We report results from a case-control study on stomach cancer conducted in 1980-81 in Cracow, Poland, where the risk of this disease is among the highest reported world-wide. One hundred and ten cases from a surgical clinic were matched by age and sex to the same number of controls from the same hospital. A matched series from a population-based health survey was also considered. After adjusting for residency, smoking and various food items, cases and hospital controls showed significant differences in consumption of fruits (RR rarely vs. daily: 3.24; 95% CI: 1.56-6.77), joint consumption of vegetables, salads and fruits (RR low vs. high: 4.23; 95% CI: 1.41-12.63), and consumption of protein-containing foods (RR low vs. high: 0.23; 95% CI: 0.08-0.61). Consumption of strong alcoholic beverages on an empty stomach (before breakfast) was associated with an RR of 2.09 (1.04-4.22). The findings are discussed in relation to the apparent urban/rural difference in stomach cancer incidence in Poland and the possible underlying etiological factors involved.