Spatial scale break in ecological strategies for host use by plant viruses.

Processes that involve performance traits such as host range underlie pathogen diversity and infection risk. Host range performance (HRP) is expected to be distributed across spatially explicit conditions, but rarely considered in microbial ecology. Merging of spatially discrete processes in epidemiological model selection exacerbates error propagation. We combine high throughput sequencing and reverse transcriptase polymerase chain reaction approaches to test whether HRP of 18 plant virus species influences species diversity across spatial scales. The results indicated a scaling break in the effect of HRP between the individual host and community scales. Regression analyses showed weaker (i.e., random) associations between viruses in individual hosts than at broader spatial scales of cooccurrence among hosts. Contrasting variances in effect sizes between the spatial scales indicated HRP informs on disease risk at the plant community level, but is a poor predictor of infection in individual plants. The evidence of contrasting HRP between the domains of scale set in this study, suggests heterogeneity among multiple processes across scales drive virus species diversity. Epidemiological model selection should consider variation in species trait performance across scales.