Contemporary temperature-driven divergence in a Nordic freshwater fish under conditions commonly thought to hinder adaptation.

Background: Evaluating the limits of adaptation to temperature is important given the IPCC-predicted rise in global temperatures. The rate and scope of evolutionary adaptation can be limited by low genetic diversity, gene flow, and costs associated with adaptive change. Freshwater organisms are physically confined to lakes and rivers, and must therefore deal directly with climate variation and change. In this study, we take advantage of a system characterised by low genetic variation, small population size, gene flow and between-trait trade-offs to study how such conditions affect the ability of a freshwater fish to adapt to climate change. We test for genetically-based differences in developmental traits indicating local adaptation, by conducting a common-garden experiment using embryos and larvae from replicate pairs of sympatric grayling demes that spawn and develop in natural cold and warm water, respectively. These demes have common ancestors from a colonization event 22 generations ago. Consequently, we explore if diversification may occur under severely constraining conditions.

Results: We found evidence for divergence in ontogenetic rates. The divergence pattern followed adaptation predictions as cold-deme individuals displayed higher growth rates and yolk conversion efficiency than warm-deme individuals at the same temperature. The cold-deme embryos had a higher rate of muscle mass development. Most of the growth- and development differences occurred prior to hatch. The divergence was probably not caused by genetic drift as there was a strong degree of parallelism in the divergence pattern and because phenotypic differentiation (Q(ST)) was larger than estimated genetic drift levels (microsatellite F(ST)) between demes from different temperature groups. We also document that these particular grayling populations cannot develop successfully at temperatures above 12°C, whereas other European populations can, and that increasing the muscle mass development rate comes at the cost of some skeletal trait development rates.

Conclusions: This study shows that genetically based phenotypic divergence can prevail even under conditions of low genetic variation and ongoing gene flow. Furthermore, population-specific maximum development temperatures along with musculoskeletal developmental trade-offs may constrain adaptation.