Host species composition, not priority effects, determines infection risk in multihost-multiparasite amphibian communities

Novel parasite emergences are posing uncertain threats to wildlife communities worldwide. Because parasites typically infect multiple host species and interact with other parasites, the degree of threat posed by these emergences is known to depend both on a communitys host species composition and history of exposure to other parasites. What is not known is how these two factors - multi-host effects and multi-parasite effects - interact to shape the outcomes of parasite emergences. We experimentally investigated the outcomes of sequential emergences by two amphibian parasites of major concern to global biodiversity, Batrachochytrium dendrobatidis (Bd) and Ranavirus (Rv). Mimicking sequential Bd and Rv emergence in communities comprising either larval common toads (Bufo bufo), larval common frogs (Rana temporaria), or both host species, showed that host community composition influenced subsequent patterns of Bd and Rv infection much more than the presence of the other parasite. Bd prevalence varied predictably with host community composition; toads exhibited consistently higher Bd infection probabilities than frogs across all scenarios. Variation in community-level prevalence therefore was explained by differences in the relative abundance of toads. In contrast, community-level Rv prevalence was not consistently driven by a single host species. Going from single-host to two-host communities reversed relative Rv infection probabilities in frogs versus toads; when kept separately, frogs had a higher Rv infection probability than toads, but toads had a higher infection probability than frogs when the two species were co-housed. Host species composition can therefore alter multi-parasite infection dynamics not only through changes in the abundance of primary host species, but also through changes in which species appear to act as the primary hosts. These insights suggest that Bd-Rv dynamics in single-host scenarios may not reflect those occurring in natural multihost communities, and the observed context-dependency in the apparent primary host species for Rv highlights an overlooked factor potentially shaping biodiversity-disease relationships.