ALTHOUGH the classification of bone tumors is still difficult and unsatisfactory, a definite improvement in this respect has occurred during the past few years. Moreover, it is gratifying to know that radiology has played a valuable part in this improvement. The roentgenographic features of such abnormal growths are often distinctive, but sometimes they are not and an absolute diagnosis cannot be made. When tissue from such a tumor can be removed for microscopic examination, the growth can usually be identified. In certain cases, unfortunately, different pathologists cannot agree on the essential character of the tumor. As the experience of radiologists has increased, their attention has been attracted by the fact that certain tumors of bone are sensitive to radium or roentgen rays, while others are much less sensitive. By radiotherapeutic tests alone, or by correlating such tests with the pathologic data, it is now possible to distinguish certain tumors from others with which they had previously been confounded. Thus, one of the features which influenced Ewing in recognizing diffuse endothelioma of bone from osteogenic sarcoma was the peculiar radiosensitiveness of the former in contrast to the relative resistance of the latter. To one who has had sufficient experience this feature is so striking and, on the whole, so uniform that it constitutes a most valuable point by which these two varieties of tumor can be distinguished. As time goes on, the relative radiosensitiveness of different varieties of tumor bids fair to assume an increasingly important rôle in the diagnosis and, perhaps, in the classification of tumors arising in the human skeleton. Physicians who are not well versed in radiology often mistake radiosensitiveness and radioresistance as absolute expressions. To them a tumor is either sensitive or resistant to irradiation, and intermediate gradations are not recognized because of lack of experience with this phase of medicine. Experienced radiologists, on the contrary, use these terms-only in a relative sense because they know that, just as different varieties of normal cells vary greatly in sensitiveness to roentgen rays or radium, tumors vary equally in this respect. Moreover, it has been established that the radiosensitiveness of any tumor corresponds to that of the cells of which it is chiefly composed. In the case of a mixed tumor, its sensitiveness to irradiation represents a composite of its cellular elements; when sensitive cells predominate, the degree of sensitiveness is greater than when the proportion of sensitive cells is small. It is clear that the scale of radiosensitiveness of different varieties of neoplasms must, and in fact does, correspond to the degree of sensitiveness of its essential cellular constituents.