Biogeographic and experimental evidence for local scale dispersal limitation in central Panamanian forest birds.

摘要

I examined the avian biogeography of the islands of Lago Gatun, Panama, in an effort to better understand the effects of forest fragmentation in this biodiverse region, and specifically to understand the importance of fragment isolation and the mechanism behind its effects on tropical bird distribution. I combined exhaustive surveys of 29 islands with an experimental release program. Analyses of these data were conducted in a hierarchical fashion to evaluate evidence for avian dispersal limitation at multiple scales. First, I assessed the potential roles of area and isolation in determining avian species richness across this series of fragments, and how these relationships may differ for edge and forest dependent birds. Next, I analyzed community structure and species distributions to determine the relative contributions of island area and isolation in creating patterns of species-specific insular occurrence. Lastly, I assessed the results of the experimental investigation to evaluate the hypothesis that dispersal limitation explains the insular distribution patterns of several tropical forest birds. Isolation is a significant predictor of avian species richness, but only after accounting for the stronger effect of area. Species-isolation relationships are different in this archipelago for birds that rely on forest and edge habitats, respectively. Species-specific distributions are significantly nested when islands are ordered by area, and by isolation once the effect of area is considered. Occurrence of most forest species is sensitive to area and isolation of the archipelago. Examination of guild structure suggests that multiple mechanisms are responsible for these occurrence patterns. There are distinct species-specific differences in ability to cross small gaps, and species which are better able to cross these gaps are more widely distributed across the archipelago than those species that negotiate the same barriers poorly. Species that performed uniformly well in release experiments were much less likely to have suffered insular extinction in the preceding 25 years than those species that showed moderate to poor experimental performance. There is strong evidence of a morphological basis for the patterns. The cumulative evidence from these analyses is the most comprehensive evidence to date of local-scale dispersal limitation in volant birds.