Published online by Cambridge University Press: 22 January 2009
The archaeological study of African agricultural history has concentrated more on origins and the identification of crop remains than on farming techniques and ancient fields. The latter rarely survive, and even then may be of indeterminate age and apparently atypical, such as hillside terracing or irrigation systems. The subject has been further bedevilled by these unusual instances of preserved fields being regarded as relics of ‘intensive’ cultivation by ‘vanished civilizations’. However, a clearer understanding of African agriculture, through ethnographic and ecological approaches, reveals not only its basically extensive character but also the infinite variety of local specializations (or cultural–environmental adaptations, combining ancient African domesticates and introduced crops which have been ‘Africanized’). This provides a perspective for examining peculiarities of the present and past and claimed instances of ‘intensification’. Conversely the concept of specialization allows us to use, with caution, the preserved remains of old field systems to illustrate more typical ones. Many of the archaeological survivals were not so much ‘intensive’ as ‘over-specialized’, often in isolated and circumscribed situations, notably remote hills in both western and eastern Africa. A moderate example is Inyanga in eastern Zimbabwe with its extensive terraced hillsides of the later Iron Age. Here most of the terraces were not irrigated, but there are hints of complex seasonal arrangements and field techniques. A more extreme, even ‘ultra-specialized’ agricultural system, also of the later Iron Age, which was abandoned in the seventeenth or eighteenth century, is that of Engaruka in the northern Tanzanian rift. This was an essentially isolated and self-sufficient settlement in dry country, absolutely dependent on its exquisite irrigation devices. Eventually this community expired, as its soil was exhausted and its water supplies declined.
Finally, there are instances in nineteenth-century East Africa, and from earlier in West Africa, of more open cultural–economic systems producing a surplus for caravans, markets and towns. Technologically these have been no more accomplished or ‘intensive’ than the specializations discussed, but developmentally their achievement has been more effective.
2 See in particular Harlan, J. R. et al. (eds.), The Origins of African Crop Domestication (The Hague, 1976),CrossRefGoogle Scholar especially chapters by Thurstan Shaw and J. Desmond Clark; also Shaw in Megaw, J. V. S. (ed.), Hunters, Gatherers and First Farmers beyond Europe (Leicester, 1977).Google Scholar
3 This distinction is made by, among others, Sauer, C. O., Agricultural Origins and Dispersals (Cambridge, Mass., 1952).Google Scholar The further dichotomy made between prepared fields of sown grains and more complex (and seemingly informal) gardens of planted, root, ground, vine and tree crops, remains useful impressionistically for tropical and equatorial climes. In general, planted crops and garden-tending prevail in wetter and more wooded environments. But since the introduction to Africa of cassava and maize from the New World, and with the increasing international popularity of rice, a wet grain, the neatness of this set of distinctions is somewhat eroded.
4 See Harlan et al., Origins, especially editors' introduction and contribution by D. R. Harris; also, for world background, Harlan, , ‘Agricultural centers and noncenters’, Science, CLXXIV, 468–474.Google Scholar It is expected that a promised volume edited by J. D. Clark and S. Brandt, Causes and Consequences of Food-production in Africa, will carry the discussion further. In recognizing complexity, factual and conceptual alike, it is important not to throw precision to the winds.
5 As Langland did for western England in the fourteenth century in his Piers Plowman.
7 The obvious introduction is Allan, William, The African Husbandman (Edinburgh, 1965).Google Scholar
8 For a stimulating exploration of this distinction see Goody, Jack, Technology, Tradition and the State in Africa (Cambridge, 1971), 25 f.Google Scholar
9 Compare midland England with its mediaeval ridge-and-furrow, and superimposed on that the hedgerow patterns of the enclosure movement.
10 Murdock (Africa, pp. 23, 68–70) remains useful for a reference list, if not reliable on every detail. What is more misleading in his reconstruction is his espousing of the concept of agricultural ‘hearths’ or ‘centres’, and with that the supposition that a particular period and region of West Africa should be identified for the domestication of a whole range of crops, which derive in fact from different ecological zones. On sorghum, see the final chapters of Harlan, et al. ; and Stemler, Ann B. L. et al. in J. Afr. Hist. XVI (1975), 161 f.Google Scholar Arguments for the antiquity of the exploitation of sorghum and certain millets in the north-eastern part of sub-Saharan Africa are presented in this Journal by L. Krzyzaniak (XIX, 1978, 159f.) and C. Ehret (XX, 1979, 161 f.), the one using new archaeological evidence on the Middle Nile, the other comparative linguistics in Ethiopia.
11 Murdock, Africa, ch. 25; Allan, African Husbandman, 199 f., 279, partly based on Davidson, Basil, Old Africa Rediscovered (London, 1959).Google Scholar
12 E.g. Jensen, A. E., Im Lande des Gada (Stuttgart, 1936), 576f.Google ScholarSummers, R., Inyanga (Cambridge, 1958), 247 f.Google ScholarNetting, R. McC., Hill Farmers of Nigeria (Seattle and London, 1968), 3–4;Google ScholarHale, G., ‘Cultivation terraces of western Darfur’ (Ph.D. thesis, University of California, Los Angeles, 1966).Google Scholar See my discussion in Farrington, , Prehistoric Intensive Agriculture in the Tropics; and for eastern Africa, Sutton in Azania, IV (1969), 1 f.Google Scholar
13 Formerly classified as Coleus dazo: see Codd, L. E. in Bothalia, XI (1975), 377–378.Google Scholar This plant is not to be confused with Solenostemon rotundifolius (formerly Coleus dysentericus, the ‘Hausa potato’, Ibid. 438) cultivated in West Africa. ‘Potato’ is hardly an appropriate name for Plectranthus esculentus; the crop resembles more a cluster of parsnips. (Another name, ‘scrambled egg’, refers presumably to its profuse yellow flowers.) In discussion of African domesticates, this cultigen (like the sub-equatorial regions generally) has received scant notice. It does not appear to be native to the open Inyanga uplands where it is cultivated, but rather to medium-altitude Brachystegia/Julbernardia woodlands covering large parts of eastern and southern Africa, including ironically the adjacent Inyanga ‘lowlands’ where it is rarely if at all cultivated. Wild varieties are eaten in various places, notably southern Tanzania: Greenway, P. J., E. Afr. Agric. J. X (1944), 3–4.Google Scholar It is cultivated in localities of north-central Nigeria: Dalziel, J. M., The Useful Plants of Tropical West Africa (London, 1948), 459.Google Scholar I am indebted to Bud Payne of the Rhodes-Inyanga Experiment Station and Bob Drummond of the National Herbarium in Harare for their assistance and discussion.
15 Randall-MacIver, D., Mediaeval Rhodesia (London, 1906), chs. 1–3;Google Scholar Summers, Inyanga, passim. The ‘pit-structures’ were demonstrably homestead compounds containing a central pen for small stock (the ‘pit’). In a number of instances a direct association between these features and adjacent abandoned terracing can be satisfactorily determined by ground inspection. While the archaeological findings generally corroborate the ethnohistorical indications of a late Iron Age dating, precise evidence is scant. However, the occasional radiocarbon determination and some tentative tree-ring tests are reasonably consistent. (See Huffman, T. N. and Storry, J. G. in S. Afr. J. Science, LXXI (1975), 247–248;Google Scholar also Beach, 345.) The more detailed chronology of Inyanga, and Summers' arguments that the remains in the ‘uplands’ are generally older than those in the ‘lowlands’, based partly on the climatic sequence as understood in the 1950s, need further scrutiny. Ethnohistorically, such a view looks reasonable.
16 Summers, (Inyanga, 233 f., 318)Google Scholar argued, on the grounds of building styles and associations, that it was possible to identify rare early to middle Iron Age (‘Z2 culture’) terraces, and to distinguish these from those of the later ‘ruin culture’. It was not possible to confirm this during the recent fieldwork: stylistic differences are perhaps more likely due to local factors than to date. But these doubts do not rule out the possibility of the terrace system having evolved over many centuries – possibly a thousand years.
17 See my article in Azania, XIII (1978), 37–70;Google Scholar and for earlier work and illustrations, Sassoon, Hamo in this Journal, VIII (1967), 201–217.Google Scholar A report on further work at Engaruka by Peter Robertshaw and myself is being prepared. Radiocarbon tests on new samples from one of the village sites fall around the sixteenth to eighteenth centuries A.D. These are consistent with most of the previous radiocarbon results and corroborate our impression, from the pottery and other considerations, that the Engaruka remains belong to the later Iron Age.
18 For a brief survey of irrigation in East Africa with references, see my Archaeology of the Western Highlands of Kenya (B.I.E.A. memoir 3, Nairobi, 1973), 70–3.Google Scholar
19 Hennings, R. O., African Morning (London, 1951), 202–205.Google Scholar A recent survey and valuable map of the Marakwet furrows by Soper, Robert has appeared in a volume on Kerio Valley: Past, Present and Future edited by Kipkorir, B. E. et al. (Institute of African Studies, University of Nairobi, 1983), 75–95 and pull-out.Google Scholar
21 These are the faced and usually rectangular ‘cairns’ of previous reports. Closer study, and the recognition of incipient examples, now explain that these features accumulated gradually in the corners of fields as neat dumps of excess stones. Suggestions that they may have been working platforms or alternatively defensive devices seem therefore to be ruled out.
22 I am indebted to David Anderson for discussing the findings of his recent historical study of Njemps (Il-Chamus). The irrigation is distinct from that of nearby Marakwet, both in its situation and in its furrow construction.