Much is known about natural genetic variation controlling the onset of flowering in annual Arabidopsis thaliana. The situation is more complicated, however, in short-lived perennials, where age at first reproduction is a fundamental life history trait reflecting resource allocation as well as time of flowering. We are studying local adaptation of flowering time across an elevation gradient in the Northern Rocky Mountains, where populations show a significant negative correlation between flowering time and their elevation of origin. Thus, high elevation populations experience short summers under natural conditions, and these genotypes flower earlier under controlled conditions. In order to understand the evolutionary forces which influence this ecologically important variation, we are studying molecular polymorphism for flowering time genes within populations, across an elevation gradient, and across a latitudinal gradient from Colorado to Montana.