Local knowledge, and the local institutions associated with it, have been regarded as an important foundation for community- based natural resource management and biodiversity conservation. However, this has been challenged as a romantic view, dependent on conditions of low population density, lack of modern technology and limited consumer demand (Attwell and Cotterill, 2000). The overexploitation of natural capital has been widely attributed to a number of factors, including the loss of social institutions at the community level. In some cases this arises from changes in local systems of administration and governance. In India, the breakdown of regulations on livestock resulted in unregulated grazing (Pretty and Ward, 2001), while water resource degradation followed the replacement of collective irrigation systems by private ownership. Similarly, the failure of many formal attempts to halt rotational shifting cultivation in Thailand, Laos and Vietnam was, at its most fundamental level, associated with differing perspectives. That is, "policy makers believed that shifting cultivation was the main cause of environmental problems such as floods and landslips" (Bass and Morrison, 1994) whereas others recognized the dynamic and diverse types of shifting cultivation in which farmers engaged, and the associated economic, social, cultural and environmental values.

Institutions are crucial for sustainable development; the innovation systems approach offers more insights than previous paradigms into the complex relationships of technology development and diffusion.

The linear model of research and extension in which innovations are transferred as products from researchers to farmers via intermediaries in extension, has been challenged by experience showing that the pathways for technical changes are more diverse. In the last 15 years, the importance of knowledge in innovation processes has been more clearly recognized (Engel and Röling, 1989; Röling, 1992). Knowledge is considered as a factor of production; considered by some to be more important than land, capital and labor. More recent approaches view innovation as a complex social process (Luecke and Katz, 2003) which takes multiple forms and involves the participation and interaction of a diversity of key actors and organizations (Sibelet, 1995; Spielman, 2005). These relationships or networks, "the innovation system", operate within specific institutional and cultural contexts. Similarly, evaluation approaches have shifted from focusing on impacts of research to tracking the institutional changes and effective operation of the innovation systems (Hall et al., 2003). The innovation systems approach emphasizes continuous learning and knowledge flows, interaction of multiple actors and institutional change. Innovation Systems thinking has encouraged greater awareness of the complexity of these relationships, the processes of institutional learning and change, market and non-market institutions, public policy, poverty reduction, and socioeconomic development (Hall et al., 2003; Ferris et al., 2006). However, the approach does not explicitly engage with poverty and development agendas by examining the relationship between innovation systems, economic growth and the

distributional effects on poverty reduction and policy options which would support this (Spielman, 2005).

Devolution of resource management to local institutions has been successful where targeted support and enabling conditions were in place.

Local institutions have the capacity to manage local resources and avert possible "tragedies of the commons" (Ostrom, 1992). Rules can be created to accommodate the heterogeneity found within communities (Agrawal and Gibson, 1999; Ostrom, 2005) and there are opportunities for comanagement with government (Balland and Platteau, 1996). In conservation programs, the participation of the range of stakeholders in consensus building and consideration of benefit distribution reduces the risk of conflict and the costs of implementation and control, and increases the chances of sustainability (Borrini-Feyerabend, 1997; Guerin, 2007). In some cases, (e.g., the transfer of irrigation management to communities) the drive to establish local management has led to rigid, hierarchical user associations with functional and democratic shortcomings (Agrawal and Gupta, 2005). However, research in the irrigation sector has identified that a supportive legal policy framework, secure water rights, local management capacity development and favorable cost/ benefit relationships, are conditions favoring the successful transfer of management to communities (Shah et al., 2002). These characteristics encourage farmers' contributions and create a strong sense of ownership, which together lead to better subsequent operations and maintenance (Bruns and Ambler, 1992). Finally, research has shown the diversity and complexity of water rights in many developing countries and the importance of recognizing both formal legal rights and customary or indigenous rights in a "pluralistic" approach (Bruns, 2007).

Local or informal seed systems provide most seed used by farmers and are increasingly being used to deliver new varieties to farmers.

Nearly all developing country farmers depend on their own seed, or seed obtained locally from relatives or markets, for planting (Almekinders and Louwaars, 1999; Tripp, 2001). In contrast, most new varieties released in developing countries originate from public sector organizations, Hybrid maize is the exception; it originates from the private sector and seeds are delivered through commercial networks (Morris, 2002), although these are not tailored to specific local situations. Local seed systems are therefore very important. Typically they support the local economy and are very robust and effective. Studies in India have shown that seed can move many kilometers through these informal systems, and that local entrepreneurs quickly act to meet a demand for seed (Witcombe et al., 1999). Consequently, a number of initiatives have built on informal seed systems to distribute seed. For example, relief agencies promote these systems by using seed vouchers in times of drought or civil