Cover Image
close this bookThe Improvement of Tropical and Subtropical Rangelands (BOSTID)
close this folderPart I
close this folderThe social context for rangeland improvement
View the document(introduction...)
View the documentProduction systems in tropical and subtropical regions
View the documentContext of environmental degradation
View the documentReferences


The loss of desirable vegetative cover is a threat to world food supplies, to the quality of human life, and to the environment. Desertification, erosion, and the loss of useful plant species can be arrested through revegetation. However, revegetation efforts have often ended in failure or have had limited impact. This has been particularly true in semiarid and marginal lands where the reestablishment of plants is a delicate process.

Revegetation often is required to correct the abuse of existing resources by people and their livestock. More often than not, the success or failure of revegetation schemes is also determined by human activities. Normally, adequate protection of an area is possible only if the people who use the land alter their behavior. If they are unwilling or unable to do so, revegetation efforts become more expensive or even impossible. Far too often, planners and conservationists ignore the human ecology of an area and fail to appreciate the importance of project lands for the survival of human populations. This chapter outlines the relationship between human activity and vegetative change.

In most instances, environmental degradation is a product of human activity. In the regions of Africa and Asia that are the focus of this study, overgrazing, the excessive cutting of fuelwood, and the cultivation of fragile lands - abuses often precipitated by the openaccess provisions of colonial public-domain law and subsequent lack of governmental management and control and economic differentiation associated with commercialization - have led to a loss of plant cover and required the development of government revegetation programs. To fully appreciate why this has occurred and how this process can be reversed, we must first understand how human beings have adapted to specific environmental settings.

Production systems in tropical and subtropical regions

The lands considered here are those in which permanent, sustainable crop production is not possible because of soil and climatic conditions. These regions have, however, supported substantial human populations for thousands of years. In these areas, people have developed production systems adapted to the low and variable productivity of these lands. In semiarid regions and marginal areas, one can find many kinds of production systems - hunting and gathering, agricultural, pastoral, and agro-pastoral systems. Pastoralism and agro-pastoralism are probably the most common production systems in these regions; this is because domestic animals can convert vegetation on land unsuitable for agriculture into food and fiber. In Asia and Africa, agro-pastoral and pastoral societies take many forms. The exact organization of these production systems is influenced by local environmental factors, by history, by culture, by economic considerations, and by level of technology. In addition to these differences, there are similarities that must be understood if successful revegetation is to take place. Traditionally, people who live in semiarid and marginal lands have relied on two strategies - diversification and mobility - to cope with the erratic and generally low productivity of their lands. Mobility is perhaps the most important characteristic of these production systems. By moving about, one can take advantage of the spatial and seasonal variation of plant production. In crop production, systems of shifting and opportunistic cultivation are examples of strategies based on mobility. Land is cultivated for several years, and then it is abandoned to fallow and new land is cleared. The shifting nature of cultivation permitted natural revegetation processes to occur - provided that the fallow cycle was long enough.

Livestock are particularly mobile. Not only can they move about to "harvest" sparse vegetation, but they convert grasses and shrubs into useful products. They also can harvest perennial shrubs and plants that are less susceptible to annual variations in weather than are annual food crops.

Livestock herds in semiarid and marginal regions are rarely confined to the same pastures year-round. Seasonal movement is a common feature of livestock production even among sedentary groups (figure 2-1). A noteworthy difference between agro-pastoralists and pastoralists is that the former do not often move as complete families with their herds, whereas family members often accompany herds among pastoralists. The movement of animals may be as little as a few kilometers or as much as hundreds of kilometers. Movement, however, permits herds to take advantage of seasonally rich pastures, helps to adjust to spatial variances in precipitation, and reduces the stress that is placed on vegetation through constant grazing and trampling (Wagner, 1983).

Mobility is probably the key production strategy for pastoral nomads. Strategies of mobility, such as nomadism and transhumance, are particularly prevalent in areas where the pattern of rainstorms is such that there can be wide differences between the amount of rain received by two plots a few kilometers apart (Gilles and Jamtgaard, 1981). Nomadism serves to reduce environmental stress and personal risk, but it is also more productive than settled livestock husbandry. In eastern Africa, in areas of Masai pastoralism, the grazing capacity of the land is increased 50 percent because of herd mobility (Western, 1982). In the important livestock-producing areas of Africa, comparisons of the productivity of mobile and sedentary herds have indicated the superiority of mobility. In Sudan, Haaland (1977) noted substantially higher mortality rates among sedentary herds than among mobile ones. Breman and de Wit (1983) studied the migratory system based in the Inland Delta of the Niger River in Mali, and found that its productivity often exceeds that of Australian and North American ranches.

Other studies in West Africa have indicated that a disproportionate number of sedentary cattle were lost in the 1968-1974 Sahelian drought. Losses of herds that quickly moved into rainier regions in response to the drought were minimal (Gallais, 1977; Sall, 1978). Not only were the impacts of drought less severe on mobile herds, but migrating herds caused less environmental degradation. Loss of vegetation around boreholes where herds permanently congregated was so severe that it could be easily recognized from satellite photos. Mobility is an important aspect of production systems in semiarid and marginal areas.

Mobility is just one way to cope with a harsh environment. Diversification of subsistence activities is another. For farmers, the ownership of livestock is one diversification strategy. Animals may survive even when grain yields are quite low, so livestock may represent a store of capital that can be used in years of drought. Farmers may also grow a number of different crops to reduce the risk of crop failure. Wheat and barley or maize and sorghum may be grown together because one species tolerates drought better than another. Diversification goes beyond the diversification of agricultural enterprises. Farmers may also have secondary occupations, engage in trade, or in migrant labor to reduce their dependency on a fickle pastoral environment. In traditional subsistence-oriented societies, this diversification lessened the dependence on the immediate environment and lessened the probability of ecological disaster.

At first glance, traditional pastoralists would appear to have been a highly specialized group dependent solely on livestock production, but in reality these societies were highly diversified. First, multiple species of animals were raised (figure 2-2). Camels, cattle, sheep, and goats all have different water requirements, feed preferences, and reproductive rates. Browsers - camels and goats - are less affected by annual fluctuations in rainfall and in grass production than cattle and sheep. Small stock such as sheep and goats have high reproduction rates when they are well nourished. They can thus be used to build up herds rapidly after droughts or to take advantage of two or three consecutive wet years. Not only did traditional pastoralists diversify their herds, but they also had other sources of livelihood. Pastoralists in the Sahara, Asia, and the Andes were often heavily involved in long distance caravan trade, in the mining of salt, and/or in military pursuits. Often they ruled or exacted tribute from sedentary groups, which provided them with agricultural products. As a result, pastoralists, like agro-pastoralists, developed diversified sources of livelihood to prevent over-reliance on any particular aspect of the environment. One consequence of this diversification was to reduce the impact of man on any single ecological niche.

Societies living in marginal areas have many institutions to facilitate diversification and mobility. One important institution is the land-tenure system. In general, the private ownership of land in such regions is rare, except in those places where irrigation or other conditions made permanent cultivation possible. Land ownership in these areas was, and still is to a large extent, collective. In areas of shifting cultivation, the cultivator had use rights to a piece of land as long as it was cultivated, but did not have an inalienable right to that land. Such rights belonged to a large group - a village, commune, clan, or tribe.

Rights to grazing lands and forest lands are also collective. However, in this case there are no user rights to individual pieces of land. Although an individual might habitually use a pasture or forest, mobility is essential to responding to fluctuations in precipitation and plant production, making exclusive assignments of land impractical. Often the boundaries between the territories of different pastoral groups are imprecisely defined, and relations of kinship and reciprocity exist that permit groups to temporarily use the pastures and forests of others. Collective ownership of pasture and forestlands is also more economical than individual tenure. The low and variable annual productivity of these lands makes the cost of maintaining fences and access roads to individual plots prohibitive. Under these conditions, if mobility is not impeded by private ownership of lands, all users of collective lands benefit from higher levels of production.
As the discussion above indicates, collective ownership of land facilitated both mobility and diversification. Therefore, a large proportion of range and forest lands remains today under the control of localities or as part of the public domain in Europe, Japan, and North America.

To say that lands are collectively owned does not imply completely open and unregulated access. That would lead to a "tragedy of the commons" situation such as that described by Hardin (1968), where individuals would each increase their herds or their use of the forests until the productive capacity of the resource was destroyed. Such unregulated exploitation of the environment ignores the fact that members of subsistence groups depend upon each other for their survival and are not individuals single-mindedly pursuing personal gain at all costs (Runge, 1981). Also, it is illogical to suggest that any group would stand by and let their subsistence base be destroyed.

The "tragedy" historically appears to occur where competition over land and its resources increases, and where differential access and market opportunities and political control reduce the effectiveness of prior regulatory procedures. Studies of traditional management systems indicate that in those areas where disease and warfare do not prevent overgrazing, a variety of institutions regulate the use of common resources. First, these lands are not open to all potential users, but are either used exclusively by certain groups, or at the very least, some groups are given priority over others. In the case of cropland, people usually need permission from local leaders or councils to use land. Even when access to pastures and woodland was technically open to all, controls over access to water, shelter, and minerals was controlled by localities or owned by individuals. For example, wells and springs are often "owned" by individuals or by small groups (Helland, 1982). Without access to water or to shelter, no one can use pastures, even if they are technically common resources.

Subsistence-oriented groups in semiarid lands do not necessarily live in harmony with nature. Pastoralists and agro-pastoralists often significantly alter the vegetation and fauna of the areas in which they live. Sometimes they do destroy the resources upon which they depend. If they do so, they quickly destroy their ability to survive and are either forced to move on or disappear. Direct dependency on the immediate environment for most subsistence needs is a strong incentive for the development of institutions to protect the environment. While most groups living in marginal areas have only rudimentary institutions, in some areas, such as southern Africa and the Atlas Mountains of Morocco, elaborate institutions evolved to regulate pasture use (Bourbouze, 1982; Gilles, 1982; and Odell, 1982). In recent years, many of the mechanisms that have traditionally served to protect vegetation have become less effective. The reasons for this decline are discussed below.

Context of environmental degradation

Typically, lands requiring revegetation have had their plant cover destroyed through improper farming methods, the extensive gathering of wood for fuel or construction purposes, or through overgrazing. In terms of destructive impact, cultivation and the gathering of woody species probably have had more impact on the environment than have grazing animals. Livestock are the primary cause of desertification only in areas where large numbers of grazing animals are concentrated, such as around boreholes. Overgrazing is, however, a major cause of vegetative change and often inhibits the restoration of plant cover.

Although these three actions of man are the main causes of environmental de gradation , the reasons for increased degradation are still debated. Four common reasons for environmental deterioration are climatic change, population growth, economic change, and human fallibility. Usually these factors interact. Over the past 30 years, the human and animal populations using semiarid and marginal lands have grown, thereby putting more pressure on the environment. Given the cycles of wet and drought years common to semiarid regions, such as the West African Sahel, this led to population growth that could only support the population in wet periods. The shortsightedness of governments and donor agencies also has contributed to environmental deterioration. In Africa in particular, ill-conceived water and livestock development projects contributed substantially to overgrazing (Bernus, 1971; Haaland, 1977). These factors have all contributed to the need for revegetation programs, but merely listing the mechanisms and factors leading to environmental problems does not explain the process by which this has occurred. Also, in many cases, destruction of plants in marginal areas has been occurring at a faster pace than have climatic or population changes. This has been due to factors that have reduced both the mobility and the diversity of traditional economies. These factors also undermined traditional mechanisms of environmental protection. The need for revegetation and conservation has been accelerated by the growth in government power, in modern economies, and in the use of improved technologies.

Changing political conditions have had severe impacts on many pastoral societies of Africa and Asia. In many cases, national boundaries were created in such a way that grazing lands used by one people were split between two or more nations. Over time, nations have incressingly restricted the movement of people and animals across frontiers . Such restrictions reduce the diversity of ecosystems available to herders and lead to herds spending longer periods of time in marginal areas than they had in the past. For example, it has been argued that the imposition of the frontier between Uganda and Kenya was the reason for overgrazing and the destruction of the pastoral economy of the Karamajong of Uganda (Quam, 1978). The imposition of national boundaries also had some impact on trading activities.

The growth of state authority has impinged on pastoral production systems in several other ways. To a large degree, governments have eliminated the raiding and warfare that often characterized the relationship between herders and their neighbors. While pacification is in itself quite desirable, it often had the effect of opening grazing lands to groups that previously did not have access to them. Often the state did not give title or the means to restrict access to grazing lands to anyone. This in effect made it impossible for communities to enforce local regulations designed to protect pastures or woodlands. Often government planners felt that local rules concerning pasture use prevented efficient meat production or impeded nation building (Sall, 1978; Cole, 1981). They wished to create a common pasture situation in which any individual who wished to raise livestock was free to do so, both to expand meat production and to combat tribalism. In many instances, these changes were accompanied by changes in herd composition. For example, cattle production increased dramatically in the northern Sahel at the expense of less destructive, better adapted forms of livestock, such as the camel (National Research Council, 1983).

The reduction of intergroup hostilities and the introduction of government land resource planning had additional impacts on the viability of traditional pastoral economies. Governments have generally sided with agriculturalists in disputes between them and pastoralists. The cessation of raiding by pastoral groups led to projects to expand agricultural production, and population pressures have contributed to the expansion of cultivated are as - further reducing the mobility of traditional pastoralists. Generally, those lands that are occupied by farmers are marginal croplands but are among the best watered and most productive pastures (figure 2-3). They are generally those used during the dry season or in winter months when the productivity of other pastures is low. The loss of these lands to farmers forces animals to remain longer on more marginal lands and increases the likelihood of erosion and desertification due to overgrazing. Even where arrangements can be made for the pasturing of animals on stubble, as is the case of much of West Africa, mobility is reduced. Such arrangements can often be developed for regular seasonal usage of pastures but not usually for occasional or for emergency use.

While governments frequently have reduced pastoral groups' land rights, in some countries there have been attempts to protect pastoralists by adjudicating land rights. French colonial authorities in North Africa attempted to adjudicate tribal boundaries, and, in East Africa, post-independence governments have attempted to delimit group ranches and grazing blocks. These attempts have in some cases restricted the growth of herds and have probably reduced overgrazing, but they also can restrict mobility if strictly enforced. As mentioned earlier, the fluid, often vague, boundaries between the areas used by different pastoral groups facilitated mobility. Overly rigid enforcement of these rules can confine animals to too small an area or, if not enforced, boundaries may be ignored. If local groups still manage resources, the fluidity of boundaries may not be a problem, but if reforms have eliminated or modified the ability to control grazing, then once again the creation of a common resource may be required where one previously did not exist.
The growth of market economies, and the adoption of new technologies that this growth permits, have also reduced the viability of tradition al resource management strategies . In a subsistence economy, one's survival is directly linked to the local environment. One exploits a wide variety of resources, but one's survival over time depends on the sustained productivity of the immediate environment. The introduction of a market economy changes this. First, one can specialize in animal production or in the cutting of fuelwood. To do so means that one can increase one's standard of living by intensely exploiting a particular environmental niche. If one is selling for cash, the feedback loop between subsistence levels and environmental conditions is less effective. If demand for one's product is rising, price increases can more than cover the loss of productivity because of overexploitation of the environment. For example, as a pasture deteriorate, a subsistence herder may only have milk and meat to eat, while a commercial beef producer may for a long time experience stable or even rising income levels.. Free labor markets also reduce the risk of degradation for the individual. The destruction of the land may cause hardship, but the possibility of wage labor in the city always exists.

In the past, some form of “passive" management occurred when quantities of stock died as a result of drought. Today, in many parts of Asia and North Africa, herders can maintain herd numbers, even when pastures and water are totally exhausted, by trucking water and feed to their animals until rains restore pastures. The purchase of feed and the delivery of water, often subsidized by governments during droughts, leads to levels of overgrazing that would be impossible for traditional subsistence pastoralists. An unintended consequence of improving veterinary services and reducing disease is to remove this “natural" regulator of herd size. Another consequence of the growth of the market economy is that individuals enjoy increased economic independence. In traditional groups, each individual family is dependent on others for survival. In such a setting, social pressure and the threat of ostracism may be sufficient to prevent deviant behavior. The development of a market economy increases economic differentiation and may reduce consensus on resource management questions.

Modernization has also contributed to the degradation of the environment in some areas Improved medical and veterinary techniques have reduced the constraints that disease placed on human and herd numbers. The development of roads and the introduction of motor transport have caused some nomads to become more dependent on herding as caravans have become less profitable and have probably encouraged the switch from camels to cattle. Roads and trucks have also made it profitable to cut fuelwood or produce charcoal at great distances from cities, and trucks make it possible to increase the use of remote or poorly watered pastures (Thalen, 1979). In many cases, mechanized plowing and sowing have made it profitable to plow up rangelands where rainfall is so erratic that only one year in three witnessed successful harvests. The introduction of new technologies often requires changes in traditional institutions, hence an unintended consequence may be a weakening of those institutions that have in the past protected the environment. In this light, publicly funded revegetation programs may be seen as attempts to correct some of the excesses of rapid social change.

As we can see, then, desertification and the destruction of plant cover have been caused by a number of factors. It is important to remember from these examples that environmental deterioration has been accelerated because the mechanisms that formerly helped people adapt to semiarid and marginal environments have been weakened. Diversification and mobility have been limited, and the feedback from man's use of the environment has been distorted. If revegetation efforts are to be successful, they must create a sustainable human ecology as well as stable, productive environmental systems. Too often, projects have undermined themselves by ignoring people, or by inadvertently accelerating the processes of declining diversity and mobility in production systems.

Successful revegetation requires changes in land use patterns so that the reestablishment of vegetation is encouraged. In the past, attempts have been made to control access to revegetated areas by changing land tenure arrangements. Nomads have been settled, private and group ranches have been created, and forest reserves have been legislated, all in attempts to control access to project lands by reducing animal movement and by restricting people to particular parcels of land. As previously mentioned, the reduction of mobility may threaten the viability of traditional subsistence systems. If their livelihood is threatened, people may resist overtly or may passively resist by bribing forest guards or by grazing or cutting revegetated areas when they are not being properly guarded. Conflict between traditional users at the very least raises the cost of revegetation substantially, and may in many instances negate project efforts.

In some cases, the failure to understand the importance of mobility can mean disaster even when project goals are attained. Boreholes are examples of efforts to increase available pastures that, in fact, led to local desertification and to heavy livestock losses during droughts (figure 2-4). At other times, the success of programs in one area may lead to larger levels of environmental deterioration outside a project are a . Herds that are required to leave the are a of a range or reforestation project must go somewhere, hence the revegetation projects may accelerate the processes that they are intended to reverse. The settlement of nomads may increase overgrazing, as we saw in the Sudanese example (Haaland, 1977). The creation of private ranches or group ranches may improve the conditions of ranges in their boundaries, as it has in some parts of Kenya (Hopcraft, 1981; and case study 10). If ranchers are not excluded from common pastures they may use their individual pastures as reserves, which permits them to exploit other lands more intensively (Little, 1983). In a similar vein, people may preemptively destroy an area rather than have it come under the control of a public range or forestry program. Pascon (1980) cites the example of herders in Morocco who, when presented with the successful establishment of wheat grass on overgrazed plain, chose to plow up the entire region and plant wheat rather than give up control of their resources to a range management scheme. These examples, though perhaps more graphic than most, are typical of many attempts at revegetation.

Those who plan revegetation efforts often face a dilemma. Successful programs may require the use of coercion and force, which, in turn, raises the cost of revegetation, reduces the extent of the area that can be treated, and reduces cooperation. This is one part of the dilemma - coercion reduces the program area. On the other hand, success in a limited area may be illusory; vegetation may be protected at the cost of widespread environmental destruction in adjacent areas. This is a cruel dilemma.

In part, this difficulty can be overcome if rehabilitation efforts are carefully reconciled with local systems of production. If one understands how a revegetation program will impact on an area, one may be able to make adjustments in other parts of the local production system to compensate for disturbances caused by a program. Instead of paying money for guards, it may be possible to plant highly valued, multiple-use species that would strengthen and diversify the local economy, thereby justifying protection by local populations. Approaches can be developed that enhance the advantages of mobility and diversity for production systems in these areas. The creation of new jobs or economic activities may have a greater impact on the environment than the creation of forest or grazing reserves.

Given sufficient time and money, it is possible for planners to characterize a production system and to design appropriate revegetation programs. An easier approach may be to reduce technical input, but to work closely with local populations to identify appropriate types of interventions and to monitor the program. Such an effort may succeed in areas where government policies have often undermined local institutions.

It is of particular importance that environmental rehabilitation projects yield multiple benefits. Multiple uses of vegetation should be encouraged. Local involvement should reduce management costs through increased self-enforcement of conservation rules. Finally, the project should help reestablish a local sustainable resource system that is not dependent on the vagaries of public funding and political will. There may often be some trade-offs between the efficiency of revegetation and local involvement. There may be more efficient and more effective ways of conserving and protecting plant cover than those acceptable to local populations. For example, the policies developed by ranchers and the Grazing Service in the United States under the Taylor Grazing Act did not satisfy many conservationists, but they could be implemented effectively and did lead to improved range conditions in the western United States (Foss, 1960; U.S. Forest Service, 1979). The goal of any revegetation program should be to create a viable environment for plants, animals, and people. This can be done only by placing revegetation efforts within the context of local and regional production systems.


Bernus, E. 1971. Possibilities and limits of pastoral watering plans in the Nigerian Sahel. Paper presented at the FAO Symposium on Nomadism, Cairo, Egypt.

Bernus, E. and G. Savonnet. 1973. Les probls de la seresse dans l'Afrique de l'Ouest. Prnce Africaine 88:112-138.

Bourbouze, A. 1982. Dacements des troupeaux et utilization des parcours dans le Haut Atlas Central. Product 'on Pastorale et Soci 10(Spring):34-45.

Breman, H. and C. T. deWit. 1983. Rangeland productivity and exploitation in the Sahel. Science 221(1):1341-1342.

Cole, D. P. 1981. Bedouin and social change in Saudi Arabia. Journal of Asian and African Studies 16(1-2):128-149.

Foss, P. O. 1960. The Politics of Grass: Administration of Grazing on the Public Domain. University of Washington Press, Seattle, Washington, USA.

Gallais, J., ed. 1977. Strates Pastorales et Agricoles durant la Seresse 1969-1974. Centre d'udes de Graphie Tropicale, Talence, France.

Gilles, J. 1982. Organizing for pastoral development: Themes from traditional sytems. Agricultural Administration 11:215-225.

Gilles, J. L. and K. Jamtgaard. 1981. Overgrazing in pastoral areas: The commons reconsidered. Sociologia Ruralis 21:129-141.

Haaland, G. 1977. Pastoral systems of production: the sociocultural context and some economic and ecological implications. Pp. 179-193 in Land Use and Development, P. O'Keefe and B. Wisner, eds. International African Institute, London, England.

Hardin, G. 1968. The tragedy of the commons. Science 162:1243-1248.

Helland, J. 1982. Social organization and water control among the Boran. Development and Change 13(1):239-258.

Hopcraft, D. N. 1981. Economic institutions and pastoral resource management: considerations for a development strategy. Pp. 224-243 in: The Future of Pastoral Peoples J.G. Galaty, D. Aronson, P.C. Salzman and A. Chouinard, eds. International Development Research Centre, Ottawa, Canada.

Little, P. D. 1983. Businessmen and part-time pastoralists: some factors affecting drought and overgrazing in Baringo District, Kenya. Paper presented at the 149th Annual Meeting of the American Association for the Advancement of Science, Detroit, Michigan, USA.

National Research Council. 1983. Environmental Change in the West African Sahel. Board on Science and Technology for International Development. National Academy Press, Washington, D.C., USA.

Odell, M. J. 1982. Location, institutions and management of communal resources: lessons from Africa and India. ODI Pastoral Network; Paper, no. 14e, Overseas Development Institute, London, England.

Pascon, P. 1980. La comption des veurs dans la ron d'Azrou. Pp. 61-72 in udes rurales: Id et enqus sur la campagne Marocaine. Soci Marocaine des iteurs Ris, Rabat, Morocco.

Quam, M. D. 1978. Cattle marketing and pastoral conservatism. African Studices Review 21:49-71.

Runge, C. F. 1981. Common property externalities: Isolation, assurance, and resource depletion in a tradition al grazing context . American Journal of Agricultural Economics 63:595-606.

Sall, A. 1978. Quel amgement pastoral pour le Sahel. Revue Tiers Monde 19(73):161-169.

Thalen, D. C. P. 1979. Ecology and Utilization of Desert Shrub Rangelands in Iraq. Dr. W. Junk, The Hague, Netherlands.

U.S. Department of Agriculture Forest Service. 1979. Rangeland Policices for the Future. Government Printing Office, Washington, D.C., USA.

Wagner, F. J. 1983. Nomadic pastoralism: Ecological adjustment to the realities of arid environments. Paper presented at the 149th Annual Meeting of the American Association for the Advancement of Science, Detroit, Michigan, USA.

Western, D. 1982. The environment and ecology of pastoralists in arid savannas. Development and Change 13(2):183-211.