Evolvability of marine dispersal traits related to larval swimming and reproductive strategy.

Abstract

While theory has long explained how selection can favor dispersal, we know much less about the capacity for dispersal to evolve in natural populations. One challenge is to identify how heritable genetic variation in multiple dispersal traits aligns with directions of selection on these traits. Therefore, we used an experimental pedigree in the solitary ascidian Molgula occidentalis and constructed Bayesian quantitative genetic models to estimate evolvability of dispersal traits. We compared the evolvability of dispersal traits by (1) assessing the average evolvability of the G-matrix, (2) considering specific directions of selection hypothesized from biomechanical and life history reasoning, and (3) estimating the predicted response to selection from the observed additive genetic covariance between dispersal traits and fitness components. There was greater additive genetic variation in tail length than in trunk length, as there was for maternal effect and dominance variance. As a result, selection for longer tails (faster swimming) was associated with a higher evolvability compared to selection for shorter tails and longer trunks (energetic efficiency) or shorter tails and trunks (fecundity). The predicted response to selection at hatching and settlement, based on observed additive genetic covariances, was similar for trunk and tail length. Overall, evolutionary change in dispersal distances would occur more rapidly under selection on swimming ability than on energetic economy and fecundity. Furthermore, multiple lineages within Molgulidae have repeatedly diverged along the larval morphology axis in which we measured the highest evolvability. Our results demonstrate the utility of combining estimates of evolvability with hypothesized and observed selection gradients.