A major goal of ecologists has been to understand how interactions between species affect the morphology, behavior, and composition of species in a community. Plants and their pollinators have provided model systems for achieving this goal, because the close correspondence between pollinator and floral morphologies has been interpreted by many ecologists as the product of reciprocal selection. Studies of plants and pollinators at the scale of local populations, however, have often failed to find reciprocal selection, calling this interpretation into question.
A recent conceptual advance in coevolutionary thinking has been the recognition that spatial variation in natural selection and gene exchange across a geographic landscape can shape local coevolutionary dynamics, termed the Geographic Mosaic Theory of Coevolution. The proposed research evaluates predictions of the theory through studies of a hummingbird - plant interaction across a Caribbean island archipelago. First, observational and field experiments will test the theory's prediction that interactions between plants and hummingbirds are well-matched and result in reciprocal selection on some islands, but not on others. Second, molecular genetics will be used to examine the theory's prediction that genetic exchange, as well as natural selection, may shape the genetic structure of plants within and between island populations. The collaboration between an ecology professor at an undergraduate institution and a plant systematist at a research institution will train a minimum of eight undergraduate students one graduate student, and workers from four host countries in observational, experimental, statistical, and molecular genetics approaches to evolutionary ecology, and will contribute to the conservation and management of tropical rainforest pollination systems.
