Hybridization and/or incomplete sorting of ancestral polymorphism are commonly implicated to explain discordant phylogenetic analyses of closely related species complexes. One genus in which these phenomena have been suggested to have played major roles based on phylogenetic data is Conradina, a genus of mints (Lamiaceae) endemic to the southeastern USA containing several endangered species. The goals of this study were to use microsatellite data to better understand patterns of genetic structure in Conradina, to test hypotheses of recent or ancient hybridization and incomplete lineage sorting, and to clarify species boundaries. Individuals from 55 populations representing all Conradina species were genotyped using 10 microsatellite loci. Analyses of the patterns of genetic structure in Conradina revealed a clear differentiation of populations following recognized species boundaries, indicating that species have diverged from one another genetically and interspecific hybridization has not occurred recently. Neither ancient hybridization nor incomplete lineage sorting is supported as the sole cause of species nonmonophyly, suggesting that both may have contributed to patterns found in phylogenetic trees; however, analyses of other types of data may be more appropriate to distinguish between these two hypotheses. Because all described species appear to be valid entities, the current listing status of most endangered species of Conradina is appropriate; however, populations of Conradina canescens are genetically differentiated into three groups, each of which may merit species status, and several recently discovered populations of Conradina in Dunn's Creek State Park in Florida are highly differentiated genetically and also appear to represent a new species.