A case-control study of breast cancer was carried out in La Plata, Argentina, where the incidence of the disease is comparable to the highest rates recorded worldwide. One hundred and fifty incident cases were identified through major hospitals. For each case, a hospital control, matched by age and hospital, and a neighbourhood control, matched by residential area and age, were also chosen. Cases and controls were interviewed to obtain information on past diet, as well as demographic and socio-economic characteristics, reproductive and menstrual history and other potential breast-cancer risk factors. The dietary information was obtained from questions on the consumption of specific food items and information on portion sizes from an earlier study was used to estimate intake of calories and selected nutrients. There was a substantial excess energy intake among cases as compared to both control groups, which was present across all 3 major macronutrients which contribute to total calories. Among the food groups, the consumption of eggs was a risk factor for breast cancer, and whole-milk products and green leafy vegetables were protective. After adjusting for the calorie difference in multivariate statistical analyses of nutrients, fibre and beta-carotene consumption were weakly protective. The results are discussed with reference to possible methodological difficulties and previous studies of diet and breast cancer.