Abstract
Forecasting which species/ecosystems are most vulnerable to climate warming is essential to guide conservation strategies to minimize extinction. Tropical/mid-latitude species are predicted to be most at risk as they live close to their upper critical thermal limits (CTLs). However, these assessments assume that upper CTL estimates, such as CTmax, are accurate predictors of vulnerability and ignore the potential for evolution to ameliorate temperature increases. Here, we use experimental evolution to assess extinction risk and adaptation in tropical and widespread Drosophila species. We find tropical species succumb to extinction before widespread species. Male fertility thermal limits, which are much lower than CTmax, are better predictors of species' current distributions and extinction in the laboratory. We find little evidence of adaptive responses to warming in any species. These results suggest that species are living closer to their upper thermal limits than currently presumed and evolution/plasticity are unlikely to rescue populations from extinction.