Printable View of:Lectures

File: Lecture 7 Coastal Migration and Aquatic Resources

The Evolution of the Human Diet Lecture 7 Coastal Migration and Aquatic Resources Anthro 4962/5962 Instructor Helen Alvarez

 

The mosaic habitats shown in this photograph of Rajegwesi in Meru Betiri National Park on the south coast of Java illustrate the paleoecology hypothesized for H. erectus in Java by O. Frank Huffman and his colleagues. The photo is from from www.eastjava.com

Populations of H. erectus were the first human emigrants from Africa. Traveling a route that must have been much the same as that hypothesized for moderns, Lecture 6, they reached the islands of Indonesia (review the map in lecture 5) where fossil crania have been found at Trinil, Ngandong, Sangiran, and Perning near the city of Mojokerto. A group of investigators led by O. Frank Huffman and Yahdi Zaim have just undertaken a major multi-disciplinary project to properly describe the Mojokerto site where the partial skull of a juvenile was found. In the latest publication of this project the authors describe a paleolandscape similar to a modern landscape from Mojokerto pictured above.

“The good condition of the skull and the large size of the ancient Mojokerto Delta favor the conclusion that the hominin died in the deltaic environment in which it was deposited. The Mojokerto child therefore provides evidence for a seacoast Homo erectus population in Southeast Asia, and raises interest in the role that maritime adaptation might have played in the dispersal and paleoecology of early hominins (Huffman et al. 2006 449).”

The preliminary analysis of the fauna and flora in the beds at the site indicate an environment very much similar to that pictured above where mangrove vegetation, grassland and montane forest provided a variety of resources including small and large bovids, Asian elephants, deer, and turtles.

Bantengs (wild oxen) would have come to gaze on the grassy plains between the water and the montane forest inhabited by large cats, monkeys, and pigs. The oxen shared their grazing ground with muntjaks, antelope, rhinos and hippos, some of which foraged on smaller shrubs. Preliminary analysis of fossil teeth from the site indicate that the grasses on these grazing grounds were primarily C4 tropical grasses. The analysis of phytoliths in the sediments reveal two types of grasses, open-land taxa from the subfamilies Panicoidae and Arundinoidae and the temperate climate subfamily Pooideae which probably drifted in from montane habitats. Notice from the opening photograph in this lecture the habitat diversity within a short distance from shore line to ridge line. Early humans might have varied their foraging strategies with the seasons depending upon which resources were most abundant in each habitat. Notice also that the presence of large herbivores near the shore doesn’t rule out big game hunting in habitats characterized by aquatic resources that women and children could collect.

 

As rivers flowed off the highlands they formed deltas characterized by mangrove swamps with several species of trees including Nypa fruticans palms and the sugar palm Arenga, with flowers that produce a sugary juice, and shrubs of Passiflora with edible passion fruit, and edible climbing ferns, Stenochlaena palustris. The trunks of the Nypa are submerged so the fruit, which is edible when it is immature, would have been accessible to terrestrial hominids. The delta facies contain shells of edible mollusks including oyster shells but at low density and scattered distribution. Crocodile and turtle fossils have been reported from the same sediments at other places in the Perning district. To date only one fish fossil has been identified. The photos of bantengs and nypa palm were taken from a web site of photographs of Ujung Kulon National Park but unfortunately the web site is no long available.

At other locations in Java where crania of H. erectus have been found, a succession of dry and wet climatic regimes can be documented in the fossil pollen record. The stratigraphy of the Sangiran dome, near where Sangiran 17 was found, documents a record from 2.6 M. A to 0.2 M. A. (Semah et al. 2003 ). The sediments at the Plio-Pleistocene boundary indicate the retreat of the seas and emergent land bridges between mainland Southeast Asia and the Indonesian islands. At all times, the habitat of East Java was characterized by high diversity with the hominid fossils dating to 800,000 years associated with pollen documenting a drier climate but tropical rain forest taxa still present. “In the area outside Sangiran, an Indonesian-French team has worked in the Southern Mountains of Java documenting a long sequence of occupation and adaptation to wet tropical environments from the late middle Pleistocene to the Holocene (Semah et al. 2003 p. 161).”

The second wave of human emigrants from Africa, H. sapiens, surely passed along these shores as they reached southern Australia by at least 40,000 years ago. The use of coastal, marine, and estuary resources by sapiens in Southeast Asia has been little considered as most of the focus in Asian archaeology has been on H. erectus in China and later populations transitioning from hunting-gathering strategies to farming. Erlandson (2001) attributes this neglect of coastal and aquatic adaptations to two factors, the widespread perception that hominids did not adapt to aquatic habitats before around 15,000 years ago and the obsessive attention to the idea that male-dominated big game hunting explained the origin of tool use, the formation of the nuclear family, and in its most recent incarnation the evolution of large brains. Erlandson labels one set of ideas about aquatic resources the “Gates of Hell” model as theorists proposed that humans only began to use lower quality aquatic resources when forced into marginal habitats by declining returns from hunting and/or by density dependent population growth, that is when they were forced from more desirable habitats. In the next section we are going to ask if foraging returns derived from one habitat can be used to guide inferences about returns in another time and place. This issue is especially relevant to the claims that evidence for the use of aquatic resources is evidence for habitat stress and decline of higher ranked resources.

Last resort scenarios ignore the evidence for the nutritional value of aquatic resources, the cultural complexity of societies subsisting on those resources, high human population densities in some of the most widely exploited riverine and coastal habitats, and the emerging archaeological data suggesting early adoption of aquatic resources. Instead proponents, of the last resort hypothesis, cite the small size of many of the marine resources such as mollusks, crabs, sea urchins, barnacles, and shrimp while ignoring the giant clams available on the reefs of the southern oceans, the abundance of individuals in extensive mollusk beds that characterize rocky coastlines north and south, the large number of stranded fish that can be acquired in drying pools, and nesting sea turtles as a source of eggs and meat.

Since many of the aquatic resources are sessile and predictable in time and space little search and no pursuit is involved in their harvest and interactions with other predators are rare. Further shell fish gathering requires no prior tool production, maintenance or preparation. Shellfish in particular are characterized by low variance, high density, and ease of collection, the very qualities that meet the daily nutritional requirements of growing children who can gather many of these resources for themselves. Erlandson cites a study by Jones and Richman (1995) showing that “mussel beds produce one of the highest rates of biomass production on earth (Erlandson 2001 p 294).” The supposed low quality of marine resources is derived from an estimate by Bailey (1978) “that 156,800 cockles were required to provide the caloric yield of one red deer (Erlandson 2001 p. 294).” Erlandson questions the accuracy of these estimates and later notes the analysis by Lindstrom (1996) of returns from the Truckee River fishery that are higher than return rates calculated by Simms (1987) for terrestrial Great Basin habitats. Shell fish are low calorie foods so would rank low in foraging models based on net energetic return but they have high nutritional value as sources of protein, calcium, and omega fatty acids. Ordinarily I would not post information from commercial sources but the chart comes from a USDA handbook and is a useful comparison of beef and shellfish. Later in the semester we will review arguments over the value of marine resources in the diet when we consider the research from the Okinawan centenarian study.

When you read the assigned pages in Erlandson pay attention to arguments over aquatic resources, the ambiguity of the archaeological record, the theoretical prejudices of the investigators, the nature of the resource in question, and the problem of changes in sea level with high interglacial levels destroying whatever archaeological evidence there might have been from periods of lower seas. Remember from Walter et al. (2000), Lecture 6, that the Pleistocene reef terrace at Eritrea was preserved by tectonic activity that uplifted a portion of the former shallow water reef. Notice, from Table 1 page 306 Erlandson (2001), that the sites with evidence of use of aquatic strategies span the time from 2.3M to 16.5K, a considerable time frame for the occupation of “marginal” habitats. Instead of regarding these habitats and strategies as marginal, new evidence and new hypotheses call for serious consideration of aquatic resources as central to the evolution of human dietary strategies. Often the most obvious fact of human subsistence is unmentioned because it is so taken for granted. Erlandson calls our attention to the fact that freshwater for drinking is the “single most important aquatic resource for humans (Erlandson 2001 p. 293).”

Even as our earliest ancestors exploited the resources of the woodlands and grasslands, they must have stayed close to sources of water. The important archaeological sites of Africa are all in riverine or lacustrine environments. It is hard to imagine wild foragers ignoring the resources in and and around water sources especially when those resources could be captured and processed with very simple tools. Where simple tools were not adequate to the job, early sapiens were capable of making more sophisticated tools as evidenced by the hafted bone point technology from Katanda.

Although I have emphasized coastal strategies in this lecture we know that early anatomically modern humans did move into other ecological niches as shown by the occupation of Niah Cave in Borneo. The Deep Skull, tentatively dated to the time period between 43,000 – 40,000 B.P., has been found in late Pleistocene sediments of the cave. Niah cave is one of a series of caverns in sheer-sided limestone walls that rise nearly 400 m above the lowland rain forest. The present climate in the vicinity of the cave is tropical super wet with mean annual rainfall greater than 3000 mm and rare or very short seasonal dry periods. The Pleistocene climate may have been more seasonal but changes in rainfall and temperature regimes at this site, through time, have not been resolved. At the time of Pleistocene occupation, the cave may have been about 30 km from the sea. Floral and faunal elements recovered from the late Pleistocene levels suggest that “humans were foraging in a mosaic of closed forest, scrub, grassland, swamp, and freshwater lakes and rivers (Crangrook 2000:83, cited in Barton 2005:57).” Barton, along with others interested in drawing as much information as possible from the archaeological record, has been involved in the development of a new line of evidence from archaeological assemblages, the identification of starch grains through the comparison of grains found in sediments and on tools with grains from modern starches.

Though the evidence from Niah is scanty, starch grains from palm pith and deep-rooted species of yams have been identified from the site. At other sites in Melanesia, starch grains of elephant yams, taro, ginger, and swamp taro have been identified. Some of these taros require cooking, drying or leaching to remove toxic calcium oxalate crystals but the pith of certain palms can be eating raw. These palms have high energetic yields but some processing costs as the palms have to be felled and the fiber pounded or otherwise extracted from the trunk. Evidence from sites in Southeast Asia and the evidence we will discuss in Lecture 9 from the Ache demonstrates that forest environments provided rich habitats for human foragers. Recall from Lecture 6 that the islands of Indonesia were accessible by land during glacial periods when water was locked up in northern hemisphere ice shields.

 

Assigned Reading

Erlandson, Jon M., 2001. The archaeology of aquatic adaptations: Paradigms for a new millennium. J Arch. Research. 9(4): 287- 350. For this lecture read pages 287-321.

Also Recommended

Barton, H., 2005. The case for rainforest foragers: The starch record from Niah Cave, Sarawak. Asian Perspectives 44(1): 56-72. (Asian Perspectives available as electronic edition Marriott Library) Huffman, O. F., Y. Zaim, J. Kappelman, D. R. Ruez Jr., J. de Vos, Y. Rizal, F. Aziz, C. Hertler. 2006. Relocation of the 1936 Mojokerto skull discovery site near Perning, East Java. J. Hum. Evol. 50: 431-451. Huffman, O. F., and Yahdi Zaim. 2003. Mojokerto Delta, East Jawa: Paleoenvironment of Homo modjokertensis First Results Semah, Francois, Semah, Anne-Marie and Simanjuntak, Truman, 2003. More than a million years of human occupation in insular Southeast Asia. In Under the Canopy: The Archaeology of Tropical Rain Forests. pp 161-190. ed. J. Mercader New Brunswick, NJ: Rutgers University Press. (available from Marriott Library, Course Reserves)

File: Lecture 8 Optimal Foraging Theory

Anthro 4962 The Evolution of the Human Diet Lecture 8 Optimal Foraging Theory University of Utah Fall 2005 Helen Alvarez

 

Shores of Walker Lake (The North American Indian; v.15) from Northwestern University Library,  Edward S. Curtis’s ‘The North American Indian’: the Photographic Images, 2001. Curtis described the setting of this photo: Walker lake, one of numerous saline lakes remaining from a great inland sea that once covered western Nevada and northeastern California, is the seat of a major division of the Paviotso. In the western corner of Nevada it is fed by Walker river, the numerous branches of which head on the eastern slope of the Sierra Nevada in California. Although there is no outlet, the water is not too saline for the thriving of trout and suckers, which were taken on bone hooks, with double-pointed spears, and in gill-nets.

 

The lectures to date have focused on the evolution of Plio-Pleistocene diets inferred from the archaeological evidence and from paleoecology informed by modern food preferences through an implicit assumption that the environment is populated with nutritious plants and animals to be captured by individual hominids ranging over that landscape. Most of you know from other classes and from your own diets that foragers are selective in their choices. The simple coincidence of resource and forager in the same habitat tells us little about the diets of the forager. To be fair we have looked at circumstantial evidence for use of some resources such as cut marks on bones, opened and cracked shells, processed fish bodies, special tools for capturing specific animals, and isotope ratios but all of that evidence provides only a coarse grained picture of early diets. Archaeologists interested in refining their assessments of past foraging strategies have applied models from optimal foraging theory (OFT) to help them understand the past. These models were originally developed by biologists interested in explaining the specific resource choices they observed in their study populations. They were subsequently used by evolutionary anthropologists to study diet choices in modern foragers subsisting on wild resources. Most of you know that the faculty and students from the University of Utah departments of biology and anthropology were very active in developing and applying these models.

 

The first efforts toward understanding resource capture were undertaken by a Canadian entomologist, C. S. Hollings who designed a simple experiment modeling predator and prey interactions. The prey were represented by 4 cm sandpaper discs tacked to a 9 sq. ft. table. The predator was his blindfolded research assistant who tapped her fingers over the table until she encountered a disc, which she removed and set aside before returning to the task of “capture”. The trials were iterated at various prey densities and the relationship between rate of capture and disc density was plotted giving the marginal value curve shown on the right. The rate of capture does not increase in a linear fashion with the increase in the density of the prey.

Hollings’s results became the basis for the distinction between search and handling in the subsequent development of OFT. His simple experiments showed that search time was a function of prey density on the landscape while handling time after encounter was a function peculiar to the prey type. These distinctions became important in all subsequent OFT models.

Optimal foraging theory drew attention to the empirical fact that every resource has a cost of acquisition and processing. These costs make comparisons of energy yields of food meaningless unless handling costs are considered in the comparisons. Here I have posted Table 1 from Bright, Ugan and Hunsaker (2002) showing the energy yield of certain resources taken by Australian and Great Basin foragers and the Kcal/hr return for those resources after processing. I have sorted their table from highest net gain to lowest and plotted the yield with and without processing to show you how the rank of resources changes when the handling costs are debited.

Notice on the figure, drawn from the table, that resources with high energy density (red line) are not necessarily the resources that yield the highest returns after processing (blue line). From this set, the highest ranked resources are all animals with tubers ranking just below small game. Compare the yield of rabbits (#3) and pine nuts (#11) on both curves to see how processing costs change the relative rank of resources. If meat is always the highest ranked resource, across many habitats, why should human foragers ever take any other resource? You should be able to answer this questions by the end of Section II.

 

How can the information from OFT be reconciled with the Grandmother Hypothesis for the evolution of human life histories from O’Connell et al. 1999 (Lecture 5)? The fundamental premise of OFT is sound as we expect natural selection to favor individuals that allocate energy to tasks that promote somatic growth, maintenance and reproduction. In any habitat those genotypes better at allocation decisions are expected to leave the most descendants, that is have the highest fitness, yet the expectations of the diet breadth models are often violated in real populations. All models assume a generic forager but the early work with the Ache demonstrated that men routinely ignore the highest ranked resource in the forest, palm starch, in favor of hunting. This falsification of the model was as important as the confirmation of most of the predictions for other resources because it raised the question of why men hunt which led through many populations and many papers to important ideas about the role of costly signaling in male-male competition and linked sexual selection theory and foraging theory. But sexual selection on males is not the only factor determining diet choice as risk and uncertainty influence forager choices.

The idea that the value of a resource item is determined by the costs of search and handling changed the way investigators ranked prey items in the foragers environment. Many resources are energy dense but have high search and handling costs. It is not the absolute value, to a forager, of the resource but the net energy gain to the consumer after the energetic costs of pursuit and handling are debited. In most investigations of actual foragers net energy gains are calculated by debiting post encounter costs against the currency of the gain, usually calculated in caloric return per unit of time spent in pursuit, capture and handling.In the simple version of the optimal diet model resources are included in the diet based on the tradeoff between handling upon encounter and the option of continuing to search for other items. The decision is based on the mean rate of return for that particular habitat and the suite of resources within it. Be aware that post encounter returns are always much larger than returns reported with search time included unless search time is nil. The inclusion of processing costs in the equation reflects the net gain to the forager from choosing any resource and influences forager preferences for certain resources over others.

Richard Lee’s failure to include processing time in the cost of mongongo nuts seriously under estimated the contribution of women’s work to family energy budgets and led to the characterization of the Dobe !Kung as the the original affluent society, meeting their subsistence needs with few hours of work. I have assigned two short readings, Sih and Milton (1985) and Hawkes and O’Connell (1985) debating the the use of OFT to study human resource choices and specifically the ranking of mongongo nuts in !Kung diets. These two papers will give you a sense of the early arguments over the application of OFT models to human foragers and the value and problem of simplifying assumptions. As you have already read in Erlandson, the failure to report search time for resources leads to the over estimation of returns from resources such as collared peccaries and makes comparisons of post encounter returns of resources, such as red deer and mussels from two very different habitats, meaningless.

Note in all of the lectures that follow I have tried to maintain the notation, calories, cal/hr, Cal/hr, or Kcal/hr used in the assigned papers. The notation is confusing, but the link provides a simple explanation of the differences. For our purposes we assume that investigators are referring to Kcal when they use cal, Cal or calories to indicate the energetic gains from any resource.

 

Risk, Uncertainty and the Forager’s Energy Balance

 

Winterhalder, Lu and Tucker (1999 p. 302-303) distinguish between risk and uncertainty by defining risk as “unpredictable variation in the outcome of a behavior, with consequences for an organism’s fitness or utility.” Further “with risk the probability distribution of outcomes is in some sense known to the organism, but stochasticity makes any particular outcome unpredictable.” “Uncertainty refers to incomplete knowledge of outcome probabilities. Uncertainty can be overcome by acquiring information about an environment; risk cannot.” They make the point that risk as used in behavioral ecology and economics does not “mean exposure to danger.” The authors point out that simple foraging models assume the forager always experiences the average conditions of its habitat where decisions have predictable outcomes. This assumption is central in all debates over the role of big game hunting in the evolution of the human diet.

When we examine Hadza foraging strategies we will consider the high variance in game capture and contrast that to figures for the average amount of meat in the diet. In contrast to deterministic models, risk sensitive models assume organisms adjust to a range of possibilities arrayed along a probabilistic distribution. In addition risk-sensitive analysis specifies the relationship between the outcome and its fitness value resulting in a sigmoid curve. To the left of the inflection point of the curve value rises with increasing resources but at an accelerating marginal rate when the resources are scarce and at a decelerating marginal rate when they are abundant. In simple language, with too little food “added increments have high value and with too much, added increments count for little (Winterhalder et al. 1999 p 304).” The work of Curaco and his colleagues, with yellow-eyed juncos, showed that the utility function for the birds takes the sigmoid form. There is no reason to suspect that hominids are any less responsive to variable outcomes. If you are interested in further exploration of these ideas the figures presented in the Winterhalder et al. (1999) paper will give you a much better sense of the issues. I have posted the paper under also recommended for those of you interested in a more detailed explanation.

Since value is measured as a rate, usually in calories gained per unit time, the sigmoid value function is time sensitive and this sensitivity varies with respect to the physiology of the organism and the character of the resource. Think back to Lecture 1 and the link to the Moran Eye center Web vision site with the graphic demonstrating the daily rebuilding of the pigment containing discs of the rods and cones. Stores of vitamin A must be sufficient to support the daily demands of this process. Winterhalder et al. (1999) make the same point about water acquisition. The preference for a reward of k liters/hr and equal probabilities of 0 or 2k liters/hr will depend upon whether the result will be suffered for hours, days or weeks before choice and outcome are reiterated.

Responses of the organism to risk are determined by design by natural selection over the long time frame and the physiological state of the organism over shorter time frames. In general females are expected to be more risk adverse than males and organisms in negative energy balance are predicted to be more risk prone. The latter prediction was confirmed in experiments with yellow-eyed juncos whose foraging choices were observed under different temperature regimes that influenced energy balance. Given two foraging options with constant or variable rewards, birds in negative energy balance favored the high variance reward, presumably in the attempt to gain a reward sufficient to change their energy balance. The excitement generated by these results was tempered by a number of experiments that failed to replicate the results. Winterhalder et al. (1999 p. 323) conclude; “These studies suggest that the expected energy budget rule may apply only rarely to hominids, nonhominid primates, and modern humans which are omnivorous and relatively large species.”

They might have qualified their statement to include only adult humans as anthropologists and archaeologists, with few exceptions (cf. O’Connell et al. 1999), have ignored the consequences of short periods of negative energy balance on the growth and development of neonates, infants, and juveniles. Conversely, resistance to short-term perturbations in energy balance may be an advantage of large body size that contributed to the success and expansion of H. erectus. We need to be careful here not to attribute the evolution of large body size to a resistance to short term negative energy balance. Such resistance can be a consequence of large body size but the causal arrow on the evolution of large body size runs from mortality risks in the environment of the species. Following Charnov’s (1993) mammalian life history model, large body size is a consequence of more time to grow larger because lower average adult mortality, from extrinsic causes, favors delaying the change from somatic to reproductive allocation. Organisms that experience low average adult mortality rates, grow longer, mature later and have longer lifespans. These are the derived traits that distinguish us from our nearest primate relatives. These arguments were reviewed by O’Connell et al. (1999).

 

Great Basin Archaeology and Optimal Foraging Models

 

The last assigned reading for this lecture is a paper by Elston and Zeanah in which they use models from behavioral ecology to understand the Holocene transition evident in the archaeological record of the Great Basin. Most of you know that our department has been in the forefront of extending foraging theory to hunter-gatherers but you may not be aware of the innovative ways the models are being applied to solve longstanding problems in Great Basin archaeology. The paper by Elston and Zeanah is a good example of this work as well as providing a link, in this course, from ancestral populations known only from the archaeological record and the next set of papers from more recent foragers. The opening graphic of this lecture, from the work of Edward S. Curtis, is intended to emphasize the importance of Great Basin studies to applications of optimal foraging theory to human foragers and to highlight the importance of ethnographic analogs to this work.

The record of human occupation in the basin can be divided into two distinct time frames labeled Prearchaic for the period from 10,500-8000 BP and Archaic for later occupation. The early Holocene (EH) climate experienced by Prearchaic populations was cooler and wetter than climates after about 7800 BP. “The relatively cool, even EH climate, abundant surface water and complex steppe vegetation created productive habitats for a rich biota of fish, waterfowl and mammals (Elston and Zeanah 2002 p. 107).” The tool assemblages from this time period have an array of projectile points and flaked tools indicating a hunting economy but the coprolites recovered from the sites indicate a diet that included many of the resources of the lakes and marshes plus seeds and small animals, quite unlike a diet predicted from the graph of energy yield posted above. Basin occupants in this time period had access to a variety of resources within short travel distances as basins with lakes and marshes were separated by ridges with brushy steppes and juniper woodland from mid-elevation to ridge tops providing high-quality habitat for larger game animals. Moreover the scatter, size and character of the archaeological sites from this time period indicate low population density and high mobility of foraging groups.

An attempt to use the diet breadth model to explain the evidence for the addition of seeds to the diet failed as the model developed by Simms predicted that women should have by-passed seed under all conditions, including the absence of large game, a prediction contrary to the clear evidence for seed use. A second model analyzed the distribution of soil types in the basin to develop a patch choice model. The simulations based on the distribution of lowland and upland habitats predicted that women should bypass seeds in favor of other plant resources and small game and men should have targeted only large and medium size game. This model fails to explain the evidence for small animal consumption found in the coprolites. Finally a risk sensitivity analysis returned the prediction that risk prone foragers should favor the high variance option “preferentially pursuing large game over smaller game and foraging in habitats where encounters with large game were most likely (Elston and Zeanah 2002 p. 115)” while risk sensitive foragers should prefer lower-ranked resources characterized by low variance in acquisition. From this model, Pinson (1999, cited in Elston and Zeanah 2002) hypothesized that Prearchaic foragers avoided starvation risk by pursuing small game. That is they were risk adverse.

Elston and Zeanah (2002) note the following problems in Pinson’s analysis; 1) the model ignores the possibility that Prearchaic foragers might have entered basins where foraging returns did not meet daily requirements at which point the risk of starvation should have prompted the risk prone strategy of choosing the high variance option of large mammal hunting, 2) if Prearchaic foragers were risk adverse they would be expected to forgo migration from known basin habitats to new, unknown locations, and 3) if EH environments were as patchy as predicted by Pinson, the Prearchaic archaeological record should more nearly correspond to the Archaic record when climate deterioration reduced the density of high ranked prey. Increasing search time would have been a consequence under both conditions.

Zeanah et al. (1999) used some of the techniques from Pinson’s analysis of Carson Desert foraging in simulations of optimal foraging in Railroad Valley but in the latter analysis “more detailed regional palaeoenvironmental data, better information on the productive capacity of modern soil types and improved GIS (Geographic Information System) capabilities permitted a more finely tuned reconstruction of prehistoric foraging landscapes than was feasible in the Carson Desert (Elston and Zeanah (2002 p. 117).” The Railroad Valley simulations considered male and female foragers randomly encountering resources across spatial and seasonal variation. From the early Holocene (EH)to middle Holocene (MH) hunting returns for men diminished by as much as 75% in all seasons. By contrast women’s returns varied much less across the transition, increasing somewhat in autumn as pinyons entered the ecosystem. As habitats dried out in the MH, wetland resources disappeared as options for women allowing lower ranked small seeds to enter the diet in late winter-early spring.

The authors conclude that the highly productive ecosystems of the EH and the low population numbers encouraged high mobility that allowed men and women to forage in the most productive habitats for their resource choices. As the climate changed and women’s wet land resources declined, Great Basin foragers faced the risk of starvation in certain seasons. In the face of this risk women collected small seeds in seasons when they were abundant and stored them for the short season. Seed caches changed the pattern of mobility for both men and women, encouraged residential bases, lower investment in tools for large game hunting and more investment in tools such as grinding stones for seed processing. Bright, Ugan and Hunsaker (2002) predicted increasing investment in processing tools to reduce handling cost as lower ranked resources enter the diet in response to a decline in the density of higher ranked resources. The change in grinding stone technology, through time, in the Little Boulder Basin area of north-central Nevada supports that prediction. These innovative approaches linking optimal foraging theory, resource changes through time, and changing investments in technology increase our confidence that ancestral diet choices can be studied and understood. In lecture 9 we will look at how foraging models, applied to resource choices in modern hunter-gathers, have improved our understanding of human diet choices.

 

Assigned Reading

Sih, Andrew and Milton, Katherine. 1985. Optimal diet theory: Should the !Kung eat mongongos? Amer. Anthropol. 87(2):396-401. Hawkes, Kristen and O’Connell, James F. 1985 Optimal foraging models and the case of the !Kung. Amer. Anthropol. 87(2):401-405. Elston, R. G. and Zeanah, D. W. 2002. Thinking outside the box: A new perspective on diet breadth and sexual division of labor in the Prearchaic Great Basin. World Arch. 34(1):103-130.

Also Recommended

Winterhalder, B., Lu, F., and Tucker, B. 1999. Risk-sensitive adaptive tactics: Models and evidence from subsistence studies in biology and anthropology. J. Arch. Res. 7(4): 301-348.

 

 

 

 

File: Lecture 9 The Ache: Broadleaf Evergreen Foragers

Anthro 4962 The Evolution of the Human Diet Lecture 9 The Ache: Broadleaf Evergreen Foragers University of Utah Fall 2005 Helen Alvarez

 

Ache Man Carrying the Head of a Tapir photo from anthrophoto.com

The Ache are foragers of the subtropical, broadleaf, evergreen forest of Paraguay. I chose this example because Ache foraging has been intensively studied by the Utah Ache Project and as a comparison to semiarid African strategies. The warm, wet summer regime of the Ache forests are similar to habitats widely distributed from South America to Indonesia between 20 degrees north and south of the equator in regions with annual average precipitation varying from 1600 mm to 2000 mm. Temperature regimes in these habitats are characterized by warm summers and warm to cool winters with January maximums in the Ache region of around 40 degrees centigrade and July minimums of about -3 degrees centigrade. Early anatomically modern humans moving south toward Australia and north into China, Viet Nam, and Thailand must have encountered these habitats and foraged in them. The early Pleistocene archaeological record of upland Southeast Asia is little explored but the Ache can serve as a reasonable analog of human foraging strategies in humid subtropical ecosystems. Keep in mind that the distribution of these ecosystems, through time, was determined by northern hemisphere glacial cycles. On the Koppen climate classification map pictured below the Ache habitat is shown in green, warm to hot wet summer, cool dry winter.

Note the distribution of zones colored purple, pink, pale yellow and medium green. All of these zones are characterized by wet conditions during the season of maximum solar insolation.

The required reading for this lecture, Hawkes, Hill, and O’Connell (1982) is one of the earliest reports of the use of OFT to analyze resource choices of humans foraging on wild food. Even though the preliminary data in this paper has been revised and corrected in subsequent papers, I have assigned this early paper because it provides a brief history of the arguments, in anthropology, over human diets and because of the broad picture it provides of Ache foraging. The northern Ache were full-time hunter-gatherers until mid-1970 but now live at settlements sponsored by a Catholic mission. At the time the research was conducted by the University of Utah Ache project, they still engaged in foraging expeditions and lived on wild resources during their time away from camp. Hawkes et al. (1982,) contrast the views of Richard Lee, developed from his field experience with the !Kung, to those of Marvin Harris developed in a series of debates with Napoleon Chagnon over the causes of violence among the Yanomamo. Harris argued that the Yanomamo fought over scarce protein resources while Lee’s tally of foods eaten daily in !Kung camps convinced him that plant foods were more important than meat. Harris argued that meat was nutritionally superior to plants and would only be replaced by plants when large game animals had been depleted by hunting pressure. The important themes that would subsequently dominate debate over foraging strategies in modern hunter-gatherers are set out in Hawkes et al. (1982 p. 380).

“Lee argues that plant foods are favored because they are abundant, reliable, and readily located, and therefore more efficiently exploited than are animal foods. Plants are said to be low-risk/high-return resources, while animals are high-risk/low-return resources. Animals are taken in spite of the inefficiencies involved because of the taste appeal of the meat and the prestige that accrues to successful hunters.”

Hawkes et al. (1982) use two models from OFT, the optimal diet model and the patch choice model, to predict the resources the Ache should gather if they are maximizing net energy return. You were introduced to the optimal diet model by Elston and Zenah (2002), lecture 8, so you should know that the forager is assumed to make a decision about whether to take a resource upon encounter or to continue searching. Resources are ranked based on the return in calories over the post encounter processing times including both pursuit, for mobile prey, and processing to turn the resource into food. The forager is predicted to take only those resources that give a return rate equal to or higher than the average rate for resource in the optimal diet. Recall the Hollings Curve which showed that search and handling time are distinct elements of the problem, the search time is determined by the density of prey on the landscape but the net returns from that prey are determined by the handling time. It is the net return, after handling, that determines the rank of the prey in the diet.

In the simple form of the model high ranked prey should always be taken when encountered. Hawkes, Hill and O’Connell make a very important point on page 388: “Note that the resource rankings of this model say nothing about the quantitative importance of a resource to optimal foragers. High-ranked items may be so rarely encountered that they represent only a very small portion of the diet; low-ranked items in the optimal set may be encountered with sufficient frequency to contribute the bulk.” Keep this in mind when we get to the disputes over the importance of various foods for contemporary human nutrition. Three additional points, 1) when the cost of search is added to the cost of any resource, its value to the forager, in absolute terms, may decline, 2) all optimal diet models are specific to a habitat and time period, and 3) the optimal diet model assumes a fine-grained environment where resources are encountered at random.

Foraging trips with the Ache showed that the environment is coarse-grained or patchy with many resources clumped. On foraging trips the Ache often stop to harvest in some resource patches but not others. The authors suggest that the “distribution of tools” (Hawkes et al. 1982 p. 391) might account for the decision but discard that suggestion because the Ache nearly always take oranges and honey, even though both are harvested with the use of axes, but often pass palm fiber. The patch choice model predicts that foragers use patches that produce the best return when travel time, search within the patch, and handling time are considered. Notice that return figures for patchy resources include the time for searching within the patch. If both animals and oranges are considered as patches the average energetic gain return for hunting, including search is about 1115 Cal. per hunter-hour while the return for oranges is 4438 Cal per forager-hour and the return for honey is 3231 Cal/hr. Palm larvae has a similarly high return rate of 1849 Cal per forager-hour but involves the risk of high variability across logs. Returns calculated for fishing patches were similarly variable, ranging from > 2000 Cal per forager-hour to about 733 Cal per forager hour leading to two questions, when should foragers exploit palms and when to take fish. The average return rate in calories per hour of sixteen resources taken by the Ache are listed in Table 3, page 389. Notice that collared peccaries are a very high ranked resource but no search time is factored into the estimate of return.

Summary: Five Years of Research

Subsequent analyses of the data from five years of research on Ache resource choices reported in Hill et al. (1987) highlighted both the value and problem of using optimal foraging theory for analysis of human foraging choices. First detailed records showed that a theory for generic foragers failed to capture the different strategies of males and females. Among the Ache, men achieve significantly higher returns, 1253 Cal/hr, than women, 1087 Cal/hr. These figures are averaged over all resources taken by Ache men and women and include both search and handling costs. Before processing the rates are more nearly the same, 1339 Cal/hr for men and 1221 Cal/hr for women. The differences in rates indicate the high cost of processing palm fiber. Men often by pass palms but when they do take them they gain higher rates because they process the fiber faster than women. Consistent with the predictions of OFT models only one of the 26 resources taken by the Ache gave returns, after encounter, lower than the overall mean. That resource was palm larvae, mostly taken by women but sometimes also collected by men. Contrary to the predictions of the theory that foragers should preferentially target high return resources, returns from adult male hunting, 1349 Cal/hr, were lower than the caloric return, 2630 Cal/hr that could be obtained from palm starch and hearts (Hill et al. 1986).

The observations that men gain higher returns than women, on average, changes if returns are compared for days when camp is not moved. On days when the group is moving women carry babies, pets, household belongings, and any meat the men catch. Men keep hunting after turning over meat they have already caught. Because of the opportunity cost of processing, most Ache processing work is completed in camp at the end of the day. When the Ache women are moving they pass many resources they might otherwise gather so their return rates on camp move days is very low. By contrast when they stay in one camp for several days, the women achieve a return on gathering of 2804 calories per hour compared to a return rate for men, on these days, of 1344. These figures include all time spent in food acquisition and processing.

These results raised the question of why men hunt and suggested two alternative hypotheses, 1) caloric return per unit of time is not the only currency by which foraging strategies should be judged and 2) fitness payoffs for hunting are broader than meeting subsistence requirements. The first hypothesis proposes, in agreement with Harris, that meat is more highly valued than plant food because of high protein and lipid content. You can evaluate this hypothesis by recalling the human RDA for lipids and protein in the diet as reported by Conklin-Brittain et al., lecture 2 and the values of lipids and proteins that can be acquired from plant foods. When you open the link search on RDA to compare values in chimpanzee diets to RDA for humans. Also have a look at the figures and ask if annual averages reported in the charts are misleading. Monthly mean lipid levels in the diets are much more variable than monthly mean protein levels. Even though the annual mean lipid level in the chimpanzee diet seems sufficient to meet the RDA for humans the high variance may not provide adequate daily lipid consumption. The second hypothesis, suggested but not fully developed in this paper, is that men opt for a high variance strategy that promises a large payoff when successful. The payoffs for hunting include not only the calories, proteins, and lipids gained from the meat but mating benefits for good hunters. This last payoff is more fully developed in the papers on Hadza hunting.

The work with the Ache demonstrated that human foragers differ, by age and sex, in the amount of time they spend foraging, the resources they take, and the amount of time they spend processing. The resources a forager takes on any day is not only a function of the costs of pursuit and handling but of the foragers state and the time frame over which the subsistence needs exist. Lower ranked fruit that can easily be gathered and eaten while moving will often be collected to satisfy short term needs while large game and palm starch are taken to satisfy needs over several days. In all cases the ranking of resources is habitat specific and even in the relatively stable habitat of the lowland deciduous forest can vary by season. The Ache only hunt and dig armadillos in the late wet season when the animals are fat and the average return is 3948 cal/hr compared to a return of 1220 cal/hr in the early wet season.

The emphasis on highly ranked resources might leave the impression that the Ache diet is very narrow but the list of wild plants and animals taken by the Ache suggests otherwise. Have a look at Table 2 from Hill et al. (1987), below to get some idea of the variety in the diet. Remember larva are the only resource that reduces the average return, over all resources included in the diet. The optimal diet model predicts that the Ache should never take bamboo larva but ethnographic observations indicate larva are often collected. Do the Ache take larva because they value the high fat content of larva? This question is still the subject of debate so no clear answer but exceptions to the predictions of OFT do not diminish the usefulness of the model. Instead they open new windows into human foraging strategies.

Hill, Kaplan, Hawkes, and Hurtado (1987) Foraging decisions amount Ache hunter-gatherers: New data and implications for optimal foraging models. Ethology and Sociobiology 8: 1-36.

The Ache are exceptional, among studied hunter-gatherers, in the number of calories they consume daily, 3610, and in the amount of meat they eat, 80% of calories from meat. Notice in Table 2, the difference in return rates depending upon the included costs. The return for White-tipped peccary is 5323 calories/hour if time spent tracking is debited but as high as 8755 calories per hour if only post encounter rates are considered. Similarly Ache men obtain a much higher return for 9-banded Armadillos they encounter on the surface compared to those they dig up. The variation in these rates makes it nearly impossible to compare return rates across foragers if the ethnographers don’t clearly distinguish between the methods of calculating rates. Keep these problems in mind as we survey foraging strategies across habitats.

Assigned Reading

Hawkes, K., Hill, K., and O’Connell, James F. 1982 Why hunters gather: optimal foraging and the Ache of eastern Paraguay. Am. Ethnol. 9(2): 379-397.

Also Recommended

Hill, K., Kaplan, H., Hawkes, K., and Hurtado, A.M. 1987 Foraging decisions among Ache hunter-gatherers: New data and implications for optimal foraging models. Ethology and Sociobiology 8:1-36.

 

 

 

 

File: Lecture 10 Semiarid African Strategies

Anthro 4962 The Evolution of the Human Diet University of Utah Fall 2005 Helen Alvarez Lecture 10 Semiarid African Strategies

 

!Kung Mother with Her Baby on Her Back Gathers Berries photo by Marjorie Shostak from www.anthrophoto.com

In this lecture we focus on the foraging strategies of the !Kung and Hadza hunter-gatherers of the semiarid region of Africa. The work of Lorna and John Marshall and Richard Lee made the Kalahari San the archetype for hunter-gatherers and the evolution of human foraging strategies. Their work in Southern Africa was complemented by the early work of James Woodburn among the Hadza of Tanzania and the detailed work of James O’Connell and Kristin Hawkes among the same group. As a consequence of these efforts we know more about human foraging in semiarid environments than in almost any other habitat. You viewed the map of present climatic regimes in the last lecture. Go back and have a look at the distribution of semiarid lands in Africa, the bright yellow and three shades of orange areas. The !Kung San live in the Kalahari desert of southwestern Botswana and eastern Namibia and the Hadza occupy the area around Lake Eyasi just south of the equator in Tanzania. Both areas are characterized by seasonal rainfall and shortages of surface water in the dry season. In these semiarid, seasonal rainfall climates, precipitation is highly variable from year to year and across the region in any one rainy season. In the Kalahari, over eleven years, precipitation varied from a low of 252.1 mm in 1962 to high of 788.9 mm the next year. Temperatures in the Kalahari vary from a high of 40 degrees centigrade in November-December to a low of nearly -10 degrees centigrade in July and August. The climatic regime of the two regions is very similar but the topography couldn’t be more different. These topographic differences have important implications for differences in foraging strategies of !Kung and Hadza children. In the lecture on children’s strategies we will consider those differences. This is a long lecture but the reading assignment is not excessive and I promise you a short Lecture 11.

Hadza Country Hadza country is characterized by rocky hills covered with mixed Acacia scrub and grass leading down to narrow, grass covered valleys. Foragers walking through this landscape to hunt or gather have vistas and landmarks to guide them. In this area children are able to forage away from home and find their way safely back to camp. We will come back to the habitat differences when we discuss juvenile provisioning and ask why the !Kung children are not able to gather more of their own resources.

!Kung Bushland The Kalahari is a vast plateau at 1100 m above sea level covered by grasses and woodland trees growing to about 30 m in height. It is a nearly featureless plain with few landmarks to guide the hunters and gatherers who make their living from its resources. Even though habitat variability seems low there is some topographic relief provided by lower lying, shallow, seasonal river beds, called molapo, and frying pan shaped dry lakes which hold water for a few days after a heavy rain but are otherwise hard and dry most of the time. After the rains, the Kalahari plains are grass covered. The higher land supports low shrubs that give way to taller woodland trees on the sand dunes and finally, on the outermost zone, open scrub plain. Each of these habitats provides a unique suite of resources for the !Kung. The woman introducing this lecture is gathering berries that grow on the scrub plain.

The widely read early work and films from Lorna and John Marshall and later the careful measures of !Kung foraging returns by Richard Lee established the !Kung as archetypal human hunter-gatherers even though their habitat represents a very small percent of the habitats occupied by early H. sapiens. The notion that early humans evolved in an ecological niche similar to those occupied by the !Kung and Hadza focused the attention of researchers on big game hunting as the principle dietary strategy of humans to the exclusion of other foraging strategies. You were introduced to this problem in the reading from Erlandson (2001), lecture 7. Review it here so that you can put the dietary strategies of semi-arid lands in the context of variable human responses to the opportunities presented over the long migration from Africa. Once Richard Lee measured the actual food brought into a !Kung camp, in the Dobe area of the Kalahari, he came to the conclusion that plant foods represented the principal component of the diet. The Dobe !Kung men do hunt but the returns are highly variable. In spite of the variable nature of returns from meat, it is a highly desired resource. When a large game animal comes into camp everyone eats meat but on the many days when there is no meat in camp, men, women and children do not go hungry. They enjoy other foods gathered from the sandy plains of their homeland.

I have assigned chapter 2 of Jiro Tanaka’s 1980 report on the !Kung San of the central Kalahari. You can download the reading from Marriott Course reserve. Tanaka’s field work was accomplished in the sixteen months he spent in the field between 1966 and 1968, sixteen months from April 1971 to August 1972 and two more months in 1974. In his introduction to San, Hunter-Gatherers of the Kalahari, Tanaka makes the point that desert is a misnomer for this region as it is an expansive, monotonous, grassland dotted with open woodland trees and shrubs characterized by a summer rain season from December to March. Tanaka gives you a study of diet in an area where there are no mongongo nuts, the most important dietary staple in the area where Richard Lee worked. Richard Lee made the San famous through his measurements of the number of hours spent gathering food. Since he didn’t count processing times his estimates left the impression that the !Kung could satisfy their dietary needs in no more than 2-3 days per week. As you know from reading the Hawkes and O’Connell (1982) response to Sih and Milton (1982) Lecture 8, mongongo nuts have particularly high processing costs.

The principal foods of the central Kalahari San are listed in Table 8. Even though Tanaka did not use the framework of OFT to study resources gathered, the first eleven foods listed in Table 8 constitute the important resources in the diet. Notice the importance of melons and roots, that store water, for a people who have no access to surface water through much of the year. Tanaka attributes the ability of the San to survive in this very habitat to the technology of the digging stick which allows humans to access deeply buried tubers that cannot be gathered by any other primate, including savannah baboons who live on shallow rooted species in other habitats but need access to surface water so are unable to colonize the central Kalahari. Four of the important resources are shown in the next table. I was unable to find a good picture of Bauhinia beans. The shrub is a beautiful ornamental of the nursery trade so all of the online images feature the flowers and not the pods and beans.

The four resources pictured here are not quite to scale but give you an idea of the foods listed in Table 8. First is Citrullus lanatus, the wild version of our cultivated watermelon. Second in the top row is Acanthosicyos naudiniana, the Kan melon. Surprisingly these melons can be gathered and kept for many months to supply water and some nutrients. Notice the relative size of these two melons in Tanaka p 60. The tubers that are so important when melons are not available are illustrated in the next panel. This tuber is probably Coccinia rehmannii. The final panel shows a Moroccan truffle of the same genus as the Kalahari truffle, Terfezia. North African truffles are now marketed to the global gourmet food trade.

Consistent with OFT, Tanaka (p 59) reports that beans of Bauhinia petersiana are the “central food item for several months” making “up nearly the whole of the San’s diet” in months when they are ripe and water is available. “Even though this is the period of greatest abundance and variety of food types during the whole year, people often totally ignore other foods.” This latter is somewhat surprising as the beans must have high carrying and processing costs since the inedible pod accounts for 75% of the whole weight so approximately 5 kg of pods and beans are carried back to camp to yield 1 kg of edible beans. Table 12 in Tanaka gives nutrient analysis of the important plant foods and Table 13 provides an estimate of the daily caloric intake per person. The figures are not intended to be totaled but to show what might be gained if 5 kg of Kan melon were eaten each day.

On page 74, Tanaka gives an estimate of 2,000 kcal for the per capita yield of food taken during the study period and finds that return consistent with the caloric needs of individuals of the size and weight of adults in the population. Compare the protein content of the dried beans of T. esculentum and B. petersiana and dried Terfezia to the protein composition in animal foods, Table 12 and Table 14, Tanaka. We can’t be sure the units reported are the same but if we assume g/100 grams in each case the plant foods are comparable in protein gain to the animal foods. T. esculentum, marama bean, is being developed as a commercial crop in arid regions of Australia and Texas as it “is an excellent source of good quality protein and compares well with other protein foods including soybeans. Its oil is rich in mono- and di-unsaturated fatty acids and contains no cholesterol. It is also a good source of calcium, iron, zinc, phosphate, magnesium, B vitamins and folate (quoted from the web abstract). ”

The number of game animals taken during the study period is shown in Table 11. Giraffes are the largest animals hunted by the San but they are rarely captured in the very dry habitat of the Central Kalahari. More commonly captured are the gemsbok, eland and kudu each weighing 200 to 300 kg. When the amount of meat taken is averaged over all camps, Tanaka estimates that a group of 50 San kill and share approximately 5,600 kg of game in a year or about 112 kg per person or .30 kg per day. From his description we know that meat comes into camp rarely so averages are hardly adequate to estimate the amount of meat eaten. When a large animal is captured people eat nothing but meat until the animal is finished and then go without meat for many days. On page 67, Tanaka describes the sharing pattern for game, large game animals weighing 100 to 300 kg are shared among all the members of the camp, smaller animals such as duiker and steenbok are distributed over a smaller circle of families and still smaller game, birds, springhares and hares are “usually consumed within the family.”

It is difficult to get a sense of the size of the eland in this photo but if you convert 300 kg to lbs these animals weigh about 660 lbs. Compare that to beef cattle weighing from 600 to 900 lbs. Lee estimates the edible fraction of game taken by the Dobe !Kung to be 50% of the live weight or 150 kg for an eland. In Table 5, reproduced below, he lists per capita consumption of meat at 230 grams. Calculating .230 kg per consumer day a large eland would provide 652 consumer days of meat. Dividing that by a camp size of 30 gives 21 days. However we know that the meat would spoil long before it was consumed at that consumption rate. Even though meat can be dried in the hot sun of the Kalahari it is most often consumed within a few days of capture.
In this photo of gemsbok you get a good sense of the red sand of the Kalahari and the formidable task of a hunter taking these impressive animals with only bow and arrow. Adult gemsbok males weigh from 167 to 209 kg with females smaller at 116 to 188 kg. Tanaka (p 77) reports that a gemsbok was brought into camp on Oct. 12 and nearly all consumed by the end of the next day. Again if we use Lee’s figures for edible yield and a mid-range figure of 188 for a male animal we can estimate 94 kg of meat in camp. On that same page, Tanaka reports 16 providers and 14 dependents so 30 people consumed, on average, 1.57 kg of meat each day for the two days the meat lasted. To convert these figures to caloric gain we can use Lee’s figures from Table 5 of 690 calories in 230 grams of meat or 4710 cal each of those days from meat.
The figures in the above panel give you some idea of why meat is so highly regarded. Remember they are not figures that could be used to rank meat in an optimal diet because they include no post encounter processing costs. But we might think of the costs in another way. The days with meat in camp are days that men do not go hunting nor women go out to gather. For those days no one pays the energetic costs of traveling to foraging patches, collecting and carrying in the hot sun. Tanaka reports a weight of 200 kg for the greater kudu, pictured here. In the last estimate of meat shared around the camp I neglected the portion eaten by the hunters in the bush. The amount of meat eaten by camp consumers should be reduced by the weight of internal organs and ribs the hunters sometimes eat before they transport the game (see Tanaka in the 3 pages on hunting). Hawkes et al. (1991) report slightly higher yield of edible fraction from live weight for Ache prey, 69 to 88% edible. An experiment at the University of Nevada Reno reported “dressing yield” of over 60% for grass fed beef. Keep in mind that hunters eating wild game would eat more parts of the animal than reported as yield by contemporary americans.

The !Kung Bushmen who live in the Dobe area of the Kalahari were studied by Richard Lee from October 1963 to January 1965. Dobe is in the northwest corner of the Kalahari where there are some permanent waterholes. During the dry season, May to October, groups cluster around the permanent water holes, moving out each morning to collect resources within about a 6 km radius of the camp. In the summer rain season between December and April they scatter across the landscape visiting the pools of water that collect from December through April. The population is more widely dispersed at this time of year. In the season of the rains the women gather fruits, berries, melons, and leafy greens. In the dry season they gather roots, bulbs and resins. Remember from Lecture 2, that chimpanzees add resins to their diets in the dry season. Mongongo nuts are gathered year around as the nuts that fall to the ground in one season are preserved by the very hard shell surrounding the kernel. Richard Lee (1968 p 33) makes the point that “food is a constant but distance required to reach food is a variable, it is short in the summer, fall and early winter and reaches its maximum in the spring” that is before the rains begin.

Recall the debate from the Sih-Milton/Hawkes-O’Connell exchange, assigned in lecture 8, over the high processing cost of mongongo nuts. That cost, when only time is considered, needs to be reviewed in the context of other resources that might be gathered in each season and against the background of high variability in game capture. It is almost certain than no dry season resources provide a higher return, otherwise the women would not pay the opportunity cost of processing mongongo. Lee reports the nuts account for 50% of the vegetable diet by weight with the average daily consumption of 300 nuts yielding about 1260 calories and 56 grams of protein. Lee (1968 p 33) estimates that 7.5 ounces of nuts “contains the caloric equivalent of 2.5 pounds of cooked rice and the protein equivalent of 14 ounces of lean beef.”

At the end of the dry season when resources around the water holes have been exhausted the !Kung women have to walk as much as 10 to 15 miles to the nut groves carrying babies as well as the nuts they gather. On these journeys the women have to make a second energetic choice, do they carry toddlers or suffer the time and energy constraint of going at the slow pace dictated by the capabilities of the youngsters. Older children are left in the camp with other caretakers but nursing babies are never left at home. In groups subsisting entirely on wild foods, interbirth intervals (IBIs) are nearly 4 years and babies are at least 3 years old before weaning. Even though !Kung babies are smaller than bottle fed American babies these women are carrying more than 30 lbs of baby and nuts.

Child care constraints always have to be factored into foraging decisions made by !Kung women. How far to travel, what to pursue, how much to carry, how long to gather are all decisions a women faces in the tradeoff between when to have another baby and how to feed the ones she already has. In the photo on the right a women is digging tubers. Blurton Jones and Sibly considered these factors in the backload model to predict the birth spacing that would optimize a !Kung women’s reproductive success. Considering the rigors of walking long distances in the hot sun and the number of nuts needed to feed a family, they predicted interbirth intervals of approximately 4 years, a prediction that matched the actual IBIs fairly closely.

 

Proteins and Calories from Vegetable and Animal Resources

The figures in this table, copied from Richard Lee (1968) What Hunters Do for a Living, or, How to Make Out on Scarce Resources In Man the Hunter eds. R. B. Lee and I De Vore. pp 30-48. Chicago: Aldine, show that mongongo nuts are an important source of both protein and calories in the diet of the Dobe !Kung. Even though vegetable foods comprise from 60-80 per cent of the total diet by weight, meat is still highly valued and widely sought. During the period charted in Table 5 !Kung hunters brought 410 lbs of meat into camp. Lee assumed that meat was shared equally by all consumers in camp, during that month between 23 and 40 individuals, distributing 410 lbs over an average of 31.8 persons per day over the 28 days between July 6 and August 2 gives an average of just over 7 ounces per person per day. In Table 5 Lee reports the weight in grams (230) which is just over 8 ounces. Beware of these average reports. Looking at the large animal prey pictured above this might be one medium sized animal. Just after the animal is brought in people eat much more meat per day than 8 ounces and when the carcass is fully consumed they go without meat for many days. However, since the !Kung hunt small game as well, the total figure may represent many small animals. San Bushmen may hunt singly or go out in groups but once an animal has been spotted and hit the hunters return to camp to rest. Since the poison used on the arrows is slow acting, the wounded animal will wonder over the landscape for some time before it dies. The !Kung men know they will have a long tracking job, the next day, before they find the dying animal.

I have copied a few pages from Tanaka describing the hunt. The copy I have posted for you is missing the figures of hunting equipment but the description gives you a good sense of the difficulty of procuring a gemsbok that has only been hit in the leg. Since the amount of poison on an arrow is small and the poison is slow acting the wounded animal may wonder a long distance before it dies. Notice that the !Kung men of the central Kalahari also scavenge prey from other carnivores and take many small species. According the Tanaka’s description, the internal organs and the ribs are cooked and eaten at the spot of the kill and the bones are smashed to get at the marrow. The uneaten portions are cut up for transport back to camp. The San consume the entire animals except for hooves, horns, bones and stomach contents.

Hadza Hunter-Gatherers of Lake Eyasi Basin

Unlike the !Kung, the Hadza focus exclusively on large game animals. Over 256 days of observation in the years from 1985 through 1989, the Hadza killed or scavenged 72 large animals for an average of one animal every 3.6 days. But considering there were 6 hunters in camp during the wet season and 10 in the dry season, individual hunter success rates were not high. The average of 1 large animal every 4 days seems to guarantee that Hadza individuals eat meat every day but averages obscure the high variance in acquisition. The researchers calculate the chance of failure for any one hunter on any given day as 97%. They compare this to a success rate for the Dobe !Kung of one animal every four man-days and for the Ache, 2-3 successful days out of every 4. In the case of both the !Kung and Ache the success rate reported here includes all the small game added to the diet. Each Hadza hunter maximizes his mean rate of return by focusing on large game but at the cost of high variance, the probability of failure on any day. Remember from earlier discussions that one of the problems with the simple OFT models is they ignore variance in acquisition. The only measure in the OFT models is the rate of return, in calories per hour, for that resource after the time cost of handling and processing is debited. Remember also that the high variance in individual hunter acquisition doesn’t necessarily translate into high variance in individual consumption because a large animal taken by any hunter is shared over all the individuals in the camp as well as others who come from longer distances when the “bush radio” indicates there is meat in the camp. The chance of consuming meat, in a camp with 6 hunters, is about once a week. For a large animal, such as a zebra, the meat might last 3 days increasing the probability of eating meat. In the picture below a Hadza hunter carries a portion of meat that was butchered at the kill or scavenging site.

I have assigned, O’Connell, Hawkes and Blurton Jones (1988) for an argument using the ethnographic example of Hadza scavenging to assess the probability that ancestral Homo might have regularly eaten meat even though the weapons they had were much less effective than the weapons of the modern Hazda. In the final exam you should be able to integrate the ideas and information from O’Connell et al. (1999), Lecture 5, with the information in this lecture’s reading assignment. In the 1988 paper the researchers report a subset of the observations I have summarized above. In the period from 1985 through 1986 the Hadza took 54 medium/large animals of which 14% by carcass weight were acquired by scavenging. Table 2 shows the high variability of success in both hunting and scavenging. The Late dry season, Sept.- Oct. 1985 was the most successful of the four seasons tallied. During that period Hadza hunters brought 28 carcases into camp compared to 13 in the late dry, Aug.-Oct. 1986. Pay attention to “day-to-day availability” reported in column one, mid-page 359. Note the variation from one season to the next. Average income from scavenging was very high in in the early dry season 1986, 245 g per camp resident per day, but all the gain came from one animal, consumed in 3-4 days, leaving consumers many days without meat in spite of high average income from scavenging.

Can you reconcile the information in Table 2 with the graph in Figure 1? The Table shows that the Hadza are most successful at scavenging large game in the late dry season but the Figure shows that large herbivore biomass increases with annual rainfall. Two points should be taken from this observation 1) large game animals may have been relatively more abundant at the Plio/Pleistocene boundary when the genus Homo diverged from the australopithecines so more scavenging opportunities and more opportunities to eat meat, but 2) game animals, predators and humans are more dispersed over the landscape in the wet season and more concentrated around the water holes in the dry season. By now you should expect human diets to vary with the seasons. Recall Stewart’s hypothesis, from Lecture 5, for seasonal dependence on fish easily taken from drying ponds. No matter if fish or mammalian flesh provide the meat component of the diet when no meat comes into camp people still need to eat.

As reviewed in O’Connell et al. (1999) berries and tubers provide the dependable daily nutritional requirements of the Hadza. Hadza children are effective foragers on berries, baobab fruit and some shallow rooted tubers but they do not have the strength to gather the deeply rooted tubers that provide the highest returns for adult women. This photo of a senior Hadza women illustrates the strength needed to gain access to these resources. Hadza women, with unweaned babies and children over the age of 6, leave camp in early morning to gather. They are usually accompanied by armed teenage boys and perhaps an adult male as protection against pastoralists who range through the same area. Men and boys who accompany women on these trips sometimes take honey from wild bees. Honey is a valued resource with high caloric return. Hadza men can achieve a mean daily return rate of .78 kg per man-day by spending an average of 41 mins a day in honey collecting.

The women walk from 4 to 60 mins. to the tuber patch where they spend the day collecting. Within the patch they stop about mid-day to build a fire and cook some of the tubers they have collected. In the afternoon they continue collecting, stopping again to cook and eat before they gather up the tubers they are taking back to camp for the evening meal. About 39% of the tubers gathered are eaten in the field. During the wet season women add berries to their foraging schedule, spending about half their foraging time in berry patches. Women gain very high returns, measured in grams per hour in the berry patches but they spend about 34% of the foraging time traveling to the patches and 23% walking within the patch between bushes. I estimated the daily gain for berries based on 6 hours of foraging in the wet season at 26% of 6 hours times a gain of 4017 cal/hr to be 6268 cal. for a days work picking berries. In the dry season women spend 4 hours foraging for tubers. On those trips 38% of the time is spend digging for an average of 1.52 hours digging tubers that provide a return of 2000 kcal/hr or 3040 kcal/day. The foraging hours reported here are for child-bearing women. In general senior, post-cycling women forage 22-52 % longer than women of child-bearing age. When you consider the “take home calories” remember that the women and children over 6 that accompany them have been eating in the field so their need for daily subsistence from the take home “bag” is less than that for individuals who were not in the patch. You can see from these estimates that changes in the distance to tuber or berry patches could have considerable effects on the returns that women are able to gain from their daily foraging round. In general the travel distance to berry patches is higher than the distance to tubers because tubers are more densely distributed in the environment. These estimates are reported by Hawkes, O’Connell and Blurton Jones (1989) for observations October to November 1985 and March to April 1986 ( you can find the reference in O’Connell et al. 1999).

Energy Conservation?

How many calories do foragers need? How many hours should they work for a net energetic gain? Was Marshall Sahlins right, needing little they meet their needs in a few hours of work and have many hours of leisure? If we remove our western lens and think about foraging through the lens of OFT we get a new perspective. The theory predicts that foragers should target resources that provide a net energetic return after the costs of pursuit and handling are considered. We have already learned that search time is not debited from the gross return, however all researchers know that walking time is energetically expensive. As a consequence of occupying a niche different from that of our nearest relatives we range more widely in the food quest but because we are bipeds we may be more efficient over large ranges. Leonard and Robertson (1997) modeled the energetic cost of walking for a generalized quaduped compared to a biped and found that male and female bipeds are more efficient over all speeds from 2.4 to 6.0 km/hr. They estimate that the evolution of obligate bipedality produced an energetic savings of 40 to 47% for male and female hominids compared to large bodied primates. Since modern hunter-gathers have an average range of 13.10 km/d compared to 1.77 km/d for large bodied apes an increase in locomotor efficiency is particularly important to occupation of the human foraging niche.

However the cost of search is related not only to distance traveled but to other variables such as temperature, humidity, topography, and load carried. Nursing infants are a particularly important component of the female forager’s load which also includes food transported back to camp where other dependent children are waiting to be fed. Unlike chimpanzees, human males carry tools, tool stone and large portions of meat. Could it be that foragers allocate their efforts to maximize calorie gain while minimizing energetic expenditure. Are foraging strategies a tradeoff between energetic returns from the resources in their environment and the energetic cost of searching for, handling and processing those resources? Are there seasons of the year and periods of the day when it pays a forager to conserve energy by not foraging? In the extreme temperatures of the Kalahari resting in the shade during the hottest part of the day may be a very efficient strategy.

Lee reports that Dobe !Kung women work an average of 2-3 days per week but Hawkes and O’Connell (1985) point out that Lee ignores the cost of processing mongongo nuts, 5 hours of cracking and pounding to produce 1 kg of nutmeat. Lee reports per capita consumption of 210 grams of mongongo nuts so a women could produce enough nuts for approximately 4 days of consumption if she only cracked for herself. However she cracks for her family dependents as well. If we compare the work effort of the foragers we have studied so far we find differences in their foraging patterns and the number of hours worked. Tanaka (Table 16) reports large variation in male work effort, hours out of camp searching for food, among the !Kung he studied, 9.35 to 4.15 hours but less variation in women’s effort; 4.25 to 1.50. The Ache move camp almost daily, walking during the day, processing food, cooking, visiting and resting during the evening and night. The !Kung gather resources during the morning hours, unless they are tracking game, and process food, rest in the shade, and visit during the mid-day and evening hours. The Hadza also begin gathering early, stopping at mid-day and sometimes gathering late in the afternoon but often spending no more than 6 hours in the field. Hadza men are more similar to Ache men in the habit of going out to hunt every day. Variation in work effort and foraging returns across habitats would make a good term paper topic.

Assigned Reading

Tanaka, J. 1980. Chapter 2 Subsistence Ecology In The San Hunter-Gathers of the Kalahari: A Study in Ecological Anthropology. pp 53-91. Tokyo: University of Tokyo Press. Go to Marriott Library Course Reserves to download the electronic file. Tanaka, J. 1980. Hunting 3 pages from Chapter 1 subsection hunting. In Subsistence Ecology In The San Hunter-Gathers of the Kalahari: A Study in Ecological Anthropology. pp 30-35. Tokyo: University of Tokyo Press. Available from the Readings section webCt. O’Connell, J. F., Hawkes, K. and Blurton Jones N. G. 1988. Hadza scavenging: implications for Plio/Pleistocene hominid subsistence. Current Anthropology 29(2): 356-363. O’Connell, J. F., Hawkes, K. and Blurton Jones N. G., 1999. Grandmothering and the evolution of Homo erectus. J. Hum. Evol. 36:461-485. Posted under readings Lecture 5 Review p 465 from the Grandmother Hypothesis to p 468 Applying the argument to Homo erectus and p 470 beginning with Climate change and “children’s” resources to p 475 An evolutionary scenario grounded in the Plio-Pleistocene.

 

File: Lecture 11 Meriam Aquatic Foragers

Anthro 4962 The Evolution of the Human Diet Lecture 11 Meriam Aquatic Foragers University of Utah Fall 2005 Helen Alvarez

 

Mer Island Home of the Meriam photo © Great Barrier Reef Marine Park Authority 1996.

Mer island is located in the Torres Strait on the northern end of the Great Barrier Reef 142 km southeast of Papua New Guinea. With this lecture we revisit the sub-humid tropics for a study of foraging strategies of a population surrounded by the sea. The island is fringed by reef, the light blue-green color center left and zigzagging toward the center of the photo to the right of the island. The foragers of Mer practice a subsistence strategy that includes marine foraging, growing of yams, manioc, and bananas as well as keeping of pigs and chickens and subsidies from the Australian government. The gardens and domesticated animals provide a small fraction of the diet as most carbohydrates are now purchased from a small store on the island and most of the meat in the diet comes from marine resources. Three distinct strategies are found in the marine niche; hand-line fishing and netting on the nearshore at high tide, shell fish collecting and spear fishing on the reef flat at low tide, and deep and shallow water fishing, lobster collecting and turtle hunting offshore, and hunting and collecting turtles in the nesting season.

This lecture has two purposes; 1) to introduce sub-tropical marine collecting strategies, and 2) to consider an important body of ideas advanced to explain widely noted falsifications of the predictions of optimal foraging strategies. Those falsifications were first noted in the finding that Ache men rarely collect palm starch in spite of the high caloric returns provided by that resource (Hill et al. 1987). The evidence that men pursue resources characterized by high variance and wide sharing, once captured, was included in the discussion by Elston and Zenah (2002), developed by O’Connell et al. (1999) in the review of the grandmother hypothesis as an explanation for the evolution of the genus Homo, and evidenced in the hunting strategies of the Kalahari San Bushmen and the Hadza big game hunters. You were prepared for these ideas by the four hypotheses advanced by Mitani and Watts (2001) to explain hunting by chimpanzee males. By now you should suspect that male foraging is about more than satisfying personal nutritional needs.

Everywhere humans forage, men and women target different resources. The resources taken by women are characterized by low day to day variance in acquisition rate, by small package size and by limited distribution within the family. By contrast men target resources characterized by high variance in acquisition, large package size, and distribution to all in the vicinity of the catch. The fact that men do not discriminate between their wives and offspring and the families of other men provides strong evidence that male strategies are not about parental investment or offspring provisioning. The argument that male hunting can best be explained as sexual selection for male-male competition is set out in the assigned papers for this lecture.

Reef Flat Foraging

Both men and women forage in the shallow water of the reef flat in the 2 to 4 hour low tide periods from March to end of September. Men walk to the reef edge and fish with spears catching a variety of small and large fish. The largest fish they take is giant trevally, pictured on the left, which can grow to 1.7 m, but they also take a variety of very small fish such as spine foot, sweet lip, and mullet about 38 cm, 45 cm and 78 cm, respectively. The catch for the day is displayed openly as the men leave the reef at the end of the fishing bout. If you can imagine these small fish, pictured in order below, flashing through the coral and the distorted optical perspective from the surface of the water to the shallow depths where fish school and hide, flash and dart you can get a sense of the skill needed to take these fish with wooden spears. In the methods section of the assigned reading, Bliege Bird, Smith and Bird (2001), the authors note that the spearfishing forager ignores other prey types as he travels across the reef. During the low tide men spend 63% of their time spearing and 31% of their time shell fish collecting. This time allocation bears no relationship to the potential returns from the prey, 292 plus or minus 135 kcal/h for spearing and 1492 plus or minus 173 kcal for shell fish collecting.

 

Spinefoot to 38 cm	Sea Mullet to 78 cm	Sweet lips to 45 cm

Notice that women allocate very little time to spearing, 9% of their time on the reef compared to 76% of the time shell fish collecting. Because of the high cost of transporting the shells of the very large clams they collect, women process as they collect. Women collect only three species, Hippopus hippopus, Tridacna spp. and Lambis lambis, the spider conch which is very much smaller than the conchs found in tropical Atlantic waters. The largest tridacna clams can reach 140 cm, over 4 feet on the longest dimension.

Women forage on the dry reef carrying a bucket, knife, hammer and a small spear. As they collect they cut out the meat from the shell and drop it in the bucket using the hammer to crack conch shells and the spear for balance, although they do stop to spear small fish or octopus trapped in the tide pools. When the man and woman return home from the reef he shares his fish with the neighbors and she cooks her share of the fish and the resources she has collected from the reef for the family dinner. Over a sample of 44 women, female forages brought in 1962 grams/per bout, on average, while spearfishermen (n = 36) brought in 356 grams. Study Table 1 (Bliege Bird, Smith and Bird 2001) where they report returns from the ebb tide reef flats according to activity. Reef collecting of large clams gives the highest return in calories, proteins and fat. Even though men and women spend the same time on the reef, their energetic gains are quite different. Figure 1, in this same reading, illustrates the gains achieved by forages targeting shellfish compared to gains from spearfishing. Pay attention to the kcal scale on each panel. Although the curves seem comparable the gains for spearfishing vary from 0 to 1700 kcal while those from shellfish collecting vary from 0 to 5500 kcal for just over 140 mins of foraging time. Notice also the high variance in returns on panel A. The extreme outlier of just under 1800 kcal in 25 minutes of time is not representative of the average gain reflected in the regression through the mean. The returns plotted in panel B cluster more tightly around the mean. This evidence lends support to the hypothesis that spearfishing is a costly signal of male talent. Keep this example in mind when we get to the lecture on children’s foraging and the argument that male foraging strategies explain the evolution of long juvenile periods in ancestral H. sapiens.

If foragers process as they go, they leave no record of past meals. Think about how we derive information on ancient diets. The archaeological record consists of remains left at central processing spots or middens, kill sites, camp sites, and dumping sites. The record is over-weighted with resources with durable discard. When foragers process as they forage, inedible portions are scattered across the landscape. Absence from the archaeological record never implies absence from the diet of ancestral humans. You can begin to appreciate the importance of ethnographic evidence drawn from foragers who still collect wild foods. The power of the simple OFT models helped ethnographers understand the evidence they collected. Even where modern foragers practice a mixed subsistence, including cash purchases of food, a great deal has been learned about the energetic costs of wild foods from the use of OFT models to understand foraging strategies for wild resources.

On their small, 2.8 by 1.7 km, bounded world the Meriam practice a seasonal round. As noted above, March to early October the Meriam forage on reef flats at ebb-tide. In October through April they switch to collecting nesting turtles and eggs. Everyone participates in turtle collection during this season as the turtles come onto the sandy beaches to dig nests and lay their eggs. The foragers go out at night or early morning and wait on the beach for turtle to come ashore. When a turtle is spotted, it is flipped over, the flippers tied back and the turtle hauled home by boat. Nesting season collections are undertaken mostly for household provisioning but raw meat from the butchered turtle is always shared out with neighbors. In some cases nesting turtles will be taken for pre-arranged feasts.

Throughout the year men take turtles in the open water. There is a rich description of this type of hunting in the required reading for this lecture. You might be surprised that a reputation for a good hunter falls on the leader of the boat who directs the driver, makes decisions about pursuit and directs the chase rather than on the jumper who actually goes into the water and wrestles the turtle. Turtle hunting is open to all men between the ages of 16-47 but participation varies. “Thus 44.5% of the 90 Meriam males ages 16-47 hunted at least once in the study period, but the 3 most active participants (3.3% of males) were over 5 times more likely to engage in a turtle hunt than the average Meriam male in this age range (Bliege Bird et al. 2001 p 14).” These same 3 men were named as the best turtle hunters in a series of interviews with other island residents. Hunting in open water is more costly than hunting on the beaches. Leaders of open water hunts are older and more experienced while young men who are jumpers in open water hunts are active participants in nesting season captures perhaps as a way of working toward hunt leadership.

 

Hunting Returns from turtle collecting and hunting are reported in Table 2. The after-sharing returns during the nesting season are comparable to the returns from reef shellfish collecting. Notice the large variation between return rates from collecting in the nesting season and hunting in the hunting season, 21,875 kcal/h vs. 4922 kcal/hr. The very large difference is caused by the inclusion of search time and energy expended in travel. The energetic costs of travel in the hunt are calculated from the conversion of the cost of fuel for the boats into calories of meat that could have been purchased at the store. When per capita returns are reported after-sharing, open water hunting results in a net loss to the hunters because of the time and fuel costs of the activity. Notice that hunting in the hunting season is more costly than hunting in the nesting season. “In this season, hunting is much easier and hunters take on fewer costs: turtles are found on nearby reefs waiting to crawl onto the beaches to lay eggs at night, the tradewinds have largely ceased, and in between monsoon storms, the water is clear, calm and visibility is excellent, allowing hunters to dog turtles more closely, to lose fewer and to more finely discern size and sex (Bliege Bird et al. 2001 p 14).”

 

Men often hunt on open water for other resources such as tuna, mackerel and large marine sea mammals. This picture of men displaying their spanish mackerel catch comes from Rebecca Bliege Bird’s home page at Stanford. This looks like an impressive catch but you can see, in Fig. 7 below, how large fish rank in return compared to other resources. The returns graphed in Fig. 7 represent after-sharing returns. The bar represents mean returns with the long line to the right of the bar representing one standard deviation. The longer the line the higher the variance in return. In the case of large pelagic fish, the standard deviation indicates that men may achieve a positive or negative 5500 Net E/hr return per hour. Remember a negative return results from an unsuccessful hunt as the cost of fuel, calculated in purchasing power of store-bought meat, is debited from the return to quantify the costs of search. Only returns from netting sardine, 11,008 kcal/h before sharing, are reported in the assigned reading (p 14) but the figure from Bliege Bird and Bird (1997) illustrates returns from the most important resources. Both men and women use casting nets to intercept schools of sardine along the perimeter of the island and hand-line fishing using hook, line and bait to catch reef carnivores, herbivores and surf fish from the tidal margin. On this chart the returns from sardine netting are lower than those reported above but I assume acquisition costs are not debited from the figure of 11,008 kcal/hr.

In all discussion of optimal foraging across environments and in the same environment across methods we need to exercise care to make certain we are comparing net returns calculated in the same way. If search costs are debited, as they are in Fig. 7, the returns will be much lower than returns reported as post encounter return rates, the method of calculating returns using the classical optimal foraging models.

Return rates

 

Costly Signaling

 

One of the features that distinguishes humans from other primates may well be male-male competition in the arena of provisioning others. Proponents of costly signaling hypotheses to explain the evolution of men’s work propose that human males signal their quality as mates and allies and their danger as competitors by providing high variance, high return resources that involve skill in acquisition and some personal risk. The caloric returns from the two types of signaling explored in this population are vastly different. The return from spearfishing is insignificant over the total collected resources but turtle hunting in the open water provides a very high caloric return shared over many individuals in very public feasts. In the first case, the spearfishermen signal the personal physical qualities that make them successful hunters of small mobile prey; hand-eye coordination, patience, and endurance. Their successes on the reef provide few caloric benefits to observers, but provide useful information to potential competitors and allies. Those who attend to the signal receive honest information about the quality of the man because the task is too difficult to be accomplished by those of lesser talent. In the second case, turtle hunting in the open water, observers gain not only information but valuable calories and opportunities for communal feasting. Pay attention to the predictions Bliege Bird et al. (2001) derive from the costly signaling hypotheses and how they are tested. I have also assigned Hawkes and Bliege Bird because they review Zahavi’s handicap principle, link it to the analysis of chimpanzee hunting by Mitani and Watts (Lecture 2) and to Veblen’s classic analysis of conspicuous consumption. Finally, Hawkes and Bliege Bird (2002) argue that reciprocal altruism fails as an explanation of most human sharing patterns because pair-wise exchanges for large meat packages have not been demonstrated. These are important arguments and you should be clear about them before the second exam. If you are not familiar with modern arguments linking conspicuous consumption and sexual selection in humans, have a look at Geoffrey Miller’s prize winning essay Waste is Good. This essay should be especially interesting to those of you majoring in psychology or marketing.

 

Assigned Reading

Bliege Bird, R., Smith, E. A., and Bird, D. W. 2001. The hunting handicap: costly signaling in human foraging strategies. Behav Ecol Sociobiol 50:9-19. Hawkes, K. and Bliege Bird, R. 2002 Showing off, handicap signaling, and the evolution of men’s work. Evol Anthro 11: 58-67.

Also Recommended

Bird, R. 1999. Cooperation and conflict: The behavioral ecology of the sexual division of labor. Evol Anthro 8(2): 65-75. Smith, E. A., Bliege Bird, R. and Bird, D. W. 2003. The benefits of costly signaling: Meriam turtle hunters. Behav Ecol 14(1):116-126.

File: Lecture 12 Juvenile Foraging

ANTH 4962 The Evolution of the Human Diet Lecture 12 Juvenile Foraging University of Utah Fall 2005 Helen Alvarez

 

Overlapping Generations photo by Marjorie Shostak @ http://www.anthrophoto.com/

 

Unlike other primate females, human mothers have overlapping dependents, a new child is born before the previous child becomes independent. A number of hypotheses have been developed to explain who feeds a women and her child when her foraging returns are constrained by child care. As illustrated in this photo and those in previous lectures, women take their nursing infants with them but here you see a !Kung women from the Dobe area carrying two of her dependents plus the food she has just gathered.

In the early versions of the man the hunter hypotheses, investigators imagined that women with dependent children remained in a central place caring for children while men provisioned the family with large game. A more recent hypothesis links the evolution of long juvenile dependencies to food sharing between older female kin and young dependents (see O’Connell et al 1999). In environments of low extrinsic mortality young animals can risk devoting more time to personal growth, growing longer and larger before maturing. In the mosaic habitats of the Plio-Pleistocene boundary mothers and children were able to move into new habitats because long-lived grandmothers helped provision weaned infants. In turn the benefits of provisioning increased longevity genes in these populations, further reducing average adult mortalities and promoting later age of maturity. It is late age of maturity that provides long juvenile periods. The consequence of selection for longevity may well be long learning periods but as always we should be careful to separate cause and consequence in evolutionary puzzles. The authors assigned in this lecture review these arguments over and over. Be sure you understand the different hypotheses because they are all related to the occupation of new habitats, the provisioning of juveniles, the evolution of the human diet, and the evolution of longevity in our lineage.

By now you have ample evidence, from ethnographic studies, demonstrating that the man the hunter scenario under-estimates the work of women who forage for themselves and their children even when diets are supplemented by widely shared resources provided by hunting men. The second proposition of the hunting scenario, that human children are dependent until they become adults is true for the !Kung and some South American subsistence horticulturists but quantitative studies of other foragers demonstrate that young children and juveniles provide some of their own resources. Differences across habitats and subsistence strategies are indicators that juvenile strategies might be structured by ecology and physical development and not by learned capabilities. The differences, across habitats, bear upon the modern version of the man the hunter proposal, the embodied capital hypothesis, which proposes that the long juvenile period in humans evolved in response to the value of a long period of learning for mastering adult foraging skills. Once individuals, males in this particular hypothesis, master hunting skills they are able to provision mates and juvenile dependents. The evidence that children can be effective foragers at a young age challenges this proposition. The various explanations for the evolution of our unique life histories are well rehearsed in the assigned readings for this lecture.

 

Hadza Children’s Foraging

 

Hadza children as young as 5 years old forage near camp for resources to supplement their diets. In certain seasons, children are left at home when mothers, carrying only nursing infants, go to dig tubers. Detailed observations reported in Blurton Jones, Hawkes and O’Connell (1989) showed that infants older than 2 1/2 years rarely accompany mothers to tuber foraging patches. Children left in camp forage on tubers growing near the surface, the honey of stingless bees, and baobab fruits. The latter fruits have high processing costs as the pods need to be cracked and the pith pounded into a dry powder. This powder is moistened with water or eaten dry with water. The seeds can be cracked and the kernels eaten raw. In spite of high processing costs, children aged 5-10 years can gain 629 cal/h from baobab, while those 10-15 get about 1014 cal/h. However children do not allocate many minutes to foraging so it would be a mistake to imagine that Hadza children entirely support themselves. Honey is mostly taken by boys between the ages of 12-15 if one can get an axe to chop out the nest. Honey yields about 339 cal/h but boys who accompany women on foraging trips often acquire 650-1350 calories on these trips. Blurton Jones et al. (1989 p 380) estimate that “If children spent just 2 hours per day foraging for baobab, or (for older boys with access to an axe) honey, they would acquire around 800-1000 calories, almost half the calories they need.”

In a subsequent field season, the same research team completed time allocation studies to determine how mothers change their own foraging patterns in response to opportunities presented by resources they can gather with their children (Hawkes, O’Connell and Blurton Jones 1995). Contrary to the expectation from optimal foraging theory that mothers choose foraging patches that optimize their own return rates, they found that women, in certain seasons, choose to forage in berry patches with their children instead of traveling to patches to collect tubers. Mothers maximize team foraging rates by focusing on resources that children can gather with adults. “If higher rates of food acquisition are advantageous to women because with higher rates they can feed children more, or feed more children, then when children are active foragers themselves a woman’s children will consume food at a higher rate if she chooses the strategy that maximizes the team rate she and her children earn collectively, even if the rate she earns herself is less than the maximum possible (Hawkes et al.1995 p 695 emphasis original).” The team rates achieved for mothers and children depend upon the number of children and the time allocated to foraging. The more children a mother has the higher the team rate. If the team forages for longer than 556 mins, berries are the best choice. If a women wants to maximize her own return rate collecting tubers she needs to forage for an additional 2.5 hours. Over the mean length of observed trips, women and children do best focusing on berries. The foraging strategies of mothers and kids might have been influenced by the scarcity of baobab fruit during the late dry season of 1988 when the time allocation studies were completed.

In spite of children’s abilities to forage for themselves around the Hadza camp, they still depend upon mothers. The time allocation data show a significant difference between foraging time for children with co-resident mothers and those without. Notice on Table 2, copied from Hawkes et al. (1995), that an 8 year old girl with no mother or father in camp spent more time foraging than any other child. In a sample of 20 juveniles, aged 3.5 to 17 years old, 9 had no coresident mother in camp.

 

Table 2

 

The caloric returns children achieve from their own foraging efforts depend upon the resources they target. The two types of tubers, Makalita and //ekwa, gathered by children yield 73 to 85 Cal/100 g and children over the ages of 7 and 8 can gather 598 and 314 g/hr, respectively, of these two resources gaining about 200 calories for 29 mins of digging Makalita. When children accompany mothers to the berry patch they can gain from 964 to 2,223 Cal/hr depending upon which berry, Salvadora persica or Cordia sp. they are picking. Children gain about 50-70% of the adult rate picking S. persica berries, Table 6 (Hawkes et al. 1995). Since Cordia berries were not fully ripe until the end of the observation period, rates by age were not available for the berries yielding the higher caloric return. There are two important take home points from the Hadza field work on children’s foraging. 1) Hadza children’s foraging choices are consistent with optimal foraging theory predictions, near camp children generally choose to forage on the tubers with the highest caloric return just as they do in the berry patch. Before Cordia berries ripened children picked berries giving a lower return but once the Cordia ripened they ignored S. persica berries even though they were still available. 2) Mothers adjust their foraging strategies to include resources their children can gather. In berry picking season mothers forego tuber picking, where they earn high returns, to travel long distances to berry patches in order to maximize the team rates they can gain with their children.

The stashing rate shown in Table 6 is defined as the “total weight of fruit accumulated for transport per hour in the resource patch. They represent the minimum rate that would have been taken had foragers eaten none of the berries they picked (Hawkes et al. 1995 p 693).” Actual picking rates are a combination of eating and stashing rates. The last column is an attempt to measure the amount of time spent picking, the stashing rate and consumption rates are added and divided by the tin measured rate. The results show that women are the most steady pickers and men, boys and children the least effective. Men may have spent more time in the berry patch in this particular field season as “hunting was poor(only eight large animals taken over 43 days of direct observation)(Hawkes et al. 1995 p 689.).” Men pick at very high rates but they also consume at a high rate so their stashing rate is lower than that of women and girls.

 

Table 6

 

The foraging efforts of Hadza youngsters stand in sharp contrast to those of Dobe !Kung children who do almost no foraging. Instead !Kung children contribute to their own subsistence by cracking mongongo nuts carried home from distant patches by their mothers. Hawkes, O’Connell and Blurton Jones (1995) attribute the difference to the scarcity, in the Kalahari, of near camp resources that children can gather and the high, dry season temperatures and lack of water that preclude taking children on long walks to the nut groves. Instead !Kung mothers and kids do better if toddlers and older youngsters stay in camp and crack nuts that mothers have carried home. Recall the topographic differences between Hadza country and the Kalahari with no prominent landmarks to guide children from foraging sites to home camps. Blurton Jones et al. (1989) suggested that the harsh dry season temperatures keep !Kung children from foraging with mothers and the danger of getting lost in the flat environment restrain children from foraging alone near camp.

 

Juvenile Foraging Size or Experience?

 

At both of their field sites, Mer Island and the Western Desert of Australia, Douglas and Rebecca Bliege Bird conducted quantitative studies of juvenile foragers to determine if optimal foraging theory could explain choices made by children. At both sites they found that children take resources that give them high rates of caloric return as predicted by theory but their foraging returns are constrained because the rate at which they encounter resources is determined by size and walking speed rather than age or experience. Their evidence indicates that physical abilities rather than learned capabilities determine juvenile foraging strategies.

On Mer Island, mixed-age groups of children, independent of adults, forage on the reef flats after school or on weekends. On these trips they collect some of the same resources collected by adults but they encounter them at a lower rate so their overall returns, in the patch, are not reduced by stopping to take smaller shellfish such as black lipped conch and small top-shell. Post encounter processing costs of the large shellfish are higher for children because they do not have the upper body strength to cut out and remove edible flesh from the largest shells. These differences, between adults and children in physical strength rather than skill, explain why children take resources ignored by adults. Further Bliege Bird and Bird, in Children on the reef (2002), suggest that children are very efficient at extracting meat from small shells because they can more easily reach in the small valve openings with their small fingers to pull out the meat.

Mer children also fish, from the beach, a foraging strategy requiring some skill. Across the Torres Strait islands, Meriam kids are known for their skill in hand-line beach fishing. Children begin as young as five years old and quickly master the technique. After an initial steep rise in efficiency there is little effect of age on efficiency. However, at every age there are large differences in individual skill and efficiency such that most of the variation in this task is explained by individual differences with one 62 year old man out-ranking all others in return rate. Figure 1 taken from Bliege Bird and Bird (2002b) illustrates the scatter and the large disparity when a very good older fisherman (the open circle top right) is included in the analysis. Residuals of overall return rates are plotted on the y axis to remove the effects of varying sample numbers across foragers. In spite of the title, return rates are calculated from both small-hook and large-hook fishing methods. Men and children of both sexes choose large-hook fishing and women choose small-hook. One difference between children and adults is that children usually have fewer choices of line size. From the figure, we might conclude that only the most skilled children choose to hand-line fish but without knowing the proportion of children who choose to fish it would be impossible to say the sample is self selected by only the best children.

 

Figure 1 Age effects on large-hook beach fishing efficiency

 

Martu Children’s Hunting

 

Martu (sometimes Mardu) children of Northwest Australia are often left at home when mothers hunt goanna lizards in the sand hill flats but these children are active forages in patches distinct from those targeted by their mothers. Bird and Bliege Bird (2005) argue that size determines the rate at which foragers encounter prey in the two patches. Adults are able to move at faster speeds across the sand flats encountering goanna at about .90 items/hour searching whereas children would encounter prey in this same patch at .68 items/hour because of their slower walking speeds. Larger children gain nearly the same caloric gain foraging on the sand plain as in rocky outcrops but smaller children do much better on the rocky outcrops, 448 kcal/hour compared to 385 kcal/hour on the sand plain. The authors suggest children forage in patches that minimize time and effort expended. “Although the gross foraging return rates are the same, foraging in sandhills as compared to rocky outcrops requires 2.3 times greater time investment and 2.2 times more walking as does rocky outcrop foraging for only 1.3 times as many calories per hour (Bird and Bliege Bird 2005 p 142).” The young girls in this picture, copyright Rebecca Bird from her home page at Stanford, are holding a small lizard they have dug from a rocky outcrop burrow.

The analyses, of Martu children’s foraging success, was designed to tease apart the factors which influence foraging choices of juveniles. The regression plots show that height has a larger effect on return rates than age (compare Figs 6.2 and 6.3). The proponents of the embodied capital hypothesis for the evolution of human life histories propose that the long human juvenile period evolved to provide a long period of subsidized dependence during which children forego supporting themselves in favor of learning the foraging skills that are later cashed out as productive adults. In this scenario juvenile dependency is subsidized by paternal support of wives and children. Bird and Bliege Bird show youngsters can be effective foragers using the same tools and techniques as their mothers but foraging in different patches. Since within patch encounter rates constrain juvenile choices, size not experience makes a difference to children’s effectiveness. These results are important because they undercut the proposition that long juvenile periods, in humans, evolved in response to the value of learning adult foraging skills. Instead they support the proposition that long juvenile periods evolved in response to the benefits of growing larger before maturing.

The question remains, why don’t children forage longer hours? The work effort of the 8 year old Hadza girl with no co-resident parents suggests children are capable of longer work hours. It is possible that children allocate energy to fogging effort while still conserving energy for growth. Humans are determinate growers which means we reach skeletal maturation and hence most of our adult height about the time of sexual maturity. Bliege Bird and Bird (2002a) suggest that foraging efficiency increases after puberty because both males and females have greater opportunity to enhance fitness gains from foraging efficiency. During the thousands of years spent subsisting on wild resources, males excelled in male-male competition through costly signaling of hunting ability and providing resources for public consumption. In this same niche, females promoted their own reproductive success by gathering low variance resources to support their dependent children and grandchildren. Foraging efficiency increases at puberty in response to natural selection on females and sexual selection on males.

 

Mikea Children

 

In Growing Up Mikea, Tucker and Young (2005) review the various arguments for the long juvenile period in human life histories and report detailed time allocation studies of Mikea tuber foraging. They emphasize what you have learned earlier in this lecture, the cost of children varies across habitats. In the dry woodland forest of Madagascar, where the Mikea live there are few dangers for children foraging in the woodland, no poisonous snakes or large predators and plenty of shade to moderate temperatures. Even though the Mikea practice a mixed subsistence economy that includes subsistence farming and herding, trade, and cash labor they still collect wild resources. Figure 7.2 shows that children spend much more time in leisure activities than other age classes but by adolescence young people spend as much time working as do married adults. Notice also that adult males and females allocate nearly the same proportion of daylight hours to food production. Among children, boys allocate nearly twice the time to food production as do girls but these figures change when allocation to tuber foraging is measured. Adolescent girls devote more time than any other age group to this task (Figure 7.3). Still across all foragers net acquisition rates increase with age and men achieve significantly higher rates than women but Table 7.2 shows that daily returns for men are low because they spend 1/3 the time that women spend foraging for tubers.

In the early dry season, Mikea children actively forage with adults and mothers, who suffer no decline in foraging efficiency when accompanied by children. Notice in Figure 7.8 that adult females forage in groups with children and adolescents more frequently than do adult males. In certain seasons adults and children forage in the same patch but when patches near home are depleted in the largest tubers adults range further to exploit new patches while children stay in the older patches harvesting smaller tubers that lie close to the surface. The ovy tuber, collected by the Mikea, grows in sandy soil so digging and collecting is relatively easy. The ease of digging ovy combined with its higher energy density produces a net return rate much higher than the tubers collected by the Hadza. Have a look at Table 7.2 and notice that children gain 505 to 537 kcal/hr gathering ovy but they gather only 35-41 minutes/day. The same question arises over and over for children, why don’t they devote more hours to foraging? Tucker and Young suggest that children are neither rate-maximizers or time-minimizers. Since they are provisioned by adults they are also not energy limited. Instead they forage for the physical and mental challenge and for the social opportunities afforded by mixed age play groups. However the ability of Hadza, Merriam, Martu, and Mikea children to collect resources for themselves suggests that they might be able to support themselves if adult provisioning failed. As illustrated from the detailed Hadza data, children often have no co-resident parents. In such cases orphans must rely on near-relatives, neighbors and their own ingenuity. We expect natural selection to favor strategies, in young mammals, that help them survive maternal accidents. To expect youngsters to forage on the same resources as adults is unreasonable given their size and physical strength. The studies reviewed in this lecture demonstrate that children can be capable foragers depending upon the opportunities in their environment but over all environments, given adult support, they choose to allocate more time to leisure.

 

Required Reading

Bird, D. W. and Bliege Bird, R. 2005. Martu children’s hunting strategies in the western desert, Australia. In Hunter-Gatherer Childhoods: Evolutionary, Developmental and Cultural Perspectives. eds. Hewlett, B. S. and Lamb, M. E. pp. 129-146. New Brunswick NJ: Transaction Publishers. Course Reserves Marriott Library Tucker, B. and Young, A. G. 2005. Growing up Mikea: Children’s time allocation and tuber foraging in southwestern Madagascar. In Hunter-Gatherer Childhoods: Evolutionary, Developmental and Cultural Perspectives. eds. Hewlett, B. S. and Lamb, M. E. pp. 147-171. New Brunswick NJ: Transaction Publishers.

Also Recommended

Blurton Jones, N. G., Hawkes, K, and O’Connell, J. F. 1989. Modelling and measuring costs of children in two foraging societies. In Comparative Socioecology: The Behavioural Ecology of Humans and Other Mammals. eds. Standen, V. and Foley, R. A. pp. 367-390. Oxford: Blackwell Scientific Publications. Hawkes, K. O’Connell, J. F. and Blurton Jones, N. G. 1995. Hadza children’ foraging: juvenile dependency, social arrangements, and mobility among hunter-gatherers. Curr Anthro 36(2): 688-700. Bird D. W. and Bliege Bird, R. 2002a. Children on the reef: Slow learning or strategic foraging? Human Nature 13(2): 269-297. Bliege Bird, R. and Bird, D. W. 2002b. Constraints of knowing or constraints of growing: Fishing and collecting by the children of Mer. Human Nature 13(2): 239-267.

 

 

Missing 8-1, 8-2, 9-1, 10-2

7-1

Journal of Archaeological Research,
Vol. 9, No. 4, December 2001 ( C 2001)

The Archaeology of Aquatic Adaptations:

Paradigms for a New Millennium

Jon M. Erlandson1

Although aquatic resources are often
seen as central to the development of post-

Pleistocene cultural complexity, most
models of human evolution have all but

ignored the role of aquatic or maritime
adaptations during the earlier stages of

human history. When did aquatic
resources, maritime adaptations, and seafaring

first play a significant role in
human evolution? I explore this fundamental question

by (1) reviewing various theories on
the subject; (2) discussing a variety of prob-

lems that prevent archaeologists from
providing a clear answer; and (3) examining

the archaeological record for evidence
of early aquatic resource use or seafaring. I

conclude that aquatic resources,
wherever they were both abundant and relatively

accessible, have probably always been
used opportunistically by our ancestors.

Evidence suggests, however, that
aquatic and maritime adaptations (including

seafaring) played a significantly
greater role in the demographic and geographic

expansion of anatomically modern humans
after about 150,000 years ago. Another

significant expansion occurred
somewhat later in time, with the development of

more sophisticated seafaring, fishing,
and marine hunting technologies.

KEY WORDS: aquatic resources; human
evolution; maritime societies; coastlines; boats.

INTRODUCTION

The average molluscan flesh is
certainly not very appealing in appearance and the earliest

humans apparently existed for
uncounted millennia before that anonymous hero ate the first

oyster. In any event, shell
middens of real antiquity are rare or absent in world archaeology

(Meighan, 1969, p. 417).

Central to the success of our
species—measured by our wide geographi-

cal range and astounding population
growth—is the combination of human

1 Department of Anthropology,
University of Oregon, Eugene, Oregon 97403-1218; e-mail: jerland@

oregon.uoregon.edu.

287

1059-0161/01/1200-0287/0 2001 Plenum Publishing
Corporation

C

288
Erlandson

intelligence, adaptive flexibility,
and technological sophistication. In a broad his-

torical or evolutionary framework,
humans are the ultimate in generalists and

opportunists, omnivores who thrived in
the widest range of earthly environments,

both natural and cultural. On a planet
whose surface is almost 75% water, where

life itself is dependent on water to
survive, and where our ancestors have success-

fully adapted for at least 2.5 million
years, it has always seemed strange to me that

modern anthropological theory has
maintained that aquatic resources and habitats

were not systematically used by humans
until relatively recently (e.g., Binford,

1968; Cohen, 1977; Osborn, 1977a,b;
Waselkov, 1987; Washburn and Lancaster,

1968; Yesner, 1987). As Bass (1972, p.
9) noted “even our land masses are crossed

and broken by rivers and streams or
dotted with lakes.” Yet among the 10 major

habitats listed by Gamble (1994, pp.
10–11, 1998) as significant to our early an-

cestors as they spread around the
earth, coastlines, lakeshores, and other aquatic

habitats are nowhere to be found.

As Washburn and Lancaster (1968,
p. 294) argued more than 30 years ago,

many archaeologists still seem to
believe that

During most of human history,
water must have been a major physical and psychological

barrier and the inability to cope
with water is shown in the archaeological record by the

absence of remains of fish,
shellfish, or any object that required going deeply into water or

using boats. There is no evidence
that resources of river and sea were utilized until this late

pre-agricultural period . . . for
early man, water was a barrier and a danger, not a resource.

More recently, Yesner (1987, p. 285)
stated categorically that the ”historical fact

that maritime resources were not
exploited until relatively late in the prehistoric

record has attracted a general
consensus. . . . A real commitment to maritime life-

ways did not precede late Upper
Paleolithic times.”

If such statements are accurate,
how did hominids spread around the globe,

colonizing much of Africa and Eurasia
by at least a million years ago, without the

aid of floats, boats, or the
capability to cross sizable bodies of water? How did

they survive in such a wide range of
landscapes when aquatic habitats were such

a physical and psychological
impediment? Why would our omnivorous hominid

ancestors—problem solvers and keen
observers of the world around them—ignore

aquatic resources when hundreds of
highly visible nonhuman predators and omni-

vores do not? Why is there so little
archaeological evidence for the use of marine

resources until postglacial times, long
after the well-documented maritime colo-

nization of island Southeast Asia and
greater Australia? Is it really possible that

aquatic resources were virtually
ignored for more than 99% of human history?

I believe the general perception
that humans only began to seriously adapt to

aquatic environments during the last
15,000 years or so has had a stultifying effect

on the evolutionary study of aquatic
adaptations and societies, maritime migrations,

and the development of boats and other
seafaring technologies. Such perceptions

peripheralize the significance of
aquatic habitats in human evolution, relegating

them to an essentially incidental role
in the broad-spectrum revolution leading to

The Archaeology of Aquatic Adaptations
289

agricultural societies and
civilizations. Thus maritime adaptations appear to play a

marginal role in a relatively brief
process during which the human developmental

trajectory departed from its natural
course as population growth forced humans

into increasingly artificial modes of
subsistence and production.

As Yesner (1987) noted, however,
the picture of aquatic resources as marginal

foods of “last resort” is out of
step with historical and archaeological data that

suggest that maritime or aquatic
hunter-gatherers were generally more sedentary,

populous, and culturally complex than
their terrestrially based interior neighbors

(Birdsell, 1953; McCartney, 1975;
Palsson, 1988; Townsend, 1980). Indeed, some

of the most complex and artistically
acomplished hunter-gatherers of all time de-

veloped in rich marine environments,
including many North Pacific peoples (the

Tlingit, Haida, Aleut, Koniag, etc.)
who lived adjacent to terrestrial environments

relatively unproductive for human
subsistence. Thus, while aquatic resources sup-

ported some of the most complex and
populous hunter-gatherer cultures on earth,

archaeological evidence for the
antiquity of aquatic resource use was extremely

limited. This results in a fundamental
paradox, where supposedly marginal aquatic

resources (although often both diverse
and abundant) appear to provide the eco-

nomic foundation for relatively complex
societies characterized by high popula-

tions and elaborated material cultures.
Despite some notable attempts to account

for problematic aspects of such models
(e.g., Osborn, 1977b; Yesner, 1987), this

aquatic paradox has yet to be
adequately explained or resolved.

In this paper, I discuss some of
these questions and problems by examining

the nature and antiquity of aquatic
adaptations. In the process, I address some of the

broader implications for our
understanding of human migrations, the evolution of

human subsistence and technology, and
current models of optimal foraging theory,

human economic intensification, and
the broad spectrum revolution. I begin with a

short summary of historical thought
about the archaeology of aquatic adaptations,

then discuss some epistemological,
methodological, and taphonomic problems

that currently prevent any real
consensus from being reached about the antiquity of

aquatic adaptations. I then review the
archaeological data available on early aquatic

resource use and maritime migrations
before discussing the broader implications

and some approaches I see as
potentially fruitful for the study of maritime and

aquatic adaptations as we embark on our
voyage into the twenty-first century.

A BRIEF HISTORY
OF THOUGHT

The study of coastal and other
aquatic societies has a long history in an-

thropology and archaeology, one that
closely reflects the general development of

the two fields. Despite this long
history, recent decades have seen a lively de-

bate about the nature of aquatic
environments, their economic productivity for

human societies, and the role they have
played in human evolution (e.g., Bailey,

1975, 1978; Binford, 1968; Claassen,
1991, 1998; Erlandson, 1988, 1994; Fischer,

290
Erlandson

1995a; Glassow and Wilcoxon, 1988;
Isaac, 1971; Jones, 1991; Moseley, 1975;

Osborn, 1977a; Parmalee and Klippel,
1974; Perlman, 1980; Price, 1995; Quilter

and Stocker, 1983; Raymond, 1981;
Sauer, 1962; Waselkov, 1987; Washburn and

Lancaster, 1968; Wilson, 1981; Yesner,
1980, 1987). Prior to the development of

the “New Archaeology” of the 1960s
and 1970s, however, there was little or no

coherent body of theory on the broader
evolution of aquatic adaptations. The opin-

ions expressed on such matters were
generally linked to regional discussions and

varied widely (see Clark, 1936, p. 140;
Morgan, 1877; Uhle, 1907). Nonetheless,

as Clark (1936) and many others
documented the close association of abundant

and widespread shell mounds with
postglacial shorelines, the development of shell

middens and relatively intensive
aquatic economies gradually came to symbolize

an important component of the
post-Pleistocene broad spectrum revolution (see

Bailey, 1978; Binford, 1968). As an
emphasis on theory, method, and broad syn-

thesis came into vogue in the 1960s and
1970s, moreover, considerable interest

focused on more global approaches to
the nature and antiquity of human adapta-

tions to aquatic environments.

In 1994, largely for heuristic
purposes, I characterized the more polarized

viewpoints in this debate as “Garden
of Eden” versus “Gates of Hell” models

(Erlandson, 1994, p. 273). Garden of
Eden theorists, I suggested, saw coastal or

aquatic habitats as veritable
cornucopia where a diverse array of foods—essentially

inexhaustible and easily harvested—was
available (e.g., Cutting, 1962; Fischer,

1995a; Hewes, 1968; Morgan, 1877, p.
21; Okladnikov, 1965, pp. 114–115; Sauer,

1962). On a global level, such
assertions may best be illustrated by Sauer’s de-

scription of the role of the sea in
human evolution.

. . . the path of our evolution
turned aside from the common primate course by going to the

sea. No other setting is as
attractive for the beginnings of humanity. The sea, in particular

the tidal shore, presented the
best opportunity to eat, settle, increase, and learn. It afforded

diversity and abundance of
provisions, continuous and inexhaustible. It gave the congenial

ecologic niche in which animal
ethology could become human culture (Sauer, 1962, p. 45).

Similar statements linked to specific
ethnographic accounts for some coastal groups

or to regional archaeological sequences
limited to the last 5,000 years were es-

poused by a number of authors. Such
glowing assessments often ignored the fact,

however, that archaeological records
for the same regions showed little evidence

for such aquatic largesse dating back
more than a few millennia. On a global

level, moreover, the accumulation of
archaeological data and the development of

chronometric dating techniques made
such statements increasingly problematic.

If aquatic resources were so
productive, why was there relatively little evidence in

the archaeological record for their
exploitation until very late in human prehistory?

After the 1960s, following the
lead of Uhle (1907) and others, a number of

scholars explicitly asserted that
aquatic habitats and resources, when compared

to the hunting of large terrestrial
game, were relatively unproductive for human

exploitation (e.g., Bailey, 1978;
Cohen, 1977; Gamble, 1986, pp. 35–36; Hogg

The Archaeology of Aquatic Adaptations
291

et al., 1971; Osborn, 1977a). These
Gates of Hell models articulated nicely with

the prevailing view of the time that,
prior to the development of agriculture, male-

dominated big-game hunting was the
driving force in human physical, cultural,

and technological evolution. Shellfish
and other aquatic foods, generally viewed

in such models as marginal or even
starvation foods, were portrayed as small and

costly to harvest or process, poor
sources of nutrition, relatively unpredictable or

unreliable, or requiring high
technological investments (boats, etc.) to access. The

fact that collecting shellfish and
other small aquatic foods was primarily women’s

work in most ethnographic societies
further marginalized their importance in hu-

man economies (Claassen, 1998, p. 175).
Gates of Hell models proposed, there-

fore, that the archaeological record
accurately reflected the low productivity of

aquatic resources and the relatively
low value placed on them by many forag-

ing peoples. They argue that humans did
not systematically or intensively harvest

aquatic resources until the
productivity of terrestrial hunting had been reduced by

the intensive harvest pressure of
growing human populations or by the postglacial

extinction of the Pleistocene
megafauna. Thus the use of aquatic resources was

(and is) often assumed to be evidence
for population pressure and environmental

degradation. Osborn, the most ardent
advocate of this position, argued that our an-

cestors “ignored” shellfish and
other aquatic resources for 99% of human history

(Osborn, 1977b, p. 301) and that the
low productivity of marine resources was

virtually universal.

. . . marine resources are
low-return subsistence resources due to a need for labor inten-

sification, in the case of
shellfish and small food package-sized organisms, and due to

their low protein content. A
number of factors combine to create an evolutionary threshold

that is too costly for human
populations to cross unless they are experiencing density-

dependent selection. This
subsistence-related threshold is so costly to cross, in fact, that,

given the option, we should expect
to see human groups shift away from the exploita-

tion of the sea, at least in
nonindustrial societies, whenever possible (Osborn, 1977a,

p. 177).

In practice, relatively few
published opinions can easily be categorized into

such polarized schemes, and most
scholars generally recognize that the situation

is considerably more complex.
Nonetheless, something closer to the Gates of Hell

model has heavily influenced the work
of some of the most influential scholars who

have worked with or discussed coastal
or other aquatic archaeological sequences

(e.g., Bailey, 1975; Binford, 1968;
Cohen, 1977; Fagan, 2001; Gamble, 1986;

Hayden, 1981; Isaac, 1971; Kelly, 1996;
Washburn and Lancaster, 1968).

To square such a dismal view of
the prospects of aquatic peoples with the

evidence that many coastal societies
were characterized by relatively high popula-

tion densities, sedentism, and cultural
complexity—the coastal paradox—required

further explanation. Osborn (1977b)
argued that the population density of aquatic

societies was exaggerated because their
offshore territories were not included

in density calculations, but he could
not resolve the more important issues of

sedentism and cultural complexity.
Cohen (1981) argued that the complexity of

292
Erlandson

Northwest Coast societies was a result
of their high population densities, but

never adequately explained how they
attained such high populations in supposedly

marginal environments. Yesner (1987)
developed the most explicit and sophisti-

cated explanation for the coastal
paradox, arguing that marine and other aquatic

environments were relatively
unproductive until the post-Pleistocene period, when

a combination of megafaunal
extinctions, climatic amelioration, sea level stabiliza-

tion, and the development of mature
coastal habitats allowed coastal populations

to bloom. Thus, he argued, humans did
not intensively utilize aquatic resources

until relatively late in human history,
but the growing productivity of postglacial

aquatic habitats ultimately fostered
the high populations, sedentism, and complex-

ity typical of many Middle or Late
Holocene coastal societies. Problems with this

model include significant variation in
the patterns and timing of megafaunal ex-

tinction or survival, the considerable
evidence for aquatic adaptations prior to such

widespread extinctions, and little
evidence that marine and other aquatic resources

were relatively unproductive prior to
sea level stabilization.

As we shall see, none of these
explanations adequately accounts for either the

basic paradox of supposedly low aquatic
productivity versus high human popula-

tions and cultural complexity, or for
the emerging archaeological data that suggest

that aquatic adaptations developed
earlier and were more widespread than pre-

viously believed. Nor do they explain
how archaeologists armed with essentially

identical data sets can come to such
radically different conclusions about the de-

velopment of such basic aspects of
human economies. To explore these problems,

however, we must first review some of
the different perspectives on the nature

of aquatic resources, then examine some
epistemological problems that inhibit a

comprehensive understanding of the
evolution of aquatic adaptations.

AQUATIC
RESOURCES

Much has been said about the
productivity of various classes of aquatic re-

sources: shellfish, fish, sea
mammals, waterfowl and seabirds, amphibians, plants,

and others. I do not review these
arguments in detail, for such a task could easily

be the subject of an entire paper. It
is important to my later arguments, however,

to examine some of the divergent
opinions expressed about the nature of various

classes of aquatic resources. Also
significant is the fact that aquatic resources are

often collectively lumped as “small”
resources, with the unwarranted assumptions

that they are therefore less productive
than terrestrial game animals for human

subsistence and their presence in
archaeological sites represents de facto evidence

for resource stress or economic
intensification. In this section, I examine some dis-

parate viewpoints about the major
classes of aquatic resources, recognizing that

classes of aquatic organisms not
discussed (amphibians, reptiles, insect larvae,

plants, etc.) may also be significant
resources in some areas.

The Archaeology of Aquatic Adaptations
293

Freshwater

Curiously, perhaps the single
most important aquatic resource for humans,

freshwater for drinking, is seldom
discussed. This may be because our dependence

on water is so fundamental and so
crucial to survival that it is taken for granted. It

is significant, however, because the
almost daily need for drinking water tethered

our ancestors to aquatic habitats for
most of human history. More than any other

resource, drinking water determined
where they settled and where they went,

especially in relatively arid regions.
Maintaining this crucial lifeline to aquatic

habitats, hominids would have spent a
great deal of time observing the behavior of

animals in such environments, including
many terrestrial and amphibious predators

or scavengers that fed on aquatic
animals (see Erlandson and Moss, in press).

Under these circumstances, it seems
unlikely that hominid hydrophobia would

have prevented similar opportunistic
harvesting of shallow water fauna by some

of our earliest ancestors living along
the shores of African lakes. With general

similarities between many of the
animals (fish, shellfish, birds, etc.) that live in

lakes, rivers, estuaries, and marine
habitats, it also seems unlikely that a significant

learning curve would have been required
to transfer such skills between aquatic

habitats. The intensity of such aquatic
harvesting probably varied tremendously,

of course, depending on the relative
productivity of such activities compared to

the other subsistence pursuits
available to a group at various times.

As noted above, the notion of a
long-standing inability of hominids to cope

with aquatic habitats is also difficult
to reconcile with the fact that our human

ancestors now appear to have spread
from Africa into southern Eurasia by about

1.7 million years ago. How did they
accomplish such extensive and early migrations

if they were afraid of the water and
incapable of either swimming or constructing

simple rafts, boats, or other flotation
devices?

Shellfish

No class of aquatic resources has
generated more debate among archaeol-

ogists than shellfish (e.g., Bailey,
1975, 1978; Buchanan, 1988; Claassen, 1991,

1998; Erlandson, 1988, 1991; Glassow
and Wilcoxon, 1988; Jones and Richman,

1995; Meehan, 1977, 1982; Meighan,
1969; Moss, 1993; Noli and Avery, 1988;

Osborn, 1977a; Parmalee and Klippel,
1974; Quilter and Stocker, 1983; Waselkov,

1987; Yesner, 1987). The generic term
shellfish is usually used to refer to a vari-

ety of aquatic invertebrates, dominated
by molluscs (bivalves and univalves), but

also including crabs, sea urchins,
barnacles, shrimp, and other relatively common

organisms. Although the size of
shellfish taxa utilized by humans varies consider-

ably, from large octopi or giant clams
to very small bivalves or gastropods, most

shellfish are relatively small
organisms. What they lack in size, however, many

294
Erlandson

shellfish make up for in quantity and
accessibility—many types are found in large

and sessile aggregations. While most
shellfish provide nutritious sources of com-

plete animal proteins and some vitamins
or minerals, most are relatively low in fat,

carbohydrates, and calories (see
Sidwell, 1981). Although shellfish beds have of-

ten been portrayed by anthropologists
as relatively unproductive, biological studies

indicate that mussel beds produce one
of the highest rates of biomass production

on earth (Jones and Richman, 1995).

Since at least the early 1900s,
many archaeologists have depicted these diverse

and seemingly innocuous creatures as
marginal, secondary, or even starvation foods

for humans.

. . . procuring the essentials of
life by collecting shells in itself indicates a low form of human

existence. In all parts of the
world, even today, people may be seen on the shore at low water

gathering for food the shells
uncovered by the retreating tide . . . these people always belong

to the lower classes of society
and lead in this manner a primitive as well as a simple life

(Uhle, 1907, p. 31).

Some archaeologists bolstered such
arguments with simple comparisons of the

nutritional content of shellfish
versus large land mammals. Bailey (1978, p. 39)

calculated, for example, that 156,800
cockles were required to provide the caloric

yield of one red deer. Some of these
comparisons were inaccurate, and others

ignored the fact that shellfish may
sometimes have been used primarily as a protein

source or that they were often a
relatively predictable and readily available meat

source that could be gathered by
virtually all members of society, including women,

children, and the elderly (see
Erlandson, 1988; Glassow and Wilcoxon, 1988;

Meehan, 1977). The fact that shellfish
gathering was done primarily by women in

most ethnographic societies (Claassen,
1998, p. 175; Moss, 1993, p. 632), in fact,

suggests that such comparisons of
shellfishing versus hunting yields may often be

inappropriate.

Some scholars have also argued
that the small size of shellfish, their relatively

low caloric content, and their
generally high ratio of shell to meat meant that they

were relatively laborious to process
(e.g., Osborn, 1977a; Waselkov, 1987). Others

countered that they required little
search time or technological investment and

could provide highly reliable and
relatively large meat yields that could buffer

the high failure rates of hunting
forays (e.g., Jones, 1991; Meehan, 1982). While

some researchers extolled shellfish as
an efficient protein source (Erlandson, 1988),

others noted that a heavy reliance on
lean shellfish meats could produce “protein

poisoning” (Noli and Avery, 1988),
and still others pointed out that a reliance

on many land mammals (bison, rabbits,
etc.) could produce the same problem

(Buchanan, 1988). While some criticized
shellfish as a resource highly susceptible

to periodic El Nino, storm, or red tide
events, others pointed out that such problems

could sometimes be predicted and
controlled for (Moss, 1993, pp. 640–641) and

that agricultural products and other
terrestrial resources were equally susceptible

to floods, droughts, disease, and
other problems (Quilter and Stocker, 1983).

The Archaeology of Aquatic Adaptations
295

Finally, though Osborn (1977a,b)
and others have used ethnographic or

historical accounts to support the
notion that shellfish were marginal or starvation

foods, Moss (1993) clearly exposed the
complexities and potential androcentrism

often inherent in such accounts. I was
present during her interview of an elderly

Tlingit friend, Richard Newton, who
insisted shellfish were not a major food for

his people. Responding to questions
about the incredible abundance of shellfish

remains in Tlingit village and camp
sites, however, Mr. Newton eventually char-

acterized their dietary role as similar
to bread and butter—long a staple in western

society (Moss, 1993, p. 643). When
asked about this contradiction, he patiently

explained that the ideal Tlingit man
needed to work hard to succeed, that shellfish

encouraged laziness because they were
too easy to collect, that they were gathered

primarily by women, but that they also
were regularly gathered and consumed

by Tlingit men (Moss, 1993). Certain
types of shellfish were especially prized by

Tlingit men, in fact, because they were
said to enhance the libido.

All this debate has had little
effect on the pervasive notion that shellfish

and other small resources are lower
ranked by human foragers (e.g., Broughton

and O’Connell, 1999; Renfrew and
Bahn, 1996, p. 282). Fagan (2001, p. 341)

concluded, for instance, that “no one
can believe that mollusks were the staple

diet” of any ancient society.
Following such assumptions, many archaeologists

continue to view the appearance of
shell middens in archaeological sequences

as evidence for human demographic
pressure, environmental degradation, and

economic intensification (e.g., Cohen,
1977; Hayden, 1981; Waselkov, 1987). The

postglacial florescence of shell
middens adjacent to aquatic habitats around the

world, therefore, has become
essentially synonymous with the anthropological

notion that human economies were
transformed by a global and relatively recent

broad-spectrum revolution.

Fish

Similar debates have taken place
over the nature and productivity of fishing

(e.g., Butler, 1996; Clark, 1948;
Garson, 1980; Kelly, 1996; Limp and Reidhead,

1979; Lindstrom, 1996; Morgan, 1877;
Osborn, 1977b; Rick and Erlandson, 2000).

Literally thousands of different
varieties of fish inhabit the wide range of aquatic

habitats, from the deep abyssal floors
of the oceans to high mountain lakes. Even as

adults, these fish range in size from
tiny gobies to the gigantic whale shark. Some are

largely solitary and relatively rare,
while others are incredibly abundant and swim

in concentrated schools numbering in
the millions. Some aquatic communities,

moreover, are characterized by a
diversity and abundance of fish; others contain

only one or two species and even these
are relatively rare. Still other communities

have low species diversity but a
relatively high piscine biomass.

The nutritional value of fish
also varies considerably, especially the fat and

calorie content of various species.
Although generally low in carbohydrates, fish

296
Erlandson

are a relatively nutritious source of
protein, vitamins, and minerals (Sidwell, 1981;

Watt and Merrill, 1975). Fish eggs,
which can sometimes be harvested in large

quantities, are also generally very
high in protein and calories. Fish flesh and protein

are also highly digestible and
metabolized more efficiently by the human body

than the meat of land mammals. High
rates of fish consumption in modern human

populations—especially certain fish
oils—also seems to be generally correlated

with lower rates of disease and greater
longevity.

Opinions expressed about the
economic productivity of fishing vary widely.

Some accounts have portrayed fishing
as an extremely productive activity, begin-

ning with Morgan’s idealized
statement that “Fish were universal in distribution,

unlimited in supply, and the only kind
of food at all times attainable” (Morgan,

1877, p. 21). In contrast, in comparing
fishing to more traditional hunting activities,

Kelly (1996, p. 209) stated that

. . . fish are different. Some
species, especially surface feeders, will give away their presence,

but not bottom feeders. And fish
cannot be tracked—this is a particular problem in exploiting

oceanic fish. The forager can
only go to a likely place to find fish, then begin searching

randomly. If there are no fish
there, the forager could waste quite a bit of time before

accepting this as likely.

Kelly’s characterization, however, is
at odds with many types of marine fishing,

including the extremely productive and
predictable fishing that can characterize

halibut or cod banks, kelp beds, and
some other nearshore habitats.

Others have argued that fishing
requires relatively sophisticated knowledge

and high technological investments.
Experimental work by Limp and Reidhead

(1979) suggested, however, that under
the right circumstances riverine fishing

could be extremely productive even
without complex technologies. In some aquatic

habitats, the seasonal drying of ponds
or pools can strand fish in shallow water or

on mud flats where they can be easily
collected. In some lakes, periodic hypersaline

or anoxic conditions can also lead to
massive fish kills in which large windrows

of dead fish are deposited on the
beach (e.g., Butler, 1996, p. 701). Quilter and

Stocker (1983, p. 549) described an
apparently regular Peruvian phenomenon

known as “anchovy beaching,” in
which hundreds of thousands of small fish strand

themselves on the beach roughly four
times a year. Spawning fish (salmon, herring,

lamprey eels, grunion, and many others)
also can be highly vulnerable to human

predation, and such spawning runs are
often highly predictable, facilitating the

logistical planning required for mass
harvesting and the processing of fish for

storage.

Even when more sophisticated
technologies are required to capture fish,

these need not be especially elaborate
or expensive to produce. Dip nets or small

tidal weirs, for instance, can greatly
facilitate the mass harvest of small fish in

truly impressive yields. Before
commercial overexploitation devastated many of

California’s marine fisheries,
enormous schools of sardines and anchovies were

available in nearshore and estuarine
habitats. With the aid of boats and dip nets,

The Archaeology of Aquatic Adaptations
297

huge quantities of these small fish
could be captured quickly and easily dried for

later consumption. Still, considering
the lack of evidence for weaving techniques

prior to the advent of the Upper
Paleolithic, even relatively simple fishing tech-

nologies involving cordage, baskets,
nets, or composite projectiles may have been

beyond the capabilities of hominids
prior to the appearance of anatomically mod-

ern humans. And some fishing
activities—especially those requiring large nets,

sophisticated boats, or elaborate weir
structures—would have required consider-

able investment in materials, labor,
and maintenance, as well as intellectual and

communication skills that may have been
beyond the capabilities of our archaic

ancestors.

Despite such technological
constraints, a number of cultural ecological stud-

ies have modeled the productivity of
various fishing activities relative to alterna-

tive terrestrial subsistence pursuits
(e.g., Osborn, 1977b; Perlman, 1980; Simms,

1987). Some of these estimates are
based on incomplete data or the use of inap-

propriate technologies in potentially
depleted modern environments, but they are

still informative, suggesting that the
productivity of fishing varies tremendously.

The most sophisticated analysis of
which I am aware is Lindstrom’s study of the

Truckee River fishery in the western
Great Basin (Lindstrom, 1996), which sug-

gests that fishing harvests using a
number of different aboriginal techniques were

higher than the return rates calculated
by Simms (1987) for terrestrial hunting.

Lindstrom’s projected return rates
varied considerably, however, and some meth-

ods of fishing produced yields that
were considerably less productive than many

terrestrial alternatives.

Aquatic Mammals

There has also been considerable
debate about the nature, antiquity, and eco-

nomic productivity of aquatic mammal
use, especially marine mammals such as

whales, seals, sea lions, sirenians
(sea cow, manatees, etc.), and sea otters (e.g.,

Clark, 1946, 1947; Colten and Arnold,
1998; Erlandson et al., 1998; Hildebrandt

and Jones, 1992; Jones and Hildebrandt,
1995; Lyman, 1995; Osborn, 1977b;

Workman and McCartney, 1998). Aquatic
habitats also are home to a variety of

freshwater animals (hippopotami,
beavers, otters, etc.) of various sizes, which

spend varying amounts of time in the
water and, like some marine mammals, may

sometimes be taken on land. Some
aquatic mammals are also not easily catego-

rized as clearly marine or freshwater:
some seals or dolphins swim considerable

distances up rivers; seals live
permanently in Lake Baikal, the Caspian Sea, and

other European lakes (Reeves et al.,
1992); some manatees are equally at home in

salt- or freshwater habitats; and river
otters and other typically freshwater mammals

may also spend time in brackish or
saltwater habitats.

Clearly most aquatic mammals are
not small resources. They include many

of the largest animals on earth, which
until devastated by commercial whaling or

298
Erlandson

hunting also were relatively abundant
along many of the world’s coastlines. Many

marine mammals weigh well over 500 kg,
with the largest whales weighing over

100,000 kg. Osborn (1977b) argued that
most aquatic mammals occupy positions

relatively high in the food chain,
which limits their numbers relative to the primary

productivity of the world’s oceans.
Such global modeling probably had little or

no relevance, however, to maritime
peoples such as the Koniag or Aleut, who

lived in proximity to biannual
migrations of hundreds of thousands of whales and

pinnipeds (see Haggarty et al., 1991).

Like virtually all mammals, the
meat and organs of aquatic mammals are

relatively rich sources of nutrients,
high in protein, vitamins, and minerals (see

Heller and Scott, 1967; Osborn, 1977b;
Sidwell, 1981). Many aquatic (especially

marine) mammals also have a thick layer
of subcutaneous blubber that provides

them with insulation and human hunters
with a rich source of fat and calories. These

fat deposits can also be rendered into
oil that may be stored for later consumption

or used in lamps as a source of heat
and light. The skins, bones, teeth, ivory, and

baleen of many aquatic animals also
provide valuable raw materials used in a

variety of technologies (houses, boats,
clothing, tools, ornaments, etc.). Among

many societies that actively pursue
large aquatic mammals, successful hunters may

also gain significant status,
prestige, and possibly even reproductive advantages.

Such potentially lucrative
economic payoffs must be measured against the

costs and risks of procuring aquatic
mammals. Sea mammal hunting can be a dan-

gerous and seasonal pursuit, especially
in offshore marine settings, and successful

hunting forays are often relatively
rare. Like fishing, some forms of aquatic hunting

may also require relatively complex and
expensive technology, including seawor-

thy boats and related hunting gear that
represent a significant investment of energy

to produce and maintain. This is
particularly true for many types of sea hunting

recorded among ethnographic marine
hunters. Many of these peoples had high

population densities and had hunted sea
mammals for millennia, however, with

negative effects on the distribution
and density of local prey populations (Jones and

Hildebrandt, 1995; Lyman, 1995). Prior
to such impacts, many pinnipeds may have

been taken relatively easily while
hauled out in breeding or birthing colonies on

islands or other isolated coastal
locales. Where abundant, scavenging of cetacean

and pinniped carcasses off the beach
could also provide large (and potentially

huge) subsistence dividends, with
minimal technological or search costs (Smith

and Kinahan, 1983). Finally, the costs
of manufacturing and maintaining boats

must be measured against the greater
overall efficiency achieved in a variety of

hunting, fishing, and transportation
activities.

As with virtually all classes of
aquatic and terrestrial resources, there is con-

siderable variability in the
characteristics of aquatic mammals and their economic

potential. This includes aspects of
their biology and behavior, their abundance and

availability to humans, the methods
used to procure them, and the relative produc-

tivity of various procurement
strategies versus subsistence alternatives. Given this

The Archaeology of Aquatic Adaptations
299

diversity, it should be no surprise
that different researchers have reached quite dif-

ferent conclusions about the general
role of aquatic mammals in human economies.

In recent discussions, for instance,
various researchers have viewed sea mam-

mals as either central or peripheral to
the development of maritime adaptations

along the Pacific Coast of North
America. Hildebrandt and Jones (1992; Jones

and Hildebrandt, 1995) proposed that
because of their large size and vulnerability

to predation in rookeries, some seals
and sea lions were the focus of early ma-

rine hunters, with later technological
developments (boats, etc.) representing labor

intensification as human impacts on
pinniped populations increased and hunting

strategies changed. Colten and Arnold
(1998) and Erlandson et al. (1998) noted

little evidence for an early focus on
pinniped hunting in the area, however, and

suggested that its general economic
importance may have been overemphasized

(see also Kent, 1989, p. 5; Workman and
McCartney, 1998, p. 362). Central to

resolving such debates are problems
related to recovering and interpreting repre-

sentative samples of sea mammal remains
and estimating their dietary contribution

within the larger economies of human
societies.

PROBLEMS IN
PARADIGMS

Underlying such debates, but
often pushed well into the background, is the

ambiguity of the archaeological record
itself. In some cases, diverging opinions

have been supported with data from
different regions. In others, nearly opposite

conclusions were drawn from virtually
the same archaeological record. How is

it possible for researchers to reach
such different conclusions based on the anal-

ysis of the same body of data? The
answer to that question lies in a variety of

taphonomic, methodological,
interpretive, and theoretical problems that make our

reconstructions of the history of
aquatic societies fraught with uncertainties. The

divergence of opinions about the
antiquity of aquatic adaptations can be attributed

to a variety of problems with the
archaeological record itself, to differences in

the way individual archaeologists
believe the record should be interpreted, and to

differences in the preconceptions of
various researchers.

Definitions

In part, different opinions can
be attributed to the general lack of defini-

tion for what constitutes a dietary
staple, systematic or intensive resource use,

or terms such as coastal, aquatic,
littoral, or maritime adaptations (Workman and

McCartney, 1998). Definitions for
coastal or maritime adaptations have varied, for

instance, from those groups who procure
some portion of their sustenance from

the sea to those who go to sea in boats
and rely on other specialized technologies.

Recognizing the complexity and
diversity inherent in northern cultures, Fitzhugh

300
Erlandson

(1975, p. 344) tried to bring some
order to the classification of maritime societies

by defining five broad adaptive
types: modified interior, interior-maritime, mod-

ified marine, maritime, and riverine.
In his work on the Oregon coast, Lyman

(1991) differentiated littoral from
maritime adaptations, the latter representing

groups who went to sea to obtain much
of their sustenance. Finally, in an attempt

to operationalize the definition of
maritime societies for anthropologists, Yesner

(1980, p. 728) defined “fully
maritime” peoples as those obtaining at least 50% of

their calories or protein from marine
sources. This definition is easily adapted to

riverine or lacustrine peoples, but in
practice it is difficult to accurately or precisely

quantify the dietary contribution of
aquatic versus terrestrial resources. Isotopic

and trace element studies of human bone
have improved our ability to quantify

general aspects of ancient diets, but a
variety of problems (diagenesis, varying

photosynthetic pathways, etc.) continue
to limit such studies.

At times, we must even confront
the issue of what constitutes an aquatic

versus terrestrial resource. How do we
classify a salmon or other anadromous fish

that may be caught in the ocean one
week, in a river or lake the next week, or

scavenged from the shoreline the next?
How do we classify the beaver, hippopota-

mus, crocodile, land otter, or many
other animals that spend a good deal of time

in aquatic habitats but may also be
captured on land? Are seabirds (or their eggs)

taken from terrestrial colonies aquatic
or terrestrial resources? What about seals

or sea lions taken from onshore
rookeries? Finally, how do we classify a deer or

elk captured—as they were sometimes
taken along the Northwest Coast of North

America—while swimming to or from
islands (Tveskov, 2000, p. 131) or nearly

paralyzed by the cold on the beach just
after such a swim? It might be argued that

such ambiguous cases are relatively
unusual, but I suspect they are more com-

mon than many of us recognize, and they
blur the arbitrary distinctions already

drawn between aquatic and terrestrial
resources or marine, estuarine, riverine, and

lacustrine habitats. If such
ambiguities can be recognized in modern habitats and

behaviors, moreover, how can we hope to
differentiate between such ambiguous

cases in the archaeological record?

Changing Sea Levels, Coastal
Erosion, and the Archaeological Record

Despite such ambiguities, the
single greatest problem in evaluating the his-

tory of aquatic adaptations lies in the
fact that sea and lake levels have varied

tremendously over the past 2 million
years, and erosion during high stands has

repeatedly obliterated the
archaeological record where evidence for early aquatic

resource use is most likely to be
found. Sea level today is among the highest of the

Quaternary, exceeded only by Last
Interglacial levels about 6 m higher than today.

Many scholars are rightfully hesitant
to assume that Pleistocene shell middens

were once widespread along submerged
shorelines. Geologically, however, there

The Archaeology of Aquatic Adaptations
301

is ample reason to believe the
archaeological record of coastal adaptations is se-

riously underrepresented (Kraft et al.,
1983). During the last glacial about 20,000

years ago, world sea levels stood
between about 100 and 125 m below present,

exposing broad coastal plains around
the world that have virtually all been inun-

dated as seas rose to their present
levels. Similar cycles have occurred numerous

times during the Plio-Pleistocene,
causing enormous and highly variable changes

in coastal geography around the world.

Worldwide, only Africa and
Eurasia were occupied by hominids when sea

levels were last comparable to today.
Along such Old World coastlines, the Last In-

terglacial sea stand of 125,000–130,000
years ago cut erosional platforms that may

have destroyed most evidence for
earlier coastal occupations. In fact, each time

global sea levels have risen
significantly the record of hominid occupation associ-

ated with lower shorelines has either
been inundated, destroyed by coastal erosion,

or both. Even today, with sea level
roughly 6 m below the Last Interglacial high,

many important coastal sites (e.g.,
Klasies River Mouth caves, Gorham’s Cave,

Grotta dei Moscerini, Daisy Cave)
occupied between about 125,000 and 10,000

years ago are being destroyed by marine
erosion. Uplifted shorelines associated

with earlier interglacials are present
in some areas, but the periods of sea level

maxima represent just a small fraction
of the Pleistocene. It should be no surprise

that associated occupation sites (e.g.,
Terra Amata) are rare. Much more common

are localities such as those in North
Africa and the Levant, where Lower Paleolithic

artifacts (hand axes, etc.) have been
found redeposited on raised marine terraces,

testifying to the destruction of
ancient sites located in coastal or pericoastal settings

(e.g., Bar-Yosef, 1994; Howe, 1967).

Equally important for
understanding the evolution of coastal and aquatic

adaptations are the effects of sea
level change on the paleogeography of coastal

localities. As sea levels rise or fall,
coastlines move laterally in response to such

changes; the environmental setting of
archaeological sites can change dramatically.

Reconstructions at coastal sites with
long occupational sequences have shown

that the exploitation territories of
many sites located on the modern coast were

entirely terrestrial during earlier
occupations (e.g., Parkington, 1981; Shackleton

and van Andel, 1980). The maximum
lateral movements of coastlines during the

last 20,000 years, for instance, have
varied from as much as 1000 km in some

areas (e.g., northern Australia) to
less than 1 km in others. Areas where shorelines

have moved less than about 10 km are
unusual and tend to be strongly correlated

with relatively early evidence for
coastal occupations (Erlandson, in press; see

ahead). Reconstructing the
paleogeography in the vicinity of coastal sites is crucial,

because a cave or open site located on
the coast today may have been 5, 10, 50 km,

or more from the coast at various times
during the last 25,000–125,000 years.

Study of modern coastal
hunter-gatherers suggests that they rarely travel more

than about 5 or 10 km from a home base
to gather foods (Bigalke, 1973, p. 161;

Meehan, 1982). When they do hunt or
forage further afield, the skeletal remains of

302
Erlandson

shellfish, fish, or sea mammals are
often not transported back to a residential base.

In most situations, therefore, sites
located more than about 5–10 km from an ancient

shoreline are unlikely to contain
substantial evidence for marine resource use.

Distances of even 1 or 2 km can
dramatically reduce the density of aquatic faunal

remains (Wing, 1977). During periods of
shoreline transgression or regression,

the intensity of aquatic resource use
at any given site should fluctuate depending

on its proximity to coastal habitats.
After the dramatic postglacial sea level rise

of the last 17,000 years, coastal sites
with long occupational sequences may show

evidence for an intensification of
marine resource use related primarily to changes

in local environments rather than a
regional diversification or intensification of

human subsistence (see Bailey, 1983a;
Parkington, 1981; Shackleton, 1988).

Some may argue that the loss of
early coastal sites can be mitigated by ex-

amining the antiquity of the human use
of lacustrine or riverine resources, but two

problems inhibit such comparisons.
First, it is not clear that the productivity and

diversity of most freshwater habitats
is comparable to marine or estuarine commu-

nities. Second, it is not clear if the
archaeological record of riverine or lacustrine

habitats is any more representative.
Such freshwater environments are also highly

dynamic, and climatic, glacial, and sea
level changes have had profound effects

on their structure and productivity.
Fluctuating lake levels also are common, and

shoreline erosion can produce
geological features essentially identical to marine

shorelines. In riverine systems,
moreover, erosive cycles can rapidly destroy sites

while depositional cycles can bury them
under large quantities of sediment. Thus

preservation and visibility problems
may be just as significant in some freshwater

systems as they are in marine
environments.

Differential
Preservation, Recovery, and Reporting

Another problem lies in the
differential preservation, recovery, and reporting

of organic remains. As we all know, the
shell and bone remains that constitute

the primary record of human use of
aquatic resources are not preserved in many

archaeological sites. Acidic soils, for
instance, or the gradual action of humic

acids in neutral soils, commonly lead
to the deterioration of shells and bones

in archaeological sites. In
comparatively recent sites, especially those occupied

by relatively sedentary peoples, the
accumulation of substantial shell middens can

mitigate the effects of soil acidity or
other factors that lead to the destruction of shell

or bone. For the Paleolithic or
Paleoindian periods, however, when most scholars

believe humans were relatively mobile,
the shell in many low-density middens may

have been insufficient to counteract
soil acidity. The same may be true of pericoastal

or other sites located some distance
from aquatic habitats, where the density of

aquatic food remains was limited by
transportation costs. My experiments with

shells and bones exposed to dilute acid
solutions also showed that shells generally

deteriorate faster than bones, probably
due to their higher calcium carbonate and

The Archaeology of Aquatic Adaptations
303

lower collagen or lipid content. At a
number of archaeological sites, including

Hidden Falls in southeast Alaska
(Erlandson, 1989, p. 139) and Die Kelders in

South Africa (Goldberg, 2000),
moreover, researchers found bone still recoverable,

while shells had either disintegrated
or were too deteriorated to recover or identify.

In the case of Die Kelders, despite the
fact that calcareous rock was abundant in the

site strata, decalcification
completely destroyed the shellfish remains in portions

of the site while bone fragments were
still relatively well preserved.

Among animal bones alone, the
denser and thicker bones of large land mam-

mals are more likely to be preserved in
most archaeological contexts (see Butler

and Chatters, 1994). There has been
relatively little experimentation on the compar-

ative survivability of skeletal remains
from terrestrial versus aquatic vertebrates,

but differential bone density is a
significant factor in preservation. The bones of

aquatic vertebrates generally have
lower densities and are probably more suscep-

tible to chemical dissolution and
mechanical breakdown. Except for the teeth of

some taxa (sharks, etc.), fish bones
are especially lightly built and often have very

high surface area to volume or mass
ratios, suggesting that they would be highly

vulnerable to chemical deterioration.
Some economically important fish (sharks,

rays, sturgeon, lamprey eels, etc.)
also have cartilaginous skeletons with very few

bony parts, and small bony fish (i.e.,
sardines, anchovies) are often eaten whole.

The bones of many aquatic mammals are
also relatively porous and may be prone

to differential deterioration from
mechanical and chemical processes.

Numerous studies clearly show
that the recovery techniques used by archae-

ologists dramatically affect the
interpretations drawn from the recovered assem-

blages. Studies of faunal recovery, for
instance, show that large proportions of the

fish bone and shellfish remains in
many sites are lost during screening of exca-

vated sediments through coarser (0.25
in. or larger) mesh sizes (e.g., Erlandson,

1994; Garson, 1980; Koloseike, 1968;
Moss, 1989). This is a crucial problem in

evaluating the evidence for aquatic
resource use in many early excavation reports,

where researchers had limited interest
in subsistence, faunal remains often were not

systematically recovered, or fine-screen
samples were not collected. Many inves-

tigators now routinely collect faunal
and floral samples through fine screening and

flotation, but others still rely on
cheaper and less systematic recovery techniques.

Because the importance of hunting
or scavenging large game animals has long

been emphasized, there have sometimes
been biases in the analysis or reporting

of other faunal remains from
archaeological sites. In many studies of Middle or

Upper Paleolithic subsistence, in fact,
the only subsistence remains reported on are

large land mammals (e.g., Barker, 1974;
Wolf, 1988), even in early coastal sites

that produced a variety of faunal
remains. Years ago, while visiting early sites in

Gibraltar, I was surprised to find a
number of bluefin tuna and mackeral vertebrae

in the Gibraltar Museum, materials
excavated from early Upper Paleolithic strata

at Gorham’s Cave. For some reason,
these fish bones were never mentioned in

any of the site publications, even
though reports on mammals, tortoises, birds, and

shellfish were all published (see
Baden-Powell, 1964; Eastham, 1968; Waechter,

304
Erlandson

1951, 1964; Zeuner and Sutcliffe,
1964). A similar problem is encountered for the

Middle and Upper Paleolithic levels at
Mugharet el‘Aliya, located near Tangier in

Morocco (Howe, 1967; Howe and Movius,
1947). One of the few Last Interglacial

sites from the south coast of the
Mediterranean, the Paleolithic cave deposits pro-

duced seal, fish, and “a series of”
mollusk remains (Howe and Movius, 1947, p.

21). Although vertebrate remains were
not quantified, they were at least identi-

fied (Arambourg, 1967). Description of
the shellfish remains was limited to the

statement that a “number of mollusks
were found in Layers 5, 6, and 9 of the ar-

chaeological deposits in the Mugharet
el‘Aliya, and were submitted to Dr. William

J. Clench of the Museum of Comparative
Zoology at Harvard. Nothing of value

for our purposes came of this however”
(Briggs, 1967, p. 187).

Such problems may have been due,
in part, to the dearth of specialists who

could identify and analyze the remains
of aquatic fauna. They are symptomatic,

however, of the lower priority
archaeologists traditionally assigned to resources

such as shellfish and fish that were
considered economically marginal or unimpor-

tant. In Howe’s synthesis of the
Mugharet el‘Aliya investigations (Howe, 1967),

for instance, the description of stone
tools is over 31 pages long, the vertebrate

remains are relegated to 5 pages in an
appendix, and the shellfish merit a single

short and obscure paragraph.

The
Hunting Hangover

Even if we can overcome such
analytical hurdles, another serious problem

still confronts us. This is the
persistent effect of the ”Man the Hunter” paradigm

on archaeology. The historical
overemphasis on hunting as central to early human

economies has been dealt with at length
elsewhere (e.g., Slocum, 1975; Zihlman,

1997). The remnants of this outdated
view are still with us, however, more than a

decade after most scholars recognized
that scavenging probably supplied much of

the meat early hominids consumed and
that gathering was much more important

than recognized in earlier
anthropological models. Comparative anatomy also tells

us that human dentition is
fundamentally adapted to omnivory and a relatively

eclectic diet (Scott and Turner, 1997,
p. 81). Evolutionary theory tells us that over

the long haul species are rarely well
served by excessive specialization. Modern

medicine and nutritional studies show
that dietary diversity is fundamental to

human health, growth, and reproductive
success. And common sense tells us that

as our hominid ancestors spread around
the globe, a fundamental part of their

success was their ability to adapt to a
variety of environments or situations and

their relatively eclectic and
opportunistic subsistence economies.

Nonetheless, many theoretical
discussions of subsistence inappropriately

compare the yields of shellfishing or
fishing to those of large-game hunting. If

many hominids relied heavily on
scavenging rather than hunting, for instance,

the relative productivity of gathering
shellfish should be compared to scavenging

The Archaeology of Aquatic Adaptations
305

yields in such cases and must have been
higher than previously estimated. Many

predictions based on optimal foraging
principles also inappropriately treat early

human societies as groups of generic
individuals, ignoring the gender or age-based

divisions of labor in hunting and
gathering activities typical of most recent foraging

cultures. Even for Holocene peoples,
therefore, comparisons of shellfishing and

hunting yields to predict dietary
breadth and subsistence choices may be inappro-

priate, since large-game hunting was
often a primarily male pursuit, and shellfish

and some other aquatic resources were
often collected mostly by women, children,

and older individuals. There is little
doubt, in fact, that the historical devaluation

of shellfish gathering in human
history is related to the fact that it was primarily

the work of women or commoners, to an
androcentric fascination with hunting,

and to biases in historical and
ethnographic accounts recorded primarily by men

(Claassen, 1991, pp. 278–279; Moss,
1993).

Until anthropology transcends some
pervasive misconceptions, the signifi-

cance of aquatic adaptations will
continue to be underemphasized in our recon-

structions of human evolution. These
misconceptions include (1) the notion that

large land mammals were virtually
always the most productive and highly ranked

resources for our hominid ancestors;
(2) that male-dominated hunting was always

the central force that shaped human
subsistence, settlement, and technological de-

velopments; (3) that the utilization of
aquatic resources is automatically evidence

for demographic pressure or resource
stress; and (4) that the archaeological record

preserves a representative picture of
our past.

ARCHAEOLOGICAL EVIDENCE FOR
THE ANTIQUITY

OF AQUATIC
RESOURCE USE

Given the nature and ubiquity of
such problems, is it any wonder that we

know so little about the history of
aquatic resource use? To understand the devel-

opment of aquatic adaptations we are
burdened with several fundamentally flawed

theoretical assumptions and blessed
with a relatively small number of assemblages

where faunal preservation is
exceptional and the full range of faunal remains were

systematically recovered and completely
reported. At the same time, any current

synthesis must rely on an
archaeological record that comes almost exclusively

from sites preserved above modern sea
level even though virtually all coastlines

dating between about 120,000 and 15,000
years ago now lie submerged and distant

from the modern coast.

Despite such problems, numerous
early sites with evidence for aquatic re-

source use have been listed over the
years by Osborn (1977a,b), Perlman (1980),

and Waselkov (1987) or mentioned by
others (e.g., Claassen, 1998; Klein and

Scott, 1986; Yesner, 1980). None of
these lists were exhaustive when they first

appeared, and additional data have
continued to accumulate in subsequent years.

In Tables I–III, I have compiled my
own lists of early “aquatic” sites—those that

306
Erlandson

Table I. Some Early Old World
Localities With Possible Evidence for Aquatic Resource Use

Description
of aquatic fauna

Locality/site and
associations Age (yr) Reference

Homo habilis

Senga 5, Possible use of
freshwater fish, 2.3–2.0M Harris et al., 1990;

Semliki molluscs, and
reptiles associated Meylan, 1990

River, Zaire with Oldowan
tools.

Olduvai Gorge, Possible use of
freshwater fish, 1.8–1.1M Leakey, 1971;

Tanzania crocodiles,
turtles, amphibians, Stewart, 1994

and molluscs.

Homo erectus

Olduvai Gorge, Possible use of
freshwater fish, 1.1–0.8M Leakey, 1971,

Tanzania crocodiles,
aquatic mammals 1994; Roe,

(hippo),
turtles, amphibians, 1994, p. 304;

molluscs, and
possibly salt. Stewart, 1994

Kao Pah Nam, Pile of
freshwater oyster shells 700K Fagan, 1990,

Thailand against cave
wall, associated with p. 120; Pope,

hearth and
land animal bones. 1989

Holon, Israel Freshwater turtle
(Trionyx sp.) shells 500–400 K Bar-Yosef, 1994,

and hippo
bones in Middle p. 246

Acheulian
assemblage of mostly

scavenged (?)
land mammals.

Mas des caves, Seal remains
found in cave site now ca. 400K Cleyet-Merle and

Lunel-Viel, located ca. 10
km from Madelaine,

France Mediterranean
coast. 1995, p. 306

Archaic Homo sapiens

∼350–300K

Hoxne, England Remains of fish,
otter, beaver, and Singer et al.,

waterfowl
associated with 1993; Stuart

Acheulian
deposits; distributions et al., 1993

similar to
artifacts, suggesting a

cultural
origin.

∼400–200K

Duinefontein 2, Sea bird
(penguin, cormorant) Klein et al., 1999a

South Africa remains in
Late Acheulian site

dominated by
land mammal bones.

∼300–230K

Terra Amata, Shellfish and
possibly fish remains de Lumley, 1969;

France associated
with multicomponent Villa, 1983

coastal
campsite.

∼186–127K

Lazaret, France Marine shellfish
in late Acheulian Cleyet-Merle and

context.
Madelaine,

1995

∼150 ± 50K?

Ramandils, Marine shellfish
(>300 fragments) in Cleyet-Merle and

France Middle
Paleolithic strata, probable Madelaine,

food remains.
1995

150 ± 50K

Kebibat, Rabat, Aterian shell
midden on Atlantic Souville, 1973,

Morocco coast,
associated with Neandertal pp. 73–81

remains.

∼130–40K

Presqu’ile du Aterian site on
coast near Berard, Roubet, 1969

Canal, contains
unspecified numbers of

Berard, limpets.

Algeria

∼130–50K

Haua Fteah, Marine shellfish
in Last Interglacial McBurney, 1967

Cyrenaica, strata.

Libya

The Archaeology of Aquatic Adaptations
307

Table I. (Continued)

Description
of aquatic fauna

Locality/site and
associations Age (yr) Reference

∼125–40K Arambourg, 1967;

Mugharet el ’Aliya, Marine
shellfish, fish, and monk seal

Morocco remains in
Mousterian/Aterian Howe, 1967

strata.

∼125–140K D´ b´ nath and

La Grotte Zouhrah, Aterian
assemblage with marine ee

Rabat, Morocco shellfish
(limpets, mussels, and Sbihi-Alaoui,

crab), Homo
sapiens remains. 1979

∼127–40K Roche and Texier,

Grotte des Aterian shell
midden on Atlantic

Contrebandiers, Coast
associated with Homo 1976; Souville,

Morocco sapiens
remains, abundant limpets. 1973, p. 112

∼125–50K Garrod et al., 1928

Devil’s Tower, “Thick
layers” of mussels over

Gibraltar Mousterian
hearths, and a “large

heap” of
marine shells.

∼125–50K Baden-Powell,

Gorham’s Cave, A variety of
marine shellfish remains

Gibraltar from several
Mousterian 1964; Waechter,

occupation
levels. 1951, 1964

∼115–65K Stiner, 1994

Grotta dei Diverse marine
shell remains (3100

Moscerini, fragments),
dominated by mussels

Latium, Italy and clams.
High rates of burning

suggest human
predation; chipped

shell tools.

>45K

Vanguard Cave, Mousterian strata
containing “clear Barton et al., 1999

Gibraltar evidence”
for marine shellfish use

by
Neandertals; includes mussels,

limpets,
cockles, etc., some burned.

>40K

Ras el-Kelb, Mousterian
occupation of coastal or Copeland and

Lebanon pericoastal
cave site, with small Moloney, 1998;

numbers of
marine shells recovered Reese, 1998

from various
occupation levels.

>40K

Salzgitter- Freshwater fish
and mollusk remains Butzer, 1971,

Lebenstedt, associated
with Mousterian p. 477; Cohen,

Germany assemblage.
1977

>37K

Grotta Breuil, Small numbers of
clam and limpet Stiner, 1994,

Latium, Italy shells from
Mousterian strata; p.189

probably not
an “economically

significant”
resource.

Gruta da Figueira Marine shells
(Patella sp.) in 31–30K Straus et al., 1993,

Brava, Portugal Mousterian
levels; density, origin, p. 15

and other
constituents unknown.

Anatomically Modern Humans (Homo
sapiens sapiens)

Middle Stone Age
use of shellfish, sea ∼130–55K Singer and

Klasies River

Mouth, South mammals, and
flightless birds. Wymer, 1982

Africa

Middle Stone Age
shell midden with ∼130–>40K Klein, 1999,

Boegoeberg II,

South Africa numerous
cormorant bones. p. 455; Klein

et al., 1999b

Abdur, Eritrea Middle Stone Age
shell midden? 125K Walter et al., 2000

Early Middle
Stone Age shell midden ∼120–80K Brink and Deacon,

Herolds Bay Cave,

South Africa with mussels
(Perna perna), other 1982

shellfish,
and otter remains

associated
with hearths.

(Continued)

308
Erlandson

Table I. (Continued)

Description
of aquatic fauna

Locality/site and
associations Age (yr) Reference

∼90–75K Brooks et al.,

Katanda 9 and 16, Thousands of fish
bones associated

Semliki River, with MSA barbed
bone harpoon 1995; Yellen

Zaire points in
riverine setting. et al., 1995

∼75–55K Marean et al.,

Die Kelders 1, Sea mammals,
birds, and shellfish

South Africa remains
abundant in MSA cave 2000; Tankard

deposits;
shellfish remains are and Schweitzer,

poorly
preserved. 1974

∼70–60K Volman, 1978

Hoodjies Punt, Open air MSA site
with evidence for

South Africa shellfish, sea
mammals, and fish.

∼70–60K Volman, 1978

Sea Harvest, South Open air MSA site
with evidence for

Africa the use of
shellfish, sea mammals,

and fish.

∼60–50K Henshilwood and

Blombos Cave, MSA shell midden
strata, with variable

or >100K

South Africa densities of
marine shell (mussels, Sealy, 1997;

limpets, etc.),
fish remains, and personal

formal bone
tools. communication,

2000

∼50–15K Johnston et al.,

Willandra Lakes, Abundant shellfish
and fish remains

Australia from numerous
lakeside camps, 1998

associated with
terrestrial fauna and

mixed economy.

Ksar ‘Akil, Numerous
freshwater and marine 43–22K Altena, 1962;

Lebanon shellfish
fragments in Early Upper Ewing, 1947;

Paleolithic
strata; pelican, swan, Kersten, 1991

goose(?), and
duck also found.

∼40–15K Arambourg, 1967;

Mugharet el ‘Aliya, Marine shellfish
and fish remains in

Morocco undated Upper
Paleolithic strata. Howe, 1967

New Britain, Several early
sites containing shell 36–15K Allen et al.,

Melanesia middens, fish
bones, etc.; several 1989a,b

substantial sea
voyages required for

colonization of
archipelago.

Riparo Mochi, Early Aurignacian
stratum produced 35–32K Kuhn and Stiner,

Liguria, Italy almost 5000
pieces of marine food 1998; Stiner,

shell (MNI ca.
500), plus 240 shell 1999

ornaments made
from 43 taxa.

∼35K

Castonet Shelter, Greenland seal
(Phoca hispida) bones Cleyet-Merle and

France in early
Aurignacian stratum. Madelaine, 1995

Mandu Mandu Low density
midden with shellfish, 34–20K Bowdler, 1990;

Rockshelter, crab, and fish
remains at pericoastal Morse, 1988

Western site ca. 5 km
from coast during early

Australia occupation.

Leang Burung, Abundant
freshwater shellfish remains 31–19K Glover, 1981

Sulawesi in cave site.

Gorham’s Cave, Numerous marine
shellfish remains in 30–25K Waechter, 1964;

Gibraltar Early Upper
Paleolithic levels; some Zeuner and

sea bird, seal,
and fish remains. Sutcliffe, 1964

Kilu Rockshelter, Shell midden with
fish bones and other 29–20K Wickler and

Solomon Islands, fauna;
colonization of island Spriggs, 1988

Melanesia required
several substantial voyages

by maritime
peoples.

The Archaeology of Aquatic Adaptations
309

Table I. (Continued)

Description of aquatic fauna

Locality/site
and associations Age (yr) Reference

Shuwikhat-1, Catfish and
large mammal remains at 25K Vermeersch and

Upper Egypt fishing and
hunting station. Van Peer, 1988

Ishango 11 and 14, Abundant fish
remains and some 25–16K Brooks et al.,

Semliki River, shellfish,
crab remains with barbed 1995; Yellen

Zaire bone points
in early LSA et al., 1995

assemblages
in riverine and

lacustrine
setting.

Site 1017 (Khor Khormusan
campsite produced 22.7K Greenwood, 1968,

Musa), Egyptian numerous
catfish bones as part of p. 100

Nubia mixed
economy.

Ohalo II, Jordan Thousands of
fish bones associated 21–18K Nadel and Werker,

Valley, Israel with house
floor on south shore of 1999

Sea of
Galilee.

La Riera, Asturias, Upper
Paleolithic cave strata with 21–14K Straus et al., 1981

Spain shellfish,
fish, and rare seal remains.

Ballana (Site Halfan
campsite with large quantities 19–18K Greenwood, 1968,

8859), Egyptian of burned
bone, mostly freshwater p. 108;

Nubia fish
(catfish, etc.). Wendorf, 1968,

p. 797

∼18–17K Straus, 1976–1977

Altamira Cave, Solutrean use
of shellfish and seal

Santander, Spain within a
predominantly terrestrial

site
economy.

∼17K

Balmori Cave, Upper
Paleolithic conchero containing Clark, 1974–1975

Asturias, Spain hundreds of
marine shells, mostly

limpets.

Coberizas Cave, Shellfish
remains and occasional fish 17–15K Clark and

Spain bones in
Upper Paleolithic strata. Cartledge, 1973

Small numbers
(n = 44) of salmon

Cueva Ambrosio,
16.5K L´ pez, 1988

o

Almeria, Spain vertebrae
and several hundred marine

shell
fragments—ornamental and

nonornamental—in Solutrean levels

of cave ca.
60 km from modern coast.

Note. M = million years; K = thousand
years.

have produced possible evidence for the
use of aquatic foods, other resources, or

maritime activities. These lists, too,
are illustrative rather than comprehensive—I

have compiled such data for years but
still frequently encounter sites with ap-

parent evidence for aquatic resource
use that I was unaware of. Because of the

proliferation of such sites, in fact, I
have limited myself to Old World localities

more than 15,000 years old and New
World sites more than 8,000 years old. There

is no question that aquatic resources
were systematically used in these areas af-

ter these times, and the different
thresholds for the Old and New Worlds also

help compensate for the fact that the
two areas were first colonized by humans at

very different times. Even so, early
aquatic sites are too numerous to discuss or

list individually. Instead, I first
discuss the evidence for the use of aquatic foods

310
Erlandson

Table II. Some Early New World
Localities With Evidence for Aquatic Resource Use

Description of aquatic fauna and

14 C

Locality/site
association age (Kyr) References

Monte Verde, Chile Pericoastal
site with evidence for 12.5? Dillehay, 1997

coastal
contact (seaweeds, etc.).

Broken Mammoth, Abundant
waterfowl remains, some 11.6–9.6 Yesner, 1996

Alaska fish,
otter, and beaver in mixed

economy.

Tule Lake, Fish and
waterfowl as a primary 11.4 Beaton, 1991

California resource in
basal layers of SIS-218

rockshelter.

Lewisville, Texas Several Clovis
hearths associated with 11 Storey et al., 1990

freshwater
shellfish, turtles, and fish

remains
within diversified economy.

Quebrada Jaguay, Faunal
assemblage dominated by fish, 11.1–9.9 Sandweiss et
al.,

Peru shellfish,
and seabird remains. 1998

Pedra Pintada Freshwater fish,
shellfish remains in 11.3–10 Roosevelt et al.,

Cave, Brazil several
Paleoindian occupation 1996

levels.

Marmes Use of
freshwater mussels and salmon 11–10 Caulk, 1988

Rockshelter, along with
terrestrial resources.

Washington

Healy Lake, Alaska Possible
freshwater fish use. 10.9 Borden, 1979

Quebrada Seabird, fish,
and shellfish use. 10.8–10.5 Keefer et al., 1998

Tacahuay, Peru

Kanaka Rapids site, Isotopic
signature of Buhl woman 10.7 Carlson, 1998;

Idaho skeleton
suggests marine (salmon?) Green et al.,

component
in Paleoindian diet. 1998

Ring site, Peru Basal levels of
multicomponent shell 10.5 Richardson, 1998;

dated to
terminal Pleistocene. Sandweiss et al.,

1989

Dalton Complex,
fish as a primary meat 10.5–9.9 Goodyear, 1982

Rodgers Shelter,

Missouri source.

10.4–7.8 Erlandson et al.,

Daisy Cave, San Abalones,
mussels, turbans, and other

1996; Rick et al.,

Miguel Island, shells in
island Paleoindian site;

in press

California Early
Holocene component rich in

shellfish,
fish, and pinniped remains,

with shell
beads, fish gorges, etc.

49-PET-408, Human skeletal
remains with strongly 9.2 Dixon, 1999,

southeast Alaska marine
dietary signature. pp. 180–181

Island
occupation and probable 9–8 Davis, 1989

Hidden Falls,

southeast Alaska maritime
economy—faunal remains

poorly
preserved.

9.6 Dunbar, 1997

Cutler Ridge, Shell midden
with tuna and shark

Florida remains,
located adjacent to narrow

continental
shelf.

9–8 Erlandson, 1994;

California coast Numerous Early
Holocene shell

Erlandson and

middens on
islands and mainland

with
diversity of maritime Moss, 1996

adaptations.

Sabine River site, Submerged Gulf
Coast shell midden 8.5 Dunbar, 1997

Texas with burned
and unburned fish bone.

Chuck Lake II, Island shell
midden with abundant fish 8.2 Ackerman et al.,

southeast Alaska remains
1985

Note. Kyr = thousand years.

The Archaeology of Aquatic Adaptations
311

Table III. Islands
Colonized or Explored by Pleistocene Seafarers

Locality
Description of evidence Date (Kyr) References

Flores Southeast Possible evidence
for Homo erectus 800? Morwood et al.,

Asia crossing of
initial water gap from 1998; Sondaar

Sunda to
Flores. et al., 1994

New Guinea and Oldest sites in
Sunda are the earliest 60–40 Clark, 1991; Groube

Australia evidence for
planned maritime et al., 1986;

voyaging,
involving several sea Roberts et al.,

crossings up
to 90 km long. 1990

∼50

Crete, Greece Homo sapiens
sapiens remains with Facchini and

poorly
documented context; Giusberti, 1992

calcareous
breccia in which bones

were cemented
dated by Pa/U to

51,000 ±
12,000 BP; colonization of

Crete
apparently required several short

sea crossings.

Bismarck Shell middens,
fishing, and seafaring at 35 Allen et al.,

Archipelago, several sites
dated from 15–35 Kyr, 1989a,b; Wickler

Melanesia with voyages
up to 140 km long. and Spriggs, 1988

Sicily, Italy Aurignacian
assemblage from 30 Chilardi et al., 1996

Mediterranean
Island involving short

voyage.

Ryukyu Islands, Human skeletal
remains found in 32–15 Matsu’ura, 1996

Japan Yamashita-cho
and other caves on

Okinawa and
other islands; involves

voyages of ca.
75–150 km.

Kozushima Island, Upper Paleolithic
peoples on Honshu 25–20 Oda, 1990, p. 64

Japan crossing 50 km
wide channel to obtain

obsidian.

Melos Island, Travel across ca.
24 km of open water to 13 Cherry, 1990

Greece obtain
obsidian for mainland trade.

Admiralty Islands, Settlement of
Manus Island required 12 Allen and Kershaw,

Melanesia 200 km voyage.
1996

Cyprus Occupation of
Aetokremnos site, 10.3 Cherry, 1990, p. 151

Akrotiri
Peninsula on southwest coast

of Cyprus.

Channel Islands, Boat and marine
resource use by coastal 11–10 Erlandson et al.,

California Paleoindian
groups, with sea crossings 1996; Johnson

of at least 10
km. et al., 2000; Orr,

1968

Southeast Alaska Presence on
islands indicates a maritime 10–9 Davis, 1989; Fedje

and British lifestyle and
seafaring capabilities. and Christensen,

Columbia
1999

Note. Kyr = thousand years.

during various stages of human
evolution, examining several key sites along the

way. After reviewing such “direct”
evidence for aquatic subsistence, I show that

questions often remain about the
cultural origin of the aquatic (and other) fau-

nal remains found in such sites.
Finally, I discuss some other lines of evidence

for early aquatic adaptations,
including early seafaring and maritime adaptations,

sites submerged on continental shelves
around the world, and the significance

312
Erlandson

of pericoastal sites that indicate some
use of coastal or other aquatic habitats or

resources.

Old
World Localities

For the Lower Paleolithic,
relatively little is known about hominid subsis-

tence. Preservation problems are
especially serious for sites of such antiquity, and

taphonomic issues related to the origin
of faunal remains and their association with

evidence for hominid activity are
paramount. The earliest evidence for the possible

use of aquatic resources by hominids
comes from East African Rift Valley localities

where the remains of a variety of
aquatic or amphibious fauna have been found with

stone tools between about 2.5 and 1.7
million years old (e.g., Auffenberg, 1981;

Greenwood and Todd, 1970; Harris et
al., 1990; Leakey, 1971, 1994; Meylan,

1990; Stewart, 1994). Probably left
primarily by Homo habilis, the contents of

these lacustrine sites record the
scavenging and foraging activities of early ho-

minids, as well as the background noise
of natural accumulation processes. Most

researchers today believe the remains
of large land mammals found at such sites

were accumulated primarily via
scavenging of animals killed by more efficient

predators or other natural causes.
Fernandez-Jalvo et al. (1999) have suggested,

however, that some of the small mammals
represented at such sites may have been

hunted by hominids. Several early Rift
Valley sites have also produced the bones of

aquatic or amphibious animals,
including hippos, crocodiles, fish, frogs, shellfish,

etc. (Leakey, 1971). Because many of
these sites formed in dynamic lakeshore

settings, however, any clear
association of aquatic (and terrestrial) fauna with ho-

minid activities is difficult to
demonstrate. At some sites, the remains of fish appear

to be closely associated with hominid
artifacts, but in others fish bones are rela-

tively abundant in both cultural and
natural strata. Greenwood and Todd (1970,

p. 240) and Stewart (1994) have argued,
however, that fish (especially the cat-

fish, Clarias sp.) would have been
relatively easy to procure in some Rift Valley

aquatic settings and are unlikely to
have been ignored by early hominids. This

seems logical, especially for hominids
living in lakeshore settings with economies

based on opportunistic scavenging and
foraging.

For Homo erectus, a series of
East African sites has produced similar asso-

ciations of artifacts and aquatic or
amphibious fauna. At Olduvai, Leakey (1994)

reported that the bones of catfish and
hippos are ubiquitous in artifact-bearing

sediments dated between about 1.1 and
0.4 million years ago, and the remains

of crocodiles, aquatic turtles, and
shellfish also are found in some sites. As was

the case with much of the Olduvai
fauna, Leakey (1994, p. 142) recognized the

difficulty in determining whether
these aquatic taxa were deposited by hominids,

but she argued that a cultural origin
for the catfish was most likely given their

fragmentary condition and close
association with artifacts (see also Auffenberg,

1981; Roe, 1994; Stewart, 1994). In Bed
III at Olduvai, dated between about 1.1

The Archaeology of Aquatic Adaptations
313

and 0.8 million years ago, Leakey
(1994; see also Roe, 1994) also found a series

of distinctive pits and furrows
possibly associated with the evaporative production

of salt by Homo erectus.

Along the coastlines of Africa
and the Middle East, there is also relatively

widespread evidence for Lower
Paleolithic occupation (e.g., Bar-Yosef, 1994,

p. 214; Howe, 1967; Wulsin, 1941). Most
of these localities are poorly dated,

however, and contain choppers, hand
axes, and other stone tools found in raised

interglacial beach deposits. Although
many of these clearly document the oc-

cupation of coastal plains, the precise
age and environmental context (coastal,

pericoastal, inland?) of such
occupations is not clear.

What appears to be relatively
unambiguous use of aquatic resources by

Homo erectus in Southeast Asia comes
from the site of Kao Pah Nam, a lime-

stone cave in northern Thailand
occupied about 700,000 years ago (Pope, 1989).

According to Fagan (1990, p. 120),
“considerable numbers” of freshwater oyster

shells were found piled against the
cave wall. In the same level, stone tools, a

cobble-ringed hearth, and the remains
of hippo, ox, deer, porcupine, and rat were

found.

Evidence for aquatic resource use
increases somewhat with the appearance

of archaic Homo sapiens after about
400,000 years ago. It is not clear, however,

whether this increase represents real
behavioral or environmental shifts or the

better preservation and greater
visibility of more recent occupations. At the Lower

Paleolithic site of Hoxne in England,
Clactonian artifacts and faunal remains have

been found in what have been
interpreted as lakeshore and alluvial deposits (Singer

et al., 1993). Although the dating of
the Hoxne occupations is still somewhat

tentative, much of the Clactonian
occupation appears to have occurred during an

interglacial period between about
350,000 and 300,000 years ago. The associated

fauna are dominated by large land
mammals (especially horse and deer), but include

numerous specimens of freshwater fish
(pike, roach, stickleback, etc.) and beaver,

and smaller numbers of otter and
waterfowl (Stuart et al., 1993). The cultural

origin of the aquatic and other faunal
remains, like those from the Olduvai sites,

has not been firmly established, but
Stuart et al. (1993, p. 198) noted

that the distributions of all of
the beaver Castor fiber and extinct beaver Trogontherium

cuvieri material . . . and most of
the fish material . . . follow the same broad distribution

pattern as the larger bones,
stones, and artifacts. This suggests that the remains of these taxa

also might be food remains
accumulated by man. . . .

Also in England, excavations at the
Lower Paleolithic site of Clacton-on-Sea pro-

duced fish and freshwater mussel
remains (Singer et al., 1973), although the dating

of the site (ca. 425,000 years (Singer
et al., 1993, p. 219) or ca. 250,000 (Gamble,

1986, p. 140)) and the cultural origin
of the aquatic fauna remain uncertain.

About 300,000 years ago, archaic
Homo sapiens also occupied Terra Amata

along the Mediterranean coast of France
(de Lumley, 1969; Villa, 1983). Mussels

314
Erlandson

and other marine shells were found at
Terra Amata, but their context and quan-

tity are poorly documented. Other early
Old World evidence for shellfish use

comes from several North African Middle
Paleolithic or Aterian sites like Haua

Fteah in Libya (Klein and Scott, 1986;
McBurney, 1967), Mugharet el’Aliya in

Morocco (Howe, 1967), and several sites
in Morocco and Algeria (D´ b´ neth and

ee

Sbihi-Alaoui, 1979; Roche and Texier,
1976; Roubet, 1969; Souville, 1973). In

southern Europe, Mousterian use of
shellfish is suggested by assemblages from

Monte Circeo (Stiner, 1994) and
Grimaldi caves (Stiner, 1999) in Italy, Ramandils

in France (Cleyet-Merle and Madelaine,
1995), and Devil’s Tower Rockshelter

(Garrod et al., 1928), Gorham’s Cave
(Waechter, 1964), and Vanguard Cave

(Barton et al., 1999) in Gibraltar. At
the Italian cave of Grotta Moscerini, ma-

rine shells with flaked edges suggest
that shell tools were used by Neandertals

between about 60,000 and 80,000 years
ago (Stiner, 1994, pp. 187–188). For

Neandertals and other archaic Homo
sapiens living in coastal areas, there is little

evidence for the exploitation of fish
(but see Cleyet-Merle, 1990; Cleyet-Merle and

Madelaine, 1995), and the exceptional
cases may represent scavenging from the

beach. Pinniped bones also are rare in
Middle Paleolithic sites and may represent

scavenging of stranded animals or
carcasses. Nonetheless, there is little doubt that

archaic Homo sapiens occupying the
Mediterranean littoral actively foraged for

shellfish and other intertidal
resources (Stiner, 1994, p. 216).

With the appearance of
anatomically modern humans (Homo sapiens sapiens

or AMH), beginning about 125,000 years
ago, Old World evidence for the use of

aquatic resources increases
dramatically. This disparity is even more pronounced

if the Aterian sites of northwest
Africa are considered to be associated with early

or nearly modern Homo sapiens sapiens
groups (see Klein, 1999). The earliest

evidence for such associations may come
from a recently reported locality near

Abdur in Eritrea along the Red Sea
coast, where what are described as Middle

Stone Age (MSA) stone tools were found
with the remains of marine shells and

other aquatic fauna in strata dated to
about 125,000 year ago (Stringer, 2000;

Walter et al., 2000). With the
information currently available, however, it is not

clear whether the stone artifacts were
left by anatomically modern humans or if

the faunal remains represent the food
refuse of hominids. More secure and better-

documented associations and evidence
come from a series of MSA coastal sites

in South Africa dating between about
120,000 and 50,000 years ago, including

Klasies River Mouth caves (Deacon and
Deacon, 1999, pp. 102–106; Singer and

Wymer, 1982), Die Kelders cave (Marean
et al., 2000; Tankard and Schweitzer,

1974), the Sea Harvest and Hoodjies
Punt sites near Saldanha Bay (Volman, 1978),

Herolds Bay Cave (Brink and Deacon,
1982), the Boegoeberg 2 rockshelter (Klein,

et al., 1999b), and Blombos Cave
(Henshilwood and Sealy, 1997). At these sites,

the earliest evidence for relatively
diversified coastal (or mixed) economies is

found, including the relatively
intensive use of shellfish, pinnipeds and cetaceans,

and flightless seabirds (i.e.,
penguins). Fish remains are virtually absent from these

coastal MSA localities (Klein and
Cruz-Uribe, 2000), except for Blombos Cave

The Archaeology of Aquatic Adaptations
315

where a significant number of large
fish bones have been found in MSA shell

midden strata associated with bone and
stone projectile points (Henshilwood and

Sealy, 1997). Initially estimated to be
between about 50,000 and 60,000 years

old, sediments capping the MSA levels
at Blombos Cave have now been dated

via thermoluminescence (TL) to
approximately 100,000 years ago (Vogel et al.,

1999). The dearth of fish in most
South African sites led Klein (1995, 1998) and

Klein and Cruz-Uribe (2000) to suggest
that fishing may have been beyond the

intellectual or technological
capabilities of early anatomically modern humans. It

is possible, however, that the higher
technological costs of marine fishing generally

discouraged such activities, just as
fishing seems to have been limited at most sites

along the California coast during the
early Holocene (Erlandson, 1994, but see

Rick et al., in press). Along with the
Blombos Cave fish remains, support for this

latter idea comes from the carefully
made barbed bone harpoon points found with

the remains of numerous large
freshwater fish at two MSA sites at Katanda on the

Semliki River in Zaire (Brooks et al.,
1995; Yellen, 1998; Yellen et al., 1995). Dated

to about 80,000 years ago, the Katanda
harpoons represent the earliest evidence for

complex composite fishing technologies
in the world and add to the evidence for a

significant expansion of aquatic
resource use among anatomically modern humans.

Similar barbed bone points also
have been found associated with numerous

fish bones at the Late Stone Age site
of Ishango 14 on Lake Rutingaze (Edward)

in Zaire, in strata dated to about
20,000 radiocarbon years before present (RYBP)

(Yellen, 1998). Fish bone is relatively
abundant at some other Late Pleistocene

African sites, including the White
Paintings rockshelter in Botswana (Stewart,

1994; Yellen, 1998) where the lower
levels are tentatively dated to ca. 20,000

years ago, and a series of Nile River
sites dated between about 40,000 and 15,000

RYBP. In coastal areas, little is known
about aquatic resource use during this time

period because sea levels were deeply
depressed during the Last Glacial and most

African coastlines were far removed
from sites now located along the modern

shore (see van Andel, 1989).

This same interval in southwest
Asia and Europe also is problematic due to

lowering sea levels and extensive
glaciation. Numerous Upper Paleolithic sites in

southern and southwest Europe have
produced evidence for shellfish collection

and consumption. Shellfish densities
increase in many of these sites near the end

of the Pleistocene (e.g., Straus, 1990;
Straus et al., 1980, 1981), but it is not

clear if this represents an
intensification of shellfishing in response to population

growth, increased sedentism, changes in
marine or estuarine environments, or

a combination of such processes (see
Bailey, 1983a,b; Clark and Straus, 1983;

Straus and Clark, 1983). Numerous
interior or pericoastal Upper Paleolithic sites

in Europe and southwest Asia also have
produced beads or other ornaments made

from marine shells or artistic
depictions of aquatic animals (Cleyet-Merle and

Madelaine, 1995; Clottes and Courtin,
1996; White, 1993). The presence of sizable

numbers of marine shell ornaments, in
some sites obtained from both Atlantic and

Mediterranean coastlines more than 100
km distant, suggests that interior people

316
Erlandson

traveled to the coast seasonally or
actively traded with peoples living along these

coasts.

In southern Asia, there is only
limited evidence for aquatic resource use

from this time period. In Indonesia, a
freshwater shell midden known as Leang

Burung attests to the systematic
exploitation of shellfish as much as 31,000 RYBP

(Glover, 1981). At Longrien, a long
Upper Paleolithic sequence contains very

limited evidence for aquatic resource
use, but produced a few bivalves from a

layer dated to about 30,000 RYBP.
Despite the current dearth of evidence, there

can be little doubt that maritime or
other aquatic peoples lived in Southeast Asia

since at least 50,000 years ago.

The peopling of Australia and New
Guinea testifies to this, since migrating

from Southeast Asia to Sahul would have
required several substantial sea crossings

even during periods of much lower sea
level (Clark, 1991). Not surprisingly, early

evidence for the use of freshwater fish
and shellfish comes from Australia, which

now appears to have been settled by
maritime peoples between about 50,000

(Roberts et al., 1990) and 60,000 years
ago (Thorne et al., 1999). Numerous

freshwater shell middens from the
Willandra Lakes area of southeast Australia have

been radiocarbon dated between 38,000
and 15,000 RYBP; Thorne et al. (1999)

recently argued that some of these
lacustrine occupations may date to as much

as 60,000 years ago. Although evidence
for intensive marine resource use in late

Pleistocene Australia is lacking,
several sites from western Australia have produced

limited amounts of marine shell from
strata dated between about 20,000 and 36,000

RYBP (e.g., Bowdler, 1990; Morse, 1988;
O’Connor, 1989; Veth, 1993). At Mandu

Mandu Creek rockshelter, located only
about 4–5 km from the coast just prior to

the Last Glacial, a low-density midden
deposit includes the remains of shellfish,

crab, fish, and terrestrial fauna
(Bowdler, 1990; Morse, 1988). These sites could

be interpreted as evidence for limited
Pleistocene use of marine resources, but sea

level and shoreline reconstructions
show a strong correlation between the presence

and density of marine resources and the
variable distance of each site from the sea.

The apparent abandonment of most of the
sites during the height of the last glacial,

and the fact that they were reoccupied
when sea levels again approached modern

levels, can be interpreted as evidence
that the lateral migration of coastal habitats

strongly influenced local settlement
and subsistence patterns. Several saltwater

shell middens located on the Melanesian
islands of New Ireland, New Britain,

and the Solomons—islands that
required additional sea voyages of 80–100 km to

reach—have been dated between about
35,000 and 15,000 RYBP (Allen et al.,

1989a,b; Wickler and Spriggs, 1988).
The aquatic focus of these early Melanesian

occupations is attested to not just by
the seafaring required to settle the islands,

but also by the abundance of marine
shellfish and fish remains found in the site

deposits. The presence of such sites in
western Melanesia, in contrast to Australia

and New Guinea, is due to the steep
local geography, where the bathymetry plunges

rapidly into deep water and changes in
sea level have had relatively limited effects

on the local shorelines and the coastal
archaeological record.

The Archaeology of Aquatic Adaptations
317

New
World Localities

In the New World, most early
evidence for human use of marine resources

comes from the Pacific Coast, where
relatively steep bathymetry also has limited

the lateral displacement of postglacial
shorelines (Erlandson, in press; Richardson,

1998). The earliest sites currently
come from South America. In Chile, the con-

troversial pericoastal site of Monte
Verde has been dated to ca. 12,500 RYBP and

reportedly contains evidence for
coastal foraging, including four types of seaweed

(Dillehay, 1997). At the coastal site
of Querero, which has produced a suite of

dates between about 11,600 and 10,900
RYBP, marine shellfish, sea lion, and

whale remains were all found associated
with those of mastodon, deer, and other

land mammals (Nu˜ ez et al., 1994). At
Quebrada de las Conchas on Chile’s north

n

coast, Llagostera (1979) also
documented the existence of a diversified maritime

economy including the use of a variety
of shellfish and fish between about 9700

and 9400 RYBP.

Along the south coast of Peru,
Sandweiss et al. (1998) reported an early

component from Quebrada Jaguay, where
shellfish, fish, and sea bird remains

have been found in strata dated between
about 11,100 and 9,900 RYBP. The

faunal remains at Quebrada Jaguay
suggest an almost exclusive reliance on marine

animals, but the presence of obsidian
from a distant interior source suggests that the

site may be just one aspect of a
seasonal round that included interior sites as well

(see Richardson, 1998). Also located on
the southern Peruvian coast, and nearly

as old (10,800–10,500 RYBP), is
Quebrada Tacahuay, where the faunal remains

from the earliest occupation are
dominated by sea bird (cormorant, booby, and

pelican) and fish (anchoveta, anchovy)
bones, with a few shellfish (clam, mussel)

remains (Keefer et al., 1998). Of the
3,775 faunal elements recovered from the

basal stratum at Quebrada Tacahuay,
only eight (0.2%) were from terrestrial taxa.

A third site on the south coast, the
Ring site, contains a shell midden that first may

have been occupied as early as 10,600
RYBP (Sandweiss et al., 1989). Along the

north coast of Peru, Richardson (1998)
has described several ephemeral camps

of the Amotape complex, where unifacial
tools have been found associated with

the remains of mangrove shellfish
(Anadara tuberculosa) dated to about 11,200,

10,000, 9200, and 9000 RYBP (see also
Llagostera, 1992). In Ecuador, coastal

shell middens of the Las Vegas complex
are now dated as early as 10,800–10,100

RYBP (Richardson, 1998; Stothert,
1985).

The meticulous work of Roosevelt
et al. (1996) on Paleoindian components

at Pedra Pintada cave in Brazil dated
to ca. 11,000 RYBP also shows that freshwa-

ter fish were an important component
of an early Amazonian economy that was

relatively eclectic and focused on
smaller plant and animal resources.

Along the Pacific Coast of North
America, the earliest and best-documented

maritime sites currently come primarily
from California. On San Miguel Island off

the California coast, Daisy Cave
contains a thin dark soil containing a few chipped

stone artifacts and a low-density shell
midden containing abalone, mussel, turban,

318
Erlandson

and other shellfish remains dated to
about 10,400 RYBP (Erlandson et al., 1996).

That humans were on California’s
Channel Islands by the end of the Pleistocene

has long been suggested by Orr’s 14 C
dating of the Arlington ”Man” (probably a

woman) skeleton to ca. 10,000 RYBP
(Orr, 1968). Recent redating of this skeleton

suggests that Arlington Woman actually
may have died closer to 11,000 RYBP

(Johnson et al., 2000), but a precise
date has yet to be established. Since the

Channel Islands have been separated
from the California mainland throughout the

Pleistocene, these two sites
demonstrate that Paleoindian peoples had seaworthy

boats during the terminal Pleistocene
and leave little doubt about their maritime

capabilities. Along the California
coast, there are also dozens of shell middens

dated between about 9,700 and 8,000
RYBP (Erlandson, 1994; Erlandson and

Moss, 1996; Jones, 1991). One of the
best examples comes from Daisy Cave,

where stratified shell midden deposits
dated between about 9,700 and 7,800 RYBP

contain abundant shellfish and fish
remains, smaller numbers of pinniped and sea

bird remains, numerous bone fishing
gorges and shell beads, and woven artifacts

made from sea grass (Connolly et al.,
1995; Erlandson et al., 1996).

Along the coastlines of northern
California, Oregon, and Washington, there

are only two shell middens reliably
dated to about 8,000 RYBP, Duncan’s Landing

Rockshelter on the northern California
coast and the Indian Sands site on the Ore-

gon coast (Erlandson, 1994; Erlandson
and Moss, 1996; Lightfoot, 1993; Moss

and Erlandson, 1995). The dearth of
early sites in this intermediate area of the

Pacific Coast now appears to be
related to a long history of occasional massive

subsidence earthquakes along the
Cascadia Subduction Zone, tectonic events com-

monly associated with tsunamis and
severe marine erosion (Erlandson et al., 1998;

Minor and Grant, 1996). In British
Columbia and southern Alaska, a number of

early coastal sites dated between about
8,000 and 10,000 RYBP have been docu-

mented (Carlson, 1998; Erlandson and
Moss, 1996; Fedje and Christensen, 1999;

Moss, 1998), including portions of a
human skeleton found in a cave known as 49-

PET-408 (On-Your-Knees Cave) on Prince
of Wales Island dated to approximately

9,200 RYBP (Dixon, 1999, p. 118). The
isotopic composition of this skeleton is

consistent with a diet comprised almost
entirely of marine foods. A bone tool

manufactured from a land mammal rib
found in another part of the same cave

has been dated to about 10,300 RYBP
(Dixon, 1999, p. 181), suggesting that the

site may have been occupied even
earlier. This terminal Pleistocene date is similar

to the estimated age of a basalt flake
recovered from the surface of a paleodelta

deposit located on the continental
shelf off the Queen Charlotte Islands of British

Columbia (Fedje and Christensen, 1999),
although these early dates should be

regarded as very preliminary.

Adjacent to the generally broader
and shallower continental shelves of the

Gulf of Mexico and Atlantic coasts,
early coastal archaeological sites are much less

common. On the Louisiana coast, where
shorelines of the Mississippi delta have

been prograding for millennia, Gagliano
(1970) reported estuarine shell associated

The Archaeology of Aquatic Adaptations
319

with an 11,000-year-old archaeological
site at Avery Island. Off the Gulf Coast,

near the intersection of a creek and
the submerged channel of the Sabine River by

the Louisiana and Texas border, Dunbar
(1997) noted the presence of a submerged

shell midden dated to about 8,500 RYBP
that contains both burned and unburned

fish bone.

Along the Atlantic Coast of North
America, shell middens dating earlier than

about 8,000 years are extremely rare.
Along most of the Florida coast, for in-

stance, the Clovis-age shoreline is
believed to have been between about 50 and

150 km offshore (see Dunbar et al.,
1992, p. 125). Consequently, postglacial shore-

line changes have been dramatic in most
areas, Florida coast shell middens more

than about 5,000 years old are highly
unusual, and a number of submerged shell

middens have been found. One exception
to this pattern is the Cutler Ridge site,

located adjacent to a narrow stretch of
continental shelf near Miami, where lateral

shoreline changes associated with
postglacial sea level rise have been minimal.

This important site, dated to as much
as 9,600 RYBP but largely unpublished,

reportedly has produced the remains of
a variety of marine fish (tuna, shark, etc.)

and shellfish (Dunbar, 1997).

Although interior Paleoindian
groups are often portrayed as relatively special-

ized big-game hunters, there is
evidence for the use of aquatic resources at a number

of early sites. These include the
Broken Mammoth site in south-central Alaska,

where two well-stratified terminal
Pleistocene components have been identified,

one dating between about 11,800 and
11,000 RYBP and another between about

10,300 and 9,600 RYBP (Yesner, 1996).
Faunal remains are well preserved in these

early components. Identifiable
elements from the older component are dominated

(>60%) by aquatic birds (swan,
geese, and ducks), but also include some large

and small land mammals (wapiti, bison,
etc.). The younger of these components

is dominated by the remains of large
ungulates, but also contains about 30% small

mammals, 10% waterfowl, and smaller
numbers of salmonid, beaver, and otter

remains. Another Paleoindian component
containing the remains of aquatic fauna

is the Lewisville Clovis site located
along the Trinity River in north-central Texas,

where archaeological deposits
associated with numerous hearths yielded a diverse

array of plant and animal remains. Of
the 16 hearths excavated, 9 contained the

remains of freshwater mussel and snail
shells—many of them burned. Also recov-

ered were the remains of box turtle,
fish, amphibians, prairie dog, rabbit, tortoise,

egg shells, raccoon, snake, etc. (Story
et al., 1990). At the Horn Shelter 2 site lo-

cated along the Brazos River,
Clovis-age deposits also yielded the remains of land

turtles and a few fish remains. A
younger component at the site, dated between

about 10,000 and 9,500 RYBP, also
produced a diverse faunal assemblage, includ-

ing the remains of many freshwater
mussels and fish such as drum, gar, and catfish,

along with a double human burial
associated with numerous beads made from the

marine shells Oliva sayana and Neritina
reclivita (Story et al., 1990, pp. 203–204).

Another early North American site is a
small rockshelter (CA-SIS-218) located

320
Erlandson

on the shore of Tule Lake in northern
California, where Beaton (1991) identified

a hearth dated to 11,450 ± 340
RYBP associated with charcoal, ash, and fish, wa-

terfowl, and mammal bones. At Marmes
Rockshelter in Washington, freshwater

mussels and salmon bones are reported
from deposits dated between about 10,000

and 11,000 RYBP.

Shellfish
Feeders and Carrion Eaters

Globally, the growing number of
early sites known to contain the remains of

shellfish, fish, sea birds, sea
mammals, and other aquatic fauna may indicate that

aquatic resources were used relatively
early in human history, by Homo habilis,

H. erectus, and H. sapiens. Due to a
variety of questions about the context, taphon-

omy, recovery, and interpretation of
many ancient faunal assemblages, however,

it is difficult to evaluate how
significant aquatic resources were in early hominid

economies. Moreover, lists like those
presented here suffer from another problem

that must be addressed before we can
conclude that even incidental use of aquatic

resources was both early and
widespread. This problem is the possible role various

animals and other noncultural processes
may have played in the accumulation of

aquatic animal remains in early sites
(see Butler, 1993; Erlandson and Moss, in

press). Although recent taphonomic
studies show that a wide range of scavengers

and predators transport bones into
caves and other sites, few have considered the

possibility that animals and not humans
may have transported marine shells, fish

bones, or sea mammal remains into early
coastal sites.

After visiting several
Paleolithic cave sites in Gibraltar in the mid-1980s, I

did not initially question whether the
remains of marine shellfish and fish found

in the site deposits (other than a
clearly defined Last Interglacial beach) could

have been deposited by anything other
than humans. In her detailed taphonomic

analysis of faunal remains from the
Monte Circeo caves in Italy, Stiner (1994)

considered a host of possible sources
for animal bones but considered only cul-

tural mechanisms for the accumulation
of shellfish remains. Recent research has

shown, however, that a wide range of
predators and scavengers—bears, hyenas,

coyotes, badgers, cats, and a variety
of birds—transport the remains of aquatic

vertebrates (seals, fish, birds, etc.)
and invertebrates (shellfish, etc.) to terrestrial

landforms (e.g., Erlandson and Moss, in
press; Jones and Allen, 1978; Klein et al.,

1999b). Caves and rockshelters, in
particular, provide shelter for a wide variety

of mammals and birds that hunt or
scavenge in aquatic habitats and may deposit

carcasses or skeletal remains at site
locations where they can be mixed with fau-

nal remains left by hominids.
Archaeologists, therefore, must carefully evaluate

the nature of terrestrial and aquatic
faunal remains found in both cave and open

sites to determine whether the activity
of nonhuman predators or scavengers has

contributed significantly to the
faunal remains present at a site.

Unfortunately, such careful
evaluations have rarely been done, and it is either

difficult or impossible to evaluate
the cultural origin of the aquatic fauna in many of

The Archaeology of Aquatic Adaptations
321

the sites listed in Tables I and II.
With a more critical eye toward the origin of aquatic

remains in early sites, the evidence
for early aquatic resource use at some key

localities may need to be reassessed.
Gorham’s Cave produced hundreds of marine

shells, for instance, but my
observations suggest that these were widely scattered

in the cave deposits. Gorham’s Cave
also produced the remains of a wide variety

of birds, including seagulls and others
known to feed on and transport shellfish

(Erlandson and Moss, in press). Without
further evidence to link these aquatic fauna

to cultural activities, we cannot be
certain how significant aquatic foods were to the

Neandertal and Upper Paleolithic cave
occupants (but see Barton et al., 1999). At

present, similar questions can be
raised about virtually all of the Lower Paleolithic

sites listed above, as well as the
Mugharet el’Aliya in Morocco where monk seal,

fish, and shellfish remains were
found in Paleolithic layers (Arambourg, 1967;

Howe, 1967). In the New World, similar
questions have been raised about some

Pleistocene or Early Holocene “shell
middens” located on California’s northern

Channel Islands (e.g., Erlandson, 1994,
pp. 183, 196; Erlandson and Morris, 1992;

Erlandson and Moss, in press).

For other sites, the Middle Stone
Age middens of South Africa promi-

nent among them (but see Klein et al.,
1999a,b), the evidence linking hominids

with aquatic resource use seems much
more secure. In the Mousterian levels at

Devil’s Tower, for instance, Garrod
et al. (1928, p. 42) described “thick layers” and

a “large heap” of shells associated
with hearths. At Grotta Moscerini in Italy, Stiner

(1994, pp. 181–184) found that a
significant percentage of the marine shells was

burned, suggesting that they too were
deliberately collected by Neandertals. Other

Old World examples include many of the
freshwater shell middens of Willandra

Lakes in Australia and the Pleistocene
middens of Melanesia (Allen et al., 1989a,b;

Wickler and Spriggs, 1988). In the New
World, there seems to be little question

about the predominantly cultural origin
of the aquatic fauna found at most of the

open air middens along the Pacific
Coast. Early components at Daisy Cave, Broken

Mammoth, and Lewisville also seem
relatively secure.

The Distribution
of Early Coastal Localities

Even allowing for such
uncertainties about the origin of aquatic faunal

remains—often even more serious for
the remains of terrestrial fauna found at

early sites—a significant number of
Paleolithic localities with secure evidence for

systematic early aquatic resource use
are now relatively well documented. The

spatial and temporal distribution of
these early sites, particularly the coastal ex-

amples, is of special interest. Yesner
(1987, 1998, p. 205) suggested, for instance,

that such sites are exceptional and are
located in areas of upwelling and unusu-

ally high marine productivity. Thus
such early coastal sites are often viewed as

rare examples of relatively intensive
aquatic resource use in a Pleistocene world

otherwise dominated by terrestrial
economies.

322
Erlandson

This association holds for some
early coastal localities, but it does not explain

the evidence for early marine resource
use at several early Mediterranean sites

in Italy, Lebanon, Libya, and Algeria
(see Klein and Scott, 1986; McBurney,

1967; Stiner, 1994), where marine
productivity is comparatively low by global

standards. My own comparison of the
distribution of early coastal sites leads to a

different conclusion. While a number of
early sites are found in areas of intense

upwelling (Peru, California, Gibraltar,
etc.), many others are not. Comparing the

distribution of coastal sites to
various physical and biological characteristics in

an atlas of the world’s oceans
(Couper, 1989), I found no clear correlation with

intensive marine upwelling, exceptional
primary (phytoplankton) or secondary

(zooplankton) productivity, sea
temperature, salinity, latitude, tidal range, tectonics

or vulcanism, marine habitat, or
terrestrial habitat. In fact, relatively early sites are

found in areas of coral reefs,
temperate seas, and even arctic or subarctic coasts

(by 8,000–10,000 years ago). They are
found adjacent to tundra environments,

boreal forests, savanna, chaparral, and
hyperarid landscapes, including some where

contemporary interior sites contain
relatively abundant remains of large terrestrial

game.

I found only one trait that seems
to link the early coastal localities: steep

bathymetry. From California to Florida
and from Melanesia to the Mediterranean,

all the early sites are located along
relatively steep shorelines where the offshore

topography drops off rapidly. The
opposite also holds true, with areas of broad and

shallow continental shelves generally
producing only relatively recent evidence

for marine resource use, regardless of
the intensity of marine upwelling. This

is due to the simple fact, clearly
demonstrated by several elegant studies (e.g.,

Parkington, 1981; Shackleton et al.,
1984), that most localities situated along

modern coastlines were far removed from
coastal habitats during most of the last

250,000 years and more. Studies of
historical foragers in coastal habitats have

shown that the skeletal remains of
edible aquatic animals are rarely transported to

residential sites more than about 10 km
from the coast (Bigalke, 1973; Meehan,

1982), except for those that have
ornamental or other utilitarian value. Where

shorelines are steep, however, sites
still preserved above sea level may sometimes

be found within the foraging radius of
ancient coastal habitats. The occupants of

sites located along shallow continental
shelves, on the other hand, may only have

had access to marine resources for the
last 5,000–8,000 years, as local sea levels

and shorelines approached the modern
condition.

This general bathymetric
correlation (which I call Richardson’s Rule)—in

which steep shorelines are associated
with relatively early evidence for marine

resource use, while shallow shelves
yield relatively recent evidence—is a much

stronger predictor of the location of
early coastal sites than upwelling or any of the

other aquatic or terrestrial traits I
examined. Furthermore, Richardson’s Rule helps

explain some puzzling anomalies. It
explains, for instance, why early coastal sites

are much more common along the
generally steep Pacific Coast versus the relatively

The Archaeology of Aquatic Adaptations
323

shallow Atlantic Coast of the United
States. It explains why along the Peruvian

coast, all of which is characterized by
upwelling and high marine productivity,

the earliest coastal sites are
differentially distributed in areas of relatively steep

bathymetry (Richardson, 1998). Finally,
it helps explain why along most of the

Florida coast, where beaches were as
much as 100 km offshore about 14,000 years

ago, the modern shoreline has produced
evidence for maritime adaptations no more

than about 5,000 years old, except for
the steeply plunging shorelines near Miami

where the Cutler Ridge site contains
evidence for marine fishing and other coastal

foraging dated to about 9,600 RYBP.

The correlation between steep
bathymetry and the location of early coastal

sites also seems to contradict two
tenets of traditional theories about maritime

adaptations, (1) that steep bathymetry,
which generally limits the extent of inter-

tidal and nearshore habitats, reduces
the productivity of such marine environments

and renders them relatively
unattractive to humans; and (2) widespread maritime

adaptations only developed in the
Holocene after sea level stabilization led to the

development of relatively broad,
shallow, and productive nearshore habitats. My

analysis of the distribution of early
coastal localities suggests that many coastal

habitats are more productive than
previously envisioned, that Pleistocene mar-

itime adaptations were more widespread
than previously thought, and that the

archaeological record for the antiquity
of coastal adaptations is fundamentally

biased in most parts of the world.

The
Antiquity of Seafaring

Further support for this
viewpoint comes from recent evidence for a relatively

early development of seafaring in
several parts of the world, including evidence

for Pleistocene maritime voyaging in
areas where the oldest coastal shell middens

date to the Holocene. For decades, the
idea that our Pleistocene ancestors may

have made substantial migrations by
boat suffered from the same theories that

marginalized maritime adaptations in
general and argued that our ancestors were

relatively unsophisticated
technologically. There is little question that hominids

must have crossed rivers and other
short water barriers in spreading out of Africa

and through Eurasia. Prior to 1980,
however, there was virtual unanimity that boats

were a very recent addition to human
technologies (e.g., Bass, 1972; Greenhill,

1976; Johnstone, 1980, p. xv). Due to
preservation problems, evidence for the

earliest use of boats—as opposed to
simple logs or floats that allowed hominids

to cross small water barriers while
partly submerged—remains lost in obscurity,

depending primarily on indirect
evidence for the colonization of island groups

(Table III). Except for the
long-distance voyaging evident among Austronesian

and other peoples in the last 5,000
years or so, such evidence requires the presence

of not-too-distant islands that have
been separated from continental land masses

in recent geological times, criteria
many regions of the world cannot meet.

324
Erlandson

Archaeologists have long argued
inconclusively both for and against the idea

that Homo erectus was capable of making
the relatively short crossing (<20 km)

of the Straits of Gibraltar from
Morocco to the Iberian Peninsula (see Cachel

and Harris, 1998; Rolland, 1998). As
evidence accumulates for a relatively long

isolation of Neandertals in western
Europe (e.g., Krings et al., 1997), however,

it seems increasingly unlikely that
archaic Homo sapiens had the capability to

routinely cross the potentially
hazardous Straits of Gibraltar. Elsewhere in the

Mediterranean, there is limited but
more convincing evidence for occasional island

exploration by Neandertals (Cherry,
1990). For Southeast Asia, recent evidence

may indicate that Homo erectus reached
the Indonesian island of Flores as much as

700,000–800,000 years ago (Morwood et
al., 1998, 1999; Sondaar et al., 1994), and

Bednarik (1998) and Bednarik et al.
(1999) proposed that relatively sophisticated

seafaring and maritime adaptations date
back a million years or more. So far,

however, there is little evidence for
any systematic use of seaworthy watercraft

by Homo erectus or archaic Homo
sapiens, and their voyaging capabilities appear

more likely to have been relatively
rudimentary.

Evidence for Pleistocene
seafaring by anatomically modern humans is much

more compelling and more widespread,
involving the dispersal of hominids across

a number of unequivocal and substantial
water barriers (Clark, 1991; Erlandson,

in press; Irwin, 1992). Evidence for
systematic and sophisticated Pleistocene voy-

aging comes primarily from eastern
Asia, Australia, and Melanesia, where voy-

ages in excess of 20–200 km have now
been widely documented between at least

50,000 and 15,000 years ago. The proof
that seafaring extended well back into

the Pleistocene requires a fundamental
paradigm shift, not yet fully realized, since

maritime voyaging was once thought to
be strictly a Holocene phenomenon. By

the 1970s, terminal Pleistocene
seafaring had been documented by the presence

of obsidian from the Mediterranean
island of Melos in strata at Franchthi Cave in

mainland Greece dated to as early as
13,000 RYBP (Cherry, 1990). The antiquity

of seafaring was extended with the
discovery that humans had reached Australia

by 20,000 years ago (Lampert, 1971), a
date rapidly pushed back to 33,000 years

ago (Bowler et al., 1970), 40,000 years
ago (Groube et al., 1986), and now as much

as 50,000–60,000 years ago (Roberts
et al., 1990; Thorne et al., 1999). Regardless

of the route chosen, colonization of
New Guinea and Australia required several

separate sea crossings, including
voyages of at least 80 km (Clark, 1991; Irwin,

1992). As a result, the colonization of
Australia is now widely viewed as the ear-

liest evidence for planned maritime
voyaging in human history and possibly some

of the earliest evidence for
anatomically modern human behavior anywhere in the

world (Davidson and Noble, 1992).

For a time, two puzzling facts
allowed some scholars to believe the Pleistocene

colonization of Australia may have been
accomplished by accident. First, in historic

times Australian Aborigines reportedly
had no sophisticated watercraft capable of

making substantial sea crossings
(Flood, 1990, p. 36), which raised questions

about their ability to travel through
island Southeast Asia by boat. Like much of

The Archaeology of Aquatic Adaptations
325

the rest of the world, Australia also
had no true coastal shell middens or other

direct evidence for maritime
adaptations dating to the Pleistocene. In fact, the vast

majority of such sites were less than
about 5,000–6,000 years old.

With the discovery in the late
1980s of several Pleistocene shell middens in the

Bismarck Archipelago and the Solomon
Islands in western Melanesia (see Allen

et al., 1989a,b; Wickler and Spriggs,
1988), any doubts about the role of deliberate

maritime voyaging in the peopling of
Australia essentially vanished. Settlement

of these islands, now dated to at least
35,000 RYBP (Allen and Kershaw, 1996,

p. 185), added several significant
maritime crossings to those already required to

reach Australia and New Guinea. More
importantly, these islands contain rela-

tively impoverished terrestrial flora
and fauna, and the sites themselves contained

the marine shellfish, fish, and other
remains expected of a maritime people. The

Melanesian evidence also suggests that
maritime voyaging capabilities improved

significantly between about 35,000 and
15,000 years ago. While the initial settle-

ment of New Guinea, New Britain, and
New Ireland required voyages of up to

100 km, colonization of Buka in the
Solomon Islands at least 28,000 years ago re-

quired a minimum sea voyage of 140 km
and possibly 175 km (Irwin, 1992, p. 20).

By 15,000 years ago, moreover,
Melanesian seafarers had reached Manus Island in

the Admiralty group, which required an
uninterrupted voyage of 200–220 km, 60–

90 km of which would have been
completely out of sight of land (Irwin, 1992, p. 21).

Further evidence for Pleistocene
seafaring comes from the islands of Japan.

Japan itself was connected to the Asian
mainland during periods of very low sea

level, so its settlement did not
necessarily require boats. Fagan (1990, p. 191) ar-

gued, however, that new blade and
edge-grinding technologies introduced about

30,000 years ago when Japan was
separated from the mainland probably involved

maritime contacts. This idea may be
supported by the discovery of human bones

beneath a charcoal-rich stratum in
Yamashita-cho Cave on Okinawa dated to

about 32,100 RYBP (Matsu’ura, 1996,
p. 186). Human remains dated between

about 15,000 and 26,000 RYBP also have
been found in several other limestone

caves on Okinawa and the smaller
islands of the Ryukyu chain (Matsu’ura, 1996),

which stretches southward from Japan
nearly to Taiwan. At Pinza-abu Cave on

Miyako Island, human remains found
below a calcareous flowstone were associ-

ated with charcoal dated to about
26,000 RYBP (Matsu’ura, 1996, p. 187). Given

the bathymetry of the Ryukyu Islands,
several sea voyages would have been re-

quired to reach Okinawa from Japan,
including a crossing about 75 km long.

Reaching Miyako Island, from either
Japan or Taiwan, would have required even

longer voyages of up to 150 km. In
Japan itself, archaeological evidence suggests

that by about 21,000 RYBP, maritime
peoples from Honshu were using boats to ob-

tain obsidian from Kozushima Island
approximately 50 km offshore (Oda, 1990).

Similar to Australia, despite
considerable evidence for Pleistocene seafaring, the

oldest shell middens in Japan date to
the Holocene (Aikens and Akazawa, 1996,

p. 224; see also Aikens and Higuchi,
1982). It seems likely, therefore, that earlier

coastal sites have been submerged by
rising sea levels.

326
Erlandson

The evidence for Pleistocene
seafaring in Japan is also significant because

it places competent mariners in the
cool waters and boreal climates of the North

Pacific at a date early enough to have
contributed to the initial colonization of the

Americas (Engelbrecht and Seyfert,
1995; Erlandson, 1994, in press). From Japan,

the Kurile Islands stretch to the
northeast like stepping stones to the Kamchatka

Peninsula and the southern shores of
Beringia. With the now well-documented

Pleistocene seafaring capabilities of
Homo sapiens sapiens, the presence of Pleis-

tocene seafarers in Japan, and the
geography of the North Pacific, maritime peoples

appear to have had the capabilities to
follow a coastal pathway to the Americas.

Whether they made such a journey is
still unknown, and the evidence that could

resolve the issue—like so many
questions related to the evolution of maritime

adaptations—lies largely unstudied
and submerged on the continental shelves of

the North Pacific.

Other Evidence for
Early Aquatic Adaptations

Two other sources of data need to
be considered in any comprehensive eval-

uation of the antiquity of aquatic
adaptations: the archaeological record from sub-

merged or “drowned” terrestrial
sites, and the nature of pericoastal sites that show

limited evidence for marine or
estuarine resource use. Although a detailed exam-

ination of either topic is beyond the
scope of this paper, it would be a mistake not

to consider such evidence at all.

Submerged
Terrestrial Sites

Scholars have long known that
human occupation sites lie submerged on con-

tinental shelves around the world (see
Negris, 1904; Smyth, 1854) and that these

may fundamentally bias our
understanding of the development of aquatic adapta-

tions (e.g., Emery and Edwards, 1966;
Flemming, 1998, p. 130; Kraft et al., 1983;

Richardson, 1981; Shepard, 1964). What
to do with such knowledge, however,

raises fundamental problems. We must be
governed by some rules of evidence,

after all, and simply assuming that
ancient shell middens lie submerged off coast-

lines around the world and that coastal
adaptations have always been a part of

the human story leaves many scholars
with a very uncomfortable feeling. On the

other hand, assuming that the
archaeological record is representative in the face of

clear evidence to the contrary is
equally problematic. The obvious solution is to

examine submerged coastal landscapes
for the presence or absence of submerged

shell middens or other evidence for
early aquatic resource use.

Unfortunately, this is not as
easy as it sounds, and such programs have been

limited so far. During marine
transgressions, the submergence of most terrestrial

sites would have been accompanied by
their essential destruction. Just as it is

destroying countless coastal sites
around the world today, wave erosion would

The Archaeology of Aquatic Adaptations
327

have redeposited most older sites into
the intertidal zone, leaving only lag deposits

on wave-cut platforms such as the Lower
Paleolithic localities documented on Old

World marine terraces. Along the
predominantly erosional outer coasts around

the world, intact submerged sites would
be preserved only in special situations

where local landforms are protected by
offshore islands, archaeological deposits

are cemented or sealed under
erosion-resistant strata, or earthquakes caused a rapid

subsidence of sites into the intertidal
or subtidal zone. In estuarine or lacustrine

settings, where wave energy and
shoreline erosion are generally less severe, the

potential for the preservation of
submerged sites is considerably better (Flemming,

1983, 1998). Even in these settings,
however, relatively little archaeological work

has been accomplished on submerged
terrestrial sites (but see Fischer, 1995b;

Masters and Flemming, 1983). Due to
technological and financial constraints,

moreover, the work that has been done
has been limited primarily to sites found

comparatively close to shore and in
relatively shallow water. Because sea levels

have risen up to 150 m in the past
17,000 years, these limitations have so far

prevented effective undersea
reconnaissance along shorelines more than about

10,000–12,000 years old.

Nonetheless, impressive numbers
of submerged coastal sites have been found,

and the number of sites is rapidly
growing (see Fischer, 1995b; Flemming, 1998).

In a recent summary, Flemming (1998, p.
129) noted that roughly 550 submerged

human occupation sites (SHOS) have been
located in coastal settings around the

world, about 100 of which are older
than 3,000 years. These include Acheulian

hand axes found in sediments underlying
a historical shipwreck 8 m below sea level

off the South African coast and
thousands of Mousterian artifacts eroding from a

creek bank submerged 18 m below sea
level near Cherbourg on the Atlantic coast

of France (Flemming, 1998, pp.
135–136). Elsewhere, numerous submerged sites

dating to the middle and early Holocene
are now known, and artifacts of Pleistocene

age also have been recovered from the
sea floor (e.g., Dunbar et al., 1992; Faught

et al., 1992; Fedje and Christenson,
1999; Flemming, 1983, 1998; Sanger, 1995;

Stright, 1990). In a number of
protected coastal areas, submerged sites that contain

evidence for aquatic adaptations have
been found. Some of these submerged sites

feature remarkable preservation, as
shown by the intact structural remains, human

burials, canoes, canoe paddles,
hearths, and other materials recovered from early

Holocene sites such as Tybrind Vig in
Denmark (Andersen, 1985, 1987) and Newe

Yam off the coast of Israel (Raban,
1983; Wreschner, 1983).

One of the earliest submerged
sites, and surely one of the most remarkable,

is the Upper Paleolithic Cosquer Cave
discovered in 1991 on the Mediterranean

coast of France (Clottes et al., 1992;
Clottes and Courtin, 1996). Cosquer Cave is a

partly submerged limestone cavern, the
small mouth of which lies 37 m below sea

level. A narrow and gradually rising
shaft extends for approximately 140 m before

opening into a large cavern, only parts
of which remain above sea level. Over 250

engraved or painted motifs have been
documented in the unflooded remnants of the

cavern, and radiocarbon dates indicate
that these were executed primarily during

328
Erlandson

two periods about 27,000 and 18,500
RYBP (Clottes and Courtin, 1996). Upper

Paleolithic artistic representations of
marine animals are rare in Europe, but at

Cosquer they make up about 12% of the
animal images and include depictions of

seals, the great auk (Pinguinus
impennis), and possibly fish and jellyfish. Although

the cave appears to have been located
more than 10 km from the sea during the

height of the Last Glacial, these
images testify to the significance of marine animals

to the artists. As Clottes and Courtin
(1996, pp. 44–45) noted, several Upper

Paleolithic skeletons excavated in the
late 1800s from the Grimaldi Caves about

150 km to the east were associated with
hundreds of marine shell ornaments, also

testifying to the symbolic importance
of the sea among Upper Paleolithic people

of the area.

Clearly, submerged terrestrial
sites do exist and may be preserved under the

right conditions. The questions that
remain are whether such submerged coastal

sites represent the proverbial tip of
the iceberg or isolated cases, whether even

earlier coastal sites lie offshore in
deeper water, and whether such sites can be

found and sampled to help unravel some
of the mysteries that remain about the

role of the sea in human history.

Pericoastal Sites

Like Cosquer Cave and the
Grimaldi Caves, there are scores of Pleistocene

sites around the world that are located
in pericoastal or even interior settings

that lack dense accumulations of
aquatic food remains, but nonetheless testify

to linkages to aquatic habitats. These
include numerous coastal sites listed in

Table I, which at various times in
their occupational histories appear to have been

located some distance from the coast.
In a number of well-dated sites with detailed

paleogeographic reconstructions, these
periods seem to correlate with reduced

densities of marine food remains, the
presence of strictly ornamental or utilitarian

objects (shell beads, baler shells,
etc.), a complete reliance on terrestrial foods, or

site abandonment. For a number of other
sites, less well documented or located

further from aquatic habitats, such
relationships are not as clear. Sites that contain

small amounts of aquatic food remains
can be viewed as evidence that aquatic

resources were relatively unimportant,
as part of a seasonal round that included

residential periods on the coast, of
exchange with more maritime people living

closer to the coast, or some
combination of such inferences. A similar range of

arguments has been made for interior
sites in Upper Paleolithic Europe where

marine shell ornaments are found far
from the coast. Such sites clearly indicate

some economic or ritualized use of
aquatic resources, but their significance can be

argued depending on the theoretical
stance of individual investigators.

In this regard, I find the
Australian case particularly compelling, where sev-

eral pericoastal sites (Mandu Mandu,
Shark Bay, etc.) more than 30,000 years

old contain small numbers of shells,
often ornamental or utilitarian types traded

The Archaeology of Aquatic Adaptations
329

between interior and coastal groups
ethnographically. It is currently impossible to

know for certain whether these shells
are evidence that early terrestrially adapted

Australians occasionally visited the
coast, that coastal people occasionally vis-

ited the interior, or that they
represent the exchange of goods between discrete

groups residing separately in coastal
and interior areas. How can a continent like

Australia, which we now know was
colonized by boat at least 50,000 years ago,

have so few early coastal sites? Did
maritime peoples reach Australia then aban-

don its coastlines in favor of more
productive terrestrial habitats and resources

for tens of thousands of years? If so,
how do we account for the evidence for the

systematic use of shellfish and fish
in the Willandra Lakes area at least 40,000

years ago? Why do we see further
evidence for maritime voyaging into western

Melanesia by 35,000–40,000 years ago?
To me, it seems most likely that the early

pericoastal sites in Australia are the
remnants of inland settlement and subsistence

by coastally adapted Pleistocene
peoples whose main settlements were submerged

or destroyed by rising postglacial
seas.

DISCUSSION

To some, none of the individual
lines of evidence I have examined may provide

a compelling argument for the early
development of aquatic adaptations. When

the theoretical and methodological
issues I have raised are combined with current

knowledge about world sea levels and
coastal paleogeography, as well as a variety

of lines of archaeological evidence, I
believe there are compelling reasons to doubt

the veracity of the current consensus
model. Comparatively speaking, it is still true

that there is only limited evidence for
the intensive use of aquatic resources prior

to the end of the Pleistocene. This is
not surprising in the New World, which

appears to have been colonized near the
end of the Pleistocene when sea levels

were much lower than they are today.
Even on a global scale, however, it is not clear

that this pattern accurately reflects
changes in human subsistence through time.

To understand the history of aquatic
adaptations, globally or in any particular

region, we must first determine if the
patterns evident in the archaeological record

result from actual changes in human
behavior, patterns imposed by geological

or taphonomic forces, or the recovery
and analytical methods of archaeologists

themselves. Unfortunately, except for
areas first occupied by humans relatively

recently (i.e., Polynesia), such
evaluations are fraught with difficulty and have

rarely been done in a manner that
inspires confidence in the results.

A New
World Test Case

It is possible, however, to
return to some of the fundamental tenets of recent

theory about aquatic adaptations and
examine them in light of current archaeolog-

ical data from the New World. If
shellfish and other aquatic foods are generally

330
Erlandson

smaller and less productive than
terrestrial alternatives—and their systematic use

reflects demographic pressure,
resource stress, or economic intensification—then

the antiquity of coastal adaptations in
the Americas should provide an excellent test

case. Traditional theory, including
many recent applications of diet breadth-prey

ranking models, suggests that until
population growth produced sufficient de-

mographic pressure to reduce the
productivity of “high-ranked” terrestrial game,

aquatic resources would not be
systematically used. Thus the relatively recent peo-

pling of two vast continents with
highly diverse and productive terrestrial land-

scapes, especially by small
hunter-gatherer groups generally thought to have mi-

grated through an interior route into
the heart of North America, should show little

evidence for marine resource use or
even coastal settlement until quite recently. If

aquatic resources are not necessarily
marginal, then we should find relatively early

evidence for their use, at least in
areas where the economics of aquatic resource

use compare favorably to those
available in adjacent terrestrial habitats.

To adequately evaluate the
evidence, of course, we need to know when humans

first settled the Americas and whether
the archaeological record is representative

of the full range of early adaptations,
especially in coastal zones. At present, we

cannot answer either question with any
certainty. In 1977, Osborn used a regional

analysis of the Peruvian archaeological
record to argue that marine resources were

inferior in the arid western slopes of
the Andes. When Osborn (1977a) evaluated

the Peruvian evidence, he believed
there was a lag between the earliest interior ver-

sus coastal occupation equal to more
than half of the total cultural sequence for the

region, well over 12,000 years. This
apparent gap was explained by proposing that

the earliest occupants of the region
did not systematically use marine resources until

their population densities had
effectively reached the carrying capacity of the terres-

trial environment. Thus the Andean
coast—one of the richest marine environments

on earth—was characterized as an
environment marginal for human occupation.

In retrospect, we now know
Osborn’s analysis had significant problems (see

Erlandson, 1988; Perlman, 1980; Quilter
and Stocker, 1983; Yesner, 1980). First,

he assumed the Andean archaeological
record was representative, based on the now

disproved claim that tectonic uplift of
the Peruvian landscape had outpaced sea

level rise since the last glacial.
Second, he assumed that the regional demographic

clock began with initial human
occupation of the Andean uplands at least 23,000

years ago, based on claims by MacNeish
(1971) for the presence of “artifacts”

(made from the same rock as the cave
walls) in the lower levels of Pikimachay

(Flea) Cave. Even today, with the
“Clovis barrier” seemingly broken, few scholars

accept the dubious evidence for
pre-Clovis occupation at Flea Cave (e.g., Dixon,

1999, pp. 100–101).

Today, the earliest widely
accepted (there are still doubters) Andean archaeo-

logical site is Monte Verde (Dillehay,
1997), which appears to date to about 12,500

RYBP. Monte Verde is located in a
valley roughly 30 km from the coast of Chile.

While there is evidence that big game
was hunted from the site, there is also ev-

idence for a relatively eclectic
economy in which plants and smaller resources

The Archaeology of Aquatic Adaptations
331

played a significant role. The
presence of seaweed also suggests that the site oc-

cupants had links to the coast. Recent
research also has pushed back the earliest

occupation of the Andean coast to
approximately 11,200–10,500 RYBP (Keefer

et al., 1998; Richardson, 1998;
Sandweiss et al., 1998), and the earliest site, Que-

brada Jaguay, shows interior links
(Sandweiss et al., 1998) that may represent

the opposite end of a seasonal
subsistence round represented at Monte Verde. If

we accept that Monte Verde is one of
the earliest sites in the Andes (which seems

likely) and assume that its occupants
had no coastal neighbors and used few aquatic

resources themselves (which seems less
likely), the age differential between the

earliest coastal versus interior
settlements has withered to less than 2,000 years.

This might be enough time for
terrestrial foragers to shift to a partly littoral or

maritime focus, but it seems highly
unlikely that a shift to marine resources about

11,000 years ago was due to demographic
pressure on the vast Andean landscape.

In North America, the situation
is very similar, with increasing evidence

for early coastal settlement and use of
marine and other aquatic resources. The

presence of people on California’s
Channel Islands 10,500–11,000 years ago

(Erlandson et al., 1996; Johnson et
al., 2000; Orr, 1968), for instance, is extremely

difficult to account for as a response
to human pressure on the highly produc-

tive and diverse terrestrial resources
of the adjacent mainland. The earliest people

of the Channel Islands, moreover, seem
to have subsisted primarily on shellfish,

small fish, plant foods, and
occasional seals or sea lions. There is currently no evi-

dence that they were big-game hunters
drawn to the islands by pygmy mammoths

or massive elephant seals. Rather, they
seem to have been eclectic foragers who

relied on a variety of resources,
including abalones, mussels, small turban snails,

and a variety of small fish. The very
presence of such people on the islands at

this early date suggests that shellfish,
fish, and other marine resources were highly

ranked, highly regarded, and highly
relied on. Along the Atlantic and Gulf coasts

of North America, the evidence for
early coastal adaptations is neither as early nor

as widespread, but in these areas the
presence of submerged sites and Paleoindian

artifacts on the continental shelves
indicates that we are missing early components

of the coastal archaeological record.

To me, the available data suggest
that marine and other aquatic resources

were an integral part of many early New
World economies, that their signifi-

cance has been underemphasized in
previous models, and that their presence in

archaeological sites does not
necessarily indicate the existence of environmental

deterioration, population pressure,
resource stress, or economic intensification. I

am not suggesting—as many have done
for large land mammals and other ter-

restrial resources—that aquatic
resources were universally productive. Nor am I

arguing that the use of aquatic
resources was not, at times, associated with resource

stress and economic intensification.
What I am suggesting is that, in the New World

and the Old World, the factors that
govern human decisions about what resources

will be used, when, and by whom are
highly complex and situational. Recognizing

this complexity, the diversity of
environments (terrestrial and aquatic) encountered

332
Erlandson

by our ancestors, and the flexibility
and opportunism of hunter-gatherers, I believe

global or universal statements about
the productivity of aquatic resources do not do

justice to the diversity and complexity
that we should expect of the archaeological

record.

SUMMARY AND
CONCLUSIONS

I began this paper by suggesting
that general conceptions of the history and

nature of aquatic adaptations have
marginalized the study of coastal, riverine, and

lacustrine societies, relegating them
to the last 1% of human history. The view that

aquatic resources are marginal and that
aquatic adaptations developed relatively

recently renders their study
essentially peripheral to many of the most compelling

issues addressed by archaeologists:
human evolution, early hominid migrations,

the appearance of anatomically modern
humans, peopling of the Americas, the

development of agriculture, the rise of
civilization, and others. I also argued that

a variety of taphonomic processes,
epistemological issues, methodological prob-

lems, and data gaps raise serious
questions about assertions that widespread and

systematic aquatic adaptations
developed only since the end of the last glacial.

Specifically, I suggested that

1. postglacial sea level rise has
submerged most of the shorelines older than

about 10,000 years along which
the evidence for earlier coastal occupa-

tions would logically be
found;

2. differential preservation,
recovery, or reporting of site constituents has

selectively underemphasized
the importance of aquatic resource use in

archaeological sites around
the world;

3. traditional models of
hunter-gatherer behavior have overemphasized the

role of hunting in general,
and large land mammal hunting in particular,

in many ancient societies;

4. normative cultural ecological
reconstructions have too often treated hu-

man societies as aggregations
of generic individuals rather than groups of

diverse people (men, women,
and children; young and old, rich and poor,

etc.) who often were engaged
in different activities;

5. prior to the development of
relatively effective hunting technologies, ho-

minids relied to a significant
extent on scavenging behavior that would have

increased the relative
productivity of and reliance on aquatic resources that

required no specialized
technologies to obtain or process;

6. our hominid ancestors have
always, with rare exceptions dictated by un-

usual environmental
conditions, been highly opportunistic and relatively

eclectic omnivores, an
economic orientation fundamental to our extraor-

dinary success in colonizing
virtually every habitable land and seascape

on earth.

The Archaeology of Aquatic Adaptations
333

With these issues in mind, I
reviewed the evidence for early aquatic resource

use in archaeological sites around the
world, focusing on Old World sites dating

earlier than about 15,000 years ago and
New World sites more than 8,000 years

old. I concluded that a variety of
unresolved problems continue to prevent us

from determining when aquatic
adaptations developed, how widespread they were,

and how important they were in the
broad scheme of human evolution. Some

of the earliest archaeological
localities associated with Homo habilis and Homo

erectus in Africa contain the remains
of aquatic or amphibious animals such as

fish, crocodiles, molluscs, and
hippos, as do some early European sites associated

with late Homo erectus or early archaic
Homo sapiens populations. Although the

distribution of fish and other aquatic
remains in some of these early sites coincides

closely with artifacts and other faunal
remains, the cultural origin of the faunal

remains (aquatic and terrestrial) and
their nature (scavenged or hunted) are difficult

to prove. Less equivocal evidence for
the use of shellfish by Homo erectus and

archaic Homo sapiens also is found at
several Old World sites. There is no doubt

that these hominids occupied coastal
and other aquatic habitats and little reason to

doubt that aquatic resources were used
by them at least occasionally. At present,

the intensity of such use remains
unknown, just as the overall nature of Lower

Paleolithic subsistence remains largely
obscure.

Hominids clearly crossed aquatic
hurdles in spreading out of Africa and

through much of Eurasia, indicating
that rivers and even some straits were not

necessarily the physical or
psychological barriers sometimes imagined. Prior to

the appearance of anatomically modern
humans, however, the use of aquatic re-

sources may have been limited largely
to shellfish and occasional “low-tech” uses

of fish, birds, mammals, and other
resources that could be collected without spe-

cialized technologies. Only with the
appearance of anatomically modern humans

do we find the evidence for a more
intensive use of shellfish and a wider range

of marine or aquatic resources. Not
surprisingly, this economic diversification co-

incides with the first evidence for
the development of a number of “modern” or

transitional technologies, including
the earliest relatively intensive use of chipped

stone blade and geometric or
microlithic industries, the first formal bone tools, the

earliest widespread evidence for the
use of red ochre, and probably the first use of

relatively sophisticated boats. In the
context of such transitional Middle Stone Age

technologies at Klasies River Mouth
caves, Die Kelders, and other Last Interglacial

localities in South Africa, for
instance, AMH appear to have regularly eaten a va-

riety of shellfish, marine mammals,
and flightless birds (Klein and Cruz-Uribe,

2000), although some or all of the
larger vertebrates may have been scavenged

rather than hunted (Binford, 1984).
From the Semliki River area in Zaire comes

the earliest evidence for complex
aquatic hunting gear, ∼80,000-year-old barbed

harpoons from Katanda associated with
numerous large fish remains (Yellen et al.,

1995). From Blombos Cave in South
Africa comes evidence for Middle Stone Age

marine fishing, probably dating to
over 60,000 years (Henshilwood and Sealy,

334
Erlandson

1997). And from the Boegeberg 2 shell
midden comes possible evidence for the

relatively intensive MSA use of
cormorants (Klein, 1999, p. 456).

Dated to about 90,000 years ago
are the earliest skeletal remains of Homo

sapiens sapiens found outside of
Africa—the Qafzeh and Skhul skeletons from

coastal Israel—suggesting that early
modern populations had begun to move out

of Africa by this time. Although
anatomically modern humans do not appear to

have moved into most of Europe for
another 50,000 years (Klein, 1998), current

evidence suggests that they spread into
southern Asia at least 60,000 years ago.

From there, within just a few
millennia, they probably made the multiple maritime

voyages through island Southeast Asia
required to settle Australia and New Guinea.

By about 30,000–35,000 years ago,
seafaring AMH peoples also had colonized

western Melanesia and the Ryukyu
Islands south of Japan.

Although the current state of our
knowledge remains somewhat fluid, the

earliest subsistence strategies that
included relatively eclectic and intensive use

of marine or other aquatic resources
may be associated with the appearance of

anatomically modern humans. When such
aquatic adaptations were combined with

the exploitation of a range of
terrestrial plants and animals, a more diversified and

stable resource base would have
resulted. Such economies may have contributed

significantly to the development of
greater sedentism (see Kelly, 1995, p. 125),

to the reproductive success of Homo
sapiens sapiens, and to our apparently dra-

matic demographic and geographic
expansion over the last 150,000 years. In this

regard, it is worth noting that current
(“Out of Africa”) models for the rapid spread

of anatomically modern humans allow
only about 10% of the time available in

multiregional models for this
demographic and geographic expansion (Erlandson,

in press). Wherever anatomically modern
humans first appear, they seem to have

carried with them a penchant for art
and symbolism, technological innovation, and

complex problem-solving and
communication skills (see Davidson and Noble,

1992; Klein, 1998; Mellars, 1998).
Aquatic adaptations and Pleistocene seafar-

ing played a more significant role
than previously supposed in the demographic

expansion, the geographic spread, and
the phenomenal success of our species.

The available archaeological
evidence contradicts aspects of both Gates of

Hell and Garden of Eden models, with
the most likely scenario falling—like

Aristotle’s “golden mean”—somewhere
in between. Despite a number of cate-

gorical statements to the contrary, we
simply do not know when aquatic resources

were first widely used by our hominid
ancestors or how important they were in hu-

man evolution. However, it makes no
sense that hominids would have completely

ignored aquatic resources for more than
2 million years. As long as scavenging

was a significant hominid pursuit, in
fact, it seems likely that aquatic resources

found in shallow water or on the shore
would have been utilized when reason-

ably abundant and available without
complex technologies. There undoubtedly

has been some intensification of
aquatic resource use during human history, but

it also seems likely that our ancestors
used such resources opportunistically and

situationally whenever and wherever it
made economic sense to do so. If aquatic

The Archaeology of Aquatic Adaptations
335

resources sometimes compare favorably
to terrestrial subsistence alternatives, it

raises significant questions about the
emblematic role shell middens have played

as anthropological indicators of the
broad spectrum revolution and postglacial

economies (see Bailey, 1978). If shell
middens are not diagnostic of postglacial

economies, is the broad spectrum
revolution still a revolution?

We cannot afford to ignore the
fact, however, that the efficient or intensive

exploitation of many types of marine
and other aquatic (and terrestrial) resources

requires relatively complex
technologies (e.g., sophisticated boats, nets, harpoons,

hook-and-line) that currently appear to
have been beyond the capabilities of ho-

minids other than anatomically modern
humans. We should also recognize that

aquatic habitats are extremely
variable, that they are juxtaposed with equally var-

ied terrestrial habitats, and that
together these offer such a diverse range of envi-

ronments that they defy broad
generalization (Erlandson, 1994, p. 278; Perlman,

1980). Given the nearly endless
diversity in the relative productivity and accessi-

bility of aquatic versus terrestrial
habitats around the world, it seems likely that

the antiquity and intensity of aquatic
adaptations varied widely through both space

and time. Today, rather than searching
for general rules of human behavior in

aquatic settings, we should be working
to overcome the taphonomic problems

that currently inhibit our
interpretations so we can more effectively document the

diversity of aquatic adaptations. More
interesting questions should take the place

of dichotomized debates about whether
the world’s aquatic habitats were Gardens

of Eden or Gates of Hell. Once we
recognize the diversity of aquatic habitats

through space and time, as well as the
almost limitless combinations of mosaic en-

vironments that result from juxtaposing
such aquatic habitats with equally diverse

terrestrial habitats, we can focus on
the complexity of human responses to aquatic

environments that took place as our
ancestors developed increasingly sophisticated

subsistence strategies on both land and
in the water. Given this diversity and the

innumerable adaptive responses possible
under various intellectual, technologi-

cal, demographic, and sociopolitical
circumstances, a search for a global model or

universal laws of aquatic adaptations
is almost certainly fruitless.

As we move into the twenty-first
century, I hope we can transcend the simple

models and polarized arguments that
have often characterized scholarly debates

about the evolution of aquatic
adaptations. Surely, as Claassen (1991, p. 275)

has suggested, “it is time to put to
rest the generic clam,” as well as the generic

fish, sea mammal, or coastal forager.
In the last century or so, archaeologists have

made great strides in understanding the
development of maritime and other aquatic

adaptations in human history. As our
studies continue in the next century, there are

numerous issues yet to be resolved and
numerous productive avenues of inquiry to

be studied. To truly understand the
role of the sea (and aquatic habitats) in human

history, however, a number of issues
need to be resolved.

Perhaps the most pressing are
questions related to the antiquity of seafaring

and the search for ancient sites
located along Pleistocene shorelines beneath the

sea. It is time to extend the search
for submerged terrestrial sites to a wider range

336
Erlandson

of shorelines around the world and into
deeper waters to look for coastal sites

dating to the crucial period between
about 60,000 and 10,000 years ago. Utilizing

new technologies, offshore
archaeological survey might be particularly productive

along steeply dipping shorelines of the
Mediterranean, where sites like Cosquer

Cave and submerged caves off Gibraltar
(Waechter and Flemming, 1962) have

been identified in limestone bedrock
where dripstone formations may have helped

to preserve evidence for early coastal
adaptations despite the problems of rising

sea levels and coastal erosion
(Flemming, 1998). Underwater reconnaissance and

excavation work might also be
particularly fruitful off some of the more protected

shorelines of the Japanese archipelago,
where evidence for the Upper Paleolithic

antecedents of the Jomon peoples may
lie submerged. We need to know not only

where such sites are located and when
they were occupied but also whether aquatic

resources played a significant role in
the lives of the occupants and how widespread

such adaptations were.

Critical to evaluating the idea
that anatomically modern humans may have

moved rapidly out of Africa along the
coastlines of southern Asia into Australia

and beyond is a search for early shell
middens or other sites on land, in ar-

eas associated with Last Interglacial
shorelines of East Africa, southern Asia,

and the islands of Southeast Asia.
Considering the spatial distribution of early

coastal sites elsewhere in the world,
such efforts are most likely to succeed if they

are focused along coastal stretches
characterized by relatively steep bathymetry,

where lateral shoreline movements
associated with sea level fluctuations have been

minimized.

Also needed are renewed
excavations at early sites that have already produced

the remains of aquatic resources, with
more sophisticated excavation and analytical

techniques, including fine screen
recovery, flotation, and more critical evaluation

of the origin of aquatic and other
faunal remains. We need more taphonomic and

actualistic studies to help distinguish
between aquatic animal remains of cultural

versus natural origin, work that will
complement the extensive studies that have

been done for terrestrial fauna in
interior areas.

Ultimately, we need more data
from “aquatic” sites of all ages and in all

regions to better document the
variability in aquatic adaptations through space

and time. In the twenty-first century,
the study of maritime peoples and aquatic

adaptations should focus on documenting
the remarkably diverse role that aquatic

resources played in human history as
hominids and humans spread around the

world, from sea to shining sea.

ACKNOWLEDGMENTS

This paper is dedicated to the
memory of J. G. D. Clark, who years ago

recognized the limits of what we could
reasonably say about the development

of aquatic adaptations. I would also
like to recognize the work of Carl Sauer,

The Archaeology of Aquatic Adaptations
337

Alan Osborn, Geoff Bailey, and David
Yesner, scholars whose provocative

views stimulated tremendous progress in
the study of coastal and aquatic adap-

tations. For freely sharing ideas and
information that contributed to my research,

I thank Virginia Butler, James Dunbar,
Anders Fischer, Leland Gilsen, Michael

Glassow, Nina Jablonski, Antoinette
Jerardino, Alan McCartney, Jerry Moore,

Greg Nelson, John Parkington, Geoffrey
Pope, Douglas Price, Mark Raab, Jim

Richardson, Torben Rick, Anna
Roosevelt, Dan Sandweiss, Judith Sealy, Ren´
e

Vellanoweth, Larry Wilcoxon, John
Yellen, and David Yesner. Jim Richardson,

Stephen Nash, and an anonymous reviewer
provided constructive criticisms that

helped me revise an earlier draft. I am
also grateful to Gary Feinman and Douglas

Price for inviting me to write this
paper, for their editorial comments and as-

sistance, and for their patience. Most
of all, however, I am deeply indebted to

Madonna Moss—my wife, colleague, and
best friend—for sharing so many of

my coastal journeys (physical and
intellectual) over the years and for her detailed

comments on an earlier version of this
paper. Only I am responsible, of course, for

the opinions expressed in this paper.

REFERENCES CITED

Ackerman, R. E., Reid, K. C., Gallison,
J. D., and Roe, M. E. (1985). Archaeology of Heceta Island: A

Survey of 16 Timber Harvest Units
in the Tongass National Forest, Southeastern Alaska, Project

Report No. 3, Center for Northwest
Anthropology, Washington State University, Pullman.

Aikens, C. M., and Akazawa, T. (1996).
The Pleistocene-Holocene transition in Japan and adjacent

northeast Asia: Climate and biotic
change, broad-spectrum diet, pottery, and sedentism. In Straus,

L. G., Eriksen, B. V., Erlandson,
J. M., and Yesner, D. R. (eds.), Humans at the End of the Ice

Age: The Archaeology of the
Pleistocene-Holocene Transition, Plenum Press, New York, pp. 215–

227.

Aikens, C. M., and Higuchi, T. (1982).
Prehistory of Japan, Academic Press, New York.

Allen, J., Gosden, C., Jones, R., and
White, J. P. (1989a). Pleistocene dates for the human occupation

of New Ireland, northern
Melanesia. Nature 331: 707–709.

Allen, J., Gosden, C., and White, J. P.
(1989b). Pleistocene New Ireland. Antiquity 63: 548–561.

Allen, J., and Kershaw, P. (1996). The
Pleistocene-Holocene transition in greater Australia. In Straus,

L. G., Eriksen, B. V., Erlandson,
J. M., and Yesner, D. R. (eds.), Humans at the End of the Ice

Age: The Archaeology of the
Pleistocene-Holocene Transition, Plenum Press, New York, pp. 175–

199.

Altena, C. O. V. R. (1962). Molluscs
and echinoderms from Paleolithic deposits in the rock shelter of

Ksar ‘Akil, Lebanon. Zoologische
Meddelingen (Leiden) 38(5): 87–99.

Anderson, S. H. (1985). Tybrind Vig: A
preliminary report on a submerged Ertebolle settlement on the

west coast of Fyn. Journal of
Danish Archaeology 4: 52–69.

Anderson, S. H. (1987). Mesolithic
dug-outs and paddles from Tybrind Vig, Denmark. Acta Archaeo-

logica 57: 87–106.

Arambourg, C. (1967). Appendix A:
Observations sur le faune des Grottes d’Hercule pres de Tanger,

Maroc. In Howe, B. (ed.), The
Paleolithic of Tangier, Morocco: Excavations at Cape Ashakar,

1939–1947, Bulletin 22, American
School of Prehistoric Research, Peabody Museum, Harvard

University, Cambridge, pp.
181–186.

Auffenberg, W. (1981). The fossil
turtles of Olduvai Gorge, Tanzania. Copeia 3: 509–552.

Baden-Powell, D. F. W. (1964). Gorham’s
Cave, Gibraltar: Report on the climatic equivalent of the

marine mollusca. Bulletin of the
Institute of Archaeology 4: 216–218.

Bailey, G. N. (1975). The role of
shellfish in coastal economies: The results of midden analysis in

Australia. Journal of
Archaeological Science 2: 45–62.

338
Erlandson

Bailey, G. N. (1978). Shell middens as
indicators of post-glacial economies: A territorial perspective.

In Mellars, P. (ed.), The Early
Post-glacial Settlement of Northern Europe, Duckworth, London,

pp. 37–63.

Bailey, G. N. (1983a). Problems of site
formation and the interpretation of spatial and temporal dis-

continuities in the distribution
of coastal middens. In Masters, P. M., and Flemming, N. C. (eds.),

Quaternary Coastlines and Marine
Archaeology: Towards the Prehistory of Land Bridges and

Continental Shelves, Academic
Press, New York, pp. 559–582.

Bailey, G. N. (1983b). Economic changes
in Late Pleistocene Cantabria. In Bailey, G. N. (ed.),

Hunter-Gatherer Economy in
Prehistory: A European Perspective, Cambridge University Press,

Cambridge, pp. 149–165.

Barker, G. W. W. (1974). Prehistoric
territories and economies in central Italy. In Higgs, E. S. (ed.),

Paleoeconomy, Cambridge University
Press, Cambridge, pp. 111–175.

Barton, R. N. E., Currant, A. P.,
Fernandez-Jalvo, Y., Finlayson, J. C., Goldberg, P., MacPhail, R.,

Pettitt, P. B., and Stringer, C.
B. (1999). Gibraltar Neanderthals and results of recent excavations

in Gorham’s, Vanguard and Ibex
caves. Antiquity 73: 13–23.

Bar-Yosef, O. (1994). The Lower
Paleolithic of the Near East. Journal of World Prehistory 8: 211–265.

Bass, G. F. (1972). Introduction. In
Bass, G. F. (ed.), A History of Seafaring, Walker and Company,

New York, pp. 9–10.

Beaton, J. M. (1991). Palaeoindian
occupation greater than 11,000 years B. P. at Tule Lake, Northern