Consistent differences in the physiological performance of wild populations of closely related plant taxa may be the result of environmentally induced phenotypic plasticity or adaptive evolution (or a combination of the two). Here we report the results of a field study of physiological and fitness-related traits in geographically proximate sister taxa in the annual wildflower genus Clarkia (Onagraceae) and interpret the differences between them in light of their ecological and reproductive differences. Within two pairs of taxa, the predominantly autogamous (self-fertilizing) taxon flowers and completes its life cycle before its pollinatordependent (predominantly outcrossing) counterpart growing in sympatry or at similar elevations in the southern Sierra Nevada. Selfers generally exhibited higher rates of photosynthesis and transpiration than their outcrossing sister taxa, and, except for the earliest-flowering (autogamous) taxon, both photosynthetic and transpiration rates tended to decline as the season progressed. Within taxa, high photosynthetic rates were positively correlated with lifetime fruit production, and selfers had lifetime fruit production equivalent to or higher than that of outcrossers, despite the fact that the latter had higher aboveground stem biomass. These patterns are consistent with the hypothesis that natural selection has favored higher gas exchange rates in selfers to allow them to achieve their faster life cycles and so escape seasonal late-spring drought. An alternative explanation is that the differences in gas exchange rates represent environmentally induced plastic responses to the cooler temperatures and higher soil moisture content in early spring. Further experimental work is necessary to distinguish between these hypotheses.

No abstract

We detected several examples where selection was consistent with the phenotypic divergence between sister taxa, but there were also numerous instances that were equivocal or in which selection did not predict the realized phenotypic difference between taxa.

Pollen attrition was mediated by style length, but the function of style length was primarily to increase the number of germinating pollen grains, which affected attrition rates either through stigma clogging or pollen-pollen interactions. Style length may have a direct effect on pollen receipt due to the stigma's position relative to pollinator body parts, but traits correlated with style length may also directly affect pollen receipt.

We tested three predictions regarding the joint evolution of pollen performance and mating system. First, due to the potential for intense intrasexual competition in outcrossing populations, we predicted that outcrossers would produce faster-growing pollen than their selfing relatives. Second, if elevated competition promotes stronger selection on traits that improve pollen performance, then, among-plant variation in pollen performance would be lower in outcrossers than in selfers. Third, given successive generations of adaptation to the same maternal genotype in selfers, we predicted that, in selfing populations (but not in outcrossing ones), pollen would perform better following self- than cross-pollinations. We tested these predictions in field populations of two pairs of Clarkia (Onagraceae) sister taxa. Consistent with our predictions, one outcrosser (C. unguiculata) exhibited faster pollen germination and less variation in pollen tube growth rate (PTGR) among pollen donors than its selfing sister species, C. exilis. Contrary to our predictions, the selfing C. xantiana ssp. parviflora exhibited faster PTGR than the outcrossing ssp. xantiana, and these taxa showed similar levels of variation in this trait. Pollen performance following self- vs. cross-pollinations did not differ within either selfing or outcrossing taxa. While these findings suggest that mating system and pollen performance may jointly evolve in Clarkia, other factors clearly contribute to pollen performance in natural populations.