Studies have demonstrated repeatedly that air pollution in Athens is associated with a small but statistically significant increase in mortality. Extremely high air temperatures can also cause excess mortality. This study investigated whether air pollution and air temperature have synergistic effects on excess mortality in Athens. Data concerning the increased number of deaths in July 1987 (when a major "heat wave" hit Greece) were compared to the deaths in July of the 6 previous years. This comparison revealed a greater increase in the number of deaths in Athens (97%), compared to all other urban areas (33%) and to all non-urban areas (27%). Data on the daily levels of smoke, sulfur dioxide, and ozone; the number of deaths that occurred daily; and meteorological variables were collected for a 5-y period. The daily value of Thom's discomfort index was calculated. Multiple linear regression models were used to investigate main and interactive effects of air temperature and Thom's discomfort index and air pollution indices. The daily number of deaths increased by more than 40 when the mean 24-h air temperature exceeded 30 degrees C. The main effects of an air pollution index are not statistically significant, but the interaction between high levels of air pollution and high temperature (> or = 30 degrees C) are statistically significant (p < .05) for sulfur dioxide and are suggestive (p < .20) for ozone and smoke. Similar results were obtained when the discomfort index was used, instead of temperature in the models.