The strength of the hypothesis that fibre reduces the risk of breast cancer is its biological plausibility, which is supported by experimental and interventional findings and by the coherence of observational studies. However, at least half the available epidemiological studies have failed to show a decreased risk in breast cancer for an increased fibre intake. But intervention studies taking the plasma concentration of oestrogens as an end-point showed significantly lower levels of breast cancer in women with a high-fibre and low-fat diet than in women with usual Western diets. Any reduction in breast cancer risk appears to be significantly dependent on the level of fibre intake. Several explanations can be proposed, including measurement errors in food intake, the insufficiency of food-composition tables, the difficulty of allowing for the diversity of fibre intake and the complexity of the natural history of breast cancer. More research is needed, not only better nutritional surveys for the different types of fibre intake and improvements in food-composition tables, but also epidemiological studies with the power to control for all the eventual confounding risk factors. Although the scientific evidence is not complete, recommendations for a fibre-rich diet should be made, both for cereals and for fruit and vegetables, in part because such diets at least do no harm, but also because fibre is known to be protective against other pathological conditions.