This interdisciplinary project traced the effects of the introduction, spread, and abandonment of agriculture at six U.S. long-term ecological research (LTER) sites, with cross comparisons in Mexico and France. Agrarian transformations represent the most pervasive alteration of the Earth's terrestrial environment during the past 10,000 years. Many current conceptualizations of these transformations, however, assume a simple linear model -- change is driven by present-day economic, demographic, and technological conditions. This project incorporates a more integrated and long-term cycle: of land-use change affecting landscapes, of altered landscapes affecting ecological processes, and of both influencing the ways in which humans monitor and respond to their surroundings, engendering further cycles of change. The central objective of this research is to identify and quantify the ways in which these integrated cycles differ across cultures, across biogeographic regions, and across time. A suite of quantitative and narrative analyses will be used to identify the prime determinants of long-term dynamics, present-day patterns, and reservoirs of ecological and social resilience in these systems. Analytical approaches will include structural-equation modeling, analysis of spatial and causal effects, and cross-site comparisons of case studies. As a practical test of the project's results, approaches and insights will be examined in the context of conservation planning at The Nature Conservancy (TNC) that includes an emphasis on eco-regional planning and scenario building. This investigation will contribute to both science and society in six ways. First, it will demonstrate the importance of social-science information and approaches in ecosystem investigations, expanding the results of the LTER network and breaching the divide between social and natural science. The data protocols developed will also benefit other communities of social and natural scientists through the involvement of ICPSR, the main national repository of social science data. Second, this project will help to develop general theories on how socio-ecological legacies, as well as lags in the recognition of and response to change, vary across space and time. Third, through detailed case histories and quantitative analyses, the project expects to provide convincing evidence that humans act not only to disturb ecosystems, but also monitor ecosystem values and respond to maintain stability and minimize crises. Fourth, project results will provide information of direct use to policy makers, TNC, and land managers by using an approach that explicitly relates socio-ecological processes to varying levels of political organization. Fifth, the cross-scale data collection and analyses are expected to demonstrate that some patterns of human-ecological interactions are surprisingly long term, vary across space and time, and are non-linear. The greatest contribution will be through education at a variety of levels; this project will train new interdisciplinary scientists at all levels of the educational spectrum, inform public officials, and contribute to more effective land management practices.