It is widely believed that habitat fragmentation during the Pleistocene initiated speciation events in many songbird genera. One such vicariance model for avian speciation in the Pleistocene was developed by R. M. Mengel for North American birds. This model suggests that the first Pleistocene glacial advance reduced the area of an extensive, eastern North American deciduous forest, forcing adaptation by some species to boreal forest. This, in turn, facilitated the development of transcontinental range expansions during interglacials. Subsequent glacial advances repeatedly fragmented the ranges of these species into eastern and western populations; western isolates speciated to form the multispecies groups observed among various North American birds. We used mtDNA restriction site data to reconstruct the phylogeny of the black-throated green warbler complex-the group that Mengel considered the best fit to his model. Contrary to Mengel's model, the phylogeny indicates that not all western endemics were derived from an eastern ancestor. Instead, our results imply a mix, wherein some western endemics were budded off an eastern source, as Mengel posits, while others probably resulted from intermontane isolations in the west.