The relationship between intake of dairy products and risk of breast cancer was studied in 4697 initially cancer-free women, aged 15 years or over. During a 25 year follow-up period after the collection of food consumption data, 88 breast cancers were diagnosed. Intakes of foods were calculated from dietary history interviews covering the habitual diet of examinees over the preceding year. There was a significant inverse gradient between milk intake and incidence of breast cancer, the age-adjusted relative risk of breast cancer being 0.42 (95% confidence interval=0.24-0.74) between the highest and lowest tertiles of milk consumption. The associations with respect to other dairy products were not significant. Adjustment for potential confounding factors, i.e. smoking, body mass index, number of childbirths, occupation and geographic area, resulted in only a minor change in the milk intake-breast cancer relation. Nor did adjustment for intake of other foodstuffs and nutrients, e.g. energy, carbohydrates, protein, fat, vitamins and trace elements, alter the results. No significant interactions between milk intake and demographic or dietary variables or time of cancer diagnosis were observed. Our data suggest that there is a protective effect, dietary or habitual, associated with consumption of milk that overwhelms the associations between different other factors and risk of breast cancer.