correlation coefficient between residuals and observed, and between residuals and fitted values. In addition to these analyses of residuals, the estimated surface is examined by considering the density and pattern of sample collection points relative to estimated contours, and by considering prior information available on the concentration surface of 24lam in Area 13. At the Area 13 study site, concentrations of *+!am in soil are known to be correlated with 239°249py concentrations (Gilbert et al., 1975, p. 403). Before sampling began, information on the concentration surface of 24lam was obtained from portable field instrument (FIDLER) surveys of the Project 57 site. The FIDLER was used to take readings (one foot above the soil surface) of 241am (net 60 KeV) at 400-foot intervals over the entire area and at 100-foot intervals near GZ. These readings give a general indication of the concentrations of 239°24 py in soil in Area 13 and were used to define the strata in Figure 1. Gilbert e¢ al. (1975) estimated the correlation between FIDLER readings and 2397240py concentrations to range from near zero for stratum 1 to about 0.70 for stratum 5 and 6. METHODS Sampling Design Surface soil samples (0-5 cm) were collected at random locations within each stratum (Figure 1) according to a stratified random sampling plan. The number of samples allocated to the strata were determined using as a guide the optimum allocation formula for stratified random sampling (Cochran, 1963, p. 97, equation 5.20). Details of the design and allocation are given in Gilbert et al. (1976a). A total of 173 samples was collected, of which three were lost leaving 170 for statistical analysis. Samples were dried for 24 hours at 105° and ball-milled for five hours (Kayuha et al., 1974). Ten-gram aliquots of ball-milled soil were analyzed for 239°240pu using wet chemistry techniques. data in units of uCi/m™ tion locations The resulting are listed in Appendix A along with their collec- (Nevada Grid Coordinates) .* *Two observations in stratum 6 (identified by +t in Appendix A) were approximated using the average 239°240py/24lam ratio for the study site and estimated concentrations of 24!am obtained for the two samples. Due to inappropriate amounts of tracer added to the aliquots, 239°240py con- centrations were not available for these two samples. All 2*!am concentrations were obtained using a Ge(Li) counter. The average ratio + standard error was 9.4 + 0.14 (from Table 18 in Gilbert et aZ., 1975) computed from 2397240py and 2414m counts on the same aliquots. footnote in Appendix A for further information. 323 See +t