evident near ground zero (GZ). This probably results in part from (i) using an approximation when transforming concentration estimates from logarithmic to arithmetic scale, and (ii) ignoring any "drift" (systematic change over distance) in plutonium concentrations within 100- x 100-foot cells. The "corrections" to the initial plutonium estimates obtained in Stage 1 are relatively small (less than factors of 2) except in areas to the south and east of GZ where the corrections increase estimated mean concentrations by factors of between 2 and 4. The corrections south of GZ are due to 4 Pu concentrations in that area that are unusually large relative to grid FIDLER readings taken previously in the vicinity. It may be desirable to take additional FIDLER readings on a finer grid in this region to confirm these Pu results. Additional soil samples may then be advisable to further refine inventory estimates. Estimates of plutonium inventory reported by Gilbert (1977) and Gilbert et al. (1975) are compared with those obtained here using kriging. The agreement is within 0.5, 0.3, and 0.3 curies for strata 3, 4, and 5, respectively. However, in stratum 6 (the stratum nearest GZ), the kriging estimate is less than half the inventory estimate reported by Gilbert (1977). This may be due at least in part to ignoring the "drift" that occurs within the 100- x 100-foot cells near GZ. Inventory estimates obtained using strata mean concentrations are much more sensitive to extreme concentrations than those obtained using kriging. A number of statistical problems associated with the analysis of plutonium data for estimating spatial distribution are discussed. These include (a) transformation of skewed data, (b) interpretation of the regression coefficient relating log Pu to log FIDLER data, (c) bias of the estimated coefficient due to measurement errors in both plutonium and FIDLER data, and (d) bias problems involved in transforming estimates obtained in logarithmic scale back to arithmetic scale. The theory of kriging is briefly outlined, and the structural analysis of the Area 13 FIDLER and plutonium data (necessary for kriging) is explained and presented in some detail. A brief review of past efforts at estimating plutonium inventory and spatial distribution at Area 13 is given. No attempt is made here to estimate average plutonium concentrations for strata 1 and 2, regions of relatively low plutonium levels. Any method (including kriging) that attempts to make use of a regression relation-~ ship between plutonium concentrations and FIDLER readings may not be applicable in these low-level areas since in these areas the correlation between these two measurements is not very strong. The suggestion is made that kriging may be most useful in those regions of a study site that lie between the immediate GZ area and the much lower~level areas removed from GZ. Whether or not kriging is used at a particular study site must depend upon an evaluation of the additional information expected to be gained relative to the increased cost of 362