Logo: to the web site of Uppsala University

Elevated temperatures are expected to rise beyond what the physiology of many organisms can tolerate. Behavioural responses facilitating microhabitat shifts may mitigate some of this increased thermal selection on physiology, but behaviours are themselves mediated by physiology, and any behavioural response may trade-off against other fitness-related activities. We investigated whether experimental evolution in different thermal regimes (Cold: 15 degrees C; Hot: 31 degrees C; Intergenerational fluctuation 15/31 degrees C; Control: 23 degrees C) resulted in genetic differentiation of standard locomotor activity in the dung fly Sepsis punctum. We assessed individual locomotor performance, an integral part of most behavioral repertoires, across eight warm temperatures from 24 degrees C to 45 degrees C using an automated device. We found no evidence for generalist-specialist trade-offs (i.e. changes in the breadth of the performance curve) for this trait. Instead, at the warmest assay temperatures hot-selected flies showed somewhat higher maximal performance than all other, especially cold-selected flies, overall more so in males than females. Yet, the flies' temperature optimum was not higher than that of the cold-selected flies, as expected under the 'hotter-is-better' hypothesis. Maximal locomotor performance merely weakly increased with body size. These results suggest that thermal performance curves are unlikely to evolve as an entity according to theory, and that locomotor activity is a trait of limited use in revealing thermal adaptation.