Women who bear their first child at an early age have lower breast-cancer incidence-rates than do women who are older at first birth or who remain nulliparous. The urine "oestriol ratio", the concentration of oestriol relative to the sum of the concentrations of oestrone and oestradiol, is inversely related to a population's breast-cancer rate. To evaluate the relationship between these two breast-cancer risk indicators the urine oestriol ratio was determined for recently delivered uniparous women aged 19-23, 25-27, and 29-34 years and nulliparous women of comparable ages. In the follicular phase, the youngest parous women had an oestriol ratio 40% higher than, and significantly different from, the ratios of all other groups which were otherwise quite similar. In the luteal phase, the oestriol ratio of the youngest parous women was again distinctively raised and generally the oestriol ratios of parous women were higher than those of nulliparae. The results are not explained by differing frequencies of ovulation among the groups or by confounding from several breast-cancer risk indicators. These findings support the hypothesis that oestrogen metabolism, as reflected by the urine oestriol ratio, is a determinant of breast-cancer risk.