Mind the gap! Integrating taxonomic approaches to assess ant diversity at the southern extreme of the Atlantic Forest

Abstract

Understanding patterns of species diversity relies on accurate taxonomy which can only be achieved by long‐term natural history research and the use of complementary information to establish species boundaries among cryptic taxa. We used DNA barcoding to characterize the ant diversity of Iguazú National Park (INP), a protected area of the Upper Paraná Atlantic Forest ecoregion, located at the southernmost extent of this forest. We assessed ant diversity using both cytochrome c oxidase subunit 1 (COI) sequences and traditional morphological approaches, and compared the results of these two methods. We successfully obtained COI sequences for 312 specimens belonging to 124 species, providing a DNA barcode reference library for nearly 50% of the currently known ant fauna of INP. Our results support a clear barcode gap for all but two species, with a mean intraspecific divergence of 0.72%, and an average congeneric distance of 17.25%. Congruently, the library assembled here was useful for the discrimination of the ants of INP and allowed us to link unidentified males and queens to their worker castes. To detect overlooked diversity, we classified the DNA barcodes into Molecular Operational Taxonomic Units (MOTUs) using three different clustering algorithms, and compared their number and composition to that of reference species identified based on morphology. The MOTU count was always higher than that of reference species regardless of the method, suggesting that the diversity of ants at INP could be between 6% and 10% higher than currently recognized. Lastly, our survey contributed with 78 new barcode clusters to the global DNA barcode reference library, and added 36 new records of ant species for the INP, being 23 of them new citations for Argentina.

Ants in Australia’s Monsoonal Tropics: CO1 Barcoding Reveals Extensive Unrecognised Diversity

“…By way of comparison, a recent CO1 analysis of the ant fauna of Brazil's southern Atlantic Forest indicated that species richness was only 10% higher than previously appreciated [85], rather than the several-fold increase that is apparent for the AMT. Our findings call for a re-evaluation of the recent assessment based on the TERC collection that 900 species are known from the Top End of the Northern Territory [40].…”

“…All four CO1 trees showed moderate (70)(71)(72)(73)(74)(75)(76)(77)(78)(79)(80)(81)(82)(83)(84)(85)(86)(87)(88)(89) to high (≥90) Bayesian posterior probability at the tips, while basal node support was generally lower and poorly resolved in all except Camponotus. Some poorly resolved clades were nested among other better supported clades (e.g., Cardiocondyla and Camponotus), and were supported by differences in locality and morphology.…”

Intra‐ and interspecific variation in trophic ecology of ‘predatory’ ants in the subfamily Ponerinae

“…Twenty‐two ponerine species were included in the trophic position analysis, representing c . 60% of their diversity for the region (Hanisch et al ., , ). As predicted, most ponerines were estimated to be predatory, with δ 15 N values placing them at TL3 or TL4.…”

“…Ants were identified to species using available keys (Brown, ; Wild, ; Jiménez et al ., ; MacKay & Mackay, ; Dash, ; Lenhart et al ., ; Fernandes et al ., ), and other taxa to the lowest taxonomic level possible. To improve taxonomic information for all ponerines and select insects, we sequenced a 658‐bp fragment near the 5′ end of the COI gene following standard protocols developed for DNA barcoding (see Hanisch et al ., ). All trace files, collection data, taxonomic information, and images were uploaded to the online Barcode of Life Data Systems (BOLD, http://www.boldsystems.org; Ratnasingham & Hebert, ) where they are publicly available in the dataset DS‐PONERECO.…”

“…In short, each sequence uploaded to BOLD is automatically assigned a BIN based on a specific algorithm (refined single linkage, or RESL), or alternatively a new BIN is created if the specimen's sequence is different enough from all BINs in the database (Ratnasingham & Hebert, ). Different BINs within a single species could indicate the presence of cryptic species and are particularly useful for understudied groups or biodiversity assessments (Ratnasingham & Hebert, ; Hanisch et al ., ). We then generated a neighbour‐joining tree with software mega7 using Kimura two‐parameter substitution model and pairwise deletion.…”