Salamander morph frequencies do not evolve as predicted in response to 40 years of climate change

The global climate is changing rapidly, yet biotic responses remain uncertain. Most studies focus on changes in species ranges or plastic responses like phenology, but adaptive evolution could be equally important. Studying evolutionary responses is challenging given limited historical data and a poor understanding of genetically variable traits under selection. We take advantage of a historical dataset to test for an adaptive response to climate change in a widespread, polymorphic amphibian, the eastern red‐backed salamander Plethodon cinereus. We resurveyed color morph frequencies across New England to test for an adaptive shift in response to climate change. We modeled historical and present‐day morph proportions as a function of climate and tested the accuracy of predictions both within and across different time periods. Our models showed moderate accuracy when predicting morph frequencies within time periods, but poor accuracy across time periods. Despite substantial changes in climate and significant relationships between morph frequency and climate variables within periods, we found no evidence for the predicted shift in morph frequencies across New England. The relationship between climate and color morph frequencies is likely more complex than originally suggested, potentially involving the interplay of additional factors such as microclimate variation, land use changes, and frequency‐dependent selection. Model extrapolation and changes in the correlation structure of climate variables also likely contributed to poor predictive ability. Evolution could provide a means to moderate the effects of climate change on many species. However, we often do not understand the direct links between climate variation, traits, and fitness. Therefore, forecasting climate‐mediated evolution remains an ongoing and important challenge for understanding climate change threats to species.