GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 279 APTARA(GRG QUARK)

12

Human Impact on Precolonial West Coast Marine Environments of

Antonieta Jerardino, George M. Branch, and Rene Navarro

ITH VERY EXTENSIVE, diverse, and ogy to modern society extends beyond its per- W productive coastlines, it is no surprise ceived status as a highly specialized and rather that South Africa offers a superb opportunity to esoteric field (Erlandson and Rick, this volume). understand how marine ecosystems function, As demonstrated below, we also propose that and the effects of people on these environments. this multidisciplinary dialogue benefits the way This prospect is heightened by the realization of coastal archaeological sites are studied, by bring- the tremendous time depth of human occupa- ing a more exhaustive and systematic approach tion, including the first emergence of modern to answering questions, and by highlighting humans in Africa and, by default, in the world their main trait as archives of both human and (Erlandson 2001; Marean et al. 2007). This environmental history. Seen in this perspective, endeavor is multidisciplinary by necessity: the protection, conservation, and management marine ecology and archaeology go hand in of coastal archaeological sites must be a priority hand, along with other closely related specialities as promoted by the current South African her- such as oceanography, geology, and palaeo- itage legislation (National Heritage Resources environmental studies. While ecological studies Act, No. 25, 1999). can provide an understanding of the ecology of In this chapter, we explore the relationship present species, their vulnerability to harvest- that people established with the marine envi- ing, and the way they are influenced by physical ronment on the South African West Coast over factors, archaeology and palaeoenvironmental several millennia. Such effects have frequently studies offer a window into the past where such been demonstrated for terrestrial environments ecological relationships and physical variables worldwide (Redman 1999), but few convincing can be seen changing through time. The result cases have been made for marine systems. A of this multidisciplinary dialogue not only feeds broad account of the long-term precolonial much needed academic debate but also brings human settlement of the West Coast of South new and important dimensions into marine Africa sets the background for a more focused conservation policies and fisheries manage- analysis of selected archaeological sequences in ment. Because of this, the relevance of archaeol- the Lamberts Bay area. An integrated approach

279 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 280 APTARA(GRG QUARK)

FIGURE 12.1. Map of the study area showing location of sites and places mentioned in the text.

that combines sampling of faunal remains employed to answer questions regarding from archaeological sites, current knowledge of human impacts on marine harvested species: the ecology of the species exploited in the past, not all changes in populations are necessarily and detailed palaeoenvironmental reconstruc- related to the same causes. Consequently, all tions is employed as a powerful tool for unrav- possible lines of evidence and hypotheses eling long-term variability in marine ecosys- deserve serious and equal attention when tack- tems and their responses to human intervention. ling these types of questions. A case is made for human harvesting having been responsible for the local depletion of black PREHISTORIC EXPLOITATION OF mussels and limpets stocks, as reflected by MARINE INVERTEBRATES ON THE WEST shrinking mean sizes at a time when human COAST OF SOUTH AFRICA population densities were peaking and when reliance on marine resources was greatest. A Some of the earliest evidence for the exploita- concurrent decline in rock lobster size for the tion of marine resources in the world is found Lamberts Bay area (Figure 12.1) is interpreted in South Africa, mostly from deeply stratified differently, given the likely resilience of this sequences along the South Coast dating to species to subsistence harvesting and its sus- between 165,000 and 120,000 years BP (Jacobs ceptibility to environmental changes. This adds et al. 2003a, 2003b; Marean et al. 2007; Thack- a note of caution in terms of the methodology eray 1988; Vogel 2001). A few Middle Stone

280 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 281 APTARA(GRG QUARK)

Age (MSA) West Coast sites confirm that sub- period. Moreover, there are many factors other sistence harvesting is very ancient, including than exploitation that influence the size compo- reasonably well developed shell middens near sition of marine invertebrates, including differ- the towns of and Saldanha dated to ences or changes in environmental conditions about 70,000 years ago (Klein et al. 2004; Vol- and the intrinsic properties of the species them- man 1978). It is possible that marine resources selves. As shown in this chapter, zonation, were collected before this time too, but many of exposure to wave action, aquatic productivity, these sites are likely to have been washed out and turbidity can profoundly affect the size due to a high sea-level stand before the last composition of limpets and black mussels. It interglacial period around 127,000 years ago also remains to be explained why black mussel (Klein et al. 2004; Parkington 2003). sizes remained relatively constant through time The number of marine species present at despite the facts that (1) the black mussel is the these early MSA sites is similar to that observed most abundant species in MSA and LSA sites, in younger and nearby Later Stone Age (LSA) and (2) rocky-shore mussels are susceptible to sites. Only a few species such as rock lobster overexploitation as reflected by reductions in (Jasus lalandii) and fish are either not present their mean sizes caused by modern subsistence in MSA sites or their presence there cannot be harvesting (Lasiak 1992). Data on early shell- attributed to human agency with certainty. The fish gathering behavior in southern Africa is species diversity, however, appears to differ still preliminary. It needs to be recognized that between MSA and LSA sites (Klein et al. 2004). the number, and at times the size, of available Whether this reflects behavioral or environ- MSA shellfish samples is small when com- mental fluctuations is yet to be established, and pared to those analyzed from LSA sites resolution will require more systematic excava- (Jerardino 1993, 1997, 2007; Jerardino and tion of MSA sites and analysis of larger shell- Yates 1997; Tonner 2005). Thus, more variables fish samples. Another pattern emerging from need to be assessed when explaining the West Coast MSA sites is the significantly larger observed reductions in limpet sizes between mean sizes of at least three limpet species when MSA and LSA sites, and more; as well, larger compared to those of LSA sites (Halkett et al. samples of MSA shellfish are also necessary to 2003; Parkington 2003; Steele and Klein 2005). support any interpretation. No changes in the mean sizes of black mussels A considerable time gap separates coastal (Choromytilus meridionalis), however, are occupation of MSA sites and those of earliest observed in previous comparisons of MSA and LSA age, probably because most, if not all, for- LSA sites. Less-intense harvesting of limpets mer coastal sites dating to this gap were drowned during the MSA due to lower human popula- by rising sea levels, and only a few of these may tion levels at that time, and greater ecological have survived on the Atlantic continental shelf resilience of black mussels have been invoked (Miller 1990; van Andel 1989). Given the to explain these patterns (Klein et al. 2004; focused research efforts in the and Parkington 2003; Steele and Klein 2005). Lamberts Bay areas for the last 30 years (Figure Although this scenario seems plausible, it is 12.1), it is not surprising that the earliest LSA also important to consider other explanations, radiocarbon-dated evidence for the exploitation such as marked differences in the residential of marine invertebrates along the West Coast of permanence of human groups with similar South Africa comes from several caves and shel- population levels during both MSA and LSA ters at these adjacent locations. Shell midden times. In other words, low exploitation pressure horizons appear for the first time in Elands Bay on MSA limpets could well have resulted from Cave and Tortoise Cave around 12,000 and shorter visits by people to the coast, while rela- 7700 BP (all dates presented in this chapter are tively longer visits characterized the latter uncalibrated), respectively (Parkington 1981;

PRECOLONIAL IMPACT IN SOUTH AFRICA 281 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 282 APTARA(GRG QUARK)

Robey 1987), and well-developed shell lenses vesting along the West Coast of South Africa to dating to ca. 8400 BP have been excavated from have a more solid empirical foundation. Some Steenbokfontein Cave (Jerardino and Swanepoel of these subregions have been sampled and 1999) (Figure 12.1). Bedrock at this latter site is studied more extensively than others, with the far below these deposits, and further excava- Lamberts Bay and Elands Bay area (hereafter tions are likely to uncover older shell lenses. referred to as “the study area”) offering more Early Holocene evidence for shellfish collec- numerous and chronologically deeper se- tion might also be present in Spoegrivier Cave, quences. This chapter presents a case study located about 200 km north of Lamberts Bay, from this particular subregion and attempts to although no radiocarbon dates are yet available ascertain the nature of the interaction between for these basal deposits (Webley 2002). Once humans and their marine environment during sea level transgressed to within a few kilome- the Late Holocene. Multiple lines of evidence ters of these sites during the Early Holocene, a are used to achieve this goal, including data on wide range of marine invertebrates (and sev- population levels and dietary mix, palaeoenvi- eral vertebrate species) were exploited, includ- ronmental reconstructions, the biology and ing limpets, mussels, whelks, winkles, chitons, ecology of rocky shore marine invertebrates, and rock lobsters. All of these species would current understanding of the effect of modern have been collected from rocky shores, al- harvesting on similar species along other South though sandy beaches appear to be exploited to African shorelines, and statistical analyses on a small but significant extent, which was not metrical observations of body size for four repeated during subsequent occupations species of marine invertebrates recovered from throughout the Late Holocene (A.J., personal several archaeological sites. observation). After about 5000 BP, coastal shell middens SETTLEMENT AND SUBSISTENCE abound along the West Coast, reflecting a full PATTERNS AT LAMBERTS AND ELANDS BAY range of settlement and subsistence choices and new cultural and economic developments The chronological record of the study area (Buchanan 1988; Buchanan et al. 1978; Conard shows that relatively few sites were occupied et al. 1999; Jerardino 1996, 2007; Jerardino between 8,000 and 4,500 years ago (Figure and Yates 1996, 1997; Parkington et al. 1988; 12.2). Between 4,500 and 3,000 years ago, Robertshaw 1978, 1979; Sadr and Smith 1992; human settlement occurred in cave sites and Sealy et al. 2004; Smith et al. 1991). With few shelters, with volumes of deposit ranging exceptions, coastal archaeological sites accumu- mostly between 1 and 10 m3. Human impact on lated in close proximity to rocky shores from marine resources was probably negligible at this where much of the marine subsistence was time. Around 3500 BP, new sites were being obtained (Jerardino 2003). Although sketchy, occupied for the first time. Rates of deposition differences in the above social and economic started to increase along with the size of floor variables are apparent among subregions along areas at sites that had been previously occupied. the West Coast (e.g., Namaqualand, Elands Bay Longer residential permanence is inferred from and Lamberts Bay area, Penin- higher densities of artifacts that, according to sula). These differences seem to have been dic- ethnographic observations, were manufactured tated by environmental variability (rainfall, geo- and/or lost or discarded during longer visits to morphology, and availability of resources), sites (Jerardino 1995a, 1996; Jerardino and changing human population levels, and cul- Yates 1996). Volumes of deposits range between tural contact situations. Much work remains to 10 and 100 m3 per site around this time. These be done in each of these subregions for this are the first signs of population increase in the emerging and variable picture of coastal har- study area. Subsequently, between 3000 and

282 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 283 APTARA(GRG QUARK)

Number of middens

Volume of middens (1000 × m3)

% open sites

δ13 C

% exotic raw material Tortoise Cave Steenbokfontein

FIGURE 12.2. Intensity of site occupation in the study Sea surface temperature (°C) area (uncalibrated radiocarbon years BP).

2000 BP, settlement focused on very large open Turbidity (g sediment / Kg mussel) shell middens situated immediately behind rocky shorelines (Figures 12.2 and 12.3). Tons of marine shell and low densities of artifacts and terrestrial fauna characterize these enormous shell middens (Jerardino and Yates 1997). The term “megamiddens” was coined for these site FIGURE 12.3. Summary of settlement, diet, and palaeo- types when first observed 30 years ago (Parking- environmental changes in the study area (uncalibrated ton 1976). Volumes of deposit range from a radiocarbon years BP). hundred to several thousand cubic meters per site. A range of activities was performed at these sites, including the processing of vast to be as extended as before or even longer peri- quantities of shellfish, possibly for drying and ods of habitation were involved (Jerardino 1996, later consumption, stone knapping, and pro- 1998; Jerardino and Yates 1996). After 2000 cessing of terrestrial animal foods (Hen- BP, population densities declined sharply as shilwood et al. 1994; Jerardino and Yates 1997). reflected by the overall lower number of sites Dating of caves and shelters has also shown that and reduced volumes of deposits. Settlement only two such sites were occupied during the also returned to caves and shelters over the last megamidden period, namely Steenbokfontein 2,000 years (Figures 12.2 and 12.3), and short Cave and Pancho’s Kitchen Midden (Figure visits seem to have characterized this period 12.1). Either residential permanence continued (Jerardino 1996, 1998).

PRECOLONIAL IMPACT IN SOUTH AFRICA 283 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 284 APTARA(GRG QUARK)

Changes in hunter-gatherer mobility can be topic evidence based on carbon isotope meas- inferred from fluctuations in the frequency of urements on collagen and bone apatite from lithic raw materials from Steenbokfontein and human skeletons buried along the West Coast Tortoise caves, both with sequences that extend shows an increase in marine food intake to the Early Holocene. Although never domi- between 3000 and 2000 BP, the so-called nant, exotic raw materials such as silcrete and megamidden period (Figure 12.3). Much of the indurated shale (known locally as “hornfels”) protein and a significant portion of energy-rich were more commonly used between 8,000 and foodstuffs were obtained from marine re- 4,000 years ago with frequencies of 13 to 36 per- sources during this millennium. In contrast, cent (Jerardino 1996; new data) (Figure 12.3). the last 2,000 years saw a greater contribution Around 3500 BP, these materials were still used of terrestrial proteins and carbohydrates into regularly, but in a highly variable fashion, with people’s diets (Lee-Thorp et al. 1989; Sealy frequencies of 3 to 21 percent. During the subse- 1989; Sealy and Van der Merwe 1988). These quent megamidden period, exotic raw materials subsistence trends are closely mirrored by were rarely used at Steenbokfontein Cave, com- changes in the density of dietary remains that prising 6 to 8 percent of the flaked stone assem- have been preserved in archaeological sites for blage. Slightly smaller frequencies are recorded which we have sufficiently detailed observa- from contemporary occupations at Pancho’s tions. The dietary mix, as reconstructed from Kitchen Midden (Jerardino 1998). Exotic raw density ratios of marine and terrestrial re- materials continued to be used infrequently over sources from Steenbokfontein Cave and Pan- the last 2,000 years as shown by Tortoise Cave cho’s Kitchen Midden show an increase in and other sites dating to the last 1,000 years (Jer- marine foodstuffs during the megamidden ardino 1996, 2000; Orton 2006). From this, a period, particularly around 2,600 and 2,500 coherent picture of shifting settlement patterns years ago (Jerardino 1996, 1998; Jerardino and emerge for the study area. As the land became Yates 1996; new data). Thus, both archaeologi- more populated with groups settling for longer cal and isotopic evidence are in agreement, sug- periods around 3,500 years ago, and reaching a gesting that hunter-gatherer diet during the maximum between 3,000 to 2,000 years ago, megamidden period was more marine than hunter-gatherer mobility became increasingly ever before or after. Given the magnitude and restricted to the coastal margin. Contact with nature of these trends in population levels, set- inland areas and beyond the Oliphant River to tlement patterns, and subsistence, it is reason- the north and Berg River to the south was not able to suspect that people would have had their only rare during the megamidden period, but greatest impact on the local marine fauna dur- also after 2000 BP, when human occupation of ing the megamidden period. the study area waned. Lack of contact between the coast and the interior during the megamid- MEASURING HUMAN IMPACT den period is also supported by the dearth of ON MARINE INVERTEBRATES radiocarbon dates from hinterland sites (2 out of 42) falling between 3000 and 2000 BP Marine ecologists working with invertebrate (Jerardino 1996:87). Hence, the suggestion by species have several complementary ways of Henshilwood et al. (1994) that dried shellfish exploring hypotheses related to human impact would have been transported from the coast to on rocky shores (Lasiak 1992; Siegfried et al. the interior for their final consumption during 1994). These include (1) the quantification of this millennium is not supported by the available the volumes or mass of harvested resources and evidence. their recruitment levels through time; (2) Concomitant changes in subsistence are species composition of catches; and (3) compar- also reflected in the archaeological record. Iso- ison of richness, abundance, and size-frequency

284 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 285 APTARA(GRG QUARK)

distribution of species in areas where human elements of the biological community. Urchins harvesting is prevented, versus areas where har- harbour juvenile abalone, so any diminish- vesting takes place. These observations are eval- ment of urchins results in a decline in juvenile uated against an understanding of the influence abalone, with repercussions for the adult popu- of the physical environment on the collected lation. Moreover, consumption of grazers such species, and the biology and community struc- as winkles allows algae to proliferate (Day and ture of the species under study. Unfortunately, Branch 2002; Mayfield and Branch 2000; archaeologists have a narrower set of choices Mayfield et al. 2000). As a result, reductions in when looking for ways to answer similar ques- the abundance of rock lobsters have the capac- tions, but species composition of the catches ity to completely alter the nature of benthic and body-size observations can be retrieved communities. fairly directly from the archaeological record. Analyses of shellfish samples from Steen- Proxies can also be established for the physical bokfontein Cave, Malkoppan, and Grootrif D environment (e.g., degree of exposure of shore- megamiddens followed the methodology out- line, water turbidity, and sea surface tempera- lined by Jerardino (1997). Size observations on tures). Although archaeologists are able to eval- limpet shells were obtained by measuring the uate their observations through time as marine total lengths of unbroken shells. Body sizes of biologists do, the chronological control allowed black mussels and rock lobsters were derived, by radiocarbon dating does not match that avail- respectively, from measurements of prismatic able to researchers working in the present day. band widths and calcareous mandibles. These Nevertheless, archaeologists have managed over initial observations were then transformed to the last 15 years to extricate exciting observations body-size estimates with the use of morpho- reflecting the ability of people to exert a tangible metric equations (Buchanan 1985; Jerardino et impact on marine invertebrate species (Jer- al. 2001). The statistical significance of any ardino et al. 1992; Spennemann 1986; changes in the mean sizes of these four inver- Swadling 1976). tebrate species was tested using one-way analysis of variance (ANOVA; Zar 1984) fol- ARCHAEOLOGICAL OBSERVATIONS lowed by Tukey-Kramer multiple comparisons FROM THE STUDY AREA (Stoline 1981). Black mussels were the most heavily col- Three molluskan species, namely Choromytilus lected shellfish between 3,000 and 2,000 years meridionalis (the black mussel), Cymbula gra- ago, with relative frequencies of 70 to nearly natina (the granite limpet), and Scutellastra 100 percent of the weight at sampled megamid- granularis (the granular limpet), and one dens. Black mussels were also abundant in species of crustacean, J. lalandii (the West shellfish samples recovered from Steenbok- Coast rock lobster), are considered in this chap- fontein Cave, with frequencies covering 40 to ter. These species belong to three different 90 percent. On the other hand, limpets were trophic levels, and some of them have direct almost absent from most megamiddens during ecological links. Black mussels are filter feed- this millennium, although important excep- ers, the limpets are intertidal grazers, and rock tions are Malkoppan and Grootrif D megamid- lobsters are top predators with a powerful abil- dens (Figure 12.1). Limpets were also collected ity to modify the relative abundance of their from Steenbokfontein Cave before and during prey and that of other species associated with the megamidden period. Limpet frequencies them (Castilla et al. 1994). In particular, rock spanned 5 to 30 percent in both of these lobsters consume mussels, urchins, and win- megamiddens, and 2 to 38 percent in Steen- kles. Elimination or depletion of these groups bokfontein Cave. Of the two limpet species, by rock lobsters has powerful affects on other Cymbula granatina was collected more regularly

PRECOLONIAL IMPACT IN SOUTH AFRICA 285 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 286 APTARA(GRG QUARK)

than S. granularis, most probably because it pro- Choromytilus meridionalis duces higher meat yields per individual (Tonner (n = 770 - 2360) 2005). Rock lobster remains were found in almost all sites that have been systematically excavated (Horwitz 1979; Jerardino 1996, 2000; Jerardino and Navarro 2002). However, large enough samples of mandibles for valid statistical manipulation are present at only two sites where the megamidden period is repre- sented, namely Steenbokfontein Cave and Grootrif D. Cymbula granatina (n = 53 - 409) In the case of the black mussel, statistically significant changes can be detected through time, particularly reductions in mean shell length during the megamidden period, as illus- trated by data for Steenbokfontein Cave (Figure 12.4). Similar changes occurred at Malkoppan as well. Significant reductions in the mean sizes of the granite limpet and the rock lobster were also detected for the megamidden period Jasus lalandii at Steenbokfontein Cave and Grootrif D mega- (n = 42 - 237) midden (Figure 12.4). A decreasing trend in shell length during the megamidden period was also detected for S. granularis in Steenbok- fontein Cave, although this trend was not sta- tistically significant. No significant changes in the mean sizes of S. granularis were detected with Grootrif D data either. It is also interesting that the mean sizes of the three quantified mol- lusk species from Steenbokfontein Cave were smaller around 4000 BP, before the start of the FIGURE 12.4. Changes in mean sizes of black mussels megamidden period (see later). (Choromytilus meridionalis), granite limpet (Cymbula gra- natina), and rock lobster (Jasus lalandii) at Steenbokfontein Cave (uncalibrated radiocarbon years BP). FAUNAL CHANGES IN THE LIGHT OF ENVIRONMENTAL CONDITIONS areas relative to protected areas testifies to this There are convincing changes through time in (Branch 1975; Branch and Odendaal 2003; Kyle the size composition and the relative propor- et al. 1997; Lasiak 1992; Lasiak and Dye 1989). tions of species in middens, and the immediate Changes in modern biological community temptation is to ascribe this to human harvest- composition have also been recorded, with con- ing. Support for this view comes from declines vergence on a relatively uniform composition in the size of harvested species in middens else- in harvested areas (Hockey and Bosman 1994; where in the world (Jerardino et al. 1992). Addi- Lasiak and Field 1995). tional support comes from the documentation Interpretation of the causes of these modern of comparable modern faunal changes that can faunal changes benefits from concurrent obser- be linked to subsistence harvesting. Declines in vations at harvested and protected areas, thus the sizes of limpets and mussels in harvested eliminating or at least reducing possible

286 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 287 APTARA(GRG QUARK)

confounding factors that may cloud the influ- the elevated nutrients. This water is later re- ence of human harvesting. No such luxury is turned to the shore downstream of the upwell- possible when deducing the effects of human ing centers, and particulate food is often more harvesting from archaeological middens. A concentrated there than at the upwelling cen- range of factors other than human harvesting ters (Wieters et al. 2003). Correspondingly, may influence patterns in size composition, growth of mussels may be expected to be including temperatures, productivity, sea level, greater downstream than at the focal points of turbidity, wave action, storms, red tides, and the upwelling. inherent relative resilience of different species Could temporal differences in productivity to harvesting. Without at least a consideration have driven the patterns of diminishing sizes of the potential effects of these factors, it is recorded in middens over the period 2400– impossible to be confident that harvesting is the 3500 BP, rather than human harvesting? It factor that drives faunal changes in middens. seems unlikely. First, temperatures were declin- Regional differences in temperature serve ing over this period (Figure 12.3), indicating ele- as a useful proxy for productivity because sea vated rather than diminished productivity; so temperatures are inversely related to nutrient size should have increased rather than levels (Nielsen and Navarrete 2004). Nutrient decreased if nutrient-fueled productivity caused levels are in turn correlated with growth rates the reduction in size. Second, there was a cor- of primary producers, notably phytoplankton responding decrease in temperature during the and benthic algae, which underpin the growth Little Ice Age (ca. 500 BP), yet mussel sizes of secondary consumers such as mussels that increased during that period (Jerardino 1997), so feed on particulate matter and limpets that opposite responses were recorded during the consume algae. Growth rates are positively two periods of temperature decline. correlated with maximum sizes that species Sea level was at a maximum about 3800 BP, achieve. Thus, long-term shifts in tempera- dropped to approximately present levels between ture, or regional differences in temperature, 3300 and 2400 BP, and then rose again by could plausibly be linked to differences in the about 1 m before declining again (Jerardino sizes attained by mussels or limpets in mid- 1995b). There is no intrinsic reason why sea dens. The West Coast of South Africa as a level per se would have affected sizes of organ- whole is a region of intense upwelling, and isms in middens, but it is possible that associ- limpet biomass and sizes are larger there than ated factors were at work. One option is that on the South and East coasts, where upwelling turbidity altered with sea level and could have is infrequent or absent (Bustamante et al. influenced growth rates and size. Patellid 1995b). However, the West Coast also has focal limpets are vulnerable to sand cover (Marshall points of upwelling at particular sites (Shan- and McQuaid 1989) and are likely to achieve non 1985), and growth of mussels and maxi- smaller sizes in sand-inundated conditions. mum sizes of limpets have both been shown Mussels feed less successfully when particulate to be greater at upwelling centers than down- matter rises above threshold (Stuart et al. 1982) stream, where upwelling is less marked and therefore grow slower and reach smaller (Menge et al. 2003; Xavier et al. 2007). The sizes. An indirect measure of turbidity can be relationship between upwelling and growth derived from archaeological middens by quanti- (or size) is, however, not necessarily a positive fying a kind of sediment (water-worn shell and one. Upwelling does enhance nutrient levels, shingle) originally trapped by the byssal threads but it also translates into advection of surface of mussels and retained later in the middens waters, so the nutrient-rich upwelled waters (see Jerardino 1993: Figure 11.3). The data sug- are shifted offshore, taking with them any phy- gest that turbidity was high around 4000–3000 toplankton growth that has been spurred by BP and could explain why sizes of mussels and

PRECOLONIAL IMPACT IN SOUTH AFRICA 287 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 288 APTARA(GRG QUARK)

limpets were low at that time. This does not sword when it comes to mussel size. Storms explain the decreases in size in the subsequent may remove mussels en masse (Branch and period, when turbidity seemed to be at a mini- Steffani 2004; Steffani and Branch 2003) but mum. Again, this allows us to reject turbidity as sometimes dump huge quantities of large, sub- a cause of size changes between 3300 and 2400 tidal mussels in the intertidal zone where they BP (Figures 12.3 and 12.4). can take hold, increasing both the quantity and Wave action has profound effects on body sizes of mussels available to intertidal har- sizes, as movements of most limpets are inhib- vesters (G. M. Branch and S. Eekhout, unpub- ited by wave action, and they achieve smaller lished data). If the frequency or intensity of sizes on wave-exposed than sheltered shores storms varied systematically over time, they (e.g., Branch and Odendaal 2003). This is partic- could have influenced body sizes of both ularly true of both S. granularis and C. granatina, limpets and mussels. However, there is no evi- the two species of greatest interest in West dence of such systematic changes over the Coast middens. Conversely, mussels, which period when sizes of these animals were declin- favor wave-exposed shores because wave action ing. There is no easy prehistoric measure of suspends and replenishes greater concentra- storms, but turbidity serves as a proxy, and it tions of particulate food (Bustamante and shows no variations over time that clearly cor- Branch 1996a, 1996b), attain greater sizes, relate with limpet and mussel sizes. growth rates, and cover on exposed shores than Thus far, we have focused on possible envi- in sheltered areas (Branch and Steffani 2004; ronmental effects that could have influenced Steffani and Branch 2003). Two lines of thought body sizes of mussels and limpets, but similar argue against wave action as an explanation for questions can be raised about rock lobsters. To diminished sizes of limpets and mussels in what extent is the decline in rock lobster size middens. First, all middens were associated over the period 3500–2400 BP likely to reflect with relatively short outcrops of rocks where the environmental conditions rather than harvest- magnitude of wave action is unlikely to have ing? Two important possibilities exist. The first changed in any systematic manner that can be is that rock lobster growth has declined since related to sea-level changes over the period the late 1980s (Johnston and Butterworth when limpet and mussels sizes were declining. 2005; Pollock et al. 1997). The causes remain Theoretically, changes in sea level could have unresolved, but reduced food supply or envi- altered the coastal topography and thus affected ronmental changes have both been invoked as wave action, but rises and falls spanning 3 m explanations. Reductions in the rate of growth would probably have had too small an effect on will result in smaller size and lower productiv- local topography to bring about significant ity, so the question that immediately arises is changes in wave action. Second, and more con- whether any past variations in growth could vincingly, even if wave action was altered, we have influenced the productivity and size com- would have predicted opposite responses from position of rock lobsters found in shell mid- limpets and mussels, yet they both declined in dens. So far, there is no easy way of testing this size. In short, alterations of wave action seem possibility, although it seems that prehistoric an implausible cause of body-size declines in sizes of rock lobsters were often substantially mussels and limpets. greater than in modern times (A. Jerardino, Storms could have had more subtle effects unpublished data). Secondly, mass “walk-outs” on the size composition of mussels and limpets of rock lobsters have been recorded on the that would have been available for people to West Coast of South Africa, during which harvest. Periodic storms may eliminate large (mostly small) lobsters have moved inshore to limpets by physically removing them from avoid oxygen-depleted waters and have ended up rocks (Denny et al. 1985) but are a two-edged becoming stranded on the shore in spectacular

288 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 289 APTARA(GRG QUARK)

quantities—up to 2,000 tons in one episode shore and on wave-beaten shores, where harvest- (Cockcroft 2001). Such events appear to be trig- ing is more hazardous (Branch and Marsh 1978; gered by upwelling followed by prolonged qui- Bustamante et al. 1995a). Mussels also occur in escence, which concentrates phytoplankton the intertidal zone, but they extend down into the inshore in bays, leading to depletion of oxygen subtidal zone. In the midshore they are small but as the blooms decay. Ensuing walkouts will easy to collect, but at the bottom of the shore they influence the availability and size composition are larger although less accessible. In the subti- of rock lobsters in two ways. They provide a dal, they cannot be harvested without diving. brief bonanza, but they also deplete stocks. Rock lobsters are even less vulnerable to subsis- There is no way of telling if such walk-outs were tence harvesting because they live entirely subti- more or less frequent in the prehistoric past, dally and extend offshore for several kilometers, but if their frequency has changed it would into depths of about 80 m (Griffiths and Branch have powerfully affected the amounts and sizes 1997; Heydorn 1969). There is scant osteological of lobsters that could have been harvested. In evidence that prehistoric harvesting on the West the case of rock lobsters, we are therefore on Coast of South Africa involved diving (A. Morris, much less certain ground in attempting to personal communication, 2006). Fishers may relate declines of size in middens to harvesting have used simple gourds and twine but would pressure. It appears that at least two plausible, have been limited to very close to the shore. As but ambiguous, environmental factors may bet- far as is known, no watercraft or offshore fishing ter explain these changes in size. was developed in the region during prehistoric times. Rock lobsters could have been captured in CHARACTERISTICS OF THE SPECIES very shallow water only by wading or by luring individuals with bait. This would have meant that In evaluating the potential effects of human har- only a tiny proportion of the population would vesting on the abundance and sizes of target have been exposed to harvesting, making it species, consideration also needs to be given to highly unlikely that harvesting could have dented both the relative vulnerabilities of different the size composition of the species, in contrast to species to harvesting, and human preferences the impacts that may have been inflicted on inter- (Lasiak 1991). Not all species are equally vulner- tidal species. able to harvesting pressure, and their vulnera- Two key features emerge. First, the apparent bility depends on a suite of biological properties. absence of watercraft and means of fishing away from the shore would have curbed the capacity of harvesters to influence the popula- Accessibility tion structure of subtidal species. Second, Species that are confined to the intertidal zone, species that have spatial refugia where they can- such as C. granatina and S. granularis, are readily not be harvested will be relatively less vulnera- accessible to harvesters every time the tide ble to the effects of fishing. recedes sufficiently to expose them. During approximately five days every fortnight the entire Mobility shore is exposed during low spring tides, and it is then that human harvesting is most intense. Not Capacity for movement will influence the ability all intertidal limpets will, however, be equally vul- of animals to recover after being harvested. nerable. S. granularis occurs high on the shore Limpets are sedentary and mussels are sessile. and will be most accessible; C. granatina occurs Rock lobsters move inshore and offshore in regu- lower down but is most abundant on sheltered lar annual migrations (Heydorn 1969), however, bays, where it is also easy to harvest; but S. and can replenish their shallow-water numbers if argenvillei is characteristically found low on the harvesting takes place there alone. Again, this

PRECOLONIAL IMPACT IN SOUTH AFRICA 289 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 290 APTARA(GRG QUARK)

points to rock lobsters being relatively immune to closely related species, wide differences may exist the attentions of shore-based harvesters. in growth rate. The limpets Cymbula oculus and C. granatina grow fast, reaching maturity within 2 or 3 years; but other limpets such as Scutellastra Larval Dispersal argenvillei and S. cochlear grow agonizingly slowly, Both the mode and frequency of larval dispersal maturing after about 6 years and attaining ages of will also influence the vulnerability of species to up to 35 years (Branch 1974; Eekhout et al. 1992). harvest pressure. Mussels have a widely dis- Mussels tend to be fast growing, being har- persed planktonic larval stage but often experi- vestable after 1 to 2 years and living for about 5 ence intermittent recruitment. Years may pass years, but the rock lobster matures only after 7 to with little or no recruitment, interspersed with 15 years and lives up to 40 years, making it much bouts of intense recruitment. Moreover, settle- more prone to the effects of harvesting and slow ment takes place mainly into existing beds of to recover (Pollock et al. 2000). adult mussels (Harris et al. 1998). No replenish- ment can take place in years with no recruit- Sex Change ment, and even when settlement does take place, recovery is slow where adult beds have been Some species undergo sex change as they age. stripped by overharvesting (Dye et al. 1997), thus The result is that older age groups are dominated increasing the chances that harvesting will influ- by one sex. As there are inevitably fewer individ- ence population structure. Species that are most uals in these older age groups, this automatically vulnerable to overharvesting are those that have skews the sex ratio. None of the species of central very limited dispersal. A classic example is the interest here undergoes sex change, but C. oculus, solitary ascidian (Pyura stolonifera), colloquially a close relative of C. granatina, is male during its known as “red-bait,” which is harvested as a first 1 to 3 years of life and then becomes female source of food by subsistence fishers on the east for the rest of its life. Females are consequently coast of South Africa and used as bait for fishing not only more rare than males but are prime tar- elsewhere. It has a larval stage that lasts only gets for harvesters because of their larger size. minutes, so its larvae settle within meters of the On the east coast of South Africa, where this adults that produce them (Griffiths 1976). As a species is heavily fished by subsistence fishers, it result, depletion of adults reduces local settle- is threatened with extinction because of the ment of larvae, and replenishment by larvae pro- depletion of females, which have declined from duced afar is impossible. By contrast, the rock on average of 36 percent of the sexually mature lobster has a prolonged larval life lasting 9 to 11 population in protected areas down to 9 percent months and is widely dispersed (Silberbauer in harvested areas (Branch and Odendaal 2003). 1971), so replenishment is possible both by adult Clearly, sex change heightens the vulnerability of movements and by larval settlement. species to harvesting.

Growth Rate and Longevity HUMAN PREFERENCES

Species that are fast growing can recover quickly In addition to the characteristics of the species after being depleted, but they tend to be short- considered above, human preferences will influ- lived so that the size-composition of their popula- ence the relative impact of harvesting on differ- tions is made up of a small number of year-groups ent species. Particular species may be harvested and is inherently unstable from year to year. From more intensely than would be predicted based a management perspective, there is thus a trade- on their abundance. Factors affecting preference off: fast growth translates into rapid recovery but include accessibility, ease of procurement, trans- high variability in stocks. Even among groups of portability, relative size, yield in relation to effort,

290 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 291 APTARA(GRG QUARK)

palatability, nutritional value, toxicity, spoilage of the severity of the impact (e.g., high popula- rate, and desirability. Not all of these aspects can tion pressure), the vulnerability of the species to be considered here, but some may have played harvesting, and the extent to which extraneous important roles in determining the rate at which factors could have influenced the resources. different species were prehistorically harvested Regarding the first, caution must be exercized on the West Coast of South Africa. For example, when proposing high human population levels the nutritional value of mussels varies on a sea- in the landscape as the main factor behind sonal basis. Just prior to spawning, mussels are declining species’ body sizes. Each case needs to plump and the energy content of the flesh is be considered and assessed according to its own high due to the buildup of gonads. After spawn- merits: in some instances population pressure ing, they are scarcely worth collecting. Modern may explain much of the observed variability, but subsistence fishers are well aware of these phe- in others, population pressure may be irrelevant. nomena and time their harvesting accordingly, Most important here is the need to present inde- often using environmental cues such as the sea- pendent evidence for population growth, other- son when particular trees flower as an indication wise, circular logic could feed an argument that that mussels are “ripe” (Harris et al. 2003). ends up presenting population pressure as a self- Another important issue is that periodic fulfilling hypothesis. blooms of noxious algae on the West Coast of We have shown that declines in the average South Africa can result in mussels becoming sizes of mussels, two species of limpets, and the lethally toxic to humans (Matthews and Pitcher rock lobster all coincided with a period when 1996; Pitcher 1999). In Elands Bay Cave, an human occupation of the study area was inten- abrupt hiatus in the harvesting of mussels takes sifying, and when reliance on marine resources place at about 9500 BP (Parkington 1981), was increasing (Sealy and van der Merwe 1988). during which harvesting switched for a brief The coincidence implies that human harvesting time to focus on much less abundant species, was responsible, but this conclusion needs to be namely, the whelk Burnupena and limpets. evaluated by asking whether these species were Although speculative, it is not beyond the sufficiently vulnerable for their populations to bounds of possibility that harvesters were have been influenced by harvesting, and struck by a harmful algal bloom that made whether there were any other factors that could mussels toxic, compelling a switch of diet have explained the trends in size. (Parkington 1981; Parkington et al. 1988). In the case of both the mussels and the Finally, during the megamidden period of limpets, we could find no convincing evidence 3000–2000 BP, there is good evidence that that environmental conditions could have caused mussel were dried and stored for later consump- the declines in size during the megamidden tion (Henshilwood et al. 1994). This would have period. Environmental changes did take place allowed more extended use of mussels, possibly over the period when sizes were declining, but tiding people over periods when the mussels none of the factors examined provided a plausi- were in poor condition, and overcoming limita- ble explanation for a diminishment in size in tions imposed by the greater frequency of toxic both groups. Limpets and mussels belong to algal blooms in summer and storms during win- two very different trophic groups, being, respec- ter (Jerardino 1996; Parkington et al. 1988). tively, grazers and filter feeders. This alone is significant, because changes in many of the INTEGRATING THE EVIDENCE environmental factors examined should yield AND CONCLUSIONS opposite outcomes for these two trophic groups. Moreover, both mussels and limpets are highly Clearly, the potential impacts of human exploita- vulnerable to the effects of harvesting tion on marine resources will depend on a blend because they are accessible, nonmobile as

PRECOLONIAL IMPACT IN SOUTH AFRICA 291 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 292 APTARA(GRG QUARK)

adults, have intermittent larval recruitment, ACKNOWLEDGMENTS

and to a large extent lack refuges beyond the Financial support for archaeological excavations of reach of harvesters. We would expect them to several sites, processing and dating of the material, be depleted by intense harvesting. Was har- and building of the database was received from the vesting intense enough to accomplish this? History of Marine Animal Populations Project (Inter- Rough calculations based on midden sizes national Consortium), the National Research Foun- dation (NRF, South Africa), Swan Fund (Oxford, and shell densities leads to the conclusion UK), University Research Fund (UCT), and Wenner- that about 1,666 kg (wet whole mass) would Gren Foundation (Chicago, Illinois). The biological have been removed per kilometer per year research was funded by the NRF, UCT, and the (Griffiths and Branch 1997). This compares Andrew Mellon Foundation. Many thanks to J. with a figure of 5,500 kg km–1 y–1 for highly Erlandson, R. Klein, G. Sampson, C. Marean, and R. intense modern subsistence fishing on the Yates for discussions on the subject matter of this chapter, to R. Klein and C. Marean for making bibli- southeast coast of South Africa, where severe ographic references available, and to J. Sealy for mak- depletion of stocks has been recorded ing isotopic data on West Coast human skeletons (Hockey et al. 1988). Given the combination available. Thanks are also extended by one of us (A.J.) of intense harvesting during the megamid- to J. C. Castilla for introducing me to one of the most den period from limited available rocky fascinating scientific endeavors and for helping me to forge what is needed to follow this path. shores, high vulnerability of mussels and limpets, and an absence of alternative expla- nations for declines in mean sizes of these REFERENCES CITED species, harvesting remains the most parsi- Branch, G. M. monious and robust explanation for the 1974 The Ecology of Patella Linnaeus from the declines in their mean sizes over this period. Cape Peninsula, South Africa. III. Growth Rates. However, there is a caveat. Over the same Transactions of the Royal Society of South Africa time, there was a significant decline in the 41:161–193. sizes of rock lobsters. It is extremely unlikely 1975 Notes on the Ecology of Patella concolor and Cellana capensis, and the Effects of Human Con- that this decline can be attributed to shore- sumption on Limpet Populations. Zoologica based harvesting. It is hard to imagine more Africana 10:75–85. than a tiny fraction of the rock lobster popula- Branch, G. M., and A. Marsh tion being harvested from the shore. The bulk 1978 Attachment Forces and Shell Shape in of the population lives in the subtidal zone, Patella: Adaptive strategies. Journal of Experi- mental Marine Biology and Ecology 34:111–130. where it would have been inaccessible, and Branch, G. M., and F. Odendaal even if the shallow-water sector was harvested, 2003 Marine Protected Areas and Wave Action: it would have been replenished by movement Impacts on a South African Limpet, Cymbula of adults. In short, rock lobsters would not oculus. Biological Conservation 114:255–269. have been sufficiently vulnerable to shore- Branch, G. M., and C. N. Steffani based fishing for the population to have been 2004 Can We Predict the Effects of Alien Species? A Case-History of the Invasion of South Africa dented sufficiently by harvesting to alter the by Mytilus galloprovincialis (Lamarck). Journal of size composition. Additionally, the availability, Experimental Marine Biology and Ecology 300: productivity, and distribution of rock lobsters 189–215. all depend strongly on environmental condi- Buchanan, W. tions, as demonstrated for modern popula- 1985 Middens and Mussels: An Archaeological Enquiry. South African Journal of Science 81: tions. There are thus good reasons for distrust- 15–16. ing prehistoric harvesting as a cause of the 1988 Shellfish in Prehistoric Diet: Elands Bay, SW Cape decline in rock lobster size, even though it too Coast, South Africa. Cambridge Monographs in coincides with the period of intensification of African Archaeology 31, BAR International Series harvesting. 455. British Archaeological Reports, Oxford.

292 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 293 APTARA(GRG QUARK)

Buchanan, W., S.L. Hall, J. Henderson, A. Olivier, J. M. Eekhout, S., C. M. Raubenheimer, G. M. Branch, A. L. Pettigrew, J. E. Parkington, and R. T. Robertshaw Bosman, and M. O. Bergh 1978 Coastal Shell Middens in the Paternoster 1992 A Holistic Approach to the Exploitation of Area, South-. South African Archae- Intertidal Stocks: Limpets as a Case History. South ological Bulletin 33:89–93. African Journal of Marine Science 12: 1017–1029. Bustamante, R. H., and G. M. Branch Erlandson, J. M. 1996a Large Scale Patterns and Trophic Structure 2001 The Archaeology of Aquatic Adaptations: Par- of Southern African Rocky Shores: The Roles of adigms for a New Millennium. Journal of Geographic Variation and Wave Exposure. Jour- Archaeological Research 9:287–350. nal of Biogeography 23:339–351. Griffiths, R. J. 1996b The Dependence of Intertidal Consumers 1976 The Larval Development of Pyura stolonifera on Kelp-Derived Organic Matter on the West (Tunicata) from the Cape Peninsula. Transactions Coast of South Africa. Journal of Experimental of the Royal Society of South Africa 42:1–9. Marine Biology and Ecology 196:1–28. Griffiths, C. L., and G. M. Branch Bustamante, R. H., G. M. Branch, and S. Eekhout 1997 The Exploitation of Coastal Invertebrates and 1995a Maintenance of an Exceptional Grazer Bio- Seaweeds in South Africa: Historical Trends, mass in South Africa: Subsidy by Subtidal Kelps. Ecological Impacts and Implications for Man- Ecology 76:2314–2329. agement. Transactions of the Royal Society of Bustamante, R. H., G. M. Branch, S. Eekhout, B. South Africa 52:121–148. Robertson, P. Zoutendyk, M. Schleyer, A. Dye, D. Halkett, D., T. Hart, R. Yates, T.P. Volman, J. E. Parking- Keats, M. Jurd, and C. D. McQuaid ton, J. Orton, R. G. Klein, K. Cruz-Uribe, and G. Avery 1995b Gradients of Intertidal Productivity around 2003 First Excavation of Intact Middle Stone Age Lay- the Coast of South Africa and Their Relation- ers at Ysterfontein, Western , South ship with Consumer Biomass. Oecologia 102: Africa: Implications for Middle Stone Age Ecology. 189–201. Journal of Archaeological Science 30: 955–971. Castilla, J. C., G. M. Branch, and A. Barkai Harris, J. M., G. M. Branch, B. L. Elliott, B. Currie, A. 1994 Exploitation of Two Critical Predators: The Dye, C. D. McQuaid, B. Tomalin, and C. Velasquez Gastropod Concholepas concholepas and the Rock 1998 Spatial and Temporal Variability in Recruit- Lobster Jasus lalandii. In Rocky Shores: Exploita- ment of Intertidal Mussels Around the Coast of tion in Chile and South Africa, edited by R. W. Southern Africa. South African Journal of Zoology Siegfried, pp. 101–130. Springer-Verlag, Ecologi- 33:1–11 cal Studies, Berlin. Harris, J., G. M. Branch, C. Sibaya, and K. Bill Cockcroft A. C. 2003 The Sokhulu Subsistence Mussel-Harvesting 2001 Jasus lalandii “Walkouts” or Mass Strand- Project: Co-management in Action. In Waves of ings in South Africa during the 1990s: An Change.Coastal Co-management in Southern Africa, Overview. Marine and Freshwater Research 52: edited by M. Hauck and M. Sowman, pp. 61–98. 1085–1094. University of Press, Cape Town. Conard, N. J., T.J. Prindiville, and A. Kandel Henshilwood, C., P. Nilssen, and J. Parkington 1999 The 1998 Fieldwork on the Stone Age 1994 Mussel Drying and Food Storage in the Late Archaeology and Palaeoecology of the Geelbek Holocene, SW Cape, South Africa. Journal of Dunes, West Coast National Park, South Africa. Field Archaeology 21:103–109. Southern African Field Archaeology 8:35–45. Heydorn, A. E. F. Day, E., and G. M. Branch 1969 The Rock Lobster of the South African West 2002 Effects of Sea urchins (Parechinus angulo- Coast Jasus lalandii (H. Milne-Edwards). 2. Popu- sus) on Recruits and Juveniles of Abalone lation Studies, Behaviour, Reproduction, Moulting, (Haliotis midae). Ecological Monographs 72: Growth and Migration. Investigational Report of the 133–149. Division of Sea Fisheries, South Africa 71:1–52. Denny, M. N., T.L. Daniel, and M. A. R. Koehl Hockey, P. A. R., and A. L. Bosman 1985 Mechanical Limits to Size in Wave-Swept 1994 Man as an Intertidal Predator in Transkei: Dis- Organisms. Ecological Monographs 55:69–102. turbance, Community Convergence, and Manage- Dye, A. H., T.A. Lasiak, and S. Gabula ment of a Natural Food Resource. Oikos 46:3–14. 1997 Recovery and Recruitment of the Brown Hockey, P. A. R., A. L. Bosman, and W.R. Siegfried Mussel Perna perna (L.) in Transkei: Implica- 1988 Patterns and Correlates of Shellfish Exploitation tions for Management. South African Journal of by Coastal People in Transkei: An Enigma of Protein Zoology 32:118–123. Production. Journal of Applied Ecology 25:353–363.

PRECOLONIAL IMPACT IN SOUTH AFRICA 293 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 294 APTARA(GRG QUARK)

Horwitz, L. Jerardino, A., and R. Yates 1979 From Materialism to Middens: A Case Study 1996 Preliminary Results from Excavations at at Eland’s Bay, Western Cape, South Africa. Steenbokfontein Cave: Implications for Past and Unpublished B.A. (honors) dissertation, Depart- Future Research. South African Archaeological ment of Archaeology, University of Cape Town. Bulletin 51:7–16. Jacobs, Z., G. A. T. Duller, and A. Wintle 1997 Excavations at Mike Taylor’s Midden: A Sum- 2003a Optical Dating of Dune Sand from Blombos mary Report and Implications for a Re-charac- Cave, South Africa. II. Single Grain Data. Jour- terization of Megamiddens. South African nal of Human Evolution 44:613–625. Archaeological Bulletin 52:43–51. Jacobs, Z., A. G. Wintle, and G. A. T. Duller Jerardino, A., J. C. Castilla, J. M. Ramírez, and N. 2003b Optical Dating of Dune Sand from Blombos Hermosilla Cave, South Africa. I. Multiple Grain Data. Jour- 1992 Early Coastal Subsistence Patterns in Central nal of Human Evolution 44:599–612. Chile: A Systematic Study of the Marine-Inverte- Jerardino, A. brate Fauna from the Site of Curaumilla-1. Latin 1993 Mid- to Late-Holocene Sea-Level Fluctua- American Antiquity 3:43–62. tions: The Archaeological Evidence at Tortoise Jeradino, A., R. Navarro, and P. Nilssen Cave, South-western Cape, South Africa. South 2001 Cape Rock Lobster (Jasus lalandii) Exploita- African Journal of Science 89:481–488. tion in the Past: Estimating Carapace Length 1995a The Problem with Density Values in Archae- from Mandible Sizes. South African Journal of ological Analysis: A Case Study from Tortoise Science 97:59–62. Cave, Western Cape, South Africa. South African Johnston, S. J., and D. S. Butterworth Archaeological Bulletin 50:21–27. 2005 Evolution of Operational Management Pro- 1995b Late Holocene Neoglacial Episodes in cedures for the South African West Coast Rock Southern South America and Southern Africa: A Lobster (Jasus lalandii) Fishery. New Zealand Comparison. The Holocene 5:361–368. Journal of Marine and Freshwater Research 1996 Changing Social Landscapes of the Western 39:687–702. Cape Coast of Southern Africa over the Last Klein, R. G., G. Avery, K. Cruz-Uribe, D. Halkett, J. E. 4500 years. Unpublished Ph.D. dissertation, Parkington, T. Steele, T.P. Volman, and R. Yates Department of Archaeology, University of Cape 2004 The Ysterfontein 1 Middle Stone Age site, Town. South Africa, and Early Human Exploitation of 1997 Changes in Shellfish Species Composition Coastal Resources. Proceedings of the National and Mean Shell Size from a Late-Holocene Academy of Science 101:5708–5715. Record of the West Coast of Southern Africa. Kyle, R., B. Pearson, P. J. Fielding, and W. D Robertson Journal of Archaeological Science 24:1031–1044. 1997 Subsistence Shell-fish Harvesting in the 1998 Excavations at Pancho’s Kitchen Midden, Maputaland Marine Reserve in Northern Western Cape Coast, South Africa: Further KwaZulu-Natal, South Africa: Rocky Shore Observations into the Megamidden Period. Organisms. Biological Conservation 82:183–192. South African Archaeological Bulletin 53:17–25. Lasiak, T.A. 2003 Precolonial Settlement and Subsistence 1991 The Susceptibility and/or Resilience of Rocky along Sandy Beaches South of Elands Bay, West Littoral Mollusks to Stock Depletion by Indige- Coast, South Africa. South African Archaeological nous Coastal People of Transkei, Southern Bulletin 58:53–62. Africa. Conservation Biology 56:245–264. 2007 Excavations at a Hunter-Gatherer Site Known 1992 Contemporary Shellfish-Gathering Practices as “Grootrif G” Shell Midden, Lambertsbay, of Indigenous Coastal People in Transkei: Some Western Cape Province. South African Archaeo- Implications for Interpretation of the Archaeo- logical Bulletin 62:162–170. logical Record. South African Journal of Science Jerardino, J., and R. Navarro 88:19–28. 2002 Cape Rock Lobster (Jasus lalandii) Remains Lasiak T.A., and A. Dye from South African West Coast Shell Middens: 1989 The Ecology of the Brown Mussel Perna Preservational Factors and Possible Bias. Journal perna in Transkei, Southern Africa: Implications of Archaeological Science 29:993–999. for the Management of a Traditional Food Jerardino, A., and N. Swanepoel Resource. Biological Conservation 47:245–257. 1999 Painted Slabs from Steenbokfontein Cave: Lasiak, T., and J.G. Field The Oldest Known Parietal Art in Southern 1995 Community-Level Attributes of Exploited and Africa. Current Anthropology 40:542–548. Non-Exploited Rocky Infratidal Macrofaunal

294 PRECOLONIAL IMPACT IN SOUTH AFRICA GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 295 APTARA(GRG QUARK)

Assemblages in Transkei. Journal of Experimen- Materials, Artefacts and Sporadic Change. tal Marine Biology and Ecology 185:33–53. Southern African Humanities 18:1–28. Lee-Thorp, J., J. Sealy, and N. Van der Merwe Parkington, J. 1989 Stable Carbon Isotope Ratio Differences 1976 Coastal Settlement between the Mouths of between Bone Collagen and Bone Apatite, and the Berg and the Olifants Rivers, Cape Province. Their Relationship to Diet. Journal of Archaeolog- South African Archaeological Bulletin 31:127–140. ical Science 16:585–599. 1981 The Effects of Environmental Change on the Marean, C. W., M. Bar-Mattews, J. Bernatchez, E. Scheduling of Visits to the Elands Bay Cave, Cape Fisher, P. Goldberg, A. Herries, Z. Jacobs, A. Jer- Province, S. A. In Patterns of the Past, edited by I. ardino, P. Karkanas, N. Mercier, M. Minichillo, P. J. Hodder, G. Isaac, and N. Hammond, pp. 341–359. Nilssen, E. Thosmpson, C. Teibolo, H. Valladas, Cambridge University Press, Cambridge. and H. Williams 1995 Elands Bay Cave: A View on the Past. Unpub- 2007 Early Human Use of Marine Resources and lished Manuscript. University of Cape Town. Pigment in South Africa during the Middle Pleis- 2003 Middens and Moderns: Shellfishing and the tocene, Nature 449:905–908. Middle Stone Age of the Western Cape, South Marshall, D. J., and C. D. McQuaid African Journal of Science 99:242–247. 1989 The Influence of Respiratory Responses on Parkington, P., C. Poggenpoel, W. Buchanan, T. Robey, the Tolerance to Sand Inundation of the Limpet A. Manhire, and J. Sealy Patella granularis L. (Prosobranchia) and 1988 Holocene Coastal Settlement Patterns in the Siphonaria capensis Q. and G. (Pulmonata). Jour- Western Cape. In The Archaeology of Prehistoric nal of Experimental Marine Biology and Ecology Coastlines, edited by G. Bailey and J. Parkington, 128:191–201. pp. 22–41. Cambridge University Press, Cam- Matthews, S. G., and G. C. Pitcher bridge. 1996 Worst Recorded Marine Mortality on the Pitcher, G. C. South African Coast. In Harmful and Toxic Algal 1999 Harmful Algal Blooms of the Benguela Current. Blooms, edited by T. Yasumoto, Y. Oshima and Y. Sea Fisheries Research Institute, Cape Town. Fukuyo, pp. 89–92. Intergovernmental Oceano- Pollock, D. E., A. C. Cockcroft, and P. C. Goosen graphic Commission of UNESCO, Paris, France. 1997 A Note on Reduced Rock Lobster Growth Mayfield, S., and G. M. Branch Rates and Related Environmental Anomalies in 2000 Interrelations among Rock Lobsters, Sea Southern Benguela. 1988–1995. South African Urchins, and Juvenile Abalone: Implications for Journal of Marine Science 18:287–293. Community Management. Canadian Journal of Pollock, D. E., A. C. Cockcroft, J. C. Groeneveld, and Fisheries and Aquatic Science 57:2175–2185. D. S. Schoeman Mayfield, S., G. M. Branch, and A. Cockcroft 2000 The Commercial Fisheries for Jasus and Pal- 2000 Relationships among Diet, Growth Rate, and inurus Species in the South-East Atlantic and Food Availability for the South African Rock South-West Indian Oceans. In: Spiny Lobsters: Lobster, Jasus lalandii (Decapoda, Palinuridea). Fisheries and Culture, edited by B. F. Phillips and Crustaceana 73:8115–834. J. Kittaka, pp. 105–120. Blackwell, Oxford. Menge, B. A., J. Lubchenco, M. E. S. Bracken, F. Redman, C. L. Chan, M. M. Foley, T.L. Freidenburg, S. D. Gaines, 1999 Human Impact on Ancient Environments. G. Hudson, C. Krenz, H. Leslie, D. N. L. Menge, R. University of Arizona Press, Tucson. Russel, and M. S. Webster Robertshaw, P. T. 2003 Coastal Oceanography Sets the Pace of Rocky 1978 Archaeological Investigations at Intertidal Community Dynamics. Proceeding of Lagoon, Cape Province. South African Archaeo- the National Academy of Science 100:12229–12234. logical Bulletin 10:139–148. Miller, D. 1979 Excavations at Duiker Eiland, Vredenburg 1990 A Southern African Late Quaternary Sea-level District, Cape Province. Annals of the Cape Curve. South African Journal of Science 86:456–458. Provincial Museums 1:1–26. Nielsen, K. J., and S. A. Navarrete Robey, T. 2004 Mesoscale Regulation Comes from the Bot- 1987 The Stratigraphic and Cultural Sequence at tom Up: Intertidal Interactions between Con- Tortoise Cave, Verlorenvlei. In Papers in the Pre- sumers and Upwelling. Ecology Letters 7:31–41. history of the Western Cape, South Africa, edited by Orton, J J. Parkington and M. Hall, pp. 294–325. BAR 2006. The Later Stone Age Lithic Sequence at International Series 332 (ii). British Archaeolog- Elands Bay, Western Cape, South Africa: Raw ical Reports, Oxford.

PRECOLONIAL IMPACT IN SOUTH AFRICA 295 GRBQ335-3427G-C12[279-296].qxd 01-31-2008 01:45 PM Page 296 APTARA(GRG QUARK)

Sadr, K., and A. B. Smith Steffani, C. N., and G. M. Branch 1992 Final Report of the Vredenburg Peninsula 2003 Growth Rate, Condition, and Shell Shape of survey 1991/1992. Unpublished report to Mytilus galloprovincialis: Responses to Wave Expo- Anglo-American Chairman’s Fund. University sure. Marine Ecology Progress Series 24:197–209. of the Witwatersrand, Johannesburg. Stoline, M. R. Sealy, J. 1981 The Status of Multiple Comparisons: Simul- 1989 Reconstruction of Later Stone Age Diets in the taneous Estimation of All Pairwise Comparisons South-western Cape, South Africa: Evaluation and in One Way Anova Designs. American Statisti- Application of Five Isotopic and Trace Element Tech- cian 35:134–141. niques. Unpublished Ph.D. dissertation, Depart- Stuart, V, J.G. Field, and R. C. Newell ment of Archaeology, University of Cape Town. 1982 Evidence for Absorption of Kelp Detritus by Sealy, J., and N. Van der Merwe the Ribbed Mussel Aulacomya ater Using a New 1988 Social, Spatial and Chronological Patterning 51Cr-Labelled Microsphere Technique. Marine in Marine Food Use as Determined by 13C Mea- Ecology Progress Series 9:263–271. surements of Holocene Human Skeletons from Swadling, P. the South-western Cape, South Africa. World 1976 Changes Induced by Human Exploitation in Archaeology 20:87–102. Prehistoric Shellfish Populations. Mankind 10: Sealy, J., T. Maggs, A. Jerardino, and J. Kaplan 156–162. 2004 Excavations at Melkbosstrand: Variability Thackeray, J. F. among Herder Sites on . South African 1988 Molluskan Fauna from Klasies River, South Archaeological Bulletin 59:17–28. Africa. South African Archaeological Bulletin 43: Shannon, L. V. 27–32. 1985 The Benguela Ecosystem. Part I. Evolution of Tonner, T.W. W. the Benguela, Physical Features and Processes. 2005 Later Stone Age Shellfishing Behaviour at Oceanography and Marine Biology Annual Review Dunefield Midden (Western Cape, South Africa). 23:105–182. Journal of Archaeological Science 32:1390–1407. Siegfried, W. R, P.A. R. Hockey, and G. M. Branch Van Andel, T.H. 1994 The Exploitation of Intertidal and Subtidal 1989 Late Pleistocene Sea Levels and the Human Biotic Resources of Rocky Shores in Chile and Exploitation of the Shore and Shelf of Southern South Africa: An Overview. In Rocky Shores: South Africa. Journal of Field Archaeology 16:132–153. Exploitation in Chile and South Africa, edited Vogel, J. C. by R. W. Siegfried, pp. 1–15. Springer-Verlag, 2001 Radiometric Dates for the Middle Stone Age Ecological Studies, Berlin. in South Africa. In Humanity from African Nais- Silberbauer, B. I. sance to Coming Millennia, edited by M. A. Raath, 1971 The Biology of the South African Rock Lob- J. Maggi-Cecchi, and G. A. Doyle, pp. 261–268. ster Jasus lalandii (H. Milne-Edwards). I. Devel- Florence University Press, Florence, Italy. opment. Investigational Report of the Division of Volman, T.P. Sea Fisheries, South Africa 92:1–10. 1978 Early Archaeological Evidence for Shellfish Smith, A. B., K. Sadr, J. Gribble, and R. Yates Collecting. Science 201:911–913. 1991 Excavations in the South-western Cape, Webley, L. South Africa, and the Archaeological Identity of 2002 The Re-excavation of Spoegrivier Cave on the Prehistoric Hunter-Gatherers within the last West Coast of South Africa. Annals of the Eastern 2000 years. South African Archaeological Bulletin Cape Museums 2:19–49. 46:71–91. Wieters, E. A., D. M. Kaplan, S. A. Navarrete, A. Spennemann, D. H. R. Sotomayor, J. Largier, K. J. Nielsen, and F. Véliz 1986 Effects of Human Predation and Changing 2003 Alongshore and Temporal Variability in Chloro- Environment on Some Mollusk Species on Ton- phyll a Concentration in Chilean Nearshore gatapu, Tonga. In The Walking Larder: Patterns of Waters. Marine Ecology Progress Series 249:93–105. Domestication, Pastoralism, and Predation, edited Xavier, B. M., G. M. Branch, and E. Wieters by J. Clutton-Brock, pp. 326–335. Unwin 2007 Abundance, Condition, Growth Rate and Hyman, London. Recruitment of Mytilus galloprovincialis along the Steele, T.E., and R. G. Klein West Coast of South Africa in Relation to Upwelling. 2005 Mollusk and Tortoise Size as Proxies for Marine Ecology Progress Series 346:189–201. Stone Age Population Density in South Africa: Zar, J. H. Implications for the Evolution of Human Cul- 1984 Biostatistical Analysis. Second edition. Pren- tural Capacity. Munibe 57:5–21. tice-Hall, Englewood Cliffs, New Jersey.

296 PRECOLONIAL IMPACT IN SOUTH AFRICA