Facultative pathogens, by definition, can persist away from their hosts and thus their capacity to cause disease is greatly impacted by the stability of virulence functions during host association and in their host-independent reservoir(s). Agrobacterium tumefaciens is a facultative pathogen that causes crown gall disease on plants through horizontal gene transfer (HGT) directed by virulence functions that are predominantly encoded on the tumor-inducing (Ti) conjugative plasmid. The Ti plasmid and another conjugative mega-plasmid known as the At plasmid impose significant carriage costs on A. tumefaciens that are exacerbated by nutritional stress. Virulence (vir) gene expression is tightly regulated by plant-released signals. Plant tissues expressing A. tumefaciens transgenes excrete semi-private nutrients called opines that support agrobacterial rhizosphere populations proximal to the infection site. Our studies in laboratory culture have revealed that Ti plasmid fitness costs are very high when virulence is activated, and mutations disabling vir genes, as well as Ti plasmid curing, are common after cultivation under vir-inducing conditions. However, whole genome sequencing of A. tumefaciens populations infecting plants suggests that the bulk of the population does not experience elevated plasmid loss or mutation of virulence functions. Clay Fuqua and graduate student Ian Reynolds hypothesize that there is a small infecting population in intimate physical association with infection sites. This spatially distinct subpopulation drives the HGT process and incurs the fitness burden but is greatly outnumbered by the bystander population that can benefit from the result of pathogenesis and do not experience its cost. A genetic approach is being developed to test this hypothesis and interrogate specifically the subpopulation that is activated for virulence and cross-kingdom horizontal gene transfer.