When 16S fails: assessing mitochondrial markers for species delimitation and identification in Pithecopus (Anura: Hylidae: Phyllomedusinae)

Abstract

Reliable species delimitation remains one of the most critical and persistent challenges in amphibian taxonomy, especially for groups with recent diversification, cryptic lineages, and low levels of morphological differentiation. The Neotropical genus Pithecopus represents a prominent example of these issues, with a history of taxonomic instability and unresolved species boundaries across multiple lineages. In this study, we conducted a comprehensive evaluation of three widely used mitochondrial markers (16S rDNA [16S], cytochrome b [Cyt-b] and cytochrome c oxidase subunit 1 [CO1]). These markers were analyzed in combination with single-locus species delimitation methods (General Mixed Yule-Coalescent, Multi-rate Poisson Tree Processes [mPTP], and Assemble Species by Automatic Partitioning) to assess their relative effectiveness in delineating species boundaries in addition to generating objective, replicable molecular diagnoses within Pithecopus. Our results reveal a substantial variation in performance among both markers and methods. CO1 consistently outperformed Cyt-b and 16S in delimitation accuracy, showing the highest congruence with current taxonomic hypotheses and delivering the most molecular diagnoses based on both robust diagnostic nucleotide combinations and minimum diagnostic characters. Cyt-b showed moderate performance, especially when paired with mPTP, while short 16S fragments (∼500 base pairs)—still commonly used in amphibian studies—performed poorly across all analytical approaches, generating high error rates, limited phylogenetic resolution, and unreliable diagnostic characters. The variability in delimitation outcomes across methods further underscores the importance of adopting discovery approaches with lower error rates, especially given that errors introduced at this initial step tend to propagate during validation phases. Our findings advocate for the use of consensus among multiple delimitation methods as a starting point for taxonomic hypothesis testing, particularly in groups such as Pithecopus, where population structure and speciation signals may be confounded. Most importantly, this study emphasizes the need for taxonomists to integrate explicit, character-based molecular diagnoses—rather than genetic distances—into species descriptions and taxonomic workflows. Doing so enhances replicability, facilitates specimen identification in future surveys, and provides a foundation for ongoing taxonomic refinement. The molecular diagnoses generated here serve as a reference framework for future research on Pithecopus and may inform best practices for molecular taxonomy in other Neotropical amphibians. Until genomic-scale data become more broadly accessible, we recommend the preferential use of CO1 for species discovery and identification in Pithecopus, coupled with carefully selected delimitation methods that balance error rates and diagnostic power.