Knowledge of genetic variability within a crop species is essential to its improvement. The objectives of this study were to assess restriction fragment length polymorphism (RFLP) patterns for oat (Avena sativa L. and A. Byzantina K. Koch) cultivar characterization, to estimate genetic relationships among oat cultivars based on RFLP patterns, and to contrast these estimates with genetic relationships obtained from pedigree information. Eighty-three North American oat cultivars were subjected to RFLP analyses using 48 oat and barley (Hordeum vulgare L.) cDNA clones selected to represent a large portion of the genome. The 46 polymorphic clones detected a total of 278 fragments, 205 (73.7%) of which were polymorphic. The number of fragments detected per polymorphic clone varied from 2 to 11 with an average of 6.0 fragments per clone (4.5 polymorphic fragments per clone). The average number of phenotypes produced per clone was 8 (range of 1-25). All cultivars could be distinguished with a minimum of 11 clones. The RFLP banding patterns were used to generate a genetic distance matrix, which was used in both cluster and principal coordinate analyses. Both analyses clearly separated the spring and fall sown groups of cultivars, and further subgroupings were consistent with available pedigree information. The RFLP-based genetic distance matrix was compared with a matrix of coefficients of parentage for a subset of 55 cultivars. The Mantel statistic was highly significant (P < 0.01) but low (r = -0.32) indicating a moderate relationship between genetic distance derived from RFLP patterns and the pedigree-based coefficient-of-parentage. The RFLP patterns could be used to predict genetic relationships between germplasm of unknown or ambiguous pedigrees as well as for the fingerprinting of North American oat cultivars.