Is There a Barcoding Gap in Arsenura Duncan? Implications for Saturniid Systematics and DNA Barcoding

Résumé

The hope that DNA barcoding would accelerate the documentation of biodiversity has materialized with mixed implications for the discipline of taxonomy. Of concern is the conflation in the taxonomic literature of two of the main goals of DNA barcoding: species discovery and diagnosis. Recent taxonomic work in the saturniid genus Arsenura, offers an example of the potential for overreliance on the barcode, and the resulting proliferation of species names without assessing the reliability of COI for either species discovery or diagnosis. Because the use of barcodes in taxonomy is likely to grow, we explored its performance as a diagnostic tool in Arsenura based on a dataset of 468 individual barcodes, specifically by estimating genetic distances within and between species to assess the existence of a barcoding gap. Parsimony-generated trees—interpreted as summaries of character state distributions—of 170 unique barcode haplotypes were compared to BIN clusters generated with neighbor joining in BOLD. We found a considerable overlap between intraspecific and interspecific genetic distances. The latter were as low as 0.8%, while more than 33% of intraspecific comparisons showed greater than 2% divergence, and 21% exceeded 3%, together demonstrating the lack of a barcoding gap in the genus, the criterion for robust species recognition used by BOLD. Although in some instances NJ analysis oversplits relative to parsimony, parsimony analyses of barcodes allow us to determine that 92% of BOLD's BINs are supported by nucleotide character states, and that conspecific barcodes are successfully recognized as such. Two exceptions involved samples of A. jennettae and A. aspasia, in which barcodes lacked diagnostic characters and thus failed to distinguish them from A. armida and A. biundulata, respectively. Given the lack of a barcoding gap and the somewhat inconsistent performance of neighbor-joining BINs in reflecting diagnosable entities, we caution against over-reliance on genetic distances as arbiters of the taxonomic choices. Barcodes should ideally be corroborated by morphological, behavioral, or geographical differences. Finally, we call attention to the need for adequate sampling to ensure the broadest possible representation of intraspecific diversity in analyses and stress the importance of transparent, empirically sound interpretations of character data.

Auteurs, date et publication :

Auteurs Andrea C. Jiménez-Bolívar , Ivonne J. Garzón-Orduña

Publication : The Journal of the Lepidopterists' Society

Date : 2025

Volume : 79

Issue : 4

Pages : 215-237