BINFORD_Mobility, Housing, And Environment, A Comparative Study

Mobility, Housing, and Environment: A Comparative StudyAuthor(s): Lewis R. BinfordReviewed work(s):Source: Journal of Anthropological Research, Vol. 46, No. 2 (Summer, 1990), pp. 119-152Published by: University of New MexicoStable URL: http://www.jstor.org/stable/3630069 .Accessed: 24/11/2011 23:19

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

University of New Mexico is collaborating with JSTOR to digitize, preserve and extend access to Journal ofAnthropological Research.

http://www.jstor.org

http://www.jstor.org/action/showPublisher?publisherCode=unm

http://www.jstor.org/stable/3630069?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

MOBILITY, HOUSING, AND ENVIRONMENT: A COMPARATIVE STUDY

Lewis R. Binford

Department of Anthropology, University of New Mexico, Albuquerque, New Mexico 87131

A cross-cultural survey of hunter-gatherers is conducted with particular emphasis on housing, mobility, and subsistence as these features vary with ecological settings and with particular environmental variables. Implications are drawn for investigations of variability as it is documented archaeologically. Particular emphasis is given to the features listed above, and to arguments in the literature that cite these variables and seek to evaluate the relative "complexity" of ancient sociocultural systems known from archaeological materials.

A MAJOR DEBATE IN archaeology is currently shared with many ethnologists and cultural anthropologists. Two issues are central to this debate: how does inequality come into being, and, more important, are changes toward inequality in any way understandable in terms of fundamental man-environment relationships (so-called adaptational approaches) as opposed to the intrinsic dynamics of "social relations" proceeding independently of external conditions? This debate is commonly conducted with representatives of both positions present, and these joint meetings and conferences frequently result in a "feel- ing" of agreement on what complexity is, as well as in some suggestions about how it might be manifest archaeologically. Nevertheless, much disagreement remains over the "causes" of complexity, a condition that is not surprising since, in my view, the discussion is commonly conducted as a kind of value judgment regarding the quality of a systems state.

"Complexity" is not a variable in any scientific sense of the word, and it is certainly not a dimension with a unique instrument for measurement (e.g., the unique instrument for measuring weight is a scale). It is therefore not surprising to hear that "prime movers have become obsolete along with the trend toward downplaying linear evolutionary typologies" (Brown and Price 1985:437). One can hardly seek an analytical explanation for any judgment without adding and subtracting "conditions" unique to each case, as well as identifying the criteria guiding each judge. This dilemma in no way impacts the analytical approaches of science, which operates with the understanding that system states are unique to systems, but nevertheless we can explain variability in systems organizations if we understand the causal processes that condition the expression of several variables as they may be manifest in a configurational sense in any given system. Minimally we have to identify what it is we wish to investigate about the external world, and then we can proceed. Thus, to approach our particular problem, we can ask what are some of the most commonly cited phenomena used to judge whether something is "complex."

In a major book devoted to this debate, the following "domains" have been identified as contributors to the judgments passed as to whether a particular

119

120 JOURNAL OF ANTHROPOLOGICAL RESEARCH

case is "complex": "population density, maximum settlement size, permanent shelter, permanent ceremonial grounds, art styles, and differences in burials" (Brown and Price 1985:437). In spite of admonitions about avoiding "prime movers," we are told that "cultural complexity in one sense is a response to the problems resulting from decreased mobility" (Brown and Price 1985:437). In this paper I will accept this statement as a guide to further research; however, study must be aimed at explaining variation in the variables that characterize organizational dynamics, not at building static functional models. In short, we want to know how evolution works. Only then can we "understand" its products.

We all bring our prior knowledge and understanding to any learning situation; hence we see different points of interest in our intellectual explorations. I have invested many years in the study of both the archaeological record remaining from the past and modern hunting and gathering peoples from the perspective that an archaeologist might have-that is, in terms of the material consequences of their organized strategies and tactics for solving problems posed by different environments and in different social settings. If we accept the proposition that our prior knowledge strongly conditions what we will accept as feasible and is, in fact, what we invoke in warranting arguments about the "meaning" of observations we may make, then we should be quite concerned with the quality of our prior knowledge. Archaeologists seeking to interpret their site plans and the patterning observed at specific sites rarely have very robust prior knowledge regarding the classes of phenomena they nevertheless interpret.

I have chosen to discuss three aspects of the archaeological record that have received little comparative attention, and my concerns are with three basic issues. (1) How do hunter-gatherers solve their housing needs in different environmental settings? (2) How are these varying solutions related to mobility strategies and subsistence strategies as these solutions may also be conditioned by the environment (including other groups of people)? (3) Finally, in light of these two points, how do we deal with the problem of "complexity"?

HOUSING AND MOBILITY

Let's begin with the relationship between housing and mobility. I think it is fair to say that all else being equal, there is a very general inverse relationship between mobility and investment in housing. Certainly this is the case among tropical and subtropical hunter-gatherers, who are generally very mobile but paradoxically do not move over very large ranges. In the tropics, mobility takes the form of frequent residential camp moves (with one place rarely occupied more than three to five days), yet the total amount of area covered is consid- erably less than that we see in other environmental settings. Variation in labor investment in shelter among such people is primarily related to providing relief from the oppressive heat in dry regions and some protection from drenching rains in other settings. In almost all of these cases, housing is expediently produced; that is, the materials used to construct shelters are not transported

MOBILITY, HOUSING AND ENVIRONMENT 121

from one place to another by the hunter-gatherers as they move. It is therefore not surprising that these people regularly make use of natural shelters, such as caves and rockshelters as well as cliff bases and large rocks. These locations provide both shade and protection against wind.

In the subtropics, particularly where rain is seasonally more common during the cooler months, mobile groups often position themselves in the habitat in order to make use of natural shelters and at the same time construct expediently built shelters nearby, which serve slightly different functions. When I was studying the Australian aboriginal peoples, they commonly constructed brush windbreaks on flat locations with soft sediments, often at the base of talus deposits, above which were rock exposures with numerous rockshelters. This positioning and the uses made of these natural and constructed shelters are good illustrations of the different dimensions in terms of which shelter is both sought and constructed.

In arid to semiarid settings, extreme contrasts between daytime and nighttime temperatures are common. During the Australian dry season, which is also their winter, and among the well-documented San peoples of the Kalahari, we note an interesting phenomenon that is frequently considered strange to "Western" ways of thinking-namely, winter or cool-weather shelters frequently have no roof and are some combination of a windbreak and a lean-to form of shelter, whereas during the very warm seasons, both groups commonly construct roofed shelters. These often-photographed "beehive" shelters are of two general types: a stick-and-brush structure or a frame-and-thatch structure. The roof clearly serves as a shade from the sun or alternatively as protection from rain. In these situations the roof is in no way directly related to the need for conservation of heat, as is the common Westerner's assumption.

Conversely, the absence of a roof during the cold season is related to thermal conservation. Given extreme variation between daytime and nighttime temperatures, the primary concern is to prevent loss of body heat while the occupants are sleeping. Among the native Australians and the San, windbreaks are positioned in order to expose the soft-sediment "floors" to maximum sun- shine. This has the effect of warming the surfaces upon which the occupants will sleep during the night. It is not uncommon to deposit white ash from the hearth over these surfaces in the late afternoon, thereby insulating the surface that was exposed to the sun during the day. In cold nighttime situations, fires may be kindled as darkness approaches, and sand piled near the fire to warm it. This warmed sand may then be spread over a sleeping surface along with warm ash to reduce heat loss further. Sleeping under these conditions is commonly a very intimate affair, since occupants often sleep very close to one another to conserve body heat. Still other precautions may be taken: for example, rubbing the body with grease, which serves as an insulator and also "collects" white ash on the body, further insulating the sleeper.

With this brief introduction to housing in the tropics, it is time to introduce some general knowledge. (1) There are no known cases among modern hunter- gatherers where shelter is not fabricated in residential sites (anywhere that


hunter-gatherers plan to sleep), regardless of the expected occupational duration, and only in rare instances are sites of any kind produced by hunter- gatherers where no shelter is provided for the occupants. (2) Archaeologists viewing the remains of expediently constructed shelters often see only hearths and lithic scatters. Most archaeologists consider these scattered features and artifacts to be "special-purpose" sites when they are encountered in the temperate and higher latitudes. However, it is quite likely they are residential sites of highly mobile people who do not transport shelter and who expediently use local materials to provide shelter during highly mobile phases of their settlement round.

Having made this point, we may now turn our attention to housing and mobility as it is documented among historically known hunter-gatherers. Over the years I have accumulated a file of data that, at present, boasts 198 different cases distributed worldwide, but as you might imagine, all cases cannot be equally studied for all the varying properties that we might wish to discuss. This corpus of data does, however, provide us with the opportunity to look for patterning in a number of properties that are of interest to hunter-gatherer studies. More important in this case, the data also serve as a provocative summary of the behavior of fully modern humans, a summary against which to play off our archaeological evidence from earlier time periods and differing environments. I will use data from this file to address the question of housing in the terms outlined in the above paragraphs.1

The initial work of Murdock (1967) was the foundation of this corpus of data, but I have added both cases and variables over the years-particularly variables descriptive of the environments in which the ethnographically documented hunter-gatherers lived. A first step in the examination of the relationships between housing and mobility can be made using Murdock's original data. I will examine several attributes of house form relative to a rough ordinal scale of mobility.

The terms fully mobile or nomadic imply that residential mobility is high throughout the year. Seminomadic describes the situation, common in the high latitudes, where relatively localized "winter" house sites are lived in continu- ously during the coldest periods of the year. These sites are located adjacent to stored foods or reliable "patches" of available food (e.g., near breathing holes where seals can be hunted). We can also imagine this pattern in environments where social conditions, such as demographic packing (see Binford 1983:208-13), limit mobility and force intensive use of localized resources, as might occur processually with a shift to the use of domesticated plants or increasing intensification in the exploitation of wild foods.

Semisedentary refers to a mobility situation in which relatively permanent house sites are maintained and returned to frequently, but where groups make seasonal forays from these hubs, moving their residential camps from place to place. Fully sedentary refers to groups who maintain living sites that are regularly used and to groups who do not move their residences from year to year, although task units may travel out periodically.


Table 1 displays a sample of the world hunter-gatherer data tabulated by the shape of the ground plan of each group's most substantial (primary) houses relative to the mobility scale described above. We note immediately a strong set of relationships between mobility pattern and house plan. Fully nomadic people tend to construct houses with a circular or semicircular plan, as do seminomadic people, while both semisedentary and fully sedentary people favor rectangular house plans, although semicircular plans continue as a secondary form. Elliptical forms appear to be more characteristic of seminomadic hunter- gatherers. This pattern is an empirical generalization. Can we introduce more observations in the hope of gaining a sufficient understanding of the variables conditioning house design and enhancing our understanding of the relationships between housing and mobility?

One of the obvious variables that those familiar with either archaeological or ethnographic data will recognize is the placement of the house relative to the ground surface. Table 2 summarizes the comparative hunter-gatherer data on this property. Is the ground surface modified prior to house construction, or is the house simply built on the natural ground surface? Among hunter- gatherers the most common situation is simply to place the house on the existing ground surface (perhaps cleaned and leveled slightly), yet some people invest considerable labor in digging out house pits so the house floor is con- siderably below the natural ground surface. Still other groups raise their houses on earth mounds or built-up pilings. Can we understand this differential investment of labor?

In Table 2 we see the relative amount of investment in the preparation of the house site. Clearly, most historically known hunter-gatherers (67.6 percent) simply build their houses on the ground surface, yet we note that this

TABLE 1 Ground Plan of Primary Structure vs. Hunter-Gatherer Mobility

Ground Plan Semi-

Mobility Circular circular Elliptical Rectangular Complexa Total

Fully 4 22 0 5 0 31 nomadic (12.9) (71.0) (16.1)

Semi- 0 65 10 18 0 93 nomadic (69.9) (10.7) (19.4)

Semi- 0 12 1 18 0 31 sedentary (38.7) (3.2) (58.1)

Fully 0 3 0 17 1 21 sedentary (14.3) (80.9) (4.8)

Total 4 102 11 58 1 176

Note: Numbers in parentheses are percentages calculated in terms of row totals. a. Complex ground plans are polygonal or quadrangular and may include an interior court.


TABLE 2 Placement of Primary Structure vs. Hunter-Gatherer Mobility

Structure Placement Ground Semi- Raised

Mobility Surface subterranean Earth Pilings Total

Fully 29 1 1 0 31 nomadic (93.6) (3.2) (3.2)

Semi- 66 27 0 0 93 nomadic (70.9) (29.1)

Semi- 15 16 0 0 31 sedentary (48.4) (51.6)

Fully 9 11 0 1 21 sedentary (42.9) (52.4) (4.8)

Total 119 55 1 1 176 (67.6) (31.3) (0.5) (0.5)

Note: Numbers in parentheses are percentages calculated in terms of row totals.

is vastly more common among more mobile people than among less mobile people, with the latter more regularly preparing the house site for construction. Can we begin to view some of this formal variability in terms of variables that might be of explanatory value?

Mobile people seem to practice two basic strategies: they either transport their basic housing materials, or they regularly construct houses from materials available in their immediate environment. Australian Aborigines do not transport housing materials but instead construct shelters from brush and "thatching" materials readily available in their environment. By way of contrast, the classic North American Plains bison hunters transported both the materials used to cover their houses as well as the wooden poles used to support the covering. This was also true of the Nunamiut Eskimo with whom I worked almost twenty years ago.

These very different tactics represent the different technological consequences of planned strategies among mobile people. In the case of high-investment housing, the amount of investment made in nontransportable housing must bear some relationship to the planned duration of stay or planned reuse of locations in the landscape. In the case of low-investment housing, the amount of investment is related to the scale of mobility and to the transport costs and the potential of the technology. What is being suggested here is simply that the transport of housing is costly in terms of energy expenditure. The frequency with which housing materials are transported is likely to vary with the technology of transport itself and the scale of mobility involved.

As a check on our thinking thus far, let us examine another formal characteristic of hunter-gatherer housing-whether the walls of the house are made


of the same material as the roof. It could be argued that if groups were carrying the materials for house construction with them as they moved across the landscape, the same material would be used for both the walls of the house and the roof. Use of different construction materials for roofs and walls would be more likely to occur when a major labor investment was made in housing or when locally obtainable materials were being used in construction. On the other hand, if only the house cover was transported, the supports or framing material might be collected locally. Depending on the transport potential of the technology, other alternatives are possible as well. Table 3 summarizes the world hunter-gatherers according to whether the same or different materials are used to cover the walls and the roofs of their houses. Primary houses, often built to withstand winter conditions, are the most substantial; secondary, or alternative, houses, when used, are often constructed during less harsh seasons and are occupied for shorter periods of time.

As in earlier examples, we note a very strong set of relationships between the formal features of house design and mobility. As anticipated, more mobile people tend to cover both the roofs and the sides of houses with the same material, whereas less mobile people tend to differentiate the sides from the roofs in terms of the covering materials used, except in the case of the secondary housing among semisedentary people. On the other hand, when there is an alternative form of the principal house type (listed as secondary housing on Table 3), we note that among more mobile people the percentage of secondary forms manufactured of differing materials tends to be greater than that

TABLE 3 Wall and Roof Material for Primary and Secondary Housing vs.

Hunter-Gatherer Mobility

Roof and Sides Roof and Sides of Same Material of Different Material Total

Mobility Primary Secondary Primary Secondary Primary Secondary

Fully 29 8 1 1 30 9 nomadic (96.6) (88.9) (3.4) (11.1)

Semi- 82 48 11 13 93 61 nomadic (88.2) (78.7) (11.8) (21.3)

Semi- 13 12 17 5 30 17 sedentary (43.3) (70.6) (56.7) (29.4)

Fully 3 0 18 4 21 4 sedentary (14.3) (85.7) (100.0)

Total 127 68 47 23 174 91 (73.0) (74.7) (27.0) (25.3)

Note: Numbers in parentheses are percentages calculated in terms of row totals by type of housing (primary vs. secondary).


of primary forms (fully mobile, 3.4 percent compared to 11.1 percent, and seminomadic, 11.8 percent compared to 21.3 percent, for primary and secondary houses, respectively). With semisedentary people the shift is in the other direction-that is, a lower percentage of secondary housing, as compared to primary housing, has walls and roofs manufactured of different materials. In semisedentary groups, the secondary housing is indicative of the mobile phases of their adaptation, which are by definition quantitatively secondary to the periods spent in the primary houses.

These patterns make a great deal of intuitive sense and can be seen as relating to the importance of mobility relative to the transportation potential that a particular technology affords. Examples of alternative housing that employ different materials for the walls and the roofs most often occur when what I have previously called "site furniture" is accumulated at regularly used locations.

As a further clue to understanding the variation in hunter-gatherer housing, let us examine the kinds of primary housing roofing materials used by groups characterized by different levels of mobility. Table 4 summarizes the worldwide record for this property. It is immediately clear that among fully mobile people the roof covering is either transportable (e.g., hides) or locally accumulated (e.g., vegetation or bark). The choice made among these materials most likely reflects differences in environments on the one hand and the relative costs and advantages of transporting housing on the other. Before we consider this suggestion, however, we need to consider the nature of the relationship between mobility and the very presence of alternative housing (Table 5).

What is interesting in the tabulation provided in Table 5 is that alternative

TABLE 4 Roofing Material (Primary Housing) vs. Hunter-Gatherer Mobility

Roofing Material

Mobility Hides Grass Bark Ice Stone Earth Mats Wood Total

Fully 16 13 2 0 0 0 0 0 31 nomadic (51.6) (41.9) (6.5)

Semi- 12 34 7 5 1 15 13 6 93 nomadic (12.9) (36.6) (7.5) (5.4) (1.1) (16.1) (14.0) (6.5)

Semi- 0 4 4 0 0 9 2 12 31 sedentary (12.9) (12.9) (29.0) (6.5) (38.7)

Fully 0 4 5 0 0 1 1 10 21 sedentary (19.0) (23.8) (4.8) (4.8) (47.6)

Total 28 55 18 5 1 25 16 28 176 (15.9) (31.3) (10.2) (2.8) (0.6) (14.2) (9.1) (15.9)



TABLE 5 Presence of Alternative House Forms vs. Hunter-Gatherer Mobility

Alternative Housing

Mobility Absent Present Total

Fully 21 9 30 nomadic (70.0) (30.0)

Semi- 32 62 94 nomadic (34.1) (65.9)

Semi- 14 17 31 sedentary (43.3) (56.7)

Fully 16 5 21 sedentary (76.2) (23.8)

Total 83 93 176


house types are most characteristic of seminomadic and semisedentary hunter- gatherers and least characteristic of fully mobile and fully sedentary people. Since the archaeological consequences of different house types may be very confusing, this situation presents the archaeologist with many problems, particularly with regard to seminomadic people. My comparative files do not contain data on the numbers of alternative house forms (something to be corrected), but Reinhardt's (1986:197, 213) interesting comparative study of housing among arctic peoples shows that the number of alternative house forms varies inversely with sedentism and with measures of "complexity. "The greater the seasonal contrasts in mobility, and the greater the social and activity-related variability within a seasonal round, the greater the number of alternative forms of housing.

Groups using alternative housing are presented several choices with regard to the construction and transport of its supports and covering. In the High Arctic, small tents made of skins are erected with a minimum of structural support, and the tents tend to be "ridged" rather than conical. Two families can easily put two such units together to make a larger living unit; however, the height is limited by the design of the supporting frame. One common way of making a higher and generally larger house in these settings is to construct a raised frame with locally available materials.

In the Central Arctic, particularly during spring, it is not uncommon to see the normal "summer tent" placed on top of a raised wall base made of blocks of snow. These raised "tent bases" can be found in other settings, such as regularly used summer camps where stone might be piled up as a raised base for the tent. Fall locations similarly might have such raised bases, particularly if groups were engaged in obtaining furs for winter clothing and "sewing camps"


were established where groups of women worked together in the "larger" houses (Boas 1888; Gronnow, Melgaard, and Nielsen 1983; Mathiassen 1928). Remains found at ancient archaeological sites, where durable materials (stone and whale bone) had been used in the construction of winter houses, also were commonly "recycled" by later occupants for use with mobile tent covers during spring, summer, and fall.

People living in the boreal forest do not transport frame supports for their conical houses (tipis) since young trees can be obtained locally throughout most of the area. In these high-primary-biomass forests, alternative covering materials, such as bark and brush (instead of hides), are also readily available, particularly during warmer months. We would expect to see greater use made of these expedient, nontransported materials by occupants of temperate and boreal settings. On the bison plains of North America, however, poles and hides were transported since the grassland offered few materials for tent supports or covering and the documented groups living there used dogs as draft animals to minimize the transportation costs.

Alternative houses, then, may take many different forms. In a fairly productive environment, mobile groups can count on finding usable material for house covering almost anywhere they camp, whereas in less productive settings the house covering (hides) may be prepared and then transported when the group moves. This type of planning is also related to the distances of moves and the costs of transporting housing materials regardless of the environment. Table 6 indicates that as compared to primary (usually cold-season) housing, alternative, usually warm-season housing among mobile people shows an increase in the use of vegetation as a covering.

TABLE 6 Roofing Material (Alternative Housing) vs. Hunter-Gatherer Mobility

Roofing Material

Mobility Hides Grass Bark Ice Stone Earth Mats Wood Total

Fully 0 6 2 0 0 1 0 0 9 nomadic (66.7) (22.2) (11. 1) (9.7)

Semi- 10 18 11 0 0 5 13 5 62 nomadic (16.1) (29.0) (17.7) (8.1) (21.0) (8.1) (66.7)

Semi- 2 2 6 0 0 1 4 2 17 sedentary (11.8) (11.8) (35.3) (5.9) (23.5) (11.8) (18.3)

Fully 1 3 0 0 0 0 0 1 5 sedentary (20.0) (60.0) (20.0) (5.4)

Total 13 29 19 0 0 7 17 8 93 (14.0) (31.2) (20.4) (7.5) (18.3) (8.6)

Note: Numbers in parentheses are percentages calculated in terms of row totals, with the exception of the total column where percentages are calculated by column.


The difference in roofing material between primary and secondary (alternative) housing is very important (see Tables 4 and 6). What it tells us is: (1) if mobility is high and portable housing is being used, the same type of housing is most commonly used in all seasons-that is, alternative housing is generally not employed (see Table 5); and (2) if mobile people do have alternative house types, the houses are almost exclusively roofed with nontransportable materials, such as vegetation or bark. As previously noted, alternative housing is characteristic of seminomadic people, and for these groups we see some interesting shifts: as compared to primary housing, the use of hides for alternative housing goes up slightly, the use of vegetation goes down slightly, and the use of both bark and mats increases slightly. Dramatically reduced is the use of earth, as one might expect, since we are primarily looking at the differences between winter (primary) and summer (alternative) housing. Among semisedentary people, a dramatic shift occurs from earth and wood in the winter to bark and mats in the summer. Among fully sedentary people, summer houses are primarily roofed with vegetation, whereas winter houses are roofed with wood.

My comparative data file is still incomplete; however, a subset of the file has been coded for transport. Table 7 summarizes the relationships between two interesting variables, the portability of housing (generated as a cumulative score for structure supports, covering, and foundation) versus the technology of transport-that is, whether humans, humans assisted by pack dogs, or true draft animals are used for transport. The cases in the final category represent both historical and protohistoric groups on the North American bison plains.

Clearly the sample of cases thus far coded is insufficient to make very secure generalizations; nevertheless, tendencies are apparent in the sample. Simply put, the greater the technological aids to transport, the greater the portability in

TABLE 7 Portability of Housing vs. Transport Technology

Portability of Housing Transport Moderately Moderately Agents Low Low Medium High High Total

Humans 3 8 5 11 1 28 (10.7) (28.6) (17.9) (39.3) (3.6)

Humans and 0 0 1 0 1 2 pack dogs (50.0) (50.0)

Humans and 1 0 0 5 3 9 draft animals (11. 1) (55.6) (33.3)

Total 4 8 6 16 5 39



the housing. Within the same sample, I could demonstrate this tendency equally well using the character of facilities for transportation on water.

Summary Our review of at least some features of the relationships between mobility

and housing allows several generalizations to be made. I will concentrate here on seminomadic people since in later discussions this adaptation takes on greater importance.

1. Houses of seminomadic people tend to be semicircular in ground plan and to be constructed directly on the ground surface.

2. The roofing material of winter (primary) housing tends to be the same as that used to cover the walls of the house-either hides or locally obtained vegetation. The kind of material used varies with the degree of mobility, planning, and the availability of suitable local materials in the habitat.

3. The use of multiple alternative house forms depends primarily on the variation in mobility, group size, and activities during an annual subsistence round.

4. In roof covering and shape, alternative housing tends to have many of the characteristics of primary housing, as is implied by similar covering materials being used for both roofs and walls; however, many ingenious forms make use of similar materials to vary house size and thermal efficiency as well as transport costs. Locally collected materials, particularly bark, vegetation, and mats, tend to be more important as covers for alternative (summer) housing.

5. The relationship between the portability of housing and the transport technology appears to be very strong indeed. When groups are limited to human transport only, we can expect minimal use of well-designed portable housing units, such as large tents made of skin and sturdy frame supports. (When substantial natural shelters are readily available, such as the rockshelters and caves of Western Europe, animal domestication for assistance in transport could be delayed.)

ENVIRONMENT, SUBSISTENCE, MOBILITY, AND HOUSING

In order to understand housing further, we must understand something about the sheltering requirements being met by persons in different environments, as well as the materials available for use in meeting these requirements. To examine environments, I employ a very useful measure of biologically relevant temperature developed by Bailey (1960). Effective temperature (ET) is a simultaneous measure of solar-radiation intensity (overall "warmth") and the length of the growing season. The coldest place on earth (Antarctica) has an ET value of 8.0, the biological boundary between boreal forest and tundra is ET 10.0, and equatorial systems (with a 365-day growing season) have an ET value between 18 and 26. Temperate settings range in ET values between 18


TABLE 8 Length of Growing Season and Overall "Warmth" (ET) vs.

Hunter-Gatherer Mobility

Effective Temperature (ET)

Mobility 9 10 11 12 13 14 15 16 >17 Total

Fully 2 0 0 1 3 1 2 1 11 21 nomadic (20.0) (4.5) (10.3) (5.5) (14.3) (33.3) (64.7)

Semi- 6 8 14 13 18 9 11 1 5 85 nomadic (60.0) (57.1) (56.0) (59.1) (62.1) (50.0) (78.6) (33.3) (29.4)

Semi- 1 2 8 7 4 3 1 0 1 27 sedentary (10.0) (14.3) (32.0) (31.8) (13.8) (16.7) (7.1) (5.9)

Fully 1 4 3 1 4 5 0 1 0 19 sedentary (10.0) (28.6) (12.0) (4.5) (13.8) (27.8) (33.3)

Total 10 14 25 22 29 18 14 3 17 152

Note: Percentages are calculated by column in this table.

and 10. Table 8 arrays the historically documented cases of hunters and gatherers for which I have sufficient environmental data against ET values calculated from weather records for the ranges within which these various groups lived.

Two very important facts are illustrated here: (1) Fully nomadic people are extremely rare in the temperate and arctic zones; even the two cases tabulated for the Arctic could be disputed. On the other hand, as the growing season lengthens, nomadism increases in a very even gradient. (2) Seminomadic people are the most common of all the documented hunter-gatherers and dominate the mobility pattern in higher-latitude settings. This makes a very straightforward pattern. The more severe the winters, the less mobile one can reasonably be; therefore, tactical investment in obtaining sufficient foods is essential during the seasons when mobility does not place one in either subsistence or personal jeopardy-that is, before the onset of severe winter conditions. This means that, in general, storage tactics are essential to the successful practice of a seminomadic mobility strategy in a northern setting. (Exceptions include the Central Eskimo, who exploit seals at breathing holes during the "nonmobile" winter phase of their adaptation.) The above generalization also means that the farther one is from the equator, the more one is dependent on nonplant resources (see below); it is primarily hunted foods that are processed for storage, making possible reduced mobility during the severe winter months.

I am suggesting a rather straightforward set of broad, determinant relationships conditioning the global patterns of mobility, dependence on storage, and exploitation of nonplant foods. Clearly there are two basic types of nonplant foods: terrestrial animals and aquatic animals. In the exploitation of aquatic versus terrestrial resources, one can expect increasing exploitation of aquatic animals to be a function of (1) decreasing ET values and (2) the contrasts in


mobility costs relative to storage potential in a given area (this boils down to the patchiness of the distribution of resources and the productivity of those patches). The above expectation assumes that aquatic resources are the targets of exploitation for winter stores. Of course, there are environments in which aquatic resources may be exploited seasonally yet not targeted for storage. This habitat trade-off (between aquatic and terrestrial resources) offers the potential for great variability in mobility and subsistence strategies, particularly among people living in the higher latitudes.

Finally, semisedentary and sedentary peoples are not randomly distributed with respect to the earth's environments. Among sedentary groups, a clump of cases clearly occurs in the high latitudes (ET 10); the frequency then decreases, only to rise again at roughly the earth's mean biotemperature (ET 14) and to increase toward the equator. Is this a sampling bias, an accident of history, or is something important indicated here?

Before we proceed, it is important to examine the empirical basis for the above argument and to provide a more understandable environmental context for further discussion. Table 9 presents the hunter-gatherers as classed by


Hunter-Gatherer Dependence on Plant Foods

Percentage of Depenrc detane ofEffective Temperature (ET) Dependence on Plants 9 10 11 12 13 14 15 16 ~17 Total

0 3 2 0 0 0 0 0 0 0 5 (42.9) (14.3)

10 4 6 6 4 1 0 0 0 0 21 (57.1) (42.8) (30.0) (17.4) (3.6)

20 0 6 9 6 2 0 0 0 1 24 (42.8) (45.0) (26.1) (7.1) (6.7)

30 0 0 5 11 8 1 0 0 6 31 (25.0) (47.8) (28.6) (5.9) (40.0)

40 0 0 0 1 7 10 2 1 5 26 (4.3) (25.0) (58.8) (14.3) (20.0) (33.3)

50 0 0 0 1 10 5 3 0 2 21 (4.3) (35.7) (29.4) (21.4) (13.3)

60 0 0 0 0 0 1 7 2 1 11 (5.9) (50.0) (40.0) (6.7)

70 0 0 0 0 0 0 2 0 0 2 (14.3)

80 0 0 0 0 0 0 0 2 0 2 (40.0)

Total 7 14 20 23 28 17 14 5 15 143

Note: Percentages are calculated by column in this table.


Murdock for their dependence on plant foods arrayed against ET values. The most obvious feature of this table is that the cases are arranged in a remarkably diagonal pattern; for every increase in ET, case positions shift downward, thus documenting a greater dependence on plant foods throughout the temperate zones. In the tropics things become somewhat complicated; elsewhere I have suggested some of the conditioners of this variation (Binford 1989b). Our concern here is with nonequatorial environments, however, and it is apparent that the length of the growing season is a primary conditioner of this very regular pattern. Contrary to a previous suggestion by Lee (1968), the pattern does not have reference to a reduction in the absolute number of plants present- only to their properties as they are adapted to differing growing season lengths and an autocorrelation with their suitability as human food.

For instance, more than 223 species of plants are available within a five- mile (8 km) radius of Anaktuvuk Village, a tundra setting in the center of Alaska's Brooks Range. Of these, however, only 7 species are considered to yield edible products; and these 7 species (1) are dispersed, (2) yield very small edible parts, (3) are available for very short periods of time, and (4) are not distributed in high-yield patches. We are considering here the structural variability in plant communities as it affects the production of foods suitable for humans, not a simple quantitative measure.

This strong pattern is an environmental baseline conditioning all uncon- strained hunter-gatherer adaptations in the nonequatorial zones. As mentioned above, this zonal pattern describes a basic set of adaptive problems presented to human occupants of these variable zones. All else being equal, as one moves farther from the equatorial zone, one is forced to exploit increasing numbers of animals in order to meet the food needs of a human community. The options that are played out in the documented sample, as well as in the history of modern human presence in the temperate latitudes, are between terrestrial adaptations and increasing dependence on aquatic resources. But these alternative strategies are not simply options; instead they are different responses to different environments in the nonequatorial zones.

We could strongly argue that hunter-gatherers could not exist in High Arctic settings without the presence of a productive aquatic alternative to terrestrial foods. Put another way, the farther from the equator one goes, the lower the productivity of the biome as a simple function of the decrease in solar energy available for biological conversion. All else being equal, two features of hominids become increasingly critical as one moves toward the poles: (1) the relatively constant demand for food does not change, and (2) the presence of foods requiring relatively minimal search time decreases in a graded fashion. These two features force an increased dependence on terrestrial mammals. Depen- dence on these mammals, because they are mobile, vastly increases the search time required by hominids to meet their relatively inelastic food demand curve. Similarly, as gross primary productivity goes down, the scale of animal move- ment increases, further increasing the time hominids must spend to locate potential foods. In practical terms this means a drastic increase in the mobility


costs for the hominids. The returns per unit increase in mobility, nevertheless, continue to go down as one moves toward the poles, thus providing a very strong selective context favoring any alternative food-getting strategy that will reduce the search time and simultaneously increase the return rate per unit of search time. In broad terms this is the selective context for expanding diet breadth to include aquatic resources and for increasing the effectiveness of a technology that permits the accessing of aquatic resources. Fundamentally, as one moves towards higher latitudes, the search time and the attendant mobility costs of exploiting terrestrial mammals are expected to increase logarithmically, given a constant food demand. Thus both the historical and geographical differences between terrestrially oriented and aquatically oriented subsistence strategies can be explained ecologically.

The argument given above anticipates a regular environmental gradient in the degree to which aquatic resources can be expected to substitute for terrestrial mammals in their role of offsetting the environmental shortfall of reliable and usable plant foods so graphically documented in Table 9. A test of this expectation is perhaps best accomplished by summarizing the sample data on the relative dependence of hunter-gatherers on aquatic resources versus effective temperature, already shown to be a sensitive measure of environmental potential with regard to plant use (Table 10).

Immediately we can appreciate that the vast majority of the historically documented hunter-gatherers compensate for the decreasing potential of plant resources in higher latitudes by increasing the exploitation of aquatic resources. The diagonal distribution of aquatic dependence is obvious and represents regular increases with decreasing effective temperature on a global basis. This pattern has enormous importance for our use of historically documented hunter- gatherers as analogies to Pleistocene people, who, it is well known, favored terrestrial resources. It is also clear that during major periods of increased cold weather in the Pleistocene, many areas appear to have been abandoned (Jochim 1987; Soffer 1985:247), whereas in the historical sample, analogous environments are occupied and some, at least, support rather "complex" hunter- gatherers generally dependent on aquatic resources. In fact, if we examine the historically documented peoples, we note that very few solely terrestrial hunters are included: the Nunamiut and Caribou Eskimo, the Chipewyan, and the historical Naskapi, all caribou-reindeer hunters, and then the large group of people inhabiting the North American Plains and heavily exploiting bison. All the others have gone through the "aquatic transition," as we can see in Table 10. I consider the understanding of this temporally correlated shift to the increasing use of aquatic resources to be one of the major problems archaeologists have yet to address realistically in terms of the issues of complexity and human evolution. Importantly, this shift would be favored in simple energetic terms and would not necessarily be a response to density-dependent "frustration" of mobility--the fundamental hunter-gatherer positioning strategy for gaining a living.

Returning to the issue more germane to the materials to be considered



Hunter-Gatherer Dependence on Aquatic Foods

Percentage of Dependence on Aquatic Effective Temperature (ET) Resources 9 10 11 12 13 14 15 16 >17 Total

0 0 0 0 0 1 0 8 3 5 17 (6.0) (47.0) (18.0) (29.0)

10 0 0 0 1 3 3 1 0 1 9 (11.0) (33.0) (33.0) (11.0) (11.0)

20 1 1 3 2 8 3 3 0 5 26 (4.0) (4.0) (12.0) (8.0) (31.0) (12.0) (12.0) (19.0)

30 1 1 1 2 8 7 1 0 3 24 (4.0) (4.0) (4.0) (8.0) (33.0) (29.0) (4.0) (13.0)

40 1 1 6 6 4 1 1 0 1 21 (5.0) (5.0) (29.0) (29.0) (19.0) (5.0) (5.0) (5.0)

50 1 6 5 8 2 3 0 0 0 25 (4.0) (24.0) (20.0) (32.0) (8.0) (12.0)

60 1 4 5 3 2 0 0 0 2 17 (6.0) (24.0) (29.0) (18.0) (12.0) (12.0)

70 1 0 0 1 0 0 0 0 0 2 (50.0) (50.0)

80 1 1 0 0 0 0 0 0 0 2 (50.0) (50.0)

Total 7 14 20 23 28 17 14 3 17 143


shortly, we might ask what the relationship is between terrestrial hunting and housing. I have chosen to illustrate this relationship in an economical fashion by using the portability measure previously described when transport technology was used as an indicator of housing design (Table 7). This same measure is arrayed in Table 11 against the dependence of groups on terrestrially hunted foods.

Here we see a very important pattern. There is an unmistakable relationship between dependence on hunting and the portability of primary housing. We have previously seen a similarly strong relationship between the transport technology and the portability of housing. If we knew nothing else, we would immediately infer that heavy dependence on hunting is strongly correlated with transport technology. This is in fact the case. The correlation is so strong that there are no known examples of historically documented people whose dependence on hunting terrestrial animals exceeds 40 percent and who do not use dogs and/or horses or reindeer as draft animals. In short, in the modern world all people who are heavily dependent on hunting also have a relatively


TABLE 11 Portability of Housing vs. Relative Dependence of Hunter-Gatherers on

(Terrestrial) Hunted Foods

Percentage of Portability of Housing

Dependence Moderately Moderately on Hunting Low Low Medium High High Total

0-10 2 3 0 1 0 6 (33.3) (50.0) (16.7)

11-20 8 9 3 6 2 28 (28.6) (32.1) (10.7) (21.4) (7.1)

21-30 10 18 19 11 4 62 (16.1) (29.0) (30.7) (17.7) (6.5)

31-40 1 4 6 19 2 32 (3.1) (12.5) (18.8) (59.4) (6.3)

41-59 0 2 1 9 3 15 (13.3) (6.7) (60.0) (20.0)

60 plus 0 0 1 3 20 24 (4.2) (12.5) (83.3)

Total 21 36 30 49 31 167


elaborate transport technology and capability. If I had included transportation on water, the complexity of the transport technology would be even more impressive.

Summary In our brief look at environment, subsistence, mobility, and housing, we

have learned some very interesting things. 1. Among historically documented hunters and gatherers, fully nomadic peo-

ples (read foragers) are very rare in temperate and high-latitude settings. In turn, seminomadic peoples are most common and in fact dominate the hunter- gatherers found in those settings. There are two interesting and very provocative exceptions-the Ona of Tierra del Fuego, who are both hunters and foragers, and the Tasmanians. We will consider these exceptions later.

2. When we investigated the relative dependence on plant foods, we noted a strong and environmentally determined relationship among hunter-gatherers: as one moves farther from the equator, the use of plant foods decreases regularly as a simple function of the length of the growing season and the amount of solar radiation reaching the earth's surface. This ecologically understandable relationship can be expected to apply to ancient hunter-gatherers


as well as to modem ones, since the causal conditions can reasonably be accepted to have operated in the past as well as the present.

3. In light of these strongly ecologically determined conditions, we have noted essentially two major strategies for procuring the needed substitutes for plant foods in temperate and colder environments: the exploitation of aquatic animals and the exploitation of terrestrial animals. When the historically documented cases are surveyed relative to the question of which alternative is most common, the answer is a resounding response in favor of aquatic resources. In fact, terrestrial hunters are rare in the temperate and colder settings, being largely restricted to the continental prairie grasslands and the tundra. Thus, the seminomadic people noted for the temperate and higher- latitude zones are preponderantly also heavy exploiters of aquatic resources.

This situation is very interesting in light of the fact that early hominids, as well as early fully modem populations, successfully occupied temperate and higher-latitude settings and did not depend heavily on aquatic resources. In. fact, although aquatic foods occur in the archaeological record fairly early (Kla- sies [Klein 1974, 1976; Voigt 1973; Volman 1978], Gibraltar [Garrod 1928], Haua Fteah [McBurney 1967]), not until the terminal Pleistocene-early Ho- locene do we see convincing evidence for increasing regular and heavy reliance on aquatic resources in some areas (Clark 1987; Brown and Price 1985). All our evidence points to the terrestrial alternative as the dominant tactic of early populations of fully modem humans, as well as of still earlier hominid populations. This means that if there are major differences between the exploitation of aquatic resources and the exploitation of terrestrial resources, most of the historically documented peoples in our sample are irrelevant as analogs for Pleis- tocene terrestrial hunters.

The major variables that are suspected to differentiate between aquatically oriented hunter-gatherers and terrestrial animal hunters are mobility and land- use patterns. Actual mobility data are very hard to find in the ethnographic literature-that is, how many times per year and how far residences were moved. Table 12 summarizes all the cases from northern temperate to arctic settings that my students and I have been able thus far to document as far as actual mobility is concerned. In all cases the figures refer to the economically organized "on-the-ground" group, not the "ethnic group."

I think it should be clear from even these few cases that terrestrial hunters make many more residential moves per year, travel much greater distances over an annual round, and in turn exploit vastly larger areas than do aquatic resource exploiters. The ecological reasons for this are essentially the same ones that explain why carnivores in general have very large ranges, even in animal-rich environments: the prey animals move, are differentially responsive to differential productivity in the plant community, and are difficult to kill. All else being equal, these facts ensure that hunters of terrestrial animals will exploit large ranges and will be quite mobile. It should be kept in mind that the local group sizes are roughly similar in all these cases except for the Piegan, whose group sizes are slightly larger.


TABLE 12 Measured Mobility for All Northern Temperate and High-Latitude Groups for Which Terrestrial Hunting or Aquatic Hunting and Fishing Represents

More Than 50 Percent of the Total Diet

Combined Percentage of Number of Annual Size of Dependence Residential Distance Group on Animalsa ET Moves/Year Covered (km) Range (km2)

Terrestrial hunters Nunamiut 87 9.8 12 720 6000 Ona 70 9.1 60 ? 790 Mistassini 50 10.8 10 510 3385 Piegan 80 11.4 28 840 8500 Montagnais 60 11.6 13 470 2700 Micmac 50 12.7 6 314 1800

Aquatic hunters and fishers So. Tlingit 60 10.9 3 ? 820 Tshimshian 60 11.1 3-5 87 600 Makah 60 11.3 2 15 190 So. Kwakiutl 50 11.6 3-4 35 727

a. Both aquatic and terrestrial

The interesting exception to the above generalization regarding mobility and resource exploitation is the Ona, who are pedestrian hunters unaided by either pack dogs or draft animals--all transport is by human carrier. They are also exceptional in that they practice little storage as a basic overwintering strategy. Their residential mobility pattern is more like that of equatorial hunter-gatherers (foragers, sensu Binford 1980) known from the historical period, and their group ranges are as small as the ranges of equatorial foragers. Other evidence clearly marks the Ona as foragers: "The fact of having killed three guanacos was sufficient motivation for the transfer of their dwelling to the locality where the guanacos had fallen, as this was easier than carrying the meat to the camp" (Gallardo 1910:240-41).

This fascinating case saves us from making a mistake about the sample of modern hunter-gatherers available for us to study. If it were not for the Ona and a few cases from the Central Arctic (Netsilik and Copper Eskimo), as well as the Tasmanians, one could make the case that foraging strategies were environmentally determined in ways analogous to that argued for the dependence on plant foods documented above (see Table 9, for instance), although I think the theoretical justification for this assertion would be difficult or im- possible to offer. I suspect that logistical strategies are the consequence of two major evolutionary changes that occurred long ago: (1) the "aquatic resource revolution," with its early occurrence primarily in higher latitudes, and (2) the perfection of transport technologies, particularly water transport vessels and the use of pack and draft animals.


The difference between foragers and logistically organized peoples (whether consumers are moved to resources or resources are moved to consumers), other things being equal, is the differential cost of transport. Are the costs less to move consumers to goods, or goods to consumers? In the case of most aquatic resources, one cannot, technically speaking, ever move consumers to goods, only to the best points of access to high-quality resources from a habitat that man cannot inhabit (shellfish and a very few others are the exceptions). I used the phrase "other things being equal" above since under conditions of demographic packing (see Binford 1983:195-213), which inhibits mobility and renders conspecific competitive costs great, humans may be forced to exploit certain resources logistically and pay the increased costs in labor if these resources are important; hence most sedentary people (agricultural or not) exploit wild terrestrial animals logistically.

All these observations and arguments make us suspect certain things about the past. (1) Foraging strategies may have been much more common in temperate and cold settings in the past. (2) If so, then most historically documented hunter-gatherers from these zones are inappropriate analogs for Pleistocene cases in any direct sense. This does not mean that logistical tactics were not employed in the ancient past, but if they were, it would have been a matter of the costs of transport being played off against the costs of residential moves. Thus far we have little information to guide our understanding of these relationships except in terms of bulk or weight units and the costs of their transport. (See, for instance, O'Connell, Hawkes, and Blurton Jones 1988 on the costs of transporting different anatomical parts from animals of varying body weights.) What then about the patterns of mobility and storage?

If our arguments can be sustained thus far, we can offer some problematic expectations about ancient hunter-gatherers living in moderate to cold environments but not heavily dependent on aquatic resources. (1) If transport assistance in the form of domesticated dogs is absent and water transport technology is not developed, then high residential mobility is expected, with a foraging organization being most likely (the Ona are a good modern example of such an adaptation). (2) The degree to which we can expect "seminomadic" mobility patterning is unclear from the sample of historically documented peoples; however, since severe winters make mobility difficult, one can imagine a strong selection favoring such a pattern even if the normal mobility tactics were those of foragers. Our uncertainty on this issue arises from the fact that most documented peoples of the High Arctic, as well as of the boreal and northern temperate zones, are heavily dependent on aquatic resources and are predominantly seminomadic, providing us with a strong environmental correlate to this pattern. (Correlations are not necessarily causally indicative; they are clues to the operation of systemic phenomena in the world. They are what needs explanation-they are not explanations in and of themselves.) (3) I think we can estimate with some confidence the levels of dependence on plants for food that were typical of ancient peoples, since this characteristic seems to be strongly conditioned by understandable ecological-energetic relationships that would have been the same in the past as they are today.


Given what we have learned about hunter-gatherers in general, our ability to link the archaeological consequences of storage to the archaeological cor- relates of mobility patterning is crucial to our understanding of past systems. Only then are we reasonably able to say whether we can expect people in the ancient past to have been seminomadic and when and how we can expect to see the evolution of "complexity."

STORAGE AND ENVIRONMENT

Ironically, many researchers have assumed that modern, "post-aquatic resource transition" hunter-gatherers existed prior to that very transition. Sim- ilarly, they have assumed that logistical organization, with its important transport- related cost budget, is an "alternative" form of organization to be expected or looked for prior to any evidence for draft animals, pack animals, or conditions that make logistics imperative, such as the exploitation of animals of the aquatic biome. On the other hand, storage has been assumed to be recent and to signal a major set of evolutionary changes leading to social complexity and nonegal- itarian societies. I will seek to show that storage, like subsistence in general, is related to environmental variation and in the ancient past can be expected to have been conditioned by very different variables than those that condition "social complexity."

Recent literature by intellectual idealists has generally taken a very different view, which either relates storage to prestige seeking and social manipulation (Bender 1978, 1981) or makes what I have previously termed the "Garden of Eden" assumption a prerequisite for both storage and complexity. The latter view is well illustrated by Testart (1982:523):

Where some natural food resources are bountiful but seasonal, they can be gathered en masse while available and stored on a large scale....

The central idea . .. is that the massive stockpiling of staples constitutes the material base for a possible development of socio-economic inequalities to the extent that the bulk of the production is thenceforward trans- formed (by techniques appropriated and accumulated differentially by individuals or by groups).

Simply put, where the environment is "rich," storage and in turn social hier- archies will result. For the extreme essentialists, one only needs a sufficient environmental base for human greed and prestige seeking to be permitted to flourish.

As in the case of dependence on plants for food, I view storage as a response to environmental conditions. (Ironically, I am in agreement with Ingold [1983] on this point.) I fully recognize that under some demographic conditions and geographic constraints on mobility, storage may be an appropriate response quite independent of climatic concerns. Nevertheless, storage under these conditions simply reflects the character of the selective environment under


which the people were living and is therefore a consequence of selection, as are other frequently correlated social changes.

In order to demonstrate this point, I have prepared two tables from the comparative data file. I have sought to hold environmental productivity constant by selecting cases of hunter-gatherers from relatively unproductive, primarily dry grassland to semidesert settings (where there is less than 100 mm possible annual runoff from realized rainfall). I am concentrating on relatively poor environments (contra the arguments by Testart and others regarding rich environments) in order to examine the potential role of storage as a security mechanism in settings with shorter growing seasons. I have information on storage for forty-nine such cases in my sample. Since the environmental conditions are such that little aquatic dependence would be expected as a result of locally earned water, this suite of cases is also interesting in terms of terrestrial adaptations. Of these forty-nine cases, let us examine initially only those for which there is no aquatic dependence (Table 13).

We immediately note that all cases falling in ET ranges greater than 15.8 practice no storage. Almost all cases falling between ET 15.3 and 14.6 store

TABLE 13 Storage Practices of Nonequatorial Hunter-Gatherers in Settings with

Less Than 100 mm Annual Runoff from Realized Rainfall and Exhibiting No Dependence on Aquatic Resources

Cases ET Storagea Method Facilities

Walbiri 17.9 none none none Hadza 16.2 none none none Dobe !Kung 16.9 none none none G/wi 16.6 none none none Aranda 15.9 none none none Dieri 15.9 none none none

Yavapai 15.3 seeds drying ? Serrano 15.1 seeds, meat drying granaries,

baskets Kaibab 15.1 seeds drying baskets Walapai 15.1 seeds drying baskets Panamint 15.0 pifion nuts drying granaries,

rock crevices

Southern Ute 14.6 bison meat, drying racks, seeds "caches"

Cheyenne 13.9 bison meat, drying racks, berries skin bags

Tehuelche 12.8 meat drying racks, skin bags

a. "None" refers to storage that does not normally exceed a period of three to five days after procurement.


only plant foods, whereas all cases falling at or below an ET value of 14.6 store meat along with some plant products. Based on these data it is tempting to make the following generalizations. (1) At around ET 15.8 there is a threshold, related to the length of the growing season, for the use of storage as a security tactic. (2) Groups in settings warmer than approximately ET 14.6 exhibit some specialization in the storage of plant foods, which are used as reserves primarily during the winter. (3) Groups in settings cooler than ET 14.6 exhibit special- izations in the storage of animal foods as the primary buffer against overwintering shortfalls of food.

At this point these suggestions are certainly tentative; however, additional data are available. There are thirty-five cases of hunter-gatherers living in settings with less than 100 mm of runoff from annual rainfall but variously dependent on aquatic resources as a function of their geographic position adjacent to either oceans or rivers that move locally unearned water through or adjacent to their ranges. Table 14 summarizes the available information on these cases.

This group of cases, all of which occupy relatively unproductive environmental settings with regard to terrestrial resources, strongly demonstrates

TABLE 14 Storage Practices of Nonequatorial Hunter-Gatherers Living in Settings

with Less Than 100 mm Annual Runoff from Realized Rainfall and Exhibiting Varying Levels of Dependence on Aquatic Resources


Seri 18.3 none none none Karaera 18.1 none none none

Luisefio 15.9 minimal (acorns, drying baskets plum seeds)

Dieguefio 15.8 minimal (acorns, drying baskets seeds)

Kiliwa 15.5 minimal drying ? Tubatulabal 15.1 acorns, seeds, drying granaries,

fish baskets Nomlaki 14.9 acorns, seeds drying granaries Kawaiisu 14.9 seeds, acorns drying ? Washo 14.8 pifion nuts, drying baskets, pits

seeds, fish Yulki proper 14.8 acorns, seeds drying pits, baskets Lake Yokuts 14.7 tule roots, drying inside house

acorns, seeds Wukchumni 14.7 acorns, seeds, drying inside house

salmon Miwok 14.3 acorns, seeds drying granaries Owens Valley 14.3 piiion nuts, drying pits, baskets

Paiute berries, acorns Continued on next page


TABLE 14-Continued


Sinkyone 14.2 acorns, seeds, drying, racks, baskets deer meat smoking

Karok 13.6 acorns, piiion drying, granaries, nuts, salmon, smoking inside house deer, eels

Uintah 13.5 pifion nuts, drying ? seeds

Modoc 13.3 camus, seeds, drying pits fish

Umatilla 13.3 salmon, seeds drying pits (?), inside house

Uncompahgre 13.1 bison, seeds, drying skin bags, racks pifion nuts

Bannock 13.1 bison, roots drying skin bags, racks Agaiduka 13.1 salmon, roots, drying caches, racks

Paiute seeds Harney Valley 13.0 wada, berries, drying pits, inside

Paiute sunflower and house chenopodium seeds

Gosiute 13.0 pifion nuts, drying pits, baskets various seeds

Kutenai 12.7 camus, fish drying pits, inside house

Tenino 12.4 fish, meat from drying racks, caches winter hunt

Shuswap 12.2 fish, camus drying ? Klamath 12.2 fish, seeds drying racks

Flathead 12.1 bison, deer, elk, drying skin bags, racks camus, berries

Dogrib 11.1 caribou, fish drying, racks, caches smoking, freezing

Nabesna 10.9 caribou drying, racks, caches freezing

Yukaghir 10.7 fish, elk drying, racks freezing

Ket 10.5 reindeer, elk, drying, racks fish freezing

Copper Eskimo 9.6 seal, minimal drying, racks, caches fish freezing

Polar Eskimo 8.7 seal, fish freezing racks, caches, cellars

a. "None" refers to storage that does not normally exceed a period of three to five days after procurement.


several interesting points. Storage is not a function of the "Garden of Eden," where resources are abundant and readily obtained. Most of these cases (all in rather poor environments) demonstrate heavy investments in mobility for procurement, labor for processing, and the construction of facilities for pro- tecting the stores. Storage is a tactic related to ensuring against consumption shortfalls during the nongrowing season (winter).

This table displays interesting patterning with regard to the regular variation it exhibits related to effective temperature gradients. We again note that cases with high ET values exhibit no storage of food. Cases between 15.9 and 14.3 primarily store plant foods, although in a few cases aquatic resources can be added to the list. Preparation is uniformly by drying, and stores may be kept in pits (15 percent), above-ground granaries (25 percent), baskets (45 percent), and within the house (10 percent).

Around ET 14.2 we note two changes: terrestrial mammal meat appears as a stored food, and racks appear consistently as a storage facility. Drying con- tinues to be the exclusive preparation tactic. Cases with ET values between 14.2 and 12.2 are generally consistent in that plant foods continue to be a major secondary storage target. Storage facilities become more varied, and there is a distinct shift in modal form. Racks (50 percent) and pits (31.25 percent) are now the dominant facilities, with in-house storage and skin bags (25.0 percent each) close behind. Baskets and caches are also mentioned, while granaries fall off to 6.25 percent.

Freezing appears for the first time in cases with ET values below 12.1, and thereafter the storage of plant foods does not occur. Animal products become the exclusive target for storage. As with the earlier environmentally related suite of cases, the modal character of storage facilities changes; racks now are used in 100 percent of the cases, followed by caches (57 percent), skin bags (14 percent), and prepared cellars (14 percent).

With regard to the types of storage facilities, Table 15 summarizes the data thus far discussed. Corresponding to the shift noted above in the character of the foods stored is a parallel shift around ET 14 in form and use of varied facilities.

Summary From this comparative exercise I think we can conclude that (1) storage is

expected to increase in importance as effective temperature decreases, and (2) there will be a shift in the foods targeted for storage around ET 12-in environments with lower ET values, plant foods will play a small role. Storage is not rare or unique to "complex hunter-gatherers, " and even more important, the meaningful patterning that is evident in the practice of storage strategies is correlated to environmental variables. This correlation is one of the major features of the mobility-subsistence pattern interaction. Storage represents a play-off between the costs of directly procuring food during the winter, when mobility is constrained by the environment, and the costs of expending additional effort during the growing season, either on plants or on animals that are


TABLE 15 Summary of Information on Storage Facilities from 49 Cases Arranged

According to ET Zones Defined in Tables 13 and 14

Percentage of Cases ET Zones

Storage 25-16 15.9-14.3 14.2-12.2 12.1-8.5 Facility (N= 6) (N = 20) (N= 16) (N= 7)

Baskets 0 45.0 12.50 0 Granaries 0 25.0 6.25 0 Pits 0 15.0 31.25 0 In-House 0 10.0 25.00 0 Racks 0 5.0 50.00 100.00 Caches 0 5.0 12.50 57.14 Skin bags 0 0 25.00 14.28 Cellars 0 0 0 14.28

themselves responding to the growing season. Storage does not imply logistical organization (contra Gladkih, Kornietz, and Soffer 1984; Jones 1983). We could cite many examples of moving consumers, as labor, to resources, exploiting more than is needed for immediate consumer needs, and then transporting the "extra" to other places for storage as security against overwintering shortfalls. The simple practice of storage is environmentally conditioned.

On the other hand, the type of labor organization is related to the tactics used to obtain foods regardless of whether they are consumed immediately or destined for stores. We can cite many concrete examples of moving consumers (potential labor) to locations of potential high-bulk yields. If procurement is successful, the high-bulk products are either (1) heavily processed for transport or (2) cached/stored at the point of procurement for subsequent transport when transport costs are less. This strategy, which is not logistical in form, is most common among hunters of herd mammals or very large animals, such as whales.

The archaeological consequences of this strategy are obvious. The greatest concentration of storage facilities is at the point of procurement for bulk resources, a location which in turn is associated with the most portable and hence lowest investment forms of housing. I have made this point before (Binford 1989a) and need only cite the density of storage facilities associated with bison drive sites on the plains of North America (Davis 1983) and the analogous high density of high-investment storage structures associated with spring caribou drive sites among the Nunamiut (Binford 1978). The most spectacular example is the huge number of storage facilities associated with whale beaching sites in the Central Arctic, where, as in the other examples, the housing was in- variably the most ephemeral and portable form (tents), even though the same people built substantial winter houses and were seminomadic. At the locations with substantial housing, storage facilities were few and are not archaeologically


obvious since they tended to be above-ground platforms where racks for freezing or drying meat were the most common facility.

Although storage cannot be seen as a determinant of logistics, the two may be strongly correlated. Logistical organization may be more common among storage-dependent peoples if their stores are inadequate for overwintering since the costs of residential mobility for consumers in these settings are high during the severe winter, substantial investments made in housing must be abandoned, and many dependents (i.e., children and older persons) do not contribute to a reduction in search time, which is the major food-procurement cost in low food density settings.

In summary, logistics pays off in the context of transport costs, considered relative to the security risks of residential mobility. Storage pays off as a simple function of the decreasing length of the growing season, which is frequently correlated with reduced productivity in the habitat.

CONCLUSIONS

Implications for Human Evolution I have examined three variables-mobility, housing, and storage-as they

interact among themselves and with the environment known for hunter-gatherers in the near-modern world. The very strong patterning among these cases is provocative both for understanding the archaeological record and the evidence for human evolution as well as for seeking explanations of variability in systems state conditions, such as "complexity."

The patterning has been presented. Can we tentatively formulate arguments as to how this patterning is related to the time-space processual dynamics we seek to understand regarding the "human condition"?

I. We can expect that in the past as in the present, the regular graded reduction in the length of the growing season and the decreasing productivity of habitats as one moves toward the poles would have strongly favored early appearance of (a) increased dependence on terrestrial mammals, (b) storage as an overwintering tactic, and (c) increased investment in substantial winter housing regardless of the mobility strategies used.

II. Given that early hominids and early populations of fully modem humans were essentially terrestrial animals, we can expect the following dynamics to have accompanied the successful penetration of increasingly high-latitude and continental environments:

A. Increasing dependence on terrestrial mammals in a gradient from equatorial to polar settings, and, accompanying these subsistence shifts,

B. Increasing mobility with its attendant costs to the security of the hominid/ human populations.

As a consequence of these costs, we expect selection to have favored the following techniques and strategies strongly:

1. Techniques for reducing mobility costs, which include ingenious use of local materials for housing, experimentation with the development of transport


technology and the construction of passive facilities (such as traps) for food procurement, and experimentation with means of obtaining additional energy sources for transport, such as domesticated dogs, reindeer, or horses.

2. Strategies for reducing mobility by an expansion of the diet breadth to include aquatic resources and the development of techniques and technology for accessing those resources from a habitat that humans cannot directly pen- etrate.

All else being equal, we can therefore expect historical trends in the archaeological record to demonstrate a chronological and geographical patterning such that systematic use of aquatic resources will have occurred first in the polar and near-polar margins of the human population (environment permitting) and will exhibit a graded temporal and geographical pattern, becoming more recent in time and less common geographically as one moves toward the equator. This pattern will also parallel innovations in (a) transport technology, (b) use of domesticated draft animals, and (c) the pattern of technological elaboration for accessing aquatic resources, as well as reducing the search time for terrestrial game.

The anticipated graded pattern will be "punctuated" and irregular as a function of any geographical and environmental variation that constrains mobility, such as islands, peninsulas, coasts adjacent to major mountain ranges, or major drainages where anadromous fish regularly spawn. In these settings we may anticipate otherwise "early" shifts to aquatic resources, and the attendant technological changes.

III. In continental settings, we can expect transport technology and draft animals to have been highly favored. When such innovations appear, we can expect (a) shifts to the increasing use of highly portable housing materials and (b) a standardization of housing in the direction of portability. From an archaeological perspective these changes may seem to indicate the "disappearance" of more substantial housing and a reduction in the repetitiveness with which winter dwellings are placed in the same locations, given seminomadic land-use patterns. Where aquatic alternatives were possible, we can expect trends in the opposite direction.

IV. Given that "access windows" for penetration of the aquatic biome are less ubiquitous and more reliable than those that characterize the terrestrial biome, as a direct interactive function of the character of the coastal "edge" and the mobility and productive rhythms of the aquatic species present, settlement becomes increasingly "tethered" to those relatively few reliable "access windows," which both natural biogeography dictates and technology renders usable. Thus we can expect (a) that redundancy in settlement location would have increased and (b) that a corresponding shift would have occurred toward housing that is both substantial and nonportable.

Increased densities of archaeological remains can be expected in fewer and fewer places on the landscape as a function of subsistence "tethering"; however, this high density of remains should not be seen as convincing evidence for sedentism or "complexity," since it would be expected even if there was a


seminomadic pattern of land use. Such a pattern could be simply more "an- chored" or "routed" (Graham and Roberts 1986) by increased seasonal dependence on aquatic resources.

All of these changes could be expected to have occurred independently of "density-dependent" selection. They can be anticipated simply as a function of increased mobility as a necessary concomitant of the terrestrial exploitation in more and more polar extensions of the human range.

Implications for the Development of Complexity The above arguments imply that for those who expect directional moves

toward sedentism-broadly defined and statically conceived-the archaeological record may appear to document these shifts (in more substantial housing, increased volume of archaeological debris at specific locations, etc.) but with no prior or linked "density-dependent" causal implications. On the other hand, I have argued that there are strong environmental conditioners for shifts to aquatic resources in areas distant from the equator. Thus both the character of the subsistence base and the mobility strategy seem to respond very ex- plicably to ecological variation.

For those interested in such issues as the origins and sociocultural consequences of agriculture or the origins of social complexity, with the attendant emphasis on inequality, several lessons can be learned from these data. Perhaps a clue rests in the distributions noted in Table 8, where cases of fully sedentary hunter-gatherers arrayed against ET values exhibited a bimodal distribution- one group was concentrated in environments with an effective temperature of 10 and another in environments with effective temperatures of 14 or more. The cases clustered around ET 10 consist of the classic "complex hunter- gatherers" of the American Northwest Coast. I have already suggested that strong selective forces would have favored the shift to aquatic resources quite independently of density-dependent conditions in these environments. Never- theless, these same cases are also noted for their high population densities. This point was recently emphasized by Keeley (1988:395): "The strong association between demography and socioeconomic complexity also suggest that any claims for the prehistoric development of complexity unaccompanied by increases in population pressure are to be treated with extreme skepticism."

This means that in the higher latitudes the high population densities and consequent social complexity are results of reduced mobility and increased subsistence security, a point also supported by Keeley's (1988:393) recent work. That increased subsistence security resulted in intense selection against mobility in settings characterized by low terrestrial productivity should be obvious. Thus, the increased population sizes are a consequence of the increased security that aquatic resources afforded in some settings. Complexity in these settings can be profitably studied in terms of demographic variables quite independent of mobility considerations per se.

On the other hand, for those sedentary cases in environments with ET values of 14 or more, demographic factors are the causes of the increased


costs of mobility itself relative to its returns (see Belovsky 1988 for an explo- ration of these ideas). These cases would fall under the population-packing argument I have previously presented (Binford 1983) and would be expected to move eventually in the direction of intensifying labor inputs to compensate for the decreasing returns from and higher costs of mobility. In short, agriculture is a density-dependent response and can be expected to occur primarily in environments exhibiting temperatures around and just above the earth's mean biotemperature. (Agriculture offers less and less as the growing season becomes shorter and shorter.) Plant and animal domestication is a response to decreasing returns accompanied by increasing costs of mobility brought about by overexploitation and competition for ranges. We can expect that, where possible, aquatic resources will be used first under conditions of demographic packing, and only after substantial periods of time would agriculture appear. On the other hand, in settings with little aquatic potential, rather "quick" moves directly to agriculture can be expected under packed conditions. In either case, demographic forces would be at the root of subsistence changes, whereas in the higher latitudes one can expect analogous subsistence changes in the absence of these initiating demographic forces, as argued above.

I would argue that we are not dealing here with an orthogenetic directional pattern moving toward "complexity" but with one that must be understood in terms of the factors conditioning variation in the organizational characteristics that have been chosen as the topics for research. When one views process in a dynamic mode, the usefulness of "criteria for recognition" for interpreting static systems-state conditions begins to fade. Thick middens can mean different things in different settings; similarly, substantial housing may vary with the ecological setting and may carry no attendant static implications for such phenomena as social "complexity." Each phenomenon must be studied analyt- ically since only the resulting understanding can guide interpretations of archaeological manifestations. If one wishes to make judgments concerning the "levels" of complexity represented, one must understand what variables and interactions condition the variation in the materials being interpreted.

For instance, we have recently been told that a bimodal distribution in house size indicates the presence of larger "corporate groups" who have achieved "differential control over trade or resources, with some individuals assuming important administrative and power roles" (Hayden et al. 1985:183). With perhaps even less justification, Soffer (1985:479) suggests that differences in the size of and degree of labor investment in housing on the central Russian plain represents "public work and monumental architecture." Without any understanding whatsoever of the factors that condition differential investment in housing and different sizes of houses, archaeologists are simply free to make up the past as they go along. We need hardheaded analytical study of the variable characteristics remaining for us to see; only with some understanding of the causes of the observed variation can we make accurate statements about the past and sort out the interactions of variables necessary for the successful building of robust theory.


NOTE

1. The data presented in this paper include work done by myself and various graduate students representing more than fifteen years of hunter-gatherer seminars I have taught at the University of New Mexico. With the help of several work-study students, June- el Piper brought the data together from many disparate sources and prepared the data base file I used in the analyses. My companions on a recent trip to the Soviet Union read an early (and much longer) draft of this paper, which was prepared at the request of the organizers of the Models in Environmental History conference held at Bad Hamburg, West Germany, in February 1989. Their comments, especially those of Clive Gamble, are gratefully acknowledged. Although the data collection and editorial assistance I have received is greatly appreciated, the analyses and interpretations are my responsibility alone.

REFERENCES CITED

Bailey, H.P., 1960, A Method of Determining the Warmth and Temperateness of Climate. Geografiska Annaler 43(1): 1-16.

Belovsky, G.E., 1988, An Optimal Foraging-Based Model of Hunter-Gatherer Pop- ulation Dynamics. Journal of Anthropological Archaeology 7:329-72.

Bender, B., 1978, Gatherer-Hunter to Farmer: A Social Perspective. World Ar- chaeology 10:204-22.

Bender, B., 1981, Gatherer-Hunter Intensification. Pp. 149-57 in Economic Ar- chaeology: Towards an Integration of Ecological and Social Approaches (ed. by A. Sheridan and G.N. Bailey). British Archaeological Reports, International Series 96. Oxford, Eng.

Binford, L.R., 1978, Nunamiut Ethnoarchaeology. New York: Academic Press. Binford, L.R., 1980, Willow Smoke and Dogs' Tails: Hunter-Gatherer Settlement

Systems and Archaeological Site Formation. American Antiquity 45:4-20. Binford, L.R., 1983, In Pursuit of the Past. New York: Thames and Hudson. Binford, L.R., 1989a, Bones for Stones: Considerations of Analogs for Features

Found in the Central Russian Plain. Paper prepared for the joint Soviet-American archaeological symposium on Paleoindian-Upper Paleolithic Adaptations, sponsored by the Subcommission on History and Archaeology, American Council of Learned Societies, and the SSSR Commission on the Humanities and Social Sciences, July 9-23, 1989, Moscow and Leningrad.

Binford, L.R., 1989b, Isolating the Transition to Cultural Adaptations: An Organi- zational Approach. Pp. 18-41 in The Emergence of Modern Humans: Biocultural Ad- aptations in the Later Pleistocene (ed. by E. Trinkaus). Cambridge, Eng.: Cambridge University Press. Reprinted 1989, in Debating Archaeology (by L. R. Binford), pp. 464- 81, New York: Academic Press.

Boas, F., 1888, The Central Eskimo. Pp. 399-669 in Sixth Annual Report of the Bureau of American Ethnology, 1884-1885. Washington, D. C.: Smithsonian Institution.

Brown, J.A., and T.D. Price, 1985, Complex Hunter-Gatherers: Retrospect and Prospect. Pp. 435-42 in Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity (ed. by T.D. Price and J.A. Brown). Orlando, Fla.: Academic Press.

Clark, G.A., 1987, From the Mousterian to the Metal Ages: Long-Term Change in the Human Diet of Northern Spain. Pp. 293-316 in the Pleistocene Old World: Regional Perspectives (ed. by O. Soffer). New York: Plenum Press.


Davis, L.B., ed., 1983, From Microcosm to Macrocosm: Advances in Tipi Ring Investigation and Interpretation. Plains Anthropologist 28(102, pt. 2), Special Memoir 19.

Gallardo, C.R., 1910, Los Onas. Buenos Aires: Cabaut y Cia. Garrod, D.A.E., 1928, Excavation of a Mousterian Rock-Shelter at Devil's Tower,

Gibraltar. The Journal of the Royal Anthropological Institute of Great Britain and Ireland 58. Reprinted, The Bobbs-Merrill Reprint Series in the Social Sciences A-294, Indi- anapolis, Ind.

Gladkih, M.I., N.L. Kornietz, and 0. Soffer, 1984, Mammoth Bone Dwellings on the Russian Plain. Scientific American 251(5): 164-75.

Graham, M., and A. Roberts, 1986, Residentially Constricted Mobility: A Preliminary Investigation of Variability in Settlement Organization. Haliksa'i: UNM Contributions to Anthropology 5:105-16. Department of Anthropology, University of New Mexico, Al- buquerque.

Gronnow, G., M. Meldgaard, and J.B. Nielsen, 1983, Aasivissuit-The Great Sum- mer Camp: Archaeological, Ethnographical, and Zoo-archaeological Studies of a Caribou- Hunting Site in West Greenland. Meddelelser om Gronland, Man and Society 5. Co- penhagen.

Hayden, B., M. Eldridge, A. Eldridge, and A. Cannon, 1985, Complex Hunter- Gatherers in Interior British Columbia. Pp. 181-99 in Prehistoric Hunter-Gatherers: The Emergence of Cultural Complexity (ed. by T.D. Price and J.A. Brown). Orlando, Fla.: Academic Press.

Ingold, T., 1983, The Significance of Storage in Hunting Societies. Man 18:553-71. Jochim, M., 1987, Late Pleistocene Refugia in Europe. Pp. 317-48 in The Pleistocene

Old World: Regional Perspectives (ed. by 0. Soffer). New York: Plenum Press. Jones, K.T., 1983, Forager Archaeology: The Ache of Eastern Paraguay. Pp. 171-

91 in Carnivores, Human Scavengers, and Predators: A Question of Bone Technology (ed. by G.M. LeMoine and A.S. MacEachern). Calgary: Archaeological Association, University of Calgary.

Keeley, L.H., 1988, Hunter-Gatherer Economic Complexity and "Population Pres- sure": A Cross-Cultural Analysis. Journal of Anthropological Archaeology 7:373-411.

Klein, R. G., 1974, Environment and Subsistence of Prehistoric Man in the Southern Cape Province, South Africa. World Archaeology 5:249-84.

Klein, R. G., 1976, The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape Province, South Africa. South African Archaeological Bulletin 31:75-98.

Lee, R.B., 1968, What Hunters Do for a Living, or How to Make Out on Scarce Resources. Pp. 30-48 in Man the Hunter (ed. by R.B. Lee and I. DeVore). Chicago: Aldine.

McBurney, C.B.M., 1967, The Haua Fteah (Cyreneica) and the Stone Age of the South-East Mediterranean. Cambridge, Eng.: Cambridge University Press.

Mathiassen, T., 1928, Material Culture of the Iglulik Eskimos. Report of the Fifth Thule Expedition, 1921-1924, vol. 7, no. 1. Copenhagen: Gyldendalske Boghandel, Nordisk Forlag.

Murdock, G. P., 1967, Ethnographic Atlas. Pittsburgh: University of Pittsburgh Press. O'Connell, J., K. Hawkes, and N. Blurton Jones, 1988, Hadza Hunting, Butchering,

and Bone Transport and Their Archaeological Implications. Journal of Anthropological Research 44:113-61.

Reinhardt, G.A., 1986, The Dwelling as Artifact: Analysis of Ethnographic Eskimo


Dwellings, with Archaeological Implications. Ph.D. diss., University of California, Los Angeles.

Soffer, 0., 1985, The Upper Paleolithic of the Central Russian Plain. New York: Academic Press.

Testart, A., 1982, The Significance of Food Storage among Hunter-Gatherers: Res- idence Patterns, Population Densities, and Social Inequalities. Current Anthropology 23:523-37.

Voigt, E.A., 1973, Stone Age Molluscan Utilization at Klasies River Mouth Caves. South African Journal of Science 69:306-9.

Volman, T.P., 1978, Early Archaeological Evidence for Shellfish Collecting. Science 210:911-13.

BINFORD_Mobility, Housing, And Environment, A Comparative Study

Documents

numbers parentheses

percentages

research journal

moves towardthe

andenvironment

housing environment

anthropological

journal anthropological