To study the relationship between mortality and temperature in Cantabria, a Spanish region that includes both rural and urban areas. Meteorological data (2003–2006) were obtained from the Spanish Meteorological Agency and daily numbers of deaths were obtained from the Spanish Institute for Statistics. A graphical approach using locally weighted regression smoothing was used to explore the relationship between mortality and temperatures and to identify temperature thresholds; we estimated the excess of mortality due to extreme temperatures in both warm and cold periods using Poisson regression models, and we simulated a situation with increased temperatures. Raising maximum or minimum temperatures by 1ºC was associated with a 2% excess in mortality risk in the whole population throughout the warm period, and we found no effect in mortality on the cold season; almost all changes in mortality occur in people aged 65 or more. Women are more sensitive to temperature changes in the warmer months.The deleterious effect of increasing temperatures in summer is more pronounced than the beneficial effect of a similar increase in winter.