In thirty-one premenopausal patients with carcinoma of the breast the plasma prolactin was measured after mastectomy. A highly significant correlation between tumour size and plasma prolactin levels (p less than 0.002) was observed after adjustment for age at diagnosis and parity. At the time of the prolactin determination no clinical signs of metastatic disease were evident, suggesting that the prolactin levels were unrelated to the tumour burden.