Accounting for local physiological adaptation in bioenergetic models: testing hypotheses for growth rate evolution by virtual transplant experiments

Abstract

We constructed bioenergetic models for locally adapted populations of Atlantic silversides, Menidia menidia, from different latitudes (Nova Scotia and South Carolina) to determine how genetic variation in growth physiology affects model parameters and predicted growth and to test two hypotheses on the evolution of countergradient variation in growth rate. Model parameters were estimated simultaneously for each population through a penalized likelihood approach incorporating laboratory measurements of metabolism, specific dynamic action, consumption, and growth. The resulting population-specific parameters differed by an average of 28%. The models were validated by successful (R2 > 0.9) prediction of growth in independent experiments under natural light and temperature conditions and by predicting growth in the field (R2 > 0.95). We then performed virtual reciprocal transplant simulations to test the alternative hypotheses that growth rate along a latitudinal gradient evolves in response to temperature or resource availability. Predictions for each transplanted population deviated significantly from observed growth for each native population, demonstrating the importance of accounting for interpopulation variation in model parameters. Our results indicate that the latitudinal cline in growth rate cannot be explained solely by thermal adaptation but may have arisen owing to the combined effects of temperature and food availability.