PART A: To objectively determine the changes in chemical character of an area subjected to mining and reclamation, prior information is needed. This study represents a broadscale inventory of total chemical composition of the surficial materials of the Northern Great Plains coal region (western North and South Dakota, eastern Montana, and northeastern Wyoming); data are given for 41 elements in A and C soil horizons. An unbalanced, nested, analysis-of-variance design was used to quantify variation in total content of elements between glaciated and unglaciated terrains, for four increasingly smaller geographic scales, and to quantify variation due to sample preparation and analysis. From this statistical study, reliable maps on a regional basis (>100 km) were prepared for C, K, and Rb in A and C soil horizons; for N a, Si, Th, D, and Zn in A-horizon soil; and for As, Ca, Ge, and Mg in C-horizon soil. The distribution of variance components for the remaining 29 elements did not permit the construction of reliable maps. Therefore, a baseline value for each of these elements is given as a measure of the total element concentration in the soils of the Northern Great Plains coal region. The baseline is expressed as the 95-percent range in concentration to be expected in samples of natural soils. PART B: A reconnaissance study of total concentrations of 38 elements in samples of soils (0-40 cm deep, composite) from the Bighorn and Wind River Basins of Montana and Wyoming indicates that the geographic variation for most elements occurs locally (5 km or less). However, in the Bighorn Basin, Zn exhibits significant regional variation (between geologic units); and in the Wind River Basin, AI, Cr, K, Mn, Mo, Ni, U, and V exhibit similar variation. For the remaining elements, the lack of regional variation suggests that a single summary statistic can be used to estimate a baseline value that reflects the range in concentration to be expected in samples of soils in each basin. The concentrations of most of these elements in both basins are not much different from those measured independently in the Powder River Basin of Wyoming or in the Western United States. In addition, data from an analysis of variance provide an estimate of the number of random samples within an area of specified size (10 km square, approximately a township) that are needed to prepare a reliable map of total element concentration in soils for each of the elements in each of the basins.