Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, ANTHROPOLOGY ( (c) Oxford University Press USA, 2020. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 26 October 2020

Shell Middens and Coastal Prehistory

Abstract and Keywords

Shell middens, the residues of shellfish gathering, consumption, and disposal in the past, have attracted the attention of archaeologists for more than one hundred and fifty years. Although there has been a tendency to view these sites as simply waste heaps, it is increasingly clear that this is usually not the case and that, sometimes, spatially meaningful arrangements of domestic debris of all kinds (fireplaces, artifacts, cooking and sleeping areas) are recognizable if excavations are sensitive enough. Some issues are as relevant and as intransigent as they have been from the beginning: Are they really food waste or could they be natural shell accumulations? Were people living at these sites or are they simply large piles of waste resulting from shell processing? In what ways and how fast did the middens accumulate? How are shell middens related to other archaeological sites inland, contemporary but without shell food waste? Because shell middens are found on all continents except Antarctica and throughout the Holocene time period (the last twelve thousand years), the literature on their excavation and interpretation is enormous and illustrates that archaeologists worldwide engage similarly with counting, measuring, weighing the shellfish, and associated faunal and artifactual remains from these sites. Often, the research involves developing proxies for the kinds of invisible but interesting aspects of the lives of the shellfish gatherers, such as: How many people lived here? How long did people stay at this site? Why did they come when they did and leave when they did? Where else did people live? While Holocene shell middens are ubiquitous, it is also clear that Pleistocene shell middens, while fairly widespread, are found more commonly in coastal areas where early modern humans have dispersed early in their migrations across the globe. It is likely that these traces, in Africa, in Europe, in island South-East Asia and Australia, and along the shores of western North America mark the routes whereby our earliest modern human ancestors peopled the world.

Keywords: shell middens, archaeo-malacology, deposition rates, seasonality, site function, very large shell mounds, processing, quantification


Coastal shell middens, that is, archaeological deposits that are visually and materially dominated by discarded marine shells (shell matrix deposits in another terminology; Hausmann et al. 2019; Villagran 2019), are valuable opportunities for reconstructing relationships between hunter-gatherers, landscapes, and resources (Claassen 1998; Erlandson 2001; Parkington 2006; Shiner et al. 2013; Stein 1992; Waselkov 1987). Shellfish remains in these middens are easy to sample (there are lots of them), easy to identify to species (shell forms are quite distinctive even when fragmented), easy to provenance on the shore (rocky or sandy substrate; upper intertidal, lower intertidal, subtidal), relatively easy to locate on the terrestrial landscape, easy to count, measure, weigh, and turn into kilojoule or macronutrient contributions to diets; and do not come with the many complicated post-collection issues that plague the understanding of large-bodied, variably useful, hard-to-transport food parcels. This is not to say they are straightforward to analyze, as shellfish are heavy when freshly mass-collected and may need to be field processed to reduce weight before transport (Metcalfe and Barlow 1992), leading to a potential disjunction between shell discard locations and flesh consumption places.

It is no wonder, then, that shell middens have attracted the attention of archaeologists the world over and continue to do so. Ertebolle is almost as famous as Olduvai Gorge or Mount Carmel to a Stone Age archaeologist. For a hundred years after the first historic excavations in Danish kitchen middens, it appeared that shellfish gathering was limited to Holocene times, with the main issue there being the relationship between Mesolithic shell-fishing hunter-gatherers and incoming Neolithic farmers (Andersen 2008). In a neat parallel to the archaeology of the Western Cape, Andersen explained, “During an excavation in the Krabbesholm 1 shell midden in 1889, it was observed that the cultural layer consisted of two different horizons: a lower one dominated by oysters and an upper layer characterized by cockles, ash and ‘burned stones’ (potboilers). The deepest layer contained thick-walled and undecorated pottery from pointed-bottomed vessels, while the upper horizon only had thin-walled, nicely decorated sherds. Today [this word is in a quoted extract]—in retrospect—we know that this sequence is a nice example of a typical Danish, stratified shell midden with a succession of Late Mesolithic Ertebolle culture (bottom) covered by a horizon of Early Neolithic Funnel Beaker culture (top). In other words, the Krabbesholm 1 kokkenmodding [kitchen midden] demonstrates a stratigraphic sequence covering the transition from the Mesolithic to the Neolithic in Denmark” (Andersen 2008, 67).

However, by the late 20th century, it had become clear that shell middens were not restricted to Holocene times but that, prior to the lowered sea levels of the Last Glacial Maximum, there had been later Pleistocene, likely Marine Isotope Stage 5 (MIS5), shell fishing at least in southern Africa (Singer and Wymer 1982; Volman 1978). Because some of the earliest modern-looking human skeletal remains came from these early shell middens, archaeologists began to ask about the relationship between encephalized early modern people and shellfish gathering and marine food consumption in general (Klein et al. 2004; Marean et al. 2007; Parkington 2001, 2010). Soon after this, it became clear that shell fishing had been widely practiced as early as MIS5 in and around the Mediterranean as well, in what appears to be a symmetrical pair of developments at the winter rainfall extremes of the African continent (Colonese et al. 2011; Cortés-Sánchez et al. 2011; Marean et al. 2007; Steele and Alvarez-Fernandez 2011). Given the association of Neanderthal populations with the early Mediterranean marine food remains at Gibraltar (Stringer et al. 2000, 2008), and perhaps elsewhere (Stiner 1994), it may be the case that two groups of early Late Pleistocene humans (Neanderthals and Cro-Magnons) were shifting into more extensive marine food exploitation (but see Klein and Steele 2008).

The relationship between marine fatty acid intake and increased intelligence remains a topic of interest (Eaton et al. 1998; Cordain et al. 2005; Kyriacou et al. 2016; Parkington 2003; Speth 2010), but research has expanded to include the role of near-shore shell fishing in facilitating the expansion of modern humans “coasting out of Africa” (Stringer et al. 2000), around the Indian Ocean (Armitage et al. 2011; Balme et al. 2009), into the Australian continent, as well as around the northern Pacific rim (Erlandson et al. 2015) and southwards through the Americas (Erlandson 2010). These narratives are often based on a view that human movements along shorelines, even when they encounter different species, were easier than those that crossed terrestrial ecosystem boundaries, with presumed greater resource differences.

Surviving shell middens, however, are almost all Holocene in age, with most research in shell midden studies practiced by archaeologists with strong Holocene landscape interests (Antczak and Cipriani 2008; Bailey et al. 1988; Bicho et al. 2011; Erlandson 2010; Estévez et al. 2001; Holdaway et al. 2017; Jerardino 2016b; Orquera et al. 2011). Many, but certainly not all, of these field programs are underway in postcolonial contexts where the shell middens were built up by precolonial hunting and gathering communities heavily impacted and dispossessed by the colonizing powers. In these circumstances (California, Canadian west coast, Northern Australia, New Zealand, South Africa), as well as others (Patagonia, Brazil, Scandinavia, Scotland, Spain, Portugal, North Africa, the Red Sea, and elsewhere), methods are similarly oriented toward integrating detailed and quantitative analyses of midden contents with models of coastal resource exploitation drawn from optimal foraging theory or relevant historical or ethnographic records. Shell midden archaeology is extremely widespread, the published literature voluminous, but the methods and approaches are fairly universal.

Informative though they are, shell middens still have to be coaxed into answering some of the more interesting questions archaeologists would like to pose. These include: How long did this midden take to accumulate? How many people lived at the site to produce the midden? What was the impact of gathering on intertidal resources? How often and at what times of the year did the occupants visit? Where else on the landscape did the shellfish gatherers live? How often and why did they move? Why is the shell midden where it is? In this article, we review some attempts to develop indices and proxies to answer these kinds of questions by South African coastal archaeologists, although the literature from other areas is replete with similar debates. Underlying the examples that follow is the challenge of saying something interesting about past societies from the material traces that survive. What we have to work with are volumes of deposit excavated, numbers of stone and other tools, weights and identifications of terrestrial and marine animal bones, shellfish of different species recovered, amounts of charcoal or ash associated with these things, and a scattering of radiocarbon dates to provide a chronological framework.

The key to these case studies lies in the substantially anthropogenic origin of both the depositional matrix (gathered and discarded shell) and contained materials (animal bones, ornaments and artifacts of stone, bone and other materials, charcoals and other combustion debris). This has encouraged shell midden archaeologists, more so than archaeologists dealing with inland depositional circumstances that are more substantially geogenic, to develop proxies that supposedly inform on the otherwise unquantifiable dimensions of prehistoric settlement (length of stay, timing of stay, number of occupants, function of visit, and other details). Such objectives often spawn debates around “time spans,” “durations,” and “rates of discard,” leading to “occupational intensity,” “intensification,” or even “population increase” when linked to sets of radiocarbon dates. The indices reveal numbers that vary between depositional units, between time frames, and between different locations that need to be assessed for behavioral (often assumed to be the same as statistical) significance. Ideally, a requirement met in some regional studies, but not all, excavations at multiple but neighboring places allows an analyst to compare indices among locations, making the testing of competing hypotheses possible.

Here are some specific studies that integrate general themes, for which we use the South African examples as a springboard.

The Problem with Density Values

In their analyses of shell fishing and shell midden construction in the Saloum Delta, Senegal, Hardy and colleagues (2016) have highlighted the integration of observations on midden size, midden location, species diversity, and the lack of substantial domestic behavioral waste as signs that accumulations may have resulted from trade or exchange rather than immediate consumption, noting that “today, shellfish collecting has little direct link to the local diet” (2016, 22). “It appears very likely that the large single species deposits in shell middens are linked to trade rather than subsistence, particularly where these accumulate relatively rapidly” (Hardy et al. 2016, 31). The question remains how do we evaluate “relatively rapidly” and how do we estimate the rate of deposition of both shell midden and embedded artifacts?

A basic issue often tackled in the analysis of shell midden contents is the appropriate method to use in comparing numbers of artifacts or weights of food waste remains from different midden units within and between sites. As Jerardino (1995, 21) puts it: “Densities offer limited interpretative power when rates of deposition in small-scale excavations are unknown and only poor chronological control is available.” Our objectives, she argues, would be served not by calculating the densities of items but “better, by estimating their total number present in different deposits and discarded within a period [my emphasis]. These estimates can be calculated by multiplying the densities of artifacts and faunal remains (n/vol) by the rate at which the corresponding deposits were accumulated (vol/unit time). Numbers of items per unit of time and not per unit of volume is the appropriate, and here preferred, unit of comparison” (Jerardino 1995, 23). The key difference is between a density, measured against volume, and a rate, measured against time. “Here” in this case is Tortoise Cave.

Tortoise Cave is a small, low rock shelter about 3.5 km upstream of the mouth of the Verlorenvlei, close to its southern bank near the Cape Atlantic shore (see Figure 1). At the time of excavation in the early 1980s, it had a relatively modest area of deposit under the roof of the shelter (about 18 m2) that is almost too low to stand up in (1.65 m roof to bedrock) and would shelter only a handful of people. Outside this shelter is a far more substantial talus, perhaps some 500 m2, of shell midden spreading downslope, nearly half a meter thick for ten meters, then trickling less intensively down further under the influence of gravity. First excavated and analyzed by Tim Robey as a master’s degree project (Robey 1984, 1987), the contents of the site were thoroughly re-examined by Antonieta Jerardino as a part of her doctoral dissertation (Jerardino 1996). The associated radiocarbon dates show that deposition here, as elsewhere in the Holocene Western Cape, and likely in all coastal circumstances, was episodic: probably short visits separated by periods of absence of variable lengths (Parkington 2016).

Shell Middens and Coastal Prehistory

Figure 1. The distribution of coastal shell middens along the Atlantic coast of South Africa from south of the Verlorenvlei River mouth to the mouth of the Olifants River. The sites and locations mentioned in the text are abbreviated on the map as follows: Dunefield Midden (DFM), Diepkloof Rock Shelter (DRS), Elands Bay Cave (EBC), Grootrif G (GDG), Mike Taylor’s Midden (MTM), Pancho’s Kitchen Midden (PKM), Steenbokfontein (SBF), Scorpion Shelter (SS), Tortoise Cave (TC), Verlorenvlei (VV). Produced by the authors, using ESRI free map data and the QGIS Open Source Geographic Information System.

The key to turning density into discard rate is having a secure knowledge of the amount of time it has taken for a particular depositional volume to accumulate (occupation time), which is not the same as knowing the amount of time between accumulations (elapsed time). Probably not the easiest site to start this kind of initiative, Tortoise Cave has a particularly patchy and disturbed depositional history with at least three very obvious hiatuses and a good deal of burrowing and repositioning of deposit by the occupants (Robey 1984). There are sixteen radiocarbon dates deemed reliable by the analysts, falling into groups at 760, 1580 to 1800, 3160 to 4020, 4200 to 4330, and beyond 6900 years ago. There appears to be no occupation deposit between 760 and 1580 years ago, none between 1800 and 3160 years ago, and none between 4330 and 6900 years ago, although many smaller intervals without visits also seem likely. Under these episodic circumstances, the question becomes: how could we hope to estimate the duration of visits, not, of course, the amount of elapsed time that produced the artifacts and food waste in the depositional episodes?

Leaving aside the earlier Holocene events, Jerardino (1995, 24) defines three episodes of occupation (760 to 1800, 3160 to 4020, and 4190 to 4330), each located on a different part of the depositional pile (cave, talus, outer cave), and presents density values by episode for various items (ostrich eggshell [oes], tortoise bone, mammal bone, fish, stone artifacts, unfinished oes beads, and finished beads and pendants). Next, “the rates of deposition were calculated for each of the episodes of occupation at Tortoise Cave with the average depth of the corresponding deposits observed in section drawings (Robey 1984, own data) and the radiocarbon time spans (Table 4a)” (Jerardino 1995, 24). This is a cavalier approach to the significance and application of radiocarbon dates in that it appears to understand “time span” as the subtraction of one date from another (compare the approaches of Hausmann et al. 2019; Holdaway et al. 2017; Shiner et al. 2013). Radiocarbon dates, at best, indicate “moments in time,” approximate temporal pegs when people leaving the dated charcoal or shell were present, but surely cannot inform on the intervals between dated visits or the duration of visits. The difference between 760 and 1800 cannot be considered a time span for the episodic depositional events of the inner cave and is more likely a record of absence than one of presence. At best, it is an indication of elapsed time but not of occupational time, nor is it likely to be an unproblematic indication of depositional rate.

Using these figures, the deposits of cave, talus, and outer cave are then ascribed “rates,” which are the estimated volumes divided by the “calculated” time spans. The weights of oes, tortoise bone, mammal bone and fish, and the numbers of stone tools, unfinished oes beads, and finished beads and pendants can then be turned into rates of discard “expressed as mass or number per 100 years” (Jerardino 1995, 25, tables 5 and 6). As the author correctly but ironically observes, “the reconstruction of site formation processes at Tortoise Cave vary enormously when inferences are drawn from density values or from rates of discard and areal extent of settlement” (Jerardino 1995, 25). The problem is that dividing an estimate of volume by a poorly conceived time span generates a meaningless value so that the “enormous variance” is not in a helpful direction in which to go.

This appears to misunderstand what a radiocarbon date is and how it can be used and reflects a poor appreciation of the episodic nature of successive visits to sites. The recognizable units we distinguish and excavate separately in these cave shell middens are probably deposited in days or weeks, less likely months, rather than the decades or centuries that are accessible to a radiocarbon date. We actually have no way of knowing how many visits of whatever duration took place between two stratigraphically superimposed radiocarbon dates. As the issue of interest is the rate of discard of bones, stone tools, or other cultural items, it is obviously essential to relate them to the time it took for the containing deposit to accumulate on one or more visit(s). This is not captured by the subtraction of radiocarbon dates and may even be unknowable. What is knowable is the relative discard rate of objects and food waste from a shared depositional volume, as we show. The result of this inadmissible transformation from density (number or weight per unit volume) to rate (number or weight per unit time) is not inconsequential as it provides a proxy for measuring behavioral change that will become entangled with others; in this case in support of a hypothesis of population increase.

The Relative Lengths of Visits

Can we distinguish shorter from longer visits to a site or tell visits by small groups from visits by larger ones? In pursuit of these goals, Jerardino (1995, 21) has written, “It is suggested that, when evaluated interactively, three parameters (area of settlement, rates of accumulation of unfinished ostrich eggshell beads and finished beads and pendants, as well as rates of accumulation of domestic debris) can provide useful insights as to how densities were generated in archaeological contexts. Each indicator can provide meaningful estimates of the relative size of visiting groups, time spent at the site and intensity of site occupation.” The solution, essentially drawn from Yellen’s (1977) ethno-archaeological study of !Kung communities of the Kalahari, is to devise reliable measures of comparison to throw light on these interesting aspects of site history. A higher rate of oes bead discard is taken as a proxy for “occupational intensity” or “residential permanence,” effectively a measure of length of stay, although numbers of occupants seems also to be implied (Jerardino 2018, 161–162).

Yellen is an understandable choice in the search for the “most appropriate ethnographic parallel” (Jerardino 1995, 22), as he himself noted that “it is more toward archaeological methodology that this book is directed” (Yellen 1977, xii). He attempts in Chapter 4 to “examine the relationship among three variables: season, the length of time a camp is occupied (for that part of the year when more than one option is available), and number of occupants” (Yellen 1977, 67). His sample is a set of sixteen camp occupations by two brothers and their families recorded between January and June 1968. Using Yellen’s observations, Jerardino and Yates propose (1996, 11) the following: “the manufacture of OES beads is a very time-consuming task and it is likely that such an activity was undertaken during longer occupations at a site,” and that “the rate of deposition of unfinished OES beads and unfinished beads and pendants can be used as an index of the relative length of visits at any stratified site.” Because this is a potential material proxy for an interesting piece of information, we need to examine its credentials carefully.

Throughout Chapter 4 of his book, Yellen makes much of the difference in patterning between subsistence and manufacturing activities, often referring to “difficulties” in predicting the undertaking of manufacturing activities from the nature or context of a campsite (Yellen 1977, 73, 76, 81, 82). The problem arises, he says, because

without exception, the raw materials—dried animal skins, pieces of bone and ostrich eggshell, sinew and vegetable fibre for twine, wood for spear shafts, bows, mortars and pestles and the like—are light and even in their unaltered form may be easily transferred. Manufactured articles often take a long time to complete; people work sporadically for brief periods and carry partially finished goods from one camp to another. Also, it strikes me that the spacing of manufacturing activities is irregular and not possible to predict; at some camps occupied by several families for more than a week, none take place.

(Yellen 1977, 76)

He does, though, offer a glimmer of hope: “I would offer only one general rule: The longer a camp is occupied, the greater the probability that any particular activity will occur there” (Yellen 1977, 77).

His conclusions on the archaeological implications of his work are worth quoting in full, as they may easily be misunderstood.

At camps occupied for only a brief period of time, the range of subsistence activities represented are likely to be “typical” for a camp located at that specific area during that particular season. Thus, inter-camp comparisons are likely to yield valid information about how particular areas were used, and differences from one camp to another are likely to reflect culturally and environmentally significant facts. As the number of man-days [sic] increases and more kinds of subsistence activities are represented, the overall picture will be drawn in finer detail but probably not be drastically altered. For manufacturing activities, however, this relationship will not hold [our emphasis]. The smaller a camp and the briefer its occupation, the more difficult it will be to predict what, if any, manufacturing activities took place. Comparisons among camps will likely give misleading results. Differences among camps will be both great and culturally meaningless. [. . .] It would be possible perhaps to identify “ostrich egg shell bead making sites,” “twine making sites,” “quiver making sites,” and so on. While these identifications would be valid, the natural tendency to leap from there to a site categorization would do little to elucidate cultural patterning. The effect would be to codify what may best be conceived as an almost random process and best be appreciated through rules of probability and chance. And the smaller the camp and the briefer the occupation, the greater the problems of comparison.

(Yellen 1977, 82–83)

Yellen’s eyewitness observations suggest that although bead making may be a lengthy and time-consuming task, it is effort likely to have been spread out across successive campsite moves and is not predictably scheduled for longer stays. As presented (Jerardino and Yates 1996, 11), the relationship between oes bead loss and duration of stay stands and falls by Yellen’s guiding framework in which there is little apparent support for their model. Once again, the key lies in the difference between a density, measured against depositional volume, and a rate, measured against time. If “rates of accumulation of unfinished oes beads and finished beads and pendants, as well as rates of accumulation of domestic debris can provide useful insights as to how densities were generated in archaeological contexts,” then they need to be reliably calculated.

Notably, the west coast shell midden sites with the lowest densities of bead making debris are the megamiddens themselves, which, by the Jerardino and Yates (1996) argument, would be the places most briefly occupied despite their massive sizes. This is in contrast to the argument that these sites reflect lengthy, semi-sedentary occupations (Jerardino 2010a, 2010b, 2012). Under these circumstances, the best use of Yellen’s observations may be to argue that all visits to megamidden locations were very short, with no attention to bead making, but very frequent, leading to massive volumes (as may have been the case in the small shoreline mounds on the Farasan Islands [see Hausmann et al. 2019]). Although longer visits to the shore and visits by larger groups produce larger midden volumes, an alternative may be regularly repeated logistical, but brief visits by groups to the same location (as Hardy et al. [2016] observed in the Saloum Delta). We now argue in support of this latter view.

Large Middens and the “Sheer Volume of Shell Deposited”: Distinguishing Domestic from Logistical Visits

Can we distinguish different kinds of visits, say domestic family residences from logistical working stays? For some years now, there has been a debate about the interpretation of a set of very large (“mega”) shell middens, all of them dating between two thousand and three thousand years ago, along the South African west coast between a few kilometers south of Elands Bay and a little north of Lamberts Bay (Henshilwood et al. 1994; Jerardino 2010a, 2012; Jerardino and Navarro 2018a, 2018b; Jerardino and Yates 1997; Parkington 2012; Parkington et al. 1988), although there may be others further south (Yates 2004). The disagreement lies in the question of whether these sites are evidence of population increase or whether they reflect the residues of shellfish processing, and one aspect of the argument hinges on the implications of the low densities of cultural items and faunal remains in the sites.

An obvious, but not necessarily definitive characteristic of these sites is size or depositional volume (Parkington 1976; Parkington et al. 1988), although, of course, they are so big we have no measurements, only estimates. For reasons that will become clear, it is not impossible to have a small megamidden, but the fact is that all those we have are extremely large. It is likely that this is partly the reason why no megamidden is located in a cave or rock shelter; all are open sites in the near-shore dune cordon. Estimates of volume vary from a thousand to more than ten thousand cubic meters (Jerardino 2010a, 2012; Jerardino and Navarro 2018a; Parkington et al. 2015), although excavations in none of them exceed five cubic meters. This very modest sampling proportion is made a little more acceptable by a second characteristic, one that is definitive. Even if early descriptions slightly exaggerated this, these sites are remarkably homogenous, with relatively little variability in content from bottom to top or across their substantial area, most clearly shown in comparison with the highly heterogeneous local cave and rock shelter shell middens nearby (Parkington 2012). Shell, ash, and charcoal dominate the depositional matrices of megamiddens, and bone and stone assemblages reflect minimal traces of human activity and may, in fact, be background noise rather than anthropogenic traces, as shown later in this section.

“Homogeneity” applies most obviously to shellfish content, where all megamiddens are completely dominated by black mussel shells, comprising more than 85 percent and usually more than 90 percent of shellfish remains (Jerardino 2010a, 2012; Jerardino and Navarro 2018a, 2018b). There are, it is true, shellfish assemblages with higher counts of limpets or other gastropods in megamidden locations, but these almost always predate three thousand years or postdate two thousand years, are above or below more extreme mussel-dominated layers, and reflect site reuse patterning by collectors outside the limits of the megamidden period. For this reason, it is important to distinguish megamidden, megamidden location, and megamidden period (form, space, and time; Spaulding 1960). This preference for mussels is, we argue, a key defining characteristic of megamiddens, as is, of course, the locational preference for an open dune swale immediately adjacent to a productive intertidal reef with abundant intertidal mussel populations.

Jerardino and Yates (1997, 50) sum up the disagreements succinctly as follows:

Megamiddens have usually been described [by those who favor the shellfish processing interpretation] as very poor in faunal and cultural remains. This is certainly true when densities of fauna and artifacts are compared with those of cave sites and rock shelters. However, a different picture emerges when comparisons are based on calculations of total mass of bone and numbers of artifacts present at the site, instead of relying on densities.

This view is repeated almost exactly word for word in later expressions of the view that megamiddens are domestic sites where some processing took place (Jerardino 2010a, 2010b, 2012, 2013; Jerardino and Navarro 2018a, 2018b) rather than logistical processing sites where people stayed briefly to process (Parkington 2012).

This former is a position that argues that the function of a site is better assessed on the basis of the total assemblage of food waste or artifact debris rather than the ratio of these to one another or to the overall volume of deposit containing them. By multiplying up the weights and numbers from small excavated volumes to large estimated site volumes, substantial evidence of animal food consumption and stone tool making is derived (Jerardino 2012, 88; 2013, 213), more substantial, it is claimed, than in nearby much smaller but obviously domestic caves and rock shelters. The “significant domestic component (stone tools and vertebrate remains) has been masked by the sheer volume of shell deposited” (Jerardino and Yates 1997, 50). Elsewhere, and more recently, “it is also clear that the sheer quantities of marine shell present at megamiddens biased their characteristic as sites nearly depleted of faunal and cultural remains” (Jerardino 2012, 2295) and “based on estimates of total mass of bones and total number of lithic artifacts, it is also likely that the sheer volume of shell deposited at other sampled megamiddens conceals the domestic component at these sites” (Jerardino 2018a, 103).

Two points need to be made here, one about the argument, the second about the evidence. The notion that site function is best understood in terms of overall quantities and that, in this case, the “sheer volume of shell” is a hindrance to analysis rather than an observation of direct relevance, is obviously wrong. We would not argue for the difference between hospital-ness and school-ness on the basis of the overall number of beds or desks but surely on the ratio between the two? Hospitals have lots of beds and fewer tables, schools the reverse (but lots of schools have lots of beds); processing sites have much more discarded shell in relation to stone tools, or hunted game animals, than do domestic sites? The “sheer volume of shell” is precisely why these megamiddens are processing sites. Compare, for example, the conclusion of Hardy and colleagues that the large Saloum middens “reflect collection principally for trade rather than food” (2016, 28). We would substitute “processing and transport” for “trade” but agree with the analytical framework.

On the evidence, we need to look at the small assemblages that are multiplied up, before comparing them with the assemblages of bone and stone from neighboring caves. One comparison may be between the assemblages from Mike Taylor’s Midden (MTM), a large megamidden 2 km south of Elands Bay Cave (EBC), and Pancho’s Kitchen Midden (PKM), about a kilometer and a half directly inland from MTM, restricting the comparison to the interval from three thousand to two thousand years ago (Jerardino 2012, 88). These occupations are more or less contemporary but exhibit massive differences, of many kinds. The MTM stone tool assemblage recovered from 3 m3 of excavated shell midden comprises thirteen pieces (two quartz chunks, one quartzite chunk, two quartz pebbles, four quartzite pebbles, and four shale pebbles). From 2.2 m3 excavated from occupations at nearby PKM came over nine hundred pieces of flaking debris, including thirty-three cores, forty-six retouched pieces, and thirty-three utilized pieces, none of which featured in the MTM assemblage. Interestingly, “pebbles,” of course, do not even appear on the listed items at PKM as they show no signs of human involvement. There were also twenty-one oes beads and shell pendants from this deposit at PKM. The faunal assemblage from the same 3 m3 at MTM consisted of seventy-six bone fragments, almost all very small (cormorant, miscellaneous bird, tortoise, small bovid, miscellaneous mammal, fish, snake, and micro-fauna). At PKM, excavators recovered over 3 kg of tortoise bone, nearly a kilogram of mammal bone (including small, small medium, medium, large medium, and large bovids), 55 g of oes, and 225 g of fish bone, all from a volume of deposit about two-thirds that of the MTM excavation. This is not a sample size issue.

The MTM “stone tool” set is, first of all, hardly an assemblage at all, as it contains no cores or worked pieces, is dominated by unretouched pebbles, and is almost certainly reflective of what is lying around on the local near-shore dune landscape. It is arguably not anthropogenic and hardly the basis for generating a four thousand-piece “total assemblage” for MTM by multiplying up. No amount of multiplying up pebbles, chips, and chunks will produce an assemblage of cores and retouched formal tools. Similarly, the list of animals recovered from the MTM excavations is dominated by ones whose bones are littered across the dunes nearby and need not be strictly “domestic” at all. Faunal remains from these sites, for example, reflect very low densities of small bovids, rodents, tortoises, snakes, dune mole rats, and (likely marine) birds. Foot surveys on the surfaces of megamiddens reveal many small bovid feces areas, dune mole rat burrows, rodent pathways, snake and rodent holes, and fragmentary bird carcasses, which may suggest that the fragmentary faunal remains in the megamiddens may, at least in part, be background noise rather than anthropogenic traces. Even if some of these animal bones do have an anthropogenic origin, they lack the clearly food waste character of the far more substantial fauna from PKM. Despite this, the “multiplied up” faunal total from MTM exceeds that of PKM (Jerardino 2012, 88, table 2; although somewhat different numbers are used in Jerardino 2013, 213, table 18.3).

A second, perhaps more salient example of the dangers of this mode of thinking is provided by the excavated results from Grootrif G, a near coastal site some 10 km south of Lamberts Bay (Jerardino 2007). At Grootrif G, a small excavation revealed a megamidden-like horizon with a radiocarbon date of 2380 ± 60 (layer 3) and on top of it a very different event with a radiocarbon date of 690 ± 40 (layer 1). The volume excavated from layer 1 was five times that of layer 3, but the differences in assemblage numbers and weights were considerably greater. The weight of tortoise bone (20123.3 g in layer 1, 31.7 g in layer 3), terrestrial mammal bone (2897.9 g in layer 1, 4.8 g in layer 3), seal bone (1965 g in layer 1, nothing in layer 3), and oes (46.9 g in layer 1, nothing in layer 3), and the numbers of rock lobster mandibles (182 in layer 1, 5 in layer 3), flaked stone (248 in layer 1, 11 in layer 3), and beads and pendants (4 in layer 1, none in layer 3) all indicate a strong domestic signal in layer 1 but not in layer 3. Again, the anthropogenic nature of the signal from layer 3 is barely visible.

It is hard to avoid the conclusion that Grootrif G layer 1 was the same “location” as, but an entirely different “place” to Grootrif G layer 3, that visits were occasioned and formed by entirely different concerns and motives, and that this is an excellent example of the difference between space and place. No amount of multiplying up can make these the same kinds of places. The “sheer quantity” of marine shell, combined with the shift from total (97 percent) mussel content in layer 3 to almost total limpet (85 percent) content in layer 1 underlines the fact that shell matrix is not a masking but an informing context. The earlier site function was as a processing station for mussels; the later site function was as domestic campsite, where limpets were preferred over mussels and where a variety of manufacturing and subsistence activities were scheduled.

Huge volumes of shelly waste, composed almost exclusively of the shells of a single species discarded within a few hundred meters of the source of the collected mollusks (see Figures 2 and 3), and with minimal food waste and artifactual content (Jerardino and Navarro 2018a, 2018b), are the repeated characteristics of processing sites as reported and described in the Farasan Islands (Hausmann et al. 2019) and the Saloum Delta (Hardy et al. 2016). It is hard to imagine a better signature of processing rather than domestic activity in the millennium from three thousand to two thousand years ago along the Cape west coast. The millennia before and after, in these same diagrams, offer a contrasting, more likely domestic-driven signature.

Shell Middens and Coastal Prehistory

Figure 2. Distribution of shell waste volumes before, during, and after the period of megamidden usage.

Created by the authors.

Shell Middens and Coastal Prehistory

Figure 3. Mussel and limpet contributions to shellfish waste before, during, and after megamidden usage.

Created by the authors.

Large Middens and Shellfish Consumption

The impact of stable carbon isotope readings from human skeletal remains on reconstructions of coastal settlement, especially the possible seasonal use of nearshore marine resources, has been substantial (Sealy et al. 1986) and the responses of archaeologists no less so (Milner et al. 2004; Parkington 1991). Given the currently available advice on how to read these isotope results (Lewis and Sealy 2018), it is important to reconcile excavated evidence for marine food consumption at particular sites with the analytical results of collagen signals in associated individuals. Recently (Colonese et al. 2009, 2012; Hausmann and Meredith-Williams 2017; Loftus et al. 2019), we have been able to add another signal that helps us understand past human behavior at the coast. Sophisticated measurements of oxygen isotope ratios, viewed in the context of locally established seawater temperature changes, offer an insight into shellfish use from the perspective of the prey rather than the predator. The recovery of fairly resolved seasons of death, and therefore collection, of shellfish from archaeological sites allows us to set measures of site food use from faunal remains, measures of marine protein intake among collectors, and measures of gastropod death together, and with mutually significant implications.

In our attempt to understand coastal resource use along the Cape west coast, we are fortunate to have two signals of shellfish, or at least marine food, consumption: the piles of shellfish remains we refer to as shell middens, and the stable carbon isotope readings from human skeletal remains. These signals refer to different aspects of food consumption: the former is a reflection of group movements, landscape preference, and site-specific behavior; the latter is an indication of differences in the ratios of marine and terrestrial proteins consumed by individual hunter-gatherers. The contrast between a site and an individual focus offers interesting possibilities for hypothesis testing. Although the shell midden sample is very good, some 215 localities over an 80 km stretch of shoreline, that of individuals analyzed isotopically is not, only about twenty-five individuals from 4000 km2 of landscape.

A graphic used several times (Jerardino 2010a, 2295; 2010b, 29; 2012, 84; 2013, 217; 2016a, 88; Jerardino et al. 2008, 283; 2013, 98) refers to the stable isotope readings that measure marine protein intake in local human skeletal remains. Using observations published some time ago (Lee-Thorp et al. 1989; Sealy 1989; Sealy and Van der Merwe 1988), Jerardino argues that the figures on the whole “show an increase in the intake of protein and energy-rich foodstuffs from marine origin during the megamidden period” (2010a, 2296–2297). The implication is that the isotope values support the hypothesis that these very large middens are domestic locations at which larger populations consumed very large numbers of shellfish, almost exclusively mussels.

Recently, a larger sample of stable carbon isotopic readings on human skeletons from the past five thousand years, mostly buried at coastal locations, reveals a rather different picture (Parkington 2012), although it does seem clear that the graphic illustrates a good deal of overlap between pre-, post-, and megamidden-aged values. Stable carbon isotope analysis of bone collagen allows us to estimate the relative amounts of terrestrial and marine protein in the diet of an individual assessed (Sealy and van der Merwe 1986), but see the caution expressed in Lewis and Sealy (2018). Without going into detail, all of which is available elsewhere (Dewar and Pfeiffer 2010; Parkington 2012; Sealy 1989), we use here the empirical endpoints of -19‰ for a completely terrestrial protein intake and -9‰ for a completely marine protein intake.

The stable carbon isotope readings for the six individuals measured who could be considered contemporary with megamidden accumulation, and who are all buried within 5 km of them, are -11.8‰, -13.9‰, -12.8‰, ­16.6‰, -13.6‰, and -15.5‰. This means that these people averaged about -14.0‰, right in the middle of the proposed scale, meaning a roughly 50:50 mix of marine and terrestrial proteins. Given that the food waste composition of the millennium from three thousand to two thousand years ago comprises almost entirely (about 95 percent) megamidden, made up almost entirely (again about 95 percent) of mussel flesh, extreme enriched marine values should be reflected in the consumers. The fact that they are not is a result of great moment. The values recorded from these six individuals are far from being the most “marine” of those measured along the west coast of the Cape (Parkington 2012, 1527–1528, table 2; Pfeiffer and Sealy 2006, 4, table 1). The “megamidden folk” were not consuming the diet entrenched in the megamiddens, but one that included far more terrestrial proteins. These massive piles of shell, we have argued, are not the food waste of groups living locally, but the processing heaps of groups who were based elsewhere, likely in the interior Sandveld. The flesh from the processed mussels was, we suggest (Henshilwood et al. 1994), dried and transported inland where it was shared with relatives and friends. These conclusions are similar to those reached by Hardy et al. (2016) in the Saloum Delta and Hausmann et al. (2019) in the Farasan Islands from sites they presume to be processing locations.

Jerardino and Navarro have recently (2018b) indicated that some level of processing is, indeed, reflected in the megamidden signal but have disputed the idea that dried shellfish flesh was transported inland. They have suggested rather that “reserves of dried shellfish could have been kept by coastal groups for times when shellfish collection was hampered by bad weather and strong swells, high tides or toxic red tides when molluscs are rendered poisonous as a result of blooms of a variety of microorganisms” (Jerardino and Navarro 2018b, 120). Although this is a possible scenario, it would still mean that the enormous bulk of megamidden mussel flesh was consumed by a local coastal population, a result that is not confirmed and, in fact, is contradicted by the stable carbon isotope readings.

Surveying the results for individuals measured over the whole of the Cape west coast and taking in periods before and after the megamidden millennium, it is clear that the megamidden folk exhibited a similar mean value to those elsewhere and at other times where and when megamiddens are not found (Parkington 2012; Parkington et al. 2015, 125). The presence of megamiddens, it seems, has little or no impact on the marine food intake of locally buried people, a result that can only mean that the excess flesh and its isotopic impact was being exported and shared. There is no support in the isotope readings for an interpretation of these large sites as reflective of larger populations as “hunter-gatherer mobility became increasingly restricted to the coastal margin” (Jerardino et al. 2008, 284). Rather, the isotope studies, along with other clues in these local shell middens, imply that coastal visits were brief, probably very largely scheduled for low tides when productive mussel colonies could be exploited, processed, and transported back to inland bases.

The admittedly few stable carbon isotope results from burials from Faraoskop Rock Shelter (Manhire 1993), some 30 km inland of local megamiddens, are entirely consistent with our suggestion. Faraoskop has a set of human skeletal remains for which we have not only direct radiocarbon dates but also stable carbon isotope readings. The radiocarbon suite from these human remains (Manhire 1993, 19; 2130 ± 60 [Pta 5281], 2000 ± 50 [Pta 5283], 2110 ± 70 [Pta 5284], 2150 ± 70 [Pta 4964], 2090 ± 60 [Pta 4965], and 2130 ± 45 [Pta 4967]) is almost exactly contemporary with most of those from a megamidden 2 km south of Elands Bay Cave (Jerardino and Yates 1997, 44; 2130 ± 25 [Pta 7013], 2090 ± 50 [Pta 3659], 2000 ± 25 [Pta 6690], 2070 ± 25 [Pta 6705], 2160 ± 50 [Pta 6707], 2100 ± 30 [Pta 6683], and 2220 ± 60 [Pta 6694], the last five being shell dates “corrected for the apparent age of sea water [-400 years]”).

Manhire’s (1993, 13) excavation at Faraoskop produced only from 100 to 200 g of marine shell remains per 100 buckets of deposit removed, an amount that he suggested reflected as much artifactual use as food intake. The marine shell fragments were indeed dominated by the black mussel, Choromytilus meridionalis (Manhire 1993, 13), but the numbers and densities, like those at Diepkloof and elsewhere in the Sandveld, are far short of those of a shell midden at the coast. Apart from these shells, there are no marine food items in the faunal list (Manhire 1993, 17, table 13). Clearly given the dominance of megamidden waste at this time, Faraoskop individuals were living in a “megamidden age” but were not buried at a “megamidden location” nor, from the faunal assemblage at Faraoskop, eating a megamidden diet. This opens up the possibility of assessing the consumption of marine foods including, but not exclusively, shellfish, not from shell midden debris but from isotope evidence.

We have stable carbon isotope readings for nine of the twelve Faraoskop individuals, ranging from -18.8‰ to -16.5‰, with an average for the group of -17.3‰ (Manhire 1993, 19, table 16). Comparatively, returning to our earlier established endpoints for the greater region of -19‰ for a completely terrestrial protein intake and -9‰ for a completely marine protein intake, it places these nine individuals at a point on the scale that makes it extremely likely that they consumed at least some form of marine protein. There are, it is true, assumptions being made to support this view such as the defined endpoint for what constitutes a fully terrestrial diet and, conversely, a fully marine diet. We turn to Hedges (2004, 35) for some guidance: “If the terrestrial end-point is uncertain to ± 1‰, this allows a difference of about 20 percent in deciding what is a minimum detectable level of marine consumption.” Although a rough guideline, if we then increase our generously low endpoint of -19‰ to -18‰, we still arrive at a mean carbon isotope reading for this group of individuals that warrants at least some amount of marine protein in their diet, and this amount cannot be solely attributed to the miniscule amount of marine protein represented in the deposits at Faraoskop.

This brings us to a situation which is in marked contrast to the carbon isotope readings on coastal skeletons that are contemporary with and buried close to the megamiddens (Parkington 2012, 1529). Among the coastal individuals, readings are far too depleted for the diet implied by the megamiddens (too much shell): at Faraoskop, the readings are too enriched for the diet implied by the food waste at the site (too little shell). There seem to be alternative interpretations of these opposed mismatches. Either the Faraoskop individuals occasionally visited the coast where they consumed marine protein, or the shellfish meat removed from the shells at megamiddens were taken inland and consumed by groups such as those buried at Faraoskop, or some variations of both. Because the Faraoskop individuals are less enriched than the coastal individuals, it seems likely they ate less marine protein and may have been the recipients of transported shellfish flesh. As a test of the “processing model” for megamidden debris, the Faraoskop stable carbon isotope readings are consistent, if not conclusive. Critically, the imminent addition of oxygen isotope readings on the shellfish from these Cape west coast sites will add another valuable strand of evidence for use in deciding between alternative reconstructions, as has been done many times elsewhere in the world (see Brockwell et al. 2013; Hausmann and Meredith-Williams 2017; Kennett and Voorhies 1996; Rick et al. 2006, to name a few).

Anthropogenic Deposits, Multiple Viewpoints, and Spatial Excavations

In his then comprehensive summary of research, Shellfish Gathering and Shell Midden Archaeology, Gregory Waselkov (1987, 140–141) clearly showed that many of the issues we struggle with today have characterized archaeologists’ work from the very beginning of midden research. An eerie and very early parallel to the debate played out along the Cape west coast and detailed here is the exchange between Dall (1877) and Walker (1877) over the question of the rate of midden accumulation. Whereas Dall thought he could estimate this from evidence on consumption among contemporary Aleutian shellfish gatherers, Walker responded that “the growth of a shell heap depended on the abundance or scarcity of shellfish, the regularity with which a site was reoccupied, the length of occupation, the number of people living in the vicinity, and other variables” (Waselkov 1987, 141). Needless to say, the argument rumbled on in North America (Cook 1946; Cook and Treganza 1950; Gifford 1916; Nelson 1909), in New Zealand (Davidson 1967; Lubell et al. 1976; Meighan et al. 1958; Shawcross 1967), and until recently, in Australia (Hausmann and Meredith-Williams 2017) as well as South Africa (Jerardino 2016c, table 1). Not only the issue, but the arguments too have changed little.

We conclude by exploring another long-lasting fascination of shell midden archaeologists: the implication that exploitation of shellfish can lead to an impact on molluskan stocks that is detectable in the archaeological record. Waselkov (1987, 169) noted that “several perspicacious zoologists and archaeologists noted that average shell size commonly decreased from lower to upper horizons of coastal shell middens in Japan and eastern North America (Moore 1892–1893; Morse 1879, 1882, 1925; Wyman 1875).” The suggestion that this results from (over)exploitation by human shellfish gatherers is thereafter ubiquitous, with examples throughout the history of research, across all continents with shell middens, and impinging on broader narratives of human evolution (Klein 2009, 578). Almost always the alternative explanation is that size differences reflect, rather, past changes in water temperature or some other aspect of oceanographic history (Jerardino 1993, 1997, 2017). In some cases, this latter is preferred (Sealy and Galimberti 2011).

At its simplest, this overexploitation hypothesis is a question of supply and demand (Parkington 2008). If gatherers are attracted to large individuals of a species, and if their demand outstrips the capacity of the target species to replace gathered individuals, then the sizes of shells should decrease in the record of removed and discarded shellfish. This simplicity, however, soon dissolves into a very complex situation with great variation among species and across time scales. Supply, for example, is affected by the growth rates, life span, habitat preference, and population density of prey species. Demand is similarly complicated and increased demand needs to be unpacked. Is the increase the result of population growth, more visits, longer visits, a shift in focus to a single target species, or a combination of some or all of these? These variables are well illustrated by the circumstances of the most regularly gathered mollusks along the Atlantic shore of the Western Cape, South Africa. From the mouth of the Olifants River to the Cape and from the Middle Stone Age to the recent past, three species have provided the vast bulk of molluskan shellfish for coastal gatherers: the black mussel, a filter-feeding bivalve (C. meridionalis), and two grazing gastropods, the larger granite (Cymbula granatina) and the smaller granular (Scutellastra granularis) limpets. The key contrast between mussels and limpets is that all limpets of both species dwell only in the intertidal and are therefore accessible to human predators at virtually every tide. The mussels live in the lower intertidal but are also abundant in the subtidal zone down to 20 m below sea level, where they live among the holdfasts of the kelp forests. Here, we deal with the limpets.

As we have reported previously:

From the early 1970s we have been sampling shell middens around the coastlines of the Western Cape and measuring the sizes of whole limpets in these samples (Buchanan et al. 1978; Parkington 1976). Our strategy in sampling the shell middens began as a series of grab samples from shell scatters we encountered along the shore. We simply put down a grid, chose a square meter, and identified, counted, and measured what we found in the sample area. Most of these scatters were open sites that we visited as part of general surveys of site distributions in hitherto unexplored areas. We had no idea of the spatial context of the samples in relation to site layout, rarely sampled more than two square meters in a scatter, and were almost never able to associate a sample with any remnant feature. We did, however, try to choose sample squares that included ones that were newly exposed by deflation and others that might have been substantially displaced by down-slope movement. Other than this, we assumed that the scatters were relatively uniform. Measurements and counts were done in the field and all shells were left more or less in place. In our excavations at about the same time (Parkington 1976; Robertshaw 1977, 1978), we sampled shellfish in a way not that dissimilar from that at the open sites. Choosing one of every few buckets of deposit, we classified, counted, and measured the shellfish from defined stratigraphic units. Again we assumed these units were internally uniform.

(Parkington 2008, 169)

After sampling more than five hundred stratified and surface shellfish assemblages from the later Holocene, it was clear that mean sizes of the smaller species varied between 34 mm and 42 mm and those of the larger species between 44 mm and 62 mm, with only a few means registering marginally below or above these numbers (Buchanan et al. 1978, 93; Parkington 2006, 53). It was also clear that there was a notable and sometimes statistically significant correlation between the means of the two species. Branch’s (1974) life history observations made it clear that these were the mean sizes of third-year animals and that 5-, 6-, and 7-year-old individuals were virtually absent or very rare in the archaeological samples. His work had also shown how uncommon such larger individuals were in modern limpet populations, where only a small percentage live to maximum age. Short, 10-minute forays into the intertidal areas near the sites generated mean sizes between 48 and 52 mm for the smaller species and between 71 and 78 mm for the larger. We were understandably tempted to

. . .interpret this patterning as a reflection of human impact on limpet populations, assuming the “10-minute samples” to represent unexploited (by humans at least) resources, from which large individuals could be harvested. The lower archaeological mean sizes could mean that people had eaten their way down through the less common, older individuals and had been reduced to collecting younger, smaller limpets.

(Parkington 2008, 169–170)

We had a situation not unlike those of shell midden archaeologists elsewhere in the world (Braje et al. 2007; Erlandson et al. 2008; Giovas et al. 2010, 2013; McCoy 2008; Milner et al. 2007; Swadling 1976; Zangrando et al. 2017), where we considered gathering impact the cause of reduced archaeological mean sizes. Unlike some cases, however, we could not recognize a gradual decrease (Braje et al. 2007; Milner et al. 2007) or increase (Giovas et al. 2010, 2013) in mean sizes from older to younger levels.

Despite the evidence, we were worried by the apparent time scales of these events and explanations. How, we wondered, could an impact that may take only a few years (as described in Branch 1975) be drawn out over millennia? An opportunity to seek some resolution over the timing of impact presented itself when, between 1988 and 1998, we excavated over 850 m2 of a single, apparently brief shellfish scatter at Dunefield Midden (DFM) and looked to use the spatial distribution of mean sizes of limpets as a reflection of impact (Parkington et al. 1992, Parkington, Fisher, and Tonner 2009, 2013). Because we had measured tens of thousands of limpets at DFM, we were able to look at the spatiality of mean sizes and found a pattern remarkably reminiscent of the figures from less well-contextualized surface and excavated samples. The spread of mean sizes of the two limpet species was almost exactly that of the spread of meter square samples from sampled surface sites, but now they could be viewed against the behavioral pattern of site structure. More specifically, we realized that earlier and later collecting and discarding could be recognized in discrete dumping episodes. Individual squares with low mean sizes for one species also had low mean sizes for the other and those with low mean sizes had relatively high frequencies of the smaller species (Parkington 2008), which we took to imply prey switching as targeted individuals became smaller. Squares with smaller or larger limpet means were grouped into patches (Parkington 2008, 170–171). These patterns could not be explained in terms of shifts in oceanic conditions or population growth, as only a few weeks, months at most, were reflected in the camp visit(s). The explanation had to be that earlier days of collecting resulted in larger mean sizes and later days in smaller mean sizes and higher frequencies of the smaller species. The impact was swift, as the growth rates and life tables of Branch (1974) would have predicted. This conclusion was supported when (Parkington et al. 2013) we used the spatial distributions of the seven hundred largest and smallest measured individuals from both species rather than the square means. Early and later gathered samples could be recognized in this briefly occupied campsite midden.

As we have noted:

These patterns could not have been formed unless two behavioral rules were followed by the site occupants. First, gatherers must have collected shellfish by size, similar ones on the same day. Second, each day’s collection must have been discarded in one place, or a small set of places depending on the number of collectors in the party. If either of these rules were not followed, no clustering of similar-sized individuals would accrue. The patterns are repeated, albeit in more cluttered, less resolved form due to superimposition, at other locations across the DFM site. Now, while we cannot distinguish between a strategy of gathering the smallest on the first day and the largest on the last from the reverse of this, it seems very likely that large individuals were targeted first. It is almost impossible to imagine how the reverse strategy could have been coordinated. The patterning we discern at DFM would not be visible, and thus the conclusion not possible, if shell middens are sampled in small excavations and if shellfish analyses are represented by occasional, small subsets from selected squares.

(e.g., Erlandson et al. 2008; Milner et al. 2007Parkington et al. 2013, 99)

This is a remarkable confirmation of presumptions about shellfish gathering and discard that cannot be viewed archaeologically but can only hypothesized from ethnographic records (Meehan 1977).

Middle Stone Age (MSA) shellfish assemblages along the Cape west coast also contain substantial numbers of Cymbula granatina and Scutellastra granularis, many more of the former than of the latter. The mean sizes of the larger C. granatina vary between 65 mm and 70 mm, midway between the Later Stone Age (LSA) means and those available from modern shorelines, which we take to mean that the MSA demand was lower than the later LSA one and that larger, older individuals could be more reliably sourced. Interestingly, the mean sizes of the smaller S. granularis are not significantly different from those of the modern means, and the frequencies of this smaller species in MSA samples are much lower than in the LSA (Parkington 2008, 172–173). All of this is consistent with the DFM expectation that as demand rises, the size of limpets taken gets smaller in both species, and some prey switching from larger to smaller species takes place. We believe the impact explanation for limpet sizes in archaeological samples at the Cape is persuasive, consistent with biological controls, applies across excavated, spatial, and sampled assemblages, and provides an underlying rationale for the long-term records from other continents (Braje et al. 2007; Erlandson et al. 2008; Giovas et al. 2010, 2013; Milner et al. 2007).

Ironically, the combination of very large shellfish numbers and very low artifact densities has been both a boon and a curse for shell midden archaeologists, tempting them to sample rather than expose. Waselkov (1987, 152), himself an ardent promoter of shell midden excavations, thought the excavation of whole middens a profligate dream. Effectively, though, the decision to excavate a large area at DFM rather than dig a small “test pit” or “test trench” has allowed us access both to the variable use of space and to an insight into resolved time. If we are right about the resolution into early and later collections discarded across a single, briefly occupied campsite, the impact on shellfish stocks can not only be discovered but can be shown to happen extremely quickly, in days or weeks. Under these circumstances, some possible explanations (population growth, changes in oceanographic conditions) can be eliminated, although others (shifts in gathering focus) cannot. Mean size changes for intertidal and therefore vulnerable species may result from the inadvertent sampling by archaeologists of early or later days in an occupation, especially when they vary between values below those of currently available samples. Decreases or increases in mean size (Erlandson et al. 2015), especially in marginal cases (Giovas et al. 2013), and especially the nondirectional shifts often reported (Zangrando et al. 2017), may simply reflect oscillations around a relatively stable but impacted circumstance. We strongly recommend that shell midden archaeologists make repeated excavations into regional landscapes and excavate spatially, rather than sampling, so as to increase their capacity to distinguish between competing hypotheses on past human shell-fishing behaviors. This endorses the call of many archaeologists to consider shell middens as far more than refuse heaps and to treat them as, in many cases, domestic campsites of great interest and integrity. They may look homogenous and boring, but often they are neither.

Further Reading

Antczak, A. T., and R. Cipriani, eds. 2008. Early Human Impact on Megamolluscs. BAR International Series 1865. Oxford: Archaeopress.Find this resource:

Bailey, G., and J. Parkington, eds. 1988. The Archaeology of Prehistoric Coastlines. Cambridge, UK: Cambridge University Press.Find this resource:

Bicho, N. F., J. A. Haws, and L. G. Davis, eds. 2011. Trekking the Shore: Changing Coastlines and the Antiquity of Coastal Settlement. New York: Springer-Verlag.Find this resource:

Claassen, C. 1998. Shells. Cambridge Manuals in Archaeology Series. Cambridge, UK: Cambridge University Press.Find this resource:

Erlandson, J. M. 2010. “Food for Thought: The Role of Coastlines and Aquatic Resources in Human Evolution.” In Human Brain Evolution: The Influence of Freshwater and Marine Food Resources, edited by S. C. Cunnaneand K. M. Stewart, 125–136. Hoboken, NJ: Wiley.Find this resource:

Jerardino, A. 2016. “On the Origins and Significance of Pleistocene Coastal Resource Use in Southern Africa with Particular Reference to Shellfish Gathering.” Journal of Anthropological Archaeology 41: 213–230.Find this resource:

Marean, C. W., M. Bar-Matthews, J. Bernatchez, E. Fisher, P. Goldberg, A. I. Herries, Z. Jacobs, A. Jerardino, P. Karkanas, T. Minichillo, and P. J. Nilssen. 2007. “Early Human Use of Marine Resources and Pigment in South Africa during the Middle Pleistocene.” Nature 449 (7164): 905–908.Find this resource:

Meehan, B. 1982. Shell Bed to Shell Midden. Canberra: Australian Institute of Aboriginal Studies.Find this resource:

Parkington, J. 2001. “Milestones: The Impact of the Systematic Exploitation of Marine Foods on Human Evolution.” In Humanity from African Naissance to Coming Millennia, edited by P. V. Tobias, M. A. Raath, J. Moggi-Cocchi, and G. A. Doyle, 327–336. Johannesburg: Wits University Press.Find this resource:

Stein, J. K. 1992. Deciphering a Shell Midden. San Diego, CA: Academic Press.Find this resource:

Waselkov, G. A. 1987. “Shellfish Gathering and Shell Midden Archaeology.” In Advances in Archaeological Method and Theory vol. 10, edited by M. B. Schiffer, 93–210. Orlando, Florida: Academic Press.Find this resource:


Andersen, S. H. 2008. “The Mesolithic–Neolithic Transition in Western Denmark Seen from a Kitchen Midden Perspective: A Survey.” Analecta Praehistorica Leidensia 40: 67–74.Find this resource:

Armitage, S. J., S. A. Jasim, A. E. Marks, A. G. Parker, V. I. Usik, and H. P. Uerpmann. 2011. “The Southern Route ‘Out of Africa’: Evidence for an Early Expansion of Modern Humans into Arabia.” Science 331 (6016): 453–456.Find this resource:

Bailey, G., J. Parkington, F. Audouze, C. Renfrew, N. Schlanger, A. Sherratt, T. Taylor, and W. Ashmore, eds., 1988. The Archaeology of Prehistoric Coastlines. Cambridge, UK: Cambridge University Press.Find this resource:

Balme, J., I. Davidson, J. McDonald, N. Stern, and P. Veth. 2009. “Symbolic Behavior and the Peopling of the Southern Arc Route to Australia.” Quaternary International 202 (1–2): 59–68.Find this resource:

Braje, T. J., D. J. Kennett, J. M. Erlandson, and B. J. Culleton. 2007. “Human Impacts on Nearshore Shellfish Taxa: A 7,000 Year Record from Santa Rosa Island, California. American Antiquity 72 (4): 735–756.Find this resource:

Branch, G. M. 1974. “The Ecology of Patella linnaeus from the Cape Peninsula, South Africa, 3. Growth-Rates.” Transactions of the Royal Society of South Africa 41 (2): 161–193.Find this resource:

Branch, G. M. 1975. “Notes on the Ecology of Patella concolor and Cellana capensis, and the Effects of Human Consumption on Limpet Populations.” African Zoology 10 (1): 75–85.Find this resource:

Brockwell, S., B. Marwick, P. Bourke, P. Faulkner, and R. Willan. 2013. “Late Holocene Climate Change and Human Behavioral Variability in the Coastal Wet-Dry Tropics of Northern Australia: Evidence from a Pilot Study of Oxygen Isotopes in Marine Bivalve Shells from Archaeological Sites.” Australian Archaeology 76 (1): 21–33.Find this resource:

Buchanan, W. F., S. L. Hall, J. Henderson, A. Olivier, J. M. Pettigrew, J. E. Parkington, and P. T. Robertshaw. 1978. “South African Archaeological Society.” South African Archaeological Bulletin 33 (127): 89–93.Find this resource:

Claassen, C. 1998. Shells. Cambridge Manuals in Archaeology Series. Cambridge, UK: Cambridge University Press.Find this resource:

Colonese, A. C., M. A. Mannino, D. B. Y. Mayer, D. A. Fa, J. C. Finlayson, D. Lubell, and M. C. Stiner. 2011. “Marine Mollusc Exploitation in Mediterranean Prehistory: An Overview.” Quaternary International 239 (1–2): 86–103.Find this resource:

Colonese, A. C., S. Troelstra, P. Ziveri, F. Martini, D. L. Vetro, and S. Tommasini. 2009. “Mesolithic Shellfish Exploitation in SW Italy: Seasonal Evidence from the Oxygen Isotopic Composition of Osilinus turbinatus Shells.” Journal of Archaeological Science 36 (9): 1935–1944.Find this resource:

Colonese, A. C., E. Verdún-Castelló, M. Álvarez, I. B. Godino, D. Zurro, and L. Salvatelli. 2012. “Oxygen Isotopic Composition of Limpet Shells from the Beagle Channel: Implications for Seasonal Studies in Shell Middens of Tierra del Fuego.” Journal of Archaeological Science 39 (6): 1738–1748.Find this resource:

Cook, S. F. 1946. “A Reconsideration of Shell Mounds with Respect to Population and Nutrition.” American Antiquity 12 (1): 50–53.Find this resource:

Cook, S. F., and A. E. Treganza. 1950. The Quantitative Investigation of Indian Mounds. Berkeley, CA: University of California Press.Find this resource:

Cordain, L., S. Boyd Eaton, A. Sebastian, N. Mann, S. Lindeberg, B. A. Watkins, J. H. O’Keefe, and J. Brand-Miller. 2005. “Reply to S. C. Cunnane.” American Journal of Clinical Nutrition 82: 483–484.Find this resource:

Cortés-Sánchez, M., A. Morales-Muniz, M. D. Simón-Vallejo, M. C. Lozano-Francisco, J. L. Vera-Peláez, C. Finlayson, J. Rodríguez-Vidal, A. Delgado-Huertas, F. J. Jiménez-Espejo, F. Martínez-Ruiz, M. A. Martínez-Aguirre, A. J. Pascual-Granged, M. M. Bergadà-Zapata, J. F. Gibaja-Bao, J. A. Riquelme-Cantal, J. A. López-Sáez, M. Rodrigo-Gámiz, S. Sakai, G. Finlayson, D. A. Fa, and N. F. Bicho. 2011. “Earliest Known Use of Marine Resources by Neanderthals.” PLOS ONE 6 (9): e24026.Find this resource:

Davidson, J. M. 1967. “Midden Analysis and the Economic Approach in New Zealand Archaeology.” Records of the Auckland Institute and Museum 6 (3): 203–228.Find this resource:

Dewar, G., and S. Pfeiffer. 2010. “Approaches to Estimating Marine Protein in Human Collagen for Radiocarbon Date Calibration.” Radiocarbon 52 (4): 1611–1625.Find this resource:

Eaton, S. B., A. J. Sinclair, L. Cordain, and N. J. Mann. 1998. “Dietary Intake of Long-Chain Polyunsaturated Fatty Acids during the Paleolithic.” World Review of Nutrition and Dietetics 83: 12–23.Find this resource:

Erlandson, J. M. 2001. “The Archaeology of Aquatic Adaptations: Paradigms for a New Millennium.” Journal of Archaeological Research 9 (4): 287–350.Find this resource:

Erlandson, J. M. 2010. “Food for Thought: The Role of Coastlines and Aquatic Resources in Human Evolution.” In Human Brain Evolution: The Influence of Freshwater and Marine Food Resources, edited by Stephen C. Cunnane and Kathlyn Stewart, 125–136. Hoboken, NJ: Wiley.Find this resource:

Erlandson, J. M., T. J. Braje, K. M. Gill, and M. H. Graham. 2015. “Ecology of the Kelp Highway: Did Marine Resources Facilitate Human Dispersal from Northeast Asia to the Americas?” Journal of Island and Coastal Archaeology 10 (3): 392–411.Find this resource:

Erlandson, J. M., T. C. Rick, T. J. Braje, A. Steinberg, and R. L. Vellanoweth. 2008. “Human Impacts on Ancient Shellfish: A 10,000 Year Record from San Miguel Island, California.” Journal of Archaeological Science 35 (8): 2144–2152.Find this resource:

Estévez, J., E. Piana, A. Schiavini, and N. Juan‐Muns. 2001. “Archaeological Analysis of Shell Middens in the Beagle Channel, Tierra del Fuego Island.” International Journal of Osteoarchaeology 11 (1–2): 24–33.Find this resource:

Gifford, E. W. 1916. Composition of California Shell Mounds. University of California Publications in American Archaeology and Ethnology 12 (1): 1–29.Find this resource:

Giovas, C. M., M. Clark, S. M. Fitzpatrick, and J. Stone. 2013. “Intensifying Collection and Size Increase of the Tessellated Nerite Snail (Nerita tessellata) at the Coconut Walk Site, Nevis, Northern Lesser Antilles, AD 890–1440.” Journal of Archaeological Science 40 (11): 4024–4038.Find this resource:

Giovas, C. M., S. M. Fitzpatrick, M. Clark, and M. Abed. 2010. “Evidence for Size Increase in an Exploited Mollusc: Humped Conch (Strombus gibberulus) at Chelechol ra Orrak, Palau from ca. 3000–0 BP.” Journal of Archaeological Science 37 (11): 2788–2798.Find this resource:

Hardy, K., A. Camara, R. Piqué, E. Dioh, M. Guèye, H. D. Diadhiou, M. Faye, and M. Carré. 2016. “Shellfishing and Shell Midden Construction in the Saloum Delta, Senegal.” Journal of Anthropological Archaeology 41: 19–32.Find this resource:

Hausmann, N., and M. Meredith-Williams. 2017. “Seasonal Patterns of Coastal Exploitation on the Farasan Islands, Saudi Arabia.” Journal of Island and Coastal Archaeology 12 (3): 360–379.Find this resource:

Hausmann, N., M. Meredith-Williams, K. Douka, R. H. Inglis, and G. Bailey. 2019. “Quantifying Spatial Variability in Shell Midden Formation in the Farasan Islands, Saudi Arabia.” PLOS ONE 14 (6): e0217596.Find this resource:

Hedges, R. E. 2004. “Isotopes and Red Herrings: Comments on Milner et al. and Lidén et al.” Antiquity 78 (299): 34–37.Find this resource:

Henshilwood, C., P. Nilssen, and J. Parkington. 1994. “Mussel Drying and Food Storage in the Late Holocene, SW Cape, South Africa.” Journal of Field Archaeology 21 (1): 103–109.Find this resource:

Holdaway, S. J., P. C. Fanning, F. Petchey, K. Allely, J. I. Shiner, and G. Bailey. 2017. “Temporal Variability in Shell Mound Formation at Albatross Bay, Northern Australia.” PLOS ONE 12 (8): e0183863.Find this resource:

Jerardino, A. 1993. “Mid-to-Late Holocene Sea-Level Fluctuations: The Archaeological Evidence at Tortoise Cave, Southwestern Cape, South Africa.” South African Journal of Science 89: 481–488.Find this resource:

Jerardino, A. 1995. “The Problem with Density Values in Archaeological Analysis: A Case Study from Tortoise Cave, Western Cape, South Africa.” South African Archaeological Bulletin 50 (161): 21–27.Find this resource:

Jerardino, A. 1996. “Changing Social Landscapes of the Western Cape Coast of Southern Africa over the Last 4500 Years.” Unpublished PhD diss., University of Cape Town.Find this resource:

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 (11): 1031–1044.Find this resource:

Jerardino, A. 2007. “Excavations at a Hunter-Gatherer Site Known as Grootrif G’shell Midden, Lamberts Bay, Western Cape Province.” South African Archaeological Bulletin 62 (186): 162–170.Find this resource:

Jerardino, A. 2010a. “Large Shell Middens in Lamberts Bay, South Africa: A Case of Hunter–Gatherer Resource Intensification.” Journal of Archaeological Science 37 (9): 2291–2302.Find this resource:

Jerardino, A. 2010b. “Prehistoric Exploitation of Marine Resources in Southern Africa with Particular Reference to Shellfish Gathering: Opportunities and Continuities.” Pyrenae 41 (1): 7–52.Find this resource:

Jerardino, A. 2012. “Large Shell Middens and Hunter-Gatherer Resource Intensification along the West Coast of South Africa: The Elands Bay Case Study.” Journal of Island and Coastal Archaeology 7 (1): 76–101.Find this resource:

Jerardino, A. 2013. “Mega Shell Middens and Hunter-Gatherer Resource Intensification along the West Coast of South Africa.” In Shell Energy: Mollusc Shells as Coastal Resources, edited by G. N. Bailey, K. Hardy, and A. Camara, 207–228. Oxford: Oxbow Books.Find this resource:

Jerardino, A. 2016a. “Beating the Same Drum for a Lifetime and Smoke Screens: A Reply to Parkington et al. (2015) [see herein].” South African Archaeological Bulletin 71 (203): 87–97.Find this resource:

Jerardino, A. 2016b. “On the Origins and Significance of Pleistocene Coastal Resource Use in Southern Africa with Particular Reference to Shellfish Gathering.” Journal of Anthropological Archaeology 41: 213–230.Find this resource:

Jerardino, A. 2016c. “Shell Density as Proxy for Reconstructing Prehistoric Aquatic Resource Exploitation, Perspectives from Southern Africa.” Journal of Archaeological Science: Reports 6: 637–644.Find this resource:

Jerardino, A. 2017. “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 427: 103–114.Find this resource:

Jerardino, A. 2018. “Shell Fragmentation Beyond Screen-Size and the Reconstruction of Intra-Site Settlement Patterns: A Case Study from the West Coast of South Africa.” In Zooarchaeology in Practice: Case Studies in Methodology and Interpretation in Archaeofaunal Analysis, edited by Christina M. Giovas and Michelle J. LeFebvre, 151–171. Cham, Switzerland: Springer.Find this resource:

Jerardino, A., and R. Navarro. 2018a. “Large-Scale Hunter-Gatherer Exploitation of Marine Resources in South Africa, Part I.” South African Archaeological Bulletin 73 (208): 93–107.Find this resource:

Jerardino, A., and R. Navarro. 2018b. “Large-Scale Hunter-Gatherer Exploitation of Marine Resources in South Africa, Part II.” South African Archaeological Bulletin 73 (208): 108–125.Find this resource:

Jerardino, A., and R. Yates. 1996. “Preliminary Results from Excavations at Steenbokfontein Cave: Implications for Past and Future Research.” South African Archaeological Bulletin 51: 7–16.Find this resource:

Jerardino, A., and R. Yates. 1997. “Excavations at Mike Taylor’s Midden: A Summary Report and Implications for a Re-characterisation of Megamiddens.” South African Archaeological Bulletin 52: 43–51.Find this resource:

Jerardino, A., G. M. Branch, and R. Navarro. 2008. “Human Impact on Precolonial West Coast Marine Environments of South Africa.” In Human Impacts on Marine Environments, edited by G. M. Branch and R. Navarro, 279–296. Berkeley, CA: University of California Press.Find this resource:

Kennett, D. J., and B. Voorhies. 1996. “Oxygen Isotopic Analysis of Archaeological Shells to Detect Seasonal Use of Wetlands on the Southern Pacific Coast of Mexico.” Journal of Archaeological Science 23 (5): 689–704.Find this resource:

Klein, R. G. 2009. The Human Career: Human Biological and Cultural Origins. Chicago: University of Chicago Press.Find this resource:

Klein, R. G., G. Avery, K. Cruz-Uribe, D. Halkett, J. E. Parkington, T. Steele, T. P. Volman, and R. Yates. 2004. “The Ysterfontein 1 Middle Stone Age Site, South Africa, and Early Human Exploitation of Coastal Resources.” Proceedings of the National Academy of Sciences 101 (16): 5708–5715.Find this resource:

Klein, R. G., and T. E. Steele. 2008. “Gibraltar Data Are Too Sparse to Inform on Neanderthal Exploitation of Coastal Resources.” Proceedings of the National Academy of Sciences 105 (51): E115–E115.Find this resource:

Kyriacou, K., D. M. Blackhurst, J. E. Parkington, and A. D. Marais. 2016. “Marine and Terrestrial Foods as a Source of Brain-Selective Nutrients for Early Modern Humans in the Southwestern Cape, South Africa.” Journal of Human Evolution 97: 86–96.Find this resource:

Lee-Thorp, J. A., J. C. Sealy, and N. J. Van Der Merwe. 1989. “Stable Carbon Isotope Ratio Differences between Bone Collagen and Bone Apatite, and Their Relationship to Diet.” Journal of Archaeological Science 16 (6): 585–599.Find this resource:

Lewis, M. C., and J. C. Sealy. 2018. “Coastal Complexity: Ancient Human Diets Inferred from Bayesian Stable Isotope Mixing Models and a Primate Analogue.” PLOS ONE 13 (12): e0209411.Find this resource:

Loftus, E., J. Lee-Thorp, M. Leng, C. Marean, and J. Sealy. 2019. “Seasonal Scheduling of Shellfish Collection in the Middle and Later Stone Ages of Southern Africa.” Journal of Human Evolution 128: 1–16.Find this resource:

Lubell, D., F. A. Hassan, A. Gautier, and J. L. Ballais. 1976. “The Capsian Escargotières.” Science 191 (4230): 910–920.Find this resource:

Manhire, A. 1993. “A Report on the Excavations at Faraoskop Rock Shelter in the Graafwater District of the South-Western Cape.” Southern African Field Archaeology 2 (1): 3–23.Find this resource:

Marean, C. W., M. Bar-Matthews, J. Bernatchez, E. Fisher, P. Goldberg, A. I. Herries, Z. Jacobs, A. Jerardino, P. Karkanas, T. Minichillo, and P. J. Nilssen. 2007. “Early Human Use of Marine Resources and Pigment in South Africa during the Middle Pleistocene.” Nature 449 (7164): 905–908.Find this resource:

McCoy, M. D. 2008. “Hawaiian Limpet Harvesting in Historical Perspective: A Review of Modern and Archaeological Data on Cellana spp. from the Kalaupapa Peninsula, Moloka ‘i Island 1.” Pacific Science 62 (1): 21–38.Find this resource:

Meehan, B. 1977. “Man Does Not Live by Calories Alone: The Role of Shellfish in a Coastal Cuisine.” In Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and Australia, edited by J. Allen, J. Golson, and R. Jones, 157–205. New York: Academic Press.Find this resource:

Meighan, C. W., D. M. Pendergast, B. K. Swartz, and M. D. Wissler. 1958. “Ecological Interpretation in Archaeology: Part I.” American Antiquity 24 (1): 1–23.Find this resource:

Metcalfe, D., and K. R. Barlow. 1992. “A Model for Exploring the Optimal Trade‐off between Field Processing and Transport.” American Anthropologist 94 (2): 340–356.Find this resource:

Milner, N., J. Barrett, and J. Welsh. 2007. “Marine Resource Intensification in Viking Age Europe: The Molluscan Evidence from Quoygrew, Orkney.” Journal of Archaeological Science 34 (9): 1461–1472.Find this resource:

Milner, N., O. E. Craig, G. N. Bailey, K. Pedersen, and S. H. Andersen. 2004. “Something Fishy in the Neolithic? A Re-evaluation of Stable Isotope Analysis of Mesolithic and Neolithic Coastal Populations.” Antiquity 78 (299): 9–22.Find this resource:

Nelson, N. C. 1909. “Shellmounds of the San Francisco Bay Region.” University of California Publications in American Archaeology and Ethnology 7 (4): 309–358.Find this resource:

Orquera, L. A., D. Legoupil, and E. L. Piana. 2011. “Littoral Adaptation at the Southern End of South America.” Quaternary International 239 (1–2): 61–69.Find this resource:

Parkington, J. 1976. “Coastal Settlement between the Mouths of the Berg and Olifants Rivers, Cape Province.” South African Archaeological Bulletin 31 (123–124): 127–140.Find this resource:

Parkington, J. 1991. “Approaches to Dietary Reconstruction in the Western Cape: Are You What You Have Eaten?” Journal of Archaeological Science 18 (3): 331–342.Find this resource:

Parkington, J. 2001. “Milestones: The Impact of the Systematic Exploitation of Marine Foods on Human Evolution.” In Humanity from African Naissance to Coming Millennia, edited by P. V. Tobias, M. A. Raath, J. Moggi-Cocchi, and G. A. Doyle, 327–336. Johannesburg: Wits University Press.Find this resource:

Parkington, J. 2003. “Middens and Moderns: Shellfishing and the Middle Stone Age of the Western Cape, South Africa.” South African Journal of Science 99: 243–247.Find this resource:

Parkington, J. 2006. Shorelines, Strandlopers and Shell Middens: Archaeology of the Cape Coast. Cape Town: Krakadouw Trust.Find this resource:

Parkington, J. 2010. “Coastal Diet, Encephalisation and Innovative Behaviors in the Late Middle Stone Age of Southern Africa.” In Human Brain Evolution: The Influence of Freshwater and Marine Resources, edited by S. C. Cunnaneand K. M. Stewart, 189–203. Hoboken, NJ: Wiley.Find this resource:

Parkington, J. 2012. “Mussels and Mongongo Nuts: Logistical Visits to the Cape West Coast, South Africa.” Journal of Archaeological Science 39 (5): 1521–1530.Find this resource:

Parkington, J. 2016. “Elands Bay Cave: Keeping an Eye on the Past.” Southern African Humanities 29 (1): 17–32.Find this resource:

Parkington, J., J. W. Fisher Jr., and K. Kyriacou. 2013. “Limpet Gathering Strategies in the Later Stone Age Along the Cape West Coast, South Africa.” The Journal of Island and Coastal Archaeology 8 (1): 91–107.Find this resource:

Parkington, J., J. W. Fisher Jr., and T. W. Tonner. 2009. “‘The Fires Are Constant, The Shelters Are Whims’: A Feature Map of Later Stone Age Campsites at the Dunefield Midden Site, Western Cape Province, South Africa.” South African Archaeological Bulletin 64: 104–121.Find this resource:

Parkington, J., P. Nilssen, C. Reeler, and C. Henshilwood. 1992. “Making Sense of Space at Dunefield Midden Campsite, Western Cape, South Africa.” Southern African Field Archaeology 1 (2): 63–70.Find this resource:

Parkington, J. E. 2008. “Limpet Sizes in Stone Age Archaeological Contexts at the Cape, South Africa: Changing Environment or Human Impact.” In Early Human Impact on Megamolluscs, edited by Andrzej T. Antczak and Roberto Cipriani, 169–178. Oxford: Archaeopress.Find this resource:

Parkington, J. E., K. Kyriacou, and J. W. Fisher Jr. 2015. “Late Holocene Settlement Along the Atlantic Coast of the Cape: Megamiddens and the Case For and Against Human Population Increase.” South African Archaeological Bulletin 70: 124–127.Find this resource:

Parkington, J. E., C. Poggenpoel, W. Buchanan, T. Robey, A. Manhire, and J. Sealy. 1988. “Holocene Coastal Settlement Patterns in the Western Cape.” In The Archaeology of Prehistoric Coastlines, edited by G. N. Bailey and John Parkington, 22–41. Cambridge, UK: Cambridge University Press.Find this resource:

Pfeiffer, S., and J. Sealy. 2006. “Body Size among Holocene Foragers of the Cape Ecozone, Southern Africa.” American Journal of Physical Anthropology 129 (1): 1–11.Find this resource:

Rick, T. C., J. A. Robbins, and K. M. Ferguson. 2006. “Stable Isotopes from Marine Shells, Ancient Environments, and Human Subsistence on Middle Holocene Santa Rosa Island, California, USA.” Journal of Island & Coastal Archaeology 1 (2): 233–254.Find this resource:

Robertshaw, P. T. 1978. “Archaeological Investigations at Langebaan Lagoon, Cape Province.” Palaeoecology of Africa 10: 139–154.Find this resource:

Robertshaw, P. T. 1977. “Excavations at Paternoster, South-western Cape.” South African Archaeological Bulletin 32: 63–73.Find this resource:

Robey, T. S. 1984. “Burrows and Bedding: Site Taphonomy and Spatial Archaeology.” Unpublished master’s thesis, University of Cape Town.Find this resource:

Robey, T. S. 1987. “The Stratigraphic and Cultural Sequence at Tortoise Cave, Verlorenvlei.” In Papers in the Prehistory of the Western Cape, South Africa, edited by J. E. Parkington and M. Hall, 294–325. Oxford: BAR International Series.Find this resource:

Sealy, J., and M. Galimberti. 2011. “Shellfishing and the Interpretation of Shellfish Sizes in the Middle and Later Stone Ages of South Africa.” In Trekking the Shore: Changing Coastlines and the Antiquity of Coastal Settlement, edited by N. F. Bicho, J. A. Haws, and L. G. Davis, 405–419. New York: Springer.Find this resource:

Sealy, J. C., and N. J. van der Merwe. 1988. “Social, Spatial and Chronological Patterning in Marine Food Use as Determined by 𝛿13 C Measurements of Holocene Human Skeletons from the Southwestern Cape, South Africa.” World Archaeology 20 (1): 87–102.Find this resource:

Sealy, J. C., N. J. Van der Merwe, K. A. Hobson, D. R. Horton, R. B. Lewis, J. Parkington, P. Robertshaw, and H. P. Schwarcz. 1986. “Isotope Assessment and the Seasonal-Mobility Hypothesis in the Southwestern Cape of South Africa [and Comments and Replies].” Current Anthropology 27 (2): 135–150.Find this resource:

Sealy, J. C. 1989. “Reconstruction of Later Stone Age Diets in the Southwestern Cape, South Africa: Evaluation and Application of Five Isotopic and Trace Element Techniques.” Unpublished PhD diss., University of Cape Town.Find this resource:

Shawcross, W. 1967. “An Investigation of Prehistoric Diet and Economy on a Coastal Site at Galatea Bay, New Zealand.” Proceedings of the Prehistoric Society 33: 107–131.Find this resource:

Shiner, J. S., P. C. Fanning, S. J. Holdaway, F. Petchey, C. Beresford, E. Hoffman, and B. Larsen. 2013. “Shell Mounds as the Basis for Understanding Human-Environment Interaction in Far North Queensland, Australia.” Queensland Archaeological Research 16: 65.Find this resource:

Singer, R., and J. Wymer. 1982. The Middle Stone Age at Klasies River Mouth in South Africa. Chicago: University of Chicago Press.Find this resource:

Spaulding, A. C. 1960. The Dimensions of Archaeology. Indianapolis, IN: Bobbs-Merrill.Find this resource:

Speth, J. D. 2010. The Palaeoanthropology and Archaeology of Big Game Hunting. New York: Springer.Find this resource:

Steele, T. E., and E. Álvarez-Fernández. 2011. “Initial Investigations into the Exploitation of Coastal Resources in North Africa during the Late Pleistocene at Grotte des Contrebandiers, Morocco.” In Trekking the Shore: Changing Coastlines and the Antiquity of Coastal Settlement, edited by N. F. Bicho, J. A. Haws, and L. G. Davis, 383–403. New York: Springer.Find this resource:

Stein, J. K. 1992. Deciphering a Shell Midden. San Diego, CA: Academic Press.Find this resource:

Stiner, M. C. 1994. Honor Among Thieves. A Zooarchaeological Study of Neandertal Ecology. Princeton, NJ: Princeton University Press.Find this resource:

Stringer, C. B., R. N. E. Barton, and J. C. Finlayson, eds. 2000. Neanderthals on the Edge: Papers from a Conference Marking the 150th Anniversary of the Forbes’ Quarry Discovery, Gibraltar. Oxford: Oxbow Books.Find this resource:

Stringer, C. B., J. C. Finlayson, R. N. E. Barton, Y. Fernández-Jalvo, I. Cáceres, R. C. Sabin, E. J. Rhodes, A. P. Currant, J. Rodríguez-Vidal, F. Giles-Pacheco, and J. A. Riquelme-Cantal. 2008. “Neanderthal Exploitation of Marine Mammals in Gibraltar.” Proceedings of the National Academy of Sciences 105 (38): 14319–14324.Find this resource:

Swadling, P. 1976. “Changes Induced by Human Exploitation in Prehistoric Shellfish Populations.” Mankind 10 (3): 156–162.Find this resource:

Villagran, X. S. 2019. “The Shell Midden Conundrum: Comparative Micromorphology of Shell-Matrix Sites from South America.” Journal of Archaeological Method and Theory 26 (1): 344–395.Find this resource:

Volman, T. P. 1978. “Early Archeological Evidence for Shellfish Collecting.” Science 201 (4359): 911–913.Find this resource:

Waselkov, G. A. 1987. “Shellfish Gathering and Shell Midden Archaeology.” Advances in Archaeological Method and Theory vol. 10, 93–210. New York: Academic Press.Find this resource:

Yates, R. 2004. Archaeological Heritage Resources Assessment: Portion 37 of the Farm Uitkomst 23, Paternoster, Saldanha Bay Municipality, Western Cape Province. Report prepared for Deon Van Zyl Consultants. Cape Town: Iziko South African Museums.Find this resource:

Yellen, J. 1977. Archaeological Approaches to the Present: Models for Reconstructing the Past. New York. Academic Press.Find this resource:

Zangrando, A. F., G. P. Vargas, and A. M. Tivoli. 2017. “Decreased Foraging Return in Shellfishing? Species Composition and Shell Size of Blue Mussel (Mytilus edulis) from a Late Holocene Site of the South Coast of Tierra del Fuego. Quaternary International 427: 160–169.Find this resource: