It was recently demonstrated that urinary prostate specific antigen (PSA) is discordant with serum PSA in many patients after radical prostatectomy. This observation led to the speculation that elevated urinary PSA in the face of undetectable serum PSA may indicate early disease recurrence. We measured urinary PSA levels in 30 patients who had undergone radical prostatectomy for prostate carcinoma and 7 patients who had undergone cystoprostatectomy for bladder cancer. PSA levels of randomly collected urine samples ranged from 0.00 to 22.9 ng./ml. and 0.01 to 8.37 ng./ml., respectively. There was no correlation among urinary and serum PSA levels, pathological stage or type of operation. In 14 patients who had undergone radical prostatectomy and who had measurable levels of urinary PSA voided specimens were divided into initial stream and end stream voided samples. The PSA levels in the end stream voided samples were significantly less than the initial stream sample in 12 of the 14 patients. In men who had undergone radical prostatectomy urethral swab samples were analyzed for PSA. Of 26 patients 24 had detectable levels of urethral swab PSA (range 0.01 to 39.04 ng./ml., median 0.93 ng./ml.). Urethral swab PSA levels did not correlate with serum PSA values or pathological stage of disease. Of 7 patients who had defunctionalized urethras after radical cystoprostatectomy 5 had significantly elevated PSA in the urethral wash or swab samples (range 4.3 to 24.5 ng./ml.). Immunohistochemical analysis of urethrectomy specimens demonstrated positive staining for PSA in 3 of 4 specimens. We conclude that the major source of urinary PSA following total prostatectomy is the urethra itself rather than residual prostate tissue. Measuring serial urinary PSA appears to have limited value in monitoring patients after radical prostatectomy. Whether this urethral PSA can ever contaminate the serum levels of PSA after radical prostatectomy is currently under investigation.