to suggest areas of research and statistical analysis that the NAEG might consider for future use. Most of our attention over the years has been directed toward soil and vegetation studies, with special emphasis on soil. We have had little if any direct involvement in the design of small mammal, cattle grazing, or resuspension studies. Probably our greatest effort has been in the design and analysis of studies for estimating the inventory (total amount) and spatial distribution of 239°249py in surface (0-5 cm) soil at safety-shot and nuclear test sites. The design and analysis of blow-sand mound studies has also been a major effort. In addition, we have contributed to the development of plans for possible cleanup efforts at the Clean Slate 2 site on TTR and to efforts at synthesizing soil, vegetation, small mammal, and cattle data at the Area 13 (Project 57) site. Our experience in the design and analysis of studies at safetyshot sites stimulated us to try new field designs and methods of statistical analysis at nuclear test sites. In our NAEG publications, we have also pointed out the effect that different statistical treatment of the same data can have on estimates of inventory and ratios of plutonium to americium or vegetation to soil concentrations. Through the years, we have developed a close working relationship with other NAEG scientists, particularly Ed Essington at the Los Alamos Scientific Laboratory. Ed's contributions are particularly evident in the field sampling protocols developed for the blow-sand mound and nuclear site studies. ESTIMATING PLUTONIUM INVENTORY AT SAFETY-SHOT SITES Historical Account One of the first NAEG objectives presented to us in 1971 was to estimate the inventory (total amount) of Pu in soil at safety-shot sites. (Throughout this paper, 239°249py and 21am will be denoted by Pu and Am). The first data available for designing such studies was a limited amount of FIDLER,* Pu, and Am data from the GMX site in Area 5. These data were used by Eberhardt and Gilbert (1972) to investigate the feasibility of using the FIDLER and/or Ge(Li) Am data in conjunction with Pu analyses to estimate inventory. It was suggested that since Am and Pu soil concentrations were highly correlated at GMX, the cost of obtaining Pu inventory estimates there would be reduced by relying primarily on Am (Ge(Li)) analyses on soil samples. The available FIDLER data also suggested that they could be a basis for dividing the GMX site into several Am activity strata within which soil samples could be chosen at random according to a stratified random sampling plan. Such a plan *Field Instrument for the Detection of Low Energy Radiation 452