The evolution of reproductive isolation (RI) is essential to the formation of new species. As species are moving at increasing rates due to ongoing human-caused global change, it is important to understand the mechanisms that create and maintain RI to determine implications of potential gene flow in areas of secondary contact. The walnut-infesting Rhagoletis suavis species group shows evidence of species movement over the last four decades with areas of co-occurence. Pre-mating, behavioral isolation is predicted to be strong in R. suavis walnut flies due to differences in morphology and courtship behavior between species. However, previous studies have shown no evidence of pre-mating isolation in R. zoqui and R. completa in Mexico. We aimed to characterize species distributions and determine the behavioral, sexual and ecological mechanisms that contribute to RI in walnut flies in the United States. We sampled 69 sites across the range of walnut species in the U.S. to determine current distributions and to compare to collections from the 1980s. We used reciprocal, pairwise mate trials to measure behavioral, premating RI and observed egg hatch rates as a mechanism for post-zygotic RI. We measured post-diapause emergence timing in the laboratory as a potential source of ecological displacement.
We found that walnut fly species distributions have changed significantly since the 1980s. In the Southwestern U.S., R. juglandis has moved to higher elevations into the range of R. boycei. In the Midwestern U.S., we found that R. completa has moved approximately 100 miles east, displacing R. suavis. Behavioral, premating isolation was surprisingly incomplete, with heterospecific matings occurring in all species combinations tested. Positive assortative mating was significant for some, but not all species pairs. With R. completa and R. juglandis, we found fewer heterospecific matings in populations that co-occur than in isolated populations and decreased egg hatching rates for hybrids, suggesting that reinforcement and post-mating mechanisms may play a larger role in RI than previously thought. Post-diapause emergence timing differed by species, but was best explained by environmental conditions and host species. Generalized linear mixed models suggest that co-occurrence may affect the emergence timing in R. completa, but not other walnut fly species. To understand the implications of secondary contact and species movement on the future of R. suavis flies, further study is needed to determine amounts of gene flow and the role of interspecific competition in areas of co-occurrence.