A case-control study of stomach cancer in relation to dietary, smoking, and drinking habits was undertaken in Saitama Prefecture, Japan. The study was based on 294 cases of newly diagnosed adenocarcinoma of the stomach at a single institution, 294 general population controls (matched by sex, age, and administrative division), and 202 hospital controls. Dietary habits were investigated based on the intake of 12 separate foods and 12 food groups in a food frequency questionnaire, together with individual food preferences. The consumption of raw vegetables was inversely related to the risk of stomach cancer, with a dose-response relation observed consistently in the comparisons with both sets of controls. Current cigarette smokers (1-29/day) had an increased risk (relative risk = 1.8, 95 percent confidence interval = 1.1-3.0) compared with nonsmokers in the general population controls, but no dose-response effect with heavier cigarette smoking. Alcohol use did not affect the risk of stomach cancer. In the multiple logistic regression, the consumption of raw vegetables showed a protective effect on stomach cancer while cigarette smoking had no significant association, in both sets of controls.