Establishing links between phenotypic and genotypic variation is a central goal of evolutionary biology, as they might provide important insights into evolutionary processes shaping genetic and species diversity in nature. One of the more intriguing possibilities is when no genetic divergence is found to be associated with conspicuous phenotypic divergence. In that case, speciation theory predicts that phenotypic divergence may still occur in the presence of significant gene flow—thereby resulting in little genomic divergence—when genetic loci underpinning phenotypes are under strong divergent selection. However, a finding of phenotypic distinctiveness with weak or no population genetic structure may simply result from low statistical power to detect shallow genetic divergences when small datasets are used. Here, we used a subgenomic dataset of 2386 ultraconserved elements to explore genome-wide divergence between two species of Antilophia manakins, which are phenotypically distinct yet evidently lack strong genetic differentiation according to previous studies based on a limited number of loci. Our results revealed clear population structure that matches the two phenotypes, supporting the idea that smaller datasets lacked the power to detect this recent divergence event (likely < 100 k ya). Indeed, we found little or no introgression between the species, as well as evidence of genome-wide divergence. One implication of our study is that the Araripe plateau may be a hotspot of cryptic-diverging forest Cerrado populations. Besides their use in biogeography, subgenomic datasets may help redefine local conservation programs by revealing cryptic population structure that may be key to population management.