A case-control study was conducted in high- and low-risk areas of Italy to evaluate reasons for the striking geographic variation in gastric cancer (GC) mortality within the country. Personal interviews with 1,016 histologically confirmed GC cases and 1,159 population controls of similar age and sex revealed that the patients were more often of lower social class and resident in rural areas and more frequently reported a familial history of gastric (but not other) cancer. After adjusting for these effects, case-control differences were found for several dietary variables, assessed by asking about the usual frequency of consumption of 146 food items and beverages. A significant trend of increasing GC risk was found with increasing consumption of traditional soups, meat, salted/dried fish and a combination of cold cuts and seasoned cheeses. The habit of adding salt and the preference for salty foods were associated with elevated GC risk, while more frequently storing foods in the refrigerator, the availability of a freezer and use of frozen foods lowered risk. Reduced GC risk were associated with increasing intake of raw vegetables, fresh fruit and citrus fruits. Lowered risk was also related to consumption of spices, olive oil and garlic. Neither cigarette smoking nor alcoholic beverage drinking were significantly related to GC risk. The case-control differences tended to be consistent across geographic areas, despite marked regional variations in intake levels of certain foods. The high-risk areas tended to show higher consumption of food associated with elevated risk (traditional soups, cold cuts) and lower consumption of foods associated with reduced risks (raw vegetables, citrus fruits, garlic). Our findings indicate that dietary factors contribute to the regional variation of stomach cancer occurrence in Italy, and offer clues for further etiologic and prevention research.