We have measured the serum concentration of prostate specific antigen (tPSA) and determined the percent free PSA (% fPSA) in serial specimens from 64 patients with prostate cancer, 35 patients with benign prostate hyperplasia (BPH), and 3 patients with prostitis. We found that the % fPSAs were not a constant for individual patients during the course of the disease. When we compared only the % fPSA of the first specimen of serial specimens from individual patients, who were largely untreated, 37% of BPH specimens were above 22%, whereas only 1.6% cancer samples were above that value. We also found that 67% of cancer specimens and 14% of BPH samples were below 8%, respectively. Although % fPSA distribution pattern remained similar between two types of specimens, less differentiation was found between BPH and prostate cancer in random specimens compared to the study using first specimens of an individual patient's serial samples. Percent fPSA apparently are affected by treatment. However, the most important benefit for the determination of % fPSA appears to be the sensitivity of % fPSA to identify occult tumors when the tPSAs were in the normal concentration range. Determination of % fPSA also seems to improve the specificity of tPSA, not only during screening for the differentiation between BPH and prostate cancer, but also during monitoring of treatment and recurrence. J. Clin. Lab. Anal. 12:26–31, 1998. © 1998 Wiley‐Liss, Inc.