New excavation at Blombos Cave, in the southern Cape of South Africa, and new radiocarbon dates for its sequence further illuminate the chronology of pastoralism in southern Africa, and the relations between pottery-using and shepherding.
Sheep in southern Africa
Sheep, not indigenous to southern Africa, were first introduced by pastoral people before 2000 b.p. (Klein 1986; Webley 1992; Sealy & Yates 1994). A number of routes for the introduction of stock to southernmost Africa have been proposed. Stow (1905) and Cook (1965) favour a westerly route from Zambia and Zimbabwe, respectively, to Namibia and along the southwest coast to the Cape [ILLUSTRATION FOR FIGURE 1 OMITTED]. From linguistic evidence, Elphick (1977) traces the origins of southern African pastoral peoples to the 'Central Bush' speakers in northern Botswana, with a southern movement to the Orange River, thence a split westwards and southwards, with both 'groups' ultimately occupying the Cape littoral [ILLUSTRATION FOR FIGURE 1 OMITTED].
No one model is correct if there were multiple introductions along various routes at different times (Klein 1986; Sealy & Yates 1994). Whatever the route, clearly pastoralism was introduced to the southernmost Cape from the north. Chronologically, then, the suite of radiocarbon dates from southern Africa sites containing sheep should reflect this southward movement.
Until recently this appeared not to be the case: the oldest dates for sheep, obtained for the layers in which they occurred, centred around 2000 b.p. at both the southernmost sites and in the more northerly regions (cf. J. Deacon 1984: 276-7; Klein 1986: 6-7; Sealy & Yates 1994: 59-62). The oldest reliable dates (cf. Sealy & Yates 1994: 59-60) for sheep in the northwestern Cape are from/Ai tomas (Webley 1992a) and Spoegrivier (Webley 1992a; 1992b) dated at 1980[+ or -]120 b.p. and 1920[+ or -]40 b.p. respectively [ILLUSTRATION FOR FIGURE 2 OMITTED]. Sites in the southernmost Cape with early sheep dates fall within the same period, notably Die Kelders, 1960[+ or -]85 b.p. (Schweitzer 1974; 1979), Byneskranskop 1, 1880[+ or -]50 b.p. (Schweitzer & Wilson 1982) and Nelson Bay Cave, 1930[+ or -]60 b.p. (Inskeep 1987). Inskeep (1987: 258), raising doubts about the early date for sheep at Nelson Bay Cave, pointed to the unreliability of dating sheep bone by association with dated units. A sheep bone, recovered from a Nelson Bay Cave layer radiocarbon dated at 193060 b.p., was submitted for direct dating by AMS which dated the bone to 1100[+ or -]80 b.p. (Gowlett et al 1987; Inskeep 1987).
Recently the accuracy of the 'indirect' method of dating sheep was queried by Sealy & Yates (1994). They submitted sheep bones from dated layers at four sites in the Cape for direct AMS dating. With the exception of one site, Spoegrivier, the oldest direct date they obtained for sheep in the Cape is from Kasteelberg A, 1630[+ or -]60 b.p. Sealy & Yates (1994: 64) cautiously interpret this new evidence as 'consistent with a west coast route' and suppose 'we may be starting to detect a southerly progression of stock'.
As archaeological evidence has supported the simultaneous arrival of pottery and sheep at the Cape, these two elements have seemed a 'package' brought to the region by migrating herder groups (cf. Deacon 1984; Klein 1986). [TABULAR DATA FOR TABLE 1 OMITTED] At c. 2000 b.p. pottery is frequent in southern African sites (cf. Deacon 1984: 276-7; Klein 1986). A number of authors record that 'early dates for pottery are more common than those for sheep' (e.g. Deacon 1984; Klein 1986; Mazel 1992; Sealy & Yates 1994: 64). Taking account of the new 'direct' dates indicating the appearance of sheep in the southernmost Cape at c. 1600 b.p. and the well-recorded presence of pottery at c. 2000 b.p., Sealy & Yates (1994: 63), suggest that sheep and pottery may not have constituted a 'package' that arrived together. …