Experimental evolution of local adaptation under unidimensional and multidimensional selection

Abstract

Local adaptation is a fundamental evolutionary process generating biological diversity and potentially enabling ecological speciation. Divergent selection underlies the evolution of local adaptation in spatially structured populations by driving their adaptation toward local optima. Environments rarely differ along just one environmental axis; therefore, divergent selection may often be multidimensional. How the dimensionality of divergent selection affects local adaptation is unclear: evolutionary theory predicts that increasing dimensionality will increase local adaptation when associated with stronger overall selection but may have less predictable effects if selection strengths are equal. Experiments are required that allow the effect of the dimensionality of selection on local adaptation to be tested independently of the total strength of selection. We experimentally evolved 32 pairs of monogonont rotifer populations under either unidimensional divergent selection (a single pair of stressors) or multidimensional divergent selection (three pairs of stressors), keeping the total strength of selection equal between treatments. At regular intervals, we assayed fitness in home and away environments to assess local adaptation. We observed an initial increase and subsequent decline of local adaptation in populations exposed to multidimensional selection, compared with a slower but eventually stronger increase in local adaptation in populations exposed to unidimensional selection. Our results contrast with existing predictions, such as the "weak multifarious" and "stronger selection" hypotheses. Instead, we hypothesize that adaptation to multidimensional divergent selection may favor generalist genotypes and only produce transient local adaptation.