The Problem The objective of this paper is to present an approach for the estimation of Pu concentration in soil at the Area 13 "safety-shot" site. We do not attempt to evaluate the Pu concentration at each "point" in space. Indeed, the variability of soil sample analyses is so large that it does Accordingly, we will not make much sense to aim at very local values. The size of these estimate average Pu concentrations over grid cells. cells may be chosen at will, in order, for example, to match the definiWith the tion of safety standards given, e.g., in acres or hectares. present data, it is convenient to consider 100- x 100-foot cells, defined by a sampling grid covering the region surrounding GZ where Pu activity is highest. Due to extreme skewness of the Area 13 data, our analyses are performed in logarithmic scale. It would be easiest for the statistician if he could also report results in that scale, but unfortunately, safety standards are not stated in log scale. In this paper, final results are given as both arithmetic and geometric means since it is not clear which estimate is to be preferred for comparison with safety standards. The serious bias problems encountered in transforming results from logarithmic to arithmetic scale are discussed. The Data In 1957, a device containing plutonium was blown apart in Area 13 by chemical explosives partly to test for "safety" against fission reactions in an accident situation involving an atomic weapon. A consequence of the test was the contamination of the immediately surrounding desert soil and vegetation with Pu and Am. The area was fe..ced off and the contamination has been monitored since that time. Since 1971, the Nevada Applied Ecology Group has studied the area by taking field instrument surveys and by collecting soil, vegetation, and animal tissue samples. This present study makes use of the following data accumulated in Area 13: 1, Pu concentrations (in uCi/m*) determined by wet chemistry on surface (top 5 cm) soil samples taken at random locations within Am activity strata (count: 2. 174). These data are plotted in Figures 3 and 4. Am concentrations in surface soil, obtained from FIDLER readings (in counts per minute (cpm)) at one foot above the surface. These were taken (a) at 145 of the above random locations and at 9 others where Pu was not measured, (b) in a 400-foot systematic grid over the entire area bounded by an outer fence (count: 352), 367 , t f , i t