(Bunch and Lopez, 1994; Hocdé et al., 2000; Hocdé et al., 2002) as were integrated rice-duck farming in Bangladesh (Khan et al., 2005) and the testing and adaptation of agricultural engineering prototypes by farmer members of the Kondomin Group network in Australia. Nongovernment organizations (NGOs), community-based organizations (CBOs), universities and the Consultative Group on International Agriculture Research (CGIAR) played key roles in elaborating effective practice and supporting local FPRE initiatives (Lumbreras, 1992; Dolberg and Petersen, 1997; IIRR, 1996, 2005).

     Participatory Plant Breeding (PPB) is a particular adaptation of FPRE: its client-oriented interactive approach to demand-driven research has been shown to be particularly effective for grains, beans and roots (de Boef et al., 1993; Sperling et al., 1993; Farrington and Witcombe, 1998; CIAT, 2001; Fukuda and Saad, 2001; Chiwona-Karltun, 2001; Mkumbira, 2002; Ceccarelli et al., 2002; Witcombe et al., 2003; Virk et al., 2005). It is a flexible strategy for generating populations, pure lines and mixes of pure lines in self-pollinated crops as well as hybrids, populations, and synthetics in cross-pollinated crops. Biodiversity is maintained or enhanced because different varieties are selected at different locations (Joshi et al., 2001; Ceccarelli et al., 2001ab). Recent assessments of over 250 participatory plant breeding projects in over 50 countries in Latin America, Europe, south and southeast Asia and sub-Saharan Africa led by farmers, NGOs or by national or international researchers or some mix of these actors (Atlin et al., 2001; Joshi et al., 2001; Cleveland and Soleri, 2002; Ashby and Lilja, 2004; Almekinders and Hardon, 2006; Mangione, 2006; Ceccarelli and Grando, 2007; Joshi et al., 2007) demonstrate that PPB is a cost-effective practice that is best viewed along a continuum of plant breeding effort. French researchers, e.g., are working with marker-assisted selection to develop virus resistant rice varieties for Central America and the Cameroon in the context of PPB activities (www .ird.fr/actualites/2006/fas247.pdf). GIS and satellite-based imaging are adding additional value to PPB activities.

     While over 8000 improved varieties of food grains with wide adaptability have been released over a 40-yr period by the CGIAR institutes (Evenson and Gollin, 2003b), PPB has shown capacity to generate multiples of this output for target environments, specific problems and the needs of farmers overlooked by conventional breeding efforts. The three major differences of PPB compared to conventional breeding are that testing and selection take place on the farm instead of on-station; the key decisions are taken jointly by farmers and breeders; the process can be independently implemented in a large number of locations. The activity also incorporates seed production with farmers multiplying promising breeding material in village-based seed production systems. The assessments also highlights the improved research efficiencies and program effectiveness gained by faster progress toward seed release and the focus on the multiplication of varieties known to be farmer-acceptable. Decentralized selection in target environments for specific adaptations allows women's seed preferences to be addressed (Sperling et al., 1993; Ashby and Lilja, 2004; Almekinders and Hardon, 2006). Sustained PPB activity has the additional advantage of bringing about the progressive empowerment of indi

vidual farmers and farmer communities (Almekinders and Hardon, 2006; Cecccarelli and Grando, 2007). However, the tightening of UPOV regulations and the increasing trend toward seed patenting and IPR over genetic material has given rise to concern (Walker, 2007) that despite PPB's demonstrated advantages in a wide variety of contexts and for multiple purposes the space for PPB may be closing.

     As the case of PPB shows, wider scale impact in the case of FPRE relies on the replication of numerous initiatives in response to specific markets and non-market demands rather than on supply-push and diffusion of messages or technologies, although diffusion processes can and do amplify the outcomes of FPRE. The process of replication can be strengthened through investment in farmer-to-farmer networking (Van Mele and Salahuddin, 2005), support to farmer driven chain development (as in poultry or dairy chains serving local markets) and in the creation of "learning alliances" among support organizations that aim to promote shared learning at societal scales (Pretty, 1994; Lightfoot et al., 2002). FPRE has proved to be cost-effective and fit for the purposes of meeting integrated development and sustainability goals (Bunch, 1982; Hyman, 1992) and for natural resource management (NRM) in agrarian landscapes (Campbell, 1992, 1994; Hilhorst and Muchena, 2000; CGIAR, 2000; Stroosnijder and van Rheenen, 2001; Borrini-Feyerabend et al., 2004). However, it has been criticized for failing in specific cases to take advantage of the "best" science and technology available, as self-indulgent by supporting farm systems that some consider insufficiently productive to provide surplus to feed the world's growing urban populations; as sometimes misreading the gender power dynamics of local communities (Guijt and Shah, 1998) and as incapable of involving a sufficient number of small-scale producers (Biggs, 1995; Richards, 1995; Cooke and Kothari, 2001). NGOs and community-based organizations have raised issues of equity. It has also been criticized as too locally focused (see critiques of Australia's Landcare experience in Lockie and Vanclay, 1997; Woodhill, 1999) and thus unable to address higher level economic and governance constraints and tradeoffs. This criticism has prompted recent institutional experimentation with applying FPRE under catchment scale regional development authorities (Australia) and in sustainable water development (South Africa and Europe) (Blackmore et al., 2007) within normative policy frameworks that explicitly seek the sustainability of both human activity and agroecologies.

     Innovations in the organization of knowledge processes also occurred in relation to farmer-developed traditions of agroecological farming (e.g., Fukuoka, 1978; Dupré, 1991; Gonzales, 1999; Furuno, 2001), gathering and domestication of wild foods and non-timber forest products (Scoones et al., 1992; Martin, 1995) and landscape management (Fairhead and Leach, 1996). For example migrants from the Susu community first encountered the rice-growing ethnic Balantes in Guinea Bissau around 1920; later on, the Susu (and the related Baga peoples) hired migrant Balantes to carry out rice cultivation in the brackish waters of coastal Guinea Conakry where the skills are now recognized as traditional knowledge (Sow, 1992; Penot, 1994).

     Indigenous long-standing technologies include the use of Golden Weaver ants as a biocontrol in citrus and mango