In a case-control study of 71 consecutive new male cases of lung cancer and 71 male hospital control patients, previous dietary and alcohol intake, smoking pattern, occupation, dust exposure, and family history of lung cancer were investigated. The cases and controls were similar in age, country of origin, area of residence, and marital status. Using a frequency-based assessment of previous dietary intake, broad food groups were similar for cases and controls. Cases had a significantly lower intake of fish than controls did (odds ratio = 0.5, confidence interval = 0.2-1.0, p = 0.05). A protective effect for fish consumption in lung cancer has not been previously reported. The dietary intake of foods containing retinol and beta-carotene and the intake of alcohol were not significantly different between cases and controls. For cases, smoking duration was longer and the time since cessation for exsmokers was shorter, cigarette pack years were longer, and the number of cigarettes smoked per day was greater. The factors of occupation, dust exposure, and family history of cancer (including lung cancer) were similarly distributed between cases and controls.