Determining organismal responses to climate change is one of biology's greatest challenges. Recent forecasts for future climates emphasize altered temperature variation and precipitation, but most studies of animals have largely focussed on forecasting the outcome of changes in mean temperature. Theory suggests that extreme thermal variation and precipitation will influence species performance, and hence affect their response to changes in climate. Using an information-theoretic approach, we show that in squamate ectotherms (mostly lizards and snakes), two fitness-influencing components of performance, the critical thermal maximum and the thermal optimum, are more closely related to temperature variation and to precipitation, respectively, than either is to mean thermal conditions. By contrast, critical thermal minimum is related to mean annual temperature. Our results suggest that temperature variation and precipitation regimes have had a strong influence on the evolution of ectotherm performance, so that forecasts for animal responses to climate change will have to incorporate these factors, and not only changes in average temperature.