Natural selection in Potentilla glandulosa revisited

Clausen, Keck and Hiesey’s studies of clonal transplants during the 1930s and 1940s are famous for demonstrating ecotypic variation. Less appreciated are their studies on natural selection, even though at the time there was little hard evidence of its operation in nature. Here we present new analyses of their study of natural selection in Potentilla (Clausen et al., 1940; Clausen and Hiesey, 1958). Although they concluded from their data that natural selection was responsible for the phenotypic differences between altitudinal races of Potentilla glandulosa, many of their conclusions lacked statistical support. We present the results of analyses of data for 107 F2 clones planted at each of three elevations, indicating significant support for their inference of natural selection in the wild, as well as revealing new findings. Survival at the high-altitude Timberline station was highest for those F2 clones with multivariate phenotypes most similar to the local alpine subspecies and lowest for clones with phenotypes most similar to the mid-elevation parent. There is evidence of strong directional selection on characteristics of plant size, but targets and intensity of selection varied among stations. At Stanford, more and longer stems were favoured together with later flowering. At Mather, selection was mainly on rosette width. At Timberline, selection was strongest and had varied targets: directional selection favoured larger rosettes, more stems and later flowering, but there was strong correlational selection among these traits as well. We also present evidence for selection on phenotypic plasticity: stabilizing selection on plasticity of flowering time, as well as selection favouring decreased plasticity of stem length and stem number.