Inconsistent results have been obtained from studies that have examined the relationship between air pollution and hospital visits for stroke. We undertook a time-stratified case-crossover study to evaluate associations between outdoor air pollution and emergency department visits for stroke among the elderly according to stroke type, season, and sex. Analyses are based on a total of 12,422 stroke visits among those 65 years of age and older in Edmonton, Canada between April 1, 1992 and March 31, 2002. Daily air pollution levels for SO(2), NO(2), PM(2.5), PM(10), CO and O(3) were estimated using data from fixed-site monitoring stations. Particulate matter data were only available from 1998 onwards. Conditional logistic regression was used to estimate the odds ratios (ORs) and their 95% confidence intervals in relation to an increase in the interquartile range (IQR) of each pollutant. ORs were adjusted for the effects of temperature and relative humidity. We found no association between outdoor measures of air pollution and all stroke visits. In contrast, elevated risks were observed between levels of air pollution and acute ischemic stroke between April and September. During this season, the ORs associated with an increase in the IQR of the 3-day average for CO and NO(2) were 1.32 (95% CI = 1.09-1.60) and 1.26 (95% CI = 1.09-1.46), respectively. CO exposures in the same season, lagged 1 day, were associated with an increased risk of hemorrhagic stroke with ORs was 1.20 (95% CI = 1.00-1.43). Our results suggest it is possible that vehicular traffic, which produces increased levels of NO(2) and CO, contributes to an increased incidence of emergency department visits for stroke.