Between 1970 and 1983, 442 patients were treated for carcinoma of the prostate at our university medical center. Of the patients 319 underwent radical prostatectomy and 159 (50 per cent) had positive surgical margins and/or seminal vesicle involvement. Of these 159 patients 46 received postoperative irradiation and the actuarial survival was 96, 90 and 90 per cent at 5, 10 and 15 years, respectively. Among the remaining 113 patients who were treated with an operation alone the corresponding figures were 82, 62 and 21 per cent, respectively (p equals 0.02). Considering deaths only of cancer, the surgery only patients had a 15-year actuarial survival of 25 per cent compared to 90 per cent for those who underwent postoperative radiotherapy (p equals 0.07). Actuarial survival free of disease for the surgery plus postoperative irradiation group at 15 years was 40 per cent compared to 28 per cent for the surgery only group (p equals 0.34). Actuarial local control in the irradiated patients was 96 per cent at 15 years versus 32 per cent for the surgery only group (p equals 0.009). Actuarial survival free of distant disease at 15 years was 42 per cent in the irradiated versus 72 per cent in the nonirradiated groups (p equals 0.104). Severe complications attributable to radiation included 3 cases of radiation cystitis, 1 patient with urinary incontinence and leg edema in 9 per cent of the patients undergoing postoperative irradiation compared to 2 per cent of those treated with radical prostatectomy only. Postoperative irradiation appears to be indicated in patients with carcinoma of the prostate who undergo radical prostatectomy and who have positive margins and/or seminal vesicle involvement. Local control is markedly improved (p equals 0.009) and actuarial survival also is benefitted. There was a trend toward decreased deaths of cancer with postoperative irradiation that approached statistical significance. Postoperative irradiation did not improve survival rates free of disease and free of distant disease over those achieved with surgery alone. This finding suggests that while postoperative irradiation may not improve the ultimate cure rate by controlling local disease, early deaths of cancer are reduced resulting in a meaningful increase in survival for these patients.