The Center for Sustainability and the Global Environment (SAGE) at the University of Wisconsin has been developing global databases of contemporary and historical agricultural land use and land cover. SAGE has chosen to focus on agriculture because it is clearly the predominant land use activity on the planet today, and provides a vital service?i.e., food?for human societies. SAGE has developed a ?data fusion? technique to integrate remotely-sensed data on the world?s land cover with administrative-unit-level inventory data on land use (Ramankutty and Foley, 1998; Ramankutty and Foley, 1999; Ramankutty et al., in press). The advent of remote sensing data has been revolutionary in providing consistent, global, estimates of the patterns of global land cover. However, remote sensing data are limited in their ability to resolve the details of agricultural land cover from space. Therein lies the strength of the ground-based inventory data, which provide detailed estimates of agricultural land use practices. However, inventory data are limited in not being spatially explicit, and these data are also plagued by problems of inconsistency across administrative units. The ?data fusion? technique developed by SAGE exploits the strengths of both the remotely-sensed data as well as the inventory data.