In estimating the effect of fallout radiation on the incidence of leukemia, it is generally assumed that there is no threshold and that the increase is proportional to the dose, even at very low levels of exposure. Based on this assumption and on the scanty available data, it may be cal- culated that exposure at the rate of 0.1 r in 30 years (the estimated gammarayfallout dose rate) would increase the annual leukemia incidence in the population of the United States by 36 cases. It has been suggested that Sr®° in the bone irradiates bone marrow in close proximity to bone and that this might induce leukemia. There is at present no evidence that leukemia can be produced in this way but the possibility cannot be excluded. It may be estimated that when the equilibrium concentration of Sr* in bone (20 ppc per gram of Ca, stated in Section 3) is reached, the dose rate to adjacent bone marrow will be about 0.03 rad per year. On the assumptions made over (and the inference that only leukemia of bone marrow origin could possibly be produced in this way), it may be estimated that at worst leukemia deaths in the population of the United States would be increased by about 160 per year, some yearsafter the above mentioned equilibrium concentration of Sr®® had been reached. According to these calculations, the total possible increase in the leukemia deaths in the United States attributable to external and internal radiation from the estimated equilibrium amount of fallout, would be 196 per year. This amounts to 1.7 per cent of the present annual leukemia deaths (11,400). It should be noted that any analysis or interpretation of the effect of fallout on the inci- dence of leukemia must take into account the fact that the reported death rates from leukemia in this country rose sharply between 1930 and 1954, as contrasted to a much more gradual increase between 1910 and 1930. Whether this increase represents an absolute increased incidence is less clear, for during these years there have been great changes in the span of coverage of reporting, in listings of cause of death, and in diagnostic criteria. Furthermore, analyses of 1954 figures demonstrate a striking age distribution with a peak during the first five years of life, a leveling off until the fifth decade, followed by a precipitous increase into the eighth decade. It should also be borne in mind that many drugs and industrial chemicals, by injuring hematopoietic organs, could be capable of inducing leukemia. 4.4 Bone Tumors The induction of bone tumors in humans by ingested radium has been established clearly. Radiographic studies of the bones of living persons, apparently in good health, with long term radium body burdens of the order of 1 pc, show small regions of damaged bone in different parts of the skeleton. Presumably these nonmalignant bone lesions are the result of local concentrations of radium. It is surmised that when bone sarcoma developsit originates in one of these regions. This is in accord with the well-known fact that cancer of the skin in grossly overexposed radiologists develops in localized areas showing persistent damage. It seemsthat the existence of a damaged region of tissue is a usual prerequisite for the development of cancer. In the case of radiation-induced cancer of the skin in radiologists, it is known that to produce the preliminary permanently damaged skin areas large doses of radiation are required (certainly more than 1000 r). Because of the high energy and the short range of the alpha rays of radium and its disintegration products, small local concentrations of radium are sufficient to deliver very large doses to the surrounding tissue in the course of the long latent period before bone sarcoma develops (15 or more years). It is well known that in general the smaller the radium body burden the longer it takes for cancer to develop. Among the radium dial painters, those who swallowed large amounts of material died within a few years of anemia, hemorrhages, and infections, rather than cancer. A radium body burden of 1 yc produces a bone dose rate of about 40 rads per year (depending somewhat on the proportion of radon decaying in situ) if it is uniformly distributed throughout the skeleton. lf, as is likely, there are local concentrations of radium, the dose rate in these regions can be much higher. The smallest body burdens of pure radium that is known to have caused bone sarcomain 24 years, is 3.6 pc at the time the tumor developed. Since a considerable fraction of the initial body burden was eliminated in this time, the bone dose in this case, for a uniform distribution, was nearly 6000 rads. In x-ray therapy it is often necessary to irradiate bone in order to deliver the desired dose to a deep-seated lesion. Occasionally tumors have appeared in such irradiated bones 273 ; Pah