We present the results of a prospective randomised trial of Estradurin, a long-acting oestrogen preparation (polyoestradiol phosphate), 160 mg per month, compared with bilateral orchiectomy in patients with advanced prostatic carcinoma (T3M1; T4MO/M1). The dose was lower than that usually recommended to induce a consistent fall in serial plasma testosterone levels to within the castrate range. Most patients treated with oestrogen showed an initial clinical and biochemical response equal to that obtained for patients undergoing orchiectomy. The inevitable relapse in hormone sensitivity sometimes occurred very soon after the start of oestrogen treatment. Many patients had poorly suppressed plasma testosterone expressed as a mean of monthly serial measurements, but then responded to secondary orchiectomy. These data only suggest that, in the treatment of hormone-sensitive prostatic carcinoma, it may be necessary to reduce plasma testosterone to midway between castrate and normal ranges. The data support the theory that response to androgen withdrawal is qualitative rather than quantative. The effective dose of oestrogen may therefore be reduced and the risk of thrombo-embolic complications minimised.