Prostatic acid phosphatase (PAP) and specific prostatic antigen (PSA) in the serum of patients with prostatic carcinoma were determined and their clinical value compared. 128 patients were examined, 60 (46.9%) of whom had prostatic carcinoma (4 in stage T2, 27 in T3 and 29 in T4) and 68 with benign prostatic pathology. ROC (receiver operating characteristic) curves were plotted from resulting data and the underlying areas calculated to evaluate the clinical accuracy of the two markers. The area for PSA (0.90 +/-0.03) was significantly greater than that for PAP (0.71 +/-0.05), showing that PSA was better at detecting patients with or without prostatic carcinoma. Maximum clinical accuracy was 0.883 obtained with discriminating values of 0.8 U/L for PAP and 10 ug/L for PSA, confirming the superiority of PSA. However, PAP determination using thymolphthalein monophosphate as the specific substrate for the prostatic isoenzyme, showed greater clinical specificity (98.5%) so that association of the two markers made it possible to eliminate false positive results. In conclusion, results suggest the possibility of using PSA for diagnosing prostatic carcinoma in a selected high risk population. However, the simplicity of the method used for determining PAP and the greater clinical specificity of PAP and PSA combined, suggest determining both parameters for diagnostic purposes.