Shellfishing and the colonization of sahul: a multivariate model evaluating the dynamic effects of prey utility, transport considerations and life-history on foraging patterns and midden composition

Archaeological evidence of shellfish exploitation along the coast of Sahul (Pleistocene Australia-New Guinea) points to an apparent paradox. While the continental record as a whole suggests that human populations were very low from initial colonization through early Holocene, coastal and peri-coastal sites dating to that time are dominated by small, low-ranked littoral taxa to the near-complete exclusion of large, higher ranked sub-littoral species, precisely the opposite of theory-based expectations, if human populations and predation rates were indeed as low as other data suggest. We present a model of shellfish exploitation combining information on species utility, transport considerations and prey life-history that might account for this apparent mismatch, and then assess it with ethnographic and archaeological data. Findings suggest either that high-ranked taxa were uncommon along the Pleistocene coastlines of Sahul, or that abundant and commonly taken high-ranked prey are under-represented in middens relative to their role in human diets largely as a function of human processing and transport practices. If the latter reading is correct, archaeological evidence of early shellfishing may be mainly the product of subsistence activities by children and their mothers.

Shellshing and the Colonization of Sahul: A multivariate model evaluating the dynamic eects of prey utility, transport considerations and life-history on foraging patterns and midden composition Brian F. Codding1;, James F. O'Connell 1 and Douglas W. Bird2 1. Department of Anthropology, University of Utah, Salt Lake City, UT, 84112, USA 2. Department of Anthropology, Stanford University, Stanford, CA, 94305, USA Corresponding author. E-mail: brian:codding@anthro:utah:edu UNCORRECTED AUTHOR'S MANUSCRIPT Citation: Codding, Brian F., James F. O'Connell and Douglas W. Bird (2014) Shellshing and the Colonization of Sahul: A multivariate model evaluating the dynamic eects of prey utility, transport considerations and life history on foraging patterns and midden composition. Journal of Island & Coastal Archaeology 9(2):238{252. Abstract. Archaeological evidence of shellsh exploitation along the coast of Sahul (Pleistocene Australia-New Guinea) points to an apparent paradox. While the continental record as a whole suggests that human populations were very low from initial colonization through early Holocene, coastal and peri-coastal sites dating to that time are dominated by small, low-ranked littoral taxa to the near-complete exclusion of large, higher ranked sub-littoral species, precisely the oppo- site of theory-based expectations, if human pop- ulations and predation rates were indeed as low as other data suggest. We present a model of shellsh exploitation combining information on species utility, transport considerations and prey life-history that might account for this apparent mismatch, and then assess it with ethnographic and archaeological data. Findings suggest either that high-ranked taxa were uncommon along the Pleistocene coastlines of Sahul, or that abun- dant and commonly taken high-ranked prey are under-represented in middens relative to their role in human diets largely as a function of hu- man processing and transport practices. If the latter reading is correct, archaeological evidence of early shellshing may be mainly the product of subsistence activities by children and their moth- ers. Keywords: Shellsh Exploitation j Coloniza- tion of Sahul (Australia/New Guinea) j Chil- dren's Foraging j Division of Labor j Human Be- havioral Ecology Introduction Archaeological evidence of early shellsh collect- ing on Sahul (Pleistocene Australia-New Guinea) points to an apparent paradox. While theo- retical predictions from foraging theory lead to the expectation that the earliest foragers on the continent would have encountered abundant large high-ranking shellsh taxa oering rela- tively high average rates of nutrient return rela- tive to time spent collecting and processing, ar- chaeological data from coastal shell middens in- dicate that foraging emphasized smaller, much lower ranked shellsh taxa. There are at least three plausible explanations for this pattern: 1. Rapidly changing Pleistocene sea levels kept the abundance of large, high-ranked, slow- growing taxa low, leaving foragers no option but to pursue smaller, lower ranked prey (Beaton 1995). 2. Early human populations were much higher than previously thought, placing sustained UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul collecting pressure on high ranked prey which rapidly reduced their numbers, leav- ing only an ephemeral archaeological signa- ture. 3. Transport considerations favored the dier- ential discard of inedible components of rel- atively high ranked prey at or near the point of acquisition, leaving little or no evidence of their use in base camp middens. The ma- terial signature of lower-ranked taxa may then be the result of foraging by children, who experience lower encounter rates with large prey due to their slower walking speed (Bird et al. 2004), and by their moth- ers, who may be more interested in low-risk than high-energy resources for provisioning young (Codding et al. 2011). Resolving this issue is crucial to understand- ing the process through which Sahul was colo- nized and the early foraging decisions of indi- viduals on a previously unoccupied landscape (O'Connell and Allen 2012). Here we address the problem with a novel model that examines prey utility, transport considerations and life- history. Combined, this approach may better predict both resource exploitation patterns and their archaeological consequences. We assess the model in light of ethnographic and ethnoarchae- logical data, then consider its implications for understanding the record of early shellsh ex- ploitation in Sahul. The Model Our model is based on well-established ideas about prey choice, central place foraging and life history. These allow us to make predic- tions about which resources will be more (or less) likely to be targeted, which will be more (or less) likely to be represented in archaeolog- ical deposits created by central place foragers, and which will be more (or less) resilient to pre- dation pressure or environmental change. The prey choice model (PCM) allows the derivation of simple predictions about which species are more likely to be pursued on en- counter (Bettinger 2009, Bird and O'Connell 2006, Charnov and Orians 2006). When search- ing a patch, a forager will encounter a vari- ety of dierent resources. If foragers are trying maximize the rate of nutrient acquisition, then they should preferentially pursue resources with higher post-encounter return rates (e/h), mea- sured as the expected nutrients (e) gained from acquiring a resource (typically measured as en- ergy) over the handling costs (h) associated with pursuing and processing that resource (typically measured in time). If higher ranking resources are suciently abundant, then foragers should only take those resources on encounter, passing up all other lower ranking resources; however, as the abundance of the highest ranking resources decline, foragers should begin to take lower rank- ing resources in rank order. From a forager's perspective, the question is: should I pursue this resource, or pass it up and continue searching for other, higher-ranked resources? The answer depends on whether the post-encounter return rate (e/h) for that resource is expected to in- crease the overall return rate (E/T) which in- cludes all energy acquired in a patch (E) rel- ative to all in-patch search and handling time (T). These model dynamics allow the derivation of predictions about which taxa a forager should preferentially pursue and a proxy for overall for- aging eciency based on the presence or absence of lower-ranking resources. Central place foraging models (CPF) allow the evaluation of the costs and benets of eld processing resources or transporting them whole from an acquisition locale to a central place (Bet- tinger 2009, Charnov and Orians 2006, Metcalfe and Barlow 1992, Orians and Pearson 1979). Once a forager has acquired a resource, they must make decisions about how to best trans- port that resource to a central place. Assuming that individuals are attempting to maximize the rate at which resources are delivered to a central place, foragers may be better o culling low util- ity items (i.e., shell or bone) in the eld and re- turning to their central place with only high util- J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul ity parts (i.e., meat) than making a greater num- ber of trips transporting both high and low util- ity parts. Assuming constant processing costs across taxa, the net gain in utility as a function of processing (the increase in high utility parts per load) can predict which items are more likely to be eld processed at a given distance from a central place. Given that individual shellsh have an edible portion (meat) and a non-edible portion (shell), meat:shell ratios provide a reli- able proxy of the potential benet gained from culling shell in the eld, thereby increasing the utility of each load returned home. Those taxa with higher proportions of meat to shell are less likely to be eld processed while those with low meat:shell ratios are more likely to be processed in the eld. Because larger taxa generally have lower meat:shell ratios and require lower pro- cessing costs than smaller taxa (e.g., Bird et al. 2002), these predictions likely hold with the in- clusion of processing costs. Life history models (LFM) provide avenues to investigate variability in the parameters that govern individual growth, maturation, reproduc- tion and mortality, which aggregate to produce population level eects. Because dierent taxa have dierent life-history characteristics, they should respond dierently to human exploita- tion, which in turn, will alter a forager's future encounter rates. With invertebrate taxa, life- history parameters such as growth and matu- ration rates should provide a reliable proxy of the relative susceptibility of a taxon to overex- ploitation. Likely the result of natural selection (Gadgil and Bossert 1970), the rate of matura- tion covaries with a number of life-history pa- rameters and can be a predictor of population level parameters (Peters 1983). As such, rates of maturation for individual taxa should provide a relative measure of the likelihood or speed at which a population could recover from exploita- tion (Whitaker 2008, Whitaker and Byrd 2012). Due to the eects of biological scaling and allom- etry, body size and post-encounter return rates (at least, for sessile resources, see Bird et al. 2009) should co-vary positively while body size, meat:shell ratios and maturation rates may co- vary negatively (Peters 1983). This could be problematic as high ranking resources may be more likely to be processed in the eld, but if particular taxa exist as outliers in these rela- tionships (i.e., high ranking, slow-growing but high meat: shell ratios), they should provide key proxies to monitor changes in foraging patterns through time. Combined, these models allow us to make pre- dictions about which species should be prefer- entially pursued, which should be more or less likely to end up in archaeological sites and which are more or less likely to be depleted by preda- tion. Following O'Connell (1995), these models can be used to derive specic deductive predic- tions that can be tested with ethnographic, eth- noarchaeological and archaeological data. Be- cause these predictions are derived from a gen- eral theory of behavior, they avoid direct ethno- graphic analogy. Predictions Imagine a parameter space represented by three variables: shellsh utility (e/h, on the Figure 1: y-axis), transport costs (operationalized by meat:shell ratios, Figure 1: x-axis) and sus- ceptibility to overexploitation (measured here through the proxy of maturation rates). How dierent shellsh taxa fall out within this pa- rameter space should allow us to identify target taxa that will allow us to investigate foraging behavior, midden composition and how they are expected to change over time. Taxa with very low meat:shell ratios (Figure 1: frame a) should rarely appear in middens due to the low net utility of a load returned to a central place with such a large amount of shell relative to a limited amount of meat. If shellsh rank varies inversely with meat:shell ratios, then evidence for the exploitation of high ranking taxa is unlikely to be recovered from archaeological deposits as their shells should be culled in the eld even over short transport distances. This is the case for many molluscan species from this region (Bird and Bliege Bird 2000). J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul Low High Low High Meat:Shell e/h Dominate early middens Absent from middens Dominate late middens (a) (b) (c) Figure 1: Figure 1. Schematic predictions of midden composition based on the combined eects of each taxon's post-encounter return rate (e/h), and likelihood of eld processing (meat:shell ratio). Taxa with low meat:shell ratios (frame a) are less likely to end up in middens as their shells should be frequently culled in the eld (unless if the shell itself provides some value). Early middens should be dominated by higher ranking taxa with relatively high meat:shell ratios (frame b). If these become less abundant in the environment due to overexploitation, then foragers should begin to take lower ranking resources which will come to dominate later middens (frame c). The rate of transition in midden dominance from b!c depends on the life-history characteristics of the taxa in question: fast growing species should be harder to deplete while slow growing species rapidly decline in abundance. Note: these values are in dimensionless space. J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul High ranking taxa with high meat:shell ra- tios (Figure 1: frame b) should always be taken on encounter by foragers and are likely to al- ways be transported whole to the central place. As such, these taxa should dominate early mid- dens. Those with slow maturation rates should be the best source to monitor the eects of re- source overexploitation through time. Those that mature early are likely to be fairly resilient to over-exploitation and may dominate middens throughout the prehistoric record. Lower-ranked taxa with high meat:shell ratios (Figure 1: frame c) should come to dominate lat- ter middens if higher ranking resources become less abundant due to overexploitation. Those that take longer to mature will be particularly useful to monitor human foraging and popula- tion pressure over time. Those that are rela- tively resilient to over-exploitation (rapid growth rates and short maturation rates) should domi- nate later shell middens (and shell mounds) fol- lowing declines of the highest ranking resources (e.g., Whitaker and Byrd, this issue). Results Resource-Specic Attributes Table 1 provides a summary of available data on shellsh genera post-encounter return rates (e/h), meat:shell ratios and age at maturity for taxa potentially exploited by the earliest inhabi- tants of Sahul. Mean values are also represented in Figure 2 following the layout of predictions in Figure 1. The two large bivalves, Tridacna and Hip- popus, would be excellent candidates to moni- tor foraging pressure given their high rank and slow growth rates. However, given their low meat:shell ratios, they are unlikely to be trans- ported to midden deposits proportionally to the frequency with which they are taken. This ex- pectation is supported by ethnoarchaeological studies indicating that the material record of Tridacna and Hippopus underrepresent the fre- quency with which they are actually acquired (see below, Bird and Bliege Bird 2000, Bird et al. 2002). Table 1: Proxy data for shellsh utility (e/h), transport cost (meat:shell) and susceptibility to overexploitation (age at maturity), ordered by mid-point return rate (e/h). Taxon e/h Meat:Shell Age at (kcal/hr) Ratio Mat. (yr) Tridacna 2,622-13,064 0.170 4-5 Hippopus 1,680-9,120 0.115 6-7 Lambis 3,412-5,106 0.092 2 Trochus 977-3,904 0.208 1-2 Cypraea 2,214 0.278 1 Chiton 446-2,228 1.159 1-2+? Nerita 42-1,106 0.301 1.6 Turbo 520-606 0.248 3-4 Strombus 294-543 0.148 1-2 Asaphis 42-78 0.338 1-2 Data on shellsh returns from Bird and Bliege Bird (2000), Bird et al. (2004), Kennedy (2005) and Thomas (2007); on meat:shell ratios from Bird and Bliege Bird (1997), Erlandson (1988); on age at maturity from Beesley et al. (1998), Munro (1993), Nash (1993) and Yamaguchi (1993). Unfortunately, within the suite of resources available to the earliest colonists of Sahul (and for which data are currently available), there are no species that can be considered high-ranking with high meat:shell ratios (i.e., falling within parameter space b in Figure 1). Evidence from North America suggests that Abalone (Haliotis spp.) likely falls within this region of the param- eter space (Kennedy 2005), but conclusive data on Austral species are lacking. The absence of species within this parameter space suggests that ecient tidal foraging in Sahul required eld pro- cessing. As a result, we are unlikely to acquire an unbiased picture of the prehistoric exploita- tion of these higher ranking taxa. With relatively low returns but high meat:shell ratios, small gastropods such as Nerita and small bivalves like Asaphis also have rapid matura- tion rates, suggesting they are resilient to overex- ploitation. As we would predict, these taxa dom- inate modern middens within permanent villages in the Torres Strait Islands, Australia and West- ern Kiribati, Micronesia where sustained forag- ing pressure does not seem to diminish their pop- J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul 0.0 0.1 0.2 0.3 1.2 0 2000 4000 6000 8000 Meat:Shell e/h Nerita Asaphis Trochus Cypraea Turbo Hippopus Strombus Lambis Tridacna Chiton Growth fast slow Figure 2: Figure 2. Shellsh genera plotted as a function of post-encounter return rate (e/h) and meat:shell ratios following predictions in Figure 1. Shading indicates growth rates measured by time till maturity from short (white) to long (black). Genera plotted by mid-point values of ranges given in Table 1. J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul ulations (Bird et al. 2002, Thomas 2007). As these taxa should be resilient to overexploitation, their abundance is unlikely to vary with human foraging and population pressure over time. The moderately sized gastropod Turbo pro- vides a welcome exception: with relatively long times to maturity and high meat:shell ratios, we should expect that Turbo will be exploited when the abundance of higher ranking resources are low, and should be one of the rst taxa within this lower-ranked set to exhibit overexploitation. As such, diachronic trends in the exploitation of Turbo should provide a key proxy to under- stand foraging pressure on shellsh species, and as such, a proxy for human population densi- ties. With by far the most extreme meat:shell ratio, chiton (class Polyplacophora) may also fall within this category. Larger species of chiton are likely to have higher return rates and slower maturation rates, while still maintaining rela- tively high meat:shell ratios; unfortunately, pre- cise data are not available for Austral species. Based on observations of Meriam shellshers by one of the authors (DWB), limited data on chi- ton exploitation by contemporary foragers may result from low encounter rates with larger chi- ton coupled with consistent decisions to pass over smaller chiton in search of more protable taxa. An Ethnographic Test Previous ethnoarchaeological investigation with Meriam Islanders in the Torres Strait explored the archaeological consequences of trade-os that shellshers face in acquiring, handling, and transporting a number of the taxa that Pleis- tocene foragers would have likely encountered on colonizing Sahul (Bird 1996, Bird and Bliege Bird 1997, 2000; Bird et al. 2002, 2004). Con- sistent with expectations framed by the prey choice model (PCM), Meriam restrict the suite of taxa they handle on encounter to those shell- sh that will increase overall foraging eciency (E/T) while collecting on the reef at or harvest- ing within the rocky shore. For example, while reef at collecting, foragers routinely select ex- posed tridacnid clams (Tridacna spp. and Hip- popus hippopus), Lambis lambis conch, and large Trochus niloticus specimens, all of which signif- icantly increase overall foraging return rates if handled on-encounter. Conversely adult shell- shers almost always pass over specimens that oer returns for time handling that will reduce foraging eciency: small Strombus (Conomurex) luhuanus conch, small Trochus, and small bi- valves embedded in the reef. Meriam also pay close attention to changes in patch return rates while intertidal foraging, and switch from reef at collecting to rocky shore harvesting (for Asaphis violascens clams and Nerita spp. snails) when an incoming tide depresses the overall re- turn rate from collecting in the mid-littoral be- low that expected from harvesting in the near shore. While observed time allocation and selectiv- ity are consistent with predictions generated by the PCM and patch residence models, they are in sharp contrast to the frequency of dierent taxa of shellsh found in either contemporary or prehistoric shellmiddens on the Meriam Islands. The most important resources for contemporary collectors (especially large tridacnids and Lam- bis lambis) are very rare in the archaeological deposits, which are dominated by small gastro- pod shells (especially nerites, small strombids, and small Trochus niloticus; Bird et al. 2002). We can account for these dierences by consid- ering the eects of dierential eld processing and age-linked dierences in foraging returns. While reef at collecting, children walk slower and encounter high ranked resources at a lower rate than adults, resulting in lower overall re- turn rates and a predictably broader range of selectivity that includes the smaller gastropods. So while both adults and children collect the higher ranked shellsh on-encounter, only chil- dren regularly collect the abundant small strom- bids and small specimens of Trochus niloticus that are common in the shell middens. But why are tridacnid and Lambis shell, which make up the bulk of shellsh harvested, so rare in the de- posits? If we consider how time spent eld pro- cessing can increase the ratio of high to low util- J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul ity parts (meat:shell), the central place foraging model discussed above generates precise predic- tions about which types of molluscs should be eld processed if a forager's goal is to increase the rate at which shellsh meat can be trans- ported home. The model well anticipates the fact that Meriam foragers almost always cull the shells of large tridacnid clams and Lambis conch while on the reef. These shellsh oer high post- encounter return rates and have low meat:shell ratios, and thus eld processing quickly frees up more space for more meat, and more time for more foraging and transport. As such, Meriam adults and children commonly forage well be- yond the predicted thresholds at which in-bulk transportation (without culling the valves) of these large shellsh will increase the home de- livery rate of meat. Conversely, the smaller reef at shellsh that children often collect, along with rocky shore taxa such as Asaphis and Ner- ita that both adults and children harvest, are always transported in-bulk, and foragers very rarely cross the predicted eld processing thresh- olds while collecting these prey. We suspect that Pleistocene foragers would have faced similar trade-os when shellshing in intertidal zones. Early Middens in Sahul Data are available for at least six Pleistocene middens (coastal and peri-coastal sites yield- ing marine shell) with dates older than 30kya. Buang Merabak (Leavesley and Allen 1998, Rosenfeld 1997) and Matenkupkum (Gosden and Robertson 1991) are located on New Ireland, Kilu (Wickler 2001) on the northern end of the Solomon Islands and Noala Cave (Veth et al. 2007), Mandu Mandu Creek (Morse 1988), and Devil's Lair (Dortch 2004) are located along the coast of modern day Western Australia.1 For more detailed summaries on the dating of each site, see Allen and O'Connell (2003, 2008), O'Connell and Allen (2004, 2007, 2012) and 1Early middens (pre-30kya) for which data were un- obtainable include other sites along the Cape Range of northwestern Australia (Przywolnik 2008, Morse 1988, Morse 1993). O'Connell et al (2010). Evidence for shellsh exploitation at many of these sites is extremely sparse; nonetheless, the available data allow us to run an initial test of these predictions. Higher ranking species with low meat:shell ratios are rare in these early middens. At Buang Merabak, chiton, Turbo and Nerita dom- inate the Pleistocene assemblage (Leavesley and Allen 1998). At Kilu, Nerita dominates assem- blages dating before the Late Holocene (Wickler 2001). At Matenkupkum, Gosden and Robert- son (1991) report extremely large Turbo shells, indicating the exploitation of lower ranked taxa, but assuming that human predation will drive down shellsh size (e.g., Klein and Steele 2013) only limited predation pressure on these taxa. While Tridacna is absent from the faunal assem- blages recovered from these early levels, Leaves- ley and Allen (1998) report that large Tridacna shells and Turbo opercula, originally recorded as white chert, were used to manufacture tools at Buang Merabak. This suggest that these taxa were taken on encounter, likely for their food and shell, but only transported and de- posited within midden contexts for purposes re- lating to the latter. This also suggests that such high-value shell might be secondarily trans- ported away from middens for subsequent use, a pattern that may further bias such a taxon's rep- resentation in the archaeological record. At peri-coastal sites along the shore of modern day Western Australia, shellsh taxa are dom- inated by dierent genera specic to the local environment; but even so, the taxa exploited at these sites share characteristics with those ex- ploited in the Bismarks. At Noala Cave on the Monte Bello Islands, foragers were exploiting a variety of whelks (Terebralia, Veth et al. 2007), which are expected to be low ranked. Perhaps at odds with ndings elsewhere, the earliest lay- ers at Mandu Mandu Creek reveal chiton and baler (Melo) (Morse 1988). While precise data on these taxa are not yet available, what we know suggests that they may be relatively high ranked with high meat:shell ratios (Figure 1, frame b). If these were large chiton and Melo, this evi- J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul BuaMnagt eMnekruapbkaukm Kilu Devil's Lair Noala Cave Mandu Mandu 60 m contour 120 m contour W A L L A C E A S U N D A S A H U L P a c i f i c O c e a n I n d i a n O c e a n B i s m a r k s S o l o m o n s Figure 3: Map of coastal and peri-coastal archaeological sites with evidence of marine shell ex- ploitation prior to 30kya. J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul dence for the exploitation of high-ranking taxa during this interval supports the idea that large, high-ranking taxa were present, pursued on en- counter, but only transported to central places when meat:shell ratios were high or when the shells has some added value. This might also be true in the southwestern corner of the conti- nent, where small fragments of abalone (Halio- tis) are some of the only taxa represented in the early levels at Devil's Lair (and at Tunnel Cave; Dortch 2004; J. Dortch, personal communication 2013). Abalone too likely falls within the param- eter space shown in Figure 1, frame b represent- ing high-rank and high meat:shell ratios, leading us to expect that it will often be transported whole. The additional value of abalone shell for ritual, ornamental or other purposes beyond sub- sistence likely increase the probability that it will often be transported whole. This is consis- tent with nds at Carpenter's Gap, where early levels contain pearl shell (O'Connor 1999). Diachronic Trends at Buang Merabak Temporal trends from the well reported nds at Buang Merabak can further our understanding of these patterns. As predicted from our model, while we should not expect high-ranking, low meat:shell taxa to be deposited at central places, we should be able to monitor the abundance of Turbo and chiton through time as an indicator of foraging pressure. As evident in Figure 4, the temporal trends at Buang Merabak show a signicant decline in the proportion of Turbo ( = 17:6; p = 0:0361) and a marginally signicant decline in the proportion of chiton ( = 4:2; p = 0:1498) through the Pleis- tocene. These are mirrored by a marginally sig- nicant increase in Nerita ( = 7:0; p = 0:0905). 2 Such patterning has also been reported at Kilu (Wickler 2001). While signicant, declines in the relative abundance of Turbo and chiton oc- cur over a long span of time, suggesting that it 2Pearson's correlation analyses run in R (R Develop- ment Core Team, 2013). Figure 4 produced using the battleship.plot function in the plotrix library. Other Nerita Chiton Turbo 1 2 3 4 5 6 Pleistocene Holocene Analytical Unit Figure 4: Battleship plot showing the propor- tion of total weight per analytical unit for im- portant Pleistocene shellsh taxa at Buang Mer- abak. Data from Leavesley and Allen (1998), shown here in Table 2. Based on radiocarbon dates form Rosenfeld (1997), analytical unit 6 (2-sigma range 38,090-34,523 cal BP), 5 (24,141- 22,505 cal BP) and 4 (24,782-23,356 cal BP) date to the Pleistocene, while analytical units 3, 2 and 1 all date to the Holocene (2-sigma ranges cali- brated in OxCal 4.2.2, Bronk Ramsey 2009, us- ing a marine curve from Reimer et al. 2009 and local correction by Petchey et al. 2004). The two most abundant taxa in the earliest Pleistocene deposits (Turbo and chiton) are also relatively slow growing. Their overexploitation through the Pleistocene is mirrored in the rising dom- inance of faster-growing Nerita. Discontinuity between the Pleistocene-Holocene record is likely due to sea level stabilization at distances closer to the site (see Lambeck and Chappell 2001). J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul Table 2: Shell weight (g) by taxa across chronological analytical units at Buang Merabak. Taxa 1 2 3 4 5 6 Cypraea 90.6 223.0 16.6 0.0 0.0 13.6 Strombus 34.2 62.8 1.2 0.6 0.0 0.0 Isognomon 15.3 66.1 4.7 0.0 0.0 0.0 Trochus 41.4 153.0 83.8 23.9 12.5 9.3 Barnacle 50.0 136.9 797.9 8.9 0.9 2.0 Limpet 7.8 26.5 33.3 24.2 0.5 0.0 Nerita 52.7 226.0 258.0 1398.4 122.0 265.8 Tectarius 2.5 3.1 11.3 79.2 11.3 49.9 Chiton 22.6 61.0 53.9 267.3 99.7 509.4 Turbo 13.1 71.1 7.9 165.5 58.9 408.9 From Balean (1989) reported in Leavesley and Allen (1998), see also Rosenfeld (1997). Given their high meat:shell ratio, chiton likely contributed more to the early diets than is represented by shell weight alone (Leavesley 2004). took most of the Pleistocene record before hu- man populations were high enough to place con- sistent pressure on these taxa. This interpre- tation is supported by the radiocarbon proba- bilities from Buang Merabak which suggest in- termittent occupations (Figure 5a) and occu- pational estimates from neighboring Australia, which suggest very low populations levels dur- ing this time (Figure 5b; Williams 2013). These ndings suggest that Pleistocene occupations at Buang Merabak were intermittent, perhaps co- varying negatively with the distance to the coast (cf., Figure 5a and 5c), and were always at a low densities. Discussion This paper began with a conundrum: the ar- chaeological record suggests that the earliest col- onizing populations of Sahul disproportionately exploited low-ranked littoral shellsh species, despite theoretical predictions to the contrary. While a number of hypotheses could be proposed to explain this conundrum, we suggest that these can be assessed by incorporating models of prey eciency, transport utility and life-history to predict the targets of foraging, the resulting mid- den composition and the potential impact of hu- man foraging on shellsh species. In light of our model, the conundrum actually makes sense: high-ranking shellsh taxa available to the early foragers of Sahul had low meat:shell ratios (Fig- ure 1: frame b). As it would not make economic sense to transport shell from these taxa (unless if the shell itself had some added value; e.g., Leavesley and Allen 1998), we should not expect middens to contain shell from these taxa. But what then accounts for the deposition of lower ranking littoral shellsh species? Given the extremely low estimated population levels and intermittent occupation of sites like Buang Merabak, it seems unlikely that the ac- quisition and deposition of low-ranked taxa was due to overexploitation of higher ranking taxa. If the abundance of these larger high-ranked species declined rapidly, then we might have ev- idence for a larger colonizing population placing more pressure on resources than traditionally as- sumed. However, given the long period of time that passed before human populations reached densities required to overexploit fairly slow grow- ing taxa like Turbo, we suggest that human pop- ulation densities were quite low, their mobility quite high, and as such, they placed minimal predation pressure on the littoral environment. Combined with evidence for the rapid dispersal of peoples along the coastlines of Sahul, this im- plies that it takes very little impact to make adja- cent, unoccupied patches more appealing to for- aging populations (O'Connell and Allen 2012). It still could be that these taxa occurred in low J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul 50 45 40 35 30 25 −120 −80 −40 0 Years cal BP Sea Level (m below present) Relative Sea Level (Huon Peninsula, PNG) 40k 20k 0 Estimated Total Population of Australia Pop Buang Merabak (a) (b) (c) Figure 5: Occupational history of Buang Merabak (a, Leavesley 2004, Leavesley and Allen 1998, summed calibrated radiocarbon proababilites from OxCal 4.2.2, Bronk Ramsey 2009, using a marine curve from Reimer et al. 2009 and a local correction by Petchey et al. 2004) relative to the estimated total population of Australia (b, assuming a founding population of 5,000 people, Williams 2013) and estimated uctuations in sea level (c, Lambeck and Chappell 2001). J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul abundance due to the natural eects of a uctu- ating coastline (Beaton 1995). While alternative lines of evidence (such as paleontological nds) will be required to test this further, we suspect that habitats (e.g. broad fringing reefs) sup- porting healthy populations of some of the high- est ranked shellsh (such as Tridacna) would have been commonly encountered when Sahul was rst colonized (Pope and Terrell 2008). But again, that such taxa are high-ranked and of- ten associated with low meat:shell ratios leads us to suspect that they would have been important prey options whose archaeological presence was heavily ltered by eld processing. This agrees with ethnographic observations on tropical reef at foraging, where highly ranked Tridacna are always exploited on-encounter, but the valves are rarely transported back to a central place (Bird et al. 1997, 2002, 2004; Thomas 2007). If our interpretations of these results are cor- rect, it seems that the most reasonable expla- nation for the high abundance of low-ranking taxa at early sites may result from the dier- ential foraging constraints faced by children and their mothers. Because children generally walk more slowly than adults and because they may have less strength and skill in processing, they are likely to have lower encounter rates with high ranked prey and are less able to eld process taxa even when it might be desirable to do so (Bird and Bliege Bird 2000, Bird et al. 2004, Meehan 1982). Further, given the limited availability of alternative caregivers expected with smaller pop- ulations (Codding et al. 2011), mothers with young children will generally experience greater trade-os between childcare and foraging. Inter- tidal shellsh foraging provides an activity with lower costs of childcare than other alternatives (e.g., sub-tidal foraging, pelagic shing, etc.) and despite being low return, shellsh (espe- cially small taxa) are characterized by low acqui- sition variance, making them ideal resources for provisioning dependents (Codding et al. 2011). Indeed, cooperative parent-juvenile units may do better overall by working together in lower- return activities than by working apart (Hawkes et al. 1995, Kramer 2011). While this seems like the most plausible explanation for the Pleistocene record of shell- shing on Sahul given the available evidence, more work is required to truly understand these patterns. Continued archaeological and pale- oenvironmental work is necessary, but research might benet most from continued ethnoarchae- ologial and simple actualistic experiments. This includes simple measures of utility, processing costs and meat:shell ratios for key species (e.g., Halitois, Melo and chiton) and additional studies of well-known taxa (e.g., Turbo, Nerita and Asaphis) to further our understanding of variability within species. Additional work along these lines should greatly improve our interpretations of the archaeological record and allow us to solve such conundrums as the record of early shellshing by the colonists of Sahul. Acknowledgments. Thanks to Richard Klein for suggesting the idea for this symposium and collection of papers. We remain indebted to the Meriam Community for hosting DWB's ethno- graphic work over many years, which was sup- ported by the National Science Foundation and the LSB Leakey Foundation. This paper has beneted from comments, suggestions and work by Jim Allen, Rebecca Bliege Bird, Shannon Arnold-Boomgarden, Joe Dortch, James Holland Jones, Peter Veth, Adie Whitaker, and David Zeanah. Three anonymous reviewers provided detailed comments that greatly improved the - nal result. References Allen, J. and J.F. O'Connell. 2003. The long and the short of it: Archaeological approaches to determining when humans rst colonised Australia and New Guinea. Australian Archae- ology 57:5{19. Allen, J. and J. F. O'Connell. 2008. Getting from Sunda to Sahul. In Islands of Inquiry: colonization, seafaring and the archaeology J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul of maritime landscapes (G. Clark, F. Leach and S. O'Connor, eds): 31{46. ANU E Press, Canberra: Australian National University. Balean, C. E. 1989. Caves as refuge sites: an analysis of shell material from Buang Merabak. BA Honours Thesis, Deptartment of Prehistory and Anthropology, Canberra: Australian Na- tional University. Beaton, J. M. 1995. The transition on the coastal fringe of greater Australia. Antiquity 69:798{806. Beesley, P. L., G. J. B. Ross and A.Wells (eds). 1998. Mollusca: The Southern Synthesis. Australian Biological Resources Study. CSIRO Publishing. Parts 1 & 2. Bettinger, R. L. 2009. Hunter-Gatherer For- aging: Five Simple Models. New York: Eliot Werner Publications. Bird, D. W. 1996. Intertidal Foraging Strate- gies Among the Meriam of the Torres Strait Islands, Australia: An Evolutionary Ecological Approach to the Ethnoarchaeology of Tropical Marine Subsistence. Ph.D. dissertation, Univer- sity of California, Davis. Bird, D. W. and R. Bliege Bird. 1997. Contem- porary shellsh gathering strategies among the Meriam of the Torres Strait Islands, Australia: Testing predictions of a central place foraging model. Journal of Archaeological Science 24(1):39{63. Bird, D. W. and R. Bliege Bird. 2000. The eth- noarchaeology of juvenile foragers: Shellshing strategies among Meriam children. Journal of Anthropological Archaeology 19(4):461{476. Bird, D. W. and J.F. O'Connell. 2006. Be- havioral ecology and archaeology. Journal of Archaeological Research 14:143{188. Bird, D. W., R. Bliege Bird, and B.F. Codding. 2009. In pursuit of mobile prey: Martu hunting strategies and archaeofaunal interpretation. American Antiquity 74(1):3{29. Bird, D. W., R. Bliege Bird and Richardson, J. L. 2004. Meriam ethnoarchaeology: Shellshing and shellmiddens. Memoirs of the Queensland Museum, Cultural Heritage Series 3(1):183{197. Bird, D. W., J.L. Richardson, P.M. Veth, and A.J. Barham. 2002. Explaining shellsh vari- ability in middens on the Meriam Islands, Torres Strait, Australia. Journal of Archaeological Science 29(5):457{469. Blurton Jones, N. 1986. Bushman Birth Spac- ing: A Test for Optimal Interbirth Intervals. Ethnology and Sociobiology 7:91{105. Bronk Ramsey, C. 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51: 337{360. Charnov, E. L. and G.H. Orians. 2006. Op- timal foraging: Some theoretical explorations. Mimeo: Department of Biology, University of Utah, Salt Lake City, UT. Digital ver- sion of 1973 edition. Available online at http://repository.unm.edu/handle/1928/1649. Codding, B.F., R. Bliege Bird and D.W. Bird. 2011. Provisioning ospring and others: risk-energy trade-os and gender dierences in hunter -gatherer foraging strategies. Pro- ceedings of the Royal Society, B. 278: 2502{2509. Dortch, J. 2004. Palaeo-environmental change and the persistence of human occupation in south-western Australian forests. British Ar- chaeological Reports. Oxford: Archaeopress. Erlandson, J. 1988. Of Millingstones and Mollusks: Cultural Ecology of Early Holocene Hunter-Gatherers on the California Coast. PhD Dissertation. Santa Barbara: University of California. J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul Gadgil, M. and W. H. Bossert. 1970. Life historical consequences of natural selection. The American Naturalist, 104(935): 1{24. Gosden, C. and N. Robertson. 1991. Models for Matenkupkum: interpreting a late Pleistocene site from southern New Ireland, paupa new guinea. In Report of the Lapita Homeland Project, number 20 in Occasional Papers, pages 20{45. Department of Prehistory, Research School of Pacic Studies, Canberra: Australian National University. Hawkes, K. J. F. O'Connell and N. G. Blurton Jones. 1995. Hadza Children's Foraging: Juvenile Dependency, Social Arrangements, and Mobility among Hunter-Gatherers. Current Anthropology 36(4):688{700. Kennedy, M. 2005. An Investigation of Hunter-Gatherer Shellsh Foraging Practices: Archaeological and Geochemical Evidence from Bodega Bay, California. PhD Dissertation. Davis: University of California. Klein, R.G. and T. E. Steele. 2013. Archaeo- logical shellsh size and later human evolution in Africa. Proceedings of the National Academy of Sciences, published ahead of print June 17, 2013, doi:10.1073/pnas.1304750110. Kramer, K. 2011. The evolution of human parental care and recruitment of juvenile help. Trends in Ecology and Evolution 26(10):533{ 540. Leavesley, M. G. 2004. Trees to the Sky: Prehistoric Hunting in New Ireland, Paupa New Guinea. PhD Dissertation. Canberra: Australian National University. Leavesley, M. G. and J. Allen. 1998. Dates, disturbance and artefact distributions: Another analysis of Buang Merabak, a pleistocene site on New Ireland, Papua New Guinea. Archaeology in Oceania 33:63{82. Meehan, B. 1982. Shell Bed to Shell Midden. Canberra: Australian Institute of Aboriginal Studies. Metcalfe, D. and K. R. Barlow. 1992. A Model for Exploring the Optimal Trade-o between Field Processing and Transport. American Anthropologist 94:340{356. Morse, K. 1988. Mandu mandu creek rockshel- ter: Pleistocene human coastal occupation of north west cape, Western Australia. Archaeol- ogy in Oceania 23(3):81{88. Morse, K. 1993. New radiocarbon dates from the north west cape, Western Australia: A preliminary report. In Sahul in Review: Pleis- tocene Archaeology in Australia, New Guinea and Island Melanesia, (M.A. Smith, M. Spriggs, and B. Fankhauser, eds): 155{163. The Aus- tralian National University, Canberra. Munro, J. L. 1993. Giant clams. In Nearshore Marine Resources of the South Pacic (A.Wright and L. Hill, eds): 431{449. Institute for Pacic Studies. Nash, W. J. 1993. Trochus. In Nearshore Marine Resources of the South Pacic (A. Wright and L. Hill, eds): 451{495. Institute for Pacic Studies. O'Connell, J. F. 1995. Ethnoarchaeology Needs a General Theory of Behavior. Journal of Archaeological Research 3(3):205{255. O'Connell, J. F. and J. Allen. 2004. Dating the colonization of Sahul (Pleistocene Australia-New Guinea): A review of recent research. Journal of Archaeological Science 31(8):835{853. O'Connell, J. F. and J. Allen. 2007. Pre-LGM Sahul (Pleistocene Australia-New Guinea) and the archaeology of early modern humans. In Rethinking the Human Revolution (P. Mellars, K. Boyle, O. Bar-Yosef, and C. Stringer, eds): J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul 395{410. McDonald Institute for Archaeological Research, Cambridge University, Cambridge. O'Connell, J. F. and J. Allen. 2012. The restau- rant at the end of the universe: modeling the colonization of Sahul. Australian Archaeology 74:5{17. O'Connell, J. F., J. Allen, and K. Hawkes. 2010. Pliestocene Sahul and the origins of seafaring. In The Global Origins and Development of Seafaring (A. Anderson, J. Barrett, and K. Boyle, eds): 57{68. The McDonald Institute for Archaeological Research, Cambridge University, Cambridge. O'Connor, S. 1999. 30,000 Years of Aboriginal occupation: Kimberley, North West Australia. Number 14 in Terra Australis. ANH Pub- lications and The Centre for Archaeological Research, Canberra: The Australian National University. Orians, G. H. and N.E. Pearson. 1979. On the theory of central place foraging. In Analysis of Ecological Systems (D.J. Horn, G. R. Stairs and R.T. Mitchell, editors): 155{177. Ohio State University Press. Petchey, F., M. Phelan and J.P.White. 2004. New R values for the southwest Pacic Ocean. Radiocarbon 46: 1005{1014. Peters, R. H. 1983. The Ecological Implications of Body Size. Cambridge University Press. Pope, K.O. and J.E. Terrell. 2008. Environ- mental setting of human migrations in the circum-Pacic region. Journal of Biogeography 35:1{21. Przywolnik, K. 2008. Long-term transitions in hunter-gatherers of coastal northwestern Australia. In Desert Peoples: Archaeological Perspectives (P. Veth, M. Smith and P. Hiscock, eds): 117{205. Oxford: Blackwell. R Development Core Team. 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Com- puting, Vienna, Austria. Reimer, P. J., M. G. L. Baillie, E. Bard, A. Bayliss, J.W. Beck, P.G. Blackwell, C. Bronk Ramsey, C.E. Buck, G.S. Burr, R.L. Edwards, M. Friedrich, P.M. Grootes, T.P. Guilderson, I. Hajdas, T.J. Heaton, A.G. Hogg, K.A. Hughen, K.F. Kaiser, B. Kromer, F.G. McCormac, S.W. Manning, R.W. Reimer, D.A. Richards, J.R. Southon, S. Talamo, C.S.M. Turney, J. van der Plicht, and C.E. Weyhenmeyer. 2009. IntCal09 and Marine09 radiocarbon age calibration curves, 0-50,000 years cal BP. Radiocarbon, 51: 1111{1150. Rosenfeld, A. 1997. Excavation at Buang Mer- abak, Central New Ireland. The Indo-Pacic Prehistory Association Bulletin (The Chang Mai Papers, Vol. 3) 16:213{224. Thomas, F. R. 2007. The behavioral ecology of shellsh gathering in Western Kiribati, Microne- sia 1: Prey choice. Human Ecology 35:179{194. Veth, P., K. Alpin, L. Wallis, T. Manne, T. Pulsford, E. White, and A. Chappell. 2007. The Archaeology of Montebello Islands, North-West Australia. BAR International Series 1668. British Archaeological Reports. Whitaker, A. 2008. The role of human predation in structuring prey populations in Northwestern California. PhD Dissertation. Davis: University of California. Whitaker, A. and B.F. Byrd. 2012. Boat-based foraging and discontinuous prehistoric red abalone exploitation. Journal of Anthropologi- cal Archaeology 31:196{214. Wickler, S. 2001. The Prehistory of Buka: Step- ping Stone Island in the Northern Solomons. Terra Australis. Department of Archaeology J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript Codding, O'Connell & Bird Shellshing and the Colonization of Sahul and Natural History, and Centre for Archaeo- logical Research, Australian National University. Williams, A. N. 2013. A new population curve for prehistoric Australia. Proceedings of the Royal Society, B.doi: 10.1098/rspb.2013.0486. Yamaguchi, M. 1993. Green snail. In Nearshore Marine Resources of the South Pacic (Wright, A. and L. Hill, eds): 488{511. Institute for Pacic Studies. J. Isl. Coast. Archaeol. 9(2):238{252 DOI:10.1080/15564894.2013.848958 UU IR Author Manuscript UU IR Author Manuscript University of Utah Institutional Repository Author Manuscript