Abstract
Since species vary in abundance and host competence (i.e., ability to get infected and transmit a pathogen), changes in species composition caused by biodiversity loss impacts disease dynamics. Forecasting effects of species composition on disease depends on community (dis)assembly, processes determining how species are added to (or lost from) communities. We simulated community assembly by planting mesocosms, nested along a richness gradient, and tested how relationships between richness, species assembly order, and overall density affect disease risk. Mesocosms with up to six crop species of varying competence were inoculated with a soilborne fungal pathogen, Rhizoctonia solani . Disease was measured as species‐level prevalence, community‐level prevalence, and total number of diseased plants. Regardless of metric, richness limited disease when species assembly order negatively correlated with competence and total density remained unchanged with richness. When density increased with richness or species assembled randomly, richness primarily correlated positively or weakly with disease. Our results align with theoretical expectations and represent the first empirical study to test the influence of species densities, assembly order, and competence on diversity–disease relationships.