Testing the predictability of morphological evolution in contrasting thermal environments

In light of climate change, the ability to predict evolutionary responses to temperature changes is of central importance for conservation efforts. Prior work has focused on exposing model organisms to different temperatures for just one or a few generations under laboratory conditions. Using a ‘natural experiment’, we show that studying parallel evolution in wild populations from contrasting thermal environments presents a more powerful approach for understanding and predicting responses to climate change. More specifically, we used a unique study system in Iceland, where freshwater populations of threespine sticklebacks ( Gasterosteus aculeatus ) are found in waters warmed by geothermal activity, adjacent to populations in ambient-temperature water. We used three sympatric and three allopatric warm-cold population pairs to test for repeated patterns of morphological divergence in relation to thermal habitat. We found that thermal habitat explained over 50% of body shape variation: fish from warm habitats had a deeper mid-body, a shorter jaw, and smaller eyes. Our common garden experiment showed that most of these morphological differences between thermal habitats were heritable. Lastly, absence of gene flow seems to facilitate parallel divergence across thermal habitats: all three allopatric population pairs were on a common evolutionary trajectory, whereas sympatric pairs followed different trajectories. Our findings therefore suggest that morphological responses to rising temperatures can be predictable when there is limited gene flow. On the other hand, migration of individuals between different thermal habitats or microhabitats can exaggerate nonparallel evolution and reduce our ability to predict evolutionary responses. ### Competing Interest Statement The authors have declared no competing interest.