Roots are an important contributor of recalcitrant organic carbon compounds for soil organic matter formation, but little is known about the composition of many species. There is a need for techniques capable of rapidly assessing significant, but often overlooked, carbon sinks such as roots. Diffuse reflectance mid-infrared spectroscopy (DRIFTS) has great potential for the analysis and characterization of plant root composition. The objectives of this research were to compare the DRIFTS spectra of roots of different species using whole root samples and root fiber fractions and to identify spectral features indicative of important root macromolecules in order to evaluate the potential of DRIFTS to determine root composition. A wide variety of roots from agronomic and horticultural crops, ornamental plants, and native plants were collected and analyzed by DRIFTS. Samples were ground in a cyclone grinder to pass a 20 mesh screen and scanned without KBr dilution from 400 to 4000 cm − 1 using KBr as the background. In addition, traditional fiber analysis of a subset of roots and DRIFTS analysis of the resulting fiber fractions (neutral detergent fiber, acid detergent fiber, hemicellulose, cellulose, lignin and wax) were utilized to identify spectral features associated with those fractions. Results indicate that the roots of the same species are similar despite differences in climate, soil and fertilization, while important differences were noted between roots of different species. Tree root lignins appeared to be similar to their above ground counterparts based on comparison with published data. Root lignins for all studied species varied by species. Spectral analysis was consistent with chemical fiber analysis composition data and revealed features that may be indicative of root suberin content. Further research is necessary to confirm that these features are related to suberin. Overall, the results of this research demonstrate the potential of DRIFTS for the characterization of plant root composition and as a tool to rapidly screen large numbers of samples for more effective utilization of more time-consuming analytical procedures.