Six subspecies of sandhill cranes (Grus canadensis) have been denoted based on perceived morphological and/or breeding locality differences among them. Three subspecies are migratory, breeding from the high arctic in North America and Siberia (lesser sandhill, G. c. canadensis), south through central Canada (Canadian sandhill, G. c. rowani) and into the northern United States (greater sandhill, G. c. tabida). A review of sandhill crane taxonomy indicates that the size variation, on the basis of which these subspecies were named, may be clinal and not diagnostic. The other three subspecies, all listed as endangered or threatened, are non-migratory, resident in Florida (G. c. pratensis), Mississippi (G. c. pulla), and Cuba (G. c. nesiotes). We used analysis of mitochondrial DNA control region (CR) sequences to determine whether haplotypes representing current subspecies show any genetic cohesion or are more consistent with a pattern of clinal variation in morphology. CR sequences indicate that only two highly divergent (5.3%) lineages of sandhill cranes occur in North America: one lineage composed only of arctic-nesting G. c. canadensis, the other of the remaining North American subspecies (we lack data on the Cuban population). The deep split between lineages is consistent with an estimated isolation of approximately 1.5 Mya (mid-Pleistocene), while the distribution of mutational changes within lineages is consistent with an hypothesis of rapid, post-Pleistocene population expansions. No other phylogeographic structuring is concordant with subspecific boundaries, however, analysis of molecular variance indicates that there is significant population genetic differentiation among all subspecies except G. c. tabida and G. c. rowani, which are indistinguishable. We suggest that recognition of the recently named G. c. rowani be abandoned.