As part of a case-control study in northern Alberta, Canada, 577 women aged 30-80 with breast cancer diagnosed during 1976-77 and a population-based age-stratified random sample of 826 disease-free female controls were questioned about certain aspects of their diet. Computing relative risks (RRs) by tertiles, significant increasing trends were found with more frequent consumption of beef (RRs of 1.0, 2.3, 1.5; test for trends, p less than 0.001), pork (RRs of 1.0, 1.6, 2.2; test for trend, p less than 0.001), and sweet desserts (RRs of 1.0, 1.3, 1.5; test for trend, p = 0.01). Elevated risks were also noted for use of butter at the table and for frying with butter or margarine, as opposed to vegetable oils. The association of total beef and pork consumption with breast cancer was not materially affected by controlling for age at first birth, family history of breast cancer, previous benign breast biopsy or socioeconomic status. Nor was the association reduced by controlling for ages of menarche and menopause, even though within the control series the intake of beef and pork reported in adult life was higher among those with a lower age at menarche or a older age at natural menopause.