Visible and near-infrared field spectral reflectance measurements of plutonic rocks were acquired in the 0.45- to 2.45-μm region with a portable fields reflectance spectrometer. These spectra were used to determine spectral signatures for the various rock types and to evaluate the separability of these rocks based on their spectral characteristics. A total of 135 samples were divided into 11 groups based on their mineralogy. These 11 groups approximately correspond to traditional rock classifications and include five granitic groups, three gabbroic groups, and three ultramafic groups. The positions, intensity, and presence of iron, CO3-2, and Al-OH and Mg-Oh absorption bands varied among the 11 groups. Each rock group also had a range of albedos that is characteristic of the group. Stepwise linear discriminant analysis was performed on the spectral data to determine the separability of the 11 groups. Classification accuracy for 30 equally spaced wavelength bands between 0.45 and 2.45 μm was 78% with 10% serious misclassifications. The same analysis was repeated, limiting the spectral data to the wavelength regions corresponding to the proposed Landsat D thematic mapper scanner. Classification accuracy was 71% with 13% serious misclassifications. To asses the effects of weathering on spectral separability of pultonic rocks, a set of 47 spectra of weathered plutonic rocks was analyzed by using the classification functions developed in the preceding analyses. Classification accuracy of these spectra could not be as critically assessed because precise mineralogy was not known; however, serious classification errors totaled 17% with the 30-channel data and 30% with the simulated Landsat D data. These results indicate that plutonic rock types can be successfully discriminated from one another with good results on the basis of their spectral characteristic in the 0.45- to 2.45-μm region.