Habitat fragmentation is the process by which large areas of continuous habitat are divided into smaller pieces (i.e. "habitat islands" or "patches"), usually due to human activity. Understanding why some species are intolerant of habitat fragmentation and disappear from isolated patches, whereas other species are widely distributed and maintain a presence even in remote fragments is a central issue in both basic community ecological research and conservation biology. A key hypothesis is that differences in dispersal abilities among species strongly influence distribution patterns. The connection between how easily species move between patches and their presence in isolated communities has often been hypothesized but rarely experimentally evaluated. The goal of this project is to uncover the relationship between dispersal ability and community composition for tropical understory forest birds in an archipelago of forested islands. This will be accomplished by evaluating the Restricted Dispersal Hypothesis, which states: Variation in behavioral or physiological traits among species of forest understory birds, such as willingness or ability to cross a habitat gap, explains biogeographic patterns of island occupancy. To assess dispersal ability, I will conduct a series of simple experiments where birds are transported by boat to various distances away from mainland forest and released to document their abilities to fly successfully back to mainland. Preliminary results show a strong positive correlation between the maximum distance flown in release trials and the distance to the most isolated island inhabited by each species. Broader scientific impacts are clear: connectivity of fragments in tropical landscapes is critical to ensure persistence of entire regional avifaunas. Without corridors of habitat through which birds may move among patches, fragmented landscapes will lack the subset of avifaunas which are poor dispersers.