Middens and Moderns: Shellfishing and the Middle Stone Age of the Western Cape, South Africa
Total Page：16
File Type：pdf, Size：1020Kb
Human Origins Research South African Journal of Science 99, May/June 2003 243 often small shelters scoured out below the Middens and moderns: calcrete. Both hyenas and humans can easily enlarge these shelters and insert shellfishing and the Middle Stone themselves into the already ancient dune topography. My colleagues and I are con- Age of the Western Cape, vinced that at most of the locations we have found, the remains of such occupa- South Africa tions are substantially later than the origi- nal dune formation. Dating such sites has to be a cautious enterprise. The rewards, John Parkington* however, are great, because the shelly sands have resulted in excellent preserva- tion of bone, marine shell and ostrich egg- HIS PAPER DESCRIBES SOME EVIDENCE OF middens we can try to understand the shell. Tshellfish gathering from what are argu- technological, social, ecological and cog- What has emerged from field surveys ably among the earliest shell middens in nitive implications of this behaviour, and over the past decade (Fig. 1) is that there human history. What makes this evidence in- their evolutionary significance. The west are different kinds of faunal assemblages teresting for scientists involved in explaining human evolutionary events is the fact that it coast of the Western Cape of South Africa in the calcrete landscape, some of which may register a key moment in the emergence represents an important opportunity are the result of human food gathering of our species. I describe the sites, list some of because there are many shell middens the archaeological remains, and speculate on from Holocene contexts and an increas- the relationship between the evidence for ing number of Late Pleistocene examples. systematic shellfish gathering and the appear- We can compare shellfish gathering across ance of hominid fossils that almost all palaeo- time in the framework of a single, albeit anthropologists would call ‘modern’. I have evolving, coastal landscape. this word in inverted commas because I believe all our definitions of ‘modern behav- Pleistocene shell middens along this iour’, and perhaps even ‘modern humans’, are coastline3–7 are deeply buried until dis- self-serving and in need of substantial un- turbed by some or other natural or indus- packing. Cynically, modern behaviour is trial events. In the north, extensive defined as likely to be reflected in the kinds of mining and, further south, coastal resort archaeological remains (worked bone, some development have added to the trunca- or other complex subsistence activity, marked tions of shoreline erosion in exposing ochre, burial) that we have in hand. It may be better to ask a less loaded question such as such sites. The key landscape features are what is the history of one of these component the dune plumes that reflect inland sand behaviours, such as inter-tidal marine food transport by strong southerly winds, acquisition. The gathering of sessile molluscs sometimes for many tens of kilometres. is, at first sight, hardly complex, but its Because the sand supply is greatest along nutritional advantages and correlates in the the long sandy beaches, the location of archaeological record might be of consider- these plumes is predictably from the 1,2 able significance. southern ends of the log-spiral, half- heart-shaped beaches. After each episode Introduction of sand transport and deposition, there First, what is a shell midden and what followed a period of soil formation in might the appearance of such things which calcium carbonate was concen- mean? By shell midden we imply an trated in sub-surface calcrete horizons by archaeological site with foodwaste visibly pedogenesis. The cycle was then repeated dominated by the shells discarded by as more sand was deposited and more cal- shellfish gatherers. The fact that we can crete formed. The harder calcrete hori- locate a concentration of such debris zons formed localized shelves under means that people in the past chose to which lie unconsolidated sands. Particu- discard shells in specific places, rather larly along the coast, where erosion than scatter them generally across the causes truncations along the shore, the re- nearshore landscape. From this we learn sultant cliffs offer many shelves and over- that the earliest such sites are not neces- hangs where animals or people can find sarily evidence for the earliest shellfish shelter. consumption, rather for the earliest local- Fig. 1. The location of MSA, LSA and modern sam- In the Saldanha Bay region and further ized discard. Nevertheless, we might con- ples referred to in the text. Large black dots are MSA south, the underlying quartz porphyry shell middens, small open circles are modern shellfish sider this important. I will assume that the and diorite has been planed down to an observations, and small triangles are LSA excavated earliest shell middens do, indeed, mark a undulating surface just a few metres shell middens. Numbers are as follows: 1, Yster- significant change in shellfish gathering fontein; 2, Hoedjiespunt; 3, Sea Harvest; 4, Paternos- above present sea level near the current behaviour, albeit one that needs careful ter; 5, Paternoster; 6, Elands Bay; 7, Liebenberg’s shore. Where sands have draped across Bay; 8, Brand se Baai; 9, Swartlintjies; 10, scrutiny. Using observations from shell these undulations, and when cementa- Boegoeberg; 11, Elands Bay South; 12, Duiker Eiland; 13, Bekbaai; 14, Northwest Baai; 15, *Archaeology Department, University of Cape Town, tion has taken place between higher Private Bag, Rondebosch 7701, South Africa. Perlemoen Baai; 16, Hospital Point; 17, Elands Bay E-mail: [email protected] standing pillars of bedrock, there are Cave and Dunefield Midden. 244 South African Journal of Science 99, May/June 2003 Human Origins Research Table 1. Shellfish weights (in g) from MSA west coast shell middens. Site* Unit P. granatina P. granularis P. argenvillei C. meridionalis Whelks Barnacles Total YZFN L and R 487.7 10.4 721.9 593.8 1.5 0 1828 YZFN 8/9 Hard 184.8 0.1 8 476.5 1.5 0.4 679 YZFN CS8 93.3 4.3 56.1 134.7 0 0 290 YZFN CS7 70.1 8 5.3 221.5 0.6 0.6 309 YZFN CS2 67.3 0.8 180.3 171.8 0.1 0.1 426 YZFN CS 183.6 0 288.3 558.4 0 0 1033 BSB5 Surface 3305.3 110.2 873.8 1787.5 0 2 6152 HDP3 SS E 116.7 6.3 69 382.5 9.6 10.9 601 HDP3 SS I 1909.3 311 391.8 3123.6 33.9 48.3 5920 HDP3 SS F 580.2 193.2 7.6 1285.8 2.4 13.8 2141 HDP3 SS G 688.3 203.5 216.7 1638.9 4.8 33.2 2870 HDP3 SS H 1083 368.8 108 2092.6 20.2 28.3 3738 HDP3 SS D 74.5 0 84.9 241.2 0.5 24.5 426 *Samples from Ysterfontein (YZFN) and Hoedjiespunt (HDP3) come from different excavated levels. BSB5 is the acronym for a surface sample from Brand se Baai. and consumption, others the accumula- at the last high stand of the sea level good and illustrate some striking pattern- tions of hyenas or owls. This is reminis- before the Holocene, which could have ing. First, the MSA shellfish assemblages cent of the taphonomic complexities of been Marine Isotope Stage 3 or part of contain negligible amounts of barnacles, the australopithecine sites, though the Marine Isotope Stage 5. crayfish or whelks, which are all regular Cape stratigraphies seem more simple. and sometimes abundant components of Even more interestingly, the stone tool Discussion west coast LSA assemblages. In fact, assemblages from the human occupa- In the relatively small excavated or sur- evident from Table 1, though somewhat tions are Middle Stone Age (MSA) in face samples we have from Pleistocene variable from site to site, is the extremely character and unquestionably associated west coast shell middens, stone tools are low diversity of MSA shell middens, with marine shell collections. Along the few in number, except at HDP1 and which are dominated by two genera, south coast of the Western Cape a similar YZFN1. At all of the sites referred to here, Patella and Choromytilus. A survey of over association appears but in the form of however, the assemblages are undoubt- 200 weighed, counted and measured deep, stratified shell middens in rock edly MSA, but have none of the distinc- samples from Elands Bay Cave (numbers shelters and caves. At sites such as those at tive tool forms of the Howiesons Poort or available on request) shows no late or Klasies River mouth8,9 and (more cau- Still Bay assemblage types. Most flakes are middle Holocene samples with as low a tiously) Die Kelders,10 hominid remains rather larger than we would expect in a combined weight of whelks and barna- associated with shellfish and MSA arte- Later Stone Age (LSA) assemblage and cles as is general in the MSA samples. At facts are described as modern. This means some platforms are faceted as in MSA the very least, MSA shellfish samples are that this particular archive contains evi- ones. Radial cores are present, but irregu- among the least diverse known from the dence that relates directly to the behav- lar forms dominate. Silcrete is seemingly LSA, comparable to those of the terminal iour of the earliest members of our own the preferred raw material, though quartz Pleistocene. As far as I am aware, no cray- species, to the origins of systematic shell- is common too. Retouched pieces are fish mandibles have been recovered from fish gathering and is, therefore, of more mostly made from silcrete.