On the ecological significance of pollen color: a case study in American trout lily (Erythronium americanum).

Evolutionary ecologists seek to explain the processes that maintain variation within populations. In plants, petal color variation can affect pollinator visitation, environmental tolerance, and herbivore deterrence. Variation in sexual organs may similarly affect plant performance. Within-population variation in pollen color, as occurs in the eastern North American spring ephemeral Erythronium americanum, provides an excellent opportunity to investigate the maintenance of variation in this trait. Although the red/yellow pollen-color polymorphism of E. americanum is widely recognized, it has been poorly documented. Our goals were thus (1) to determine the geographic distribution of the color morphs and (2) to test the effects of pollen color on components of pollen performance. Data provided by citizen scientists indicated that populations range from monomorphic red, to polymorphic, to monomorphic yellow, but there was no detectable geographic pattern in morph distribution, suggesting morph occurrence cannot be explained by a broad-scale ecological cline. In field experiments, we found no effect of pollen color on the probability of predation by the pollen-feeding beetle Asclera ruficollis, on the ability of pollen to tolerate UV-B radiation, or on siring success (as measured by the fruit set of hand-pollinated flowers). Pollinators, however, exhibited site-specific pollen-color preferences, suggesting they may act as agents of selection on this trait, and, depending on the constancy of their preferences, could contribute to the maintenance of variation. Collectively, our results eliminate some hypothesized ecological effects of pollen color in E. americanum, and identify effects of pollen color on pollinator attraction as a promising direction for future investigation.