Theory has shown that spatial structure is crucially important in driving virulence evolution: when hosts are more likely to transmit disease exclusively to close neighbors (i.e. local transmission), parasites are expected to evolve lower virulence than when hosts are likely to infect remote individuals (i.e. global transmission). However, this theory remains untested in a real-life field system. This proposal takes advantage of honeybees (Apis mellifera) and their destructive parasitic mites (Varroa destructor) to study the role of spatial structure in virulence evolution. Varroa mites are the single largest cause of honeybee colony losses worldwide, and beekeeping practices are likely to drive this parasite's virulence by routinely altering transmission conditions and population structure. Intensive beekeeping increases global mite transmission and thus has the potential to unintentionally select for devastating parasites. This proposal has three specific aims: (1) an experimental evolution study at an unprecedented scale, which will vary the relative importance of local versus global mite transmission to determine how this affects virulence evolution; (2) development of virulence evolution models to study the role of spatial structure in agricultural systems, which will be applied to the honeybee-Varroa system to make specific recommendations on beekeeping practices to prevent selection of high virulence; and (3) a large scale cross-infection experiment to test whether current beekeeping practices have selected for higher virulence, which will compare mites from intensively managed, lightly managed and feral bees. The combination of large-scale field experiments and theory development on the tractable system of bees and mites will be powerful in developing important insights in the role of spatial structure and host heterogeneity in disease transmission, epidemiology and evolution.