Competition between foragers and food producers on the Botletli River, Botswana

The immigration of food-producing groups into areas occupied by hunters and gatherers must have been a common occurrence in prehistory. How were the hunter-gatherers affected by this? I describe here two groups of Kalahari Basarwa ('Bushmen'), one living along the flood plain of the lower Botletli river, the other occupying the savanna a short distance away from the river. These two groups differed in subsistence and social organisation and were affected by immigrant herders and farmers in strikingly different ways. Today the Basarwa of the flood plain are wealthy cattle owners, whereas those of the savanna are poor and have few or no cattle. How and why did the two groups respond so differently to the same competitive threat?

Africa, 56 (3), 1986 COMPETITION BETWEEN FORAGERS AND FOOD-PRODUCERS ON THE BOTLETLI RIVER, BOTSWANA1 Elizabeth Cashdan The immigration of food-producing groups into areas occupied by hunters and gatherers must have been a common occurrence in prehistory. How were the hunter-gatherers affected by this? I describe here two groups of Kalahari Basarwa (‘Bushmen'), one living along the flood plain of the lower Botletli river, the other occupying the savanna a short distance away from the river. These two groups differed in subsistence and social organisation and were affected by immigrant herders and farmers in strikingly different ways. Today the Basarwa of the flood plain are wealthy cattle owners, whereas those of the savanna are poor and have few or no cattle. How and why did the two groups respond so differently to the same competitive threat? I begin with a brief description of the Botletli environment. I then turn to the aboriginal populations of the area and discuss their traditional economy and early history. Because our knowledge of riverine foragers is so scanty, I have given this section extra emphasis. Finally I discuss the interactions that took place between these Basarwa and immigrant Bantu-speaking peoples, showing the changes in economy and land use that took place during the nineteenth and early twentieth centuries. The ethnohistorical information in the following account is based on my interviews in 1975-77 with forty Botletli river residents, on the accounts of early explorers and, to a lesser extent, on material from the Botswana National Archives and current ethnographic sources. Interviews with people on the Nata river provided additional information about the distribution, history and early economy of Botswana's riverine foragers. Although this article is primarily historical, I have included a few summary remarks about current patterns of land use. These are based on a larger ethnographic study which is described in detail elsewhere (Cashdan, 1979). The reader may assume throughout that if I make an assertion that is neither widely known nor specifically attributed, the information derives from my own interviews and observations. background: the river and its traditional inhabitants The river region The Botletli is in some respects reminiscent of the Nile. While not a long river, the Botletli's ultimate source lies far away, some 1100 km to the northwest in the Angolan highlands. Some of this water eventually empties into the Okavango delta, a vast inland swamp in northwest Botswana (see map), and a small fraction flows eastward out of the delta and becomes the Botletli. The gradient of the Botletli is nearly flat and the current is slow. Because of both the gradient and the distance from its source, the water does not even reach the lower Botletli until the beginning of the dry season in June. It rises to its highest point late in the dry season and then begins to fall around October or November, at the start of the summer rains. Since residents can plough in the moist river banks300 FORAGERS v. FOOD PRODUCERS IN BOTSWANA as the water level drops during the rainy season, the Botletli, like the Nile, creates a fertile strip of land in an otherwise dry and barren country. The surrounding region gets only 400-450 mm of rain per year (Pike, 1971) and cultivation away from the river is chancy and unrewarding. As I have documented elsewhere (Cashdan, 1979), fields in the flood plain are far more productive than those in the savanna. Most of the water in the lower Botletli now flows into the Mopipi dam, which was built to provide a reliable source of water for the Orapa diamond mine, 50 km to the east. Formerly the Botletli spilled south to fill a large lake, Lake //Gau (spelled ‘Xau' and previously ‘Dow' on Botswana maps), the remainder of the water feeding into the Makgadikgadi Salt Pans to the northeast. Because the water is diverted to the dam, Lake //Gau is now little more than a dry, grassy lake bed used chiefly for grazing. During the nineteenth and much of the present century, however, Lake //Gau was apparently a large swamp (Chapman, 1971; Schwartz, 1926; Jeffares, 1938). The dam has also affected water levels in the Botletli river itself. Aerial photos and informants' reports indicate that before the dam was built the river usually emptied completely in the rainy season, allowing people to plough directly in the moist riverbed. In recent years water has remained in the river year round, and farmers have been forced to plough higher up on the banks. During my fieldwork the fields of many farmers remained partly or entirely under water, thus exacerbating the already acute competition for the valuable flood-plain land. It would be incorrect, however, to suppose that water levels were constant before the dam was built. Reports of early travellers and feasibility studies done in connection with the building of the dam suggest that dramatic changes in flow levels existed long before the dam was built, and have been the rule for this region at least since the beginning of this century. The ethnohistorical work of Jeffares (1938) and the hydrological studies of Gibb (1969) suggest that there have been several periods in the first half of this century during which there was no flow at all to the lower part of the river. My informants report that one such dry period in the middle of this century was followed by the emigration of most of the downriver farmers to places as far as 130 km away. Unusually high water levels can also cause economic hardship by flooding fields. The dam has caused such problems in the past decade, but problems of a similar nature occurred long before the dam was built. In the 1850s Chapman found that the river was coming down and overflowing its banks, and inundating the native cornfields, which were infested with geese and ducks. The natives here . . . have either a prolific crop or none; the latter happens when their lands remain inundated throughout the season. When this occurs for two or three consecutive years, famine results. [1971: 161] Most of the Basarwa in this area live either along the river itself or at cattle posts to the south and around Lake //Gau. There is intermarriage and trade with G//ana foragers living some 70 km to the south in the Central Reserve (Cashdan, 1984), whereas such ties are few north of the river. The current pattern of settle­ment may be due in part to political factors. The G//ana area south of the river is a game reserve, and is one of the few remaining areas in Botswana where grazing is forbidden and only traditional forms of hunting are permitted. The distribution of salty soil has probably also influenced settlement patterns. The soils of theApproximate distribution of Khoisan-speaking Negro Basarwa Ethnic names TSIBI 302 FORAGERS v. FOOD PRODUCERS IN BOTSWANA Botletli area overlie lacustrine deposits that were at one time an extension of the Makgadikgadi Salt Pans. The north side of the river is closer to these pans, and the vegetation there is sparse and salt-adapted (Bawden and Stobbs, 1963). Foragers of the Botletli river The former foragers of the Botletli-Lake //Gau area belong to a population about which little is known. The conventional wisdom about indigenous Southern African peoples divides them into (a) San, identified by their distinctive ‘Khoisan' physical type, ‘click' languages and by the fact that they are, or were until recently, hunter-gatherers, and (b) Bantu-speaking Negroes. The traditional inhabitants of the Botletli, however, like others found in a broad zone across the northern and eastern edges of the Kalahari, are an anomaly in such a classification. They are dark-skinned and apparently Negro (see Jenkins and Nurse, 1977; Nurse and Jenkins, 1977; Nurse et al., 1985; Chasko et al., 1979, for serogenetic data on related groups), but they speak a click language related to that of the Nharo, G/wi and Khoikhoi (Westphal, 1963, 1971; Kohler, 1971, 1981). Hunter-gatherers of this description include not only the traditional inhabitants of the Botletli river, but also those of the Okavango region in northwest Botswana and those living around the Makgadikgadi Pans and along the Nata river in northeast Botswana. As I have shown elsewhere (Cashdan, 1986), these groups also share a number of cultural characteristics which set them apart from the Khoisan foragers of the desert. Some of these differences stem from the fact that they inhabit a wetter region than the desert San. The region is traversed by numerous rivers, and rainfall is greater and less variable than it is to the south and west, where the desert San live (Pike, 1971). Differences between these foragers and the desert San may also be due in part to aspects of their early history (see below). Because they are physically Negro, not Khoisan, it is inappropriate to refer to them as ‘San'. In this article I will refer to these foragers simply as ‘Basarwa'. This Setswana word is used in Botswana to refer to all traditional hunter-gatherers. It includes the well-known desert San groups (IKung, G/wi, etc.), as well as the groups discussed here. The origin of these Basarwa is something of a mystery. Although they see themselves (and are seen by their Bantu-speaking neighbours) as having been hunter-gatherers and fishermen traditionally, there is some evidence that they may have had an early experience with pastoralism. I have discussed the early history of these groups more fully elsewhere (Cashdan, 1986), but a brief sketch will be helpful in understanding the later history that is the subject of this paper. Most of the early writers (i.e. Schapera, 1930; Bleek and Duggan-Cronin, 1942) and those working in the Okavango region (i.e. Seiner, 1910; Gusinde, 1966; Clark, 1951), have assumed that the groups discussed here were typical Khoisan foragers who developed Negro physical characteristics as a result of intermarriage with neighbouring Bantu-speaking peoples. Serogenetic data casts doubt on this hypothesis, however. The work of Nurse, Jenkins and their colleagues indicates that at least some of the Khoisan-speaking Negro hunter-gatherers are fully Negro populations, with no more Khoisan admixture than many Southern African Bantu­speaking groups. One possibility is that they are the descendants of a pre-iron Age Negro hunting and gathering population. Westphal (1963) notes that legends of a ‘primitive black people' who had no chiefs and lived by hunting and gathering are widespread among Bantu-speaking groups in this region. Such a populationFORAGERS v. FOOD PRODUCERS IN BOTSWANA 303 may be related to the Twa, the traditional Negro hunter-gatherers of South-Central Africa (see Estermann, 1976; also Clark, 1950; Rangeley, 1963; Miller, 1969). A second possibility, suggested by Nurse and Jenkins (1977) and Nurse et al. (1985), is that they originated from herding and farming peoples who lost their cattle and were forced to hunt and gather. Examples of pastoral peoples who adopted hunting and gathering when they fell on hard times are well known in Southern Africa (for example, the Bakgalagadi and the Khoikhoi), so such a scenario is not improbable. The hypothesis is given support by the finding that many words relating to cattle keeping and cultivation in their Khoe languages have proto-Khoe roots, and are probably derived from central Sudanic languages (Ehret, 1974, 1982; Vossen, 1984). Recent archaeological evidence also supports the antiquity of pastoralism in Southern Africa. Archaeological evidence of Iron Age sites with cattle remains have been found at several sites in northern Botswana which date to the first millennium ad (Denbow and Wilmsen, 1983; Denbow, 1984). These sites could have been inhabited by early Khoikhoi (Khoisan) pastoralists or by the hunter-gatherers discussed here. Although this article is concerned with more recent history, it should be remembered that these people may have had a history of herding and possibly agriculture at some time in the distant past, and that their ‘traditional' way of life has probably changed considerably over the centuries. This is especially true for the Bateti, who may also have possessed cattle during the early part of the nineteenth century (Livingstone, 1857) and perhaps earlier (Denbow, 1984). Because the reliability of information from oral history diminishes over time, I am beginning this account at the time of the first written records, made by Livingstone and Chapman in the 1850s. At this time there were apparently no cattle in the lower Botletli region and people lived by hunting, gathering, fishing, and cultivation. Major Botletli river Basarwa groups Name Totem Location Taxpayers in ward* //Kanikhoe Crocodile River 32 Bateti Ostrich Mopipi, river 182 in three wards (31, 56, 95) Tshaiti Hippo Mopipi, river 15 G/wiokhoe Cow //Gumtsao, Libunyani (none reported) //Gorokhoe Hare Kaongabe, /Tauhadom (none reported) Danisan Dog North of river (none reported) * Data on taxpayers are from Schapera (1952), compiled from information in the 1946 census. I am including them here because they give some idea of population and settlement earlier in the century. Schapera lists five ‘Teti' wards in the Bamangwato tribal territory, one each for those with the hippo and crocodile totems, and three for those with the ostrich totem. The Bamangwato tribal territory covers the region discussed in this paper, but not the upriver stretch of the river or the Okavango region. The Basarwa of the lower Botletli river belong to five principal groups, each having a traditional area and totem. Two of these groups, the G/wiokhoe and the //Gorokhoe (also known as /Xaiokhoe), live in the savanna to the south of Lake //Gau.2 The remaining three groups, the Bateti (also called Deti), the Tshaiti and the //Kanikhoe (also known as Tshumakhoe) live along the Botletli river flood plain. In order to minimise the confusion of speaking about so many different304 FORAGERS v. FOOD PRODUCERS IN BOTSWANA groups, I will sometimes refer to the two groups living near Lake //Gau as ‘savanna foragers' and the groups living along the Botletli flood plain as ‘riverine foragers'. The groups are listed with their totems and locations in the table above. The discussion will focus on the lower Botletli river, the area around Lake //Gau. Traditional economy I begin with a brief sketch of what life was like on the lower Botletli in the 1850s, when Europeans first entered the area. The Botletli is a meandering, reedy river and probably looked much as it does today, except that in the nineteenth century it was teeming with many species of wildlife that are either no longer found there or are found only in greatly reduced numbers. Bantu-speaking farmers (Bakhurutsi and Kalanga) lived in the area, and the Bakhurutsi chief, who lived at Lake //Gau, claimed political control over the region. Although the political prominence of the Bakhurutsi led the early explorers to describe them in disproportionate detail, we know that the Bateti, who lived with them along the flood plain, were considered to be the traditional inhabitants and ‘owners' of the area.3 While the Bateti resembled desert foragers in some respects, such as language, they differed in others. They had headmen with real power (although they were subordinate to the Bakhurutsi chief) and they cultivated gardens in the flood plain. Most of their food came from fish and game, however, and the river provided such an abundance of both that they lived in settled communities year round. Other Basarwa (‘savanna foragers') could be found away from the river and around the swampy Lake //Gau. Although dark-skinned like the riverine groups (including Bateti), they were in most respects similar to the desert San (!Kung, G/wi, !Ko). In 1852 Chapman (1971: 58) found a camp near Lake //Gau that consisted of ‘seven or eight little huts about three feet high, just a few sticks put together and grass thrown over it. ... A few bows and arrows hung in the trees. A small broken earthen pot stood by a smouldering fire, a calabash of water, a few tortoise shells for dishes, and the shells of innumerable bitter melons and a few old bones lay scattered about'. They seem to have been as politically egalitarian as the desert San, and their dependence on bush foods away from the river made them more mobile than the Bateti. A visitor to the river would have been most likely to see them in the dry season, since they returned to the Lake //Gau area when the seasonal pools away from the river became dry. At other times of the year it would have been easy to miss them entirely, and the paucity of ethnohistorical information on these savanna foragers suggests that this may often have been the case. While the savanna foragers probably did a little fishing in Lake //Gau, they depended primarily on game and bush foods. Because game travelled in a predictable fashion to the river, good use was made of pitfalls and hunting hides or blinds. Other aspects of hunting and gathering technology resembled those of the desert San. In the following sections I will amplify and document some of the points made in this sketch. Readers not interested in the details or documentation may skip the remainder of this first section without losing the thread of the argument. Hunting technology: Hunting technology was affected by the presence of permanent water, since many species remained near the river and others returned to it regularly to drink. Several early travellers (Dornan, 1925; Chapman, 1971; Livingstone, 1961) mention the use of pitfalls which were dug along animal pathsFORAGERS v. FOOD PRODUCERS IN BOTSWANA 305 near the banks of major rivers. My Bateti informants say they trapped lechwe and tsessebe in pitfalls dug in the reedbeds by the river banks, and then killed them with a spear. Other animals (buffalo, wildebeest, kudu, springbok) were also trapped in this fashion away from the reedbeds. Livingstone and Chapman make frequent mention of the many pitfalls studding the area around the Botletli river and Lake //Gau, and complain repeatedly about their oxen and men falling into them. It is usually not clear from these accounts, however, whether the pitfalls were dug primarily by Basarwa or Bakhurutsi. The pitfalls in this area are described as being 3 metres deep, were often narrowed at the bottom or had sharpened stakes planted in them, and were sometimes arranged in pairs so that an animal falling into the first plunged forward and became trapped in the second (Livingstone, 1857; Campbell, 1976). Like the desert San hunters, those in the Botletli area used fire as a hunting aid: ‘near the river the natives burn the grass, to let the young grass spring up and entice the game near the pitfalls' (Chapman, 1971: 168). The hunting implements of the Lake //Gau groups were similar to those used by the desert San. My informants say that bows and poisoned arrows were regularly used (as the scene quoted from Chapman suggests). Some informants also mentioned the use of dogs and spears, and rope snares were used to capture springbok. Much of the hunting around Lake //Gau was apparently done from hides (blinds). The hunters would hide in the area of brush and trees surrounding the lake and kill the animals after they had drunk and were returning to the desert. They also made hides near some of the seasonal pools away from the river and waited for game in a similar fashion. Ambush hunting from hides was also practised by the Nata river Basarwa and the IKung San (Crowell and Hitchcock, 1978). However, ethnographic accounts of the desert San groups (!Kung, G/wi) make little mention of this form of hunting; presumably it is of less importance where game need little or no standing water, since animal paths are less predictable. Gathering: Information on traditional gathered foods and techniques is scanty, but the Lake //Gau groups currently use many of the wild foods collected by other desert San, and presumably did so in the past. The tribes that lived along the river itself used the root of the reed Prionium serratum (tsitla) as a staple (Livingstone, 1961; Chapman, 1971), and the bulb, stalk and flower of the waterlily Nymphaea caerulea (tswii, or /wz) (Campbell, 1976). Both plants are used today along the lower Botletli. Fishing: Fishing was important to all the northern Kalahari riverine foragers and was practised along the Botletli river by ad 800-900 (Denbow and Wilmsen, 1983) and probably much earlier (Denbow and Campbell, personal communication.) It is, of course, imposible to tell from archaeological evidence which populations were responsible for the fishing remains at this early date, and evidence showing Khoi-type pottery remains at Late Stone Age sites along the Botletli river (Denbow and Wilmsen, 1983) suggests that the fishermen may have been early Khoikhoi. In any event, however, oral histories and written sources emphasise the more recent importance of fishing along the Botletli river and in the nearby Okavango region.306 FORAGERS v. FOOD PRODUCERS IN BOTSWANA The Bateti are widely known as a fishing people, although today fishing has largely been supplanted by herding. The technology of the Bateti was similar to that of the Bayei (an upriver Bantu group) and may, in some cases, have been borrowed from them. Stone fishing weirs were observed in the Bateti area by several early visitors (Livingstone, 1960; Chapman, 1971; Dornan, 1925), and my informants also report other techniques, including reed weirs, fish traps, fish spears and woven nets. Dornan (1925) mentions that fishing along the Botletli was also done by poisoning a fenced-off area of the river, by spearing and by the use of baskets, which a row of men, sitting in the river, would hold in front of them. The traps and weirs described by my informants were generally similar to published descriptions of //Kanikhoe fishing technology in the Okavango region (see Cowley, 1968; Seiner, 1910; Heinz, 1969; Gusinde, 1966; Campbell, 1976). The foragers who lived around Lake //Gau did not depend on fish to the same extent as these riverine groups, but some of my informants report that they fished with wire spears when water in the lake was ankle- to knee-deep. Neither weirs nor most of the traps are found in the region today, and the traditional grass nets have been replaced by commercial ones. Mobility: It appears that riverine foragers throughout Botswana were more sedentary and had smaller ranges than the desert San. This is not surprising, since they lived in an area of greater rainfall (hence greater productivity) and permanent water. The Bateti and //Kanikhoe whom I observed along the flood plain live year round in permanent villages and reportedly did so in the ‘old days' as well. The savanna foragers of Lake //Gau moved seasonally, but kept a home base by the lake. During the rainy season they would build a camp by the seasonal pools south of Lake //Gau, and would return to the lake in the dry season when the pools became dry. Lake //Gau was a year-round source of water, since water could be obtained from wells in the lake bed when the lake itself was dry. Today both abundant wild foods and good grazing conditions prompt this seasonal move away from the river, and presumably the first of these was an important factor even before the savanna foragers began caring for cattle. A similar sedentarism or home-base sedentarism exists on the Nata river. My Nata river informants report that in the ‘old days' (before they began caring for cattle) they would either remain near the river or would move to nearby pools during the rainy season and return to the river when the pools became dry. This is similar to the .pattern I observed in 1975-77, although today most residents do not leave the Nata even in the rainy season, and those who leave do so to graze cattle. Even the Ganade, who lived in the savanna northwest of the Nata river, say they kept a home base by the major pans in their territory. The sedentarism that I found is corroborated by accounts of foragers in the Okavango region. Seiner (1910) reports that the //Kanikhoe who lived along the Okavango river and delta were relatively sedentary, occupying dry season villages in the swamps and moving out to the edge of the savanna during the wet season, when the swamps became uninhabitable. Cowley (1968) agrees that the //Kanikhoe of the Okavango move seasonally, but argues that they move out of the swamps in the dry season, when the water level in the river and swamps reaches its height. This latter interpretation is probably correct.4 The Bateti and //Kanikhoe of the lower Botletli had no need for villages away from the river since it remained habitable throughout the year. Permanent villages are also reported for theFORAGERS v. FOOD PRODUCERS IN BOTSWANA 307 Basarwa living along the Ramokgwebana river in 1873 (Dornan, 1925) and the Ganisha riverbed (Seiner, 1910). Range size: Territory (or range) size appears to have been small, as one would expect given the more abundant and predictable resource base of these foragers (see Dyson-Hudson and Smith, 1978; Cashdan, 1983). Seiner (1910) reports that the hunting territories of the //Kanikhoe in the Okavango region were smaller than those of the desert San, extending only 3Q-35 km into the savanna. The small territory size is also stressed by Stigand (1923: 411). He states that because the //Kanikhoe were not allowed to fish outside their own district, ‘as guides . . . they are useless outside their little radius, as a rule. In the swamps you have continually ... to hunt out generally reluctant fresh pilots every 10 to 15 miles.' Indirect evidence suggests that Bateti territories were also small, with each head­man in the lower-river area controlling a strip of river land averaging 16-21 km in length. According to one informant, different headmen controlled the areas from Mmadikola to /Tumo (about 33 km), ITumo to Daokudi (about 15 km), Daokudi to /Tadi (about 5 km) and /Tadi to Phorokwe (about 9 km). The area between Daokudi and /Tadi was given by the Bateti to the Tshaiti, and a second informant did not recognise this as a separate territory. If a single headman controlled the stretch from Daokudi to Phorokwe, the territories would have been roughly equal and slightly larger. I do not know how far away from the river these territories extended, but since the Bateti did not go into the savanna to hunt, there being ample game along the river itself, the distance was probably very small. The areas claimed by the Lake IIGau groups are larger and extend far south into the savanna. I am using the word ‘territory' loosely here, since it is difficult to know whether these areas were traditionally defended, and to what degree residents had exclusive use of them. Political organisation and distribution The first known inhabitants of the Botletli flood plain were the //Kanikhoe (also called Tshumakhoe). Because the //Kanikhoe of the Okavango are comparatively well known, it should be noted here that the two groups of //Kanikhoe are probably unrelated. They have different totems, and the local history of the Okavango //Kanikhoe suggests that they are a branch of the Kxoe, and hence did not come from the Botletli river area (Kohler, 1971; Keuthmann, personal communication). Since //Kanikhoe simply means ‘river people', it is reasonable that different groups inhabiting the river areas would be called by the same name. Although more sedentary than the desert San, and possibly more territorial (see Stigand's comments above), the //Kanikhoe along the lower Botletli appear to have been politically egalitarian, lacking headmen and other formal political institutions. This was not the case for the Bateti, the other traditional residents of the Botletli river.5 Although similar to the //Kanikhoe in appearance and language, the Bateti had headmen with some political authority, and regarded the //Kanikhoe as politically subordinate to them. My informants spoke of the //Kanikhoe as having been ‘servants' of the Bateti and ‘like Basarwa to them' (see also Schapera, 1952: 83). Few individuals on the river today claim to be //Kanikhoe, perhaps because of this low social status. Interestingly, a similar situation has been reported for the Okavango region. The Gomayi (Gumahi), a Khoe-speaking group now living on the banks of the Okavango river (Heinz, 1969; Westphal, 1962), are also308 FORAGERS v. FOOD PRODUCERS IN BOTSWANA reported to have had chiefs, and to have viewed the //Kanikhoe of the Okavango as subordinate (Kohler, 1971; Stigand, 1923; Cowley, 1968). It is possible that the Gumahi are a branch of Bateti. Cowley's ethnohistorical information (1968: 207-8) indicates that the Gumahi migrated to the swamps from the Botletli river and settled on the Okavango river ‘where they subjugated such Tannekhwe as they encountered'. It is difficult to reconstruct the exact powers of traditional Bateti headmen because most of the references to local chiefs by early explorers are to those of the Bakhurutsi. By this time (1850s) the Bateti headmen apparently deferred to the principal Bakhurutsi chief; Schwartz (1926: 534) says of Tsapo, the Bakhurutsi chief who lived at Lake //Gau, that ‘all the various tribes round about put themselves under his protection.' As far as I could determine, the Bateti headmen did not have an economic redistributive function, nor did they organise group labour or activities. One informant volunteered that the Bateti chiefs tried cases and would impose simple punishments. However, the clearest evidence for their power is that they had the ability to allocate (and deny) flood-plain land to newcomers. The information concerning this was corroborated by many informants of different ethnic groups. Residents of the Botletli river say that ‘the Bateti are the owners of this place' and ‘this place belongs to the Bateti'. These statements reflect the Bateti's early political control of the river as well as their claim to be residents of long standing. This control is indicated by the fact that the various groups who later came to the region (including the Tshaiti, Bakhurutsi and Kalanga) did not simply move in but instead asked the Bateti headmen for a place to stay. The allocation of land to incoming groups appears to have been more than simply a formality. Both Kalanga and Bateti informants say that the farmland allocated to the Kalanga was insufficient for their needs, and the Kalanga therefore bought grain from the Bateti in exchange for cattle (see below). The Bateti are said to have controlled the entire river area from Mopipi to Rakops (and perhaps farther upriver), although at the time these other groups arrived the Bateti were only farming in a small part of this region. As discussed previously, different Bateti headmen had control over different parts of the river. The Tshaiti are the third major Basarwa group now living on the lower Botletli flood plain. They came to the area at least a hundred years ago from Thabatshukudu (between the Makgadikgadi Pans), probably coming initially from the Wankie district of Zimbabwe. The Tshaiti say they asked the Bateti for land and were given places along the river to live (I/oa) and to plough (/tadi). The river area was not crowded then and no favours were asked in return for the land. The //Kanikhoe, Bateti and Tshaiti have different totems and different group names, but today all three are often loosely referred to as ‘Bateti'. There are also a few isolated individuals belonging to other totems who regard themselves as Bateti; as one of them explained, ‘We came here, married with the Bateti, became mixed with them, and called ourselves Bateti'. Current similarities in settlement and subsistence are probably responsible for the tendency today to blur ethnic distinctions between these groups while strengthening the distinction between them and the savanna foragers of the lake shore. Most of the foragers around Lake //Gau that I have been calling ‘savanna foragers' belong to two groups: the //Gorokhoe (also known as /Xaiokhoe), who live on the western side of the lake, and the G/wiokhoe, who live on the easternFORAGERS v. FOOD PRODUCERS IN BOTSWANA 309 side. Like the flood-plain groups, each of the Lake //Gau groups has its own totem. Most river Basarwa totem groups are localised, in the sense that each is associated with a traditional area. This is also the case around Lake //Gau. Although there is some individual movement, the //Gorokhoe and G/wiokhoe are associated with the areas I have indicated and have lived there as long as anyone can remember. There is virtually no visiting or intermarriage between the //Gorokhoe and G/wiokhoe, in spite of their being only 15 or 20 km apart. Kin ties between them and the riverine groups (especially the Tshaiti) exist, but are not numerous. Most of the contact between flood-plain and lake-shore residents today takes place in the context of trade, in which maize grown in the flood plain is exchanged for berries, milk and firewood obtained around the lake shore (see Cashdan, 1979). Some kin ties also link the //Gorokhoe to the G//ana, a foraging population living some 70 km to the south in the Central Reserve. G//ana and Domkhoe San visit the Lake //Gau area during the dry season, when moisture in the Central Reserve is scarce, and a few have moved to the cattleposts around Lake //Gau for several years at a time (Cashdan, 1984). This is not a new pattern and is similar to that described by Passarge (1907) for the Dukwe and Tserekwe San, who live to the west. He reports that these groups live on the Nghabe and Botletli rivers during the driest months, moving south to the sand fields when the rains return. INTERACTIONS WITH BANTU PASTORALISTS The Bantu-speaking groups who came to the Botletli area in the nineteenth and twentieth centuries affected density and land use along the river and the lake shore. The effect on the savanna foragers was similar to such encounters in other parts of Botswana, but the interaction with the Bateti took a very different form. In this section I discuss these two historical trajectories, beginning with some background remarks on interactions between Basarwa and Bantu-speaking groups in Botswana as a whole. Overview of Basarzoa-Bantu interactions Bantu-speaking herders and farmers affected Botswana's hunter-gatherers both economically and politically. Much of this was indirect, in the form of competition from cattle that diminished the local supply of game. This economic competition was extremely important in the Botletli area and will be discussed in more detail below. More direct effects stemmed from the centralised political organisation and the military superiority of the Tswana and other Bantu-speaking groups. These effects varied from place to place and from tribe to tribe, but included the requirement (irregularly observed) that Basarwa pay a ‘tribute' of game skins and other goods to the Tswana chief (Mohr, 1876; Dornan, 1925; Chapman, 1971; Campbell, 1976), the capture of Basarwa for slaves (Selous, 1893; Dornan, 1925), mutual hostility arising from the occasional Basarwa practice of killing cattle, and the sometimes violent retribution of the cattle owners (Dornan, 1925; Hermans, 1977). Relations became less harsh under the chieftainship of Khama (paramount chief of the Bamangwato until 1923), who prohibited the sale of Basarwa as slaves. Even after this time, however, there is evidence of a proprietary feeling toward the Basarwa. In a 1924 letter to the British Resident Commissioner, the next Bamangwato chief argued that the Basarwa who were living on the Nata river310 FORAGERS v. FOOD PRODUCERS IN BOTSWANA were his ‘hereditary servants'. This deft phrase left the British wondering whether he meant that he had the power to collect tribute from them or the right to their labour, i.e. slavery, and led them to issue warnings against the latter interpre­tation. The issue is still of current concern: Hitchcock (1980) notes that he found Basarwa working for no pay in 1977-78, although labour without recompense has been illegal since the 1930s. Environmental changes brought about by large herds of cattle have also had a significant effect on Botswana's hunter-gatherer populations. Cattle have changed the composition of grasses and other flora in ways detrimental to many species of game, and have caused serious overgrazing around the limited sources of water. Basarwa in Botswana, and in Southern Africa generally, responded by killing the cattle that had supplanted the game (Khama, 1935; Dornan, 1925; Hermans, 1977). It was therefore to the advantage of the cattle owners to encourage Basarwa to settle on their cattleposts and become stock herders. Basarwa had long been used as hunters by these groups, but their use as herders was more recent (Mackenzie, 1883; Khama, 1935; Hermans, 1977; Hitchcock, 1982; Lee, 1979). The depletion of game, together with the imposition of strict game laws, resulted in an increasing number of Basarwa availing themselves of this option (see Hitchcock, 1982). Effects of pastoralism on the Botletli environment The influx of Bantu food producers and their cattle caused a substantial reduction in game along the Botletli, as it had elsewhere in Botswana. Information I gathered confirms the accounts of early explorers that previously the Botletli area was abundant in a great many species of game, most of which are today found only in small numbers or not at all. Livingstone (1857 and 1961) reports seeing elephant, rhinoceros, buffalo, lechwe and sitatunga along the Botletli, none of which are found along the lower Botletli today. Many of these animals had also been seen by my own informants, which suggests that their demise in this region was fairly recent. It appears that much of the decrease in game along the Botletli has been a direct result of competition from the increasing number of cattle in the area. In a discussion of ecological changes that have taken place along the Botletli river and around the southern periphery of the Okavango delta, Campbell (1976: 11-12) concludes that ‘the introduction of stock has certainly been responsible for many changes for the worse in grass composition which has resulted in range reduction of many species of wildlife such as sable, roan and tsessebe, and severe reduction in the vast herds of plains wildlife which were such a common feature 100 years ago'. The effect of overgrazing in the Botletli area has been severe. In 1934 the District Commissioner reported that conditions were so bad on the Bamangwato (southwest) side of the Botletli river between Rakops and Mopipi that cattle were being moved without authorisation to the Crown lands (northeast) side. He found on the Bamangwato side that ‘there is a lack of grass and the whole area is covered with a fine dust which reduces visibility to only a few yards when the wind blows' (Joyce, 1934). As Campbell (1976) points out, stock also have an adverse effect on foragers by reducing the areas where wild edible plants can be collected. Some informants felt that overhunting has also been responsible for the decrease in game. This is likely, given the growing use of firearms and the recognition of the commercial value of ivory, which became widespread after the visits ofFORAGERS v. FOOD PRODUCERS IN BOTSWANA 311 Livingstone and Chapman. Elephants were abundant in the Botletli area at the time of Livingstone's first trip in 1849, when ‘tusks were left in the field with the other bones', but by the following year 900 were known to have been killed (Livingstone, 1961: 160). Chapman (1971) estimated that 2000-3000 elephants were being killed annually south of Lake Ngami (upriver of the Botletli), and in 1856, just a few years after his first visit, he found elephants becoming ‘less numerous every year, and more wary, and they retreat further into the interior', where malaria and tsetse kept travellers out. Much of the hunting was done not by Europeans but by local tribes, since by this time commercial trophies were sought by the Tawana and Ngwato chiefs as part of the tribute they demanded from other groups (Chapman, 1971: 71; Campbell, 1976). The decimation of the rhinoceros population in the Botletli-Okavango area has also been blamed on excessive hunting (Chapman, 1971; Campbell, 1976). Intensive farming has also reduced game populations through habitat destruc­tion. The people who plough today in the flood plain of the river customarily burn the reeds in the areas they plough although this was not practised tradi­tionally by the Bateti. One informant claimed that after a year of no water the reeds which had been burned failed to grow back for several years. The lechwe and buffalo moved farther upriver as a result and have not been seen in the area since. The role of burning in habitat destruction is brought out by Campbell (1976: 11), who states that the Tswana and Khurutse, in burning the reedbeds for use as productive fields, ‘probably destroyed all the major reedbeds to the south of the Delta and along the Boteti'. There are, then, a number of factors responsible for the decrease in game in the area, but most appear to be related directly or indirectly to the immigration of large numbers of people, together with the marked increase in the number of cattle and fields. European influence, in the form of firearms and a market for ivory, also had an effect. The increase in cattle and decrease in game were probably a strong incentive for the Botletli foragers to substitute pastoralism for hunting and fishing. I will show below how this transition took place. The flood plain: interactions between Bantu and river foragers The first Bantu-speakers to come to the Lake //Gau-lower Botletli area in significant numbers were Bakhurutsi, who arrived there in the late eighteenth to early nineteenth centuries (Schapera, 1952; Schwartz, 1926). Schwartz (1926: 534) says of them that Lake Kumadow [//Gau] is Chapo's Marsh. He was a chief of the Bahrutsis, who came here to escape the depredations of the Matabele, somewhere about the beginning of last century. Here, amid the trackless bogs and ever-burning fires, he was safe, and all the various tribes round about put themselves under his protection.6 Although my informants say that the Bakhurutsi asked the Bateti for land when they arrived, it appears that the Bateti later acknowledged the leadership of the Bakhurutsi chief; Chapman (1971: 58) observed in the 1850s that ‘Chapo . . . has under him a large number of Botletlie, the aborigines from whom this river receives its name'. He also reports that the Bakhurutsi along the lower Botletli kept many sheep and goats, and grew maize and other grains, beans and pumpkins. They apparently did not have cattle during Livingstone's and Chapman's first visits (1849 and 1852 respectively), although they may have done so earlier.7312 FORAGERS v. FOOD PRODUCERS IN BOTSWANA The Bateti gave the Bakhurutsi flood-plain land in which to plough, intermarried with them to some extent, and were on good terms with them. The Bateti had a less amicable relationship with the Kalanga, the other major Bantu-speaking immigrants to the area. This may have been due as much to feelings of superiority on the part of the Kalanga as to the behaviour of the Bateti. One Kalanga man described the old days by saying ‘all these tribes came and married Bateti and were given land by them, but the Kalanga didn't want to marry them because they were Basarwa.' Such feelings of superiority have not disappeared today. The animosity may also have been due to increased density on the river. While some Kalanga may have settled on the river before the arrival of the Bakhurutsi, it appears that the majority of them arrived later. Burning the reedbeds, which had not been a traditional Bateti practice, may be evidence of growing density and a shortage of good farming land. In any event, informants agreed that when the Kalanga arrived in the area in the nineteenth century, having left what is now Zimbabwe to escape the depredations of the Matabele, they were not given a single location along the river for their own use but were instead asked to plough for the Bateti in exchange for small plots of land along the river. Many informants told me that the farmland given to the Kalanga was insufficient, and they were consequently forced to buy grain from the Bateti, which they paid for with cattle. The Bateti were thereby able to build up sizable herds of their own. It is worth noting that the Kalanga arrived in small groups over a number of years (see Chapman, 1971: 62 and 161). Had they come in one wave under a single chief, the Bateti may not have been able to keep them from taking what land they wanted. At the same time the Bateti would never have succeeded had they been as egalitarian as the desert San. The Bateti today do not like to speak of themselves as Basarwa, since Basarwa are poor and this implies an inferior status. Others, however, do. One Kalanga man said, ‘We call them Basarwa, but if you call the Bateti Basarwa they get angry and fight back,' and he added, ‘The Bateti are only rich because they had land.' The G/wiokhoe who live around Lake //Gau, when asked whether Bateti are Basarwa, say that the Bateti are Basarwa of the river, or that they are Basarwa batona (‘big, or important, Bushmen'). Their origins and language notwithstanding, the Bateti and Kalanga occupy a similar economic niche today. They have a subsistence based on agriculture and husbandry, and they do little hunting or fishing (as one Kalanga said, ‘Now they have cattle and don't have to fish'). To the extent that ethnic groups occupy distinct ecological niches we may expect the Bateti to succeed in losing their ethnic identity. One Bateti man said to me, ‘Our grandfathers were Bamangwato,' the Bamangwato being an important Tswana tribe. This claim is almost certainly untrue, but may be a useful fiction. The Bateti used their position as owners of the valuable flood plain to their advantage in a way that contrasts markedly with the fate of the foragers around Lake //Gau, and the interaction discussed above had little or no effect on these groups. To the fate of these lake-shore foragers I now turn. The lake shore: interactions between Bantu and savanna foragers Bantu-speaking pastoralists had minimal impact on the hunter-gatherers around Lake //Gau during the eighteenth and early nineteenth centuries. As farmers and herders, the Bakhurutsi and Kalanga preferred to settle close to the river, whereFORAGERS v. FOOD PRODUCERS IN BOTSWANA 313 they could farm in the flood plain and use the river to water livestock. In the early days the Kalanga used the scrub grassland south of the lake shore only for hunting. Most of my informants say they asked the Basarwa for permission to hunt and that it was never denied. While the //Gorokhoe and G/wiokhoe did not have the political power to refuse the Kalanga permission to hunt in their territory, they had a kind of ‘magical' power since Kalanga are of the opinion that if permission is not granted by the Basarwa and one hunts anyway, one will never be able to catch anything. The interaction between the Lake //Gau residents and the Kalanga seems to have been minimal until farmland became scarce. My informants report that, when the first Kalanga arrived, population density was low enough for stock to be grazed near the river throughout the year. As densities grew, however, we see evidence of competition for the flood-plain land. Because of the inadequacy of historical records for this area, evidence of competition must be inferred from informants' reports and historical events. One possible early sign of density pressure along the flood plain was the practice of burning the reeds to provide additional places to plough. The Kalanga, and possibly also the Bakhurutsi, engaged in this practice, but the Bateti did not do so before these groups came to the Botletli river. Clear evidence of density pressure is apparent by the mid-1900s, when much of the lower Botletli flood plain was appropriated by Kalanga from Rakops who were running out of good farming land upriver. This is a current source of bitterness to my Kalanga and Bateti informants living downriver, who claim that only some of the land they had previously ploughed had been restored to them. Although there is disagreement about whether the downriver people were forcibly evicted or left because of drought, informants agreed that virtually all the people (Bateti, Kalanga, Bakhurutsi) ploughing in the flood plain of the lower Botletli left the area. The Kalanga from the Rakops area upriver, who were growing short of land, then came down and began farming in the ‘vacated' flood- plain land. Some of the dispossessed, including virtually all the Bakhurutsi, never returned to the Botletli river area, but others returned in a few years and asked the Rakops chief for land to plough. Most were given some of their land back, but allegedly only about half as much as they had previously owned. The growing competition for flood-plain land resulted in a greater specialisation of land use in the area. Because the flood plain was most valuable for farming, cattle were moved to the lake-shore area during the rainy season when the crops were growing. Today the lake shore is used chiefly for grazing and many herds are kept there year round. As game became increasingly scarce in the Botletli area, the //Gorokhoe and G/wiokhoe began killing the cattle which had supplanted it - just as happened elsewhere in Southern Africa. A law was eventually passed that imposed strict penalties for killing cattle, and it was then (shortly after the Second World War) that these foragers began caring for cattle in large numbers. The effects of Bantu pastoralists on the //Gorokhoe and G/wiokhoe, therefore, resembled Basarwa-Bantu interations elsewhere in Botswana. The current subordinate economic status of these former foragers is also similar. Whereas the Bateti are today wealthy cattle owners, the Lake //Gau groups have few or no cattle of their own. Today 83 per cent of Sarwa households (« = 23) in the Lake //Gau area care for the cattle of wealthy cattle owners most of which314 FORAGERS v. FOOD PRODUCERS IN BOTSWANA are mafisa cattle.8 Of these cattle owners, 59 per cent today are members of Bantu­speaking groups (Kalanga, Tswana, Damara), while the rest are Bateti. The two groups also differ in their way of obtaining grain. While the Bateti plant productive stands of maize in the flood plain, the Lake //Gau groups plant only tiny gardens, principally of beans and melons, and trade wild berries, milk and firewood to the flood-plain residents in exchange for grain. Wild foods, although used more extensively by the Lake //Gau groups than by the Bateti, are today no more than a supplement to the diet (Cashdan, 1979). Foraging and food production interfere with each other in Botswana, not only because of the direct effects of cattle on native flora and fauna, but because herding in areas of limited standing water requires a degree of sedentarism that makes a dependence on hunting and gathering difficult. DISCUSSION AND CONCLUSIONS How have hunter-gatherers been affected by competition from herding and farming peoples? While the ethnographic literature shows sufficient variation in the form of such interactions to suggest caution in making generalisations, there has been little attempt to define the factors that account for this variation. The historical comparison of riverine and savanna foragers suggests factors that may be of general relevance. The Bateti differed from the Lake //Gau groups in two important ways. First, they had headmen with the power to allocate land. Second, they had valuable land to allocate. They were able to use their control over the flood plain to acquire Kalanga cattle. They are now as wealthy as Kalanga and are trying to erase their ‘Basarwa' identity. Even if they had not been able to acquire cattle by manipulating land allocation, their prior occupancy and ownership of the flood plain would have given them access to excellent farming land. This in itself would have placed them in an advantageous economic position. The uncentralised political organisation of the Lake //Gau groups, together with their lack of access to the scarce and valuable flood plain, made them powerless in the face of Bantu competition. The arid area in which they lived was good for hunting and gathering, reasonable for grazing, but poor for farming. When economic competition made full-time foraging less feasible, pastoralism became the most reasonable form of land use. Without cattle of their own, however, the shift from foraging to pastoralism meant a shift to dependency on Bantu cattle. The ethnohistorical literature presents a generally bleak picture of the effect of food producers on hunter-gatherers, and the Bateti appear to be an interesting exception. Yet they may not be exceptional when one considers prehistoric as well as extant foragers. Foragers with a centralised political organisation are known, ethnohistorically and archaeologically, from highly productive coastal and riverine habitats like the Botletli river. Because of the desirability of such environments, these foragers were affected by food producers long before the desert, Arctic and tropical foragers of our ethnographic acquaintance. Our current perspective on forager-farmer relations is no doubt biased as a result, and the prevailing bleak picture of such interactions may be applicable only to politically uncentralised foragers living in ‘marginal' areas. It would not be unreasonable to suppose that the Bantu-Bateti interaction described here has been played outFORAGERS v. FOOD PRODUCERS IN BOTSWANA 315 many times in the past, wherever food producers encroached on foragers living in productive riverine or coastal environments. NOTES ' Fieldwork on the Botletli river was supported by the National Science Foundation Grant SOC 75-02253. I am grateful for the comments of Alec Campbell and James Denbow of the Botswana National Museum. They should not, of course, be held responsible for any factual errors remaining in this paper. This paper also benefited from references suggested by Klaus Keuthmann, Trefor Jenkins, Mathias Guenther and J. Desmond Clark. I also wish to thank Terry Reisch for his help in translating the German literature. 2 The G/wiokhoe are a different group from the G/wikhoe described by Silberbauer and Tanaka; the latter are physically San while the former are not, and their traditional places are 200 km apart. 3 The names Bateti or Deti (‘the people') and Botletli (‘the river') have the same element - Bo- (as in Botswana) being the Tswana prefix for a place, and Ba- for people. Since the tl combination is not pronounced in northern Botswana, the teti/tletli element sounds the same locally. I have kept the spellings Bateti and Botletli because these are how they are usually spelled. 4 Documents from DeBeers Botswana Mining Company in 1974 and 1975 indicate that the water level at Shakawe (at the upper part of the delta) reached its height between March and May, which is the very beginning of the dry season. This is consistent with other information about flow levels in the Okavango-Botletli system. 5 There is some disagreement about the origins of the Bateti. Schapera (1952) says that the Bateti came to the Botletli river from Zimbabwe in the eighteenth century, and this is corroborated by some of my informants. Denbow (personal communication), on the other hand, thinks that they have always lived along the Botletli river. 6 Schapera, however, states that the arrival of the Bakhurutsi in the area resulted not from Matabele raids, but rather from a dispute about succession. 7 The determination of cattle ownership on the Botletli is complicated by the fact that cattle raiding was prevalent here and elsewhere in Botswana, so that cattle changed hands frequently. The first written records are those of Livingstone and Chapman in the 1850s, and neither reports seeing cattle in the lower Botletli area during their early visits. Chapman (1971: 135) reports having given the Bakhurutsi chief Chapo his first cow in 1854 to reward him for his help and honesty (‘the chief kissed my hand rapturously on becoming the owner of his first cow'). It should not be assumed, however, that there were no cattle on the river prior to this time. The Bakhurutsi probably had had cattle before their arrival on the Botletli, and may have done so for a time afterwards. Campbell (1976) says the Tawana (the Tswana tribe living in the Okavango area) found long-horned cattle among the Bakhurutsi near Lake //Gau, although by 1850 nearly all the cattle were owned by the Tawana. The Bateti may also have had cattle, although the picture is difficult to reconstruct as far back as the early 1800s. I asked five Bateti informants about this but do not feel confident of my information in this early period; in two interviews (one Moteti, one K//anihoe), I was told that the Bateti had come to the Botletli river long ago from the east (Wankie district) and had no cattle when they came. This event is dated by Schapera (1952) to the eighteenth century. However, a Sotho chief in northern Botswana told Livingstone that, after losing his cattle during a Matabele attack in the early 1800s, he ‘stocked himself again among the Botletli, on Lake Kumadau, whose herds were of the large-horned species of cattle' (Livingstone, 1857: 85). Denbow (1984) also reports being told by a Teti informant that the Bateti had cattle prior to the settlement of Bantu-speaking peoples. More oral history focusing on this early time period is needed to clarify the matter. 8 Mafisa is a traditional Tswana practice whereby wealthy cattle owners loan cattle to poor people who have few or none of their own. The borrower cares for the cattle at his place of residence, and uses them for ploughing and for milk. He also receives a share of the increase and in this fashion may build up a herd of his own. REFERENCES Bawden, M. G., and Stobbs, A. R. 1963. The Land Resources of Eastern Bechuanaland. Tolworth: Directorate of Overseas Surveys, Forestry and Land Use Section. Bleek, Dorothea, and Duggan-Cronin, A. M. 1942. The Bushman Tribes of Southern Africa. Kimberley: Alexander McGregor Memorial Museum.316 FORAGERS v. FOOD PRODUCERS IN BOTSWANA Campbell, A. C. 1976. ‘Traditional utilization of the Okavango delta.' Proceedings of the Symposium on the Okavango Delta and its Future Utilization. Botswana Society. Cashdan, E. A. 1979. ‘Trade and Reciprocity among the River Bushmen of Northern Botswana', PhD thesis, University of New Mexico. ----- 1983. ‘Territoriality among human foragers: ecological models and an application to four Bushman groups', Current Anthropology, 24, 47-66. ----- 1984. ‘G//ana territorial organization', Human Ecology, 12, 443-63. -----1986. ‘Hunter-gatherers of the northern Kalahari', in R. Vossen and K. Keuthmann (eds.), Contemporary Studies on Khoisan. Hamburg: Helmut Buske Verlag (forthcoming). Chapman, J. 1971. Travels in the Interior of South Africa. Capetown: A. A. Balkema. Chasko, W. J. Jr, Nurse, G. T., Harpending, H., and Jenkins, T. 1979. ‘Sero-genetic studies on the "Masarwa" of north-eastern Botswana', Botswana Notes and Records, 11, 15-23. Clark, J. D. 1950. ‘A note on the pre-Bantu inhabitants of Northern Rhodesia and Nyasaland', South African Journal of Science, 47, 80-85. -----1951. ‘Bushmen hunters of the Barotse forests', Northern Rhodesia Journal, 1, 56-65. Cowley, C. 1968. Fabled Tribe: a voyage to discover the river Bushmen of the Okavango swamps. New York: Atheneum. Crowell, A. L., and Hitchcock, R. K. 1978. ‘Basarwa ambush hunting in Botswana', Botswana Notes and Records, 10, 37-51. Denbow, J. R. 1984. ‘Prehistoric herders and foragers of the Kalahari: the evidence for 1500 years of interaction', in C. Schrire (ed.), Past and Present in Hunter Gatherer Studies, pp. 175-93. Orlando: Academic Press. -----and Wilmsen, E. 1983. ‘Iron Age pastoralist settlements in Botswana', South African Journal of Science, 79, 405-8. Dornan, S. S. 1925. Pygmies and Bushmen of the Kalahari. London: Seeley, Service. Dyson-Hudson, R., and Smith, E. A. 1978. ‘Human territoriality: an ecological reassess­ment', American Anthropologist, 80, 21-41. Ehret, C. 1974. ‘Patterns of Bantu and central Sudanic settlement in Central and Southern Africa (CA. 100 B.C.-500 A.D.)', Transafrican Journal of History, 3, 1-71. -----1982. ‘The first spread of food production to Southern Africa', in C. Ehret and M. Posnansky (eds.), The Archaeological and Linguistic Reconstruction of African History, pp. 158-81. Berkeley: University of California Press. Estermann, C. 1976. The Ethnography of Southwestern Angola, ed. Gordon Gibson. New York: Africana Publishing Company. Gibb, Sir Alexander. 1969. Botswana-Shashi River Feasibility Studies: report on water supply to AK 1 (Orapa). London: Sir Alexander Gibb and Partners, Consulting Engineers. Gusinde, M. 1966. Von Gelben und Schwarzen Buschmannem. Graz, Austria: Akademische Druck- u. Verlagsanstalt. Heinz, H. J. 1969. ‘Search for Bushman tribes of the Okavango', Geographic Magazine, 41, 742-50. Hermans, Janet. 1977. ‘Official policy towards the Bushmen of Botswana: a review, Part I', Botswana Notes and Records, 9, 55-67. Hitchock, R. K. 1980. ‘Tradition, social justice and land reform in central Botswana', Journal of African Law, 24, 1 -34. ----- 1982. ‘Patterns of sedentarism among the Basarwa of eastern Botswana', in Eleanor Leacock and Richard Lee (eds.), Politics and History in Band Societies, pp. 223-65. Cambridge: Cambridge University Press. Jeffares, J. L. S. 1938. ‘Report of the N'gamiland Waterways Surveys of 1937', unpublished report to the Botswana Government. Jenkins, T., and Nurse, T. 1977. Health and the Hunter Gatherer: biomedical studies on the hunting and gathering populations of Southern Africa. Basel & New York: S. Karger. Joyce, J. W. 1934. ‘District Commissioner Report', Botswana National Archives S/273.4.FORAGERS v. FOOD PRODUCERS IN BOTSWANA 317 Khama, T. 1935. Statement by Chief Tshekedi Khama, in TheMasarwa (Bushmen): report of an inquiry by the South Africa District Committee of the London Missionary Society. Lovedale Press. Kohler, O. 1971. ‘Die Khoe-sprachigen Buschmanner der Kalahari', in Forschungen zur Allgemeinen und Regionalen Geographie: Festschrift fur Kurt Kayser, pp. 373-411. Wiesbaden: Franz Steiner Verlag. -----1981. ‘Les langues khoisan', in Jean Perrot (ed.), Les langues dans le monde ancien et moderne: Part I, G. Manessy (ed.), Les langues de I'Afrique subsaharienne, pp. 460-615. Paris: Editions du Centre National de la Recherche Scientifique. Lee, R. B. \919..The IKung San: men, women and work in a foraging society. Cambridge: Cambridge University Press. Livingstone, D. 1857. Missionary Travels and Researches in South Africa. London: John Murray. -----1960. Livingstone's Private Journals 1851-1853, ed. I. Schapera. Berkeley: University of California Press. -----1961. Livingstone's Missionary Correspondence 1841-1856, ed. I. Schapera. Berkeley: University of California Press. Mackenzie, J. 1883. Day-Dawn in Dark Places: a story of wanderings and work in Bechwanaland. London: Cassell. Miller, Sheryl. 1969. ‘Contacts between the Later Stone Age and the Early Iron Age in Southern Central Africa', Azania, 4, 81-90. Mohr, E. 1876. To the Victoria Falls of the Zambesi. London: Sampson Low, Marson, Searle & Rivington. Nurse, G. T., and Jenkins, T. 1977. ‘Serogenetic studies on the Kavango peoples of South West Africa', Annals of Human Biology, 4, 465-78. Nurse, G., Weiner, T., and Jenkins, T. 1985. The Peoples of Southern Africa and Their Affinities. Oxford: Oxford University Press. Passarge, S. 1907. Die Buschmanner der Kalahari. Berlin: Reimer & Vohsen. Pike, J. G. 1971. ‘Rainfall over Botswana', Botswana Notes and Records, special edition no. 1, pp. 69-76. Rangeley, W. H. J. 1963. ‘The earliest inhabitants of Nyasaland', Nyasaland Journal, 16, 35-42. Schapera, I. 1930. The Khoisan Peoples of South Africa: Bushmen and Hottentots. London: Routledge & Kegan Paul. ----- 1952. The Ethnic Composition of Tswana Tribes. London: London School of Economics & Political Science, University of London. Schwartz, E. H. L. 1926. ‘The Botletle river', Geographical Journal, 67, 528-35. Seiner, F. 1910. ‘Die Buschmanner des Okavango and Sambesigebietes der Nord-Kalahari', Globus Illustrierte Zeitschrift fur Lander und Volkerkunde, vol. 97, no. 22. English translation by Helga Esche, Botswana Notes and Records, 1977, 9, 31-6. Selous, F. C. 1893. Travel and Adventure in South-East Africa. London: Rowland Ward. Statham, J. C. B. 1924. With My Wife Across Africa by Canoe and Caravan. London: Simpkin, Marshall, Hamilton, Kent. Stigand, A. G. 1923. ‘Ngamiland', Geographical Journal, 62, 401-19. Vossen, R. 1984. ‘Studying the linguistic and ethno-history of the Khoe-speaking (central Khoisan) peoples of Botswana, research in progress', Botswana Notes and Records, 16, 19-35. Westphal, E. O. J. 1962. ‘A re-classification of Southern African non-Bantu languages', Journal of African Languages, 1, 1-8. -----1963. ‘The linguistic prehistory of Southern Africa: Bush, Kwadi, Hottentot and Bantu linguistic relationships', Africa, 33, 237-65. -----1971. ‘The click languages of Southern and Eastern Africa', in J. Berry and J. H. Greenberg (eds.), Current Trends in Linguistics, 7: Linguistics in Sub-Saharan Africa, pp. 367-420. The Hague: Mouton.318 FORAGERS v. FOOD PRODUCERS IN BOTSWANA Resume Concurrence entre les producteurs de cultures alimentaires et de fourrages sur la riviere Botletli au Botswana L'immigration des groupes producteurs de denrees alimentaires dans les regions occupees de peuplades vivant de la chasse et de la cueillette avait du constituer un phenomene courant dans la prehistoire. Comment ce phenomene affectait-il ces peuplades vivant de la chasse et de la cueillette? Je decris ici deux groupes de Bochimans Basarwa, l'un vivant le long de la plaine d'inondation au bas de la riviere Botletli et l'autre occupant la savane, a une faible distance de la riviere. Ces deux groupes possedent des moyens de subsistance et une organisation sociale differents et furent affectes par les bouviers et les fermiers immigrants de diverses fajons. Les Basarwa de la savane pratiquaient une politique egalitaire et occupaient une region riche en paturages mais pauvre pour la culture agricole. Lorsque la chasse a outrance et la concurrence de grands nombres de bestiaux reduirent la quantite de gibier dans leur region, ils n'eurent d'autre choix que de s'occuper du betail des riches proprietaries absents. Les Basarwa vivant le long de la plaine d'inondation de la riviere Botetli etaient differents de ces groupes (et des Sans decrits dans la documen­tation) car ils possedaient des chefs qui detenaient le pouvoir de repartition des terres ainsi que la propriete des riches terres arables a repartir. Ils pouvaient utiliser leur controle sur la plaine d'inondation pour acquerir du betail et maintenant ils sont aussie riches que leurs voisins de langue bantou.