The introduction of chemotherapy dramatically changed the epidemiology of tuberculosis as the risk of infection was thereby nearly eliminated. The present paper illustrates the risk of disease under these conditions. A large and representative segment of the Danish population, a total of over 626 000 persons aged 15-44 years, was examined by a standardized technique in 1950-52 and has now been followed for 12 years. It has been possible by means of simple parameters such as infection and vaccination status, X-ray lesion and age to divide the population into groups with widely different incidence rates. The time trend in disease rates among vaccinated persons and natural reactors suggests that post-primary tuberculosis is of great significance in the present tuberculosis situation. Three-quarters of all cases stem from the natural reactors. It would have been of great practical significance to identify high-risk groups which yielded a great part of the patients. This was not possible since the majority of cases developed among reactors whose distinctive feature was that they were infected at time of examination.