Clomiphene was administered to 16 patients with elevated serum prolactin levels in doses of 100, 200 and 300 mg/day for five days in succeeding months and total urinary oestrogens estimated on days 0, 5, 8, 12 and 15 following commencement of treatment. The responses were compared with six patients who were amenorrhoeic but had normal serum prolactin levels and absent positive feedback to oestrogen. The increased outputs of oestrogens were similar in the two groups. In the hyperprolactinaemic group 5 out of 16 subjects showed evidence of ovulation whilst the remainder showed a secondary failure of response. Six subjects who failed to ovulate were treated with clomiphene and human chorionic gonadotrophin (HCG) and ovulations were induced in 31 out of 34 treatment cycles but no pregnancies were achieved. The responses to clomiphene therapy in the hyperprolactinaemic subjects were compared to the assessment of positive feedback mechanisms by means of oestrogen provocation and oestrogen amplification tests and good correlation was obtained. Only those with evidence of positive feedback to these tests were likely to ovulate on clomiphene.