In the course of a study of duodenal ulcer in a group of men undergoing periodic health examinations, it became necessary to set up a diagnostic classification which would serve as the true or absolute status of the individual. That is, we wanted to be as certain as possible of the diagnosis in order to make a study of the association of other factors with duodenal ulcer. Since the nature of the disease process requires that healing be manifested by scar tissue formation, the ultimate diagnosis of duodenal ulcer probably rests upon microscopic examination of serial sections of the duodenum. Such a procedure, however, is not generally feasible, and clinical and radiological criteria must serve to make our diagnoses. Most often, the latter is taken as the more accurate, although clinical observations occasionally prevail when the two are in disagreement. As Templeton (1) has shown, an ulcer can heal without leaving any deformity visible by radiological means. The radiological criteria of ulcer are demonstration of a crater, a niche, or a constant deformity of the duodenal bulb. Findings such as spasm, irritability, etc., are suggestive only. The accuracy of the radiologic criteria has been repeatedly estimated. These studies have usually been carried out on patients who have required hospitalization and frequently surgery. In such persons, in whom the disease has run a severe course, the criteria have been highly accurate. This would be true too of clinical criteria. It must be remembered, however, that duodenal ulcer is, for the most part, a benign condition, and the majority of cases do not come to surgery, hospitalization, or even regular medical care. In cases occupying this end of the biologic gradient of the disease process, the same degree of accuracy is not to be expected. Investigation of the accuracy of radiological criteria in individuals who were in relatively good health led to the following studies. From April 1958 through January 1959 in a periodic health examination clinic, 245 white males primarily in their fifth and sixth decades had received gastrointestinal x-ray examinations which included both the conventional spot-films and also cinefluorograms. A diagnosis had been made for the Clinic records, based upon conventional fluoroscopy and the two technics just mentioned, by one of the radiologists participating in the study. This will be referred to as the original reading. Of the 245 sets of films, 50 had been reported as positive or suggestive on the original reading. These were all taken for study. Of the 195 remaining sets, reported on the original reading as negative for ulcer, 86 or 44 per cent were randomly selected. From the 136 sets of films all identifying data were removed. Each of three radiologists, independently and without knowledge of the clinical history or previous readings, read the spot films, then the cinefluorograms, and then the two concurrently.