Host Trait-Based Patterns of Parasite Species Co-occurrence in Component Communities: A Case Study With Fleas Parasitic on Small Mammals.

We applied a novel eigenvector ellipsoid method to investigate patterns of flea species co-occurrences on small mammalian hosts from six biogeographic realms. We asked whether the application of this method that links host trait-associated requirements of flea species (represented by eigenvector ellipsoids based on the traits of hosts exploited by them) with their among-host distribution would allow us to detect the same patterns as those identified by traditional methods for detecting non-randomness in species co-occurrences. We also asked whether (a) congeneric and heterogeneric species pairs differ in their pairwise pattern of co-occurrence and (b) the patterns of flea species co-occurrences in regional compound communities are associated with the regions' environmental characteristics. Although flea species seem to be randomly associated in the majority of communities, the values of average ellipsoid overlap suggested different degrees of species aggregation rather than segregation. Clear patterns of nestedness and modularity, in among-host flea distribution, were detected in a few communities only. Congeneric species pairs showed higher ellipsoid overlap and longer distances between their centroids than heterogeneric pairs. The relationship between the overlap of host-associated flea niches in a region and regional precipitation was either positive or negative, depending on the realm. The results of the application of the eigenvector ellipsoid method appeared to be more reliable than those produced by traditional methods, which overestimated the degree of species aggregation and underestimated differences in co-occurrence patterns between conspecific and heterogeneric species pairs.