This research will document an extreme foraging specialization that is unique to diverse tropical bird communities, and then test specific hypotheses concerning the maintenance of this specialization. A dozen bird species at each of two Amazonian rainforest sites were found to forage primarily by extracting arthropods from curled dead leaves suspended in the forest understory. Because the dead leaves represent discrete resource patches, this research can quantify leaf availability and use, as well as the abundance of available prey. A preliminary empirical study suggests the following predictions to be tested in a series of outdoor cage experiments on wild-caught birds and arthropods: 1) dead leaf specialists have fixed preferences for dead leaf vs. other substrates, but are flexible with respect to particular types; 2) specialists are more efficient than co-occurring generalists at capturing prey hiding in dead leaves; 3) dead-leaf arthropods have fixed preferences for hiding in these leaves; and 4) these arthropods exhibit stereotyped behavioral responses to bird predation that in turn influence the bird's foraging efficiency. This study differs from most other insectivorous bird studies in two important ways; first, food availability and diets are measured directly rather than inferred from feeding behavior or morphology, and second, the amenability of this resource system to experimental manipulation provides a rare opportunity to evaluate the role of the various processes that may promote specialization.
