The research that led to the writing of this paper began with a puzzle. The four volumes of K. E. Laman's Kongo ethnography (1953–68) are based on texts written by BaKongo. Who were these men and why is there no explicit reference to them in the books? The lack of any indication of who wrote what and on what authority deprives this ethnography of much of its potential value (Janzen, 1972). In search of biographical information, I turned to the bulletin Minsamu Miayenge (MM), published by the Swedish Missionary Society (SMF) from 1892 onwards. From its pages emerged a clear sense of a group of men whose position in the nascent colonial order explained both their ethnographic ambitions and their subsequent disappearance from the record.

The settlements of Bantu-speaking people in Southern Africa vary widely in size and distribution, ranging from the dispersed homesteads of the Nguni to the large towns of the Tswana. These two extremes have interested Africanists since the beginning of the nineteenth century, when Europeans first encountered the Thlaping at Dithakong near present-day Kuruman. Today the contrast between Tswana and Nguni settlements are most often attributed to differences in social stratification, cultural preference or environmental conditions.

These conventional explanations provide a focus for considering the meaning of settlement patterns among the southern Bantu. I first develop a model of political and settlement hierarchies to isolate the essential differences between Nguni and Tswana communities, and then I present archaeological evidence that calls into question the conventional explanations.

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?

Recent studies of African pastoralism have come more and more to concentrate on its political economy and to note the increasing social and economic differentiation occurring within pastoral societies. As Swift and Maliki write of West Africa: ‘Since the 1973 drought, there has been an increasing process of proletarianization in the countryside which has particularly affected herders, who are in many places being transformed from independent rural producers into cowboys herding other people's animals on land they no longer control’ (1984: 2). In Kenya this process has become increasingly apparent since independence, as pastoralism has become dominated by a town-based elite (see Dahl, 1979a; Little, 1983, 1985a and b; Ensminger, 1984). In this article I trace the origins of a new kind of pastoralism in northern Kenya, and argue that poverty and dependence is becoming a permanent way of life to many pastoralists.

A considerable part of the anthropological literature on Africa is devoted to the description of African ideas about the causes of misfortune and illness and the steps which should be taken to remedy them. In the early days of African ethnography information on this aspect of culture was usually presented as part of a general ethnography or included in a more specific description of an African cosmological system. As the general ethnographic literature on Africa accumulated, and as the need for more focused information for application in health and development work arose, studies which investigated the more narrow domain of African therapeutics became more common. As Janzen has observed (1978: 121), the study of African therapeutic systems has now moved into a phase of exemplary field studies which gather data in a number of different areas: illness taxonomies, curative techniques, healing roles, health-care decision-making resource allocations, course of illnesses through distinct episodes, and the relationship between bioecological and sociocultural factors. No one would deny that an ideal study would gather data in all these areas. The study of Lugbara medicine reported on in this paper, however, was carried out under difficult conditions and was abruptly halted after one year because of the worsening political situation. No epidemiological or biomedical data were collected.