The emberizid genera Aimophila and Pipilo represent longstanding taxonomic conundrums. Each is comprised of subclades whose members appear to share diagnostic morphological and behavioral characters; however, relationships among sub-clades within each of these genera remain unclear, and numerous authors have suggested that either one or both of these genera may be polyphyletic. We addressed this taxonomic problem by sequencing and analyzing complete mitochondrial cytochrome-b and NADH dehydrogenase subunit 2 genes for all members of Aimophila and Pipilo along with 33 species representing 17 additional emberizid genera. Our maximum likelihood and Bayesian analyses indicate that both Aimophila and Pipilo are polyphyletic. Aimophila is divided into a minimum of three distinct groups. The forms notosticta, ruficeps, and rufescens are part of a well-supported clade that includes all members of Melozone and some members of Pipilo. Aimophila quinquestriata is placed within Amphispiza, and the remaining members of Aimophila are placed within a clade that includes all members of Arremonops and some members of Ammodramus. Within Pipilo, the ''rufous-sided'' and ''brown'' towhee groups do not form sister groups. Rather, the former are most closely related to the tropical genus Atlapetes whereas the latter are placed nearest Melozone and some Aimophila. Our analyses reject traditional taxonomic arrangements for both genera, and we present suggestions for a revised taxonomy for all members of Aimophila and Pipilo. These results provide further evidence of discordance among phylogenetic hypotheses based on morphological and molecular characters for groups of birds with generally conserved morphology.