Daily mortality displays a seasonal pattern linked to weather, air pollution, photoperiod length, influenza incidence and diet, among which temperature ranks as a leading cause. This study thus sought to assess the relationship between temperature, relative humidity, wind speed and mortality in the Madrid Autonomous Region (Spain) for the period January 1986-December 1992, controlling for the effects of air pollution and influenza incidence. Daily data on maximum, minimum and 24-hour mean temperature, relative humidity and wind speed were matched against daily mortality. Transfer function was identified using the Box-Jenkins pre-whitening method. Multivariate time series regression models were used to control for the confounding effects of air pollution and influenza incidence. Separate seasonal analyses were carried out for winter and summer periods. A J-shaped relationship between outdoor temperature, relative humidity and daily mortality was found. Mortality proved to be inversely related to cold temperature (4- to 11-day lag) and directly related to warm temperature (1-day lag). High relative humidity during summer periods was negatively related to mortality. Thermal variation ascribable to Madrid's mesothermal Mediterranean climate was strongly related to daily mortality, even where air pollution and influenza incidence were controlled for.