Soil sampling surveys were conducted during 1984-1986 across 50 solution-collapse features exposed on the Coconino Plateau of northern Arizona in order to determine whether soil geochemistry can be used to distinguish mineralized breccia pipes from unmineralized collapse features. The 50 sampled features represent the variety of collapse features that crop out on plateau surfaces in northwestern Arizonaoodeeplyorooted solution-collapse breccia pipes, near-surface gypsum collapses, and sinkholes. Of the 50 features that were sampled in this study, 3 are confirmed breccia pipes that contain significant uranium and base-metal minerals, I is believed to be a sinkhole with no economic potential, and 4 are stratabound copper deposits whose possible relationship to breccia pipes is yet to be determined. The remaining collapse features are suspected to overlie breccia pipes, although some of these may represent near surface gypsum collapse features. However, no exploratory drilling results or breccia exposures exist to indicate their underlying structure. The low cost and ease of soil sampling suggested that this technique be evaluated for breccia pipe exploration. This report provides the locations and geochemical results for the soil sampling surveys and brief descriptions of the 50 collapse features. The analytical results of almost 2,000 soil samples are provided in tabular hardcopy and dBase III Plus diskcopy format. The analytical data is provided in digital format to allow the reader to choose their own methods for evaluating the effectiveness of soil sampling over known and suspected breccia pipes. A pilot survey conducted over 17 collapse features in 1984 suggested that soil sampling might be useful in distinguishing mineralized breccia pipes from other circular features. Followup detailed surveys in 1985 and 1986 used a radial sampling pattern at each of 50 sites; at least one third of the samples were collected from areas outside of the collapse feature to provide background data. Samples were consistently collected from 3-4 inches depth after the pilot survey showed that metal concentrations were similar in samples from 3-4 inches and 7-8 inches depth. The geochemical analyses of the <80 mesh fractions of the soil samples were performed by the U.S. Geological Survey Analytical Laboratories and Geochemical Services, Inc. The analytical methods applied to these samples by the U.S. Geological Survey laboratories included inductively coupled plasma-atomic emission spectroscopy, X-ray fluorescence spectrometry, neutron activation, atomic absorption, delayed neutron activation, and classical wet chemistry for carbon, fluorine, and sulfur. Geochemical Services, Inc. analyzed the soil samples by inductively coupled plasma emission spectroscopy.