exposure of children less than 10 years of age, and upon data available through 1979. A central value of 4.0 cancers per million person-rad-years at risk was reported, but after review of their report, we modified the estimate to 4.9 cancers per million person-rad-years at risk. Our result, based on this modification, is discussed in the text and is noted in Table 7. The adjustment was based on weighting the risk coefficient from each study according to the number of excess cancers observed; that is, we gave more weight to cancer risk coefficients developed from studies reporting the greatest number of cancers. The BEIR risk coefficient was based on a minimum latent period of 10 years and on studies involving only external irradiation of the thyroid. Risk coefficients for external and internal radiation were given in NCRP Report 80 (2), and these coefficients were estimated for a five-year legent pgyiod- Report 80 indicated the external risk coefficient applied to I and I intake, but not for I exposure. The two short-lived isotopes of iodine were assumed to have the same effectiveness as x rays, because of the fairly uniform distribution of dose, and because of the comparatively higher dose rates (2). In our analyses, we used risk coefficients for external exposure computed for 5- and lO-year latent periods derived from the following reports. We used external risk coefficients from NCRP Report 80 because they were based on a five~year latent period, and these appear in the results section along with the coefficients developed by the Working Group, which were based on a ten-year latent period. Risk coefficient estimates, made here, were based on the total external and internal thyroid dose, the total number of cancers, the risk value published for external irradiation of the thyroid, and the partitioning of external and internal dose as follows AB+CD# (A+ CE, (1) where A = the person-rad to all thyroids from radioisotopes of iodine, B = the risk coefficient for internal exposure of the thyroid from radioisotopes of iodine, cancers per person-rad-years at risk, C = the person-rad to all thyroids from external gamma radiation, D = the risk coefficient from external exposure of the thyroid, for exam ple, 1.0x10 cancers per person-rad-years at gisk for adults, or in the case of children <10 years of age, 4.9x10°° rad-years at risk, and cancers per person- E = the risk coefficient determined from the observed health effects, the total thyroid dose, and the spontaneous rates of thyroid disease in the Marshall Islands subjects. The value of E was computed from Eq. (2-1) given in NCRP Report 80 (2). Computations of B and E were for latent periods of both 5 and 10 years, since the length of latent period affects the years at risk and the risk coefficient. Years at risk are the period from the end of the latent period to the time cancer is observed in a subject. The value for years at risk strongly affected the computation of risk coefficients. RESULTS The data in the Appendix are the result of 31 years of medical and 22