The structure and stability of small mammal faunas.

I examined questions about the local and global stability and resilience characteristics of six small mammal faunas and their relationship to connectance, average interaction strength, community covariance, and indirect competitive pathways. Community matrices estimated elsewhere were used in the analyses. The faunas had from three to nine species and represented several habitat types. The following properties were observed: (1) all community matrices were globally stable, (2) return rates to equilibrium were generally maximized, and (3) competitive interactions were strictly hierarchical. Global stability resulted from a reduction in connectance with increasing species and not from structural characteristics (i.e., community covariance). Average interaction strength did not change with the number of species in these faunas. Increased resilience was due largely to reduced community covariance. Two of the six faunas showed some evidence for indirect pathways, but this appeared to be artifactual. Randomizations of the original community matrices showed that indirect pathways have a high probability of occurrence. Monte Carlo simulations indicated that the probability of indirect pathways increases with increasing number of species or magnitude of competition. These results bring into question previous observational studies invoking indirect competitive pathways. Communities without indirect pathways have a special linear hierarchical structure. This structure has greater resilience than when indirect pathways are present. The observed patterns in community structure are discussed with regard to current theories of habitat selection.