The number of daily deaths, temperature, relative humidity, and 24-hr concentrations of main air pollutants observed during a heat wave (June 27-August 7, 1994) in Belgium were compared with those recorded before and after this heat wave. All these variables were averaged over the country. Expected mortality was calculated from daily deaths observed during the summers of 1985-1993. The influence of meteorological and air pollution variables on daily mortality was analyzed using generalized least-squares method. Mortality recorded during the heat wave was higher than expected: it increased by 9.4% in the age group 0-64 years (236 excess deaths; P < 0.001) and by 13.2% in the elderly (1168 excess deaths; P < 0.001). After the heat wave, mortality in the elderly was lower than expected (178 deficit deaths; P< 0.05); the net excess of mortality in the whole population amounted to 1226 deaths when accounting for this deficit. This increased mortality was associated with unusually high outdoor temperatures (range of daily mean: 15.3-27.5 degrees C) and elevated ozone levels (range of 24-hr concentration: 34.5-111.5 microg/m3). The duration of the ozone overexposure during the heat wave was also uncommon: half-hour concentrations of ozone exceeded, on an average, 100 microg/m3 for 8 consecutive hr. The number of daily deaths was mostly correlated with the mean daily temperature and 24-hr ozone concentration, both measured the day before. A synergy between temperature and ozone in their effects on mortality was also highlighted in both age groups. The product of the logarithm of temperature by the logarithm of ozone concentration, both measured the day before, contributed to 39.5% of the variance of the logarithm of daily deaths in elderly and to 4.5% in the age group 0-64 years. In conclusion, elevated outdoor temperatures combined with high ozone concentrations were assumed to be the likely cause of the important excess mortality observed in Belgium during the summer, 1994.