Natural Selection and Genetic Constraints on Flowering Phenology in an Invasive Plant

Theory suggests that the contemporary evolution of local adaptation may increase the rate of biological invasion, yet natural selection has rarely been measured in invasive species. A recently published model predicted that latitudinal variation in the strength of stabilizing selection on two correlated traits—flowering time and size—can result in local adaptation during plant invasion but implicitly assumed that population × environment interactions are weak. We tested for stabilizing selection and the effect of growing environment on 13 populations of the invasive plant Lythrum salicaria sampled along a latitudinal gradient. Plants were grown under uniform glasshouse conditions and in a 3-year common garden field experiment at midlatitude. Consistent with model assumptions, the ranking of population means was highly correlated across the 3 years, and with glasshouse measurements, despite significant interannual differences in flowering time and vegetative size. As predicted, the first principal component (PC1) of flowering time and size was under stabilizing selection, and directional selection was detected on PC2, thus favoring plants that flowered earlier at a larger size. Our results demonstrate the value of using a priori models of natural selection to guide hypothesis testing and support the prediction that fitness trade-offs can constrain adaptive evolution in invasive species.