Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa

Quaternary International xxx (2015) 1e12

Contents lists available at ScienceDirect

Quaternary International journal homepage: www.elsevier.com/locate/quaint

Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa Antonieta Jerardino a, b a ICREA/ Department of Experimental & Health Sciences, Universitat Pompeu Fabra, CaSEs Research Group, Ramon Trias Fargas 25-27, 08005 Barcelona, Spain b Department of Anthropology & Archaeology, University of South Africa, PO Box 392, UNISA, Pretoria, South Africa

a r t i c l e i n f o

a b s t r a c t

Article history: Available online xxx

Geoarchaeological studies today are a vital component of archaeological research. The sedimentary environments of coastal settings in particular are highly dynamic and governed by a variety of factors among which changes in sea level play an important role. The object of such studies often involves the study of sediments deposited by natural factors in overlapping geological and anthropogenic contexts. Archaeomalacological studies conducted on shell middens along the West Coast of South Africa in the last two decades have also identified the presence of additional natural components that became incorporated into archaeological sites through active but inadvertent human agency. These sediments are relatively large particles (2e20 mm) of water-worn shells and water-worn pebbles (WWSP) that became entangled among the byssus threads that rocky shore mussels use for attaching themselves to hard substrate. Prehistoric shellfish foraging and subsequent transport of rocky shore mussels along with their byssus contents to campsites ensured the inclusion of WWSP into archaeological middens. This study shows that WWSP abundances and the proportion of their organic fraction (water worn shell) can be used as proxies for coastal palaeoenvironmental reconstruction and as a complement to archaeomalacological studies for inferring technologies involved in shellfish collection and transport. Conclusions presented here probably apply to other similarly configured shorelines and shell assemblages of comparable age from elsewhere in southern Africa and beyond. Follow-up studies from this sub-region and beyond might confirm and/or qualify the use of WWSP observations in other coastal settings. © 2015 Elsevier Ltd and INQUA. All rights reserved.

Keywords: Geoarchaeology Marine sediments Mussels Molluscs Coastal foraging Netting bags

1. Introduction Sediment analyses and taphonomic studies provide archaeologists with a better understanding of environmental changes, site stratigraphy and site formation processes, as well as routinely supplying important clues on possible post-depositional disturbance and the use of space through the identification of activity areas (Stein, 1992; Butzer, 2006; Goldberg and Macphail, 2006). In a similar vein, archaeomalacological analyses have also contributed to palaeoenvironmental reconstructions based on the particular ecological requirements of mollusc species, and also towards the interpretation of depositional sequences through the study of shell fragmentation and weight loss (Waselkov, 1987; Ford, 1992;

E-mail addresses: [email protected], [email protected].

Mowat, 1994; Claassen, 1998; Faulkner, 2011; Jerardino, in press). Studies that would be able to combine both of these approaches are thus expected to increase the reliability and accuracy of behavioural and environmental inferences and make important methodological contributions to the analysis of archaeomalacological assemblages. The study of micro-sediments and mollusc assemblages from archaeological sites in the Elands Bay and Lamberts Bay areas (Fig. 1; hereafter referred to as ‘the study area’) in particular have aided in reconstructing the local cultural sequence and palaeoenvironmental changes (Butzer, 1979; Miller, 1987; Jerardino, 1993). Larger marine sediments (2e20 mm) termed here “waterworn shells and water-worn pebbles (WWSP)” have also been identified in shell middens several thousand years old, and have added new directions to the interpretation of the past (Jerardino, 1993; Miller et al., 1995). As revealed by observations on freshly collected mussels, the presence of WWSP has been explained as a

http://dx.doi.org/10.1016/j.quaint.2015.06.057 1040-6182/© 2015 Elsevier Ltd and INQUA. All rights reserved.

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

2

A. Jerardino / Quaternary International xxx (2015) 1e12

Fig. 1. Geographic setting and location of research area; a) and b) sites and places mentioned in the text: Elands Bay Cave (EBC), Elands Bay Open (EBO), Spring Cave (SC), Pancho's Kitchen Midden (PKM), Steenbokfontein Cave (SBF), and Tortoise Cave (TC); c) representation of a rocky shore mussel and its byssus sediment content (WWSP). Aerial photographs and maps are from Google Earth and SRTM (v4.1), processed by CGIAR-CSI (http://www.cgiar-csi.org).

result of having arrived in coastal sites attached to mussel threads (byssus, plural: byssi) of rocky shore mussels collected from nearby reefs (Yates, 1989). Variation in quantities of these marine sediments were initially interpreted as probably reflecting different modes of mussel collection, whereby larger quantities of these marine sediments reflected mass harvest (stripping off mats of mussels) as opposed to collecting beach-stranded mussels or targeting those growing in small clusters, both of which would have less associated quantities of byssus threads and sediments (Yates, 1989, pp. 16e18; see also; Parkington et al., 2014, p. 231). However, in terms of our current and local knowledge, mass harvest is only possible with iron technology such as spades, hoes and axes as seen

today in the East Coast of South Africa (Hockey and Bosman, 1986; Kyle et al., 1997). This technology was not available in the Western Cape until European colonization. Actualistic studies also show that stranded mussels would have been an unlikely source of food for coastal dwellers (Jerardino, 2014). Moreover, on the basis of Tortoise Cave and Pancho's Kitchen Midden observations (Jerardino, 1993, 1997) (Fig. 1), it has become clear that changes in the abundance of these sediments coincided with periods of sea level changes and coastal sediment instability. Although a good case was made in this regard around two decades ago, these results needed to be checked against data obtained from additional shell midden material analysed since then. The objective of this paper is thus to

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

A. Jerardino / Quaternary International xxx (2015) 1e12

do so by using observations from a total of six coastal sites with depositional sequences spanning the last 6000 years (Fig. 1). The interpretation of the changes in WWSP abundances is also expanded to include further palaeoenvironmental reconstructions and also behavioural aspects of shellfish collection.

3

direction owing to the prevailing southwest swell direction, although sediments can also move seasonally in a southerly direction during winter (Franceschini et al., 2003; Franceschini and Compton, 2006; Roberts et al., 2009). An active dune field and a long dune cordon are clearly visible, respectively, on the northern shore of Elands Bay and south of Baboon Point (Fig. 1a).

2. Environment: past and present 2.2. Palaeoenvironmental history The relevance of palaeoenvironmental changes in making sense of WWSP variability in local coastal assemblages is thus apparent. The physiography, geology, palaeoecology and palaeoenvironmental history of the South African West Coast has been described in considerable detail elsewhere (i.e., Shannon, 1985; Miller, 1987; Rogers, 1987; Compton, 2001, 2006; Chase and Meadows, 2007; Jerardino et al., 2013), and a brief summary on relevant observations is presented here. 2.1. Current environmental setting The South African West Coast climate is typically Mediterranean and is characterized by cold, wet winters and dry, windy, hot summers. The influence of the anticyclone high pressure system is strong during summer, with dominant winds from the south and south-east. The prevalence of winter rainfall is a product of cold frontal systems and westerly winds laden with moisture. Rainfall decreases markedly in a northerly direction, ranging from ~800 mm per year in Cape Town, 250 mm in Saldanha Bay, 170 mm in Elands Bay, 150 mm in Lamberts Bay, close to 120 mm near the Olifants River mouth and much less further north (Chase and Meadows, 2007) (Fig. 1). The Benguela Current is perhaps the most important physical feature influencing the West Coast environments (Shannon, 1985, 1989). Surface waters flow adjacent to the coast in a northenorth-westerly direction, moving progressively off-shore towards latitude 20
S. Average coastal sea-surface temperatures reach lowest values (8e10
C) during late spring and summer months (OctobereFebruary) in association with upwelling events. Relatively warmer sea-surface temperatures (11e14
C) are recorded during the rest of the year (Shannon, 1985, 1989). One of the three upwelling cells reported south of latitude 29
S is located just south of the Elands Bay area (Shannon, 1985) on the northern edge of Cape Columbine (32e33
S) (Fig. 1). The upwelling coastal waters bring enough nutrients to the sea surface and support a diverse and highly productive marine environment (Branch and Griffiths, 1988; Shannon, 1989). Cenozoic sediments cover most of the low-lying coastal platform of this stretch of the West Coast. Most of the landscape of the study area is dominated by hills and mountains of the Table Mountain Group sandstones, with faults running north-west and controlling the drainage patterns of existing rivers and streams (Rogers, 1987). The coastal plain (known locally as Sandveld) is a gently undulating landscape rising towards the interior to about 100e120 m above mean sea level. Several large and small sandstone outcrops, or koppies, add contrast to this otherwise open coastal plain (Roberts et al., 2006), the largest of which gives its form to the large headland of Baboon Point on the southern bank of Verlorenvlei (Fig. 1a). The coastline is dominated by log-spiral sandy beaches punctuated by rocky points and is often bordered by a pair of long littoral dune ridges, most of which are partially vegetated and in places also deflated by wind action (Miller, 1987; Rogers, 1987). River sand, recycled beach and older dune material, as well as the breakdown of carbonate shells and other invertebrate exoskeletons, are the most important sources of present-day beaches and dune fields. Seabed material is transported onto beaches in a net northward

The study area has experienced several minor but important environmental shifts during the Holocene as a result of climatological and related sea level changes. A trend of increasing aridity took place between 9000 cal. BP and 4500 cal. BP (Chase and Meadows, 2007; Chase and Thomas, 2007). This period is known as the Holocene Altithermal (also known as “Climatic Optimum”) and represents the warmest period of the Holocene in southern Africa. Sea level rose steadily from 125 m around 18 000 BP (Last Glacial Maximum) to its present position by 7500 cal. BP (Fig. 2). Coinciding with the Holocene Altithermal between 7000 and 4500 cal. BP, a 2e3 m sea level high stand created an open, but sheltered, estuary at Velorenvlei and placed the shoreline several hundred meters further inland than at present. For Elands Bay in particular, geoarchaeological field observations and archaeological proxy-evidence (Yates et al., 1986; Jerardino, 1993; Miller et al., 1993, 1995; Compton, 2001, 2006; Orton and Compton, 2006; Jerardino et al., 2013) are in agreement with this reconstruction. A short and minor sea level drop (6500e5500 cal. BP) to presentday levels seems to have interrupted this general trend of midHolocene sea level rise (Compton, 2001, 2006) (Fig. 2). Two subsequent minor sea level regressions around 4500e4200 cal. BP and between about 2500 and 1800 cal. BP coincided with two short, but significant, atmospheric and sea surface temperature cooling phases known as Neoglacial events (Jerardino, 1995; Compton, 2001, 2006). The third and more recent Neoglacial event is the well-known Little Ice Age (c. 600e100 cal. BP) (Cohen et al., 1992; Tyson and Lindesay, 1992; Jerardino, 1995; Chase and Meadows, 2007). Minor but noticeable northward latitudinal shifts of moisture-bearing frontal systems and expansions of polar waters into the Benguela Current were probably part of the driving mechanisms behind these climatic events. A short warming trend of modest amplitude, known as the Medieval Warm Period or Medieval Climatic Anomaly (c. 1200e650 cal. BP), is also recognized in southern Africa (Tyson and Lindesay, 1992; Chase and Meadows, 2007). The long and semi-vegetated dune cordon that parallels the coast south of Baboon Point (Fig. 1a) built up most likely between 5500 and 3000 BP, as revealed by the above-described sea level history, calibration of radiocarbon dates of shell middens on top of this land feature and geoarchaeological observations (Miller et al., 1993; Jerardino, 1993, 2003; Compton, 2001, 2006). It is likely that the dune field north of Elands Bay village (Fig. 1a) also formed in tandem with this long dune cordon. Moreover, evidence for intensified aeolian sand deposition between 4900 and 3100 cal. BP comes from Spring Cave and Elands Bay Cave (Fig. 1) where 0.4e0.5 m thick layers of fine sands with much reduced cultural remains were observed. A long and narrow lagoon with a tidal sea connection existed behind this dune cordon until about 4500 ca. BP (Miller et al., 1993). Sand supply and recent sea level changes are closely related, and the distinct Holocene dune fields commonly found immediately north of prominent headlands along the West Coast can be interpreted as representing pulses in sediment stocks (Illenberger, 1988; Bateman et al., 2011). According to this model, increases in Holocene sediment supply resulted from erosive processes during relatively minor, but important, marine transgressions that allowed

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

4

A. Jerardino / Quaternary International xxx (2015) 1e12

Fig. 2. Holocene sea level history along the West Coast of South Africa. Redrawn from Compton (2001, 2006).

the subsequent build-up of these dune fields. The Elands Bay data (see above) and sedimentological and chronometric studies of the Sixteen Mile Beach dune cordon (c. 4500 cal. BP) situated about 100 km south of Elands Bay (Fig. 1) lends support to this interpretation over that of an increase in sand supply from rivers (Compton and Franceschini, 2005; Franceschini and Compton, 2006). Consequently, the mid-Holocene 2e3 m high sea level stand (7000 and 4500 cal. BP) seems to have contributed significantly to the formation of many Holocene dune fields along the South African West Coast and also East Coast (c. 5000 BP; Illenberger, 1988; Roberts et al., 2009; Bateman et al., 2011). 3. Case study Since Tortoise Cave observations were published more than twenty years ago (Jerardino, 1993), WWSP from several local shell middens have been routinely quantified while conducting archaeomalacological analyses. WWSP are part of the coarsest natural sediment fraction associated with intertidal reefs. Fracturing and breakage of shells after mollusc death or dislodgment from fossiliferous deposits occurs often in high-energy marine settings, followed by extensive rolling and abrasion (Claassen, 1998, pp. 56e58; Franceschini and Compton, 2006, p. 1164). As a result, and depending on the local species (e.g. black mussel, white mussel, whelks and barnacles), water-worn shells (WWS) can be round, flat-oval or horseshoe-shaped shell fragments (Fig. 1c). In archaeological contexts, WWS range in size between approximately 2 to 20 mm. Water-worn pebbles (WWP) are worn and polished particles of mineral origin, most frequently including those derived from local quartzitic origin, and are probably the result of river sediment discharge and mechanical breakdown and abrasion of coastal rocks within high-energy marine environments. The size of WWP observed in archaeological contexts ranges from 2 to 15 mm. Absolute WWSP abundances (numbers of particles or their weight) are extremely low in shell assemblages where limpets dominate (see below), suggesting that these sediments are introduced in association with other mollusc species. Initial sampling of modern mussel beds for ascertaining the origin of WWSP reflects their presence in such contexts (Yates, 1989). These grains are also

unique because they tend to be larger than normal beach- or windblown sand grains and, except perhaps for kelp holdfasts, no other sea harvested material would contain them. In any case, kelp is a rare addition to archaeological middens and only observed at Steenbokfontein Cave (Fig. 1). Hence, as initially suggested, WWSP were most likely brought to the site as part of the byssus contents of mussels collected from nearby rocky reefs (Fig. 1c). Fragments of shells and other sediments have been observed accumulating among byssus threads of life mussel colonies (Barkai and Branch, 1988, p. 123), and byssus contents in archaeological deposits have also been recognized elsewhere (Waselkov, 1987, p. 100). The possibility that WWSP could have been transported to sites on people's feet and clothing is unlikely. People's feet are often kept submerged and wet during the collection of shellfish, which would preclude much of intertidal sediments clinging onto them for long. Sticky remnants of sediments would largely be shaken off when walking to camp sites at some distance from the shoreline. If clothing was a major transporting agent, a more random distribution of these sediments' abundances through time would then be expected as garments can change with the season of occupation and culturallymediated fashion, both of which are expected to be variable through time. It is thus very likely that WWSP in shell midden deposits results from mussel byssus contents being transported there as part of shellfish harvests. As indicated by archaeomalacological studies (e.g., Jerardino, 1997, 2014), rocky shore mussels were largely collected at both mid- and low-intertidal zones where they are most accessible. Possible differences in byssus quantities produced by black mussels living in different tidal zones (Barkai and Branch, 1988), and acting as larger or smaller sediment traps, would be evened out as a result of people foraging in both tidal zones. 4. Methods The samples used in this study date to the last 6100 calibrated radiocarbon years and are from six sites: Elands Bay Cave, Elands Bay Open, Spring Cave, Pancho's Kitchen Midden, Steenbokfontein Cave, and Tortoise Cave (Fig. 1). Sites were largely excavated following natural stratigraphy (Horwitz, 1979; Klein and Cruz-

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

A. Jerardino / Quaternary International xxx (2015) 1e12

Uribe, 1987; Robey, 1987; Parkington, 1988; Jerardino, 1998; Jerardino and Swanepoel, 1999), and shell samples were recovered during excavations for further study after routine on-site sieving. A chronological gap and lack of data due to an occupational hiatus dating to about 2900 and 2100 cal. BP in Elands Bay Cave, Tortoise Cave, Elands Bay Open and Spring Cave fortunately overlaps with Pancho's Kitchen Midden and Steenbokfontein Cave chronostratigraphy where this temporal gap does not exist. Except for almost all Pancho's Kitchen Midden samples sieved with a 1.5 mm mesh, and Elands Bay Cave samples sieved through a variety of mesh sizes, all other shell samples with their WWSP contents were passed through a 3.2 mm mesh. At least 90% of all shell bulk samples weigh 1e5 kg, and no shell sample weighing less than 0.5 kg was included in this study. Shell was sorted and taxonomically identified to species or genus level (depending on preservation) in the laboratory, whereupon minimum number of individuals (MNI) and weights were established. While sorting this material, WWSP were separated from the shells. WWS and WWP were counted and weighed separately. Doing so allowed the proportion of organic-derived sediments (WWS) of each sediment sample to be established. The relative abundance of WWSP is expressed here as number of sediment particles and also weight of these sediments per kilogram of associated mussel shells (Choromytilus meridionalis and Aulacomya ater). Earlier reports on WWSP abundances from Tortoise Cave and Pancho's Kitchen Midden (Jerardino, 1993, 1997; Miller et al., 1995) were based exclusively on the weight of C. meridionalis quantities due to the overwhelming dominance (95% by weight) of this species in these assemblages. However, A. ater appear in small but noticeable quantities in the additional shell samples considered here (up to 22% by weight), hence the need to include both species of rocky shore mussels

5

when computing WWSP abundances. Unfortunately, only weightbased WWSP abundances from Elands Bay Cave are available. Linear regression models on %WWS data were conducted with R statistical software, version 2.15.0 (R Development Core Team, 2012). Given the relatively small size of WWSP particles, different mesh sizes retain different quantities of these sediments. Consequently, and for comparative purposes, only those samples from Elands Bay Cave that were sieved with a 3.2 mesh are included in this study. Pancho's Kitchen Midden observations established with 1.5 mm mesh sieves can still be studied on its own for temporal trends. However, comparison among sites is desirable for at least three time periods represented in most sites considered in this study (500e900 cal. BP, 3000e3300 cal. BP, and 3700e3850 cal. BP). Hence, two of the few Pancho's Kitchen Midden samples sieved on site with a 3.2 mm mesh and dated to 540 cal. BP are also included here. Likewise, shell samples from this site initially sieved through a 1.5 mm sieve and dating to 755 cal. BP and 3805 cal. BP were passed through a 3.2 mm mesh in the laboratory and (re)quantified. Radiocarbon dates associated with each shell and sediment sample were calibrated using OxCal program and the latest updated ShCal calibration curve (Hogg et al., 2013). Calibration of marine shell dates follows Marine 13 (Reimer et al., 2013) and an added local marine reservoir value (DR ¼ 146 ± 85) is applied to computations (Dewar et al., 2012). Calibrated radiocarbon mid-points were rounded to the nearest half decade (Table 1). In two instances where stratigraphic layers were not dated (Pancho's Kitchen Midden and Tortoise Cave), it was assumed that these dated to a time chronologically equidistant between the layers above and below.

Table 1 Summary observations of WWSP abundances, proportion of organic-derived sediments (%WWS), distances between sites and coastline (meters), associated mean calibrated radiocarbon dates (m), and number of samples studied for each chrono-stratigraphic unit. Except for Pancho's Kitchen Midden samples that have been sieved with a 1.5 mm mesh, all shell samples included in this table and their WWSP contents were passed through a 3.2 mm mesh. Note that no data on WWSP particle counts is available for Elands Bay Cave. Site

Distance to nearest rocky shore (m)

Date cal BP (m)

WWS&P (n/kg mussels) ± SE

WWS&P (n/kg mussels) range

WWS&P (g/kg mussels) ± SE

WWS&P (g/kg mussels) range

% WWS (n)

% WWS (g)

EBC EBC EBC EBC EBC EBC EBC EBO EBO EBO EBO TC TC TC TC TC TC TC TC TC TC SC SC SC SC SC PKM PKM PKM PKM

725 725 725 725 725 725 725 430 430 430 430 4000 4000 4000 4000 4000 4000 4000 4000 4000 4000 825 825 825 825 825 1680 1680 1680 1680

370 500 885 1200 3065 3480 4790 565 615 1330 3010 650 1430 1475 1510 1665 3270 3750 4095 4440 4680 460 720 1015 3080 3740 540 755 2675 3030

e e e e e e e 42.9 ± 14.8 43.1 ± 11.4 57.3 ± 26.5 101.0 ± 19.6 2.9 ± 0.9 5.6 ± 1.3 3.1 ± 0.8 1.6 9.6 ± 2.6 36.2 ± 9.5 27.9 ± 1.1 27.2 ± 3.3 39.3 ± 3.6 30.1 ± 5.6 11.7 ± 8.9 16.8 ± 4.2 10.6 ± 2.4 43.5 ± 5.9 97.2 ± 44.9 25.2 ± 8.0 62.8 ± 11.8 60.5 ± 9.8 124.0 ± 12.9

e e e e e e e 12e229 36e219 20e157 109e278 6e20 16e36 3e11 9 13e62 28e162 59e88 44e65 19e50 29e91 2e3 2e25 4e9 25e115 6e116 40e384 238e534 248e612 631e1340

5.3 ± 2.3 3.6 ± 1.1 1.3 6.9 ± 0.3 18.6 16.7 ± 3.4 17.3 ± 8.1 2.9 ± 0.8 6.3 ± 0.4 6.2 ± 2.8 10.1 ± 1.5 0.4 ± 0.2 1.1 ± 0.2 0.7 ± 0.1 0.2 1.7 ± 0.4 6.6 ± 1.7 5.6 ± 0.5 5.0 ± 0.9 7.8 ± 1.2 9.6 ± 4.2 1.3 ± 0.8 2.3 ± 0.4 1.5 ± 0.7 4.7 ± 0.9 13.2 ± 6.4 0.6 ± 0.1 1.3 ± 0.2 1.5 ± 0.1 2.2 ± 0.2

0.6e4.8 1.1e2.1 0.5 2.2e2.6 9.9 6.9e13.5 3.9e13.1 1.0e12.0 3.8e19.5 2.7e16.5 11.6e25.5 0.6e2.5 7.1e3.2 1.1e2.3 1.0 2.2e11.0 4.4e29.3 12.1e14.5 7.3e11.8 4.7e12.3 3.4e50.6 0.3e0.4 0.3e3.9 0.5e1.7 1.9e13.4 0.7e16.3 2.9e8.0 4.6e11.0 7.2e16.5 12.7e23.4

e e e e e e e 29.7 36.7 47.5 73.0 61.9 64.6 49.7 66.7 61.4 51.6 61.4 75.9 72.0 80.4 41.7 77.8 66.7 83.2 74.3 32.6 19.4 56.0 72.1

45.9 26.0 20.0 27.1 48.5 61.6 48.1 29.6 30.5 39.9 61.6 40.2 47.6 57.1 52.0 49.1 36.6 44.7 65.8 49.6 73.8 41.7 78.1 50.8 76.1 62.4 34.9 43.4 52.8 65.0

± ± ± ± ± ± ±

3.4 1.8 2.1 1.3 9.1 2.1 8.4

± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

6.1 5.6 2.6 6.7 3.5 4.6 8.3 8.4 27.9 7.5 8.9 5.2 9.5 5.3 4.1

± 3.0 ± 16.9 ± 0.2 ± ± ± ± ± ± ± ± ±

6.5 6.8 6.6 2.1 4.2 4.9 9.1 6.9 11.7

± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

9.0 7.1 4.5 10.7 8.1 6.5 8.3 8.6 26.1 10.8 14.6 6.8 9.8 3.8 2.6

Number of samples 3 2 1 2 1 3 3 7 6 5 4 4 2 4 1 5 6 4 4 5 6 2 8 3 5 4 6 3 5 5

(continued on next page)

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

6

A. Jerardino / Quaternary International xxx (2015) 1e12

Table 1 (continued ) Site

Distance to nearest rocky shore (m)

Date cal BP (m)

WWS&P (n/kg mussels) ± SE

WWS&P (n/kg mussels) range

WWS&P (g/kg mussels) ± SE

PKM PKM PKM SBF SBF SBF SBF SBF SBF SBF

1680 1680 1680 2600 2600 2600 2600 2600 2600 2600

3120 3200 3805 2170 2340 2550 2650 3810 4390 6070

143.1 ± 35.7 45.3 ± 11.9 77.1 ± 6.2 4.5 ± 0.7 10.2 ± 3.5 5.6 ± 1.4 5.6 ± 0.9 19.7 ± 4.4 38.9 ± 4.3 71.7 ± 8.2

590e1687 222e458 277e520 2e21 4e37 2e21 7e34 13e99 11e168 13e254

3.1 1.1 2.3 0.6 1.1 0.8 0.6 2.5 3.0 7.6

In order to evaluate possible changes in WWSP content as a function of the distances travelled from coast to sites, the trajectories between foraging localities and campsites were established using aerial photos (Google Earth) and measuring at least four different stretches between sites and their nearest rocky reef(s) (as a straight line) and averaged. For Spring Cave, situated about 100 m above mean sea level (Fig. 1b), the tract of slope below this site was considered when calculating the average distance value for this site. 5. Results It is important to note that absolute WWSP abundances are very low in assemblages where limpets contribute substantially. A case in point are the roughly coeval Spring Cave and Elands Bay Cave shell samples dating to 720e460 cal. BP and 370e500 cal. BP, respectively (Table 1). Both sites can be compared because they are also situated at a similar distance from the nearest shoreline (825 m and 725 m, respectively; see Table 1). Black mussels contribute with small percentages to Spring Cave samples (6.4e30.0% by weight and 2.4e19.7% by MNI), while the opposite is the case for Elands Bay Cave (60.0e74.3% by weight and 45.7e61.3% by MNI). Standardizing for sample size (per kg of shell), Spring Cave WWSP weight abundances (0e1.1 g) are markedly smaller than those recorded for Elands Bay Cave (0.88e6.8). Table 1 presents summary observations of WWSP relative abundances (hereafter referred to as ‘abundances’; see Methods), associated calibrated radiocarbon dates, distances between sites and coastline, proportion of organic-derived sediments (%WWS) and number of samples studied for each chrono-stratigraphic unit at the six sites. Figs. 3 and 4 show a marked general pattern of high mean abundances of WWSP before 3000 cal. BP and low mean abundances thereafter. Variability around average values is modest in general, with the smallest standard errors among means with largest sample sizes (see Table 1: Steenbokfontein Cave), but highest among those samples dating to before 3000 cal. BP (i.e., Fig. 3: Spring Cave, Pancho's Kitchen Midden; Fig. 4: Spring Cave, Elands Bay Cave). Depending on the type of quantification, WWSP abundances after 3000 cal. BP are reduced to around half or less. The exception is evident in Pancho's Kitchen Midden only when quantification is expressed as weight of particles per kg of mussels, where the drop in abundance is not as pronounced (Fig. 4). Trends in time for this site are basically the same with both quantification methods and closely reflect observations reported previously when only black mussels were used to calculate relative abundances of WWSP (Jerardino, 1997). Quantities of organic-derived sediments in all studied samples expressed as a percentage of WWS in each sediment sample (% WWS) are shown in Fig. 5 as averages and standard errors. In Fig. 6, individual percentages for each studied assemblage per site are shown. A distinct pattern of decreasing contribution of the organic

± ± ± ± ± ± ± ± ± ±

0.1 0.2 0.2 0.1 0.2 0.2 0.2 0.5 0.3 0.8

WWS&P (g/kg mussels) range

% WWS (n)

13.0e31.2 5.3e10.2 8.1e16.8 0.2e3.8 0.4e2.6 0.1e2.5 0.3e5.3 2.4e11.0 1.0e10.0 1.3e23.2

78.4 70.1 66.4 58.5 63.7 75.5 63.2 79.9 83.2 82.4

± ± ± ± ± ± ± ± ± ±

2.0 4.2 8.6 4.7 7.9 7.0 3.6 3.8 2.0 1.4

% WWS (g) 74.8 78.3 62.2 56.4 64.4 76.5 60.3 78.7 79.1 81.0

± ± ± ± ± ± ± ± ± ±

1.5 2.6 8.9 6.8 8.9 6.3 5.5 6.6 2.6 2.8

Number of samples 3 5 5 11 8 8 9 10 9 11

sediment component is evident. Changes in %WWS is about twofold in more recent samples (<3000 cal. BP) when compared to older samples. Higher values from Spring Cave (1015 cal. BP) and Tortoise Cave (650 cal. BP) among post-1100 cal. BP observations depart somewhat from this general pattern (Table 1; Figs. 5 and 6). Despite evident data dispersion, the overall chronological trend in % WWS is statistically significant for both presentations of results (p < 0.001). Linear regressions show that logarithmic models fit the data best for the reductions in averages (by number of WWS: F ¼ 35.83, df ¼ 1 and 31; by weight of WWS: F ¼ 26.6, df ¼ 1 and 38) (Fig. 5) and individual percentage values (by number of WWS: F ¼ 86.16, df ¼ 1 and 182; by weight of WWS: F ¼ 29.78; df ¼ 1 and 197) through time (Fig. 6). Fig. 7 shows the changes in WWSP abundances in relation to average distances between sites and nearest rocky shore for three different periods of time (500e900 cal. BP, 3000e3300 cal. BP, and 3700e3850 cal. BP) and assuming negligible changes in distances due to sea level changes (Fig. 2). Variability in local settlement patterns explain the larger number of observations for these time periods because they are more frequently represented in local sites (Table 1; Jerardino et al., 2013). WWSP abundances decrease markedly as distances between coastline and shell middens increase. In general, and independently of the time period considered, the longer the distance people had to walk back between coastal foraging locations and their campsites, the less quantities of WWSP arrived there. This trend is best reflected among the younger samples for which more observations are available (Fig. 7). As with other trends described above, some measure of variability is apparent, and WWSP abundances seem to reach minimum values with distances around and beyond 1 km. 6. Discussion WWSP weights (standardized by kg of shell sample) are much lower in shell samples where mussels are infrequent when compared to those where this species is present in high numbers. These observations confirm those derived from sampling of modern mussel beds (Yates, 1989) that identify mussel byssus threads as the natural containers or ‘traps’ of these sediments. Consequently, WWSP were most likely transported to the sites as part of the byssus contents of mussels collected from nearby rocky reefs. The high WWSP abundances registered in local shell middens between 6000 and 3000 cal. BP coincide with a period of marked environmental changes along the entire South African West Coast. Locally, the mid-Holocene 2e3 m high sea level stand (Fig. 2: 7500e6000 cal. BP) brought coastal sediment erosion and the opening of Verlorenvlei mouth into a sheltered estuary. The subsequent regressive event resulted in sea levels to oscillate around the present mean from 5000 cal. BP onwards, with much sediment being transported onto the shore and coastal plain through sea and wind action. It is very likely that this net onshore sediment

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

A. Jerardino / Quaternary International xxx (2015) 1e12

7

Fig. 3. Changes in the average abundances and standard errors through time of water-worn shell and pebbles (WWSP) (numbers of particles/kg of mussel) derived from shell samples recovered from archaeological sites in the Lamberts Bay and Elands Bay areas. Note that no data is available for Elands Bay Cave sediment counts. See Fig. 1 and Table 1 for details.

Fig. 4. Changes in the average abundances and standard errors through time of water-worn shell and pebbles (WWSP) (grams/kg of mussel) derived from shell samples recovered from archaeological sites in the Lamberts Bay and Elands Bay areas. See Fig. 1 and Table 1 for details.

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

8

A. Jerardino / Quaternary International xxx (2015) 1e12

Fig. 5. Changes in the average percentages and standard errors through time of water-worn shell (%WWS) from shell assemblages recovered from archaeological sites located in the Lamberts Bay and Elands Bay areas, and results on best fitting linear models (logarithmic) derived from regression analyses; a) means and standard errors calculated on counts of sediment particles, note that no data on WWSP particle counts are available for Elands Bay Cave; b) means and standard errors calculated on mass of sediment particles.

movement led to the formation of much of present day Elands Bay dune field and the dune cordon south of Baboon Point (Fig. 1a). Clearly, the chronological patterning of WWSP abundances (larger fraction of near-coastal marine sediments) mirrors the expected behaviour of the smaller sediment fraction (coarse and fine sand) as

predicted by this current model in which sea level changes and pulses in sand supply lead to the formation of dune fields (Illenberger, 1988; Compton and Franceschini, 2005; Franceschini and Compton, 2006). In other words, when sea level transgressions occur, large quantities of coastal sediment are removed

Fig. 6. Changes in the percentages and standard errors through time of water-worn shell (%WWS) from individual shell samples from archaeological sites located in the Lamberts Bay and Elands Bay areas, and results on best fitting linear models (logarithmic) derived from regression analyses; a) raw percentages (based on counts of WWSP particles) per studied sample in each site, note that no data on WWSP particle counts are available for Elands Bay Cave; b) raw percentages (based on mass of WWSP) per studied sample in each site. See Table 1 for details.

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

A. Jerardino / Quaternary International xxx (2015) 1e12

9

Fig. 7. Changes in the average abundances and standard errors of water-worn shell and pebbles (WWSP) with distance from nearest rocky shoreline for three different time periods most commonly represented at local archaeological sites. Pancho's Kitchen Midden data includes samples dated to 540 cal. BP sieved on site with 3.2 mm mesh, and samples dating to 755 cal. BP and 3805 cal. BP initially sieved with finer mesh but subsequently passed through a 3.2 mm mesh in the laboratory. Note that no data on WWSP particle counts is available for Elands Bay Cave. See Fig. 1 and Table 1 for details.

and re-worked from the shoreline into adjacent littoral waters. As a result, some of the coarser fraction of such sediments (in this case, WWSP) also becomes more abundant in intertidal environments where they become trapped among the byssi of rocky shore mussels. However, it is possible that a “saturation point” in the amount of WWSP retained by byssi would be reached beyond an unknown concentration or quantity of marine sediments. Nonetheless, WWSP records appear to be good proxies for deriving sea level history and local Holocene geomorphological processes along the South African West Coast. Moreover, while WWSP abundances diminish between 6000 and 3000 BP, these sediments were increasingly depleted of their organic component as measured by %WWS. The timing of this change coincides broadly with the reduction in overall WWSP abundances after 3000 cal. BP, although %WWS continue to shrink after this date reaching lowest values in the last 1000 years. Spring Cave and Tortoise Cave show higher values in post 1100 cal. BP samples, and high variability in the former might well be related to the small number of analysed samples (Table 1, Figs. 5 and 6). This is

so because, unfortunately, not all available post-1100 cal. BP samples from these sites included WWSP. It is unclear whether this is due to behavioural and/or depositional processes, or perhaps because some of these samples were inadvertently ‘cleaned’ from their WWSP contents during previous archaeomalacological analyses (Yates, pers. comm., 1992). It could be argued that the progressive decrease in average % WWS is related to diminishing coastal productivity whereby an overall smaller biomass of shell-producing organisms covered intertidal rocks after 3000 cal. BP. This is, however, not likely. Three different Neoglacial episodes marked Holocene local climatic variability along the South African coast in the last 4500 cal years, bringing repeatedly cooler (Cohen et al., 1992) and therefore more productive marine conditions to intertidal ecosystems. The last such Neoglacial event (Little Ice Age) took place in the last 1000 years when lowest average %WWS are recorded. Hence, some other factor is responsible for this temporal pattern. Perhaps, relict beach shell of Holocene and (Last Interglacial?) Pleistocene age became re-worked and incorporated into sediment stocks as a result of the

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

10

A. Jerardino / Quaternary International xxx (2015) 1e12

mid-Holocene high stand, and thus greater quantities of older or fossil organic sediments were added to those that were generated as a result of the normal life and death cycle of contemporary molluscs. A test of this hypothetical scenario would be to radiocarbon date some of the WWS particles from older and younger samples through accelerated mass spectrometry (AMS). If the age differences between AMS dates on WWS and their pre-3000 cal. BP shell sample of origin are much larger than those recorded for post3000 cal. BP samples, additions of re-worked relict beach shell to sediment stocks as a result of the mid-Holocene high sea levels would be supported. A general pattern of smaller quantities of WWSP in more distant sites from the coast is apparent for three different periods of time (Fig. 7). Comparison had to be established between sites dating to the same time period because quantities of WWSP do change through time, thus, chronological variability had to be factored out. The choice of these short periods is entirely dependent on their most frequent representation in local stratified sites, ultimately a function of regional settlement and demographic patterns. As more local sites are sampled and dated, more data is likely to be available for these periods and even new time slices for inter-site comparison could be established in future. Independently of the period of time under scrutiny, WWSP abundances drop half to five times at sites situated between 1 and 2 km from the coast. Hence, a larger pool of observations could also potentially enable the development of predictive drop-off curves with which to infer distances between coast and sites in the distant past when sea level and/or bathymetric data is unavailable or their resolution is not adequate. The way mussels and their byssi were transported back to campsites, along with the rest of harvested shellfish can likely explain the diminishing quantities of WWSP with distance from the coast. For relatively less WWSP to be repeatedly reflected furthest away from collection points means that either some of these sediments (along with the byssi that contained them) could have been removed before transport to campsites or these sediments were lost while walking to the camps. It is unlikely that any real benefit could be gained from removing byssi and their WWSP as they probably add little weight to the loads of shellfish carried to campsites. Moreover, processing costs (time and energy) would only increase when doing so, an expenditure that is usually minimised among foraging groups (Bird and Bliege Bird, 1997; Lupo, 2007; Thomas, 2007). Hence, people are unlikely to have cleaned mussels from their byssi and WWSP before taking them to places for their final consumption. It is more likely that the containers that were used to transport rocky shore mussels must have had interstices or gaps wide enough for WWSP to have dropped through them and thus removed from the harvested grab of molluscs. The longer the walk back to camp, the more particles of WWSP would have dropped through the gaps. World-wide ethnographic examples of containers used to transport aquatic or marine prey are baskets or netting bags made of plant fibre or sinew (e.g., Budack, 1977; Meehan, 1982; Waselkov, 1987; Verdún et al., 2010). Knotted grass, string, twine and fragments of netting mesh are known from southern African archaeological contexts dating to the last 6300 years (Grobbellaar and Goodwin, 1952; Parkington and Poeggenpoel, 1971, 1987; Deacon, 1984; Anderson, 1991; Orton et al., 2011), and nets have also been portrayed in local rock paintings (Manhire et al., 1985). Consequently, it is likely that South African West Coast Holocene foragers used netting bags for transporting their harvested shellfish away from the coastline (Fig. 8). Because the preservation of plant material in local sites is generally less than desirable, the microscopic presence (i.e., phytoliths, micro-samples of tissue) of plant species (e.g., rushes) most likely used for manufacturing such carrying devices in archaeological sediments may reveal their widespread use in the past. A logical

Fig. 8. Ethnographic example of netting bag (Papua New Guinea, image courtesy of the Queensland Museum, Brisbane, Australia, acquired on 22 Feb 1886). Bags with a similar manufacture are likely to have been used by South African West Coast foragers for transporting marine prey such as shellfish for many millennia (see text).

starting point to further research into this issue would be to consult ethnographic and ethnohistorical records as extensively as possible, finding out about the use of plant raw materials from interviews with descendants of local indigenous people that may still be familiar with basketry and related crafts, and to build a suitable plant reference collection. The present shoreline was configured by about 7500 cal. BP (Fig. 2). Any earlier shellfish foraging would have involved longer travel distances from local sites, for the shoreline lay further west from the present coastline as a result of lower sea levels. It can thus be expected that WWSP abundances predating 7500 cal. BP in a single cultural sequence would be smaller when compared to post7500 cal. BP samples if netting bags were used throughout these millennia. The only sufficiently long local cultural sequence for which such data is available is Elands Bay Cave (Fig. 1b). As expected, average WWSP abundances drop from 16.7 to 18.6 g/kg mussel between 3000 and 4800 cal. BP (Table 1) to between 5.8 and 3.2 g/kg mussel in shell-bearing strata dating to between 8700 and 11 400 cal. BP (personal observations). It is thus likely that netting bags were used locally since at least the Pleistocene/Holocene transition. 7. Conclusions This study shows that changes in WWSP abundances are useful proxies for deriving broad sea level reconstructions and Holocene geomorphological processes along the South African West Coast. Since changes in coastal sediment budget would probably affect both the finest fraction (sand) and coarser fraction of marine sediments (WWSP) in similar ways, the amount of WWSP entangled in the byssi of rocky shore mussels can be expected to be proportional to the amount of the overall sediment availability present in the nearby and contemporary marine-aquatic environment. It is proposed that within the environmental, chronological and archaeological parameters considered here, large WWSP ratios in archaeological samples indicate relatively large supplies of sediment along the nearby contemporary coastline as a result of higher sea levels, while low WWSP ratios indicate the opposite. MidHolocene higher sea levels may also explain the greater contribution of organic sediments (WWS) to WWSP samples derived from pre-3000 cal. BP, whereby older Holocene or fossil shell accumulated higher up the beaches' profiles became re-worked and included into sediment loads. A reasonably large AMS dating program of WWSP samples would help to establish possible significant input of fossil (Last Interglacial) shell to these sediments.

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

A. Jerardino / Quaternary International xxx (2015) 1e12

Independent of age, the patterning of WWSP abundances in terms of distances between archaeological sites and nearest rocky shore (harvesting points) suggests indirectly that netting bags were used to transport rocky shore mussels and probably also other species of molluscs. These types of carrying devices are common among extant and extinct coastal foragers worldwide, and might well have been in use locally since the Pleistocene/Holocene transition. With additional shell middens studied in the same way as presented here, the absence of change or lack of any consistent pattern in the abundance of WWSP with distance would suggest alternative and/or a variety of carrying devices to transport harvested shellfish in the past. Follow-up studies on modern mussel beds for ascertaining the variability of WWSP abundances along different types of shorelines and archaeomalacological studies on shell middens with similar geomorphological characteristics in southern Africa and beyond ought to further test on the suitability of WWSP as proxies for palaeoenvironmental and behavioural reconstructions. Acknowledgements I am most grateful to Royden Yates for first drawing my attention and that of others to the presence of water worn shell and pebbles in local shell middens and for his considerable help during field work at Pancho's Kitchen Midden and Steenbokfontein Cave. I am very grateful to Debie (Lekas) Adams, Steven Lekas, Ansa Menhert, Thembi Russel and Natalie Swanepoel for their assistance with many of the shell analyses. Funding of various excavations and the analyses of samples was provided by grants from the University of Cape Town, the Centre for Science Development, Wenner-Gren (Gr. 5699, Chicago, Illinois) and the Swan Fund (Oxford). I also thank John Parkington for making Elands Bay Cave data available and for access to shell samples from various Elands Bay sites,
Navarro for Francesc Conesa for compiling Figs. 1 and 2, Rene Figs. 3e7 and statistical analyses, and Debora Zurro for palaeobotanical discussions. I am most grateful to Brit Asmussen and David Parkhill of the Queensland Museum and Sciencecentre in Brisbane, Australia, for sourcing and providing the photographic image of a netting bag (Fig. 8). Thanks are also extended to John Compton for his useful comments and suggestions on an earlier version of this paper, to Hayley Cawthra and an anonymous reviewer for their useful suggestions and bibliography. Any errors are my own. References Anderson, G.C., 1991. Andriesgrond Revisited: Material Culture, Ideologies and Social Change. Honours Dissertation. University of Cape Town. Barkai, A., Branch, G., 1988. Contrasts between the benthic communities of subtidal hard substrata at Marcus and Malgas islands: a case of alternative stable states? South African Journal of Marine Science 7 (1), 117e137. Bateman, M.D., Carr, A.S., Dunajko, A.C., Holmes, P.J., Roberts, D.L., McLaren, S.J., Bryant, R.G., Marker, M.E., Murray-Wallace, C.V., 2011. The evolution of coastal barrier systems: a case study of the Middle Late Pleistocene Wilderness barriers, South Africa. Quaternary Science Reviews 30, 63e81. Bird, D.W., Bliege Bird, R., 1997. Contemporary shellfish gathering strategies among the Merriam of the Torres Strait Islands, Australia: testing predictions of a central place foraging model. Journal of Archaeological Science 24, 39e63. Budack, K.F.R., 1977. The sAonin or Topnaar of the lower !Khuiseb valley and the sea. Khoisan Linguistic Studies 3, 1e42. Butzer, K.W., 1979. Geomorphology and geo-archaeology at Elandsbaai, Western Cape, South Africa. Catena 6, 157e166. Butzer, K.W., 2006. Archaeology as Human Ecology: Method and Theory for a Contextual Approach. Cambridge University Press, Cambridge. Branch, G.M., Griffiths, C.L., 1988. The Benguela ecosystem, part V: the coastal zone. Oceanography and Marine Biology Annual Review 26, 395e486. Chase, B.M., Meadows, M.E., 2007. Late Quaternary dynamics of southern Africa's winter rainfall zone. Earth-Science Reviews 84, 103e138. Chase, B.M., Thomas, D.S.G., 2007. Multiphase late Quaternary aeolian sediment accumulation in western South Africa: timing and relationship to

11

palaeoclimatic changes inferred from the marine record. Quaternary International 166, 29e41. Claassen, C., 1998. Shells. Cambridge Manuals in Archaeology. Cambridge University Press, Cambridge. Cohen, A.L., Parkington, J.E., Brundrit, G.B., Van der Merwe, N.J., 1992. A Holocene marine climate record in mollusc shells from the southwest African coast. Quaternary Research 38, 379e385. Compton, J.S., 2001. Holocene sea-level fluctuations inferred from the evolution of depositional environments of the southern Langebaan Lagoon salt marsh, South Africa. The Holocene 11, 395e405. Compton, J.S., 2006. The mid-Holocene sea-level highstand at Bogenfels Pan on the southwest coast of Namibia. Quaternary Research 66, 303e310. Compton, J.S., Franceschini, G., 2005. Holocene geoarchaeology of the Sixteen Mile Beach barrier dunes in the Western Cape, South Africa. Quaternary Research 63, 99e107. Deacon, J., 1984. Later Stone Age people and their descendants in southern Africa. In: Klein, R.G. (Ed.), Southern African Prehistory and Palaeoenvironments. Balkema, Rotterdam, pp. 221e328. Dewar, G., Reimer, P.J., Sealy, J., Woodborne, S., 2012. Late-Holocene marine radiocarbon reservoir correction (DR) for the west coast of South Africa. The Holocene 22 (12), 1481e1489. Faulkner, P., 2011. Quantifying shell weight loss in archaeological deposits. Archaeology in Oceania 46, 118e129. Ford, P.J., 1992. Interpreting grain size distributions of archaeological shell. In: Stein, J. (Ed.), Deciphering a Shell Midden. Academic Press, San Diego, pp. 283e325. Franceschini, G., Compton, J.S., 2006. Holocene evolution of the Sixteen Mile Beach Complex, Western Cape, South Africa. Journal of Coastal Research 22 (5), 1158e1166. Franceschini, G., Compton, J.S., Wigley, R.A., 2003. Sand transport along the Western Cape: gone with the wind? South African Journal of Science 99, 317e318. Goldberg, P., Macphail, R., 2006. Practical and Theoretical Geoarchaeology. Blackwell Science, Oxford. Grobbellaar, C.S., Goodwin, A.H.J., 1952. Report on the skeletons and implements in association with them from a cave near Bredasdorp, Cape Province. South African Archaeological Bulletin 7, 95e107. Hockey, P.A.R., Bosman, A.L., 1986. Man as an intertidal predator in Transkei: disturbance, community convergence and management of a natural food resource. Oikos 46, 3e14. Hogg, A.G., Hua, Q., Blackwell, P.G., Niu, M., Buck, C.E., Guilderson, T.P., Heaton, T.J., Palmer, J.G., Reimer, P.J., Reimer, R.W., Turney, C.S.M., Zimmerman, S.R.H., 2013. SHCal13 Southern Hemisphere Calibration, 0e50,000 years cal BP. Radiocarbon 55 (4), 1e15. Horwitz, L.K., 1979. From Materialism to Middens: a Case Study at Eland's Bay, Western Cape, South Africa. Honours Dissertation. University of Cape Town. Illenberger, W., 1988. The Holocene evolution of the Sundays estuary and adjacent coastal dunefields, Algoa Bay, South Africa. In: Dardis, G.F., Moon, B.P. (Eds.), Geomorphological Studies in Southern Africa. Balkema, Rotterdam, pp. 389e405. Jerardino, A., 1993. Mid to late Holocene sea level fluctuations: the archaeological evidence at Tortoise Cave, south Western Cape, South Africa. South African Journal of Science 89, 481e488. Jerardino, A., 1995. Late Holocene Neoglacial episodes in southern South America and southern Africa: a comparison. The Holocene 5, 361e368. Jerardino, A., 1997. Changes in shellfish species composition and mean shell size from a late-Holocene record of the west coast of southern Africa. Journal of Archaeological Science 24, 1031e1044. Jerardino, A., 1998. Excavations at Pancho's Kitchen Midden, western Cape Coast, South Africa: further observations into the megamidden period. The South African Archaeological Bulletin 53, 17e25. Jerardino, A., 2003. Precolonial settlement and subsistence along sandy beaches south of Elands Bay, west coast, South Africa. South African Archaeological Bulletin 58 (178), 53e62. Jerardino, A., 2014. Stranded rocky shore mussels and their possible procurement during prehistory on the West Coast of South Africa. Journal of Archaeological Science 49, 536e545. Jerardino, A., in press. Shell fragmentation beyond screen-size and the reconstruction of intra-site settlement patterns: a case study from the West Coast of South Africa. In: Giovas, C.M., Poteate, A.S. (Eds.), Zooarchaeology in Practice: Case Studies in Methodology and Interpretation in Archaeofaunal Analysis. Springer. Jerardino, A., Klein, R.G., Navarro, R., Orton, J., Horwitz, L.K., 2013. Settlement and subsistence patterns since the terminal Pleistocene in the Elands Bay and Lamberts Bay areas. In: Jerardino, A., Malan, A., Braun, D. (Eds.), The Archaeology of the West Coast of South Africa. Archaeopress, Oxford, pp. 85e108. Jerardino, A., Swanepoel, N., 1999. Painted Slabs from Steenbokfontein Cave: the oldest known parietal art in southern Africa. Current Anthropology 40, 542e548. Klein, R.G., Cruz-Uribe, K., 1987. Large mammal and tortoise bones from Eland's Bay Cave and nearby sites, Western Cape Province, South Africa. In: Parkington, J.E., Hall, M. (Eds.), Papers in the Prehistory of the Western Cape, South Africa. British Archaeological Reports International Series 332 (i)Oxford, pp. 132e163. Kyle, R., Pearson, B., Fielding, P.J., Robertson, W.D., Birnie, S.L., 1997. Subsistence shellfish harvesting in the Maputaland marine reserve in northern Kwazulu-

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057

12

A. Jerardino / Quaternary International xxx (2015) 1e12

Natal, South Africa: rocky shore organisms. Biological Conservation 82, 183e192. Lupo, K., 2007. Evolutionary foraging models in zooarchaeological analysis: recent applications and future challenges. Journal of Archaeological Research 15, 143e189. Manhire, A., Parkington, J., Yates, R., 1985. Nets and fully recurved bows: rock paintings and hunting methods in the Western Cape, South Africa. World Archaeology 17 (2), 161e174. Meehan, B., 1982. Shell Bed to Shell Midden. Australian Institute of Aboriginal Studies, Atlantic Highlands. Miller, D.E., 1987. Geoarchaeology at Verlorenvlei. In: Parkington, J.E., Hall, M. (Eds.), Papers in the Prehistory of the Western Cape, South Africa. British Archaeological Reports International Series 332 (i)Oxford, pp. 46e77. Miller, D.E., Yates, R.J., Jerardino, A., Parkington, J.E., 1995. Late Holocene coastal change in the southwestern Cape, South Africa. Quaternary International 29/30, 3e10. Miller, D.E., Yates, R.J., Parkington, J.E., Vogel, J.C., 1993. Radiocarbon-dated evidence relating to a mid-Holocene relative high sea-level on the south-western Cape coast, South Africa. South African Journal of Science 89, 35e44. Mowat, F.M., 1994. Size really does matter: factors affecting shell fragmentation. In: Sullivan, M., Brockwell, S., Webb, A. (Eds.), Archaeology in the North: Proceedings of the 1993 Australian Archaeological Association Conference. The Australian National University, Darwin, pp. 201e210. Orton, J., Compton, J.S., 2006. A reworked mid-Holocene lithic assemblage at Dunefield Midden 1, Elands Bay, South Africa. South African Archaeological Bulletin 61, 90e95. Orton, J., Klein, R.G., Mackay, A., Schwortz, S., Steele, T.E., 2011. Two Holocene rock shelter deposits from the Knersvlakte, southern Namaqualand, South Africa. Southern African Humanities 23, 109e150. Parkington, J.E., 1988. The Pleistocene/Holocene transition in the western Cape, South Africa: observations from Verlorenvlei. In: Bower, J., Lubell, D. (Eds.), Prehistoric Cultures and Environments in the Late Quaternary of Africa. British Archaeological Reports International Series 405Oxford, pp. 197e206. Parkington, J., Fisher, J.W., Poggenpoel, C., Kyriacou, K., 2014. Strandloping as a resource-gathering strategy in the Cape, South African Holocene Later Stone Age: the Verloren Vlei record. Journal of Island and Coastal Archaeology 9 (2), 219e237. Parkington, J.E., Poggenpoel, C.A., 1971. Excavations at De Hangen 1968. South African Archaeological Bulletin 26, 3e36. Parkington, J.E., Poggenpoel, C.A., 1987. Diepkloof Rock Shelter. In: Parkington, J.E., Hall, M. (Eds.), Papers in the Prehistory of the Western Cape, South Africa. British Archaeological Reports International Series 332(ii)Oxford, pp. 269e293. R Development Core Team, 2012. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. http://www. Rproject.org/.

Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C.,
, C., Heaton, T.J., Grootes, P.M., Guilderson, T.P., Haflidason, H., Hajdas, I., Hatte Hoffmann, D.L., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Staff, R.A., Turney, C.S.M., van der Plicht, J., 2013. IntCal13 and Marine13 Radiocarbon Age Calibration Curves 0e50,000 years cal BP. Radiocarbon 55 (4), 1869e1887. Roberts, D.L., Bateman, M.D., Murray-Wallace, C.V., Carr, A.S., Holmes, P.J., 2009. West coast dune plumes: climate driven contrasts in dunefield morphogenesis along the western and southern South African coasts. Palaeogeography, Palaeoclimatology, Palaeoecology 271, 24e38. Roberts, D.L., Botha, G.A., Maud, R.R., Pether, J., 2006. Coastal Cenozoic deposits. In: Johnson, M.R., Annhauser, C.R., Thomas, R.J. (Eds.), The Geology of South Africa. Geological Society of South Africa/Council for Geoscience, Pretoria, pp. 605e628. Robey, T., 1987. The stratigraphic and cultural sequence at Tortoise Cave, Verlorenvlei. In: Parkington, J.E., Hall, M. (Eds.), Papers in the Prehistory of the Western Cape, South Africa. British Archaeological Reports International Series 332 (ii)Oxford, pp. 294e325. Rogers, J., 1987. The evolution of the continental terrace between St Helena Bay and Lambert's Bay. In: Parkington, J.E., Hall, M. (Eds.), Papers in the Prehistory of the Western Cape, South Africa. British Archaeological Reports International Series 332 (i)Oxford, pp. 35e45. Shannon, L.V., 1985. The Benguela ecosystem part I: evolution of the Benguela, physical features and processes. Oceanography and Marine Biology Annual Review 23, 105e183. Shannon, L.V., 1989. The physical environment. In: Payne, A.I.L., Crawford, R.J.M. (Eds.), Oceans of Life off Southern Africa. Vlaeberg Publishers, Cape Town, pp. 12e27. Stein, J., 1992. Deciphering a Shell Midden. Academic Press, San Diego. Thomas, F.R., 2007. The behavioral ecology of shellfish gathering in western Kiribati, Micronesia 1: prey Choice. Human Ecology 35, 179e194. Tyson, P.D., Lindesay, J.A., 1992. The climate of the last 2000 years in southern Africa. The Holocene 2, 271e278.
s, E., Colonese, A.C., Este
vez, J., Zurro, D., 2010. MetodoVerdún, E., Briz, I., Camaro
n y ana
lisis de concheros: experiencias acumuladas despue
s logía de excavacio ~ os de estudios etnoarqueolo
gicos en la costa norte del Canal Beagle de 20 an
rvedes 6, 25e32. (Tierra del Fuego, Argentina). Fe Waselkov, G.A., 1987. Shellfish gathering and shell midden archaeology. Advances in Archaeological Method and Theory 10, 93e210. Yates, R., 1989. Mega- and Minor Middens: Shellfish and Subsistence over Four Thousand Years on the West Coast, South Africa (unpublished report). Department of Archaeology, University of Cape Town. Yates, R.J., Miller, D.E., Halkett, D.J., Manhire, A.H., Parkington, J.E., Vogel, J.C., 1986. A late mid-Holocene high sea-level: a preliminary report on geo-archaeology at Elands Bay, western Cape Province, South Africa. South African Journal of Science 82, 164e165.

Please cite this article in press as: Jerardino, A., Water-worn shell and pebbles in shell middens as proxies of palaeoenvironmental reconstruction, shellfish procurement and their transport: A case study from the West Coast of South Africa, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.06.057