Reithrodontomys megalotis and R. montanus, two species of harvest mouse that co-occur at low elevations of the southeastern Rockies and western Great Plains, are notoriously difficult to distinguish both externally and cranially, particularly for juveniles. Three external characteristics and 14 cranial measurements were used in discriminate function analyses (DFA) to detect the most robust measurements for species identification. Using 151 and 66 specimens of R. megalotis and R. montanus, respectively, from Colorado, we constructed DFA models for all specimens combined and segregated into four age classes by tooth wear. Due to substantial overlap in measurements, DFA models could not reliably predict species identity based on external or cranial characteristics alone, whereas models including all characteristics were more reliable. The most reliable DFA models were those for each age class (juvenile, subadult, adult, and old adult) using all external and cranial measurements. With various DFA models, 19 juvenile and subadult specimens of Reithrodontomys sp. from recent trapping efforts were all classified as R. megalotis with an average probability of 99.7%. Tail stripe width was also shown to be an unreliable identifier. We advocate using combined external and cranial measurements segregated by age class to robustly discriminate between these two species, particularly when identifying young individuals.