Context-dependent selection and genetic facilitation and constraint on rosette diameter and herbivore resistance across european outdoor common gardens under ambient and reduced precipitation in Fragaria vesca
The expression of plant defensive traits against herbivores often incurs costs to other essential functions, such as growth and reproduction. Understanding how selection acts on putatively functional traits that are expected to trade off across space and time is therefore critical for predicting evolutionary responses to ongoing and future environmental change. Here, we used multiple replicated genotypes of woodland strawberry (Fragaria vesca; Rosaceae) grown over two years in three outdoor common gardens in Spain, Belgium, and Sweden. In each garden, genotypes were exposed to both a reduced-precipitation treatment simulating drought and an ambient precipitation treatment. We estimated directional and correlational selection on rosette diameter (a proxy for growth) and herbivore resistance (measured as the inverse of chewing damage) using fruit and stolon production as proxies for sexual and asexual fitness across all environments (i.e., every site x year x treatment combination). We then combined selection gradients with environment-specific genetic (co)variance among genotypes to quantify the expected response to selection ({Delta}z = G{beta}) and to identify covariance-driven constraints or facilitation therein. Selection consistently favored larger rosette diameter for both fitness proxies across nearly all environments, which, in combination with genetic covariances among genotypes, resulted in a general evolutionary response toward increased rosette diameter, with the strongest response at the wettest site (Belgium). In contrast, selection on resistance and the corresponding among-genotype evolutionary responses were strongly context-dependent. Correlational selection on rosette diameter x resistance occurred in only a few environments, primarily under reduced precipitation. Environment-dependent genetic covariances constrained or facilitated selection on both traits only at the site with the highest herbivory (Sweden) under drought conditions. Overall, our results reveal a context-dependent interplay between selection and genetic architecture, underlining the difficulty of predicting evolutionary trajectories under environmental change, and highlighting how spatially and temporally variable conditions may maintain standing genetic variation in plant traits.