Eighty-one women who had received hormonal therapy for advanced disease were assessed for objective response to treatment. Another 73 patients who had received chemotherapy were similarly studied. Patients with ER - tumours had a significantly poorer response rate to hormonal therapy than those with ER + tumours (P less than 0.05) and a significant trend toward an increasing rate of response occurred with increasing quantitative levels of receptor (P less than 0.01). This trend was even more significant when the assay had been performed on recurrent tumour (P less than 0.001), suggesting that this provides a more accurate index of response to hormonal treatment than assays performed on the primary lesion. These results suggest that whilst the quantitative receptor level used to discriminate between 'negative' or 'positive' assays is appropriate, this division is an over-simplification. Oestrogen receptor was not helpful in predicting response to chemotherapy.