During a population-based screening project for breast cancer, almost 15,000 women aged 50 years and over have provided a 12 h (overnight) sample of urine for research purposes. In 3,789 women the excretion of 11-desoxy-17-oxosteroids (DOOS) and creatinine was measured. Results were analysed in terms of urinary concentrations and of a ratio between DOOS and creatinine. Age had an effect on DOOS, creatinine and their ratio. Body weight and body surface area had an effect on creatinine excretion and therefore on the ratio. The following variables did not have an appreciable effect on the above-mentioned ratio: a family history of breast cancer, parity and age at first pregnancy, menopause and oestrogenic drugs, and parenchymal pattern of the breast as observed on the xeromammogram. Breast cancer was found at first screening in 106 out of 14,697 women. In 100 of these cases DOOS and creatinine were measured. Excretion values expressed as the ratio between the two, allowing for body surface area, did not differ materially from those of 100 age-matched controls. These results lead the authors to the conclusion that the determination of androgen metabolite excretion in women over 50 years of age is of no help in selecting a group at high risk of breast cancer.