This study of 1250 gastric lesions treated in the Presbyterian Hospital in New York was undertaken to determine the frequency with which the clinician is faced with the problem of a gastric lesion of questionable diagnosis. There were 650 gastric carcinomas, of which 299 proved resectable, and of these, 150 were of the fungating or mixed categories, as classified by Golden and Stout. Of these 150, 19 were not diagnosed as carcinoma pre-operatively. Of the 351 non-resectable carcinomas, 45 were not diagnosed pre-operatively. The problem of the ulcerating lesion proved more difficult. The 600 ulcers which eventually proved benign are included in the category of lesions of questionable malignancy when 1st seen by the clinician. In addition, there were 105 resectable malignancies of the penetrating, superficial spreading, and linitis plastica classifications in which ulceration was a prominent feature. Forty-four of these were not diagnosed pre-operatively. The diagnosis was considered questionable in a total of 708 cases or 57% of the total of 1250 gastric lesions. Those resectable carcinomas which originally appeared to be benign ulcerations proved to have a better prognosis than the resectable carcinoma group as a whole.