The Sacaton Mountains are located in the Sonoran Desert, 50 km south of Phoenix. The range consists primarily of Precambrian and Laramide granitoid plutonic rocks, though previous researchers have disagreed on the exact classification and distribution of rock units at this locality. Thermal Infrared Multispectral Scanner (TIMS) data of the Sacaton Mountains were used to classify and map bedrock in the Sacaton Mountains. TIMS emissivity data were linearly deconvolved to mineral endmember abundances. An algorithm was then used to classify the granitoids from their remotely measured mineral abundances. Laboratory spectroscopy, of both weathered and fresh faces of samples, and linear deconvolution of sample spectra were performed. Spectroscopically determined modal abundances were verified by feldspar staining and point counting of sample slabs, as well as thin section analysis. Preliminary results have indicated that rock surfaces are deficient in plagioclase feldspars and, to a much lesser extent, alkaline feldspars, relative to mineral modes determined from lab spectroscopy and point counting of fresh surfaces. In this case, mineral modes determined with TIMS and laboratory spectra of weathered surfaces may not be representative of whole rock mineralogy. However, measured clay content on the surface of rocks appears to be proportional to the amount of plagioclase and total feldspar within the rock. Remotely measured abundances of plagioclase and alkaline feldpars, quartz, and clay, can be carefully interpreted to indicate whole rock modal mineralogy. This technique is applied to remote sensing data of the Sacaton Mountains, where bedrock is classified and mapped from TIMS. These results suggest that differential weathering of minerals, even in very arid climates, can lead to erroneous conclusions about surface mineralogy. However, knowledge obtained from laboratory analysis of field samples can be extrapolated to large areas, with careful interpretation of remotely measured surface mineralogy.