Bos indicus L. cattle breeds are generally believed to have first appeared in Sub-Saharan Africa between A.D. 700 and I500 and to have been steadily replacing Bos taurus L. breeds ever since (Payne i964).2 Recently excavated Bos skeletal material from Ngamuriak, a Neolithic site in southwestern Kenya, shows cranial characteristics distinctive of B. indicus. Since the site has been dated by both radiocarbon and cultural materials to ca. 2,ooo b.p., this suggests that these cattle breeds were present in East Africa at least i,500 years earlier than previously documented and that B. indicus breeds have a long history in Sub-Saharan Africa. Zooarchaeological evidence indicates that domestic cattle were present in North Africa by the middle of the 6th millennium b.c. and perhaps as early as the 7th millennium b.c. (Gautier i984, i987). On the basis of depictions of cattle on rocks, tombs, and pottery supported by small quantities of zooarchaeological data, it is thought that the earliest domestic cattle in Africa were humpless longhorn cattle, a breed of B. taurus, and that humpless shorthornL cattle occurred in North Africa somewhat later (Epstein I97I; Gautier I984, I987). Egyptian tomb paintings and records show that humped cattle, B. indicus breeds, were introduced into Egypt during the first half of the 2d millennium B.C., and it is thought that they were never present in large numbers in North Africa (Epstein I97I, Payne i964). These cattle are thought to have been imported from the Near East (Payne i964) or the south (Epstein I97i, Epstein and Mason i984). There is no direct evidence of the first introduction of humped cattle into the Horn of Africa. It has been suggested that this may not have occurred until the Ist millennium A.D. with the expansion of the Axumites; a bronze figurine of a humped cow found at Zeban Kutur, near Cohaito in Ethiopia, is dated to ca. A.D. 330 (Clark I976). In the Sudan humped cattle are not known until the I4th or isth century a.d., when humped cattle figurines first occur in the archaeological record (David i982). It is generally thought that the major influx of B. indicus breeds into East Africa came with the expansion of the Swahili-Arab coastal trade from about A.D. 700 (Epstein I97I, Payne i964), although the first evidence of humped cattle, bifid vertebral spines from the site of Lanet in the Central Rift Valley, is dated only to about I575 + ioo a.d. (Posnansky I957). Similar evidence for humped cattle from the North East Village, Hyrax Hill, is thought to date to about the same time (Leakey I945). The evidence of humped cattle at Ngamuriak, an open Elmenteitan Neolithic living site in the Lemek Valley of the Loita-Mara area of Kenya, is considerably earlier. The site has only one major occupation horizon. There are two radiocarbon dates on well-stratified charcoal samples from different areas of the site: GX-8534, I,940 ? 40 b.p., and GX-8533, 2,35 ? I40 b.p. (Libby halflife) (Marshall and Robertshaw i982). The pooled mean of these dates, 2,037.5 ? 99, when calibrated, falls between the 2d century B.C. and the Ist century A.D. (mean calibrated according to Stuiver and Reimer i986). These dates are consistent with those for Neolithic cultures from central Kenya, where there is a Neolithic sequence from the end of the 2d millennium b.c. (Ambrose i984). The bone preservation at Ngamuriak is good, and a faunal assemblage of some 62,000 specimens was studied. Ninety-eight percent of the identifiable fragments are attributable to cattle, sheep, and goat, with 2,404 sheep and goat and 2,228 cattle specimens identified. The identification of B. indicus at Ngamuriak is based upon the cranial morphology of Bos at the site and supported by postcranial measurements. Grigson (I976,i980), in a comparative study of the cranial anatomy of four species of Bos, notes that “flat orbital rims were found in the younger taurine cattle, but in almost all cases was replaced in the older age groups by a sharp rim. Thus the persistence of the flat rim in older animals could be used as a fairly good indicator for indicus skulls” (Grigson ig80:23; fig. i). This character was used in the identification of B. indicus at Ngamuriak. Orbital-rim fragments are fragile, and only four adult specimens were preserved. Two of these are flat and appear characteristic of B. indicus (fig. 2); two are fragmentary but seem intermediate in morphology between B. indicus and B. taurus. I have observed characteristics similar to the latter in modem African sanga cattle breeds, which are thought to be descendants of early B. indicus-B. taurus crossbreeds (Payne i964, Epstein I957, Manwell and Baker ig80b). One orbital-rim specimen was preserved in the faunal assemblage from the Neolithic site of Crescent Island Main (GtJ1 2) in the Central Rift Valley, which dates a little earlier than Ngamuriak, to the end of the sth century b.c. or, when calibrated, between the 8th and 4th centuries B.C. (GX-4586/A, 2,2i5 ? I2o b.p., and GX4586/B, 2,535 ? I40 b.p. [Onyango-Abuje I977], pooled mean 2,367 ? 87 b.p. calibrated according to Stuiver and i. I thank the Kenyan Government, Office of the President, and National Council of Science and Technology for research clearance and the National Museums of Kenya for facilitating this project. I am very grateful to Richard Meadow for his advice and to Diane Gifford-Gonzalez and Tom Pilgram, Stanley Ambrose, and several reviewers for their comments on versions of this paper. I am solely responsible for any errors. I thank Peter Robertshaw for collaboration in the southwestern Kenyan research and the British Institute in Eastern Africa for funding for the project. Finally, I wish to express my gratitude to Desmond Clark and to the memory of Glynn Isaac for introducing me to such topics. 2. Following convention and the work of Grigson (I976, i980) and Manwell and Baker (ig8oa, b; contra Epstein and Mason i984), I refer to domestic cattle by the Linnaean binomials of B. taurus for humpless cattle and B. indicus for humped cattle. They do, however, interbreed and produce fertile offspring.
