An essential challenge in ecological genetics is to understand the origin and maintenance of phenotypic plasticity — the capacity of some genotypes to alter one or more phenotypic traits in response to environmental variation. Meeting this challenge requires integrative methods for characterizing phenotypes, phenotypic costs, and plasticity costs at the trait and whole-organism levels. Currently, a pressing task is to interpret results properly with information about the physiological mechanisms and genes that regulate the focal traits. We have accomplished this using Arabidopsis thaliana, in which flowering time traits in some genotypes show strong plasticity to over-wintering temperatures (vernalization response); other genotypes lack this response. We have analyzed data from lab studies with four recombinant inbred (RI) populations and from parallel field studies with two of these four populations. With each data set, we have carried out selection gradient analyses, tests of the adaptive plasticity hypothesis, and analyses of plasticity costs.

Results/Conclusions

Selection, when detected, consistently favored chronologically earlier flowering and was stronger when populations were not exposed to sufficient vernalization. Selection on the chronological time of flowering was sometimes reinforced but sometimes offset by selection on correlated traits (e.g., developmental stage or size at flowering). These results contradict other published studies arguing that this syndrome of plasticity is adaptive. While there is some evidence that searching for plasticity costs in RI populations is more likely to find them, our results contribute to mounting evidence that plasticity costs are generally weak, and local rather than global. Despite their prominence in theoretical models, the role of plasticity costs in the evolution of plastic traits may be rather minor.