30 Based on the incidence of carcinoma of the thy- roid among the | 7,000 Marshallese reported by the Trust Territory, the expected incidence in the Rongelap high exposure group would be 0.056 cases over the 15-year period. The finding of 3 cases (5.6% incidence) thus represents a considerable increase over the expected number( p<0.01, x? test). Among the 157 Utirik people about 0.14 cases would be expected, and | case is reported. It seemsless likely that this single case would involve radiation etiology in view of the low dose received by this group and the fact that no nodular lesions had been noted among the children. Based on the present incidence of thyroid malignancy in the high exposure Rongelap group,the risk of this malignancy developing — per 10% persons per rad per year — is 3 cases for the children exposed at < 10 years of age, 10 cases for the older people, and 5.6 cases for the group as a whole. The risk in the Marshallese children is not inconsistent with that reported by others.34 —— eer Radioiodine exposure has been generally con- sidered less effective than x-radiation in producing thyroid lesions, primarily because few thyroid tumors have been noted following radioiodinetherapy.26 It seems likely, however, that the scarcity ofsuch findingsis related to the high doses of radiation used (5000 to 10,000 rads or morein thetreat- mentof hyperthyroidism and 50,000 rads or more for ablation of the gland to ameliorate symptoms in certain diseases). Such doses probably are so destructive that they preclude proliferative activity and malignant transformation. The increasing incidence of hypothyroidism without tumorformation, years after treatment of hyperthyroid patients with radioiodines,illustrates this point. Tumorformation in animals is not always a dose dependent phenomenon.*5 Shellabargeret al.36 showed that the incidence of breast tumors in rats reached a maximum atabout 400 rads andfell off with higher doses. Lindsayet al.?9 reported thatin rats doses of 1311 >> 200 to 400 uCi wereless carcin- ogenic than lower doses. Mark and Bustad reported similar findings in sheep.37 Although the dose to the thyroid gland in the Marshallese was generally considerably below the dose of 131] used for therapy of hyperthyroidism, the doses received by someof the children were probably above the optimum carcinogenic range and therefore the true risk per effective rad maybegreaterin this group. The paradoxical finding of greater risk in the older group appears to be in line with this reasoning. The two stunted Marshallese boys who showed almost complete atrophy of the thyroid gland with no evidence of nodular development are probably comparable with cases of hypothyroidism developing in patients years after radioiodine therapy. The thyroid exposuresofthe Rongelapese wereslightly different from those of patients treated with 131] because their thyroids were not hyperplastic when exposed, at least part of the radioiodine isotopes were ofshorter half-life than 131], and their exposure was complicated by gammaradiation. Somefactors secondary to rad... ion exposure might have enhanced the development of thyroid lesions in the Marshallese. Iodine deficiency or goitrogensin the diet did not appear to be among these. However,the physiologicalstresses of puberty and pregnancy mayhave played a role in the developmentofthelesions. For instance, the development of 10 cases of nodular goiter in the children during or nearthe time of puberty mightindicate that this stress may have enhanced nodularde-. velopment. In 3 females who later developed thyroid nodules the demand of multiple pregnancies may have beenrelated. Since the latent period between exposure and nodule discovery varied only between 10 and 13 years, it may be argued that developmentof thyroid nodules was independent of these stresses. On the other handit is noteworthy that all the women whodeveloped malignant thyroid lesions had had muliple pregnancies. The occurrence of thyroid nodules in the Marshallese was only slightly higher in females than in males (ratio of 1.3 to 1.0). This is similar to the nearly equal sex distribution reported by Toyooka et al.33 for thyroid nodules developing in persons irradiated over the neck region in infancy. However, in the case of carcinoma of the thyroid the expected female preponderance occurred.?& The insidious development of growth retardation in some of the Marshallese children before clinical evidence of thyroid abnormality or deficlency was recognized demonstrates the apparent sensitivity of growth and developmentprocesses to borderline or subclinical thyroid deficiency. All possible steps are being taken to help the children adhereto the present thyroid trez_.nent schedule so that they will achieve satisfactory growth and maturation. Careful medical surveillance of these exposed people, including the Utirik, will be essential in future years. The latent period for the develop-