G//ana Territorial Organization Author(s): Elizabeth Cashdan Source: Human Ecology, Vol. 12, No. 4 (Dec., 1984), pp. 443-463 Published by: Springer Stable URL: http://www.jstor.org/stable/4602755 Accessed: 15/07/2010 11:40

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=springer.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

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

Springer is collaborating with JSTOR to digitize, preserve and extend access to Human Ecology.

http://www.jstor.org Human Ecology, Vol. 12, No. 4, 1984

G//ana TerritorialOrganization

Elizabeth Cashdan1

This analysis of G//ana territorialorganization shows how land rights are ac- quired and how they affect patterns of land use. Both spatial and social boundaries are discussed. It is shown that the appearance of overlapping spa- tial boundaries is clarified through a focus on the land rights of individuals and a consideration of historical population movements over the region. The discussion of social boundaries shows that, while interterritorial marriage and mobility networks are extensive, increases in property andfoodproduc- tion appear to be associated with a closing of social boundaries (increased social nucleation). This is suggested by greater endogamy and a trend toward patrilineal inheritance of land rights in recent years.

KEY WORDS: territoriality; hunter-gatherers; spatial organization; G//ana.

INTRODUCTION

This paper describes the territorial organization of the G//ana, a predominantly foraging population of the Central Kalahari. Its emphasis is on how land rights are acquired, and how they determine patterns of land use. I will show, first, that this can best be understood by focusing on in- dividual land rights ratherthan on group territories,and by considering histor- ical population movements that underlie the current pattern of land rights. Second, I consider the effects of property and food production on territorial organization. Preliminary data suggest that an increased dependence on food production is associated with greater endogamy and a trend toward patrilineal inheritance of land rights. Kalahariforagers are well known for having a "fluid,"flexibly organized social and territorial organization. Territorial boundaries are said to over-

'Department of Anthropology, University of Pittsburgh, Pittsburgh, Pennsylvania 15260. 443

0300-7839/84/1200-0443$03.50/0 ? 1984 Plenum Publishing Corporation 444 Cashdan

lap, and band composition changes frequently as individuals visit kinsmen living in other regions. This characterization, which became popular with the publication of the Man the Hunter symposium (Lee and DeVore, 1968), is accurate, but the concept of "flexibility"really tells us very little. We know from ethnographic descriptions of San land rights that what appears as "flux" on the level of the band arises from a cross-cutting network of individually based land rights and kin ties (see Lee, 1972, 1979; Wiessner, 1977; Yellen and Harpending, 1972). From the perspective of the individual, it is clear that mobility, land rights, and kin ties are not simply "fluid," but are pat- terned and predictable. The implications of this for territorial boundaries have not been well understood, since bands have been considered the territory- holding unit. Do the San have flexible, overlapping territories? When one looks at G//ana territories from the perspective of the group, a picture emerges of overlapping territories shared by populations living in different areas. It will be shown below that the fuzziness of these overlapping territorial boundaries disappears when one focuses on the land rights of individuals. The demise of the patrilineal band as the model for hunter-gatherer social organization (Lee and DeVore, 1968; Lee, 1972), and the accompany- ing emphasis on flux and flexibility, has led to a picture of foragers moving freely about the landscape as ecological conditions dictate. For example, Yellen and Harpending (1972) have argued that "there is much in Bushman culture about land ownership, land inheritance, and living rights, but these serve to justify these underlying ecological processes and should be inter- preted in that light. [They] give each individual the right to live in a number of widely scattered areas, and allow him to move freely from one to another much as he pleases." It is certainly true that hunter-gatherers,especially those in highly variable environments such as the Kalahari, have available social mechanisms that allow them access to resources in often widely separate regions. At the same time, however, these social mechanisms also operate as constraints on mobility and land use. If the norms about land ownership, land inheritance, and so forth were simply justifications that allowed in- dividuals to move freely about the region, one might wonder why they would be needed in the first place. Cultural norms concerning land rights are only necessary because there is often a "conflict of interest"; while it may be advantageous for each individual to have access to as many places as possible, it may also be advantageous, especially under conditions of resource scarcity, to keep others out. As I will show, the system of G//ana land rights acts as a constraint in determining settlement patterns, just as it also is manipulated to enable individuals to move into areas of their choice. The following discussion begins with a brief ethnographic account of G//ana subsistence. The analysis of territorial organization itself is in two G//ana Territorial Organization 445 parts, the first of which considers spatial boundaries and the second, social boundaries. This distinction was made by Lee (1972, p. 126), who pointed out that "the existence of a group and a space necessarily implies the exis- tence of two kinds of boundaries: social and spatial. A social boundary can be measuredaccording to how open or closed the group is .... A spatialbound- ary can be measured along the dimension of overlap/nonoverlap." A closed social boundary, in Lee's terminology, is equivalent to what Yellen and Har- pending (1972) call a "nucleated"social organization, nucleation being a meas- ure of the degree to which social relationships form discrete clusters. In the first part of the discussion of G//ana territoriality, I consider whether the G//ana do in fact have overlapping spatial (territorial) boundaries. I argue that while a static picture that focuses on the boundaries of groups makes it appear that they do, a dynamic approach that focuses on the land rights of individuals suggests otherwise. In the second part, I turn to the interregional social ties that determine the degree of social nucleation. After a discussion of marriage and mobility networks, I consider whether food production and property are associated with increased nucleation in the G//ana population.

ETHNOGRAPHIC BACKGROUND

The G//ana, who number about 800, live in the nothern and eastern parts of 's Central Kalahari Game Reserve (see Fig. 1). They are a physicallyand culturallymixed population, tracing their origins both to Bak- galagadi pastoralists from southern Botswana and to the hunting and gather- ing San whom these Bakgalagadi married when they moved north into the Reserve over 200 years ago. G//ana territory overlaps that of the G/wi, although the 209 people discussed here live in the northeastern part of the region, some 90-120 km east of the G/wi area. The G//ana, like the G/wi, must cope with an environment characterizedby low and extremely variable rainfall, and by an absence of permanentstanding water. Mean annual rainfall in the regionis about 350 mm per year, and the coefficientof variationis 50-60% (Pike, 1971; see Silberbauer, 1972, 1981, and Tanaka, 1976, 1980, for descriptionsof the Central Reserve environment). The G//ana differ from the G/wi, however, in supplementing their diet of bush foods and game with small-scale cultivation and goat husbandry. The crops cultivated by the G//ana include marotsi (a domesticated melon similar to but larger than the wild tsama), beans, and a small amount of maize. Like the tsama melon, marotsi acts as a naturally storable form of moisture and it is grown and used by the G//ana chiefly for this purpose. The cultivated marotsi provides the G//ana with moisture when the nearby pans become dry, thereby allowing them to remain at one location during 446 Cashdan

ZAMBIA ANGOLA G AN

NAMI B IA pkga REGION REIO Moppi * I \Lake Xau I WI G/IANA REGION G/V \S REGION * Central Kalahari \ I Game Reserve BOTSWANA

0 100 200km 1f7 G//na rei Fig. 1. Map of Botswana, showing location of G//ana region.

the rains and, if the crop has been a good one, for a short period after the rains have ended. During the dry season, the G//ana generally move to tsa- ma melon areas, where they subsist on wild foods; but they may use the garden site as a base camp even during this period to store bush foods (especially dried game meat and grewia berries), cultivated crops (if any remain from the harvest), and skins and furs. If the water-bearing wild melons (tsama and mokaphane) are depleted before the rains return, the G//ana split up into smaller groups and use the various species of water root for moisture, or they may move to the Botletle River, 115 km to the north, and stay there until the rains return. There are no cattle in the Central Reserve because of the lack of water, but the G//ana do keep goats, donkeys, and a few horses in the Reserve, all of which are "watered"during the dry months on wild tsama mel- ons. The horses are used as aids in hunting, the donkeys chiefly for trans- port (including transport of bush foods when gathering), and the goats as G//ana Territorial Organization 447

Table I. Population of Bands in Study Area Location Adults Children Total Molapo 26 17 43 Moiapo 23 20 43 Molapo 11 7 18 Molapo 6 8 14 Molapo 4 5 9 Totwe 32 26 58 Kunadie 13 11 24 Goa 6 4 10

Total 121 98 219 a source of skins, milk, and meat. During my fieldwork, goats were rarely killed solely for food, but it would be reasonable to expect that they are an important source of meat when game in the camp is scarce. During my fieldwork in 1976-1977, there were six base camp locations that had been continuously occupied for at least a generation, and at least two new locations that were being used for the first time that year. My field- work was limited to the two northernmost of the main settlements, Molapo and Totwe.2 Molapo was occupied by 127 people in five camps, and Totwe was occupied by one large camp of 58 people. The remaining 24 people in the sample moved during my fieldwork from Kunadie (near Molapo) to a newly formed base camp at Meno. An additional ten people moved from the Botletle River in 1977 to form a new base camp at Goa (see Fig. 2 and Table I).

SPATIAL BOUNDARIES

One of my first attempts to understand G//ana spatial organization en- tailed learning the location of boundaries between the different base camp locations. As boundaries were not marked, I thought that I could make a useful map by ascertaining which tsama melon areas belonged to, and were regularly used by, the people from Molapo, from Totwe, and so on. The results of these efforts appear in Fig. 2, which shows the melon areas to which people from the different base camps claimed access without having to ask permission. As the map indicates, many of the tsama areas are "shared"by

'For the benefit of readers unfamiliar with Khoisan pronunciation, the marks for the "click" consonants have been eliminated in the text and figures except for names which have already appeared in the literature,e.g., G/wi and G//ana. The correct pronunciation of all place names is given in the Appendix. 448 Cashdan

RAKOPS.

/7 ' XMOPIPI

LKE XAU

- - ./ J /2\

CENTRAL / //Mopani

KALAHARI I /

GAME RESERVE a . * ,,~MOLAPO7

\ ONaroko . *Gabishi j .*/ ,Gage, OTOTWEr / -~~ *''.~)Santokwe 7

/,-- --7MTSU RU Y.METSEAMANONG

* BASE CAMP / \ *Dry Season Camp / \

0 10 20 30 40 50km

Fig. 2. Territorial "boundaries" around base-camp locations (see text for discussion). people from different base camp locations, and may be used by more than one band at a time if the melon crop is abundant. For example, Santlokwe, which has been used at various times by people from Molapo, Totwe, and Metseamanong, was reportedly used by people from all three areas in 1972. Similarly, people from Totwe joined people from Metseamanong at Diti in 1974, after they had finished the melons in their own area. It should be not- ed that the shared access discussed here arose not from interterritorial visit- ing among kinsmen (an important part of G//ana spatial organization to be considered later) but from shared rights to the territory. G//ana Territorial Organization 449

While not inaccurate, the pattern of overlapping territorial boundaries pictured in Fig. 2 is an oversimplification. This became apparent when I real- ized that a "shared"tsama area is shared not simply by all the surrounding base camps, but only by people who can claim it as part of their lefatshe (territory). An understanding of G//ana spatial organization, therefore, re- quires an understandingof how individuals use and claim rights to a lefatshe.

LEFATSHE

Lefatshe (pl. mafatshe) is a Setswana and Sekgalagadi word meaning country or territory;like the !Kung word n!ore, it means both place in a gener- al sense and home place or territory. A G//ana who claims an area as his lefatshe is saying, in effect, that he has the right to stay there and use the resources in that area without asking permission. Although the place name given as an individual's lefatshe is invariably the location of a base camp, it is understood that nearby hunting and gathering areas are included. As with other Kalahari foragers, territory boundaries are not marked or defend- ed, and "borders"between named places (including places within a territo- ry) seem to be indistinct. There is a clear patrilineal bias in lefatshe inheritance among the G//ana. Table II presents these data in a form comparable to Lee's (1979, p. 338) data on inheritance of n!ores among the !Kung; a comparison of the two groups shows the G//ana to be more patrilineal, at least for men. Although patrilineal inheritance of lefatshe among the G//ana is both the

Table II. Inheritance of Lefatshe (Territory) among Adultsa Adult males Adult females Lefatshe (n = 28) (n = 26) Total Same as father, different than mother 15 8 23 Same as mother, different than father 1 3 4 Same as both mother and father 9 10 19 Different than both mother and father 3 5 8 'A few adults claimed to have more than one lefatshe; in this table, in- dividuals are recorded as having the same lefatshe as a parent if the two have any lefatshe in common. The patrilineal trend looks similar, but slightly less marked, if one considers only the first lefatshe men- tioned. 450 Cashdan norm and the ideology, there are exceptions. A woman will sometimes claim the lefatshe of her husband if it is their new place of residence. Factors other than paternal inheritance may be used when the lefatshe of one's father is no longer the location of a base camp, since in such cases a person will want to strengthen rights to his current place of residence. A settler of a formerly unoccupied area may claim it as his lefatshe, and birthplace or long-term residence may also be used. Although there is a norm and ideology of patrilineal inheritance, the process is manipulable and allows a person the flexibility, within limits, to strengthen territorial claims to a place of his choosing. Returning to Fig. 2, then, we can say that while the foraging areas near a base camp belong to people who claim the area as their lefatshe, bound- aries overlap because individuals with claims to the "shared areas" live at different base camp locations. This is a consequence of historical popula- tion movements over the region, a topic to which we now turn.

Long-Term Mobility and Movement of Settlements

There are currently six locations in the Central Reserve that are the sites of long-standing G//ana base camps, but their number and location are sub- ject to change through a continual process of site abandonment and forma- tion of new settlements. While base camps at some of the larger pans, such as Metseamanong and Molapo, have been in existence for several genera- tions, others, like the two formed in 1976 at Meno and Goa, are new. Figure 3 illustrates this long-term mobility by showing how long the elders of the larger base camps in the northeast region have been at their current loca- tion, and where they came from. Base camp moves of this sort enable the population to adjust to changes in resources and in local population densities. The most dramatic examples of the latter are radical disruptions in population due to disease, a reason sometimes given for having moved a settlement. A man whose parents had lived at a large base camp at Santlokwe, for example, said that his parents left the area when an epidemic killed the people there. Another man, speak- ing of his former base camp at Gabishi, told a similar story, saying "those people all died, and their children are scattered." Today there are no base camps at either Santlokwe or Gabishi, although both are good tsama melon areas and are regularly used for foraging by former residents and their chil- dren. Presumably, in cases such as these, a group would move in with kins- men in another area when its population dropped below a viable level. Drought has provoked other changes in settlement location. For ex- ample, one man said that his grandparents moved to Metseamanong because of "arguments with the sun," and more recently a small band abandoned G//ana Territorial Organization 451

RAKOPS. NAN10e

MOPIPI *MABE TONGOROHA LK XAU

GWACHWVA

e

Ttabe CENTRAL Tobe KALAHARI *GOA GAME RESERVE MOLAPO Kunadie b f

eMENO TOWE antokwef //QDiti. a Kanye,e \ Nao KgakamaIi *DUMTSURU \a METSEAMANONG

*CURRENT BASE CAMP * Abandoned Base Camp

0 10 20 30 40 50km.

Fig. 3. Movement of base camps. This map indicates only a few of the major moves, where the elders of a camp moved with a number of people to a new location, abandoning the former area as a base camp. Individual changes of residence take place frequently and can- not be included here. The most recent moves to each of the major camps is marked by a letter, indicating approximately when the move was made (as of February, 1977): (a) 120 years ago, (b) 50 years, (c) 22 years, (d) 5 years, (e) 1 year, and (f) 1 year.

Totabe because of lack of water and moved north to the Botletle River. The gradual lowering of the water table around the area of Metseamanong was reportedly responsible for the initial move north into the Molapo area, some 120-150 years ago. The process continues today; a band from Totwe moved their base camp near Molapo in 1971 for the explicit reason that they wanted access to the larger Molapo pan. Current and contemplated moves of this 452 Cashdan sort suggest that people wishing to move to a new area either move near kins- men in a desired location (or attempt to forge such a kin tie), or form a base camp at an unoccupied area, perhaps one within the region of their current lefatshe. It is difficult to know in most cases whether such moves are in fact caused by environmental changes, or whether the perception of water short- age is caused by a local increase in population with no change in water avail- ability. Clearly, however, population movements of this sort are a way of adjusting the spatial distribution of the population to the available and shift- ing resources of the region.

Abandoned Settlements and Shared Melon Areas

A consideration of changes in settlement location allows us to under- stand better the "overlapping" spatial boundaries shown in Fig. 2. A loca- tion that has been abandoned as a base camp is still used for foraging and dry-season camps, and the rights to the area remain with the former resi- dents and their children. Because these former residents may have moved to widely separate locations, however, the abandoned base camp location then becomes a "shared" resource area, shared not simply by all who live in surrounding locations, but by the former residents and their descendants, who retain their rights to the area. The shared areas pictured in Fig. 2, there- fore, are not simply fluid boundaries between base camps but are used and "owned" by former residents who have moved elsewhere. The tsama melon area at Gabishi is a case in point. Gabishi was at one time the location of a large base camp, and when it was abandoned some 50 years ago because of disease, the survivors moved northeast to Molapo or south to Dumtsuru, with people at both areas retainingtheir rightsto the region around the old base camp. The migrants to Molapo formed a new base camp, joining other camps that had been at Molapo for many years. To some ex- tent, Gabishi and the surroundingareas southwest of Molapo are now regard- ed as part of the Molapo lefatshe. As one Molapo resident put it, "The parents of Gaorapelwe [the head of this band] were from Gabishi, but after that dis- ease killed the people they came here, and now all those places are all our places." It is also clear, however, that Gaorapelwe and his kinsmen retain special claims to the Gabishi area that are recognized by the other Molapo residents. Since people who wish to use an area outside their lefatshe must ask permission to do so, one indicator of such claims is the person one asks for permission. When I asked other Molapo residents whom people from another area would ask if they wished to use Gabishi or the areas near it, I was told that they would ask permission of Gaorapelwe, "becausehe is from there." If, on the other hand, these people wished to camp at Mopani, a tsama area north of Molapo, they would "ask Kanyo or Reletaka [elders of G//ana Territorial Organization 453 the two oldest Molapo base camps] because it is their lefatshe." It appears that the old mafatshe also influence, to some extent, the patterns of use. During the 1976 dry season, Reletaka's band was camped at Mopani and the bands of Gaorapelwe and Kanyo were camped at Gabishi. While in a sense all the areas surrounding Molapo are considered to be part of the lefatshe of Molapo, a long-range perspective and consideration of how these areas are used, and of whom permission is asked, reveals that it is the claims of lefatshe, and not strictly a base camp location, that is the basis of ter- ritorial rights. Ties of coresidence, which join neighboring base camps such as those at Molapo, eventually diminish these distinctions, as does intermarriage be- tween the camps and the fact that lefatshe claims do not extend back more than about two generations. As with other aspects of G//ana territorial or- ganization, the longevity of claims to abandoned base camp areas appears to be manipulable. One Molapo resident, for example, justified his claim to the abandoned base camp location of Diti by saying "my grandfather's lefatshe was Diti. Now Moeti [a man from Totwe] and I are the owners of that place, and when there are tsama melons there we all use it." In fact, neither this man nor anyone else at Molapo customarily uses Diti, although it is possible that this rather distant claim (together with charm, kin ties, and an abundant melon crop) might enable this man to do so if he wished.

Land Rights and Territorial Exclusion

While lefatshe claims are manipulable, and kin ties allow people access to areas outside their lefatshe, this type of system can nonetheless act as a means of territorialexclusion. It does so both because people are denied access to areas where they have neither kinsmen nor lefatshe rights, and because use of the resources at a kinsman's lefatshe is subject to permission and can be prevented if circumstances dictate. These mechanisms are subtle, and not easy to observe. Since people will not normally go to a place where they have no rights of access, one rare- ly witnesses overt exclusion resulting from territorial trespass. Some of the most graphic examples of hostility and territorial exclusion in the literature on the San are those that occurred when the ethnographer Heinz brought members of a !Ko band into an area where they had no land rights and no kinsmen, or where new employment opportunities brought together people from different areas (Heinz, 1972). Since the !Ko would not normally move into the territory of strangers, these examples can be regarded as "experi- ments" that would not normally take place in a traditional setting. They are, however, experiments that prove a point, namely, that sanctions for trespass exist and are used when needed. Kalahari foragers rarely move into an area 454 Cashdan unless they feel they will be welcome, but since their unwillingness stems from the existence of these sanctions, the situation may appropriately be viewed as territoriality. One band from Totwe, for example, wanted access to the larger Mola- po pan, but their lefatshe was Totwe and they had no close kinsmen at Mola- po. Only after beginning a marriage negotiation with a Molapo family did they feel comfortable moving to Kunadie, a pan near Molapo. Even this did not last long; after a few years the man who had married into the Molapo family moved into his wife's compound, while the rest of the band left for Meno, a place nearer Totwe that was part of their traditional lefatshe. It is also of interest that in 1976, while they were still based at Kunadie, they chose to make their temporary dry season camp at Meno rather than at the locations used by the nearby Molapo bands. These decisions suggest that set- tlement and mobility are constrained by lefatshe claims, even in the absence of overt territorial exclusion. The subject of interregional kin ties and mobility will be discussed in the following section, but it can be noted here that gaining access to resources in a kinsman's lefatshe is not automatic, although this can also be difficult to see, since permission is not overtly withheld. Silberbauer (1981, p. 141) explains that among the G/wi a discussion takes place concerning expected visitors, and consensus is reached before they arrive. Although "permission is never actually withheld . .. unwelcome visitors are given permission to re- main but are later eased out of the band." Wiessner (1980, pp. 772-773) describes the same process among the !Kung. This means of exclusion re- minds me of the way in which an American commune reportedly rid itself of unwelcome visitors. According to one of the members, such visitors were always given permission to stay, but would later be "vibed out" of the com- mune. Even where permission is not overtly withheld, then, unwelcome out- siders can be made to feel so unwelcome that they will leave of their own accord.

SOCIAL BOUNDARIES

In this section, I consider the interregional social ties that permit ac- cess to areas outside a person's lefatshe. Because this kind of reciprocal ter- ritorial access is a common feature of what I have called "social boundary defense" (Cashdan, 1983), I begin with a brief summary of this type of ter- ritoriality and note the circumstances under which it can be expected to oc- cur. I then present data showing the patterning and frequency of interterritorial visiting among the G//ana. Finally, I present data suggesting that these interband networks become less extensive and more nucleated as foragers acquire property and depend to a greater extent on food production. G//ana Territorial Organization 455

Social Boundary Defense

According to the economic defensibility theory of territoriality (see Brown, 1964; Davies, 1978; Dyson-Hudson and Smith, 1978; Cashdan, 1983), territoriality is least adaptive, given the existence of competition, where resources are scarce and unpredictable. This is so because territories must be larger where resources are scarce and unpredictable, and large territories are more costly to defend. Although the G//ana live in such an environ- ment, they clearly have a system of land rights which enables them to con- trol and limit access to resources. The solution to this apparent paradox is that the economic defensibility arguments assume that defense costs increase with territory size, an assumption that is valid only where people defend resources by patrolling and defending the boundaries of the territorial space itself. This, the G//ana do not do. I have argued elsewhere (Cashdan, 1983) that as resources become sparse or unpredictable and ranges become larger, making this type of "perimeterdefense" uneconomical, foragers have another option available; they can control access to resources by controlling access to the social group inhabiting the area ("social boundary defense"). Because defense costs are unrelated to the size of the territory, this strategy is well suited to foragers in environments with sparse or unpredictable resources, where ranges are typically large. For Australian foragers (Peterson, 1975, p. 60), this system of territorial defense entails making "acceptance into the local land using group a prelimi- nary requirement for using the resources in its territory." The system in the Kalahari is similar: a person wishing to use the resources in an area where he does not have land rights, i.e., lefatshe or n!ore claims, does so by visit- ing kinsmen in that area, and asking their permission to stay. Since the system requires the cooperation of the visitors in asking per- mission, it is reasonable to ask how it can work. Why not simply trespass unannounced? Understanding when there is a payoff to such a strategy will shed light on why we find this type of territoriality among the G//ana. The reasons have been discussed at length in Cashdan (1983) but can be summa- rized here. First, the admittance of outsiders to the social group typically involves the expectation that access will be reciprocated when circumstances permit, and the benefits of such "reciprocal altruism" (Trivers, 1971) will be high in patchy environments or environments in which areas of relative abundance vary from year to year. In these environments, where reciprocity is important, the short-term benefit of cheating (trespassing unannounced) is outweighed not only by the risk of expulsion if caught (a short-term cost) but by the risk of losing one's "good name" (a long-term cost). Because hu- mans exchange information about cheaters, the probability of getting caught need not be high to make it a dangerous strategy. Second, cooperation ena- bles outsiders to obtain information from residents about the resources in 456 Cashdan a territory. This should be particularly valuable in large territories with un- predictable resources, since such environments will be more difficult to ex- ploit unaided. Finally, the risk of discovery may increase when territories are so large that trespassers must remain in the territory while they exploit the resource. In the Kalahari, for example, where territories are very large, an outsider wishing to use the water or melons in another territory would have to stay there while he used them; the transport costs of carrying them home would be prohibitive. The longer such an individual remained in the territory, the greater would be his risk of discovery. Each of these condi- tions makes it more advantageous for a potential intruder to seek social rights of access than to trespass unannounced. The circumstances favoring this type of territorialsystem, then, occur where territoriesare large and resources are sparse and unpredictable, exactly the conditions one finds in the Central Kalahari.

G//ana Visiting Patterns

Moves to areas outside a person's lefatshe are frequent in both dry and rainy seasons, as Table III indicates. During the year in which these data were obtained, about 16% of G//ana women and 20% of G//ana men were visit- ing at any given time, although the motive for visits, and the frequency, varies throughout the year. Social visiting between bands takes place at all times of the year; interregional mobility for "business" reasons, e.g., trading and selling meat and skins at the Botletle River, is particularly frequent in the early dry season, and visits for the purpose of obtaining moisture are most

Table III. Visiting within and outside the Central Reservea Visiting within Visiting outside Visiting Kwikum or Central Reserve Central Reserveb Botletle R. onlyc July Males (n = 91) 4Wo 23W0o 16Wo Females (n = 98) 10%0 50o 5Wo September Males (n = 97) 2% 10% 6W0o Females (n = 107) 7WMo 707o 7?10 January Males (n = 98) 10% 12% 5%0 Females (n = 105) 14% 5%07 4%01 aData for Table III are based on censuses taken at three times during 1976-1977. July is in the early dry season, Septemberin the late dry season, and Januaryin the rainy season. bIncludes Botletle River, Ghanzi, Johannesburg, Kweneng, and Kwikum. cLocations visited for permanent water. G//ana Territorial Organization 457

Table IV. Moves to the Botletle River in the Late Dry Season, 1975 vs. 1976a Totwe 1975 Totwe 1976 Molapo & Totwe 1976 Population sampled/total adult population 23/29 27/32 71/121 Percent of sample at Botletle River 57?70 22Wo 32Wo 'Data in Table IV were obtained by interviewingadults about their mobility in the preced- ing year. Data for 1975 were obtained in July 1976, while data for 1976 were obtained in January 1977. The figures refer only to adults and indicate whether an individual went to the Botletle River at any time during the preceding dry season.

frequent in the late dry season, when tsama melons and other moisture sources in the Central Reserve become scarce. The Botletle River, some 115 km to the north, is the nearest source of permanent water for the G//ana from Molapo and Totwe, so visits to kins- men at the river in the late dry season are particularly important. The num- ber of people who visit relatives at the river in the dry season varies considerably, and depends on the degree of drought. In 1976, there was abun- dant tsama melon, and comparatively few people left the Reserve for water. In 1975, on the other hand, there was a scarcity of tsama, at least at Totwe, and the new rains did not begin until quite late. As a result, about half the population of Totwe left for the river during the late dry season of that year, twice as many as left in 1976 (see Table IV). Visits for the purpose of obtaining water or tsama melons also take place between bands in different areas of the Central Reserve itself. For ex- ample, in mid-January of 1977 (ostensibly a rainy-season month) rain was still extremely sparse in the Totwe area, and about a third of that group were away visiting relatives at Molapo, where there had been a great deal of rain. Table III describes the relative magnitude of visiting within and outside of the Central Reserve. (Readers wishing more information on G//ana mobili- ty should see Cashdan, 1984 which contains maps showing the areas visited during each of the seasonal censuses, and the number of people at each lo- cation.) The sex difference in mobility apparent in Table III deserves mention. These data indicate that women do more visiting within the Central Reserve itself, whereas men do more of the long-distance visiting to places outside the Reserve. This difference is chiefly due to the fact that much of the long- distance visiting outside the Reserve is motivated by business concerns that are primarily the province of the men, such as selling meat and skins and buying tobacco and other articles. Most of the visits made by men in the early dry season (July) to places outside the Reserve were of this nature. As Table III shows, the difference between the sexes in long-distance visiting 458 Cashdan largely disappears when one considers only those places usually visited as sources of permanent water, since women as well as men will make long- distance moves for this purpose. Interterritorial visiting among G//ana, as among other Kalahari hunter-gatherers, is nearly always based on some kinship tie, and the G//ana appreciate the value of marriage in providing alliances over the region to increase one's options in times of scarcity. The marriage links between the Central Reserve and the Botletle River can be seen in this light, as can many marriages within the Reserve itself. The degree to which Kalahari hunter- gatherers marry people from outside their own local area, therefore, is an indicator of the importance placed on interregional visiting. It is interesting to note, in this regard, that 74% of G//ana marriage partners were born in a different location from their spouse.'

Changing Patterns of Nucleation

Interregional visiting and exogamy are measures of social "nucleation" and can be expected to change with the degree of dependence on food produc- tion. Yellen and Harpending (1972, p. 248) suggest that "nucleated organi- zation is probably optimum for conserving property rights and that non-nucleated or anucleate organization is optimum for rapid information flow." They present supporting data showing that the !Kung are less endoga- mous and less inbred (hence, less nucleated) than settled agricultural peo- ples. Peterson (1980) has suggested, similarly, that the increasing importance of cultivable land among Philippine (Agta and Negrito) forager-farmers is associated with "a growing concern with consolidating access to that land through territorially endogamous marriage." Another reason for expecting greater social nucleation to accompany the transition from foraging to food production is the substitution of storage for mobility as a means of risk reduction. Interterritorial visiting enables foragers such as the G//ana to cope with unpredictablevariation in resources. Storage is another way of doing the same thing. The accumulation and storage of food associated with a dependence on agriculture, therefore, should make interterritorial visiting less necessary. Animal husbandry, because it "stores" meat on the hoof, should have a similar effect. The reduction in mobility that usually accompanies cultivation, furthermore, should decrease the trans-

3By "location" I mean the region of the home base, not the particular camp site. For example, spouses born in different Molapo camps would be considered to have been born at the same location. G//ana Territorial Organization 459

Table V. Birthplaces of Spouses Wives with same Average distance Wives with same birthplace as husband between birthplaces lefatshe as husbanda Age group (n = 39) (n = 39) (n = 34) 16-34 41Wo 38 km 73Wo 35+ 14Wo 56 km 37% aSince women sometimes take the lefatshe of their husband when it is their new place of residence, spouses are more likely to report the same lefatshe than the same birthplace. port costs associated with using stored resources. We can therefore expect storage to become a more economical means of risk reduction as hunter- gatherers adopt food production, making interterritorial marriage and mo- bility networks less necessary. The G//ana buffer fluctuations in resources not only by interterritori- al visiting but also by storage of food and wealth items, e.g., furs and skins, at reoccupied base camps and by ownership of property such as livestock and water-storage containers (Cashdan, 1980). The value of water-storage containers in this environment is obvious, but furs and skins, which are some- times sold at the Botletle River, are a source of cash and, consequently, a source of additional long-term security. There is some evidence, albeit anec- dotal, of an increase in livestock and other property items (most important- ly, 55-gallon drums used for water storage) in this area in recent years. Because they help to reduce the risk associated with local scarcity, we might expect an increase in these property items to be associated with a decrease in the need for widespread marriage partnerships. A comparison of G//ana in different age groups suggests that this has indeed taken place. As Table V shows, 41 %oof younger wives were born in the same areas as their husbands, whereas this was true for only 140%oof older wives. Data on the distance between spouses' birthplaces show a similar pat- tern, the average being 56 km for older couples and 38 km for younger ones. These data suggest a decrease in geographically widespread marriages in re- cent years and may reflect a decrease in the importance of interterritorialmo- bility that is based on such marriageties. Table V also shows a similartrend with spouses' territories, younger couples reporting the same lefatshe more fre- quently than older couples. Another indicator of greater social nucleation is the trend toward in- creased lineality in lefatshe inheritance. Table VI shows that older adults have acquired their lefatshe more or less bilaterally, while inheritance of lefatshe among the younger adults is more strongly patrilineal. This observation is based on a small sample and so should be treated with caution. However, it is consistent with the fact that territorial inheritance among the G//ana appears to be more patrilineal than it is among the less wealthy !Kung. 460 Cashdan

Table VI. Inheritance of Lefatshe (Territory) Age group Lefatshe same as father Lefatshe same as mother 16-25 lOOWo 27% 26-45 77Wo 38%o 46 + 65% 59Wo n = 54 adults

To summarize, there is evidence of an increase in social nucleation in this population, as measuredby increased endogamy and unilineality. It seems likely that this is related to a recent increase in property accumulation, which affects social nucleation in two ways. The first is through greater concern over property rights. The work of Peterson and others suggests that concern over property rights may be associated with a trend toward endogamy, and the association of unilineal inheritance with property and food production (pastoralism and horticulture) has been explained in a similar fashion.4 The second is through the use of storage and property to buffer variation in resources. As discussed previously, storage and reciprocal territorial access are two means of reducing the risk of local scarcity. It is reasonable to sup- pose that an increase in the former will make the latter less necessary. Be- cause interterritorialaccess is based on geographically widespread marriages, a trend toward endogamy should be expected when such access is no longer needed. The trend toward unilineal inheritance of territory can be interpret- ed in a similar fashion. Bilateral inheritance of territory gives an individual a choice of locations to claim as a residence, and access rights to the territo- ries of two sets of kin. Limiting inheritance of territory to one parent can be expected to reduce residential and mobility options. I suggest that Kala- hari hunter-gatherers cannot affort this restriction unless there are other sources of protection against resource variability. Food storage, water drums, livestock, and other sources of wealth can afford a measure of such pro- tection.

4Whileendogamy and unilineality may be associated with a general concern over property rights, it is unlikely that among the G//ana they are a response to scarcity of land. The G//ana densi- ty of .04 person per km2 is low, even lower than that of the neighboring G/wi (.07 person per km2; Silberbauer, 1981, p. 192), who practice little or no cultivation and appear to have fewer animals and less property than the G//ana. G//ana Territorial Organization 461

SUMMARY

In this paper, I have discussed the nature of G//ana land rights and territorial boundaries, both spatial and social. I have shown that an under- standing of spatial boundaries is enhanced by a historical perspective and a focus on how land rights are acquired and held by individuals. A synchronic perspective of spatial boundaries shows that gathering areas are shared by nearby home-base locations. These overlapping boundaries are not simply an example of flux and flexibility, however, but are a result of long-term changes in the location of home-base settlements in the G//ana region. A location that has been abandoned as a base camp is still used for foraging, and the rights to the area stay with the former residents and their children. Because former residents may have moved to base camps in widely separate locations, the abandoned region then becomes a "shared" resource area, shared not simply by all those in surrounding locations, but by the former residents and their descendants, who retain their rights to the area. Spatial and temporal variation in rainfall is considerable in this region, and areas of relative abundance change markedly from year to year. For this reason, reciprocal territorial access is an important means of coping with fluctuations in the availability of resources. The extent of such visiting, and the interterritorial marriage ties that facilitate it, indicate the nature of "so- cial boundaries" (Lee, 1972). It was shown that interterritorialvisiting is fre- quent among the G//ana, particularly during times of drought. Although visiting and marriage networks are extensive, the available data indicate a trend toward greater nucleation, as measured by increased endogamy and increased lineality of territorialinheritance. It was suggested that these trends are related to an increase in property in recent years.

ACKNOWLEDGMENTS

Fieldwork with the G//ana was supported by NSF Grant SOC 75-02253. The discussion of spatial boundaries and visiting patterns is based on a report to the Government of Botswana that I wrote in 1977. I wish to thank Pat Draper, Henry Harpending, Nancy Howell, Richard Lee, Nicolas Peterson, and John Pfeiffer for their comments on that report. The data on changing patterns of nucleation were analyzed with the support of a Faculty Research Grant from the Faculty of Arts and Sciences, University of Pittsburgh. Len Plotnicov's editorial suggestions improved the paper, and I am grateful for his help. 462 Cashdan

REFERENCES

Brown, J. L. (1964). The evolution of diversity in avian territorial systems. Wilson Bulletin 76: 160-169. Cashdan, E. (1980). Egalitarianism among hunters and gatherers. American Anthropologist 82: 116-120. Cashdan, E. (1983). Territoriality among human foragers: Ecological models and an applica- tion to four Bushman groups. Current Anthropology 24: 47-66. Cashdan, E. (1984). The effects of food production on mobility in the central Kalahari. In Clark, J. D., and Brandt, S. A. (eds.), From Hunters to Farmers. The Causes and consequences of Food Production in Africa. University of California Press, Berkeley. Davies N. B. (1978). Ecological questions about territorialbehaviour. In Krebs, J. R., and Da- vies, N. B. (eds.), Behavioural Ecology: An Evolutionary Approach. Sinauer, Sunder- land, Massachusetts, pp. 317-350. Dyson-Hudson, R., and Smith, E. A. (1978). Human territoriality: An ecological reassessment. American Anthropologist 80: 21-41. Heinz, H. J. (1972). Territoriality among the Bushmen in general and the !Ko in particular. Anthropos 67: 405-416. Lee, R. B. (1972). !Kung spatial organization: An ecological and historical perspective. Hu- man Ecology 1: 125-147. Lee, R. B. (1979). The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge University Press, Cambridge. Lee, R. B., and DeVore, I. (eds.). (1968). Man the Hunter. Aldine, Chicago. Peterson, J. (1980). Food sharing and directed marriage choice. Paper presented at the Second International Conference on Hunting and Gathering Societies, Quebec. Peterson, N. (1975). Hunter-gatherer territoriality: The perspective from Australia. American Anthropologist 77: 53-68. Pike, J. (1971). Rainfall over Botswana. Botswana Notes and Records (special edition no. 1). The Botswana Society, , pp. 69-76. Silberbauer,G. B. (1972). The G/wi Bushmen. In Bicchier, M. G. (ed.), Hunters and Gatherers Today. Holt, Rinehart and Winston, New York, pp. 271-325. Silberbauer, G. B. (1981). Hunter and Habitat in the Central KalahariDesert. Cambridge Univer- sity Press, Cambridge. Tanaka, J. (1976). Subsistence ecology of central Kalahari San. In Lee, R. B., and DeVore, I. (eds.), Kalahari Hunter-Gatherers: Studies of the !Kung San and Their Neighbors. Harvard University Press, Cambridge, pp. 98-119. Tanaka, J. (1980). The San Hunter-Gatherers of the Kalahari: A Study in Ecological Anthro- pology (translated by D. Hughes). University of Tokyo Press, Tokyo. Trivers, R. L. (1971). The Evolution of reciprocalaltruism. QuarterlyReview of Biology 46: 35-57. Wiessner, P. (1977). Hxaro: A Regional System of Reciprocityfor Reducing Risk among the !Kung San. Unpublished Ph.D. dissertation, University of Michigan. Wiessner, P. (1980). History and continuity in !Kung San reciprocalrelationships. Paper presented at the Second International Conference on Hunting and Gathering Societies, Quebec. Yellen, J., and Harpending, H. (1972). Hunter-gatherer populations and archaeological infer- ence. World Archaeology 4: 244-253. G//ana Territorial Organization 463

APPENDIX. The Following List Gives the Correct Pronounciation of Place Names Noted in the Text and Figuresa Name in text Correct spelling Totwe /"o # we Kunadie Kun//adie Goa G/oa Gabishi G/abishi Dumtsuru G/umtsuru Xau G//au Gage G/ag/e Naroko N/aroko Nowe N!oe Kanyere //'anyere Nao /hao Gwachwa G/wachwa Tongoroha /ong/oroha Tobe /"obe Totabe G/o ? abe Kwikum !'wi!um 'A"/" indicates the dental click, "?" the alveolar click, "!" the palatal click, and "//" the lateral click. A "g" in- dicates that the click is voiced. An apostrophe indicates a glottal stop, and a double apostrophe indicates a pharyngeal fricative. Names not in the table are pronounced as spelled.