Prospective studies of risk factors for breast cancer have been carried out since 1961 on a normal population which now comprises some 15,000 women. Classical (anamnestic) factors such as ages at menarche, menopause, or at first child, post-menopausal weight and high parity are not strong determinants of risk. Measurements of various aspects of endocrine function are more successful in identifying women at risk, as is assessment of mammographic parenchymal patterns but, again, these variables are not effective enough to make an intervention strategy imperative. However, the combination of classical, endocrinological, and radiological factors appears to achieve this aim. A substantial sub-set of women who have a four-fold risk of breast cancer may be identifiable.