Several themes emerged from this conference suggesting how science can be better linked to sustainability. Major themes that need to be addressed include: institutions and change; linkages among different groups involved in the food system, from producers through consumers; indicators of sustainability that encompass “quality of life”; changing the behavior of farmers and researchers to make agriculture more sustainable; and new approaches to research that are suited to alternative agricultures. Throughout the conference these themes were developed as farmers, researchers, and others exchanged their diverse viewpoints and experiences. These viewpoints are reflected both in the formal papers published in this issue and in the comments of panelists that I will draw on in trying to convey the major themes of the conference.

Researchers at land-grant universities are under pressure to widen their research objectives, which requires them to expand their inquiry from a disciplinary base to a systems level, to include farmers and nonuniversity organizations as research collaborators, and to communicate with broader sectors of the public, such as consumers and policy makers. However, there are few rewards for scientists who undertake long-term, systems based research. An approach that brings together multidisciplinary, multiprofession teams using a radial model of organization has successfully addressed these challenges in Wisconsin. A small “hub” manages the team, with the remaining participants linked to component taskforces or “satellite” projects. This enables research to be done on both multidisciplinary and related single-discipline questions. Radially organized teams can attract diverse people with different time commitments and reward incentives. The research meets the criteria of public and academic audiences alike. For innovative research and extension efforts like radial teams to reach their full potential, larger institutional changes are needed within the land-grant system and professional agricultural societies.

Research on ecological agricultural systems should not confine itself to their biological and technological aspects, but should also deal with their social and human implications. Moreover, the goal should be to improve ecological agriculture, not just to document the performance of existing systems. Comparative studies of ecological versus conventional production systems are not adequate to achieve this goal. We need a broad research program that deals with fundamental agroecosystem processes, socioeconomic questions, appropriate livestock production systems, food quality, and information transfer among farmers and others. Research strategies on several levels are needed for this purpose, such as pot and field experiments, case studies, and modeling.