Powerful tests are now available for the detection of prostate cancer, both more frequently and at an earlier and usually more curable stage. In some countries of the world, mainly the United States and some European countries, serum prostate-specific antigen (PSA) and rectal examination-based screening is routinely applied to men at risk. In other countries, particularly in Northern Europe, routine application of screening procedures for prostate cancer is not accepted for a number of reasons: knowledge about the natural history of early lesions suggests that indiscriminate use of these tests will lead to overdiagnosis and overtreatment, information on the effectiveness of treatment from randomized trials is unavailable, and no evidence exists that early diagnosis and treatment will lead to an improvement of disease-related and overall mortality. In this article a number of critical and controversial issues of screening for prostate cancer are reviewed. This includes the risk of prostate cancer to patients, the efficacy and acceptability of the screening tests, the issue of overdiagnosis in relation to the natural history, evidence concerning the effectiveness of treatment, and the chances that early treatment may lead to an improvement of prostate cancer mortality. In relation to these points it is concluded that prostate cancer is a frequent cause of death in men and that at present the possibility of early diagnosis and treatment represents the only possibility of cure. The question whether cure is necessary in every identified case or in identifiable subgroups remains unanswered at this time. Although the individual positive predictive value of the screening tests is low, a combination offers higher positive predictive values that are in the range of 70–80%. With proper streamlining of the screening tests, it may be possible in the future to detect one cancer in 2.5–3.0 biopsies. The most efficient use of the screening tests in the general population still remains to be determined. Estimates related to the amount of overdiagnosis are made. Depending on the definition used, overdiagnosis with one round of screening for prostate cancer is probably in the range two to threefold. The effectiveness of available treatment modalities compared with delayed management has not been studied in a prospective randomized manner. There is, however, indirect evidence, especially relating to the long-term survival of grade 3 patients and to the long-term normalization of serum PSA values after radical prostatectomy, that at least this treatment is effective. In conclusion, the use of the screening tests cannot be withheld from symptomatic men and those who wish to be examined for the presence of prostate cancer. Application to the general population should depend, however, on the result of a prospective randomized study that shows that early detection and treatment will decrease prostate cancer mortality.