AbstractIn the present study, we compare the results of alpha taxonomy (based on morphology), DNA Barcoding method with a 2% genetic divergence threshold, and the GMYC (General Model Yule Coalescent) technique to identify species clusters in Neoplecostomus. We used partial sequences of mitochondrial COI (Cytochrome oxidase subunit I) for 59 specimens representing 13 valid species (Neoplecostomus bandeirante, N. jaguari, N. langeanii, N. paranensis, N. yapo, N. botucatu, N. selenae, N. doceensis, N. corumba, N. ribeirensis, N. microps, N. espiritosantensis and N. franciscoensis) of Neoplecostomus collected in all its distribution area. For the analysis we used Bayesian inference of phylogeny with relaxed clock methods on an arbitrary timescale, using BEAST. The ultrametrics genes trees obtained for each tree prior evaluated (Yule, Birth-death and coalescent Population Size) were used in the GMYC analysis to identify a time in the tree when the branching rate shifts (in forward time) from Yule and Birth-Death (species) to a coalescent (population) process. We found that the GMYC model using the Yule prior identified 11 groups, closer to the current taxonomy (13 species). GMYC analyses using other ultrametric gene trees estimated under alternative prior such as Birth-Death and Coalescent Population Size, identified 9 groups, results observed in the traditional 2% genetic distance threshold, resulting in a low number of species recognized compared to the number of species identified with current taxonomy. Based on these results we conclude that the molecular approaches are helpful to distinguish species of Neoplecostomus, nevertheless it is important to combine molecular methodologies with current taxonomy in order to identify correctly species that recently originated.