Women with breast cancer whose blood and urinary androgens are at the lower end of the normal range have an early age at diagnosis, a rapid recurrence rate after mastectomy and tumours that are unresponsive to endocrine therapy. The most plausible explanation for these results is that the androgens are not involved in tumour initiation but control tumour growth rates after the malignant transformation has occurred. If this is so, endocrine case/control studies in which growth rates are not measured are an inefficient way of studying the endocrinology of breast cancer. It is suggested that the oestrogens and androgens should be considered as a single functional unit in future research.