To test the hypothesis that high levels of endogenous oestrogens increase the risk for developing breast cancer, concentrations of oestrone, oestradiol and oestriol were measured in 24 h urine samples from 1000 women participants in a prospective study of breast cancer on the island of Guernsey. Sixty-nine subjects were diagnosed with breast cancer subsequent to urine collection. Among women who were premenopausal at the time of urine collection, cases excreted less oestrogen than controls; the odds ratios (95% CI) for breast cancer in the middle and upper thirds of the distribution of oestrogen excretion, in comparison with the lower third (reference group, assigned odds ratio = 1.0), were 0.5(0.2-1.2) and 0.4(0.2-1.1) respectively for oestrone, 0.8(0.4-1.8 and 0.4(0.2-1.1) for oestradiol, 0.7(0.3-1.6) and 0.7(0.3-1.6) for oestriol and 0.9(0.4-2.0) and 0.5(0.2-1.3) for total oestrogens. Among women who were post-menopausal at the time of urine collection, the trend was in the opposite direction, with an increase in risk associated with increased oestrogen excretion; the odds ratios were 0.9(0.3-2.2) and 1.1(0.5-2.8) for oestrone, 0.8(0.3-2.3) and 1.9(0.8-4.6) for oestradiol, 1.5(0.6-3.9) and 1.8(0.7-4.6) for oestriol and 0.9(0.4-2.6) and 1.9(0.7-4.7) for total oestrogens. The trends of increasing risk with increasing oestrogen excretion among post-menopausal women were statistically significant for oestradiol (P = 0.022) and for total oestrogens (P = 0.016). We conclude that high levels of endogenous oestrogens in post-menopausal women are associated with increased breast cancer risk, but that the relationship of oestrogens in premenopausal women with risk is unclear.