Research indicates that air pollution is a risk factor of an increased occurrence of diseases. However, evidence is limited on the effects of the pollution index on disease and whether temperature modifies the effects. The objectives were (i) to explore the effects of the Air Pollution Index (API) and specific indices for pollutants (PM10, NO2, and SO2) on respiratory emergency department (ED) visits in Beijing and (ii) to investigate whether temperature modified the effects of main air pollutants on respiratory ED visits. A quasi-Poisson generalized additive model was employed to examine the association of API and indices for pollutants with respiratory disease. Bivariate response surface model and stratification model (cold days, moderately cold days, moderately hot days, and hot days) were used to analyze the modification effects of temperature on air pollution and respiratory disease. The results showed that (i) the effects of API on respiratory diseases were similar to the index for PM(10)in Beijing. (ii) API and PM(10)were associated with increased respiratory ED visits on cold days and moderately cold days. Furthermore, the effects of PM(10)on respiratory disease on moderately cold days [Relative risk (RR) = 1.006 per 10 mu g/m(3), 95% CI 1.002-1.009] were stronger than on cold days (RR = 1.004 per 10 mu g/m(3), 95% CI 1.000-1.008). (iii) PM10(API) had a greater impact on children aged 10 to 17 years and females on moderately cold days, while the elderly had an increased risk of respiratory disease to PM10(RR = 1.008 per 10 mu g/m(3), 95% CI 1.002-1.013) and API (RR = 1.013 per 10, 95% CI 1.004-1.022) on cold days. In conclusion, temperature can modify the association between API and respiratory morbidity. A stronger correlation existed between PM(10)and respiratory diseases on moderately cold days, while the effects of cold days were less than that attributable to moderately cold days.