Before any discussion of the data presented in Table 1, the reader should be reminded of certain caveats concerning the unexposed comparison populations used in these studies. These populations consisted of a group of unexposed people of Rongelap (established in 1958) that has varied in number from 150 to 200, who have been examined regularly, and additional groups of unexposed people from the atolls of Rongelap, Utirik, Likiep, and Wotje added during the past ten years. The latter groups contain many children who were not included in the present analysis, but were used in determining an all-age prevalence of thyroid abnormalities (see Appendix I). As pointed out in Appendix I, it is unlikely that the low doses from residual radiation that some of these people in the comparison groups received would produce any detectable thyroidal effects. Of the unexposed populations, the best comparison group is that established in 1958, when individuals were matched for age and sex. Unfortunately this group has suffered attrition over the years, and adequate matching of new persons to replace those missing has not been possible. The more recent thyroid comparison groups have had fewer examinations (some only one) and encompass only segments of the island populations. In addition, an element of possible bias in this group is that some people may have presented themselves for examination because of a suspected thyroid problem. This would tend artefactually to increase the incidence of thyroid disease in the control groups. An opposite bias may occur because most of the established control groups were examined many times over a long period as compared with the added control groups so that the fractional incidence of thyroid disease in the exposed population shown in Table 1, even after correction, may be too high. Table 2 compares the prevalance of thyroid nodules in the two control groups and in the combined groups. Individual island listings for prevalance of thyroid nodules can be found in Appendix I. The data in Table 2 indicate no significant difference between findings in the two comparison groups; therefore, the data from the two groups combined were used in the calculations culations for Table 1. Nevertheless, the caveats referred to above should be kept in mind in interpreting the prevalence and risk calculations. Table 1 shows a considerably greater prevalence of thyroid abnormalities (total and malignant nodules and hypofunction) in the exposed RongelapAilingnae groups than in the Utirik and age-matched comparison populations, particularly in the youngest Rongelap age group.* The Utirik group shows a slightly greater prevalence of thyroid abnormalities than does the comparison population, but the youngest Utirik age group appears to be comparatively less affected than the corresponding young Rongelap group. Table 3 shows that the observed versus expected ratios are greater for the Rongelap-Ailingnae group than for the Utirik group on the basis of prevalence in the unexposed Marshallese. *Prevalence and risks for benign nodules are not presented in Table 1 since 5 of 46 nodule cases in the exposed and 15 of 35 nodules in the unexposed (age-matched) groups did not have surgery. (See Table 1, Appendix IV.) Carcinoma prevalence is presented with the realization that values possibly represent an underestimate of prevalence in view of the unoperated cases. SOCbh TE "38