The plant genus Escallonia is widely distributed in the Neotropical mountains, from Costa Rica to Southern Chile and from sea level to the paramos. This broad geographic distribution provides an excellent opportunity to assess the influence of geographical and ecological isolation in the diversification of montane plant taxa. Biogeographic analyses revealed a deep signal of phylogenetic geographic structure indicating that geographic isolation was a major factor during the early evolutionary history of Escallonia. This pattern is consistent with the hypothesis of allopatric speciation plausibly induced by ancient niche conservatism. Although “geographic clades” later diversified in isolation, there is not strong support for the hypothesis of ecological speciation induced by niche evolution across environmental gradients along elevation. Alternatively, these findings might suggest that species limits as proposed in the current taxonomy are problematic. Combining genomic and phenomic analyses with state of the art species discovery approaches, I show that taxonomic species in Escallonia correspond poorly with the pattern of variation shown by individuals in nature as predicted by species descriptions. Moreover, I show that genomic and phenomic data rarely delimit congruent entitles suggesting that the assumption that the species delimited by traditional taxonomists are evolutionary and ecologically meaningful entities is questionable. It is plausible that the patterns recovered in Escallonia are rampant across many groups. If so, relying on taxonomic species without empirical evidence of their biological reality can have relevant implications in ecology, evolution, and conservation.