The relationship between coffee drinking and cancer of the lower urinary tract (LUT) was investigated by a case-control study of white women identified from hospitals in urban areas of Massachusetts and Rhode Island. Data on coffee drinking, tea drinking, use of coffee additives, and cigarette smoking were obtained by mail questionnaire. Information was available for 135 women with LUT cancer and 390 controls. For women who usually drank 1+ cups of coffee per day, the risk ratio of having LUT cancer was 2.1 (95 percent confidence limits, 1.1-4.3), compared to a risk of 1 for women who drank less or not at all. However, no dose-response relationship was demonstrated between LUT cancer and usual daily coffee consumption or "cup-years" of coffee drinking. The association of coffee with disease was no different, whether decaffeinated, nondecaffeinated, regular, or instant coffee had been drunk, or whether the coffee was brewed strong or weak. Use of nondairy creamers, saccharin, or cyclamates was not associated with increased risk of disease. The risk of LUT cancer for cigarette smokers relative to nonsmokers was 1.6 (1.0-2.4). The attendant dose-response relationship was statistically significant. The absence of a dose-response relationship between coffee drinking and LUT cancer suggested that the association observed was noncasual.