In a population-based case-referent study of urothelial cancer in Stockholm during 1985-87, information was obtained from 78% of 418 identified cases and 77% of 511 selected referents. The relative risk (with 95% confidence interval) for intake of vitamin A supplements was 0.5 (0.2-1.0), with a dose-response relationship with increasing frequency of consumption. Increased risks of urothelial cancer were seen for several fried foods, for example fried meat [relative risk 1.4 (1.0-1.8) for weekly intake] and fried potatoes [relative risk 1.6 (1.1-2.6) for weekly intake]. Subjects with a high intake of fried foods, as defined by a collapsed variable, had a relative risk of 2.4 (1.4-4.2). A dose-response relationship was also seen with an increasing average daily intake of fat [relative risk 1.7 (1.0-2.8) in the highest quintile], but adjusting for fried foods decreased the relative risk, and it is uncertain whether the adjustment allowed for residual confounding. No association was noted for meat other than fried, but the analysis was based on small numbers.