Interspecific competition might drive the evolution of ecological niches and result in pairs of formerly competing species segregating along ecological gradients following a process of character displacement. This mechanism has been proposed to account for replacement of related species along gradients of elevation in many areas of the world, but the fundamental issue of whether competition is responsible for the origin of elevational replacements has not been tested. To test hypotheses about the role of interspecific competition in the origin of complementary elevational ranges, I combined molecular phylogenetics, phylogeography, and population genetic analyses on Buarremon torquatus and B. brunneinucha (Aves, Emberizidae), whose patterns of elevational distribution suggest character displacement or ecological release. The hypothesis that elevational distributions in these species changed in opposite directions as a result of competition is untenable because: (1) a historical expansion of the range of B. brunneinucha into areas occupied by B. torquatus was not accompanied by a shift in the elevational range of the former species; (2) when B. brunneinucha colonized the range of B. torquatus, lineages of the latter distributions had already diverged; and (3) historical trends in effective population size do not suggest populations with elevational ranges abutting those of putative competitors have declined as would be expected if competition caused range contractions. However, owing to uncertainty in coalescent estimates of historical population sizes, the hypothesis that some populations of B. torquatus have declined cannot be confidently rejected, which suggests asymmetric character displacement might have occurred. I suggest that the main role of competition in elevational zonation may be to act as a sorting mechanism that allows the coexistence along mountain slopes only of ecologically similar species that differ in elevational distributions prior to attaining sympatry. The contrasting biogeographic histories of B. brunneinucha and B. torquatus illustrate how present-day ecological interactions can have recent origins, and highlights important challenges for testing the hypothesis of character displacement in the absence of data on population history and robust reconstructions of the evolution of traits and geographic ranges.