inside. It is tempting to think that the makers of 'Uelian' stone axes were agriculturalists, living on the north-eastern border of the rain forest, but in fact we know nothing else at all about this 'culture', neither its date nor the character of the rest of its material equipment (Van Noten 1968). Apart from these two rather ambiguous pieces of indirect evidence, and ths fact that the teeth of the skeleton from the Rop Rock Shelter in Nigeria, radiocarbon-dated to 25 B.C. ± 120 are said to be those of an agricultuialist (Gaherty 1968), we have very little idea how the practice of millet cultivation may have spread among people with a Late Stone Age technology in the savannahs of West Africa, central Africa (i.e., the Central African Republic) and parts of east Africa. Nevertheless, although we know so little as yet about the details, we can feel reasonably confident that some such spread did occur. The Kotoko living south of Lake Chad regard Pennisetum as their oldest kind of food grain (Lebeuf 1969). It is assumed that the first known iron-users in sub- Saharan Africa, the Nok people of the second half of the first millennium B.C. in Nigeria, were agriculturalists; two terracottas represent fluted pumpkins ( Telfairia occidentaliis) (Fagg 1956, p. 1086; 1959, p. 289). In most of Africa south of the equator, the archaeological record usually shows an iron technology succeeding Late Stone Age industries of a type believed to be indicative 155This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

of a hunting, fishing and gathering economy (sometimes highly specialized and per- haps semi-sedentary, e.g., Ishango (Heinzelin 1957)), rather than of a food-produc- ing one. Hence the idea has developed that in most of sub-equatorial Africa a know- ledge of iron technology and of food production were diffused together - and that the agents of this diffusion were speakers of Bantu languages (Oliver 1966(a), 1966(ò)). This idea of Oliver's was an attempt to put together the results of linguistic work done by Greenberg (1955) and Guthrie (1962) and Murdock's suggestion that the expansion of the Bantu-speaking peoples was the result of their possession of superior food-resources (Murdock 1959, pp. 290-1). Greenberg had decided that Bantu languages belonged to the Western Sudanic family and that the more easterly peoples of this language group, as a result of the general southward trend, had moved into the Congo basin and thence into the rest of Bantu-speaking Africa. Guthrie on the other hand, as a result of his linguistic studies, saw the Katanga and an east/west oval area around it as the centre of dispersion. Oliver mairied the two pieces of linguistic research by postulating a small band of pre-Bantu speakers moving rapidly through the Congo forest from a primary dispersal centre in south-east Nigeria and Cameroon to a secondary centre in Katanga. This is an attractive hypothesis, but there are a number of difficulties about it which have been enumerated (Posnansky 1968); in particular, in relation to the subject under discussion, the pre-Bantu making their quick dash through the forest are supposed to have been millet-growers and might have found this difficult in a different ecosystem (Shaw 1969(a), p. 229), whereas the pattern of distribution of Eleusine might be regarded as favouring a route around the forest along its northern and eastern margins (Portér es, 1951). Anthropological evidence confirms the general pattern of movement but gives no preference for a north-south route through the forest or a clockwise movement around it, and tells us nothing about the sources of subsistence of the travellers (Hiernaux 1968). The key to a through-the-forest route may be the Nana-Ekeia-Sanga waterway leading from forest savannah mosaic in Cameroon and Western Central African Republic to the similar zone of the Middle Congo, a distance of less than 500 miles; trading contacts of a later date along this route are suggestive that it may have been in use earlier (Vansina, 1970). Study of the early Iron Age in the pre-Bantu homeland area of Cameroon has suggested an agricultural economy tied to the river valleys (David 1971); the area of land for expansion (or even of maintenance, if soils become exhausted or weeds of cultivation become too difficult to cope with, cf. Carneiro 1961, p. 57) would in this case have been limited, and tended to push settlement up the valleys of the upper Logone/Chari system, whose headwaters rise comparatively close to those of the Nana and Lobaye, which run south into the Congo. What is of importance is that, extending in a rather broken fashion from the area of Kinshasa through north-eastern and eastern Angola into Katanga and adjacent parts of Zambia, there is a great stretch of grassland which was at one time a puzzle to geog- raphers but which has been explained as the result of early agricultural activities (Gray 1962, p. 184); and it will be noted that this is largely Guthrie's nuclear area for Bantu expansion. It now seems likely that both the northern group of early Iron Age industries in East Africa (Dimple Base Kwale, Sandaweland) as well as the southern group (Kamadzwo, Ziwa, Gokomere, etc.) in the Zambia/Rhodesia area, 156This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

are, on the basis of their pottery, derivatives of a co-tradition from another area, rather than that they influenced each other directly in a north and south direction (Huffman 1970). This other area could well be Guthrie's nuclear area. Some doubts have been expressed about the idea that it was Bantu-speakers who spread both a knowledge of iron and of agriculture throughout sub-equatorial Africa. Not many people now subscribe to the idea of earlier writers (Johnston 1913, p. 413; Wrigley 1960) that it was the possession of the iron hunting spear which gave such superiority to its possessors as to enable small parties of them to establish their language and their leadership over others by being fabulous hunters and meat- providers, but there is some evidence that iron technology may have been introduced into sub-equatorial Africa by more than one route - one from West Africa and by an eastern route from north of the Great Lakes (Summers, 1969; Posnansky, 1968, p. 11); that the earliest iron-using peoples in southern Africa were not all negroes (Gardner 1963; Posnansky 1968); that agriculture was first introduced into Bantu Africa prior to the adoption of metallurgical techniques' (Grundeman, 1968, p. 4); and that later Bantu expansion into East Africa came into an area already partly occupied by food-producing societies, especially in the central Rift Valley of Kenya (Posnansky 1968). We cannot even be sure that the presence of negroes definitely implies agriculture, as the example of the foraging Hadza shows (Posnansky, 1968, p. 9). However, recent radiocarbon dates for east Africa, Zambia, Rhodesia and down into Swaziland 'serve to emphasise the basic contemporaneity of the introduction of the Early Iron Age throughout the area of its distribution' (Phillipson, 1970a, p. 5) which means the first few centuries of the first millennium A.D. ; the archaeologist who in 1964 wrote 'Iron Age technology and economic practices spread into southern Africa faster than either the language or the physical type associated with them in the first place' (Fagan 1964, p. 359) is not inclined now to think that this is true of southern Africa to any great extent (Fagan 1971); and in this area the accumulated evidence indicates that Iron Age culture was introduced as a fully integrated system, incorporat- ing food-production, metallurgy and a tribal organization with very little participation by previously stone-using culture bearers (Vogel 1971). Hitherto, nothing has been said about tuberous crops - for the simple reason that we have no direct or indirect archaeological evidence about them at all - so we have to rely on our third type of evidence for one of the most important classes of cultigens in Africa. We do not have to concern ourselves with cassava, sweet potatoes, or Xantho - soma (coco-yam with sagittate leaf) which are introductions to Africa from the New World since its discovery, nor with Colocasia (coco-yam with peltate leaf, or taro), one of the introductions from Asia (Coursey 1967, p. 7). Coleus (Hausa potato, Kaffir potato) is probably indigenous to Africa; little is known about its area of origin (Dalziel 1955 p. 459) but Abyssinia has been suggested (Davies 1968, p. 481). The important cultigen is the yam, whose name has an interesting history (Burkill, 1938). What has tended to obscure the history of this genus in Africa is that until recent years it was believed that the use and domestication of the indigenous African yams, Dios - corea cayanensis and D. rotmdata was only brought about by the stimulus of the intro- duction of the Asian forms, D . alata and D . esculenta (Forde 1953, p. 211; Gray 1962, p. 183) although the latter is quite a recent introduction (Morgan, 1962, p. 236; Lawton 1966). Alternatively, as a result of thinking in terms of 'the introduction of 157This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

agriculture', it was envisaged that the African yams were only developed when the idea of cultivation reached the 'natural' yam-growing areas of West Africa (Morgan, 1962, p. 238), following the domestication of Oryza glaberrima (African rice; see below). Even the less good indigenous D. dumetorum is grown in West Africa as much as the Asiatic D . esculenta (Coursey 1967, p. 8). Murdock (1959, pp. 222-3) speculated that the Asiatic food plants entered Africa via the Ethiopian lowlands, travelled westwards along what he erroneously called 'the yam belt' (Morgan 1962, p. 235), thus enabling the Western Sudanic peoples to expand southwards into the forest. Oliver has criticised this concept (1966Ò, p. 365), and the Asian food crops are generally believed to have reached Africa from Indonesia via Madagascar or the adjacent part of the east coast, probably in the first three centuries of the first millennium A.D. (Gray 1962, p. 182). Nowadays it seems a more acceptable idea that yam cultivation may be very ancient in West Africa (as it was in southeast Asia), and it has been suggested, indeed, that the Sangoan pick may have been a tool for grubbing up wild species, from which practice natural vegetative reproduction, localized around living places, might result (Davies 1968, pp. 479-481). In addition, it is a characteristic of the yam that it can regenerate after the removal of the tuber if too much damage is not done to the vine and roots; thus very early collectors and gatherers might have become accustomed to the idea of returning to the same spot for a source of food, and from this practice it is a small step to a kind of proto-agriculture in which care is taken not to damage the plant in remov- ing the tuber in order to safeguard a future food supply (Coursey, 1967, p. 12). Search- ing for wild yams has been recorded among hunter/gatherers in the Congo (Buikill, 1939) and their use in time of scarcity among farmers in Nigeria (Okiy, 1960, p. 118) and by the Chokwe in Angola (Clark 1963, p. 194). In support of an early domestica- tion of yams in Africa, Coursey points to the prohibition in certain areas on the use of iron tools for the digging of yams in New Yam festivals, which strongly suggests that yam cultivation antedates the commencement of the Iron Age (Coursey 1967, p. 10). The Idoma and some Ibo prefer to use a wooden spade for digging yams, but the reason given is that an iron hoe is likely to injure the yam (Armstrong, 1971). It has been suggested that yam-grcwing was a development resulting from the stimulus of cereal cultivators to the north rather than as having an independent origin (Alexander & Coursey 1969, p. 421). But the opposite possibility has to be borne in mind - that yam-growing along the northern forest margins was older than cereal cultivation and that it was yam-growers who developed the use of local grasses encountered as weeds in their yam patches ; in Asia it seems that root and tuber-growing in the southeast were older than cereal growing in north China, and that it was the root- and tuber- growers of south China who domesticated rice - although they may have had some stimulus from cereal growers from the north (Chang 1970, p. 183). Posnansky has suggested (1969, p. 106) that West African yam cultivation began between 2500 and 1500 B.C. One might perhaps apply the Binford-Flannery model here: the savanna was the 'optimal' habitat for the hunting and gathering population and the forest a 'less favourable' habitat, which received the groups which hived off from the savanna population when this became necessary if the savanna was to maintain its capacity to carry its population density. It is just in these marginal foiest areas that the African yams and the oil palm must have been developed, first as protected, and then as 158This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

planted sources of food supply (Binford 1968; Flannery 1969). Theoretical reasons have been given for supposing that once both seed-agriculture and tuber-agriculture have been established, the former will tend to displace the latter (Harris, 1971). The demonstration that after 3000 B.C. the West African forest was occupied by Late Stone Age people who had pottery, ground stone axes and stone tools that could well have been used as hoes or as blades for digging sticks, may be of the greatest impor- tance, in suggesting that they systematically exploited tuberous crops; they probably also made use of the fruit of the oil palm. This can be regarded as a specialized form of 'gathering' at first; the crossing of the line to 'vegeculture' and 'arboriculture', beginn- ing with the mere protection of wild trees, could have been hardly perceptible. An important step in the exploitation and domestication of the yams must have been the devising of the wooden pestle and mortar which so greatly facilitates the human con- sumption of fibrous starchy foods, particularly by infants, and which therefore may be an invention as important for tropical African agriculture as the plough was elsewhere. Is it a coincidence that the densest populations in sub-Saharan Africa are in southern Nigeria where the combination of yam cultivation and the exploitation of the oil palm, providing complementary food values, has been most highly developed. Within this area all these factors are seen to their greatest extent among the Ibo, who are 'the most enthusiastic yam cultivators in the world' (Coursey 1967, p. 198) and who have converted the greater part of their rain forest into oil palm bush. Thus there is probably a quite simple explanation for the density of population in lboland - the antiquity and effectiveness of yam cultivation and the exploitation of the oil palm. It is probably significant, too, that Ibo cultural origins, associated with yam cultivation and a yam festival, seem to have been in northwest lboland on the Anambra river, on the northern forest margins (Onwuejeogwu, 1971) and here also there is a grassland fringe indicative of early agricultural clearings (Gray 1962, p. 184). We need to learn more about the early protection and 'domestication' of the oil- palm, since it is not native to the primeval West African rain-forest, requiring more light to grow than this allows and the fruit needing the heat of full insolation to germi- nate, and yet needing plenty of groundwater (Hartley 1967, p. 4);wherever it is now found in the forest area it indicates human settlement at some time (Meer 1971). Presumably therefore the oil palm also must have been first developed along forest margins and in gallery forests in the savannah. We have already noted (p. 4) a pericarp of oil palm at Shaheinab, and Raymond (1961 p. 69) has suggested that there may have been a trade in palm oil in the Nile valley 'some 3000 years B.C.', but this was on the basis of an analysis of a large jar of fat from an Abydos tomb carried out nearly 80 years ago (Friedel 1897). Other tree-crops have been protected and developed such as Butyrospermum paradoxům , the shea-butter tree (Dalziel 1955, pp. 350-4; Davies, 1968, p. 480; Hugot, 1968, p. 486), Cola acuminata and C. nitida, (Dalziel 1955, pp. 100-4) and Adansonia digitata , the baobab. The latter occurs throughout the driest parts of tropical Africa, and has many uses (Dalziel 1955, p. 112-115; Owen, 1968, 1970). It is usually assumed to be indigenous to Africa (Baker 1965, p. 193), and its introduction to India in the fourteenth century A.D. has been associa- ted with the arrrval of African slaves (Burton-Page J. 1969). However, one botanist (Jackson, 1970), considers that it was spr ead over the African continent by the activities of Arab traders from the east African coast (Madagascar being its 'centre'), but a 159This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

radiocarbon dating of wood from a 45 ft-baobab near Lake Kariba of 1010 ± 100 BP. might be thought to militate against this idea (Swart 1963). If the baobab has to be thought of as being diffused by man from the Madagascar area, perhaps it was spread by the same agency as spread the bananas. Its presence in wooded savannah commonly indicates the site of a former village (Davies 1968, p. 480), and north of Nouakchott in Mauretania they are regarded as relicts, but whether associated with cultivation is unproven (Hugot 1968, p. 487). In Nigeria a number of indigenous fruits are still used as food: wild mango ( [Irvingia spp.), African breadfruit or African rice tiee (Treculia africana Decne), fan palm {Borassus aethiopicum), native pear ( Dacryodes edulis ), Allenblackia floribunda , Parkia clapperioniana (Redhead 1971). Vavilov did pioneer work in the inter-war period on the botanical evidence concern- ing centres of domestication for cultivated plants, but he did not investigate the African yams. He was convinced that he had 'established the independence of Ethiopia in its cultivated flora and had proved beyond doubt the existence here (including the hill country of Eritrea) of an independent centre of origin of the world's cultivated plants' (Vavilov 1951, p. 37). This view was based to a considerable extent, it seems, on the existence in this area of an amazing wealth cf forms of wheats and barleys - now thought to be the result not of primary domestication there, but of the introduction and hybridisation of a variety of already cultivated forms - and perhaps of a new environment exerting selective pressures. Thus Vavilov made Abyssinia one of his eight major centres in the world for the origin of domesticated plants, also attributing to this area the origin of Sorghum sp ., Eragrostis aby ssinica, Eleusine coracana , Pennisetum spicatum, Vigna unguiculata , Ricinus communis and Hibiscus esculentus (okra) - among others (Vavilov 1951, p. 38). It is now considered that wheat and barley cultivation and the use of the plough were introduced into Ethiopia by the ancient Cushitic inhabitants who were in a good position in northern Ethiopia to have contacts with countries at the northern end of the Red Sea (Simoons 1965). A research project designed to gain data from Ethiopia on its role in the development of African agriculture has so far not found sites with sufficient time depths to give us what we are seeking, as the caves at Lalibela and Natchabiet do not go back beyond 500 B.C. ; they show the presence at that date of barley and chickpea ( Cicer arietinum ), which need not surprise us (Dombrowski 1969). Murdock posited an ancient centre of plant domestication in his 'nuclear Mande area' around the headwaters of the Niger. He based his argument primarily on linguistic distributions, but claimed 'modest support from archaeologists, on the basis of admittedly fragmentary evidence from the very few relevant excavations reported to date' (but none of which are cited) 'and from botanists who have identified the ranges of wild species from which the domesticated forms have presumably been ennobled'. However, the latter type of evidence offered by Murdock did not stand up well to examination by a botanist. Baker (1962) examined Murdock's list and on botanical grounds considered that Hibiscus sabdariffa (roselle), Sesamum indicum (Sesame) and Gossvpium herbaceum (cotton) could equally well have been domestica- ted in Asia, and if they were domesticated in Africa it need not have been in the western Sudan. (A consideration of cotton is intentionally omitted from this paper since its domestication has not been for purposes of human consumption, the subject is complex, and expert opinion about it has changed frequently in recent years. It 160This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

is now believed that cotton was domesticated in Asia and introduced to Egypt in Meroitic times (Arkell 1971) although there were wild cottons in Africa of which the seeds were collected for stockfeed in Egyptian Nubia by the middle of the third millennium B.C. (Chowdhury & Buth, 1970)). Similarly Bakei considered there was insufficient evidence to regard the following members of Murdock's list as only domes- ticated in the western Sudan: Sorghum vulgare, Pennisetum typhoideum, Coleus dazo, and Coleus dysentericus, Vigna unguiculata, Sphenostylis (yam-bean), Citrullus vulgaris (water-melon), Cucumeropsis edulis and C. mannii, Lagenaria siceraria Tamarindus indica, Hibiscus esculentus, Hibiscus cannabinus. 'If the first domestication of these species in the Nuclear Mande area is, after all, unproven' says Baker, 'we are left with only a fraction of the original list'. (Evidence has since been adduced to suggest that the cowpea ( Vigna sinensis, Vigna unguiculata) was domesticated in West Africa, most likely in Nigeria (Faris 1963); it has been grown for centuries both in Africa and India (Ames 1939; Anderson 1952).) Baker divides the remaining species into two groups: the first group of species is 'Sudanic' but virtually restricted to West Africa, and it was the existence of such 'endemics' that Murdock considered provided the strongest botanical evidence for a domestication centre in the nuclear Mande area: Butyrospermum parkii ( paradoxům ) (shea-butter tree, really a wild tree given protection), Telfairia occidentalis (fluted pumpkin, wild and cultivated), Kerstingiella geocarpa (a ground-nut), and Digitaria exilis (fonio), very close to the wild D. longi- flora. With the exception of Kerstingiella geocarpa, these species are either undifferen- tiated from wild plants still occurring in the area or else show a close and obvious relationship, and could have been added from the local environment after agricultural ideas had been introduced from elsewhere. Kerstingiella geocarpa, however, is the only species in its genus and when Baker was writing was unknown wild (both it and Voandzeia subterranea have since been found wild neat the Nigeria/Cameroun border; Hepper 1963) and he regarded it as Murdock's best bit of evidence - together with Oryza glaberrima, although the centre of origin of this form of rice is believed to have been in the inland delta of the Niger, not its headwaters area (Portères 1962, p. 237). (Inferential evidence has been adduced for the growing of rice in the Casamance region of Senegal by the beginning of the first millennium A.D. ; Linares de Sapir, 1971, pp. 41, 43). The second group remaining are savannah and woodland species, 'Guinean' rather than 'Sudanic' in distribution: Blighia sapida (Akee apple), Cola acuminata and C. nitida, Elaeis guineensis and Dioscorea spp. The species occur wild as well as planted and one cannot be sure of any great antiquity for them as domesti- cates, and these Guinea species cannot be held to contribute any evidence in favour of the Nuclear Mande domestication centre. A great deal of botanical work on centres of plant domestication in sub-Saharan Africa has been done by Portères (1962). His interest has been predominantly with grain crops, which perhaps makes him underestimate the importance and potential of the forest products, and some of his inferences appear to be based more on doubtful anthropological and archaeological ideas than on botanical evidence - such as his assertion that food production in the forest is a recent innovation, the invention of former cei eal-growing savannah peoples forced into the forest by steppe nomads. His examination of the botanical evidence led Portères to posit some eight centres of crop domestication in Africa, four in West Africa and one each in Central Africa 161This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

(Cameroon/Central African Republic), the Upper Nile, Abyssinia and East Africa. In West Africa, Portères sees the Middle Niger area as originating Oryza glaberrima, Pennisetum cinereum, five species of sorghum, Kerstingiella geocarpa, and as a secon- dary centre for the development of Coleus dazo and Coleus dysentericus; the Senegam- bian area as a secondaiy centre for Oryza glaberrima and as producing Sorghun gam- bicum, Digitaria exilis and four species of Pennisetum; the Chad region as originating the cultivation of Dioscorea cayenensis, D. rotundata, D. dumetorum, D. bulbifera and four other species of yams, as well as the oil palm, Elaeis guineensis, and the fluted pumpkin, Telfairia occidentalis. In the Upper Nile area Portères places the origin of Sorghum durra and three other species, 4 species of Pennisetum and three species of Sesamum, while the debt to Abyssinia is especially for teff ( Eragrostis abyssinica), Eleusine coracana, and the African banana, Musa ensete {Ensete edulis), while it is also regarded as a secondaiy centre for wheats and barleys. The Central African area is regarded as the primary centre for Coleus dazo, C. dysentericus and Voandzeia sub- terranea and in part foi Elaeis guineensis, and the East African centre for five species of pennisetum and three of sorghum. It should be pointed out that to call Ensete 'the African banana' is somewhat mis- leading and in one way the name 'false banana' is more appropriate, as it is a banana- like plant but one which bears no edible fruit. Nevertheless the stem and especially the root are edible and provide the staple diet for the ensete cutivatois of southwestern Ethiopia (Shack 1966, p. 1). Ensete is distributed throughout West Africa, the Congo basin and East Africa (Simmonds, 1962, p. 25) but it is only in Ethiopia that it has been domesticated. When such ensete cultivation began seems to be quite unknown, but the Sidama peoples of Southern Ethiopia are thought to have been responsible, and the claimed evidence for ensete cultivation in ancient Egypt, although firmly believed in by some (e.g., Laurent-Tackholm, 1951; Arkell, 1971), is regarded as inconclusive (Keimer 1953 ; Simoons 1965). Unfortunately for archaeologists - or perhaps fortunately, if ultimately it brings us nearer the truth and stops us wandering down misleading by-ways - the whole principle that a proliferation of variant forms indicates a 'centre of origin' now seems called in question; as has recently been written, referring to Vavilov: 'Slowly over the years almost all the points made in his centres of o.igin theory have been refuted . . . This is not to say that Vavilovian theory was not an important contribution . . . Yet, when the evidence is properly qualified and corrected there is little left of the original theory except that ciops are more variable in some places than otheis' (Harlan & de Wet 1971 ; Baker 1971). This must cast some doubt on the basis for Portères's centres of plant domestication. On the other hand, perhaps we shall gain from the application to Africa of the newly-understood principle of 'disruptive selection'. For example, this may account for there being no known wild ancestor of 'fonio' ( Digitaria exilis ) to which the nearest is D. longiflora (Dalziel 1955, p. 526). Linguistics is obviously an important source of evidence upon the history of food pro- duction in Africa, although unfortunately the study c f comparative linguistics has not yet reached a state where there is general agreement on interpretations. For example, it is difficulttoknowhowmuch reliance to place upon the time depths proposedby theexpo- nents of glottochronology. Nevertheless it is interesting that one cf them sees people in southernNigeriacultivatingyams,cottonandbeans,keepinggoatsandmakingpalm-wine 6000yearsago(Armstrongl964,p. 136). 162This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

The introduction of the Asiatic food crops in the first three centuries A.D. into Madagascar and the adjacent parts of the mainland has already been mentioned above (p. 19). One of the most important groups of these consists of the bananas and plantains, although Dalziel (1955, p. 468) speculated that the plantain was introduced into Africa 'probably through Egypt in ancient times'. The banana and the plantain were formerly called respectively Musa sapientum and Musa paradisiaca (or M. sapientum var. paradisiaca) but the Latin nomenclature is in such confusion that it is regarded as better ťto disregard all Latin names' (Simmonds 1959, p. 54). Other important Asiatic crops are the Asiatic yam, Dioscorea alata , and probably the cocoyam, colocasia esculentum (although Dalziel says (1955, p. 481) it 'was cultivated in Egypt and India from remote antiquity'). Just how these Asiatic cultigens spread from the East African coast must be a very interesting story, but we know little about it at the moment. It would have been much easier for Musa spp., being vegetatively propagated and requiring considerable moisture, to have spread up the Zambezi valley and via the East African lakes to the other suitable areas of Africa, rather than by any other route. It would have been more difficult to get these species inland from the African coast around Mombasa, and even more difficult still to bring them along the arid Sabaean lane into the Nile Valley, although this was the route favoured by Sauer (1952, p. 36), stating that root stocks can be thoroughly dried out and left exposed for months before replanting. Simmonds also states (1962, p. 31) that the Eumusas are tolerant of considerable drought, ranging far out into the seasonal climates of the monsoon lands. Unfortunately scientific archaeology in Madagascar is in its infancy, and about all we know archaeologically of its early history is that the southern part of the island was inhabited by the end of the first millennium A.D. (Verin 1966, p. 124). Deschamps has supposed that the Malayo-Polynesians who colo- nised Madagascar only did so after first reaching the east African coast (Deschamps, 1960, p. 26). The spread of the Asiatic crops on the mainland could have been rapid because they quickly came into contact with African, probably Bantu-speaking, food-producers, even if these had taken to food production comparatively recently; or the introduction of these crops and their growers could itself have been an impor- tant stimulus to the adoption of food-producing practices. However this may be, when we come to southern Nigeria, to which the Asiatic crops at some point penet- rated, it is interesting that linguistic evidence independently splits the crops nowadays grown into three groups: the crops cultivated prior to the Asiatic introductions, the Asiatic introductions and he latest comers, such as maize and cassava, introduced after the European discovery of America (Williamson, 1970). Thus the food plant names which are sufficiently well correlated with the different genetic language groups to suggest an origin in Proto-Niger-Congo, i.e., at least 4000 years ago, are the following: raphia palm {Raphia hookeri), oil palm {Elaeis guineensis ), yam ( Diocorea cayenensis and D. rotondata), and kola nut (Cola acuminata and C. nitida). The second group, which it will be observed contain the Asiatic introductions, have names which are not so neatly correlated with genetic language groups but which do not cross language boundaries in such a way as to suggest recent borrowing: cocoyam (Colocasia esculentum ), plantain (Musa sapientum var paradisiaca ), banana (Musa sapientum) and water yam (Dioscorea alata). It should be noted, however, that Blakney (1963) thinks that the Ko root for plantain or banana is so widespread that it really 163This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

belongs to the earlier Proto-Niger-Congo group. The third group of names are those

with very little correlation with the language groupings: instead the names are spread

over wide areas which can sometimes be associated with trade routes and sometimes

give an idea of the direction from which the plant was introduced: bitterleaf ( Vernonia

amygdalina and V. colorata) is probably a recent introduction to the Niger delta

from Iboland; okro {Hibiscus esculentus) is indigenous to Africa but again probably

a recent introduction from Iboland; the different names in the delta for groundnut

{Arachis hypogaea) and maize ( Zea mays) correlate with the trading centres of port

Harcourt, Onitsha and Warri; the names for cassava {Manihot esculenta ), rice

{Oryza sativa) orange, {Citrus aurantium) lime {Citrus aurantifolia ), coconut {Cocos

nucífera) are derived from Portuguese; while those for onion {Allium cepa) are either

from Portuguese or from Arabic through Hausa.

Many local plants in different parts of Africa which are not staple food-providers

have been used for making sauces, soups or condiments. One of the most interesting

is Aframomum melegueta, named 'Melegueta pepper' by the Portuguese, the derivation

of 'Melegueta' being obscure; nowadays in English it is commonly conupted to

'alligator pepper'. 'Melegueta' is probably a derivative of the Tamil /w7tfga=pepper,

but the earliest references in Europe, in the thirteenth century, probably refer toa

plant from India different from Aframomum melegueta (Mauny 1953, p. 708).

Because Melegueta pepper was called 'Grains of Paradise' in the sixteenth century,

the coast of Liberia came to be called the 'Grain Coast' - a somewhat misleading

term to anyone thinking it referred to cereals (Dalziel, 1955, p. 471).

ANIMAL SOURCES OF FOOD

Although direct evidence of animal husbandry is more often preserved in archaeological record than that for crop agriculture, in the form of bones of domesticated

animals, there has been considerable difficulty, especially in the early stages domestication, in deciding whether a particular set of remains belongs to a wild,

domesticated or feral animal (Reed 1960, pp. 123-6; Monod 1964, p. 200; Bokonyi,

1969; Higgs & Jarman, 1969, pp. 33-6). Now, however, the application to bone some of the techniques of petiological identification appear to hold out a very good

hope of determining this with much greater ease and ceitainly than before (Drew

et al., 1971). Attention has also been drawn to the danger of accommodating everincreasing

data in this field to old piovisional hypotheses (Jarman 1969, p. 260). archaeologist may be permitted to wonder sometimes at the reliability of faunal

identifications made: in a recent experiment 38 identifiable bones were submitted to

eight specialists in turn; the eight agreed on the identification of one specimen only

(Van der Merwe 1971).

The problem of distinguishing wild and domesticated forms is encountered northern Africa, where there were wild cattle {Bos primigeni us, ? 6 Bos ibericus but into which it is commonly assumed domesticated breeds {Bos taurus) were introduced

from south-west Asia; whereas in sub-Saharan Africa there was not the wild

species to make identification of domesticates difficult. There were also wild pigs

{Sus scrofa) in northern Africa, and today also in east central Sudan (Dorst 164

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Dandelot 1970), but south of this only warthog (Phacochoerus aethiopicus ), bush-pig ( [Potamochoerus por cus) and giant forest hog (Hylochoerus meiner tzhageni); it might be difficult in some cases to distinguish the bones of these species from domestic pig, but the difficulty has not so far been reported. There was no wild sheep (Ovis) in Africa after the end of the Pleistocene, and probably no goat (Capra) but the pi esence of the Barbary 'sheep' (Ammotragus) confuses the position. Thus while there are serious problems of distinguishing wild and domesticated forms in the archaeological record of northern Africa, the problem in sub-Saharan Africa is simply that as yet we have not accumulated from a wide enough area sufficient well-analysed sets of animal bone material from well-stratified and firmly dated sites. As this accumulates, so the picture will become very much clearer, although another inherent difficulty is the poor preservation or total absence of bone remains in the acid soils of the humid regions, especially West Africa and the Congo basin. Whereas sub-Saharan Africa had to find its own local equivalents for wheat and barley, and in fact found other food crops for itself as well, the four animal domesti- cates introduced were able to become adapted to tropical conditions and were suf- ficiently successful in doing so for theie to be no point in looking for local substitutes 01 additions. Plenitude of game has also been adduced as a reason for there being little incentive towards domestication of indigenous animals (Clark 1967, p. 607), although this may have been more true of the savanna lands of Africa with their very heavy animal biomass than of areas with different types of vegetation (Butzer, 1971, Table 9, p. 150). The buffalo, the elephant, the zebra, the wait-hog, the bush-pig, the hartebeeste, the eland and certain other antelopes, the bush fowl, the cane rat (cutting- grass; Thryonomys smnderianus) and the giant African snail (Achatina achatina and Archachantia marginata) (Ajayi 1971) might have been domesticated but were not. Sauer (1952, p. 35) said that 'the natives of West Africa have a partly domesticated bush pig (Potamochoerus) from Guinea down to Angola', but gives no authority for the statement). The taming during Old Kingdom times in Egypt of gazelles, Bubalus , antelopes, moufflon sheep, ibex and even hyaenas, was considered to have been short-lived (Mauny 1967, pp. 587, 594, quoting Yoyotte in Posener et al 1959, p. 101 and Brentjes, 1965), but Clark has recently suggested (1971(0)) that this represents a practice going back to before Predynastic times; H. S. Smith (1969) has emphasized the cult element in the practice, and that this may be equally ancient. Similarly, nothing lasting seems to have come from early moves in the direction of domestication in the Jebel Uweinat area, where 'the oldest rock engravings (dating from before the introduction of cattle) represent scenes where giraffes and ostriches are kept in capti- vity' (Van Noten 1971). Therefore when we come to animal sources of food south of the Sahara we are not involved in a search for evidence and dating concerning African domesticates as we were in the case of sorghum, millets and yams; we are simply concerned with the rate and pattern of a process of diffusion. As far as we know at present, only three animals have been domesticated in Africa - the ass, the cat and the guinea fowl - and only the last of these as a food souice. In the case of animal sources of food it will be more convenient to consider the evidence for each animal one by one, again distinguishing between the diiect and indirect archaeological evidence and that which comes from zoology, stock breeding and linguistics. 165This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Cattle : northern area: direct evidence

For cattle, again our evidence falls into two groups: that from north of latitude

11° N, and that from south of it; it will be convenient to consider first the northein

area. (In this paper, the view of Reed (1971) is adopted that all wild varieties of cattle

(even including Bos namadicus) weie most probably capable of interbreeding and

should be assigned to the single species Bos primigenius , while all domestic breeds

should for the same reason be assigned to the single species Bos taurus. For clarity's

sake, however, older usages are in places given in brackets or inverted commas).

Unfortunately, as so often with the earliest osteological evidence, the oldest

archaeological occurrences of bones in Africa cannot be declared to be certainly

domesticated. In the Fayum A Neolithic of the early fourth millennium B.C. bones

were reported of cattle, sheep or goat, and pig but it was stated that none 'need

necessarily be domesticated' (Caton-Thomson & Gardner, 1934, p. 89). It has now

been definitely established that these cattle, sheep and goats were domesticated. At Uan

Muhuggiag in the Acacus, a level dated to the end of the fifth millennium contained

cattle bones, one (called Bos brachy ceros) probably domesticated, and it is possible

that cattle may also be present in an older level dated to the mid-sixth millennium

(Mori 1965; Camps 1969 p. 207; Clark 1970(è), p. 197). The indirect evidence of a

'neolithic' type of material culture in these south-western Libyan sites (an eastward

extension of Tassili N'Ajjer) may indicate that cattle-keeping in this area might even

extend back into the sixth millennium B.C. The Hoggar site of Meniet provided Bos

(called ibericus) and Capra sp. from the mid-fourth millennium, but again no certainly

domesticated species were identified (Hugot 1963, pp. 148-51). Similarly the Badarian

produced bones of cattle, sheep or goat, and dog which were 'probably domesticated

- though the remains were not carefully studied and have since disappeared' (Arkell

& Ucko 1965, p. 150). There were domesticated cattle from Merimde and Omari,

and specimens described as resembling B . longifrons types from Naqada (Zeuner

1963, p. 222). The first report of the 1970 expedition to the Tenerean site of Adrar

Bous listed Barbary sheep and pig, and an almost complete articulated skeleton of a

domestic Bos; no actually associated dates were available at the time of writing that

report but there was a previous radiocarbon date for the Tenerean of 3180 B.C.±

300 (Hugot 1962, pp. 71-2) and it may well extend down towards the end of the

second millennium B.C. (Clark 1970(a), pp. 19, 29). Now the Bos has been identified

as a small shorthorn comparable to that from Uan Muhuggiag and like the West

African shorthorns and Ndama cattle; a collagen radio-carbon determination gives

a date of 3800 B.C. ±500; the pig is now identified as warthog (Clark 1971(a)).

Phases 2-8 of the Tichitt sites in southern Mauretania, extending fr om the midsecond

to the mid-first century B.C., produced Bos taurus and Capra hircus throughout

(Munson 1970). Cattle have now been definitely identified in the Kintampo

culture of Ghana at the K6 site at Kintampo (second/first millennium B.C.) and

possibly at Ntereso (Flight 1971).

Cattle burials at Jebel Moya, dating from the Mercitic period (Addison, 1956)

show their importance in the culture represented, and this is confiimed by terracotta

figurines and designs on pottery (Addison 1949, pp. 56-59, 146-149, 221). An

important piece of negative evidence comes from the Neolithic levels of Haua Fteah

166

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

showing domsticated sheep or goat but no cattle at the beginning cf the fifth millennium

B.C. in Cyrenaica (Higgs 1967(6). Similarly the absence of cattle at Early

Khartoum and Shaheinab (with dwarf goat) may be significant.

Cattle: northern area: indirect evidence

Turning to the indirect evidence for cattle in this northern area, the rock art of the

Sahara depicting cattle is most important, placed as it is after that which shows only

the large wild animals of the Ethiopian fauna and before that depicting horses, and

considered to extend from the beginning of the fourth millennium B.C. to the end of

the second (Mauny 1967, p. 592) and perhaps even beginning earlier than the fourth

millennium in the Tadrat Acacus (Mori 1965; Clark 1967, p. 603). It has been

observed that in drawings of cattle in protodynastic and old Kingdom times in Egypt

it is chiefly the long-horn ( Bos africanus) which is portrayed, while the later drawings

show both this type and the short-horn ( B . brachyceros) which superseded the former

in the Nile Valley during the Middle Kingdom (Sandford & Arkell, 1933, p. 64).

The sculptures and hieroglyphs of Pharaonic Egypt indicate four different breeds of

domestic cattle, namely hornless cattle, and Bos primigenius breeds with small horns,

lyre horns and double-lyre horns respectively (Zeuner 1963, p. 223). The zebu type,

with cervico-thoracic hump and lateral horns, was popular only from the eighteenth

dynasty ( c . 1570 B.C.) onwards (Zeuner 1963, p. 226). This type of cattle ( Bos

indicus), perhaps descended from the wild type known as Bos namadicus, was domesticated

in India not later than the early fifth millennium B.C. (Zeuner 1963, p. 216);

the ability of zebu cattle to stand the conditions of tropical climates has made their

introduction into Africaof the greatest importancefor cattle-breeding in the continent.

The short-horned, thoracic-humped zebu was probably not introduced into Africa

until after the Arab invasion of the seventh century A.D. Cattle bones occur throughout

the sequence at Daima in north-eastern Nigeria, stretching from the sixth century

B.C. to about the eleventh century A.D., but their types have not yet been identified

(Connah 1968, 1970), nor are they likely to be, since the bones are too fragmentary

(Fagan 1971). Clay figurines from this site show humpless cattle throughout, with a

possible humped specimen at the end of the sequence. Rock paintings at Birnin

Kudu in northern Nigeria show long and short-horned humpless cattle, but the

paintings are not dated (Fagg 1957).

Similarities have been pointed out between the style of the portrayal of cattle in

the rock art of the Senafe area of Ethiopia (south-east of Asmara) and of Jebel

Uweinat, on the borders of Egypt, Sudan and Libya; the Senafe cattle paintings are

not dated but are presumed to antedate the introduction of the zebu into the area

(first millennium B.C.) (Graziosi 1964). It has been suggested (Cole 1964, p. 234)

that the Harar (Dire Dawa) group of rock art is later s'nce it shows zebu cattle and

camels, but the earliest paintings in this area are all of humpless longhorns (Clarke,

1970(6), p. 207).

Cattle: stock-breeding evidence

On the basis of a consideration of the cattle found in Africa today, both Epstein

(1957) and Payne (1964) derive them all either from the Hamitic longhorn (presumably

derived from Bos primigenius or B. africanus) known in the north-eastern part of the

167

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

continent from 5000 B.C. or earlier; or from the Hamitic Shorthorn {Bos brachy ceros)

known in Egypt from Middle Kingdom times, and which, it is suggested, pushed the

Longhorns up the Nile and into the Sahara and the Maghreb; or from the zebu,

having come at different times from different countries of Asia. All sanga cattle are

mixtures of different proportions of zebu with the others. The nineteenth century

rinderpest pandemic largely wiped out the other breeds and left Africa predominantly

cattled by zebu (Payne 1 964).

Cattle: northern area: inferences

What does this evidence about cattle in Africa north of 1 1°N latitude add up to,

and how has it been interpreted? Because in Egypt we at present lack sites belonging

to the sixth millennium B.C. and earlier, since they are likely to be buried under a

considerable depth of silt (see above p. 5), we aie in much the same position as we

found ourselves in concerning cereals; that is, we do not have the evidence to say

whether Egypt merely received domesticated animals from south-west Asia or whether

north-eastern Afiica took part in a widespiead and prolonged experimentation which

led to domestication. So for the time being this must be left as an open question. It

has been argued that since longhorn cattle occur earlier in south-west Asia and those

in the early Egyptian records are always associated with goats, it is reasonable to

assume that longhorn cattle originated in Asia (Epstein 1957, p. 142). That wild

cattle were indigenous in Egypt explains the appearance of a pure and primitive

primigenius breed there, from which the later long-and short-horned breeds are both

descended' (Zeuner 1963, p. 223; Higgs, 1967(6) p. 316) but we do not know whether

this Egyptian domestication was independent, or achieved under the stimulus of ideas

from south-west Asia, or only after domesticated cattle had actually reached Egypt

from Asia. Arkell (1971) and Butzer (1971, p. 592) speak confidently of an African

domestication of the local Bos primigenius , probably in Egypt. The wild bulls of the

Delta marshes were still captured and tamed in the third and second millennia B.C.

and were distinguished from those which texts imply were usually imported from the

south (H. S. Smith 1969, p. 308). If there was an early independent domestication

of cattle in Egypt, we must keep our minds open to the possibility of such domesticated

cattle spreading into south-west Asia. There is one tantalisting piece of indirect

evidence, from the cliffs of Kom Ombo, consisting of a number of friezes and indidual

engravings of cattle. None of the animals seem to be domesticated, yet the

composition is very reminiscent of many of the Saharan pastoral friezes; they are

undated and could fit anywhere in the period 9000 to 4000 B.C., but they might

represent an early stage of domestication (Smith 1971).

From Egypt, it has been supposed that cattle spread westwards by two routes -

along the Mediterranean coast and penetrating into the Sahara over the higher,

better-watered areas like Tassili and Hoggar; the other from Upper Egypt along the

southern edge of the Sahara about latitude 20°N. (Mauny 1967, p. 583). Routes

from Egypt via Kharga oasis to Tibesti, Wanyanga and Ennedi have been suggested

(Clark 1967, p. 603). The absence of domesticated cattle at Early Khartoum and

Shaheinab suggests that the take-off points for the westward movement from the

Nile valley were in Egypt rather than Sudan. Of course this was all at a time (the fifth

to the first half of the third millennium B.C.) when there was certainly more

168

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

groundwater, and probably more precipitation as well, in the Sahara desert, with a

Mediterranean flora in the highlands and steppe-like conditions elsewhere (Hugot 1962;

Monod 1964; Butzer, 1971, pp. 332-3, 350). The suggestion has been made that the

Capsians of Tunisia and the Qadan people of Nubia were such successful hunters/

fishers/fowlers/gatherers that they were resistant to incoming 'neolithic' culture

traits and only adopted the piactice of food-production late: by contrast, the people

forced to find an existence away from these 'optimal' areas, in the highlands of the

Sahara, found that they could make a better living out of this less favourable region

by developing cattle breeding, and thus they did this early (Butzer, 1971, pp. 585-94).

On the other hand caution has been expressed concerning the earliness of the Sahara

Neolithic dates on the grounds that they may be derived from long-lived species of

wood (Clark 1971(6), p. 53). It has also been suggested that the same set of climatic

conditions allowed tsetse flies to extend further north than they do today and that

in this way a southern frontier for cattle was established around about latitude 18°N.

When the final desiccation of the Sahara set in around the mid-third millennium

B.C. this frontier shifted down to about 14°N; these two changes together exerted a

strong stimulus to the cattle-keepers to move south and probably resulted in their

being able to penetrate the tsetse-free areas of East Africa (Mauny 1967, p. 583-5,

Fig. 1). Others moved with their longhorns into tsetse infested areas and the survivors

developed into the trypanosomiasis-resistant Ndama cattle, while the West African

Dwarf Shorthorn, derived from the Hamitic Shorthorn and perhaps moving coastwise

southwards frcm the Magrheb into West Africa, also developed a resistance but

to a lesser extent than the Ndama (Payne 1 964).

The absence of cattle in the Neolithic of Haua Fteah and the earliness of the dates

for cattle in the Acacus, mentioned above, taken together with the evidence that

domestic cattle may be earlier in south-eastern Europe than in southwestern Asia

(Reed 1969, pp. 372-3) has led Higgs to consider the possibility 'of cattle herdsmen

emigrating along the northern shores and archipelagos of the Mediterranean, making

landfall in northeast Tunisia and spreading thence south and into the interior'

(Higgs 1967(a), (è); Higgs & Jarman, 1969). Thus, just as there may be a 'duality of

origin' for the Neolithic of the north African coast, so there may be more than one

source of origin for domesticated cattle, although derivatives from these may subsequently

have mingled and interacted with each other.

Cattle : southern area : direct evidence

Turning now to Africa south of 1 1°N latitude, it has been confidently stated that

'there is solid evidence' for a pastoral economy among the Stone Bowl folk of East

Africa by the end of the second millennium B.C. (Clark 1969, p. 201). It is indeed

probable that the shift in tsetse distribution made a spread of cattle into this area

from further north possible, but the direct archaeological evidence is in fact only just

beginning to appear. There was none at the Njoro River Cave (Leakey & Leakey,

1950), the Hyrax Hill Neolithic Cemetery (Leakey 1945) or the Ngorongoro Crater

sites (Leakey 1966; Sassoon 1968(a) and (b); none is reported from the Nakuru

burial site (Leakey 1931, pp. 198-201), or Ilkek (Brown 1966), but cattle bones are

said to have been associated with stone bowls and charcoal which gave a radiocarbon

date of 740 B.C. ± 80 from Prospect Farm, Elementeita (Sassoon 1968 (6), p. 22) ; another

169

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

date (N-651) new available for this site is B.C. 960+110 (Cohen 1970). A settlement

site at Narosura in the Narok district of Kenya has yielded cattle bones associated

with stone bowls, an obsidian industry, black-burnished pottery similar to some from

Njoro River Cave and a bone collagen date of 310 B.C. ±110 (Phillipson, 1970(a), p.4) ;

further dates, on charcoal (N-700-703), now available span from the fifth to the

ninth centuries B.C. (Soper 1971). Tunnel Rocksheiter (radiocarbon dated to the

first millennium B.C., with a Wilton-like industry but no stone bowls; Sutton, 1966

p. 41), the uppei levels of the Mur inga Rocksheiter and some of the large cairns in

the western highlands of Kenya (not precisely dated but unlikely to be less than

2000 years old) have also produced cattle bones (Posnansky 1967, pp. 635-6; Sutton,

forthcoming). Thus it looks as if we can be reasonably sure of pastoralism among

the Stone Bowl folk by the beginning of the first millennium B.C. There was no

evidence of food production at the Late Stone Age rock shelter of Nyangoma in the

Mwanza region of Tanganyika, with a radiocarbon date of 690 B.C.±120, but it

has been suggested that there was a frontier round about 35° E latitude separating

the Late Stone Age hunter-gatherers to the west from the Stone Bowl pastoralists

to the east (Soper & Golden, 1969, p. 41). On the other hand, cattle, sheep and goat

bones were present in the 'Kenya Wilton' site of Rangong, on the northern side of the

Kavirondo Gulf of Lake Victoria, and cattle also at two other comparable rock

shelters in the area, all dated to the second half of the first millennium B.C. However,

one must respect the caution of the excavator: 'any or all may have been intrusive

due to stratigraphie disturbances' (Gabel 1969, p. 246).

In Zambia, domestic cattle appear in the faunal sequence at Markwe at the same

level as the first Early Iron Age pottery (Phillipson 1971); the Early Iron Age site

at Kalambo Falls, lasting from A.D. 500 until the sixteenth century, produced bovid

teeth (Seddon 1968, p. 492, quoting Clark 1962, 1964); the Early Iron Age site of

Kapwirimbwe, south of Lusaka in Zambia, dated to the fifth century A.D., produced

cattle bones (Phillipson 1968); cattle bones occurred at Kumadzulo (sixth/seventh

centuries), Chundu (eigth century) and the Zambesi Farm midden (Vogel 1971). The

Kalomo culture of Zambia had cattle, and sheep/goats, dog and domestic fowl as

well, in the eight, to eleventh centuries A.D. In south-western Rhodesia, buried

horn-cores appear in Phase III of the Leopard's Kopje Culture, dated to the early

part of the second millennium A.D., and may be associated with a cattle cult (Robinson

1966, p. 25). Bone evidence shows that cattle, and sheep/goats, had reached Bambandyanalo

on the Limpopo by the late eleventh century (Gardner 1963, p. 29). A study

of the food waste from a number of Transvaal Iron Age sites showed that, in most,

50-75 per cent of the animal food came from domestic animals (cattle, sheep and

goats); only in one was the figure as high as 90 per cent (Welbourne 1971, p. 81).

Cattle: southern area: indirect evidence

The Zambian Early Iron Age sites of Kumadzulo and Chundu have produced

cattle figurines of 'Sanga' type (Vogel 1969, 1971); figurines from Isamu Pati and

Kalundu are believed to represent 'Sanga Shorthorns' (Fagan 1967, p. 66), but these,

like others from Simbusanga, Simde, Silmonga and Mulima, are all post-twelfthcentury

occurrences and related to post-Kalomo-tradition phenomena (Vogel 1971).

Figurines of cattle, sheep and goats in Phase II of the Leopard's Kopje Culture dated

170

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

to the later part of the first millennium A.D. are taken as confirming the suggestion

that the distribution of this phase on light, sandy soils is indicative of pastoralism.

Although there are human figurines on Phase I sites, there are no cattle figurines

(Robinson, 1966(6), p. 14). A 'clay ox' from Bambandyanalo confirms the bone evidence

(Fouché 1937, p. 20) and a figurine in Period II at Zimbabwe (probably tenth century)

shows a wide-horned zebu (Robinson et al. , 1961). Many dozens of well-defined

terracotta cattle figurines were recovered in the 1971 excavations at Olifantspoort

in the Western Transvaal (sixteenth to nineteenth centuries A.D.) which should enable

identification of racial variety (Mason 1971).

One might have hoped that indirect evidence from sub-Saharan Africa's wealth

in rock paintings and engravings would have helped to trace the spread of cattle,

but in this respect it is disappointing. There are inherent difficulties in trying to use

rock art in this way, since it is rare to be able to find an independent means of dating

it; more often any dating is by reference from the internal evidence of the subject

matter. In addition rock art is confined to areas with suitable rock surfaces, paintings

in the open are not likely to have survived, and rock surfaces exfoliate. The Abbé

Breuil saw a connection between the rock art of Eastern Spain, North Africa, the

Sahara and sub-Saharan Africa, but modern archaeological opinion is against this

idea of an artistic tradition carried by a continent-long ethnic movement, since there

is no direct evidence for any approximate contemporaneity in time nor of cultural

identity; rather are the different art manifestations in Africa regarded as spontaneous

local responses, even if influencing their neighbours and sharing some traditions

(Clark 1959(a), p. 260).

In Uganda and Kenya there is little rock art - there is a frieze of crudely painted

cattle on the eastern side of Mount Elgon, of apparently humpless longhorns, in

which case they are not likely to be very recent. In spite of the wealth of rock art

in parts of Tanzania no domestic animals are portrayed except in the very late style

(Fosbrooke et al. , 1950; Cole 1964, p. 235), even though Inskeep concluded, as a

result of the level in the occupation deposit of the Kisese Shelter at which he found

a painted exfoliated slab, that the age of some at least of the Kondoa paintings should

be measured in centuries rather than millennia (Inskeep 1962). It seems that the

stock-keepers were either not responsible for the art or did not commonly portray

their domestic animals. The rock art of Angola, Katanga, Zambia and Malawi is

also probably not more than a few centuries old, and is likely to have been connected

with magico-religious practices and initiation ceremonies; since most of it consists

neither of purely decorative patterns nor of representations of animals, scenes or

objects, this seems to be the best interpretation. Bantu-speaking people may have

been responsible for much of it (Clark 1959(a), 1959(ò), pp. 269-70). In spite of the

wealth of rock-paintings in Rhodesia, they do not help us very much in tracing cattle

movements (Goodall and Cooke in Summers, 1959). In South Africa the earliest Iron

Age immigrants appear at first from the rock-art to have lived peacably with the Bush/

Hottentot indigenes, bringing their herds of long-horned cattle and fat-tailed sheep;

it is only in the later Bantu period that there seems to have been conflict, with cattle

shown being protected by Basuto or Zulu warriors (Clark 1959(a), pp. 279, 283).

171

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Linguistic Evidence

Evidence has been adduced to show that cattle have been known in the more

northerly parts of East Africa for as lcng as there have been speakers of Eastern and

Central Sudanic languages in the area and probably also as long as speakers of

Southern Cushitic as well (Ehret 1967). A single people initiated the spread of cattle

further south through southern East Africa and partly into southern Africa, at a time

prior to the expansion of Bantu speakei s into those regions. (This spread was not accompanied

by a knowledge of milking, which diffused to Bantu-speaking peopleslater).

In South-east Bantu the word for cow does not have the normal Bantu root but

comes directly fiom Khoisan (Ehret, quoted in Huffman, 1970, p. 18). It has been

suggested that this group of Bantu languages may be associated with the earliest Iron

Age cultures of Rhodesia; and further, that they did not emphasize cattle in their

economy, since no conclusive evidence, either of actual remains or of figurines, has

come from any Gokomere or Ziwa phase site; evidence for cattle herding in southwest

Rhodesia only comes with the later Leopard's Kopje II culture (Huffman 1970).

Cattle: southern area: inferences

The pattern which emerges from this evidence would seem to indicate a spread of

cattle-keeping from the Sudan zone of Africa during the second millennium B.C. into

East Africa, where the stone bowls of the first millennium B.C. may or may not

reflect a remote connection to the stone craft of the Nile valley. The southward spread

may have continued, probably slowly, perhaps into north-eastern Zambia but in fact

there is little if any evidence for it (Fagan 1965, p. 46; 1971); in any case by the end

of the first millennium B.C. cattle-keeping was probably established in the Zambia/

Rhodesia area. An interesting speculation concerns the (supposedly Cushitic-speaking)

cattle-keepers who moved down East Africa during the first millennium B.C. and

their meeting at the end of the millennium with Bantu-speakers, supposedly cropgrowing

and iron-using; this would have taken place somewhere in south central

Africa (Katanga) if the latter had pierced the Congo forest from the northeast, elsewhere

if they 'overtook' the cattle-folk from the direction of the Congo/Nile watershed.

It has been suggested that this meeting should be thought of in terms of small numbers

of technologically superior immigrants being welcomed for their new skills by the

occupying hunters or pastoralists, rather than in terms of large-scale invasion (Clark,

1970(0), p. 218). By the time Europeans arrived in South Africa in the late fifteenth

century, the pastoral Khoi-Khoi (Hottentots) had large herds of long-horned cattle

and fat-tailed sheep (Fagan 1965, p. 31), yet a Late Stone Age site with their bones

is yet to be found. Where cattle are nowadays more important as wealth than as a

source of food, it is interesting to speculate as to when and how this came about.

Sheep and Goats

As is well known, although sheep and goats may be the oldest of all animal domesticates,

with sheep going back to the ninth millennium B.C. and goats to the eighth/

seventh in south-western Asia, there are great difficulties both in disentangling one

from the other and in distinguishing the bones of wild and early domesticated specimens

(Reed 1960. pp, 130-137); even the existence of a wild ancestor for the sheep

has been questioned (Higgs & Jarman, 1969, p. 37).

172

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Passing reference has already been made to some of the evidence concerning sheep

and goats in Africa when dealing with cattle. The sheep/goats originally reported

from Fayum A at the end of the fifth millennium B.C. were said to be 'not necessarily

domesticated' (Caton-Thompson & Gardner 1934, p. 89), but Gautier regards the

Capra associated with Fayum A as clearly domesticated (Wendorf, 1971). The domesticated

status of sheep/goat remains from Badarian sites was not fully established,

and there was similar uncertainty about the sheep/goats from Merimda (Reed, 1960,

p. 133); but until there is better evidence for the presence of wild goats in Africa after

the end of the Pleistocene (Reed 1960, p. 133-4), it is fair to assume that these remains

belonged to domesticated species - provided always they were determined to have

been true sheep and true goats (and not Barbary sheep or ibex).

Capra sp. was reported (together with possible Bos ibericus) frcm Meniet in the

Hoggar in the mid-fourth millennium B.C., and Capra hircus was identified in all but

the earliest stage in the Tichitt sites (associated with cattle) from the mid-second

millennium B.C. to the mid-first (Munson 1970). Domesticated sheep or goat was

present early in the fifth millennium B.C. in the Cyrenaican Neolithic of Haua Fteah

(Higgs 1967(a), 1967(¿>)); at Early Khartoum there were no remains which suggested

domesticated animals (Bate in Arkell, 1949, p. 27) but at the 'neolithic' site of Shaheinab

(with radiocarbon dates averaging to 3300 B.C.) although 98 per cent of the bones

recovered were of wild animals, there were domesticated dwarf goats and possibly

sheep (Bate in Arkell, 1953, pp. 15-18); teeth and traces of jaw-bone of 'probably a

kid' were found in one of the Protodynastic graves at Omdurman Bridge (Arkell

1 949, p. 99). At Haua Fteah and Shaheinab there were no other domesticated animals,

and it is interesting that it thus appears that the goat was able to spread in Africa ahead

of any other animal domesticate; for although it is possible that wild goats were

present in Africa at the end of the Pleistocene, from which a local domestication might

have taken place, there is no unequivocal evidence that they were (Reed 1960, p.

133-4). Wild Ovis is not known to have lived in Africa (Reed 1960, p. 135; 1971);

and there is no evidence for any domestication of the Barbary 'sheep' ( Ammotragus )

indigenous to northern Africa, although its bones are found in north Africa and Sahara

'neolithic' contexts, including the Tenerean of Adra Bous (Clark 1970(a) p. 29).

If goats at Shaheinab were spreading ahead of other domesticated animals and

continued to do so southwards into East Africa we had until recently no direct evidence

of this. The Hyrax Hill report (Leakey 1945, p. 365) refers to 'bones of sheep and

oxen' being found at the Nakuru Burial site but no identifications of the 'animal

bones' were given in the original report on the site (Leakey, 1931), and it now seems

possible that the Nakuru site was contaminated by later Iron Age material (Sutton,

1971). Until recently there was a complete gap, geographically and chronologically,

in direct evidence for the spread of sheep and goats, between Shaheinab on the one

hand and, on the other, Ingombe Ilede and the Kalomo culture of Zambia from the

ninth century A.D. onwards (Fagan 1967), the site of Twickenham Road, Lusaka,

which produced goat probably to be dated to the tenth century A.D. (Phillipson, 1 970 (¿>)),

pp. 97, 113), and Bambandyanalo in the eleventh (Gardner 1963), unless the two

immature 'sheep' mandibles from Mabveni in Rhodesia (Robinson, 1961) belong to

the horizon which has produced a radiocarbon date of A.D. 180 ± 120 (Fagan 1966,

p. 503). Now, however, the gap has been filled, since the teeth of 75 goat/sheep (as

173

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

against 52 for cattle) have been identified from the first millennium B.C. Stone Bowl

site of Narosura in Kenya (Soper 1971). Goat is reported from second/first millennium

B.C. contexts at Kintampo and Ntereso in Ghana (Flight 1971). In the linguistic

evidence for West Africa referred to above (p. 13), goats are mentioned as known six

thousand years ago, but not cattle. It is doubtful whether this type of evidence is

sufficiently reliable to say that this is another indication of the goat outstripping

cattle in their spread in Africa, but it is a possibility. An examination of the Bantu

words for sheep suggests that livestock keeping was introduced to southern Africa

by people speaking Central Sudanic languages (Ehret 1967).

The indirect evidence of the rock art provides paintings of fat-tailed sheep in

Rhodesia and in the Brandberg of Southwest Africa, and there were clay figurines of

fat-tailed sheep in the Khami ruins and at Zimbabwe. It has been posited that it was

in the seventeenth century, following the Rozwi invasions and during a slightly moister

period in the northern Kalahari desert, that the artistic shepherds of Rhodesia escaped

westwards to the Brandberg (Summers 1959) and introduced the sheep and the

polychrome paintings to that area (Willcox 1963), but Cooke does not think the

shepherds themselves were the artists: from his detailed study of the evidence he

suggests that fat-tailed sheep arrived in Rhodesia roughly fifteen hundred years ago,

and reached the Cape of Good Hope by a slow migration rather than by diffusion ;

the actual people who started the process were probably unlike the Hottentots they

became, by a process of hybridization on the way, but the sheep had become a stabilized

breed by the time they left Rhodesia (Cooke 1965).

Pigs

In common with other parts of the Old World, the archaeological record in Africa

for the domesticated pig is much more scanty than for the other three domesticates.

It is reported among the fifth millennium B.C. bones from the Fayum (Caton-Thompson

& Gardner 1934, p. 89) but, as with the sheep and cattle, 'not necessarily domesticated'.

Reed has usefully reviewed the evidence for Egypt and came to the conclusion

that domesticated pigs may well have been present before the Third Dynasty (Reed,

1960, p. 141). The pig reported for the Tenerean of Adrar Bous (Clark 1970(a)) has

turned out to be warthog (Clark 1971(a)), and for the rest of Africa there seems to be

virtually no archaeological evidence of any kind; even in areas not subject to Moslem

influence, pig keeping in Africa today is much less widespread than that of cattle,

sheep and goats.

Reed has given reasons for preferring to believe that the wild pigs of Egypt were

locally domesticated, rather than that domesticates were introduced from south-east

Asia (1960, p. 139), and attention has recently been drawn to an apparently independent

domestication of pigs in the Crimea in the eighth millennium B.C. (Higgs & Jarman

1969, p. 38). However that may be, the pig does not seem to have been an important

food source in Africa. Perhaps this is not so surprising, since the pig is a much less

efficient converter of cellulose into a form of food digestible by man than the three

ruminants; the pig, like the dog, competes directly with man for food, and thus 'only

in a community where there is an excess of human food can pigs and dogs be tolerated

as a part of the biosocial community' (Reed 1969, p. 366). Oral tradition in eastern

Zambia is adamant that pigs were introduced there by the Portuguese (Phillipson,

1971).

174

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Dogs

It has been stated that dogs were first domesticated for food, not as hunting animals

(Herre 1963, quoting Degerbol, 1962); but if this is so, it is odd that dog bones are

rare in midden material compared with those of the hoofed food animals (Reed 1971).

Cynophagy, although firmly rejected by some African peoples, has been recorded

among others; the skeletons of the Isamu Pati dogs of about 1000 A.D. in Zambia

showed no signs of their flesh having been eaten and one skeleton was complete

(Degerbol 1967, pp. 201-2). There is a reasonably definite drawing of a leashed dog

in Diana's Vow Paintings, Rusape, Rhodesia; the dog appears to be of the whippet/

?Saluki type(Cooke, 1971). Many tens of thousands of animal bones have been studied

from the Transvaal Iron Age sites by R. G. Welbourne but no clear evidence of domestic

dogs has yet appeared (Mason 1971).

Domestic Fowls

The domestic fowl, originally domesticated in India or further east, was known

in the fourteenth century B.C. in Egypt but then disappears from the record, perhaps

for religious reasons, until Ptolemaic times (Zeuner 1963, pp. 443-5). It is widespread

in Africa, but the earliest archaeological evidence for it comes from the Kalomo

culture in Zambia around A.D. 1000 (Fagan 1967, p. 70). It was also found at Bigo

in Uganda in a context thought to date from about A.D. 1200 (Shinnie 1960, p. 26).

Present evidence is inadequate to indicate whether the source of dispersal for fowls

was Ptolemaic Egypt, or the Malayo/Indonesian immigrants on the east African

coast in the first millennium A.D., or both.

The African Domesticates

It was mentioned above (p. 14) that only the ass, the cat and the guinea fowl were

domesticated in Africa, and as we are only dealing with food animals, only the latter

need concern us here. (Although, by analogy with modern Masai practice of drinking

the blood, the ass might also have been a food source at one time.)

Guinea fowl appears to have been common in north Africa formerly, but not to

have been known in ancient Egypt ; however, they were taken across the Mediterranean

and there is literary reference to them in Greece by the fifth century B.C. The Romans

knew them as a poultry bird but they seem to have become forgotten and were unknown

in medieval Europe. They were rediscovered by Portuguese explorers of the

African coast at the end of the fifteenth century, hence their name (Zeuner, 1963,

p. 457). There is as yet no archaeological or other evidence concerning the antiquity

of the domestication of the guinea fowl in sub-Saharan Africa.

CONCLUSIONS AND QUESTIONS

What conclusions can be drawn from the foregoing review of the evidence for the

intentional production of food from botanical and animal sources in Africa? It would

be possible from it to attempt yet another synthesis concerning 'the spread of agriculture

in Africa', 'centres of origin', 'nuclear areas for diffusion' and so on, even if

175

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

such a synthesis would be somewhat cautious. However, in accord with the attitude

outlined at the beginning of this paper, it may be better to realize that we are left with

more questions than answers, and to list these questions in order that research workers

may pay attention to them and look for the evidence that will throw light upon them.

There are two theoretical questions that require further attention :

1 . What can justifiably be inferred concerning historical processes of domestication

from a study of present distributions of wild and domesticated crop varieties?

2. Since it has become evident from this paper that direct archaeological evidence

is scarce, and likely to remain relatively so, are we to confine ourselves to the

minimum firm ground provided by this evidence and say that we cannot venture

further? Surely this would be a counsel of despair; rather, we have to consider

in more detail what is justified in using all the indirect evidence, both archaeological

and other, and what is the methodology to be employed.

Apart from these two theoretical questions, in the present state of our knowledge it

is better to look for the evidence to answer small, specific questions than large generalized

ones. If we got the answers to all the small questions, we should be in a fair way

to answer the big ones. In this way we are more likely to get the information which

will enable us to make progress in understanding the development of food-production

in Africa. Some of these questions are listed below; in considering them we need all

the time to be asking what ecological factors have been involved in the emergence of

food-production in different micro-environments in Africa, and what lessons can be

learnt from surviving agricultural practices.

1 . What was the relationship between early cereal growers in Egypt and the Saharan

highlands ?

2. How well is it established that cereal growing and stock-raising was earlier in the

Saharan highlands than in the Nile valley?

3. Was experimentation with tropical wild grasses anjndependent process or was

there a stimulus from wheat and barley growers further north ?

4. If it is established that Sorghum , Eleusine and Pennisetum are of African origin

but that they had been transmitted to India by the middle of the second millennium

B.C., what evidence can we find, and going back how far, for the domestication

process at the African end (i.e. in Sudan, Ethiopia and Somalia)?

5. What other evidence of cultural contacts between the areas referred to in Question

3 above are there? Did some cultural traits move in the opposite direction?

6. Was the cultivation of yams in the West African yam area independent of infliience

from cereal growers in the savannah, or did it only develop under the influence

of agricultural practices from the north?

7. Could independently-attained yam cultivation along the forest margins have led

to the cultivation of local wild grasses?

8. Can we locate more precisely the area of initial exploitation and protection of

the oil-palm?

9. What is the precise relationship in sub-equatorial Africa between the spread of

agricultural practices, the spread of iron technology and the spread of Bantu

speech?

176

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

10. Did the Asian food crops introduced into East Africa in the first three centuries

A.D. merely become added to the repertoire of existing agriculturalists, or did

they serve to initiate the practice of agriculture itself in the area of introduction?

1 1 . Did Egypt or the Sahara domesticate its own wild cattle or did it receive domesticated

cattle as an introduced innovation?

12. Was there a separate early introduction of domesticated cattle into Tunisia from

southern Europe?

13. Was the mid-second millennium B.C. introduction of the zebu into Egypt from

Mesopotamia via Sinai, from Arabia via the Red Sea, or from Arabia via the

Horn area of Africa?

14. What traces can be found of the probable spread of cattle from Ethiopia and/or

the Sudan zone southwards into East Africa during the course of the second

millennium B.C.?

15. What was the area, date and character of the meeting of the (Cushitic-speaking)

cattle-keepers of East Africa and immigrant Bantu peoples ?

16. How and when did sheep/goats spread into West Africa?

17. What were the dates and routes for the diffusion of domesticated pigs in sub-

Saharan Africa?

18. Did the spread of the domestic fowl in Africa stem from Egypt in Ptolemaic

times; or was it only introduced by the Malayo/Indonesian immigrants on the

east African coast in the first few centuries A.D. ; or were there introductions

from both these sources?

Tf in the course of the next ten years we can find the answers to some of these questions,

and give more precise chronology to what we already know, we shall certainly have

advanced our knowledge concerning the history of African food production. There

is not space here to go into the best strategy of framing research so as to obtain these

results, but it is to be hoped that there will not only be research programmes designed

to tackle some of these questions, but also that no opportunity will be lost in other

programmes to watch out for data which might contribute to the solution of these

problems.

A cknowledgemen ts

Earlier drafts of this paper were shown to the following colleagues who were kind

enough to make comments and suggestions, many of which have now been incorporated;

I wish therefore to acknowledge my indebtedness and express my gratitude

to them for this help: A. J. Arkell, R. G. Armstrong, H. G. Baker, L. Balout, R.

Braidwood, K. W. Butzer, Miss G. Caton Thompson, G. Camps, J. D. Clark, G. E.

Connah, C. K. Cooke, S. G. H. Daniels, N. David, R. Derricourt, G. W. Dimbleby,

M. Gill, J. Harlan, J. Hester, H. Hugot, Sir Joseph Hutchinson, R. Inskeep, R.

Kennedy, J. P. Lebeuf, J. Leclant, R. Mason, R. Mauny, P. J. Munson, P. Ozanne,

D. Phillipson, W. G. L. Randies, J. F. Redhead, C. A. Reed, H. Sassoon, D. Seddon,

P. Smith, R. Soper, G. Souville, R. Summers, J. Sutton, P. Ucko, N. J. van der Merwe,

F. Van Noten, J. O. Vogel, W. Watson, F. Wendorf, and F. Willett.

177

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

REFERENCES

Addison, Frank, Jebel Moya. Oxford. 1949.

Ajayi, S. S. Wild life as a source of protein in Nigeria. pp. 115-127.

Alexander, John, The indirect evidence for domestication. plants and animals. Ed. Peter J. Ucko & G. W. Dimbleby, Alexander, John & D. G. Coursey. The origins of and exploitation of plants and animals . Ed. (1969), pp. 405-25. London.

Allchin, F. R. Early cultivated plants in India and (1969), pp. 323-9.

Ames, O. Economic annuals and human cultures , 1939. Cambridge, Mass.

Anderson, E. Plants , man and life (1952). Boston.

- -The evolution of domestication in Evolution after Arkell, A. J. Early Khartoum , (1949), Oxford.

487-9.

Arkell, A. J. & Ucko, P. J. Review of predynastic development Current Anthrop. 6 (1965), pp. 145-56.

Armstrong, Robert, G. The use of linguistic and ethnographic Idoma and Yoruba history. In The Historian in Tropical R. Mauny and L. V. Thomas (1964), pp. 127-144. London.

Assemien, P. Etude comparative des flores paysages végétaux. C.N.R.S. Paris, Centre Ayrton, Edward R. & W. L. S. Loat. Predynastic (1911).

Baker, H. G. "Comments on the thesis that there was a major centre of plant domestication

near the headwaters of the River Niger",/. Afr. Hist. 3(2) (1962), pp 229-

33.

In Brokensha, D. ed. Ecology and economic development (1965), pp. 185-216. Institute for International Studies, Berkeley, Research Series No. 9.

C. W. Pennington & R. L. Rands ed. Man across London.

Balout, L. Préhistoire de V Afrique du Nord . Paris (1955).

the Nile Valley". Current Anthrop. 6(2) (1965), p. 156.

178

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Baumgartel, Elise J. The cultures of prehistoric Egypt I. Revised edition. Oxford 1955.

85(4) (1965), pp. 502-11.

Binford, L. R. "Post-Pleistocene adaptations" In Binford, New Perspectives in Archaeology Chicago (1968), Bishop, W. W. & Desmond Clark, J. {ed.) Background 1967.

Blakney, C. P. "On 'banana' and 'iron'; linguistic footprints Hartford Studies in Linguistics No. 13. Hartford (1963).

Bokonyi, S. "Archaeological problems and methods In The Domestication and exploitation of plants Ucko & G. W. Dimbleby (1969), pp. 219-29. London.

Bonnefille, R. "Atlas des pollens d'Ethiopie", Adansonia 518.

Pollen et Spores 13(1) (1971)(Z>), pp. 1 5-72.

Braidwood, Robert J. & Bruce Howe "Prehistoric Studies in Ancient Oriental Civilisation (1960), p. 31. Brentjes, B. Die Haustierwerdung im Orient. Wittenberg Bronckers, F. "Palynologie africaine VII. Bull. I.F.Brown, Jean "The excavation of a group of burial Kenya", Azania 1 (1966), pp. 59-77.

Burkill, I. H. A Dictionary of economic products of the (1935).

as "Yam" to European Languages", Proc. Linnean Soc. de Bruxelles 15 (1939), pp. 345-92.

Burton-Page, J. "The problem of the introduction of Adansonia In The domestication and exploitation of plants and animals. & G. W. Dimbleby (1969), pp. 331-5. London.

Busson, F. Plantes alimentaires de I ouest africain . Marseilles. 1965.

Anthrop. 6(2) (1965), pp. 157-8.

Butzer, Karl W. Environment and Archaeology, 2nd edition. (1971).

Camps, Gabriel "Amekni. Neolithique ancien du Hoggar", Mémoires du CRAPE.

X. Paris 1969.

Carneiro, Robert L. "Slash and burn cultivation for cultural development in the Amazon Systems in Native South America: causes Wilkers. Caracas, Sociedad de Ciencias Naturales Caton-Thompson, Gertrude & E. W. Gardner Chang, Kwang-Chin. "The beginnings of Agriculture (1970), pp. 175-85.

179

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Chowdhury, K. A. & G. M. Buth. "4,500 year old seeds suggest that true cotton is

indigenous to Nubia", Nature 227 (1970), pp. 85-6.

Clark, Grahame "Radiocarbon dating and the spread of farming economy", Anti -

quity 39(153) (1965), pp. 45-48.

Clark, J. Desmond The Prehistory of southern Africa. of Rhodesia and Nyasaland. Ed. Roger Summers, p. 207. Panafricain de Préhistoire. Ed. G. Tervuren 1962.

Lisbon 1963.

Rhodesia", Nature 4923 (1964), pp. 971-75.

Evolution in Africa , ed. Walter W. Bishop and J. Chicago (1967).

Museo do Dundo, No. 78. Lisbon (1968).

Ténéré, Republic of Niger, January to March 1970". circulated. (1970) (a).

Evidence for agricultural origins in the Nile pp. 34-79 (19716).

Coatzee, J. A. "The morphology of Acacia pp. 23-7.

Cohen, Mark "A reassessment of the Stone Bowl Cultures of the Rift Valley, Kenya".

Azania 5 (1970), pp. 27-38.

Cole, Sonia The prehistory of East Africa. London (1964).

Connah, Graham "Progress report on archaeological work in Bornu 1964-1966",

Northern History Research Scheme , Second Interim Report (1967), pp. 20-31.

Zaria.

/. Hist. Soc. Nig. IV (1968), pp. 313-20.

pp. 55-60.

Cooke, C. K. "Rock art in Matabeleland". In Summers, 1959.

Africa 35(3) (1965), pp. 263-85.

Coursey, D. G. Yams. London 1967.

Dalziel, J. M. "The useful plants of West Tropical Africa". Second reprint. London

1955.

180

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Darlington, e. D. "The silent millennia in the origin of agriculture". In The domestication

and exploitation of plants and animals. Ed Peter J. Ucko & G. W. Dimbleby

(1969), pp. 67-72. London.

David. N. "Personal communication". ( 1971).

Davies, O. The Quaternary in the Coastlands of Guinea. Glasgow. (1964).

pp. 479-82.

Davies, O., H. J. Hugot, & David Seddon "The origins of African agriculture".

Current Anthrop. 9(5) (1968), pp. 479-504.

De Candolle, A. Origin of cultivated plants. London. (2nd edition 1959). 1886.

De Wet, J. M. J. & P. Huckabay "The origin of Sorghum bicolor. Distribution and

domestication", Evolution 21(4) (1967). pp. 787-802.

Deevey, Edward S., Richard Foster Flint & Irving Rouse "Editorial Statement",

Radiocarbon 5 (1963), V.

Degerbol, M. "Z. Tierzuchtung u Zuchtungsbiologie 76 (1962), pp. 334-41.

tures". In Fagan, 1967, pp. 198-207.

Deschamps, M. "Histoire de Madagascar". Coll. Mordes Doggett, H. "The development of the cultivated sorghums". evolution. Ed. Sir Joseph Hutchinson, pp. 50-69. Cambridge. Dombrowski, Joanne "Preliminary report on excavations Caves, Begemeder". Annales ď Ethiopia 8 (1969), Dorst, Jean & Pierre Dandelot A field guide to the London. (1970).

Drew, I. M., Dexter Perkins, JR. & Patricia Daly animals : effects on bone structure". Science 171 (Ehret, Christopher "Cattle-keeping and milking history: the linguistic evidence". /. Afr. Hist., 8 (Epstein, H. "The origin and history of African cattle". British Dependent Territories in Africa , London, (1968), pp. 213-21.

Fagan, Brian "The Greefswald Sequence: Bambandyonalo J. Afr. Hist., 5(3) (1964), pp. 337-61 .

506.

Fagan, Brian, Phillipson D. W. & Daniels S. ii (1969), London.

Fagg, B. E. B. "The Nok culture, W. Afr. Review, Prehistory, Livingstone, 1955, pp. 306-312. London 1957.

181

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Faris, D. G. "Evidence for the West African origin Dissertation, University of California (1963).

Filipowiak, W., Jasnosz S. & Wolagiewicz R. "Polskogwinejskie w Niani w 1968 r (Les recherches archéologiques Niani en 1968)". Materiaty Zachodniopomorskie Flannery, Kent "The ecology of early food production 147(1965), pp. 1247-55.

In Ucko & Dimbleby ed., (1969), pp. 72-100.

Flight, C. "Excavations at Kintampo", W. Afr. Archaeol Forde, Daryll, "The cultural map of West forests and grasslands", Transactions Series II, 15 (5) (1953), pp. 206-19.

Fosbrooke, H. A., Phyllis Ginner, Leakey rock paintings: a guide and record", pp. 1-61.

Fouche, Leo, Mapungubwe. Cambridge. (1937).

Friedel, M. C. "Sur des matières grasses trouvées dans les tombes Egyptiennes

ď Aby dos", C. R. Acad. Se/., 124 (1897), pp. 648-53.

Gabel, Creighton, "Seminar on economic types in pre-urban cultures of temperate

woodland, arid and tropical areas". Current Anthrop. 1 (1960), pp. 437-38.

"Six rock shelters on the northern Kavirondo Shore of Lake Victoria", African

Historical Studies , (Boston) 29 (2) (1969), pp. 205-54.

Gaherty, Geoffrey, "The Human skeleton from Rop rock Shelter", W. Afr. Archaeol.

Newsl. 9(1968), pp. 18-19.

Gardner, Guy A. Mapungubwe , Vol. II (1963), Pretoria.

Good all, E, "The rock paintings of Mashonaland", In Summers, (1959).

Gorman, Chester, "Hoabinhian: a pebble-tool complex with early plant associations

in southeast Asia", Proc. Prehist. Soc.9 35 (1969), pp. 355-8.

Gray, Richard, "A report on the conference: Third Conference on African History

and Archaeology", /. Afr. Hist., 3 (2) (1962), pp. 175-91 .

Graziosi, P. "New discoveries of rock paintings in Ethiopia", Antiquity , 3 8 (1964)

pp. 91-8, pp. 187-90.

Greenberg, J. H. Studies in African linguistic classification. (1955), New Haven.

Grundeman, Tulli, "Wenner Gren Conference on Bantu Origins , Bull. Afr. Stud.

Assoc. U.K., 13 (1968), pp. 1-9.

Guers, J. "Palynologie africaine X", Bull . I.F.A.N, 32(A) (1970), pp. 312-65.

Guinet, P. H. "Palvncloeie africaine VIII", Bull. I.F.A.N., 30(A): (1968), pp. 151-66.

Guthrie, Malcolm, "Some developments in the prehistory of the Bantu languages",

J. Afr. Hist., 3 (1962), pp. 273-82.

Harlan, Jack R. "Personal communication", 1968.

- -"Ethiopia: a centre of diversity", Econ. Bot., 23 (4) (1969), pp. 309-314.

182

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Harlan, Jack R. & J. M. J. de Wet, "On the quality of the evidence for origin and

dispersal of cultivated plants", Circulated typescript proposed for Current Anthropology

(197 1).

Harris, David R. "The prehistory of tropical agriculture : an ethnoecological model".

Paper presented to the Research Seminar on Archaeology and Related Subjects,

Sheffield, December, 1971. Roneo'd.

Hartley, C. W. S. The Oil Palm. London, (1967).

Heinzelin de Braucourt, J. de. Les Fouilles ďlshango. Exploration du Parc

National Albert, 1950, vol. 2 Brussels, (1957).

Helbaek, H. "Ancient Egyptian Wheats", Proc. Prehist. Soc., 21 (1955), pp. 93-95.

Hemarddinquer, Jean-Jacques, Michael Keul & W. G. L. Randles Sorgho.

Documentation géographique et historique. Centre de Recherches Historiques

de l'Ecole Pratique des Hautes Etudes, Paris. 1967.

Hepper, F. N. "The Bambara Groundnut ( Voandzeia subterranea ) and Kersting's

Groundnut ( Kerstingielìa geocarpa ) wild in West Africa", Kew Bull., 16(3)

(1963). pp. 395-407.

Herre, Wolf, "The Science and History of Domestic Animals. In Brothwell, Don &

Eric Higgs ed. Science and Archaeology (London 1963), pp. 235-49.

Hiernaux, Jean, "Bantu expansion: the evidence from physical anthropology confronted

with linguistic and archaeological evidence" J. Afr. Hist., 9(4) (1968),

pp. 505-15.

Higgs, e. S. "Early domesticated animals in Libya". In Background to Evolution in

Africa. Ed. W. W. Bishop & J. Desmond Clark, 1967(a), pp. 165-73.

"Domestic Animals". In The Haua Fteah, by C. B. M. McBurney (1967(6)),

pp. 313-19. Cambridge.

Higgs, E. S. & Jarman, M. R. "The origins of agriculture; a reconsideration",

Antiquity, 43 (169) (1969), pp. 31-41 .

Hobler, Philip M. & James J. Hester, "Prehistory and environment in the Libyan

Desert", S. Afr. Archaeol. Bull., 33 (1969), pp. 120-1 30.

Huffman, Thomas N. "The Early Iron Age and the spread of the Bantu", S. Afr.

Archaeol. Bull., 25(1) (1970), pp. 3-21 .

Hugot, H. J. ed. Missions Berliet - Ténéré-Tchad. Paris. 1962.

du C.R.A.P.E. Paris. 1963.

Hutchinson, Sir Joseph ed. Essays in Crop Plant Evolution. Cambridge, Inskeep, R. "The age of the Kondoa rock paintings in the light at Kisese II Rock Shelter", In Actes du IVe Congrès Panafricain l'Etude du Quaternaire. Ed. G. Mortelmans & J. Nenquin, Jackson, G. "Personal communication (Mr George Jackson, University of Ibadan, Nigeria). (1970).

Jarman, M. R. "The Prehistory of Upper Pleistocene and Recent Mediterranean, with reference to north-west Europe", 35 (1969), pp. 236-66.

Johnston, Sir, H. H. "A survey of the ethnography of Africa: and tribal migrations in that continent", Jl Roy. Anthrop. pp. 375-421.

183

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Kaiser, W. "Zur Inneren Chronologie der Naqadakultur", Archaeologia Geographica

6 (1957), pp. 69-77.

Keimer, L. Chronique ď Egypte 27(2) (1953)., pp. 268-71.

Kennedy-O'Byrne, J. "Notes on African grasses XXIX - a new species of Eleusine

from Tropical and South Africa", Kew Bull., 11 (1957), pp. 65-72.

Laurent-Tackholm, V. "The plant of Naquada", Annales du Service des Antiquités

de V Egypte, 51 (1951), pp. 299-312.

Lawton, J. R. S. "Personal communication. (The late Dr J. R. S. Lawton, formerly

of the Botany Department, University of lbadan, Nigeria) 1966.

Leakey, L. S. B. The Stone Age cultures of Kenya colony . London, 1931 .

Leakey, M. D. "Notes on the ground and polished stone axes of East Africa",

J. East A frica and Uganda Nat. Hist. Soc., 17 (3 and 4), (1 943), pp. 1 82-1 95.

- - "Report on the excavations at Hyrax Hill, Nakuru, Kenya Colony, 1937-38",

Trans. Roy. Soc. S. Africa , 30, Part 4 (1945).

"Excavation of burial mounds in Ngorongoro crater", Tanzania Notes and

Records , 66 (1966), pp. 123-35.

Leakey, M. D. & Leakey, L. S. B. Excavations at the Njoro River Cave. Oxford

(1950).

Lebeuf, A. M. D. Les principautés Kotokos. Paris. (1969).

Linares de Sapir, Olga, "Shell middens of Lower Casamance and problems of

Diola protohistory", W. Afr. J. Archaeol. 1 (1971), pp. 23-54.

Livingstone, D. A. "Post-glacial and post-pluvial profiles from tropical Africa",

Abstracts of the 10th International Botanical Congress . Edinburgh, (1964) p. 269.

Livingstone, D. A. & Kendall, R. L. "Stratigraphie studies of East African lakes",

Mitt. Internat. Verein. LimnoL, 17 (1969), pp. 147-53.

Lobreau, D., J. Guers, P. Assemien, G. Bou, P. H. Guinet & L. Potier, "Palynologie

africaine IX", Bull. I. F. A.N., 31(A) (1969), pp. 41 1-60, pls. 167-90.

McBurney, e. B. N. The Stone Age of North Africa. Harmondsworth. (1960).

McNeish, R. S. "The Origins of American Agriculture", pp. 87-94.

Maley, J. "Contribution à l'etude du bassin tchadien Atlas du pollens du Tchad",

Bull. Jard. Bot. Nat. Belgique , 40(1) (1 970), pp. 29-48.

Marshall, Sir J. Mohenjo-Daro and the Indus Valley Civilization. London, 1931.

Mason, R. J. "Discovery of new paintings in the Brandberg, South West Africa, and

Waterberg, Transvaal". Lantern 7(4): 1958, p. 368.

Wit. I.

Mauny, Raymond. "Notes historiques autour des Afrique occidentale. Bull de V I.F.A.N. 15(2): Africa, ed. Walter W. Bishop and J. Desmond Clark, Meer, P. P. C. van, "Personal communication" (Mr P. P. of Forestry, University of lbadan, Nigeria). 1971.

184

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Mehra, K. L. Natural hybridization between Eleusine corocana and E. africana in

Uganda. J. Indian Bot. Soc. XLI (4): (1962), pp. 531-9.

Science 7: 1963(a) pp. 300-1.

pp. 189-98.

Meighan, Clement W. et al. Ecological interpretation in archaeology: Part 2.

American Antiquity 24: (1958), pp. 131-50.

Miege, J. L'appui de la palynologie dans la distinction des espèces africaines de

Dioscorea. (1965) Webbia 19, p. 491.

Monod, Th. "The late Tertiary and Pleistocene in the Sahara". In African Ecology and

Human Evolution, ed. F. C. Howell & Bourlière, (1964), pp. 117-229. Chicago.

Morgan, W. B. "The forest and agriculture in West Africa."/. Afr. Hist. 3(2): 1962,

pp. 235-9.

Mori, F. Tadrart Acacus. 1965. Turin.

Mortelmans, G. & J. Nenquin ed. Actes du lVe Congrès Panafricain de Préhistoire

et de V Etude du Quaternaire. 1962. Tervuren.

Munson, Patrick J. "Recent archaeological research in the Dhar Tichitt region of

south-central Mauretania." W. Afr. Archaeol. Newsl. 10: 1968, pp. 6-13.

Afr. Archaeol. Newsl. 12: (1970), pp. 47-8.

Murdock, G.P. Africa : Its peoples and their culture history Neustupny, Evzen. "A new epoch in radiocarbon dating." pp. 38-45.

Okiy, G. E. O. "Indigenous Nigerian food-plants", J. W. Afr. Sci. Assoc. 6(2): (1960),

pp. 117-21.

Oliver Roland. "Bantu Genesis", African Affairs , 65 (260), (1966)(a)

Olsson, Ingrid U. ed. " Radiocarbon variations and absolute York.

Onwuejeogwu, M. A. The culture history of an Anambra civilisation :Nn culture",

(1971). African Notes (Ibadan) forthcoming.

Owen, J. ''Water storage properties of Adasonia digitata 1968" (Baobab), W. Afr.

Archaeol. Newsl. 9, pp. 55-6.

in African communities. African Notes 6 (1), pp. 24-Payne, W. J. A. "The origin of domestic cattle in Africa", Agriculture 32 (126), (1964), pp. 97-113.

Peet, T. Eric & S. L. S. Loat. The cemeteries of Abydos. Phillipson, D. W. "The Early Iron Age Site at Kapwirimbwe, 3, pp. 87-105.

J. Afr. Hist. 11 (1), 1970(a). pp. 1-5.

Phillipson, D. W. & Brian fagan The date of the J. Afr . Hist. X (2), pp. 199-204.

185

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

PoRTERES, Roland, Eleusine coracana Gaertner. 1951. Bull, de VI.F.A.N. 13, pp. 1-78.

(2), pp. 195-210.

Posener, G., S. Saulneron & J. Yoyette. Dictionnaire Paris. (1959).

Posnansky, M. ' The Iron Age in East Africa", in Bishop Quetzel, P. "Flore et palynologie sahariennes", Bull. I.F.A.N. Quetzel, P. & C. Martinez. ''Etude palynologique de deux diatomites Bull. Soc. Hist . Nat. Afrique Nord 49, (1958), pp. 230-44.

gique et paléoclimatique. Libyca 6/7, pp. 21 1-27.

méridional à l'occasion de la Mission Berliet-Tchad". Berliet Ténéré-Tchad. Paris (1962), pp. 313-27.

Rachie, K. O. (International Institute of Tropical Agriculture). (1971).

Raymond, W. D. The Oil Palm Industry. Trop. Sci. 3, Redhead, J. F. (Department of Forestry, University 1971.

Reed, Charles A. A review of the Archaeological Evidence on Animal Domestication

in the Prehistoric Near East. 1960. In Prehistoric Investigations in Iraqi

Kurdistan . Ed. Robert J. Braidwood & Bruce Howe, pp. 119-45. Chicago.

domestication and exploitation of plants and animals. Dimbleby (1969), pp. 361-80. London.

Robinson, K. R. "An early Iron Age site from theChibi S. Afr . Archaeol. Bull . 16, (1961), pp. 75-102.

/. Afr. Hist. VII (2), (1966) (a) pp. 169-188.

S. Afr. Archaeol . Bull. 21 (1), (1966)(6) pp. 5-51.

Antiquities of Malawi (1970).

Robinson, K. R. & B. Sandelowsky "The Iron recent work. Azania 3, (1968), pp. 107-146.

Robinson, K. R., R. Summers & A. Whitty. 1961. Zimbabwe Excavations 1958 .

Bulawayo.

Sandford, K. S. & W. J. Arkell 1933. Palaeolothic Man and the Nile Valley in Nubia

and Upper Egypt. Chicago.

Sassoon, Hamo. "New views on Engaruka, northern Tanzania", /. Afr. Hist., 7 (2),

(1967), pp. 201-17.

3: p. 2.

186

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Records. 69 (1968)(6), pp. 15-32.

Geog . Soc. Research Reports , 1965 Projects (197 1), Sauer, Carl O. Agricultural origins and dispersals. Seddon, David. 'The origins and development of agriculture Africa", Current Anthropology 9 (5) (1968), pp. rent Anthrop. 10 (5) (1969), pp. 524-5.

Shack, William A. The Gurage. A People of the Ensete Shaw, Thurstan. "Field Research in Nigerian Archaeology", 2(4) (1964), pp. 449-464.

Anthrop . 9 (5) (1968), pp. 500-1.

Anthrop. 10 (Z-3) (1969)(a) pp. 226-231.

national ď Archéologie Africaine (1969)(6), pp. 34-73. Fort Shinnie, P. Excavations at Bigo, 1957", Uganda Journal 24(Simmonds, N. W. (1959). Bananas. London.

Simoons, Fredrick J. 4 Some questions on the economic Afir. Hist . 6 (1) (1965), pp. 1-12.

Smith, F. G. "Some pollen grains in the Caesalpinaceae Spores 6 (1) (1964), pp. 85-98.

Smith, H. S. "Animal domestication and animal cult Dimbleby ed. (1969), pp. 307-314.

Smith, P. 1971. (University of Montreal) Personal communication.

Solheim II, Wilhelm G. "Reworking southeast Asian (1969), pp. 125-139.

Soper, R. C. "Personal communication". (1971).

Soper, R. C. & Golden, B. "An archaeological survey Azania 4 (1969), pp. 1 5-79.

Sowunmi, M. A. "Pollen morphology in the Palmae, in aperture development", Rev. Palaeobot. & Palynology, Struever, Stuart "Flotation techniques for the recovery remains", American Antiquity , 33(3) (1968), pp. Stuiver, Minze & Hans E. Suess. "On the relationship and true sample ages", Radiocarbon 8 (1966), pp. Summers, R. 1958. Inyanga. Cambridge.

- - ed. 1959. Prehistoric Rock Art of the Federation Salisbury, Rhodesia.

Sutton, J. "The archaeology and early peoples Northern Tanzania", Azania. 1 (1966), pp. 37-187

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Swart, E. R. "Age of the baobab tree", Nature , 198(1963), p. Ucko, P. J. & Dimbleby, G.W. ed. The domestication and exploitation animals . London. (1969).

Van Campo, M. "Palynologie africaine I", Bull. I.F.A.N. , 19(Van Campo, M., L. Bertrand, & F. Bronckers, "Palynologie Van Campo, M., B. de Keyser, PH. Guinet & F. Roland-heydacker", Bull.

I.F.A.N . 26(A) (1964.) pp. 1037-70.

Van Campo, M., Bronckers, F & Guinet, P. "Palynologie africaine VI", Bull.

I.F.A.N ., 27(A) (1965), pp. 795-842.

Van Campo, M' Nicolas Halle, Palynologie africaine III. Bull. I.F.A.N. 21(A) (1959),

pp. 808-99.

Van Campo, M. & R. Coque, Palynologie et géomorphologié dans le sud tunésien

Pollen et Spores 2: 1960 275-84.

Van der Merwe, N. J. Personal communication (1971).

Van der Merwe, N. J. & Robert T. K. Scully, "The Phalaborwa story: archaeological

and ethnographie investigation of a South African Iron Age group",

World Archaeology, (1971).

Van Noten, Francis, The Uelian. A culture with a neolithic aspect, Uele - Basin

(N.E. Congo Republic). Tervuren. (1968).

"A ground axe from Burundi", Azania , 4 (1969), pp. 166-7

Personal communication (1971).

Van Zinderen barker, E. M. Palaeoecology of Africa 1965-69, pp. 1-5. Cape Town.

tation changes in sub-Saharan Africa. In Background Ed. Walter W. Bishop and J. Desmond Clark, Chicago, in Africa. Ed. Walter W. Bishop & J. Desmond Clark, Vansina, J. Personal communication. (Professor J, Vansina, U.S. A.) (1970).

Vavilov, H. "The origin, variation, immunity and breeding Chronica botanica , 13 (1951), (1949-50), pp. 1-364.

Vérin, Pierre, "L'Archéologie à Madagascar". Azania 1 (Vishnu-mittre, "Protohistoric records of agriculture Inst., 31(3) (1968), pp. 87-106.

Vogel, Joseph O. "On early evidence of agriculture in Anthropology, 10 (5) (1969), p. 524.

"Early Iron Age tools from Chundu Farm, Zambia", Azania, Personal communication. 1971.

Watt, G. Dictionary of the economic products of India. London & Calcutta. (1 893).

Welbourne, R. G. Provisional assessment of Transvaal Iron Age economies with

special reference to the Melville Cave 9/65 foodwaste assemblage". In

Mason, R. J. et al Prehistoric Man at Melville Koppies, Johannesburg. University

of Witwatersrand, Dept. of Archaeology, Occasional Paper No» 6. 1971 .

188

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

Wendorf, Fred. Summary of Nubian Prehistory. In The prehistory of Nubia , ÍI Ed.

Fred. Wendorf, 1968. pp. 1041-59.

ed. ThePrehistoy of Nubia , Vols. 1/ II (1968), Dallas.

Personal communication. (1971).

Wendorf, F., Said, R. & Schild. R. "Egyptian prehistory: Some new concepts",

Science 169 (1970), pp. 1161-71.

Werth, E. Angew Bot., 19 (1937) p. 42.

West, R. G. & C. N. B. McBurney, "The Quaternary deposits at Hoxne Suffolk,

and their achaeology". Proc. Prehist. Soc. 20 (2) (1954), pp. 131-154.

Willcox, A. R. The rock art of South Africa . Edinburgh. 1963.

Williamson, Kay, "Some food plant names in the Niger Delta.", Internat . /. Amer .

Linguistics , 36 (1970), pp. 156-167.

Weight, Gary A. "Origins of food production in southwestern Asia: a survey of

ideas". Current Anthropology, 12 (4-5) (1971), pp. 447-77.

Wrigley, Christopher, "Speculations on the economic prehistory of Africa",

/. Afr. Hist. 1 (2) (1960), pp. 189-203.

Yarnell, Richard A. "Comment on Davies, Hugot & Seddon", Current Anthrop .

9 (5) (1968), p. 502.

Zeuner, F. E. A History of Domesticated Animals. London, (1963).

189

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

SITES MENTIONED IN THE TEXT

This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTC

All use subject to https://about.jstor.org/terms

StTES MENTIONED IN THE TEXT 1 . Amekni 2. Fay um 3. Meniet 4. Badari 5. Merimda 6. Abydos, El Mahasna 7. Nagada 8. Shaheinab 9. Niani 10. Daima 11. Engaruka 12. Chundu 13. Inyanga 14. Mwamasapwa, Nkope, Phopo Hill 15. Early Khartoum 16. Kalundu, Isamu Pati 17. Mapungubwe, Bambandyanalo 18. Ingombe Ilede 19. Klipriviersberg, Olifantspoort 20. Khami 21. Kintampo 22. Nok 23 . Njoro R iver Cave 24. Ilkek, Hyrax Hill, Elmenteita, Nakuru 25. Ngorongoro 26. Adrar Bous 27. DharTichitt 28. Ntereso 29. Dungul 30. Jebel Uweinat 31. HauaFteah 32. Kapwirimbwe 33. Kumadzulo 34. Iwo Eleru 35. Ishango 36. Lalibela and Natchabiet Caves 37. Casamance 38. Acacus 39. Jebel Moya 40. Birnin Kudu 41. Senafe 42. Harar 43. Brandberg 44. Kom Ombo 45. Narosura 46. Tunnel Rock Shelter 47. Nyang'oma 48. Rangong 49. Kalambo Falls 50. Lusaka 51. Mabveni 52. Zimbabwe. Abydos 6 Acacus 38 Adrar Bous 26 Amekni 1 Badari 5 Bambandyanalo 17 Birnin Kudu 40 Brandberg 43 Casamance 37 Chundu 12 Daima 10 Dhar Tichitt 27 Dungul 29 Early Khartoum 15 El Mahasna 6 Elmenteita 24 Engaruka 1 1 Fayum 2 Harar 42 HauaFteah 31 Hyrax Hill 24 Ilkek 24 Ingombe Ilede 18 Inyanga 13 Isamu Pati 16 Ishango 35 Iwo Eleru 34 Jebel Moya 39 Jebel Uweinat 30 Kalambo Falls 49 Kalundu 16 Kapwirimbwe 32 Khami 20 Kintampo 21 Klipriviersberg 19 Kom Ombo 44 Kumadzulo 33 Lalibela and Natchabiet Caves 36 Lusaka 50 Mabveni 51 Mapungubwe 17 Meniet 3 Merimda 5 Mwamasapwa 14 Nagada 7 Nakuru 24 Narosura 45 Ngorongoro 25 Niani 9 Njoro River Cave 23 Nkope 14 Nok 22 Ntereso 28 Nyang'oma 47 Olifantspoort 19 Phopo Hill 14 Rangong 48 Senafe 41 Shaheinab 8 Tunnel Rock Shelter 46 Zimbabwe 52 191This content downloaded from 146.230.187.153 on Tue, 16 Apr 2019 07:10:38 UTCAll use subject to https://about.jstor.org/terms

Permitted Uses of the Content 4.1 Institutional Licensees and/or Authorized Users may search, view, reproduce, display, download, print, perform, and distribute Licensed Content for the following Permitted Uses, provided they abide by the restrictions in this Section 4, Section 5 and elsewhere in these Terms and Conditions of Use: (a) research activities; (b) classroom or organizational instruction and related classroom or organizational activities; (c) student assignments; (d) as part of a scholarly, cultural, educational or organizational presentation or workshop, if such use conforms to the customary and usual practice in the field; (e) authors or other Content creators may at their discretion incorporate their Content (other than Books) into unrestricted databases or website with prior permission from the publisher and other applicable rights holders; (f) linking (g) Data for Research as defined further in Section 14.3 below. and for Institutional Licensees and/or Authorized Users other than Register & Read Users: (h) on an ad hoc basis and without commercial gain or in a manner that would substitute for direct access to the Content via services offered by JSTOR, sharing discrete portions of Content for purposes of collaboration, comment, or the scholarly exchange of ideas; (i) downloading Content (including certain Book chapters where downloading is made available for the Book chapters on the JSTOR Platform) to view, use and display including on a personal digital device (e.g. mobile devices, e-readers, and personal computers); (j) (except for Books) in research papers or dissertations, including reproductions of the dissertations, provided such reproductions are only for personal use, library deposit, and/or use solely within the institution(s) with which the Authorized User and/or his or her faculty readers are affiliated; and (k) fair use under Section 107 of the U.S. Copyright Act, educational exceptions, or other similar provisions of the copyright laws or other intellectual property right laws in the United States. 4.2 The portions of Licensed Content that have been downloaded or printed out by an Authorized User may continue to be used in compliance with these Terms and Conditions even if such license should later terminate or expire. 4.3 Additional Permitted Uses of the Early Journal Content. JSTOR encourages broad use of the Early Journal Content. In addition to all permitted uses described above, users are free to copy, use, and redistribute the Early Journal Content in part or in whole for non-commercial purposes. JSTOR asks that you acknowledge JSTOR as the source of the Early Journal Content; if you use material from JSTOR online, we request that you link directly to the stable URL provided. If you use Early Journal Content offline, we ask that you credit the source as follows: “Courtesy of JSTOR.” Please be considerate of other users and do not use robots or other devices or coordinate activity to systematically download these works as this may be disruptive to our systems. 4.4 Use of Open Access Content. Open Access Content may be subject to different terms of use, such as a Creative Commons license, as indicated in the copyright statement for the content. If an Open Access Content item has different terms posted where the content appears, those terms shall have precedence over these Terms and Conditions of Use.

Nelisiwe Ndlovu

04/17/2019 - 14:44