The chronology of the introduction of pastoralism to the Cape, South Africa.

Sealy, Judith ; Yates, Royden

A careful survey of reports of early sheep in southernmost Africa combines with new radiocarbon dates to revise our knowledge of early pastoralism in the Cape. The new chronology shows the keeping of domestic stock and the making of pottery are not simultaneous and intertwined but separate events in a more complex history.

Introduction

Early European visitors to the Cape met local pastoralist people whom they called 'Hottentots', and who called themselves 'Khoikhoi'.

These people, and especially their sheep and cattle, were of great interest to scurvy-afflicted sailors in search of fresh meat. As a result, the surviving historical records preserve a certain amount of information about 17th- and 18th-century pastoralist communities. Since the social fabric of these societies broke down rapidly after European colonists dispossessed the local inhabitants of their lands and livestock, the many unrecorded aspects of pastoralist life and the history of these communities are accessible only through archaeological investigation.

Sheep and cattle have no wild progenitors in southern Africa, and so must have been brought into the region, presumably from further north in Africa. Stock were kept, in the drier western parts of southern Africa, by pastoralist peoples who did not grow crops, and made only very limited use of metals. In the wetter eastern and northern areas, stock-keeping was practised mainly by grain agriculturists who also smelted iron.

Investigations into the origins of pastoralism at the Cape have focused on how and when this took place. As early as 1905, Stow suggested that pastoralists from East Africa moved southwards into the area that is now Zambia and Zimbabwe, then westwards into southern Angola, and then south again until they reached the southernmost tip of Africa. Stow drew mainly on surviving fragments of oral traditions. A similar route was suggested by Cooke (1965), arguing that the environment of these areas would have provided suitable grazing for domestic animals, whereas the adjacent semi-desert countryside was too arid to support large herds of stock. In addition, this route follows the distribution of rock paintings of sheep, known mostly from Zimbabwe and from the southwestern Cape. An alternative route, suggested by Elphick (1977) on the basis of linguistic data, postulates movement from northern Botswana southwards towards the Orange River, then a split into two streams, one of which moved westwards along the river, and the other continued south through the central Karoo. These proposals need not be mutually exclusive. Domestic animals may have been brought to the Cape more than once, via different routes. Whichever way they came, though, the earliest pastoralist sites should be in the north, with progressively younger remains as one moves further south.

Archaeological research has found a large number of sites with sheep bones in western Southern Africa post-dating 1600 b.p.

Cattle, rarer in archaeological deposits than sheep, may have been a somewhat later addition (Inskeep 1969; Klein 1986). The lack of archaeological evidence for large-scale cattle-keeping, as documented in the historical records, remains a puzzle. In this article, we evaluate the earliest evidence for sheep in a Stone Age context: material presently believed to date to 2000-1600 b.p. Pottery first appears in this part of the world c. 2000 b.p.; sheep and pottery are generally assumed to be parts of a pastoralist 'package'.

We suggest that this conventional interpretation may need to be reconsidered. The case study considered here is one example of a wider issue: the extent to which different components of a food producing way of life emerged in concert in various regions around the world.

Archaeological evidence for early stock

Early pastoralist residues have been claimed for several archaeological sites in Namibia. Klein (1986) tabulates radiocarbon determinations associated with stock at sites in western South Africa and Namibia, including results in excess of 1800 years b.p. at both Falls Rock Shelter and Snake Rock. A more recent account of these sites, however, indicates that none of the faunal material has been positively identified as sheep (Kinahan 1991). At Mirabib, in the Namib, a dung layer yielded a radiocarbon result of 1550|+ or -~50 b.p. (Pta-1535), with sheep hairs identified in the dung (Sandelowsky et al. 1979). At Geduld, horizons containing dung range from before 1980 b.p. to after 1790 b.p. Sheep bones have recently been identified in the most recent of these horizons (Jacobson 1987; A.B. Smith pers. comm.).

Little Witkrans, Equus Cave and Dikbosch Shelter 1, in the northern Cape, have yielded sheep bones from excavation units with determinations of 1830|+ or -~50 (Pta-3418), 2390|+ or -~55 (Pta-2452) and 3060|+ or -~60 (Pta-1065) b.p. respectively. In all three cases, the deposits were poorly stratified and the excavators removed arbitrary horizontal spits. The associations between the charcoal on which the dates were run, and the other materials in the grid squares are questionable, as the excavators themselves have cautioned (Humphreys 1974; Humphreys & Thackeray 1983).

Further south, at the site of /Ai tomas or Vaalhoek, one distal sheep phalange was recovered from the same grid square and stratigraphic level as a radiocarbon result, on burned bone, of 1980|+ or -~120 b.p. (Pta-5530) (Webley 1992a). At Spoegrivier, a sheep phalange was stratified below a determination of 1920|+ or -~40 b.p. (Pta-4745) (Webley 1992a; 1992b). In both sites, more recent levels contain many additional sheep bones.

At Elands Bay Cave, the affinities of some of the stratigraphic units are still under investigation. The available information, however, is that sheep first appear in Layer 5, at 1120|+ or -~85 b.p. (GaK-4335) (Parkington 1977; 1981; Klein & Cruz-Uribe 1987). The nearby Tortoise Cave has approximately two dozen fragments of sheep bone associated with results of c. 1600 b.p. (Klein & Cruz-Uribe 1987; Robey 1987). Additional work on the site and excavated material is needed before we will know how much deposit (and faunal material) may relate to underlying determinations of c. 1800 b.p.

The site of Kasteelberg A is the remains of a large pastoralist encampment, with numerous sheep remains (Klein & Cruz-Uribe 1989). There are two radiocarbon results of 1860|+ or -~60 (Pta-3711) and 1790|+ or -~40 (Pta-3461) b.p. from the very bottom of the sequence, but it is unclear what volume of deposit, and hence which faunal and other material, is associated with this early period. Other dates from the site are substantially younger (Smith 1987).

The best evidence for early sheep at the Cape comes from the site of Die Kelders, on the southern Cape coast (Schweitzer & Scott 1973; Schweitzer 1979). The Later Stone Age levels at Die Kelders fall into two groups: the lower deposits, Layers 12 to 7, are bracketed between two indistinguishable radiocarbon determinations, one for Layer 12 of 1960|+ or -~85 (GX-1688), and one for layer 7/9, of 1960|+ or -~95 b.p. (GX-1687). Above Layer 7, there are a series of younger deposits clustered around 1500 b.p. Sheep first appear at Die Kelders in Layer 10. In addition to sheep bones, Die Kelders has yielded numerous potsherds: 946 from Layer 12 alone.

The nearby shell midden of Hawston contained a single sheep bone. The original report (Avery 1975) stated that this bone was stratified between two radiocarbon results of 1860|+ or -~60 and 1900|+ or -~40 b.p. (Pta-834 and 835 respectively). A subsequent account (Avery 1976) makes it clear that the specimen came from a unit overlying the 1860 determination (confirmed by the excavator: Avery pers. comm.). This bone is younger than originally believed.

Sheep bones from Byneskranskop occur only in Layer 1, which ranges from the recent past to c. 3200 b.p. (Schweitzer & Wilson 1982). During the first season of excavation in 1974, Layer i was removed as a single unit. During the second, 1976 season, when grid squares O 29-30 and O and P 31-32 were excavated, three subdivisions of Layer 1 were recognized: from top to bottom, Adam, Clara and Eva. Two radio-carbon determinations of 255|+ or -~50 (Pta-1864) and 535|+ or -~50 (Pta-1866) were obtained from charcoal samples from the uppermost sub-unit. The middle sub-unit yielded a result of 1880|+ or -~50 (Pta-1865), and the lowermost sub-unit one of 3220|+ or -~45 (Pta-1631). The excavators comment that the lowermost level probably should have been included with the underlying Layer 2, which is dated to 3400|+ or -~55 b.p. (Pta-1569) (Schweitzer & Wilson 1982: 22). Sheep remains are recorded from the lower sub-units, and may be older than 1600 b.p.

Boomplaas A yielded a single sheep upper dentition from the BLD Member, for which there is a radiocarbon result of 1955|+ or -~65 b.p. (UW-336) (Klein 1978). The excavator believes, however, that this specimen was intrusive, and belongs with the overlying stratum, the DGL member (Deacon et al. 1978; Deacon 1979). DGL ranges from 1700|+ or -~55 (UW-338) to 1510|+ or -~75 (UW-307) b.p. This site, therefore, does not have sheep from levels unequivocally dated to earlier than 1600 b.p.

Is the picture reliable?

The accepted picture is based on radiocarbon dates, usually on charcoal, and apparent associations between the dated charcoal and faunal and other material in the same stratigraphic units. Early sheep have usually been indirectly dated. Even so, there are only a handful of sites at the Cape in which sheep bones seem reliably associated with radiocarbon determinations of 1600 b.p. or older: /Ai tomas, Spoegrivier, Kasteelberg, Die Kelders, possibly Byneskranskop. In each case, there are relatively few fragments from the lower levels, and examination reveals them to be small. All of these sites have younger overlying deposits in which sheep remains are much more numerous. Given the well-documented mobility of objects within archaeological deposits (e.g. Hughes & Lampert 1977; Villa 1982; Villa & Courtin 1983; Gifford-Gonzalez et al. 1985), one wonders whether the 'earliest' sheep might in fact be more recent bones which have become mixed with older deposits. Many of the sites are coastal, and include large quantities of marine shell, making the deposits coarse-grained and porous. Anomalously early dates for domesticated grain in North Africa have been shown to be the result of small objects moving downwards through the deposit (Wendorf et al. 1984).

site |is less than~1600 b.p. |is greater than~1600 b.p.

/Ai tomas (Area 3) 56 1 (2)
Spoegrivier 52 1 (3)
Die Kelders 692 2 (10)

TABLE 1. Numbers of sheep bones (NISP) in units older than 1600 b.p., compared
with younger deposits. For the older units, the first number is the number of
bones which re-examination for this project showed to be unquestionably sheep.
The number in brackets is that from the faunal report, which in all cases
includes some probable identifications.

Data for Die Kelders is from Klein (unpublished records), for /Ai tomas and
Spoegrivier from Webley (1992a; 1992b). This information is not currently
available for Kasteelberg A or Byneskranskop.

New dates

We assembled a series of bones which could reliably be identified as sheep, from excavation units older than 1600 b.p. This comprised one specimen from /Ai tomas, one from Spoegrivier, two from Kasteelberg, two from Die Kelders and one from Byneskranskop. Additional specimens are available from the same units at Kasteelberg and Byneskranskop. Each bone was measured and photographed, and then dated by the radio-carbon accelerator technique at the Research Laboratory for Archaeology and the History of Art at Oxford University.

The specimen from /Ai tomas unfortunately contained insufficient protein for a date. The results for the other bones are presented in TABLE 2.

The two specimens from Die Kelders Layer 7 (museum accession number SAM-AA 8725) yielded results of 1325|+ or -~60 and 1290|+ or -~60 b.p. respectively. Although the site report records sheep in Layer 10 (Schweitzer 1979), we did not relocate any other bones from the lower series of deposits which could confidently be identified as sheep.

These specimens were the only two candidates for dating. We conclude they are intrusive into the layer in which they were found, and must derive from younger overlying deposits, in which sheep are much more numerous. Examination of the Die Kelders stratigraphy shows that, at least in the section drawn here, the older series of deposits (Layers 12-7, c. 1960 b.p.) seem to have been truncated by the younger series (Layers 6-2, c. 1500 b.p.), perhaps affording an opportunity for small fragments of bone from the more recent occupation to have become mixed with Layer 7. Layer 12, thought to be the same age as Layer 7, has a larger faunal sample than the whole of the rest of the site: over 17,000 fragments of bone. Yet there are no sheep in Layer 12, although there are numerous potsherds (Schweitzer 1979: 203). If sheep really were present at the site at 1960 b.p., we would expect to find them in Layer 12, as well as in the much smaller faunal assemblages of the layers above.

The specimen from Spoegrivier is, at 2105|+ or -~65 b.p., the only one of our series significantly TABULAR DATA OMITTED older than 1600 b.p. We conclude that this bone was found in situ, stratified below a hearth which yielded a result of 1920|+ or -~40 b.p. (Pta-4745). Unfortunately, it is the only specimen from this level. Site reports indicate that two overlying layers (Shelly Patch & Brown) also yielded sheep bones, a rib and a distal sesamoid respectively, neither a very diagnostic element. It is unwise to rely upon them for dating purposes (Klein pers. comm.). The many sheep bones higher up in the sequence are likely to be much younger (Webley 1992a; 1992b).

The sheep mandibular condyle from Byneskranskop gave a result of 1370|+ or -~60 b.p. This specimen comes from the middle sub-unit of Layer 1, which also yielded a radiocarbon determination, on charcoal, of 1880|+ or -~50 (Pta-1865). The excavators' lumping of the three sub-units of Layer 1 into a single entity in their report suggests some uncertainty as to their stratigraphic integrity. The relatively recent date for this specimen is not surprising.

The two specimens from Kasteelberg come from the same grid square and spit as a charcoal sample which gave a result of 1860|+ or -~60 b.p. (Pta- 3711) (Smith 1987). At two standard deviations, this determination just overlaps with the older of the two sheep results (1630|+ or -~60). The other (1430|+ or -~55 b.p.) is clearly younger. A conventional radiocarbon determination on a composite sample comprising a number of bone fragments, also from the same square and spit, yielded a value of 1230|+ or -~40 b.p. (Pta-5937). The site is an open one, in the lee of some large granite boulders. The level under discussion forms the very bottom of the sequence of archaeological deposits, directly overlying sterile gravelly sediments. There may be a very slight archaeological presence, too slight to have been detectable as a separate stratigraphic unit, represented by the 1860 b.p. result. Most of the material in the unit, however, is probably rather younger. The direct dates on the sheep bones are consistent with the general pattern mentioned above, with firm evidence for a substantial pastoralist presence only from 1600 b.p.

Other sites have yielded apparently early sheep remains which direct dating subsequently demonstrated to be out of position in the sequence. At Nelson Bay Cave, a sheep bone was recovered from a layer with a radiocarbon determination, on charcoal, of 1930|+ or -~60 b.p. (GrN-5703). Direct radiocarbon accelerator measurement of the bone itself produced a re-suit of only 1100|+ or -~80 b.p. (OxA-873) (Gowlett et al. 1987; Inskeep 1987). Striped Giraffe Shelter, in Namibia, yielded a sheep horn core from Spit 3 (Plug 1979). This horn was stratified between spits for which there are radiocarbon determinations of 4590|+ or -~100 (S.R-63) and 3080|+ or -~100 b.p. (SR-64) respectively (Sandelowsky & Viereck 1969; Vogel & Visser 1981). The horn core itself, however, is only 370|+ or -~40 b.p. (Pta-2230) (Vogel & Visser 1981). At Wonderwerk Cave, in the northern Cape, sheep hair was recovered from several units, extending as far down as the lowest sub-unit of layer 3a (Humphreys & Thackeray 1983). The top of layer 3a gave a result of 1890|+ or -~50 b.p. (Pta-2542), but the hair itself was only 460|+ or -~120 b.p. (OxA-622) (Gowlett et al. 1987).

Should we revise the chronology?

FIGURE 4 summarizes our picture of the introduction of stock to the Cape before and after direct accelerator dating. The new radiocarbon dates indicate that we have no direct evidence for stock-keeping in the southernmost Cape before about 1600 b.p. We have only a few direct determinations for sheep, of which just one is c. 2000 b.p.

Obviously, interpretations of this small sample must be extremely cautious. On current evidence, the earliest reliable determination for sheep in southwestern Africa is 2105|+ or -~65 b.p., from Spoegrivier on the northwestern Cape coast. This result is consistent with a west coast route for the introduction of sheep to the region.

Formerly, the most secure evidence for early sheep seemed to be that from Die Kelders, which is also the most southerly site. Archaeological surveys in areas to the north, along the supposed routes of introduction, failed to produce significantly earlier evidence of pastoralism (Webley 1992a; Sampson et al. 1989). It is entirely possible that stock-keeping spread so fast that we cannot resolve its expansion through radiocarbon dating, as seems the case with the spread of Matola-type ceramics associated with Iron Age farmers down the eastern coast of Africa at more or less the same time (Maggs & Whitelaw 1991). If, however, pastoralism in the southern Cape has a time depth of 1600 years, rather than 2000, while sheep were present in Namaqualand 500 radiocarbon years earlier, then we may be starting to detect a southerly progression of stock.

If stock-keeping in the southern regions of the Cape Province began only 1600 years ago, then the use of pottery and the keeping of domestic stock are not inextricably linked. We are fairly certain that pottery does date to c. 2000 b.p.: almost 1000 potsherds in Layer 12 at Die Kelders constitute a substantial body of evidence. Several authors have noted that early dates for pottery are more common than those for sheep (Sandelowsky et al. 1979; Deacon 1984; Klein 1986). Mazel (1992) has argued that, in the eastern parts of the country, use of pottery preceded the keeping of domestic stock. Our results strengthen this pattern considerably. Pottery is common at Die Kelders, and occurs at other sites, c. 2000 b.p. Domesticated animals, nowhere common until c. 1600 b.p., are from that time found in significant numbers in many different sites.

The new dates answer some rather specific questions about the sites from which they come. They raise many additional questions, and highlight the probability that the introduction of domesticated animals into southwestern Africa was a complicated process. The belief that both pottery and stock appeared in the Cape simultaneously has underlain the hypothesis that these items were brought into the region by an immigrant herder population. If there was, instead, a two-phase introduction, then this scenario is too simplistic.

An early date, c. 2000 b.p., for stock in western South Africa sets it prior to the arrival of the first iron-using farmers. A late date, c. 1600 b.p., makes a remarkable coincidence in the timing of the appearance of pastoralism in western South Africa, and agro-pastoralism in the east and north. We would like to know more about contacts between stone-tool using pastoralists in the western part of the sub-continent, and iron-using agriculturists who first settled in the eastern and northern parts of South Africa in about the 3rd century AD. Both groups made pottery, and kept domesticated sheep and cattle.

The pottery made by the western pastoralists includes so-called 'Cape Coastal' ware; thin-walled, relatively high-fired vessels of distinctive shape, often with spouts, lugs and pointed bases (Rudner 1968). Some pottery, however, is less distinctive. As Schweitzer (1979) remarked, at Die Kelders, the pottery from Layer 12 is different from the Cape Coastal ware in the overlying Layer 2. Layer 12 pots are very diverse: some sherds are thick, tempered with shale/siltstone inclusions, and were fired at low temperatures. Others are thinner and fired at higher temperatures. There are a number of bowls, and vessels with rounded, rather than pointed bases. Inskeep (1978) hypothesized that pottery might have been invented independently by sedentary hunter-gatherers; Deacon (1984) regarded this as unlikely. Schweitzer (1979) noted the similarity between Layer 12 pots and other globular pots from Oakhurst and Tsitsikama Cave, some 300 km to the east. He wrote (1979: 216):

The whole question of the relationship between the Later Stone Age Hottentot or 'Strandloper' pottery and that of the Iron Age peoples to the north is urgently in need of examination . . .

The Khoikhoi encountered by European colonists at the Cape kept fat-tailed sheep. Presumably, the ancestral sheep recovered from the archaeological sites were also fat-tailed, although there is no direct osteological reflection of this soft-tissue feature. Different breeds of cattle, on the other hand, can be separated osteologically. Historical pictures of Khoi cattle suggest mixed Sanga and Afrikaner genetic origins, and both kinds of animal are found in the faunal assemblages of the 9th century AD site of Schroda, and the 10th-11th century AD sites of K2 and Mapungubwe, all in the northern Transvaal (Voigt 1986). Cattle remains from archaeological sites in the western part of the country are, thus far, too fragmentary to be assigned to particular breeds.

Clearly, much work remains to be done. We need more sites with large assemblages of fauna and pottery, and many more direct radiocarbon dates on the bones of domestic animals. The results reported here demonstrate this limitations of the archaeological record. We cannot trace the beginnings of pastoralism on the basis of dating by association. Direct radiocarbon dates on the objects of enquiry, however, can help us to unravel the complicated interactions associated with the adoption of food production in southern Africa.

Acknowledgements. We are indebted to Richard Klein for a great deal of help with the identification of specimens, and information from his unpublished records. Graham Avery and Mike Wilson helped us to obtain material for dating, and Mike Wilson also furnished detailed information on the stratigraphies of Die Kelders and Byneskranskop.

Andrew Smith and Lita Webley supplied specimens and unpublished information from their excavations. James Brink identified the bones from /Ai tomas. Gavin Evans photographed the specimens before they were submitted for dating. Ray Inskeep began a programme of direct radiocarbon dating of sheep bones some years ago, and we have benefitted from his results, and discussion and advice. Leon Jacobson brought to our attention the date for the horn core from Striped Giraffe Shelter. Francis Thackeray supplied information on the sheep hair from Wonderwerk. The dates were funded by a grant from the Swan Fund, and run by the Radiocarbon Accelerator Unit at Oxford University. Many members of the South African archaeological community, and especially John Parkington, discussed the project design and the results with us, providing many useful insights.

Janette Deacon, Ray Inskeep, John Parkington and anonymous referees read and commented on the manuscript. We thank them all.

References

AVERY, G. 1975. Discussion on the age and use of tidal fish-traps (viswywers), South African Archaeological Bulletin 30: 105-13.

1976. A systematic investigation of open-station shell midden sites along the southwestern Cape coast. Unpublished M.A. thesis, University of Cape Town.

COOKE, C.K. 1965. Evidence of human migrations from the rock art of southern Rhodesia, Africa 35: 263-85.

DEACON, H.J. 1979. Excavations at Boomplaas Cave: a sequence through the Upper Pleistocene and Holocene in South Africa, World Archaeology 10: 241-57.

DEACON, H.J., J. DEACON, M. BROOKER & M.L. WILSON. 1978. The evidence for herding at Boomplaas Cave in the southern Cape, South Africa, South African Archaeological Bulletin 33: 39-65.

DEACON, J. 1984. Later Stone Age people and their descendants in southern Africa, in R.G. Klein (ed.), Southern African prehistory and palaeoenvironments: 221-328. Rotterdam: A.A. Balkema.

ELPHICK, R. 1977. Kraal and Castle: Khoikhoi and the founding of white South Africa. New Haven (CT): Yale University Press.

GIFFORD-GONZALEZ, D.P., D.B. DAMROSCH, D.R. DAMROSCH, J. PRYOR & R.L. THUNEN. 1985. The third dimension in site structure: an experiment in trampling and vertical dispersal, American Antiquity 50: 803-18.

GOWLETT, J.A.J., R.E.M. HEDGES, I.A. LAW & C. PERRY. 1987. Radiocarbon dates from the Oxford AMS system: Archaeometry Datelist 5, Archaeometry 29: 125-55.

HUGHES, P.J. & R.J. LAMPERT. 1977. Occupational disturbance and types of archaeological deposits, Journal of Archaeological Science 4: 135-40.

HUMPHREYS, A.J.B. 1974. A preliminary report on test excavations at Dikbosch Shelter 1, Herbert District, northern Cape, South African Archaeological Bulletin 29: 115-119.

HUMPHREYS, A.J.B. & A.I. THACKERAY. 1983. Ghaap and Gariep: Later Stone Age studies in the northern Cape. Cape Town: South African Archaeological Society. Monograph series 2.

INSKEEP, R.R. 1969. The archaeological background, in M. Wilson & L. Thompson (ed.), The Oxford history of South Africa 1: South Africa to 1870: 1-39. Oxford: Oxford University Press.

1978. The peopling of Southern Africa. Cape Town: David Philip.

1987. Nelson Bay Cave, Cape Province, South Africa: the Holocene levels. Oxford: British Archaeological Reports. International series 357(i) & (ii).

JACOBSON, L. 1987. The size variability of ostrich eggshell beads from central Namibia and its relevance as a stylistic and temporal marker, South African Archaeological Bulletin 42: 55-8.

KINAHAN, J. 1991. Pastoral nomads of the Central Namib Desert. Windhoek: New Namibia Books for Namibia Archaeological Trust.

KLEIN, R.G. 1978. A preliminary report on the larger mammals from the Boomplaas Stone Age Cave site, Congo Valley, Oudtshoorn district, South Africa, South African Archaeological Bulletin 33: 66-75.

1986. The prehistory of Stone Age herders in the Cape Province of South Africa, South African Archaeological Society Goodwin Series 5: 5-12.

KLEIN, R.G. & K. CRUZ-URIBE. 1987. Large mammal and tortoise bones from Elands Bay Cave and nearby sites, western Cape Province, South Africa, in Parkington & Hall (1987): 132-83.

1989. Faunal evidence for prehistoric herder-forager activities at Kasteelberg, western Cape Province, South Africa, South African Archaeological Bulletin 44: 82-97.

MAGGS, T. & G. WHITELAW. 1991. A review of recent archaeological research on food-producing communities in southern Africa, Journal of African History 32: 3-24.

MAZEL, A. 1979. Stow: the introduction of pastoralism to southern Africa, South African Archaeological Bulletin 34: 84.

1992. Early pottery from the eastern part of southern Africa, South African Archaeological Bulletin 47: 3-7.

PARKINGTON, J.E. 1977. Follow the San: an analysis of seasonality in the prehistory of the western Cape, South Africa. Unpublished Ph.D dissertation, University of Cambridge.

1981. The effects of environmental change on the scheduling of visits to the Elands Bay Cave, Cape Province, SA, in I. Hodder, G. Isaac & N. Hammond (ed.), Pattern of the past: studies in honour of David Clarke: 341-59. Cambridge: Cambridge University Press.

PARKINGTON J. & M. HALL (ed.). 1987. Papers in the prehistory of the Western Cape, South Africa. Oxford: British Archaeological Reports. International series 332(ii).

PLUG, I. 1979. Striped Giraffe Shelter faunal report, Cimbebasia B 3: 71-2.

ROBEY, T.S. 1987. The stratigraphic and cultural sequence at Tortoise Cave, Verlorenvlei, in Parkington & Hall (1987): 294-325.

RUDNER, J. 1968. Strandloper pottery from South & South West Africa, Annals of the South African Museum 49: 441-663.

SAMPSON, C.G., T.J.G. HART, D.L. WALLSMITH & J.D. BLAGG. 1989. The ceramic sequence in the Upper Seacow Valley: problems and implications, South African Archaeological Bulletin 44: 3-16.

SANDELOWSKY, B.H. & A. VIERECK. 1969. Supplementary report on the archaeological expedition of 1962 to the Erongo Mountains of South West Africa, Cimbebasia B 1: 4-43.

SANDELOWSKY, B.H., J.H. VAN ROOYEN & J.C. VOGEL. 1979. Early evidence for herders in the Namib, South African Archaeological Bulletin 34: 50-1.

SCHWEITZER, F.R. 1979. Excavations at Die Kelders, Cape Province, South Africa: the Holocene deposits, Annals of the South African Museum 78: 101-232.

SCHWEITZER, F.R. & K.J. SCOTT. 1973. Early occurrence of domestic sheep in sub-Saharan Africa, Nature 241: 547.

SCHWEITZER, F.R. & M.L. WILSON. 1982. Byneskranskop 1: a Late Quaternary living site in the southern Cape Province, South Africa, Annals of the South African Museum 88: 1-203.

SMITH, A.B. 1987. The economics of prehistoric herding at the Cape. Unpublished final project report to the Human Sciences Research Council of South Africa.

STOW, G.W. 1905. The native races of South Africa. London: Swan Sonnenschein.

STUIVER, M. & G.W. PEARSON. 1993. High-precision bidecadal calibration of the radiocarbon time scale, AD 1950-500 BC and 2500-6000 BC, Radiocarbon 35: 1-24.

TALMA, A.S. & VOGEL, J.C. 1993. A simplified approach to calibrating 14C dates, Radiocarbon 35: 317-22.

VILLA, P. 1982. Conjoinable pieces and site formation processes, American Antiquity 47: 276-90.

VILLA, P. & J. COURTIN. 1983. The interpretation of stratified sites: a view from underground, Journal of Archaeological Science 10: 267-81.

VOGEL, J.C. & E. VISSER. 1981. Pretoria radiocarbon dates II, Radiocarbon 23: 43-80.

VOIGT, E,A. 1986. Iron Age herding: archaeological and ethnoarchaeological approaches to pastoral problems, South African Archaeological Society Goodwin Series 5: 13-21.

WEBLEY, L.E. 1992a. The history and archaeology of pastoralist and hunter-gatherer settlement in the northwestern Cape, South Africa. Unpublished Ph.D thesis, University of Cape Town.

1992b. Early evidence for sheep at Spoeg River Cave, Namaqualand, Southern African Field Archaeology 1: 3-13.

WENDORF, F., R. SCHILD, A.E. CLOSE, D.J. DONAHUE, A.J.T. JULL, T.H. ZABEL, H. WIECKOWSKA, M. KOBUSIEWICZ, B. ISSAWI & N. EL HADIDI. 1984. New radiocarbon dates on the cereals from Wadi Kubbaniya, Science 225: 645-6.