A series of 280 cases at the Bronchoscopic Clinic affords abundant illustrations of the well known errors incident to inferential diagnosis as to the source of blood issuing from the mouth, and illustrates how a positive diagnosis can be reached in practically all cases in which the blood comes from the food passages proximally to the pyloric antrum, with all the certainty afforded by direct inspection and, in suitable cases, by biopsy. Complete histories and the serologic and other tests in use by clinicians were made in all cases; but in this presentation everything that was negative or irrelevant will be omitted. Nine cases have been selected as illustrative of the most important points. Case 1. —Hematemesis in a man, aged 50. There was a history of gradual loss of weight and strength, slight difficulty in swallowing of six months' duration, melena and occasional vomiting of dark, blood-tinged food