Relative contributions of ecological drift and selection on bat community structure in interior Atlantic Forest of Paraguay.

Despite extensive focus on numerous mechanisms that potentially structure Neotropical bat communities, understanding of the relative importance of any is still illusive. Recently, it has been demonstrated that all mechanisms used to explain community organization can be conceptualized as one or a combination of the few higher-level processes of dispersal, drift, selection and speciation. These four higher-level processes have not been addressed equally by Neotropical bat community ecologists. In particular, predictions formulated from a hypothesis of ecological drift have not been tested for any Neotropical bat community. Herein we contrast efficacy of predictions based on the higher-level processes of drift and selection in describing community structure of bats in the Atlantic Forest of eastern Paraguay. Predictions apply to species-environment interactions, patterns of trait variation and beta-diversity, predictability of dominant competitors and responses to seasonality. At best, there was inconsistent support for the operation of either drift or selection within this bat community. Selection, however, had more various forms of support including strong species-environment relationships, predictable patterns of dominant competitors and strong responses to seasonality. Despite stronger support for selection, a number of predictions of drift were supported as well. It is likely that a combination of both of these processes operates across the variable environments experienced in Atlantic Forest. Predictions of both processes are difficult to make operational. Support for drift often comes from failure to demonstrate a significant pattern and should not be considered strong support of a prediction. Similarly, many predictions of selection predict phenotypic patterns among species without specifying a particular trait. This is problematic because the phenotype is multifaceted and a lack of pattern in one measured trait might mask a strong pattern in some other unmeasured trait. Distilling mechanisms of community organization into four higher level processes is a substantial innovation in community ecology. Nonetheless, efforts need to be made to develop a suite of mutually exclusive and falsifiable predictions to facilitate future and more rapid understanding of community organization.