Genetic and fitness costs of raising wild pollinators in captivity: interaction among species, subpecies and populations of orchard bees

Managed megachilid bees are often established as small genetically uniform populations, whose fate may foretell the costs of ongoing habitat fragmentation on wild pollinator species. Within our small captive population of a managed orchard bee Osmia ribifloris, mtDNA (COI) markers show two populations, each from different subspecies, shared the closest phylogenetic kinship. In fact, the fittest adult females originated from two distant populations, Texas and California, which provisioned the largest and healthiest broods using pollen from southern blueberry cultivars. In contrast, Osmia ribifloris from northern Utah, their Mississippi progeny, along with a floral generalist from east Texas (O. lignaria) produced comparatively fewer brood. Overall, fitness of O. ribifloris also depended on a female’s mating status. For instance, mated bees were more productive than unmated bees. Mated bees nested 2 weeks earlier and produced broods that were twice as heavy. Moreover, 22% of unmated bees actually reduced the fitness of conspecifics through the parasitism of active nests, within which 67% of brood died, a huge loss of potential blueberry pollinators. Such high rates of infanticide as well as a steepening rate of male production indicate a captive bee population in decline. Thus, prompt releases of wild O. ribifloris into blueberry fields with ample forage and nest blocks are preferable to long-term captivity and its risk of unstable secondary sex ratios and facultative nest parasitism.