It should be noted that long range estimates of this kind are only intelligent guesses at best. For one thing, there is practically no information as to how Sr® will be distributed in the soil after a long period of time, which, obviously, will influence its incorporation in plants. However, it seems fair to say that most of the envisaged factors that cannot be evaluated today, will probably make the actual concentration of Sr*° in human bones lower than the estimated values, rather than higher. For the purpose of evaluating possible injury to the population of the United States, we shall assume that the average concentration of Sr*® in bone in equilibrium with fallout will eventually reach 20 pyc per gram of Ca, if tests continue at the present rate for many years. Since the average surface concentration of world-wide fallout is considerably lower outside the United States, the value applicable to the world’s population is considerably lower thanthis. The present maximum permissible concentration of Sr®° in bone for a large population is 100 puc per gram of Ca, according to the National Committee on Radiation Protection and the International Commission on Radiological Protection. The NAS Committee recommends the same concentration, but 50 pyc of Sr’ per gram of Ca is also mentioned. Therefore, the estimated (biological and radioactive) equilibrium concentration of Sr®® from fallout in human bones in the United States, is 20 or 40 per cent of the MPC for large populations, recommended by authoritative bodies. It should be noted that all figures given above are averages. Through a combination of unusual circumstances it is possible that fairly large numbers of people in some localities may accumulate Sr” in their bones to a value five or ten times greater than the average for the United States or the world, as the case may be. This would bring the bone concentration of Sr*® above the permissible limit for large populations, but still below the limit for occupational exposure. 4 ESTIMATE OF POSSIBLE DAMAGE 4.1 General Considerations The biological effects with which we are concerned in the peacetime fallout problem are those that might possibly result from long continued low level exposure to radiation (externally or internally). Our knowledge of such effects has been derived largely from animal experiments. Since the effects occur also spontaneously, it is always a matter of determining whether there is a real increase in the number of animals showing the effect in question caused by exposure to radiation. In order to obtain a statistically significant difference at very low levels of ex- posure, thousands of animals would have to be used. In practice it has been found expedient to use instead a high level of exposure to obtain statistically valid results using small numbers of animals. The question then arises as to how to estimate the effects of exposure at a much lower radiation level than was used in the experiments. In the case of gene mutations it has peen established, or at least it is believed by practi- cally all geneticists, that the number of induced mutations is proportional to the dose received by the gonads upto the time of reproduction, no matter how low the dose is and no matter how it has been distributed with respect to time. On this basis it is then a simple matter to calculate the number of mutations that would be produced by a dose of radiation, however small, once the number for a large dose is known. To apply the results of animal experiments to man, various assumptions must be made, but there is good agreement among geneticists at least as to the order of magnitude of the effect. In the case of somatic effects; that is, effects manifested in the exposed individual himself rather than in his descendants, the extrapolations to very low radiation levels and from animals to man are carried out ina similar manner. Here, however, the situation is more complex and the estimates are less reliable. The most important reason for the unreliability is that the mechanisms by which these effects are produced are not known. The extrapolation may be made by assuming direct proportionality between dose and number of individuals affected (as in the genetic case). This assumption implicitly denies the possibility that if the radiation level is low enough, a given somatic effect may not be produced at all, a conclusion that at the present time can neither be denied nor affirmed. It may be concluded, therefore, that proportional ex- 271