A hospital-based case-control study of laryngeal cancer was conducted in Bremen in 1986 and 1987 with 100 prevalent male laryngeal cancer patients and 100 male hospital controls with diseases not considered to be related to smoking, alcohol, or occupational exposures, who were frequency matched by age. The odds ratio for heavy smoking (more than 30 pack-years) reached a value of 3.5 (95% confidence limits (CL) 1.1, 7.9). Ex-smokers showed a significant decrease in risk; this reached the level of those who had never smoked about 15 years after smoking cessation. For daily consumers of alcohol an odds ratio of 3.2 (95% CL 1.4, 7.5) was observed. Among the 17 occupations in which at least ten subjects had worked, excess risks were observed for stock keeping and transportation workers, and for leather and textile workers. The odds ratio was significantly increased for the latter (p less than 0.05). Among all those persons ever employed in a priori defined-risk occupations, an odds ratio of 2.74 (95% CL 1.23, 6.06) was observed. Considering responses to an exposure check-list, no increased risks could be shown for exposure to asbestos, coal tar, or welding fumes. On the other hand, excess risks were observed for exposures to diesel oil, gasoline, and mineral oil, controlling for smoking and alcohol. The findings in occupational and exposure subgroups were based on small numbers of cases and controls and, consequently, were subject to large sampling errors. Many of the results are consistent, however, with occupational risk factors reported from other studies.