We document a facultative Bartonella-like Rhizobiales bacterium in the giant tropical ant, Paraponera clavata. In a lowland tropical rainforest in Costa Rica, 59 colonies were assayed for the prevalence of the Bartonella-like bacterium (BLB), 14 of which were positive. We addressed three questions: First, how does the prevalence of BLB within colonies vary with environmental conditions? Second, how does diet affect the prevalence of BLB in P. clavata? Third, how does the distribution of BLB among colonies reflect ambient differences in food resources and foraging habits? A variety of environmental variables that may be predictive of the presence of BLB were measured, and diet manipulations were conducted to test whether the prevalence of BLB responded to supplemental carbohydrate or prey. The ambient frequency of BLB is much higher in young secondary forests, but is nearly absent from older secondary forests. The prevalence of BLB inside field colonies increased over the duration of a 2-week carbohydrate supplementation; however, water and prey supplementation did not alter the prevalence of BLB. The diets of the colonies located in young secondary forest, compared to other habitats, have a diet richer in carbohydrates and lower in prey. The abundance of carbohydrate, or the relative lack of N, in a colony's diet influences the occurrence of the BLB microbe in P. clavata. As experimental diet manipulations can affect the facultative presence of an N-cycling microbe, a consistent diet shift in diet may facilitate the emergence of tighter symbioses.