1. Do ball-mill (BM) and <100 mesh (sieved) soil fractions contain equal plutonium concentrations? 2. If the answer to Question 1 is no, then what is the error in estimating inventory by using concentrations from sieved aliquots in place of those from BM fractions? 3. If additional information on the weights and concentrations of the <100 mesh and >100 mesh fractions is available, can they be used to "correct" the <100 mesh and >100 mesh concentrations to what would be expected if BM fractions had been analyzed? Throughout this discussion, it is assumed that BM fractions are the preferred fraction to be analyzed for plutonium and if differences in plutonium concentrations between BM and sieved aliquots are found, then the proper approach is to try and adjust the results on sieved aliquots to those obtained on BM fraction aliquots. The largest single body of data available to investigate Question 1 is given in its entirety in Table 4. These are 239-2405,, concentra- tions (nCi/g dry) in BM and sieved soil fractions from profile samples collected on the TTR. The library number in Column 1 defines a soil sample from which LFE was sent a 10-gram aliquot from the BM fraction, and LASL a 50-gram or larger aliquot from the sieved fraction. The data are tabulated by depth of soil sample and ordered from largest to smallest LASL plutonium concentrations within each depth. The most striking feature of these data are given by the ratio of LASL to LFE mean concentrations in the last columm. For each soil depth, this ratio is seen to decrease as the LASL average concentration decreases. That is, the relationship between BM and sieved Pu concen- trations is not a constant over all levels of contamination. LASL results are up to 20 or 25 times greater than reported by LFE for the highest sieved concentrations. Note, however, that the ratio is less than 1 for the lower LASL (sieved) concentrations, so that for these low-level samples, LFE is reporting higher concentrations than LASL. The logs of the mean data (Columms A and B in Table 4) are plotted in Fig. 9 and a linear regression equation is fitted. It is clear that the fitted line is not coincident with the line that should 134