Biodiversity has the potential to constrain (i.e. dilution effects) or facilitate (i.e. amplification effects) disease spread. The understanding of disease-diversity relationships is crucial for adequate management and prediction of disease in humans, wildlife and farmed species, in particular at a time of unprecedented biodiversity loss and rise in emergent infectious diseases. Theoretical models suggest that as biodiversity increases both dilution and amplification are to be expected. Yet empirical data often supports dilution being a more frequent outcome. The mechanisms behind these patterns are an issue of ongoing debate. Nonetheless, previous work has taken a predominantly host-centered approach and has often overlooked the aggregated distribution of parasites. Here we assess whether increasing parasite infracommunity diversity (Shannon’s diversity index and richness) affects the abundance of one species of ectoparasite (Gyrodactylus spp.) and four morphospecies of encysting endoparasites, in two wild poeciliid host fishes (Poecilia reticulata and Poecilia picta, n=217 and n=210 respectively) collected throughout their native range. Furthermore, we use mathematical models to simulate alternative parasite distributions (from a highly aggregated to random distribution) and assess the degree to which aggregation and sampling effects influence disease-diversity relationships.
Results/Conclusions
We find that all ecto and endoparasites are highly aggregated (k=0.04 to 0.36), and that as parasite infracommunity diversity increases, in both host species, the abundance of particular parasites often increases (P. reticulata - Shannon’s: p<0.01-0.019, richness: p<0.01; P. picta - Shannon’s: p<0.01 and richness: p<0.01, for three out of five parasite species). Thus our empirical results support amplification being a more common outcome of increasing parasite diversity. Nonetheless, our simulations show that randomly assembled infracommunities also lead to amplification, but that the likelihood of finding dilution effects increased with decreasing sample sizes and decreasing parasite aggregation. Taken together, our results indicate that amplification might be a more likely outcome of increasing parasite diversity; that the occurrence of dilution may be explained by sampling effects; and that disease-diversity outcomes may often be driven by parasite distribution patterns rather than by species interactions. Furthermore, our results underscore the importance of directly accounting for parasite distributions and sample size interactions in empirical studies. We conclude by presenting a synthesis of how parasite aggregation can explain the high frequency of dilution effects with increasing host diversity.