Generalist infectious disease agents are widespread among plant and animal hosts. Because hosts often vary in their competence to support and transmit pathogens, community composition may have important consequences for the prevalence of pathogens within the host community. Communities with higher host species richness have been shown to have reduced disease transmission, but this effect may be contingent on the identity of the component host species. We conducted field mesocosm experiments to determine the relative influences of the density of a highly competent host and overall community composition on disease prevalence and community biomass. Using a pool of eight species of grasses known to be susceptible to barley yellow dwarf virus (BYDV), we planted equal-density communities in a factorial design with species richness (seven versus three species) and exposure to BYDV (infected versus uninfected). All mixtures contained Avena fatua, the most competent host out of the species pool. Each level of species richness contained seven replicate community compositions.

Results/Conclusions

Viral prevalence was higher in three-species than in seven-species mixtures, in which the most competent host was at lower density. BYDV prevalence was similar among different compositions within the richness treatments. These results indicate a large influence of the density of the most competent host on viral prevalence in the community. Individual host species varied consistently in viral prevalence throughout the different community compositions, but the presence or absence of hosts of intermediate and low competence contributed less to variation in overall prevalence than the density of our most competent host. The impact of viral exposure varied with species richness, with BYDV altering biomass in three-species mixtures but not in seven-species mixtures. Surprisingly, biomass was higher in infected than uninfected low-richness plots, possibly because viral infection of our most competent host may have caused competitive release of less susceptible hosts in three-species mixtures. Our results support the importance of dominant reservoir species in multi-host systems.