Prostatic cancer (CaP) is a major cause of mortality and morbidity in men. Prostate specific antigen (PSA) can be used to find cases of prostatic cancer, but at some cost. The clinical biochemistry of PSA has been comprehensively reviewed in a recent issue of this journal.' The pros and cons of using PSA as a screening tool for prostatic cancer have also been extensively reviewed.r? Screening for CaP is influenced by technical factors governing test performance, the natural history of the disease, health economic and ethical factors. Purists would argue that screening is something that is applied only to populations, while others would argue that screening is looking for disease in an asymptomatic person. There is potential for conflict in balancing population interests against individual interests. The body of informed opinion is currently against the use of PSA as a screening tool for prostatic cancer. However, the numbers of PSA tests performed in clinical laboratories are rising as evidenced by a 90% increase within the authors' laboratory over the last year. In part this reflects the ad-hoc use of PSA as a screen for CaP, a practice which generally lies outwith a formal population based screening programme and is not a planned healthcare strategy. Virtually all of the prospective trial data on the use of PSA as a screening test has been derived using the Hybritech Tandem total PSA assay. A search for greater diagnostic specificity has lead to the use of PSA density, PSA velocity and, more promisingly, 'free' PSA. However, none of these techniques have been evaluated in prospective trials. In this personal view we discuss the important issues raised by ad-hoc testing, and consider unresolved issues in relation to screening and the potential problems posed by new assays. We offer some advice to the patient or doctor who asks for a PSA test and for the laboratory worker who has to perform the 'gatekeeper' role. The incidence of prostatic cancer is increasing, partly as a result of greater ascertainment but also because of increased longevity. It is second to lung cancer as a cause of mortality and morbidity in men with malignancy. The prevalence of prostatic cancer in postmortem series is reported to be as high as 32·9% compared to the prevalence of clinical disease, which is around 1·2%.5 This implies that most prostatic cancers do not progress to clinical disease. Notwithstanding this, advocates of screening point to the need for detection of localized cancer for early curative treatment. Opponents of screening point to the costs of screening and the uncertainties about the benefits of intervention. However, they do not offer an alternative strategy for reducing the mortality and morbidity associated with this common disease. The rationale for screening for prostatic cancer as with any other screening programme are based on a management strategy in which there are three components. (1) Case finding: how effective will the test(s) be at finding the disease? (2) Prognosis: how good will it be at predicting the natural history of the disease? (3) Treatment: is there an effective and acceptable treatment for the disease?