Institutional abbreviations

• EVK – Entomologischer Verein Krefeld e.V., Krefeld, Germany;
• NINA - Norwegian Institute for Nature Research, Trondheim, Norway;
• NTNU - Norwegian University of Science and Technology, University Museum, Department of Natural History, Trondheim, Norway;
• UB - University of Barcelona;
• ZFMK - Leibniz Institute for the Analysis of Biodiversity Change, Museum Koenig Bonn, Germany;
• ZSM - Zoologische Staatssammlung München, Germany.

All specimens with the ‘ZFMK-TIS-...’ tag were prepared at the molecular laboratory of the ZFMK in Bonn in the course of the GBOL III: Dark Taxa project, following the procedures as described in Jafari et al. (2023), using LCO1490-JJ as forward and HCO2198-JJ as reverse primer (Astrin and Stüben 2008). Sequencing was done at BGI (Hong Kong, China). We first assembled forward- and reverse-reads to contig files to then infer consensus sequences using Geneious Prime 2022.1.1 (Biomatters Ltd.). The DNA of specimen ‘HM128-04-CC’ was extracted at the ZFMK using HotSHOT extractions (Truett et al. 2000) by adding 25 µl alkaline lysis buffer and the full body specimen as tissue. We incubated the specimen at 70°C for 30 minutes. We added 25 µl of neutralising solution and used 1 µl of the extract in the PCR. The sample was further processed along with others not used here, using LCO1490-JJ and HCO2198-JJ primers (Astrin and Stüben 2008) with individual tags attached. We sequenced the amplicon pool using MinION technology as described in Vasilita et al. (2024). In addition to the sequences produced in GBOL III: Dark Taxa, we downloaded all Phaenoglyphis CO1 sequences of European origin and three Alloxysta sequences as outgroup sequences from BOLD (Ratnasingham and Hebert 2007, accessed on 7 Sep 2023). We aligned all sequences using Clustal Omega 1.2.2 (Sievers et al. 2011). All CO1 barcode sequences generated herein and the downloaded sequences with corresponding BOLD-ID are listed in Suppl. material 1.

Using IQ-Tree v.2.2.2.6, we reconstructed a Maximum Likelihood tree by applying the -s option and ultrafast bootstrap with 1000 replicates (Hoang et al. 2018) and without further modifications of parameters (Minh et al. 2020). We performed three molecular species delimitation algorithms on the dataset. Firstly, we applied the ASAP species delimitation algorithm on our alignment (via https://bioinfo.mnhn.fr/abi/public/asap/ accessed 16 Jan 2024, Puillandre et al. (2021)) using default settings. Secondly, we used the Cluster function of Species Identifier v.1.6.2 (Meier et al. 2006) on our alignment, set to a 3% threshold. We chose a 3% threshold as was proposed by Hebert et al. (2003) and supported as objective clustering in Hartop et al. (2022). Lastly, we performed multirate PTP (via https://mptp.h-its.org/#/tree accessed 16 Jan 2024, Kapli et al. (2017)), using default settings on the tree file previously constructed with IQ-Tree. For the tree figure, we combined the final tree and the species delimitation results using InkScape v.1.2 (Inkscape Project).

For the molecular characterisation of species, we used the distance matrix from the alignment provided in Geneious (Suppl. material 2) to extract the maximum intraspecific barcode-distance, stating the sample size in parentheses, and the minimum intraspecific distance, stating the name of the closest species in parentheses. We generated a consensus sequence by aligning the sequences of each species separately in Geneious. As the molecular characterisation is part of an integrative taxonomy approach, we only used the sequences of those specimens that we studied morphologically.

All specimens were left externally intact during DNA extraction and each was point-mounted post lysis using Shellac glue. We attempted to spread the legs and wings so that no other body part was obscured. The metasoma was aligned with the body axis. All specimens are deposited at ZFMK unless stated otherwise.

All specimens were examined using a Leica M205C stereomicroscope (JV & RSP) and Optika ZSM-2 (MFS). Morphological terms and abbreviations are taken from Paretas-Martínez et al. (2007). The width of the fore wing radial cell is measured from the margin of the wing to the beginning Rs vein. Females and males share the same character states, except where indicated.

Out of 55 specimens processed, 48 barcode sequences were generated (87% success rate). Including the publicly available barcodes from BOLD, the final dataset consists of 101 ingroup and three outgroup sequences.

The species limits established by morphological features are corroborated by the molecular results (Fig. 2). Seven Phaenoglyphis species have been identified, which is supported by the molecular species delimitations of Species Identifier, mPTP and largely conclusive with the first-ranked partition of ASAP (except Phaenoglyphis longicornis and P. salicis being clustered as conspecific). The second-ranked partition of the ASAP analysis split the included specimens into 13 separate species. The first and second-ranked partition of ASAP had an identical ASAP-score (4.0) and were, therefore, included both in Fig. 2.

We complement the previously-established morphological characterisation of the genus Phaenoglyphis and seven of its species from north-western Europe (Ferrer-Suay et al. 2018) with the first molecular characterisation of their respective DNA barcode sequences: P. belizini, P. evenhuisi, P. longicornis, P. salicis, P. stricta, P. villosa, and P. xanthochroa. This currently leaves ten species known from north-western Europe without molecular characterisation (P. abbreviata, P. americana, P. calverti, P. fuscicornis, P. gutierrezi, P. heterocera, P. nigripes, P. proximus, P. pubicollis and P. ruficornis).

We complement BOLD with additional sequences for all of the molecularly characterised taxa, except P. stricta, for which we did not provide any additional sequences. Four of the taxa within our material were not represented on BOLD before (P. belizini, P. evenhuisi, P. longicornis, and P. salicis).

Our discovery of P. belizini and P. evenhuisi in our material represents new records for Germany.

The results of the mPTP and SpID species delimitations were largely congruent with our morphological identifications. There is an apparent over-splitting by the second-ranked ASAP partition and one case of lumping in the first-ranked ASAP partition. This could be interpreted as additional evidence that it is advisable to use more than one species delimitation algorithm and is in line with previous findings (e.g. Meier et al. (2022)).

Ten species from north-western Europe are currently lacking CO1 barcode sequences and these can hopefully be added in future investigations. It is important to note that, within Charipinae, Alloxysta specimens are the most common and they are very well represented in many samples, but Phaenoglyphis is comparably rare, which makes it more difficult to acquire a good number of fresh specimens for sequencing.

The 6 bp deletion, present exclusively in the Phaenoglyphis barcodes within Charipinae, is additional evidence for the monophyly of the genus that was previously questioned (Paretas-Martínez et al. 2007). More extensive analyses, ideally based on a phylogenomics/taxonomics dataset, are needed to answer this question and steer the classification of monophyletic Charipinae genera.

Molecular characterisation of Charipinae species is still at its first steps. With this study, a significant portion of one of the main genera, Phaenoglyphis, is now ready to be included in DNA barcode-based activities. However, many species remain uncharacterised and it will be necessary to continue integrative taxonomy studies and to improve our knowledge of the genus.