The determination of serum prostate specific antigen (PSA) has made the diagnosis of prostate cancer easier. PSA is also used extensively in the follow-up of patients, both treated and untreated. On the basis of material from 308 patients who were mainly managed conservatively, we discuss the usefulness of this practice. It is important to monitor PSA after radical treatment because an increase may indicate local recurrence. In some patients this may lead to further treatment with cure as the aim. For patients under observation only, or those being treated by endocrine intervention, the value of regular PSA measurements is less certain. Where such patients were followed up for at least three years, we found considerable overlaping of PSA values among patients with different outlooks. Within the present therapeutic possibilities it may be better to base their management on clinical signs rather than on PSA. Regular measurements of PSA lead to focusing on this variable, causing unnecessary distress to patients months, or even years, before clinical progression of the disease.