Quantifying multivariate genotype-by-environment interactions, evolutionary potential and its context-dependence in natural populations of the water flea, Daphnia magna.

Genotype-by-environment interactions (G x E) underpin the evolution of plastic responses in natural populations. Theory assumes that G x E interactions exist but empirical evidence from natural populations is equivocal and difficult to interpret because G x E interactions are normally univariate plastic responses to a single environmental gradient. We compared multivariate plastic responses of 43 Daphnia magna clones from the same population in a factorial experiment that crossed temperature and food environments. Multivariate plastic responses explained more than 30% of the total phenotypic variation in each environment. G x E interactions were detected in most environment combinations irrespective of the methodology used. However, the nature of G x E interactions was context-dependent and led to environment-specific differences in additive genetic variation (G-matrices). Clones that deviated from the population average plastic response were not the same in each environmental context and there was no difference in whether clones varied in the nature (phenotypic integration) or magnitude of their plastic response in different environments. Plastic responses to food were aligned with additive genetic variation (gmax) at both temperatures, whereas plastic responses to temperature were not aligned with additive genetic variation (gmax) in either food environment. These results suggest that fundamental differences may exist in the potential for our population to evolve novel responses to food versus temperature changes, and challenges past interpretations of thermal adaptation based on univariate studies.