A principal component image enhancement has been effective in applying Landsat data to geologic mapping in a heavily forested area of eastern Virginia. A chestnut-oak forest unit, which occurs on metavolcanic rocks and some metaclastic rocks in the western Piedmont and on highly weathered upland gravel deposits in the eastern Piedmont, can be discerned on a digitally enhanced Landsat winner image. The image enhancement procedure consists of a principal component transformation, a histogram normalization, and the inverse principal component transformation. The enhancement preserves the independence of the principal components, yet produces a more readily interpretable image than does a single principal component transformation. To determine how the chestnut-oak forest unit was being enhanced, average Landsat multispectral scanner (MSS) values were extracted for four sample forest types and were calculated through the inverted principal component transformation. Slope and intercept values for the linear histogram normalization were chosen to keep the scale between the raw MSS bands and the inverted principal component (PC) bands constant. Plots of the inverted principal component data show that the most separation between forest types is in inverted PC band 5. The chestnut-oak forest unit is characterized by a high value for inverged PC band 5 as opposed to a low value for inverted PC band 4. In contrast, raw MSS band 4 is greater than MSS band 5 for the four forest types sampled in the winter image. Such observations cannot be readily deduced from analysis of only the principal component rotation matrix. They may ultimately provide a means to map the distribution of chestnut-oak forest from Landsat.