¹ There is an interesting correspondence with the ‘drought zone’ maps which figure in the international press these days. However, the incidence of drought in the 1970s, as on various other occasions in recent centuries, is due essentially to relatively short-term factors operating within a regime already established by a very much longer climatic trend. It is only such long-term trends which are relevant in the present paper.

² In the accumulating literature on this subject, the following papers are especially illuminating: van Zinderen Bakker, E. M., ‘Late Quaternary lacustrine phases in the southern Sahara and East Africa’, Palaeoecology of Africa, VI (1972), 15–27;Google ScholarFaure, H., ‘Évolution des grands lacs sahariens à l'Holocène’, Quaternaria, VIII (1966), 167–75;Google ScholarHobler, P. M. and Hester, J. J., ‘Prehistory and environment in the Libyan desert’, South African Archaeological Bulletin, XXIII (1969), 120–30;CrossRefGoogle ScholarBurke, K., Durotoye, A. B. and Whiteman, A. J., ‘A dry phase south of the Sahara 20,000 years ago’, West African Journal of Archaeology, 1 (1971), 1–8;Google ScholarGrove, A. T. and Goudie, A. S., ‘Late Quaternary lake levels in the rift valley of southern Ethiopia and elsewhere in Tropical Africa’, Nature, vol. 234 (1971), 403–5;CrossRefGoogle ScholarButzer, K. W., Isaac, G. L., Richardson, J. L. and Washbourn-Kamau, C., ‘Radiocarbon dating of East African lake levels’, Science, vol. 175 (1972), 1069–76.CrossRefGoogle ScholarPubMed The last provides a brief but valuable synthesis of the present state of the subject and cross- continental correlations. Useful for background is Butzer, K. W., Environment and Archaeology (2nd ed., 1971), ch. 20.Google Scholar

³ Waterfalls, of course, can be formidable obstacles. On the other hand, it is sometimes argued—rarely convincingly, however—that fish or their spawn can be transported across physical barriers by birds or other agencies.

⁴ There are claims that cattle-keeping in parts of the Sahara began as early as the sixth millennium: e.g. Mori, F., Antiquity, XLVIII (1974), 88;Google ScholarSmith's, P. E. L.discussion in African Historical Studies, I (1968), 34–6;Google Scholar and the survey by Thurstan, Shaw, J. Mist. Soc. Nigeria, VI (1972), 166. Even if so—and some Saharan authorities seem content to date the introduction of cattle a millennium or so later—the present argument would not be affected. In fact, the development of pastoralism in the Sahara probably corresponds with a decline of aquatic activities enforced by the rather drier conditions of the late sixth and fifth millennia.Google Scholar

⁵ This suggestion of an African invention of pottery is by no means novel: both Leakey and Arkell (see below) had thought on this; and more recently Colin Flight, while commenting on the latest Saharan radiocarbon dates, has put the essence of the point in this Journal, XIV (1973), 536–8, 554.Google ScholarCamps-Fabrer, H. and Camps, G. comment on late seventh millennium pottery in the central Sahara in Revue de l'Occident Musulman et de la MÉditerranÉe, no. 11 (1972), 26;Google Scholar as also does Gabriel, B. in Palaeoecology of Africa, VI (1972), 219.Google Scholar For the Nile valley and the Kenya rift most of the recently processed radiocarbon determinations indicating, directly or indirectly, dates of around 6,000 B.C. or a little earlier for some of the ‘wavy-line’/ ‘dotted wavy-line’ wares have still to be properly published: see J. Afr. Hist. XIII (1972), 3–4 (Nakuru);Google ScholarRobbins, L. H., Science, vol. 176 (1972), 362–3 (Lake Rudolf);CrossRefGoogle ScholarClark, J. Desmond, in Nyame Akuma (a Newsletter of African Archaeology, Calgary), no. 3 (1973), 63 (Middle Nile).Google Scholar

⁶ From this, and especially the stratified sequence of Ishango by Lake Edward in the western rift, de Heinzelin, J. (Scientific American, CCVI (1962), 106–11) offers an argument for a central African origin for the harpoons. In fact, the case for diffusion from the southeasterly end of the ‘aquatic arc’ is supported by the presumed dating of harpoons at both Ishango and from sites above Lake Rudolf (Robbins, bc. cit.) to 7,000 B.C. and earlier. But it is probably still too early to speculate usefully on where in the Middle African belt either harpoons or the pots were first made.Just as pottery appears to have been independently invented in other parts of the world, so apparently were harpoons at both earlier and later periods, for open-country hunting as well as aquatic subsistence.Google Scholar

⁷ For surveys and general descriptions of these archaeological finds see Monod, T.Mauny, R., in Clark, J. D. (ed.), Third Panafrican Congress on Prehistory (1957), 242–7;Google Scholarde, Heinzelin, loc. cit.;Google ScholarHuard, P. and Massip, J.-M., ‘Harpons en os et cÉramique à dÉcor en vague (wavy-line) au Sahara Tchadien’, Bull. Soc. PrÉhist. Francaise, LXI (1964), 105–23;Google ScholarCamps-Fabrer, H., Matière et art mobilier dans la prÉhistoire nord-Africaine et Saharienne (MÉm. C.R.A.P.E., V, Paris, 1966).It should be noted that some of the pottery and harpoons included in these surveys —and marked on Map 1 of the present article—clearly belong to the later phases of the ‘aquatic civilization’ rather than to the earlier phase around the seventh millennium B.C. In several cases the dating is unclear.Google Scholar

⁸ Arkell, A. J., Early Khartoum (Oxford, 1949);Google Scholar summarized in A History of the Sudan to 1821 (2nd ed., London, 1961), 24 f.Google Scholar For the extension of ‘dotted wavy-line’ pottery into the eastern and central Sahara, see his paper in Actes du IVe Congrès Panafricain de Prèhistoire (ed. Mortelmans, G. and Nenquin, J., Tervuren, 1962), II, 283–7;Google Scholar also Antiquity, XLVI (1972), 221–2.Google Scholar

⁹ E. g.Arkell, , History of the Sudan, 28, 31–2;Google ScholarSonia, Cole, The Prehistory of East Africa (2nd ed., 1963), 247 f;Google ScholarClark, J. Desmond, The Prehistory of Africa (1970), 171–2;Google ScholarClark, , J. Afr. Hist. V (1964), 178.Google Scholar

¹⁰ The interesting discussions generated by Murdock's, G. P. provocative book (Africa: its peoples and their culture history, New York, 1959)Google Scholar, with its speculations on the origins of African agriculture, have tackled Murdock almost entirely on his own terms of reference: e.g. Fage's, J. D.review article, J. Afr. Hist. II (1961), 299 f;Google Scholar and the conference proceedings in J. Afr. Hist. III (1962), no. 2.Google Scholar

¹¹ ‘A re-examination of the evidence for agricultural origins in the Nile valley’, Proceedings of the Prehistoric Society, XXXVII (1971), esp. 34, 39–40;Google Scholar in Gabel, C. and Bennett, N. R. (eds.), Reconstructing African Culture History (Boston, 1967), 9.Google Scholar

¹² For conceptual purposes there may be more virtue in using the term ‘neolithic’ for Stone Age communities with advanced technological features (ground stone, worked bone, pottery, etc.) rather than in reserving it for those with incontrovertible evidence of cultivated crops or domestic animals—in other words to prefer the normal Francophone to Anglophone definition (despite my contrary sentiments—or rather aversion to the word ‘neolithic’ in any sense—expressed in a paper in 1967 to the VIe Congrès Panafricain de Prèhistoire, ed. Hugot, H. J. (Chambèry, 1972), 88–90).Google Scholar The concept ‘aqualithic’ might prove useful, if only ephemerally, in considering not only the Middle African cultures described in this article, but similarly communities in other parts of the world at various periods in the last 12,000 years, whose livelihood and outlook on life revolved around lakes and rivers. There are hints of something approaching this around Lake Baikal in central Asia and in Japan, the latter with probably the earliest pottery in the world (Okladnikov, A. P., in Braidwood, R. J. and Willey, G. R., Courses toward Urban Life (VFPA, 32, 1962), 267–87;Google ScholarWatson, W., in Ucko, P. J., Tringham, R. and Dimbleby, G. W., Man, Settlement and Urbanism (London, 1972), 329–41).Google Scholar Certain facies of the Natufian of the Levant and more obviously the Maglemosian of northern Europe are clearly aquatic and close in age to the main stage of the Middle African culture. (See Clark, J. G. D., Prehistoric Europe: the economic basis (London, 1952), esp. 33–48;Google ScholarClark, G. and Piggott, S., Prehistoric Societies (London, 1965), 106–8, 144–53.)Google Scholar A more recent, but nonetheless interesting, North American example is described by Baumhoff, M. A. (‘Ecological determinants in aboriginal California populations’, University of California Publications in American Archaeology and Ethnology, XLIV, no. 2 (1963), 155–236)Google Scholar: ‘The outlook of the people of the Lower Klamath province was centered on rivers—they lived and traveled on rivers, got their livelihood from rivers and even conceived their cosmography in terms of rivers’ (p. 177). And for an Australian example: Allen, H., ‘The Bagundji of the Darling basin’, World Archaeology, V (1973), 309–22. Maritime and archipelagic communities may be something very different again.Google Scholar

¹³ Leakey, L. S. B., The Stone Age Cultures of Kenya Colony (Cambridge, 1931), ch. VII. I am taking here as the ‘classic Kenya Capsian’ phases (a) and (b) only of Leakey's ‘Upper Kenya Aurignacian’. The more poorly described phases (c) and (d)—the latter not represented in Gamble's Cave—are later, probably much later, and not relevant to this discussion.Google Scholar

¹⁴ J. Afr. Hist. XIII (1972), 3–4.Google Scholar

¹⁵ Oakley, K. P., Antiquaries J. XLI (1961), 86–7.CrossRefGoogle Scholar

¹⁶ Stone Age Cultures, 103–4; 120–1.Google Scholar

¹⁷ Lion Hill Cave: Ibid. 24–5, 247.

¹⁸ Whitworth, T., S. Afr. Arch. Bull. XX (1965), 75–8;CrossRefGoogle ScholarRobbins, , Science, vol. 176 (1972), 362 f.;CrossRefGoogle ScholarJacobs, A. H. and Soper, R. C., ‘Surface pottery from Eliye Springs’ (unpub. paper, Inst. of African Studies, University of Nairobi, 1972).Google Scholar

¹⁹ Leakey was unnecessarily apologetic in suggesting that this piece of pottery could have resulted from the burning, perhaps accidental, of a clay-smeared basket. Examination of the sherd—the original is in the National Museum in Nairobi, a cast in Cambridge— indicates that it is genuine pottery. But this technical observation does not detract from the more general one made above that the ‘wavy-line’/’dotted wavy-line’ tradition of ceramics was initially influenced, if not inspired, by basket forms.

²⁰ E.g., Cole, Prehistory of East Africa, 266 f., 337–8;Google ScholarClark, , Prehistory of Africa, 167.Google Scholar

²¹ Oschinsky, L., who later examined these skeletal materials, registered his doubts in Anthropologica, V (1963), 112–13.Google Scholar See also my Archaeology of the Western Highlands of Kenya (1973), 78–9.Google Scholar

²² Stone Age Cultures, 116–17.Google Scholar

²³ Differing levels of inference and imagination are required here. Baskets, as noted, are directly deducible from some of the pottery styles. People with baskets and cooking-pots would surely have been collecting vegetable foods. The water-lily, with its nutritious rhyzome, and certain other aquatic plants would have lent themselves to quite intensive exploitation (see Clark, , Proc. Prehistoric Soc. XXXVII (1971), 65–6).Google Scholar At Khartoum there is direct archaeological evidence of the fruit of the Celtis tree being collected and consumed (Arkell, , Early Khartoum, 108).Google Scholar Herbs and condiments would also have been appreciated in the cuisine. Other plants may have provided fish-poisons and medicines. Incontrovertible direct evidence of boats, poles or paddles remains to be found, although Robbins illustrates a ‘broken piece of worked hippo rib’ from a Lake Rudolf site, which was ‘perhaps part of a scoop or blade of a canoe paddle’: Azania, II (1967), 73Google Scholar, p1. XVI. Moreover, some sites were below cliffs and approachable only from the water. ‘Catfish Cave’ below Halfa in Nubia, for instance, with harpoons, mollusc-shells and fish-bones, must have been a boatmen's camp during the Nile flood season at times in the sixth millennium B.C. or earlier: Wendt, W. E., Postilla, no. 102 (1966), 3–24.Google Scholar It should prove possible to enlarge on the various fishing methods. The grooved sinkers, some of stone, others of baked clay (Early Khartoum, 68–9, 79–80, 107)Google Scholar, should not necessarily be interpreted as evidence of hook-and-line fishing on the Nile during the main phase, since fish-hooks have been recognized only with the later phases. Perhaps weighted nets should be imagined, baited by selected snail types (Ibid. 107), to supplement —or to facilitate—spearing and harpooning. The preferred methods would have depended on the season, the water and the species of fish.

²⁴ The responsible literature on this subject is notoriously difficult and commonly inconclusive, owing to the fragmentary nature and state of preservation of the skeletal evidence, as well as the difficulty of establishing objective criteria for description and definition. For summary see Clark, , Prehistory of Africa, 166.Google Scholar See also Chamla, M.-C., Les populations anciennes du Sahara et des rÉgions limitrophes (MÉm. C.R.A.P.E. IX, Paris, 1968);Google Scholar and de Heinzelin, , Scientific American, CCVI (1962), 111–16. Skeletal remains have also been found at one or more of the sites above Lake Rudolf. It remains to be seen how revealing these may prove.Google Scholar

²⁵ Greenberg, J. H., The Languages of Africa (Indiana and The Hague, 1963), final map.Google Scholar

²⁶ …if, of course, we can turn a blind eye to Kordofanian, or at least to Greenberg's arguments for linking it with Niger-Congo.

²⁷ Some anomalies must be admitted, however. Around Lake Chad fishing and boating seem to be monopolized not by the local Nilo-Saharan representatives, but by those with Chadic (Afroasiatic) languages, notably the Buduma. Would these be in part an old Nilo-Saharan population which was ‘Chadicized’ at some point? In East Africa the expansion of different groups of Nilotes during the Iron Age has exaggerated the extent of the Eastern Sudanic branch of Chari-Nile; however, there are traces here of older Chari-Nile-speaking populations, of both the Eastern Sudanic and Central Sudanic branches. Along the Nile cataract reaches, moreover, a prejudice against fish is reported among the Nubians (also Eastern Sudanic); and Trigger argues that River Nubian is anyway a relatively recent offshoot of Hill Nubian. But here again it is likely that older Chari-Nile languages (including Meroitic?) were supplanted: see Trigger, B. G., Kush, XII (1964), 188–94;Google ScholarJ. Afr. Hist. VII (1966), 19–25.Google Scholar

²⁸ Both Christopher, Wrigley (J. Afr. Hist. 1 (1960), 193;Google ScholarIII (1962), 271–2) and Alan Jacobs (unpub. Nairobi paper, 1972) have addressed themselves to the historical implications of the distribution of Greenberg's ‘Nilo-Saharan’ or at least of its central and easterly groupings. Jacobs, moreover, has considered possible archaeological correlations.Google Scholar

²⁹ The recognition at Khartoum of reedwork (? plaited) for matting or perhaps walling as well as wall-daub (Early Khartoum, 79, 108) is an especially encouraging clue for the development of the study of settlements in this aquatic civilization.Google Scholar

³⁰ See below. Moreover, at the western extremity of the Continent, in shell-middens of the Casamance delta, pottery with possible distant ‘wavy-line’/‘dotted wavy-line’ affinities was being made until the beginning of the Iron Age and may even have continued to influence the Iron Age wares: Linares de Sapir, O., W. Afr. J. Archaeology, 1 (1971), 23–54.Google Scholar

³¹ An Iron Age and presumably A.D. date has been claimed for some of the bone harpoons found near Lake Chad (Courtin, J., Bull. Soc. PrÉhist. Française, comptes rendus, 1965Google Scholar, LXX-LXXV; F. Treinen, Ibid. 1965, CCLXVII–VIII). But this reasoning appears to be based partly on indirect indications as well as on the problematical chronology of the ‘Sao cultures’. Dates before iron in the final millennium or so B.C. might be more reasonable: compare Lebeuf, J.-P. in Actes du Premier Colloque International d'ArchÉologie Africaine (Etudes et Documents Tchadiens, MÉm. I, Fort-Lamy, 1969), 234–41;Google Scholar and G. Connah (Ibid. 116–17) on Daima with bone harpoons in the first millennium B.C. but not in later levels. In South Africa, interestingly, there is a nineteenth-century record of Bushmen of the Upper Orange and Vaal rivers and their tributaries making harpoons from ‘long, sharp, barbed points of bone’: Stow, G. W., The Native Races of South Africa (London, 1905), 72.Google Scholar Moreover, carved wooden harpoon-heads for arrows, especially for hunting hyrax, are described for the Sandawe and ‘Kindiga’ (Hadza) of northern Tanzania: Lindblom, G., ‘African harpoon arrows’, Ethnos, IV (1939), 62–72;CrossRefGoogle ScholarReche, O., Zur Ethnographie des abflusslosen Gebietes Deutsch-Ostafrikas (Hamburg, 1914), 11. These are, conceivably, related typologically to bone examples.Google Scholar

³² One could cull ad infinitum from the ethnographic literature, both survey works and detailed local studies.Google Scholar However, the following are worth special notice: Hornell, J., Water Transport (Cambridge, 1946);Google ScholarMonod, T., L'Industrie des Pêches au Cameroun (Paris, 1928);Google ScholarFosbrooke, H. A., ‘Some aspects of the Kimwani fishing culture’, J. Royal Anthrop. Inst. LXIV (1934), 1–22 (on south-west Lake Victoria);Google ScholarWorthington, E. B., ‘Primitive craft of the central African lakes’, The Mariner's Mirror, XIX (1933), 146–63 (on East Africa);CrossRefGoogle ScholarRoscoe, J., The Baganda (London, 1911), 391–9;Google ScholarSmith, R., ‘The canoe in West African history’, J. Afr. Hist. XI (1970), 515–33;CrossRefGoogle ScholarHenze, P. B., Ethiopian Observer, XVI (1973), 89–96.Google Scholar

³³ As a qualification to—and perhaps partial explanation of—this charge of neglect, we might note the very considerable interest of culture-historians in boats of all kinds and in certain techniques and equipment used in fishing to support theories, responsible and romantic alike, of contacts between Africa and distant shores of the Indian Ocean (and even the Pacific).

³⁴ Arkell, A. J., Shaheinab (Oxford, 1953).Google Scholar

³⁵ See the works cited in note 7.

³⁶ An estuarine, if not actually maritime, reflection—with wooden and ivory harpoons—is attested from shell-mounds in the delta of the Senegal river: J. Afr. Hist. XIV (1973)Google Scholar, RavisÉ, A., Notes Africaines, no. 28 (1970), 97–102.Google Scholar Although the pottery at this site is said not to be ‘wavy-line’ or ‘dotted wavy-line’, sherds which may well belong to this tradition have been collected at Dakar: Silla, O. and Thilmans, G., Notes Africaines, no. 121 (1969), 15–18. See also note 30.Google Scholar

³⁷ For Daima and other discoveries around Lake Chad, see note 31. For Ntereso in Ghana: Carter, P. L. and Flight, C., Man, N.S. VII (1972), 277–82.CrossRefGoogle Scholar

³⁸ Leakey, M. D., ‘Report on the excavations at Hyrax Hill, 1937–8’, Transactions of the Royal Society of South Africa, XXX (1945), part 4.Google Scholar

³⁹ Leakey, L. S. B., Stone Age Cultures, 198–9, p1. XXII-III.Google Scholar

⁴⁰ Sutton, J. E. G., The Archaeology of the Western Highlands of Kenya (Memoir no. 3 of the British Institute in Eastern Africa, Nairobi, 1973), 79–80, 146, 149–50. The pottery from Hyrax Hill (Site I) and the ‘Gumban A’ collections together comprise my ‘Kenya Highlands Class B’.Google Scholar

⁴¹ Leakey, L. S. B., Stone Age Cultures, 32;Google ScholarLeakey, M. D., ‘Hyrax Hill’, 372.Google Scholar

⁴² Robbins, , Science, vol. 176 (1972), 364.Google Scholar

⁴³ Apparently conflicting with the interpretation and chronology suggested here is a date late in the first millennium B.C. for ‘Gumban A’ pottery from Seronera in northern Tanzania: Azania VII (1972), 198;Google ScholarNyame Akuma, IV (1974), 6.Google Scholar The answer may be that elements of this late aquatic tradition spanned a long period persisting into the times of pastoral dominance. Leakey, (Stone Age Cultures, 31) notes that some of his ‘Gumban A’ was found in the muds of Lake Nakuru at a level only about 45 metres above the present. This would probably be later than the occupation of the Hyrax Hill site, though unlikely to have been after 2,000 B.C.Google Scholar

⁴⁴ S. Afr. Archaeol. Bull. XIX (1964), 33.Google Scholar

⁴⁵ This surmise is based largely on environmental considerations and preliminary information on changes in lake-levels. Archaeologically, southern Ethiopia remains very poorly known. Culturally, the relevant volume of the Ethnographic Survey of Africa contains some possible clues: Ernesta, Cerulli, Peoples of South-West Ethiopia and its Borderland (International African Institute, 1956), esp. 18–19, 99–102.Google Scholar A new look at certain of Jensen's, A. E. ‘archaic peoples’ (Altvölker Süd-Äthiopiens, Stuttgart, 1959) might prove revealing.Google Scholar

⁴⁶ The Archaeology of the Western Highlands, 79–81;Google Scholar and in Ogot, B. A. and Kieran, J. A. (eds.), Zamani: a survey of East African history (2nd ed., Nairobi, 1974), 86–8.Google Scholar

⁴⁷ The Cushitic fish-taboo, covering much of Ethiopia and the Horn, Kenya and northern Tanzania, is a difficult culture-historical problem. It needs to be more satisfactorily defined and more precisely plotted geographically and ethnically. It cannot be explained away simply on environmental grounds. Nor can it be considered just an automatic response of pastoralists regardless of cultural tradition or linguistic affiliation. While there is not an exact coincidence between the taboo and the distribution of Cushitic languages (and of heavy Cushitic borrowing, most of the highlands and plains of Kenya falling now into this latter category), it looks impressive enough to assume a connexion with a formative period in Cushitic history.

⁴⁸ It appears to extend as far as the Eastern Sudanic-speaking Didinga in the south-east of the Sudan Republic: Cerulli, , South-West Ethiopia, 72–3. Avoidance of fish is found again among pastorally-inclined communities in western Uganda, Rwanda, Burundi and north-western Tanzania. Its origin here may be due to local factors in the upland grasslands rather than to a putative Cushitic connexion.Google Scholar

⁴⁹ This probably lies at the root of the success and expansion of the initial Niloticspeaking peoples (an outgrowth of Eastern Sudanic, itself a branch of Chari-Nile), some three thousand years ago.The Nuer, as described by Evans-Pritchard, E. E. (The Nuer, Oxford, 1940, esp. 18–21, 70–2, pls IX, X, XXII), illustrate this well, with their dependence on sorghum, their attachment to cattle, and their maintenance of fishing by various methods, including spear- harpoons.Google Scholar

⁵⁰ Kendall, R. L., ‘An ecological history of the Lake Victoria basin’, Ecological Monographs, XXXIX (1969), 121–76.CrossRefGoogle Scholar

⁵¹ Sutton, in Shinnie, P. L. (ed.), The African Iron Age (Oxford, 1971), 157–9;Google ScholarSoper, R. C., Azania, IV (1969), 40–1.Google Scholar

⁵² ‘The shell-mound branch of the Wilton culture’, notably on the eastern shores of Lake Victoria—Leakey, L. S. B., Stone Age Africa (Oxford, 1936), 69—Google Scholar may be very relevant, as indeed Soper notes (Azania, IV, 40).Google Scholar

⁵³ This possibility of herding seems to depend at present on the correct identification of certain potsherds and on the original stratigraphic position of cattle bones found in rock-shelters excavated on the northern side of the Kavirondo Gulf: Gabel, C., African Historical Studies, II (1969), 205–54CrossRefGoogle Scholar, esp. 246, figs. 20, 21. The case for some possible agriculture, or at least some form of intensive land-use, by Lake Victoria in the first millennium B.C. is more indirect, being based on evidence of forest-clearance: Kendall, ‘An ecological history’; Clark, , Prehistory of Africa, 205–6.Google Scholar

⁵⁴ ‘Patterns of Bantu and Central Sudanic settlement in central and southern Africa (ca. 1000 B.C.–A.D. 500)’, Transafrican J. History, III (1973), 1–72;Google Scholar following up article in J. Afr. Hist. IX (1968), 223–21.Google Scholar

⁵⁵ Roscoe, (The Baganda, 391–4)Google Scholar is highly fascinating: ‘A special pot was kept in every fishing-canoe, in which the herbs, given by the priest of Mukasa, were placed; they were smoked over the net, or were thrown into the lake, to cause the fish to enter the nets’ (p. 394). So is Fosbrooke, (J.R.A.I. LXIV, 3–4, 20)Google Scholar: ‘The spirits of all departed fishermen go to live with Mugasha… Late at night they can be heard singing their canoe-songs… These spirits use the “spirits” of canoes that have been lost in the lake, and fish for spirit fish⃜’ (p. 4). See also Cohen, D. W., J. Afr. Hist. IX (1968), 655.Google Scholar

⁵⁶ See J. Desmond Clark, ‘Africa south of the Sahara’, in Braidwood, and Wiley, , Courses toward Urban Life, 1–33;Google Scholar and in Proc. Prehistoric Soc. XXVII (1971), 71.Google Scholar

⁵⁷ This approach involves a rejection, implicit or explicit, of Murdock's complicated arguments for an independent development of agriculture in the Sudan belt as early as the fifth millennium B.C. Details of the evidence to date of the history of cultivated crops and domesticated animals in Africa, both indigenous and introduced, are condensed in Shaw, T., ‘Early agriculture in Africa’, J. Hist. Soc. Nigeria, VI (1972), 143–91.Google Scholar

⁵⁸ Thurstan, Shaw in J. Afr. Hist. XII (1971), 143–53 (here p. 151)Google Scholar, reviewing Grahame, Clark'sWorld Prehistory—a new outline (Cambridge, 1969), whose Chapter 8 sees Africa as a continent which ‘had already relapsed into provincialism during the Late Pleistocene’ (p. 181).Google Scholar