How and When Selection Experiments Might Actually be Useful.

Laboratory natural selection and artificial selection are vital tools for addressing specific questions about evolutionary patterns of variation. Laboratory natural selection can illuminate whether a putative selective agent is capable of generating long-term, sustained changes in individual traits and suites of traits. Artificial selection is the essential tool for understanding the general evolvability of traits and the extent to which genetic correlations constrain evolution. We review the contexts in which each type of experiment seems capable of offering key insights into important evolutionary issues. We also discuss theoretical and methodological considerations that play critical roles in designing selection experiments that are relevant to evolutionary patterns of trait variation. In particular, we focus on the critical role of selection intensity and the consequences of experiments with different intensities. While selection experiments are not practical in many cases, sophisticated selection experiments-designed with careful consideration of the theory of selection-should be taken beyond model organisms and used in well-chosen natural systems to understand natural patterns of variation.