Vagrant pollinators and fragrant plants - Geographic structure in floral scent despite hawkmoth-mediated gene flow linking isolated populations

Background/Question/Methods

Many bees commonly studied in pollination research are central-place foragers with small home ranges and low potential to carry pollen between isolated populations. In contrast, nocturnal pollinators such as hawkmoths frequently forage between patches, travel longer distances in search of nectar and therefore have the potential to contribute significantly to long-distance gene flow. While long-distance dispersal can constrain divergence between populations, differences in pollinator communities can result in divergent selection on traits important to pollinator attraction and fidelity. Hence pollinators that move long distances can serve as agents of selection while simultaneously constraining divergence through gene flow. Scent is an important floral trait for many nocturnal pollinators, allowing them to locate populations and individual plants, hence floral fragrance may have a significant impact on gene flow through its role in pollinator attraction. Here we compare patterns of neutral genetic variation (microsatellites) to variation in floral scent in Oenothera harringtonii (Onagraceae), an endemic to southeastern Colorado, primarily pollinated by nocturnal hawkmoths, Hyles lineata and Manduca quinquemaculata. We test the hypothesis that hawkmoth-mediated gene flow results in low levels of genetic differentiation and constrains geographic variation in floral scent chemistry.

Results/Conclusions

Consistent with theories regarding long-distance dispersal, our genetic analyses (13 microsatellite markers) provide evidence for high levels of gene flow and little genetic structure among 19 range-wide populations of O. harringtonii. In contrast, data from two field seasons and common garden experiments indicate there is geographic differentiation in floral traits and scent. Populations to the south and east of the range lack linalool, a volatile compound common to many hawkmoth-pollinated plants, but produce PHE-derived compounds (e.g. 2-phenylethanol, phenylacetonitrile), while other populations produce linalool but lack PHE-derived compounds. While O. harringtonii is pollinated primarily by the same hawkmoth species throughout its range, infrequent visits by small pollen-collecting bees are known and allow us to assess the abundance and intensity of interaction with all pollinators. We provide insight into landscape-level processes that may explain these patterns through comparisons of our genetic data, habitat preferences, and landscape characteristics. The autecology of O. harringtonii and its pollinators offers the opportunity for a comprehensive evaluation of geographic divergence in floral traits, evidence of pollinator-mediated selection, and an understanding of the role of gene flow in constraining floral divergence.