This study was undertaken in an attempt to evaluate the effect that a given number of roentgens, delivered to a carcinoma of the skin, had upon the success or failure of treatment. By comparing the results in large groups of lesions each given a different dosage, it was hoped that an optimum dose might become apparent. With the establishment of such an optimum dose, the existing confusion as to the proper roentgen dosage for skin cancer should be considerably lessened, and a standard of treatment more uniform throughout the country might be accepted. It is a well recognized fact that few patients die of carcinoma of the skin, and also that recurrence, even after months and sometimes years of apparent cure, is not uncommon. A study based on three-and five-year cures following treatment of these lesions is never entirely satisfactory, largely because approximately 25 per cent of the patients affected by skin cancer, due to their age, die of intercurrent disease within five years from the time of treatment. Also, many of these patients are too feeble for the enforcement of routine follow-up, and on this account many real five-year cures may be listed as lost.4 In the present discussion, the word “failure” will be used arbitrarily to indicate both those cases in which the lesions were not immediately destroyed by a given treatment and those in which there was later (months or years after treatment) recurrence. (Actually, most of the lesions referred to as failures were subsequently destroyed or controlled by irradiation or surgery.) In only 3 of the entire group studied could the carcinoma of the skin be considered. as a contributing cause of death. Undoubtedly, a certain number of lesions classified as successfully treated recurred without our knowledge, or may still recur. Any such inaccuracies, however, should not influence the comparative results, since they should apply more or less equally to all groups regardless of the roentgen dosage used. Approximately 1,500 carcinomas of the skin were reviewed for the purposes of this study. (Lesions of the mucocutaneous junctions are not included.) They were found on 1,300 patients. The lesions were taken in the order in which they appeared on the hospital files between the years 1930 and 1944 inclusive. The group is a selected one only to the extent that lesions which had received treatment elsewhere, prior to admission to the Massachusetts General Hospital, were discarded. This was done because the purpose of the study was to evaluate the various methods of treatment used at this hospital. Ten hundred and thirty-five lesions were irradiated with x-ray and 387 were treated surgically. Histologic examination in 630 of the total group showed epidermoid carcinoma in 183 and basal-cell carcinoma in 447. In the remaining cases, treatment was given without histologic confirmation because the clinical appearance was characteristic of carcinoma of the skin.