Peroral irradiation of tumors of the oral cavity has been practiced since the early 1900's (8). Its application was accepted because it allowed high doses to the tumor without heavy irradiation of the sensitive intervening tissues, particularly the vulnerable mandible. In late years the development of successful radium technics and the advent of megavoltage beams have reduced the indications for peroral treatment, and Paterson (10) has gone so far as to label the method “an historical relic.” Because of the many successful results observed in our institution, however, its use has been continued, and in recent textbooks (4, 9) the method has been suggested in selected cases. Peroral irradiation avoids the discomfort and attendant hospitalization of a radium insertion, as well as the sequelae of external treatment. The present review was undertaken to establish the rate of local control and of complications, such as necrosis, in relation to the treatment technic. From the study, a better knowledge of case selection, of optimum time-dose relationships, and of the five-year survival rate was anticipated. Material The records of all patients treated for oral cancer between 1940 and 1965 in the Section of Therapeutic Radiology, University of California, San Francisco Medical Center, were reviewed. Those tumors that had received a surface dose of at least 1,500 R by peroral applicator were selected for this investigation. Tumors of the tongue, floor of mouth, alveolar ridge, buccal mucosa, hard palate, soft palate, tonsillar pillar, and tonsil were included. A total of 143 patients met the criteria of the study. Follow-up information was complete on all. Peroral irradiation alone had been employed in 56 patients (37 male and 19 female). External and peroral irradiation was combined in 87 patients (77 male and 10 female). The age distribution was similar in both groups, ranging from thirty-five to ninety-four years with a mean of sixty-three years in the males and sixty years in the females. The combined mean was sixty-two years. The primary lesions were staged retrospectively, based on the measurements given in the record. The T, N, M System described by Fletcher (4) was used. T-1 indicates lesions less than 2 cm in greatest diameter, T-2 those 2 to 4 cm with minimal extension to adjacent structures, and T-3 those more than 4 cm. The T-4 lesions are massive. When any doubt existed, the lower stage was applied. No clearly described T-4 tumors were found, and all lesions greater than 4 cm were staged T-3. Metastatic lesions of the lymph nodes were classed according to appearance on the initial visit or at any time during the illness. The material is not broken down here into the finer classification of N-l, 2, and 3. The possible combinations were too many, and the numbers involved were relatively small.