The relationship between air pollution and daily mortality in 1989 was examined in two residential areas in Beijing, China. Very high concentrations of sulfur dioxide (SO2) (mean = 102 micrograms/m3, maximum = 630 micrograms/m3) and total suspended particulates (TSPs) (mean = 375 micrograms/m3, maximum = 1,003 micrograms/m3) were observed in these areas. Daily counts of deaths were regressed, using Poisson regression on the logarithm of (SO2) and/or TSPs controlling for effects of temperature, humidity, and day of week. A highly significant association was found between In(SO2) and daily mortality. The risk of total mortality was estimated to increase by 11% (95% confidence interval [95% CI] = 5%-16%) with each doubling in SO2 concentration. Separately, the association of In(TSP) with total daily mortality was positive but not significant (4% increase in mortality with each doubling in TSP; 95% CI = -2%-11%). When mortality was analyzed separately by cause, the association with a doubling in SO2 was significant for chronic obstructive pulmonary disease (29%), pulmonary heart disease (19%), and cardiovascular disease (11%), and marginally significant for the other nonmalignant causes (8%), but not statistically significant for cancer (2%). A similar association was noted for a doubling in TSP (4%, 38%, and 8% for total, chronic obstructive pulmonary disease, and pulmonary heart disease mortality, respectively), but the result was only statistically significant for chronic obstructive pulmonary disease. In the season-specific analysis, both SO2 and TSP were found to be significant predictors of total daily mortality in summer. In winter, SO2 was again significantly associated with increased mortality, but no positive association was found between TSP and mortality.(ABSTRACT TRUNCATED AT 250 WORDS)