Ucirik. Few direct measurements were made at this time. The data plotted in Fig. 6 for 600, and 37 Fe were derived from Rongelap measurements. Biological variation and errors in the collection and analysis of urine samples introduced larger errors in body-burden estimates than did direct whole-body counting. These variations can be observed in Fig. 5 where 05, data vary widely from the theoretical curve. In contrast, the 1376, data fit the curve closely. The method used to generate Figs. 5 and 6 was not chosen to minimize the weighted sum of squares of deviations of the body-burden estimates and measure- ments from the fitting function (Eq. 2). Instead average values of k and P° were selected to represent all the body-burden data. For Rongelap, the 1370, body burdens varied from the fitted function by a maximum factor of 1.7 and an average factor of 1.4.; the 905, body burdens varied from the fitted function by a maximum factor of 3 and an average factor of 1.6. These factors reflect the quality of fit for directly measured body burdens and urine derived body burdens in general. The integral intake for 50 years and the committed effective dose equiva- lent were derived quantities which depended on knowledge of k and P° for each population subgroup. The 50 year interval chosen for integral intake represented the years 1957 through 2007 for Rongelap residents. For Utirik residents, the fifty year interval represented the years 1954 through 2004. The committed effective dose equivalent was based on this cumulated intake and both values can be found in Table 4. . . tees : 6 An important result of using the fitting function (Eq. 2) was that >on and 137. were the largest contributors to dose equivalent for each population. The 65 . .os : Zn dose equivalent was greatest at Utirik because of a three month inter16