A hospital-based, multicenter, case-control study has been performed in Poland covering 741 incident stomach-cancer cases (520 males and 221 females) and the same number of controls. All stomach-cancer diagnoses were evaluated for histologic type according to the Lauren criteria. Fifty-one percent were of the intestinal type, 35 percent of the diffuse type, and 8.5 percent of the mixed type. The frequency of consumption of individual food items and several food groups was analyzed and the association of various foods with stomach cancer risk was evaluated after controlling for sex, age, occupation, education, and residency. Increased consumption of sausages was related significantly to gastric cancer risk, whereas increased consumption of cheese products, nonwhite bread, vegetables, and fruit was associated with decreased risk. A particularly strong decrease in risk was associated with consumption of radishes and onions. When consumption of fruits and vegetables, sausages, nonwhite bread, and cheese were introduced simultaneously in a multivariate model, independent effects were found only for fruit and vegetables, sausages, and nonwhite bread. The use of table salt, the frequency of eating hot meals, and an irregular eating pattern were also associated with increased risk, while additional consumption of fruit between meals showed reduced risk. If a reduction in vegetable and fruit consumption took place after marriage, an increased risk for stomach cancer was found, whereas augmented consumption of these food items after marriage decreased the risk. Separate risk models were calculated for stomach cancer of the intestinal and diffuse types, but both histologic varieties showed the same pattern of associations with dietary risk factors.