The relation between diet and breast cancer was examined in a population-based case-control study conducted in Adelaide, South Australia, involving 451 case-control pairs aged 20-74 years. Cases were identified through the state cancer registry between April 1982 and July 1984; for each case, one age-matched control was selected from the electoral register. Dietary intake was measured by self-administered quantitative food frequency questionnaires. There was little variation in risk across levels of daily intake of energy, protein, and total fat; for energy, the relative risk of breast cancer at the uppermost fifth of intake, relative to a risk of unity for the lowest fifth, was 1.22 (95% confidence interval (CI) 0.80-1.86); for protein, the corresponding relative risk was 1.09 (95% CI 0.72-1.64), and for total fat, the relative risk was 0.90 (95% CI 0.59-1.38). Variation in risk in association with sugar and starch intake was also insubstantial, while for fiber, there was a nonuniform reduction in risk at the three uppermost fifths of intake. Risk varied little with level of retinol intake, but it decreased with increasing intake of beta-carotene, a trend that was statistically significant; the relative risk of breast cancer at the uppermost fifth of beta-carotene intake was 0.76 (95% CI 0.50-1.18). Multivariate adjustment for the effects of potentially confounding variables did not alter these patterns. The study does not support a role for dietary fat in the etiology of breast cancer.