A total of 84 consecutive men with extracapsular disease after radical prostatectomy who received postoperative irradiation and no adjuvant endocrine therapy were analyzed. Failure was defined as the development of clinical disease recurrence either locally or at a distant site, the development of detectable prostate-specific antigen (PSA) when postoperatively it had been undetectable, or any rise in PSA when postoperatively it still had been detectable. Sixteen of the 84 men had nodal disease. Overall, five-year actuarial freedom from relapse was 60 percent. For node-negative men, it was 64 percent and for node-positive men 43 percent. For the 68 men with pathologic Stage T3N0 disease, five-year freedom from relapse was 73 percent when seminal vesicles were negative and 43 percent when involved. Tumor grade also predicted the likelihood of recurrence. In a multivariate analysis the time interval from surgery to radiation and the level of postoperative PSA (detectable versus undetectable) did not influence the likelihood of relapse nor the median time to relapse. Fourteen separate patients treated with radiation alone for palpable tumor recurrence were also analyzed. Fewer than 40 percent were disease-free only two years after irradiation. We conclude that when treatment failure is defined in biochemical as well as clinical terms postoperative irradiation reduces the rate of relapse at five years relative to recently reported series in which adjuvant irradiation was not given. The additional morbidity is low. Whether or not this will translate into an overall cause-specific survival gain is currently unclear.