tively on farmers’ fields. For example, most information on the contribution of legume nitrogen is from research stations where soils have sufficient P and other nutrients and is sometimes is irrigated (Mafongoya et al., 2006). Most
soil fertility research in East Africa has concentrated on recommendations for monocrop systems despite the fact that most smallholder farmers use intercropping and mixed cropping systems (Bekunda et al., 2004). Evidence suggests that involving farmers in soil fertility research improves the likelihood of recommendations that are more relevant to farmers’ situations (CIAT, 2002; Bekunda et al., 2004). Onfarm experiments are more likely to provide realistic rates of return to different technologies and therefore those that would best suit the farmers; and farmers may be more likely than on-station researchers to identify green manures with food or forage uses that are more likely to be adopted.

A number of approaches naturally lend themselves to farmer-oriented research. Production ecological approaches and conservation farming have both been promoted as approaches to reversing on-farm environmental degradation that take account of soil-water-nutrient interlinkages. A production ecological approach is one way to take account of complex biological linkages such as those between water retention and soil fertility, and between pest management and soil fertility. It requires an understanding of what is happening in the fields to orient research towards technologies that enhance productivity and profitability in an environmentally sustainable way. For example, integrated soil management requires a combination of improved soil hydraulic measures, organic fertility maintenance, and inorganic fertilizer and soil amendments (Batjes, 2001).

Conservation tillage (in which crops are grown with minimal cultivation of the soil) directly affects water infiltration and water retention in the soil, and so improves the efficiency of rainwater use, and may contribute to yield stability and food security in drought prone regions. However, more studies of sufficient size are required to determine the true benefits and constraints to the adoption of conservation farming. For example, conservation tillage has high labor requirements that may deter farmers from adopting the approach. The effectiveness of conservation tillage most likely depends on specific agroclimatic conditions—for waterconserving conservation tillage—and access to draft power influences profitability (and hence the likelihood of uptake). Moreover, the benefits of conservation tillage occur gradually over time, suggesting that poor credit-constrained and risk-averse farmers (a typical SSA farmer) will find it difficult to adopt such techniques without confidence as to their benefits and the ability to make upfront investments—such as through access to credit.

Currently the capacity for integrated soil fertility management in many countries in SSA is limited by insufficient numbers of professional personnel and the essential laboratory facilities required (World Bank, 2002). More integrated approaches require interdisciplinary teams working together, more complex institutional arrangements, and increased coordination among different agencies and organizations, particularly given that governments often separate, for example, agriculture, natural resources, and wildlife agencies. Integrated approaches may also imply new approaches to training and extension. Previously, efforts to undertake research

at the level of large complex systems have tended to result in excess amounts of costly effort to collect data, yielding few results that are of immediate practical value (Campbell and Sayer, 2003).

Livestock. The role of livestock in land degradation has been controversial: Livestock grazing and pastoralism in SSA have often been viewed as a critical factor in the interaction between agriculture and the natural resource base, and overstocking has long been blamed for the cause of extensive land degradation in rangeland areas. For example, some state that overgrazing causes 49% of soil degradation in dryland SSA, while agriculture causes 24%, and overexploitation and forest degradation 27% (Dejene et al., 1997). Many previously proposed solutions to perceived overstocking are now considered to have been misguided. For example, in Tanzania, officials have viewed large herd size and overgrazing as major causes of land degradation and so attempted to enforce destocking and also introduced zerograzing of improved dairy cows for milk. Yet livestock were moved to other areas (rather than numbers being reduced), thereby transferring the problem to different locations and also leading to increased malnutrition (Dejene et al., 1997). A lack of understanding of the social, cultural and economic roles of livestock most likely led to misguided solutions that did not have the intended effect and had overall negative consequences (Box 5-3).

There is increasing evidence that climate, rather than overgrazing, is the key cause of land degradation in rangelands. Climate change is likely therefore to exacerbate the problem of land degradation. For example, long-term monitoring by ILRI (International Livestock Research Institute) in East and West Africa has provided evidence that climate has been the main determinant of changes in arid and semiarid environments and that rangelands are resilient and capable of recovery. Indeed, strong seasonality of rangeland production in the Sahel appears to limit the environmental damage of overgrazing to short periods and confined areas (Ellis, 1992; Hiernaux, 1993).

Recent rethinking of “range ecology” suggests that the opportunistic range land management practiced by pastoral livestock farmers is indeed the appropriate response to natural conditions (Behnke et al., 1993; Scoones, 1995; Homann and Rischkowsky, 2001). Local and traditional management strategies have evolved naturally in response to knowledge of the spatial and temporal availability of natural resources,“and include mobile resource exploitation, flexible stocking rates, and herd diversification, sustained by a system of communal resource tenure” (Sandford, 1983). These strategies, however, may not be able to evolve as rapidly as needed given changing climatic conditions. Nonetheless, they can be integrated into AKST research and development if they are first documented and understood within pastoral livelihood constraints (Oba and Kotile, 2001).

In general, there is insufficient understanding of the role of livestock in livelihoods and the motivations behind pastoralist practices. Better knowledge can be incorporated into the development of technologies and approaches that enable pastoralists to manage their resource base more effectively. For example, approaches that simply encourage lower stock levels may not be sufficient, in part because of farmers’ and