The association between consumption of coffee and bladder cancer and the effect modification of cigarette smoking was investigated as part of a comprehensive case-control study. The study population consisted of 351 case patients with histologically confirmed transitional cell carcinomas of the bladder among white males and 855 white male control subjects selected from Erie, Niagara, and Monroe counties in New York from 1979 to 1985. Usual diet, coffee consumption, and cigarette use were estimated by comprehensive interviews using a detailed food frequency questionnaire. After adjustment for age, education, and dietary risk factors by multiple logistic regression, risk was found to increase with increasing pack-years of cigarette use with an odds ratio in the highest quartile of 2.7 (95% confidence interval, 1.8 to 4.0) when compared to the lowest quartile. Coffee consumption was associated with an increased risk for bladder cancer among the heaviest coffee drinkers after adjustment for cigarette smoking and other dietary risk factors (odds ratio, 2.1; 95% confidence interval, 1.3 to 3.2). The effect was more pronounced among nonsmokers, especially among those 65 years and older. These findings support the contention that coffee is a weak carcinogen.