agricultural employment opportunities, women's share in this did not greatly increase between 1990 and 2004 (UN, 2006a). In the general context of rising youth unemployment, young rural women in particular, have difficulty in entering the labor market. Some have argued that the increasing proportion of rural income from non-agricultural sources in Africa is indicative of the failure of agriculture to sustain the livelihoods of the rural poor (Reardon, 1997; Bryceson, 1999; Ellis and Freeman, 2004). There is evidence that the larger the proportion of non-farm to farm income, the larger the overall income.

Where farm size or productivity can no longer sustain the needs of the household, alternative strategies of migration or investment are likely.

Factors which increase vulnerability constitute severe challenges to the sustainability of livelihoods, e.g., population pressure, land and water shortages, declining productivity due to climate change, collapse of soil fertility, unstable and declining market prices. In these circumstances, some family members, often the young men, migrate to urban centers within or outside their country, in search of employment. These decisions are affected by generational and gender relationships (Chant, 1992; Tacoli, 1998; Bryceson, 1999), and contribute to the "feminization" of agriculture (Song, 1999; Abdelali-Martini et al., 2003), and the increasing dependence of poor rural households on remittances for their survival. Increasingly the migrants include young women, leaving the old and the very young on the farm. In some cases, this has negatively affected agricultural production, food security, and service provision. Labor constraints have encouraged investment in technologies and options which are less demanding in labor, e.g., the establishment of tree crops which are profitable with lower labor inputs (Schreckenberg et al., 2002; Kindt et al., 2004; Degrande et al., 2006). Off-farm remittances have in some cases also encouraged broader investments, e.g., in Andean rural communities, remittances are used for small-scale agriculture, living expenses, and construction and home improvements aimed at the agro-tourism industry (Tamagno, 2003). There is also some evidence for other aspects of more sustainable farming at very high population densities and dependence on migrant community members (see 3.2.2.1.6), combining intensification of production and erosion control (Tiffen et al., 1994; Leach and Mearns, 1996).

3.2.3.3 Participation and local knowledge systems

There is a growing body of work that systematically seeks to assess the impacts of participatory and gender sensitive approaches in agricultural research and development, and the role of local knowledge-for example the Systemwide Program on Participatory Research and Gender Analysis Program of the CGIAR (Lilja et al., 2001, 2004).

3.2.3.3.1 Participatory research approaches

Participatory approaches have developed in response to the lack of economically useful, socially appropriate and environmentally desirable applications from AKST generated by agricultural research and development organizations.

There is much evidence that the technological advances of the Green Revolution have sometimes led to environmental degradation and social injustice (Conway, 1997). This has stimulated interest in new participatory approaches, methods and techniques to meet sustainability criteria (Engel et al., 1988) and to contribute to a new development paradigm (Jamieson, 1987) targeting development goals (Garrity, 2004) (see Chapter 2). It has required major advances in the analysis of the behavior of the complex social systems found in rural communities. The growing interest in participatory approaches from the 1980s onwards, was in part a response to the contrast in the successes of Green Revolution technology in some contexts and its lack of, or negative, impact in others, particularly those characterized by high diversity, inaccessibility and weak institutions and infrastructure (Haverkort et al., 1991; Okali et al., 1994; Scoones and Thompson, 1994; Röling and Wagemakers, 1998; Cerf et al., 2000). Participatory approaches, in which development agencies and technical specialists participate, use existing local skills and knowledge as the starting point (Croxton, 1999). They are built around a process that enables farmers to control and direct research and development to meet their own needs and to ensure a sense of ownership in decisions and actions (Engel et al., 1988). The main advantages of participatory approaches have been their responsiveness to local ecological and socioeconomic conditions, needs and preferences; building on local institutions, knowledge and initiatives and fostering local organizational capacity. Criticisms have focused on their resource requirements, the difficulties of scaling-up successes from small focus areas (Cooke and Kothari, 2001), the lack of radical change in institutional relationships and knowledge sharing, and the limited engagement with market and policy actors.

Participatory approaches to genetic improvement of crops and animals results in better identification of farmer's requirements and preferences, leading to higher levels of adoption and benefit.

In cereals and legumes, participatory approaches have been promoted in response to perceived weaknesses in conventional variety testing and formal release procedures which have not delivered suitable varieties to farmers in marginal environments, especially, but not exclusively, small-scale farmers (Witcombe et al., 1998). Formal release systems are often centralized, use a research station or other atypically favorable environment for testing, and select for average performance. Farmers or consumers are also rarely involved in this process. Consequently, varieties from these