Following examination of the geographical pattern of stomach cancer incidence in Spain, the possible relationship between diet and stomach cancer has been assessed using an ecological model which looks into the relationship between the consumption of different foods and the nutrient intake in each of the 50 Spanish provinces during the period 1964-65, and the mortality rate for stomach cancer in these same 50 provinces during a period 20 years later (1984-86). The geographical pattern of stomach cancer mortality has been arrived at using the standardized mortality ratios (SMRs) calculated for each province. The per capita consumption of foods was obtained from the National Institute for Statistics and the calculation of nutrient intake was based on tables of nutritional composition. The results have been obtained using the statistical techniques of linear and Poisson regression. Stomach cancer mortality in Spain exhibits a stable and well-defined pattern, with less risk being apparent in the Mediterranean coastal provinces. The variability in the geographical distribution of stomach cancer mortality can, in large part (76%), be explained by the variation in the types of foods consumed between provinces, and this figure rises to 82% when one considers other socio-economic variables. A significant inverse association has been found between gastric cancer mortality and consumption of vegetables, fruits and vegetable oil as well as with vegetative growth and population density. From the multiple regression analysis, which considered both nutritional and socio-economic variables, the strongest inverse association with stomach cancer mortality was found in the consumption of green vegetables. The estimated relative risk (RR) of dying, for an increase in consumption of vegetables of 100 g per capita per day is 0.72, (95% CI: 0.53-0.91). These results are consistent with those of similar studies elsewhere.