DISCUSSION/CONCLUSION Interest in the relative risk of 131, taken internally and external radiation dose to the thyroid relates to radiation protection and medical care issues. Unfortunately for those interested in obtaining information on this important issue, the complex mixture of radionuclides taken up by the Marshallese precludes such an analysis. The results obtained for these studies are specific to the case where the thyroid dose was due to a mixture of shortlived radioisotopes of fodine, some of which were produced by the decay of tellurium within the body. Current information on animal and human data ¥39 Summarized recently in NCRP Report 80 (2). The Committee concluded that I was less then one third as effective for thyroid cancer induction as external radiation. This can not be compared dyyectly to the results of the present study because of the small amount of I in the Marshallese exposures. In most animal studies, which used rodents, high TSH levels were found to be 131, necessary co-factors for thyroid cancer induction. Thus, goitrogen plus exposures were needed to induce thyroid cancer, except in several studies using Long-Evans rats, ybich behaved differently from all other strains studied. Results of I treatment of children for hyperthyroidism were reported in two large studies. In reviewing results of treatment of nine children, Sheline et al. (9) found that all of them subsequently developed thyroid nodules and one was diagnosed as having of thyroid cancer, about which there was disagreement regarding pathology. thyroid replacement therapy after high endogenous TSH levels. None of those children received I treatment, and all presumably developed In Los iggeles, at a later date, 73 children were treated with approximately the same I dose, all were placed on thyroid replacement, and none developed thyroid nodules (10). Thus the relative risk of thyroid dose from internal emitters compared to external radiation for Marshall Islanders may be influenced by a high TSH co-factor, since thyroid replacement therapy began 11 years after exposure. Replacement therapy was recommended only for the high-dose group which, at that time, was thought to be the people at Rongelap. Also no increased incidence of thyroid cancer was seep in large numbers of human subjects exposed to similar or higher doses of I in the treatment of thyrotoxicosis (11), or in children given i3ly in lower diagnostic doses (12). Hypothyroidism is a nonstochastic effect of ionizing radiation exposure, with estimated threshold for induction of 2000 rad to the thyroid (1). In the Marshallese children, whose thyroids were exposed to doses in the several thousand rad range, hypothyroidism and increased TSH levels certainly existed in the early years following exposure. In later years, uneven acceptance of thyroid supplementation by children may have led to persistent increased TSH levels. The combination of high TSH and high internal and external radiation doses may account for the unusually high incidence of nodules in this population, and in the unusual age distribution of sensitivity. The numbers of individuals in the study are small, and statistical segregation of the interacting factors is not possible. Thus, it will be difficult to draw precise conclusions from this study with respect to apportionment of risk between internal and external doges - pyrther, the differences between the Teqdologica) characteristics of I, I, and I and the larggy doses from T and I make it difficult to assess the relative risk of I and external radiation in this circumstance. A simple statistical model was used (3) to indicate the one sigma confidence interval. This confidence interval is indicated in the following paragraph in parentheses. The standard deviation of the risk estimate, E, was 1.5 times the average value for the risk estimate, and development of this standard deviation was given by Lessard et al. (3). 28