Resource predictability and density model of Rada Dyson-Hudson and Eric Alden Smith with predicted hunter and gatherer responses. This and other models were used to help interpret the Waco Lake sites and link evidence to past human behavior. For the three sites in the study area, archeologists predicted that hunter-gatherer groups would be highly mobile and, depending on resource density, highly dispersed within loosely defined, undefended territories. View larger version of model.
Research Approaches and Models
On a site-by-site basis, the Waco Lake investigations focused on what types of activities took place, what resources were used, what kinds of technologies were employed, and, if possible, the time of year the sites were occupied. There is, however, more to this analysis than just providing a list of activities and traits or descriptions of the tools and resources used. Organized and examined within a theoretical framework, site activities and the resources and technologies used can provide a clearer picture of prehistoric hunter-gatherer behavior. In the section below, we examine some of the research approaches used to link the archeological record—the artifacts, features, and traces in the dirt—with human behavior in the past. Many of these assumptions and theories are based on ethnographic studies of modern-day hunting and gathering groups as well as historical accounts by early North American explorers who witnessed Native Americans living in traditional ways. For students of archeology as well as the public in general, this discussion provides a brief compendium of some of the key studies and ideas by leading theorists in hunter-gatherer archeology. Full citations for works referenced below can be found in the Credits and Sources section.
Hunter-Gatherer Behavior and the Environment
Hunter-gatherer characteristics and behaviors such as mobility, subsistence, social group composition and dynamics, and land tenure vary across time and space. In ethnographic, ethnoarcheological, and archeological literature, the relationship between the environment and hunter-gatherer behaviors and decision making is modeled through the examination of many types of environmental and resource variables.
One particular model by Rada Dyson-Hudson and Eric Alden Smith focuses on the density and predictability of resources and how hunters and gatherers might respond to these variables. As shown in the image above, the model is divided into four predicted hunter-gatherer responses based on the nature of the resources. The model assumes that, when resource density and predictability are low, hunter-gatherer groups will be highly dispersed and mobile across the landscape within territories that are loosely defined and not defended. Where resource predictability is low and density is high, it is expected that hunters and gatherers will be highly mobile but conducive to information-sharing with other groups. In areas of low resource density and high predictability, passively defended territories will develop, but population densities would be low, with groups tending to stay in one area of predictable resources. High resource density and dependability predicts geographically stable territories within which the movement of groups would be restricted and the territory and resources strongly defended. Of course, environmental reality cannot always be easily shoehorned into neat categories, because environments are dynamic. The Grand Prairie environment is no different, but the best characterization of this region may be one of low resource predictability and shifting resource densities.
In many environments, there is a high correspondence between seasonal changes and resource predictability. Regular or expected seasonal temperature fluctuations impose a sense of predictability for resource availability and scheduling. In the Grand Prairie, however, the links between seasonal change and resource predictability seemingly are weak or nonexistent for a couple of reasons. One is the lack of a lengthy cold or winter season. Though seasonal temperature variation is high, with the daily average difference between January and July being 22.7° C (40.9° F), an average year only has 32 days with below-freezing temperatures and a growing season of 253 days. Periods of freezing temperatures are usually short-lived, and overnight freezing temperatures often give way to sunny and warm daytime conditions during the winter months. The lack of snow cover and frozen ground would make certain roots, tubers, and bulbs available year-round, rather than being limited to warmer spring and summer months, as they are in northern temperate climates. The lack of snow cover and relatively warm temperatures also provide access to adequate forage for animals and eliminates the need for migration or hibernation for most warm-blooded species. Even cold-blooded reptilian species such as turtles can be observed out and about on warm, sunny days during the winter months.
Another reason for the low resource predictability in the Grand Prairie is the lack of large topographic features that would host vastly different ecosystems with distinct sets of resources due to differences in elevation—sets of resources that would be available at different times of the year. It is clear that the availability of many resources is not predicated on seasonal changes or distinct ecological zones of different elevations. Many resources are readily available throughout the year, therefore subsistence activities might not vary much throughout the year. That is not to say the region lacks seasonally available resources, but even the densities of these can vary in unpredictable fashions (this is discussed in more detail below). In fact, the densities of most resources in the region can vary throughout the year and from year to year.
The region’s proximity to the semiarid plains and arid deserts to the west and the more humid woodlands and forests to the east makes it particularly vulnerable to oscillating, yet relatively unpredictable, periods of rainfall deficits and surpluses. The annual rainfall pattern, while bimodal with late spring and early fall peaks, is often interrupted by prolonged droughts or short periods of torrential rains that can occur at any time of the year. The Balcones Escarpment to the south and southwest is one the most flash flood–prone regions in the United States. Chaotic rainfall patterns can deliver paralyzing blows or setbacks to many resources, affecting their density and availability. As rainfall patterns become less regular, uncertainty regarding resources and on how foragers must deal with them increases.
Even seasonally available resources such as nut crops (e.g., pecans, walnuts, and acorns), prickly pear fruits, and mesquite beans, which by being seasonally available are generally predictable, can nonetheless vary in production from year to year due to rainfall amounts and distribution. For example, in south Texas, prickly pear tuna and mesquite bean production have been found to vary considerably from year to year due to rainfall amounts during the pre-flowering and flowering periods. Acorn production is highly variable from year to year, even among trees of the same species. Oaks produce low to moderate acorn crops in most years, with abundant yields once every two to five years. Abundant crop yields may be 80 percent higher than those in a low-production year, varying by hundreds of pounds of acorns per acre. This natural cycle of production can be confounded by factors such as erratic rainfall. Poor acorn production can result from extended periods of rain or high humidity during the flowering period, which reduces the tree’s ability to pollinate. Poor acorn production would directly impact hunter-gatherer food supplies, as well as have a devastating effect on deer populations. Annual pecan production is similarly cyclical. Pecan groves do not naturally produce abundant crops every year, but in fact do so every two to four years.
So even those resources that are truly seasonal can be unpredictable in terms of their yields due to natural cycles of production and climatic factors. In short, the density of resources can remain stable, or can vary from high to low and in a largely unpredictable fashion. Given these characteristics of resource density and predictability, Dyson-Hudson and Smith’s model predicts that hunter-gatherer groups in the study area would be highly mobile and, depending on the resource density, highly dispersed within ill-defined territories that are not defended. Another scenario is that they would be conducive to information-sharing within territories that shift due to unpredictable but dense resources.
There is one last factor that we need to consider before examining the linkages between behavior and the archeological record: the cost of acquiring the resources. Central Texas and adjacent regions often have been described in bountiful terms by researchers who provide extensive lists of available food resources, but what has not been addressed is the cost of obtaining these resources. Hence, the Dyson-Hudson and Smith model is coupled here with optimal foraging theory. Optimal foraging theory is centered on the idea that to forage optimally is to be efficient relative to time or energy costs (see Robert Bettinger 1991 and Bruce Winterhalder 1981). As noted by J. R. Krebs (1978), it can be thought of as a set of “decision rules for predators.” Optimal foraging theory is derived from evolutionary ecology—a neo-Darwinian approach that focuses on the interaction of natural selection and ecological variables in the development of a biological organism’s adaptations (for overviews, see G. A, Parker and John Maynard-Smith 1990, and Eric Pianka 1978). Widely used in the biological sciences, evolutionary ecology assumes that natural selection has influenced organisms to behave in ways that enhance fitness, or in terms of hunters and gatherers, cognitive decisions regarding foraging strategies are shaped by “natural selection.” Simply put, favorable decisions or efficient behaviors promote fitness and survival, and it is this basis that supports predictive modeling.
Optimal foraging models predict that selection will favor the best strategy from among a defined set of strategies that are possible in the environmental context of interest. Optimal foraging models can deal with a variety of relationships centered on foraging strategies and diet and produce hypotheses that can be tested. One of the more common models, the diet breadth model (see Bettinger 1991; Robert L. Kelly 1995), is based on the concept of ranking resources based on the returns or benefits provided by the resource relative to the cost of acquiring and processing the resource. Some of the predictions the model makes include: (1) the optimal diet will become specialized if high-ranked resources become abundant and acquisition costs decline; and (2) lower-ranked resources may be added to the diet if higher-ranked resources decline and their acquisition cost increases. Another model, the patch choice model predicts that hunters and gatherers will abandon an area before resources are depleted, since the cost of obtaining dwindling resources is high (see, for example, Karl Butzer 1982; Kelly 1995; Winterhalder 1981). Optimization models also can be applied to hunter-gatherer technologies. Such a model might look at the relationship between the portability versus the durability and versatility of tools (e.g., Steven Kuhn 1994), predicting that the optimal tool design is one that has the greatest potential utility relative to the cost of transporting it.
These models do not replicate reality, nor do they assume that hunters and gatherers will always behave in the most efficient way possible. They do assume, however, that behaviors will tend toward a maximization of efficiency. Optimization models predict an optimal pattern of behavior (or an optimal tool design) that is influenced by selective pressures and that can be tested or compared against those patterns observed in the archeological record, as Kelly has noted. It is the explanatory and predictive powers of optimal foraging theory that make it useful. Scholars can use it to predict how hunters and gatherers might cope with a shifting (in terms of density) and unpredictable resource base and other challenges, and it is these coping mechanisms, or risk-reduction strategies, that can be tested against the archeological record.
Dynamic and unpredictable resources and the cost of obtaining these resources are dealt with through the employment of risk-reduction strategies. Basically, all hunting and gathering societies tend to work to reduce risks or avoid production shortfalls through various technological, social, and cultural mechanisms. Risk-reduction strategies such as group mobility may involve moving a foraging group to a new resource patch when the cost of obtaining resources relative to the benefit increases in the current location. When faced with shortages, a foraging group (or at least some members of the foraging group) might also take advantage of kin or economic ties with neighboring groups that may be experiencing a windfall. Or if experiencing a windfall of their own, the foraging group may invite others to cohabit and share resources to assure reciprocal actions or access to resources in the future. Regardless of the strategy employed, risk-reduction strategies imply that the amount of available food resources is ever changing relative to need or that food is often scarce relative to demand.
For the Bosque River/Lake Waco study area, the Dyson-Hudson and Smith model predicts risk-reduction strategies of high mobility, fluid and undefended territorial boundaries, and information sharing. In linking these risk-reduction strategies to the archeological record, we collapse or translate the predicted behavior of high mobility into two broad categories that the archeological record can address: duration of occupation and use intensity. Since the archeological record of Waco Lake has a limited ability to address aspects of territoriality and information sharing, these issues will not be formally addressed, but a few pieces of tantalizing data will be discussed that provide some insights into these issues. Duration of occupation and use intensity are broad in the sense that they subsume a multitude of behaviors, including aspects of subsistence and technology.
Without question, aspects of length of occupation and use intensity are interrelated, but they each make particular marks on the archeological record, which are examined in greater detail below. However, before we examine the archeological results of the risk-reduction strategies, we should articulate some basic assumptions about hunters and gatherers. These assumptions, which are primarily based on John Yellen’s 1977 study of the !Kung but come from other ethnographic studies as well (eg., Kelly 1995), relate to the composition of hunter-gatherer groups, or bands, and the structure or organization of their campsites. We assume that the most basic socioeconomic unit comprising hunter-gather groups is the nuclear family, who claim group or band membership through kinship or marriage but may also claim membership within other groups and at times may cohabit with them as well. Yellen’s work with the !Kung revealed that nuclear families are fairly unrestricted in their movements between groups. Over a four-year period, he observed a turnover rate of 80 percent in individuals of the Dobe band of the !Kung at the group aggregation camp during the dry season. Families and individuals move relatively freely about in a network of groups that has been referred to as a marriage group. The marriage group is endogamous and provides mates for most of its members. Groups within the marriage group, probably through a small core of siblings or cousins, collectively hold, move about in, and exploit a territory, which is not defended or is passively defended. The size of this territory is fluid, due largely to the comings and goings of individuals and nuclear families, but it is known by all within the marriage group that the core of the territory (e.g., a perennial waterhole in the Kalahari) is held by a particular sibling-cousin group.
We also assume that hunter-gatherer campsites at a basic social level can be divided into private and communal areas, with the private areas essentially representing individual nuclear families. Private areas are seen as containing the family hearth and shelter, and it is the idea of a family hearth area that is of particular importance to the current investigations. Many camp activities are centered around hearths. For the !Kung, most manufacturing and cooking activities take place near the family hearth. Marc Stevenson (1991) has noted that exterior hearths are frequently the loci of social life and activities in hunting and gathering societies, particularly in temperate and warm climates. Prehistorically, patterns of individual family hearths, and hence family activity areas, are interpreted for sites throughout central Texas (see LeRoy Johnson 1994) and the Plains (eg., Michael Quigg 1983), although this pattern clearly is not unique to these areas.
Group Mobility, Duration of Occupation, Use Intensity, and the Archeological Record
The topic of mobility among hunters and gatherers has received a lot of attention in anthropological, ethnoarcheological, and archeological literature. Lewis Binford (1980) discusses hunter-gatherer mobility in terms of a continuum from highly mobile foragers to less mobile logistical collectors, depending on how they acquire resources and the nature of their environment. Not unlike Binford’s forager-collector model is Bettinger’s (1991) traveler-processor model, which takes a more in-depth look at the relationships between population, resources, settlement, and subsistence via optimal foraging theory. Kelly (1983) notes five dimensions of mobility in his study: (1) number of residential moves per year, (2) average distance moved, (3) total distance moved each year, (4) total area used over a course of a year, and (5) the average length of a logistical foray. Like the density and predictability of resources, these dimensions of mobility can vary from year to year, and from circumstance to circumstance. Identifying and examining these components from the standpoint of a static archeological record can be particularly challenging, if not impossible. Fortunately, though, mobility influences the material culture of hunters and gatherers, affecting what material remains enter the archeological record and how they enter it. Tools, weapons, shelters that provide protection from the elements, personal adornment items, and features for cooking, processing, and storage all have to adhere to the demands of mobility. These are tangible pieces of evidence that can be examined by archeologists and that allow inferences about the relative length of occupations, whether movements between sites were short or long distances, and how material culture meets the demands of mobility. Tied to these issues of mobility are length of occupation and use intensity, which alludes to how a site is used, what resources are used, and what technologies are employed while a site is occupied before it is ultimately abandoned.
Several pieces of archeological data can provide insights into the relative length of an occupation of a site, or use intensity, and by inference determine how mobile groups were. One can look at the kinds of resources used and the costs of obtaining those resources. Archeological sites across the central Texas region, particularly during the Archaic period, display a strikingly consistent and similar suite of material remains, with deer remains and burned rock features being two notable examples. Deer are considered a high-ranked resource, providing many calories from meat and marrow, making search, capture, and processing costs relatively low. Burned rock features, particularly basin-shaped hearths and pits, are considered geophyte-processing features based on the recovery of various bulbs, roots, and tubers from feature contexts across the region. Although it is unlikely that these hot rock cooking features were used exclusively for baking geophytes, it is assumed that this was their foremost function, since many geophytes require prolonged periods of cooking to render them edible and palatable. The gathering and processing of geophytes can be viewed as tedious and time-consuming, especially considering the popular notion that plants are low yield–high cost foods (e.g., Kelly 1995:90). However, these gathering and processing costs can be driven down if the processing features and campsite are adjacent to dense patches of geophytes. The fact that geophytes are stationary resources also helps keep acquisition costs down. Deer and geophytes provide a good balance of calories from proteins, fats, and carbohydrates at relatively low cost, so it is reasonable to think of them as resources that are highly ranked and largely available year-round, but susceptible to erratic droughts and floods that affect their densities. These two resources probably were the initial targets of acquisition for hunters and gatherers once a campsite was chosen or resource patch entered. This action resulted in a widespread archeological pattern throughout central Texas; however, it is not to say that lower-ranked resrouces would not be initially taken if the opportunity presented itself.
With few exceptions, resources other than deer and geophytes can be considered lower-ranked resources, and their presence in archeological assemblages would suggest longer occupations. Use of lower-ranked resources would suggest that deer and geophytes were dwindling to the point that the cost of their acquisition was rising. These lower-ranked resources usually consist of small-bodied vertebrates, such as reptiles (e.g., turtles and snakes) and small mammals (e.g., rodents and rabbits) and invertebrates such as mussels. Such creatures were probably trapped, snared, or simply hand collected, rather than hunted. With the dwindling returns and increased costs of geophytes, the acquisition of these lower-ranked resources may have fallen upon the women (or older children) of the group, since men would have continued to hunt high-return large game such as deer, even if it meant traveling greater distances. This notion is also consistent with the ethnographic literature, which is replete with examples of women trapping and collecting small animals. The duration of the occupation can be extended through the use of lower-ranked resources—a phenomenon that should be evident in the archeological record. The number and variety of lower-ranked faunal remains in terms of Number of Identified Specimens (NISP)—and the ratios of lower-ranked fauna to deer or other artiodactyls—should provide insight into whether sites supported short-term occupations.
Firewood was another vital resource for the hunter-gatherer groups of the North Bosque River valley, and the taxa of the downed limb wood collected and used for campfires can be indicative of the duration of a site’s occupation. The species of downed limb wood collected for campfires at the Britton, McMillan, and Higginbotham sites should reflect the arboreal species comprising the North Bosque riparian zone, if the length of stay was relatively short. Longer stays at a camp location may deplete firewood sources in the immediate camp area, driving up the costs of collecting firewood as greater distances from camp are traveled to collect it. The use of firewood from species not present in the riparian zone (e.g., ashe juniper, mesquite, live oak) would indicate lengthier stays as nearby riparian sources were depleted and travel and collection costs for firewood outside the valley were lower than lengthy trips up and down the valley’s riparian zone to collect firewood.
Obviously, occupants of campsites generate waste materials, and longer occupations generate more debris, with all other variables being equal. Refuse has to be managed and living areas maintained to curtail interference with subsequent activities. Yellen notes that long-term occupied camps of the !Kung are better organized largely through waste management, which results in a greater number of dumps or secondary refuse piles. Refuse clearing often produces secondary refuse dumps on the peripheries of camps or intensively or repeatedly occupied activity areas (see Stevenson 1991). In cleaning, smaller objects are often left behind, no matter how thorough or often cleaning occurs, while larger objects get removed. The result is that secondary refuse dumps should consist primarily of larger-sized objects. The content and context of the materials can also aid in determining whether a feature represents a secondary refuse dump. An examination of the number, size, and location of secondary refuse dumps within the site can then provide information on the duration of occupation. The distribution of unmodified debitage size classes across a site can provide similar information. Larger debitage should be concentrated away from activity areas if these areas have been cleaned.
With the onset of site abandonment, refuse should no longer be subject to the same level of cleanup, size sorting, and dispersal as during the earlier periods of occupation. Discarded items, regardless of size, from activities that occur late in an occupation should remain grouped around hearths, whereas large objects used and disposed of earlier would be found in secondary refuse piles away from the hearth and along the periphery of the camp. As previously noted, lower-ranked resources should enter the camp in the later stages of occupation if the occupation is relatively long. Remains of these lower-ranked resources, therefore, should have a limited presence in secondary refuse piles and a more dominant presence around the hearths. Deer remains, particularly larger-sized elements, should be present in greater numbers than remains of low-ranked resources in the secondary refuse dumps if the occupation is lengthy. This assumption is not only consistent with optimal foraging theory but with ethnographic data as well.
Long-term occupations witness more activities, not necessarily in terms of diversity but in terms of actual employment. The more times an activity requiring tools is performed, the more wear and breakage of tools occur. Even though many of these tools are part of a mobile tool kit and are brought into sites with use wear and edges already resharpened, use-broken tools and those with little or no remaining use life should be more common in sites that were occupied longer, because tool use lives tend to exceed the length of occupation at short-term sites. To further examine this assertion, the ratio of unmodified debitage to finished formal chipped stone tools, which provides a measure of the frequency of tool discard, can be calculated. Worn and use-broken chipped stone tools have to be replaced, and unmodified debitage is a meaningful byproduct of this action for determining the relative length of occupation. The amount of unmodified debitage is a good measure of length of occupation because, unlike other chipped stone artifacts (e.g., formal tools), the removal of unmodified debitage far from its locus of production is less probable. Burned densities also may provide information on use intensity.
Yellen offers this obvious general rule: the longer a site is occupied, the greater the probability that any particular activity will occur there. Different activities may require different suites of tools and features, so that not only would longer occupations result in a greater number of tools and features but a greater diversity of tools and features. Assemblage diversity can be measured in terms of richness and evenness. Richness indexes measure the number of classes or categories in a sample or assemblage. The more categories or classes present, the richer it is. A richness index (R = S/sqrt N) derived by E. F. Menhinick (1964) is used in this analysis. In this equation, greater values of R indicate increasing richness, S is the number of classes or categories, and N represents the total number of specimens within the assemblage. Another assemblage aspect, evenness, describes the relative frequencies of specimens within each of the classes or categories across the assemblage, or the degree to which all classes or categories are equally represented. Unfortunately, many measures of evenness have inherent problems that make them unreliable measures of heterogeneity in most cases (e.g., see Peter Bobrowsky and Bruce Ball 1989). Due to these limitations, Daniel Kaufman (1998) uses a measure relating to variance, specifically the coefficient of variation (the standard deviation divided by the mean number of specimens per class) of each assemblage, to describe evenness. In this case, greater evenness is indicated by smaller values of the coefficient of variation. The jackknife technique is then applied to the two measures (richness and evenness) of diversity, because each measure is dependent on sample or assemblage size. This technique involves repeatedly recalculating the statistic of interest (richness or evenness), each time eliminating one of the classes or categories. These calculations produce a series of jackknife estimates, which are used to generate a set of corresponding pseudovalues, the mean of which provides the best estimate for the statistic of interest. The relationship between richness and evenness can be examined in terms of site function, where general campsites are represented by high richness and low evenness values and specialized activity sites are represented by low richness and high evenness values. In this context, we assume the former are represented by longer occupations.
If hunters and gatherers’ knowledge of the environment tells them that resource density is high in a particular area, they may anticipate lengthier stays at sites. This concept of “anticipated mobility”—length of time people expect to occupy a site—was found to be important among the San and Bantu peoples. Susan Kent’s (1991) study showed that anticipated mobility was more influential in determining site structure and the number of m2 per person, a rough index of site investment at a camp, than group size or how long the group actually lived there. All the sites in Kent’s study with an anticipated short length of occupation had a value of less than 33 m2 per person. That works out to site sizes of 825 to 2,475 m2, assuming groups or bands consist of 25–75 people as suggested by Michael Jochim (1976) and Robert L. Kelly (1995). Determining the number of m2 per person, or even the size of the area used by site occupants during a single occupation based on archeological data, is difficult at best, but there are other ways to determine anticipated mobility using archeological information. Since anticipated mobility also influences site structure, the anticipated duration of occupation should be reflected in the proximity of individual households or private family areas to each other. The closer the households, the shorter the anticipated stay. Assuming that small hearth features represent individual family hearths, the distances between hearths might provide insight into whether a group anticipated a short or long stay. Probably more important, the anticipated length of occupation might give us a better understanding of a group’s knowledge of their environment.
On a side note, if, as implied earlier, close proximity to dense patches of geophytes is a criteria for site selection, the anticipated length of stay might be determined by those in the group who would be most knowledgeable about how long the resource might sustain the group. Since anticipated mobility affects the organization of the camp, many of the decisions made regarding the layout of the camp were probably made by women.
Distances moved between campsites can vary greatly. Again this is a dimension of mobility that cannot be measured in the archeological record, but insights into whether movements were relatively short or lengthy might be possible based on tool curation and the distance to stone sources. If opportunities to replace tools are few, meaning immediate access to adequate sources of stone is limited due to distance, tools should show evidence of intense reworking. However, if a source of material is close by, tools may not display much evidence of intense curation and resharpening. In the lower North Bosque River valley in the vicinity of the Britton, McMillan, and Higginbotham sites, small chert gravels can be found on some of the gravel bars, but in limited numbers. Because of their small size and limited numbers, they may have been exploited only for the production of small expedient flake tools. The closest known sources of sizable chert clasts and nodules to the three sites range from ca. 11.5 to 20.0 km away. Unless procurement of lithic raw materials is embedded in or part of the daily foraging strategy, or the foraging radius includes adequate sources of stone, groups (campsites) would have to move closer to the source areas to obtain raw materials. Given the distance to the known sources, a high degree of tool curation in the form of intense resharpening and edge rejuvenation ought to reflect short moves between campsites in the lower North Bosque River valley, since tool replacement opportunities would rarely, if at all, present themselves. However, if tools do not display a high degree of tool curation, it could reflect longer distances moved between camps—movements that would present procurement and tool replacement opportunities. The ratio of formal tools, in this case projectile points and bifaces, that exhibit resharpening to those that do not may be indicative of the distances moved between campsites. The use of a tool form that is highly durable, versatile, and amenable to edge rejuvenation and promoting a high level of curation will negate the need for frequent tool replacement and hence frequent trips to stone sources. This idea is explored in greater detail below, but it would tend to support the notion of short-distance movements in the chert-poor lower North Bosque River valley since such tools designs are often used in areas where sources of stone are limited (see Robert Kelly and Lawrence Todd 1988).
Mobility influences material culture in the sense that material possessions are few due to the cost of transporting materials from site to site. Among the !Kung, most of a nuclear family’s possessions can be carried by a single adult. Transport costs of material possessions can be minimized in several ways. One is through the use of beasts of burden, of which the only example known in North America is the dog. The presence of dogs among hunter-gatherer groups might be indicated by a vertebrate faunal assemblage containing specimens that display carnivore gnaw marks. Surely if dogs were kept, food scraps such as bones would have been fed to them or they would have been opportunistically seized by the animals themselves.
In the absence of dogs, group members would be the only agents of transport for moving possessions between sites. If this is the case, then transported materials would probably be limited to those used to acquire resources, namely chipped stone tools and associated shafts. We assume that high mobility results in tools or tool kits that would provide the greatest potential utility relative to the cost of transporting them. Tool kits of mobile hunter and gatherers should be composed of a limited number of lightweight tools so as to minimize transport costs, while at the same time ensuring that the tools are as durable, maintainable, and multifunctional as possible. Tools should be capable of dealing with a broad and changeable set of actions or needs and designed to last until there is an opportunity to replace them. Durability and multifunctionality require increased overall tool size, but this benefit is outweighed by increased transportation costs. Smaller-sized tools, while less costly to transport, tend to have potentially detrimental implications because tool use life and functional versatility are too limited.
The optimal artifact design is thus the one that produces the greatest potential utility relative to the cost of transporting it. Bifacial tools have that capacity if made of high-quality stone. Bifaces can have fairly sharp but durable edges that can be repeatedly resharpened, and from which flakes can be removed for expedient use, all within a thin and low-mass form. If bifaces are the optimal tool form in terms of portability and potential utility, versatility, and durability, then these tool forms should easily outnumber all other tool forms at the sites if mobility is high, with high bifacial to unifacial tool ratios. These optimal tool forms should give way to more task-specific or expedient tools if mobility decreases and site use becomes intense. A utility/mass ratio, derived from Kuhn’s (1994) measure of potential utility versus transport costs of the same name, for bifacial tools may provide even more insights into the degree of mobility. Using complete or nearly complete bifacial tools, the utility/mass ratio examines a tool’s potential utility in the form of the number of working edges and their total length versus its transport costs in the form of its weight. Higher ratios would be indicative of higher degrees of mobility.
These and other models and approaches provided a framework for analysis and interpretation of evidence from the three Waco lake sites. Examples of how these analytical tools were applied to the data are provided in the three sites sections, the What Was It Like sections, and the concluding section, Bosque River and Beyond.