Pollinator community species richness dilutes prevalence of multiple viruses within multiple host species

Most pathogens are embedded in complex communities composed of multiple interacting hosts, but we are still learning how community-level factors, such as host diversity, abundance, and composition, contribute to pathogen spread for many host-pathogen systems. Evaluating relationships among multiple pathogens and hosts may clarify whether particular host or pathogen traits consistently drive links between community factors and pathogen prevalence. Pollinators are a good system to test how community composition influences pathogen spread because pollinator communities are extremely variable and contain several multi-host pathogens transmitted on shared floral resources. We conducted a field survey of four pollinator species to test the prevalence of three RNA viruses (deformed wing virus, black queen cell virus, and sacbrood virus) among pollinator communities with variable species richness, abundance, and composition. All three viruses showed a similar pattern of prevalence among hosts. Apis mellifera and Bombus impatiens had significantly higher viral prevalence than Lasioglossum spp. and Eucera pruinosa. In each species, lower virus prevalence was most strongly linked with greater pollinator community species richness. In contrast, pollinator abundance, species-specific pollinator abundance, and community composition were not associated with virus prevalence. Our results support a consistent dilution effect for multiple viruses and host species. Pollinators in species-rich communities had lower viral prevalence than pollinators from species-poor communities, when accounting for differences in pollinator abundance. Species-rich communities likely had lower viral prevalence because species-rich communities contained more native bee species likely to be poor viral hosts than species-poor communities, and all communities contained the highly competent hosts A. mellifera and B. impatiens. Interestingly, the strength of the dilution effect was not consistent among hosts. Instead, host species with low viral prevalence exhibited weaker dilution effects compared to hosts with high viral prevalence. Therefore, host species susceptibility and competence for each virus may contribute to variation in the strength of dilution effects. This study expands biodiversity-disease studies to the pollinator-virus system, finding consistent evidence of the dilution effect among multiple similar pathogens that infect "replicate" host communities.