Background/Question/Methods   Plant ecologists traditionally interpreted the plant distributions in terms of adaptations to environmental gradients like altitude, soil fertility, and succession. Whittaker and John Curtis emphasized Gleason's idea that plant species respond individualistically to these gradients and pioneered multivariate methods to describe community variation. MacArthur and Wilson later emphasized how habitat area, isolation, and colonization and extinction processes serve to predict community composition and turnover. Finally, Bell and Hubbell's neutral models explore how stochastic dynamics (drift) alone might account for community composition and dynamics. To explore long-term ecological change and evaluate the relative importance of these forces, we resurveyed more than 200 forest communities originally sampled by Curtis and the PEL lab 50+ years ago. At each site, we characterized over- and understory communities, soil and light conditions, and surrounding landscape features. We are also measuring functional traits on 232 species to infer how these influence local and regional species dynamics.

Results/Conclusions   Although plant responses to succession and edaphic conditions are still evident, these links to local site conditions are weakening. In contrast, variation in plant community composition now depends far more on landscape conditions and processes, including deer herbivory in northern forests and patch size and nearby development in southern forests. Payment of the ‘extinction debt' in southern forests reflects failed colonization more than increased local extinction. Reductions in colonization particularly threaten more sparsely distributed species with limited dispersal. Increases in common native species and invading exotics account for greater declines in beta diversity (‘biotic homogenization') than losses of rare species. Analyses of meta-community structure and dynamics show a strong stochastic element to community composition and dynamics that appears to be increasing. Nevertheless, species differ in functional traits in ways that affect their ability to persist and perform in particular environments and colonize new sites. We using the functional trait data to assess how predictably communities change in ‘trait space', whether these shifts account for changes in community composition, and to infer the particular forces driving ecological change in these forests.