Abstract
Community ecology has struggled to provide strong predictive models that link environmental drivers with community structure. Greater progress could be made by focussing on the functional traits represented in communities rather than the identities of species. We are specifically missing analyses of trait diversity at large spatial scales where dispersal between sites is rare, so that we cannot determine if functional diversity in general is determined by niche-based processes, or limited by dispersal, evolution, or biogeography. Our focal system (the faunal food web inhabiting water-filled bromeliad leaves) has been exhaustively documented by us in 12 regions in the neotropics, and the dataset (ca. 400 taxa; > 600 bromeliads; nine relevant traits; environmental variables) has been collated in a database. Our fundamental question for a new trait-based paradigm is: which processes determine the functional diversity of food webs at different spatial scales? We ask 3 sub-questions: (i) Are food webs in a single bromeliad non-randomly assembled in terms of functional traits? (ii) Within a site, does the distribution of functional traits change predictably over environmental gradients? (iii) is there convergence in the functional structure of food webs over a broad biogeographic? Based on abundance- or biomass-weighted trait matrices, we use null models and partial Mantel tests to evaluate functional trait convergence and divergence. /n
France