Geographic variation as a proxy for climate change: Forecasting evolutionary trajectories from species differentiation and genetic covariance

Background/Question/Methods:   Climate change models for California predict a warmer and drier future, potentially resulting in shorter growing seasons and stronger natural selection favoring traits associated with drought-escape, such as early or rapid flowering. In this study, we evaluated variation in floral and life history traits within and among populations of two annual herbaceous Clarkia species (the highly outcrossing C. unguiculata and its facultatively autogamous sister species, C. exilis) to determine whether life-history traits that promote early completion of the life cycle are genetically correlated with floral traits associated with self-fertilization (e.g. small flowers and short anther-stigma distances). If so, then if selection favors genotypes that complete their life cycle early, self-fertilization may evolve as a correlated trait, along with its genetic risks. We examined variation and covariation among life history and floral traits within and among 11 populations sampled across an elevation gradient in the southern Sierra Nevada, California.  Under greenhouse conditions, we cultivated maternal families from each population to examine the relationships between mean phenotype, elevation, and climatic conditions of each sampled location.  We then estimated covariation among family means in each species.  Our primary aims were to determine whether: (1) differences between species are genetically based and associated with climatic differences among the sampled sites; and (2) genetically based correlations between traits are sufficiently strong that, if selection favors genotypes with reduced exposure to late-season drought, then smaller flowers or self-fertilization will evolve as a correlated response.

Results/Conclusions: In the southern Sierra Nevada, C. exilis populations occupy warmer, drier sites than C. unguiculata; in the greenhouse, the former flowers earlier and faster, and produces smaller flowers with shorter anther-stigma distances.  In both species, among maternal sibships, petal area and the rate of sequential flower production were positively correlated with days to first flower; these correlations mirror the relationships between species.  If selection favors life history or developmental traits that promote drought avoidance, then smaller flowers, and potentially higher rates of self-fertilization could evolve as a correlated response. These results pertain most strongly to late-spring wildflower species in arid regions and suggest that the rapid evolution of selfing could be an unrecognized risk associated with climate change.