To re-evaluate the treatment of advanced prostatic carcinoma with diethylstilboestrol (DES) in low dosage in relation to the degree of suppression of plasma testosterone. The study comprised 106 patients with advanced carcinoma of the prostate (89 with T3/4 M1 and 17 with T3/4 N0/1 M0) who were treated with 1 mg/day of DES. The response was assessed clinically and by the change in plasma prostate specific antigen, prostatic and phosphatase and alkaline phosphatase, and plasma testosterone was monitored regularly. In a few patients it was possible to reduce the dose to 0.5 mg/day DES; in others, the initial response was not sustained and they were treated with an increased dose of DES or bilateral orchidectomy. Seventy patients (Group 1) showed a sustained response to 1 mg/day of DES (0.5 mg/day in three) and 50 remained in remission at a mean of 21 months of treatment. Of the 36 patients offered secondary treatment, 12 (Group 2) responded with a second remission. Only 27% of patients had mean testosterone levels in the castrate range (0-2 nmol/L) but most in Group 1 had mean levels of < 10 nmol/L, whilst in 10 of the 12 patients in Group 2, the level was > 10 nmol/L. Overall times to progression and death were comparable with the results of conventional monotherapy or combination treatment and complication rates were acceptable. Low-dose oestrogen therapy (1 mg/day of DES) is cheap, effective and caused few side-effects, none of which was life-threatening. In many patients, only minimal suppression of the plasma testosterone was required and the response appeared to be qualitative, although there was considerable variation in the threshold of response. A randomized trial of oestrogen, in a minimal dose adjusted to the requirements of the individual patient, against conventional hormone treatment now seems justified.