Previous occupational survey studies have identified 'waiter' and 'cook' as possible high risk occupations for cancer. However, few cohort studies have been performed among persons in the restaurant business, and we therefore have analyzed cancer incidence in two cohorts of Norwegian waiters and cooks. The cohorts consisted of skilled male workers, 1,463 waiters and 2,582 cooks, who received their craft certificate between 1958 and 1983. The cohorts were followed from 1959 through 1991. The standardized incidence ratio (SIR) for all causes of cancer was 1.4 (95 percent confidence interval [CI] = 1.2-1.7] for waiters, and 1.1 (CI = 0.9-1.4) for cooks. Cancers of the tongue, mouth, pharynx, larynx, esophagus, and liver were grouped together as alcohol-associated cancers. SIR for these cancers combined was 5.1 (CI = 3.4-7.4) for waiters and 4.2 (CI = 2.2-7.2) for cooks. For lung cancer, SIR was 2.0 (CI = 1.3-2.9) for waiters and 0.7 (CI = 0.2-1.7) for cooks. For alcohol-associated cancers, the analysis carried out according to number of years since first employment showed a larger number of cases than expected for both occupations in all time-periods. The excess of lung cancer cases among waiters appeared after 30 years or more of employment. The study shows that waiters and cooks are at high risk of cancers associated with alcohol consumption, and that waiters, in addition, show high rates for lung cancer. The hypothesis of an occupational lung-cancer risk in cooks was not supported by this study.