To compare the clinical usefulness of the serum markers prostate-specific antigen (PSA) and prostatic acid phosphatase (PAP), we measured them by radioimmunoassay in 2200 serum samples from 699 patients, 378 of whom had prostatic cancer. PSA was elevated in 122 of 127 patients with newly diagnosed, untreated prostatic cancer, including 7 of 12 patients with unsuspected early disease and all of 115 with more advanced disease. The PSA level increased with advancing clinical stage and was proportional to the estimated volume of the tumor. The PAP concentration was elevated in only 57 of the patients with cancer and correlated less closely with tumor volume. PSA was increased in 86 percent and PAP in 14 percent of the patients with benign prostatic hyperplasia. After radical prostatectomy for cancer, PSA routinely fell to undetectable levels, with a half-life of 2.2 days. If initially elevated, PAP fell to normal levels within 24 hours but always remained detectable. In six patients followed postoperatively by means of repeated measurements, PSA--but not PAP--appeared to be useful in detecting residual and early recurrence of tumor and in monitoring responses to radiation therapy. Prostate massage increased the levels of both PSA and PAP approximately 1.5 to 2 times. Needle biopsy and transurethral resection increased both considerably. We conclude that PSA is more sensitive than PAP in the detection of prostatic cancer and will probably be more useful in monitoring responses and recurrence after therapy. However, since both PSA and PAP may be elevated in benign prostatic hyperplasia, neither marker is specific.