Abstract
The number of species coexisting in a community may be regulated by local factors (e.g., competitive interactions), or by regional processes (e.g., dispersal from a regional species pool). The relative importance of local and regional processes can be inferred from the shape of the relationship between local and regional species richness. We investigated this relationship in communities of fleas parasitic on small mammals at two spatial scales: between the richness of fleas on individual hosts (infracommunities) and that of fleas on host populations (component communities), and between the richness of component communities and that of the entire regional species pool. We tested linearity (proportional sampling) versus curvilinearity with an asymptote (species saturation) by plotting “local” against “regional” species richness of fleas either among host species or within host species among populations. At the two spatial scales, we found consistent curvilinear relationships between species richness of the more “local” communities and richness of the more “regional” communities. This was true across all host species in the data set and for geographic subsets, even after controlling for the influence of sampling effort on estimates of species richness, and that of host phylogeny in interspecific analyses. We also tested for density compensation in species-poor communities. There was no strong evidence for density compensation at the infracommunity level, although its existence at the component community level appeared likely. Our results suggest that identical patterns in local-versus-regional species richness observed on two different spatial scales arise via different mechanisms: infracommunities appear saturated with flea species most likely because of local processes, such as host immune defenses, whereas component communities are saturated with species through interspecific competition, possibly among larval stages.