New evidence on the identity of the European Helorus species (Hymenoptera, Proctotrupoidea, Heloridae)

Abstract Background Species of Helorus Latreille 1802 are rarely collected endoparasitoids of Chrysopidae larvae (Neuroptera). Previous work on the limits between the European species of this species-poor genus, based on morphology only, has left some uncertainties. Here, we approach these cases and revisit previous taxonomic decisions using freshly collected and museum material. New information We generated the first large-scale Heloridae DNA barcode dataset, combined these with morphological data in an integrative taxonomic approach, and added information from studying all relevant type material. We found five species, Helorusanomalipes (Panzer, 1798), H.coruscus Haliday, 1857 stat. rev., H.nigripes Förster, 1856, H.ruficornis Förster, 1856, and H.striolatus Cameron, 1906, for which we provide an updated identification key. DNA barcode data are added to publicly available DNA barcode reference databases, for all species, except H.nigripes.


Introduction
The family Heloridae Förster, 1856 (Hymenoptera, Proctotrupoidea), includes one extant genus, Helorus Latreille, 1802, with currently 19 valid species worldwide (Zhang et al. 2020 ).They are endoparasitoids of the larvae of chrysopid lacewings (Neuroptera, Chrysopidae, Chrysopinae, Chrysopa species) (Van Achterberg 2006).The host is killed after the wasp larva spins a cocoon and the adult wasp emerges from the host cocoon after pupating within (Townes 1977).The family is widespread and occurs in all major zoogeographic regions, but seems to be most species-rich in the Palaearctic Region (Van Achterberg 2006, Zhang et al. 2020).
In this study, we aim to contribute to our understanding of species limits in the central European species of Helorus.To achieve this, we gathered molecular sequence data from the DNA barcode region (a portion of the 5' end of the mitochondrial cytochrome oxidase 1 gene), as well as morphological data and combined them in an integrative taxonomic approach, also considering all relevant type specimens.For DNA barcoding, we collected fresh material from different regions of Germany, complemented by material from Belgium, and processed it following the established workflow of GBOL/GBOL III Dark Taxa (e.g.Jafari et al. (2023)).For type studies, we located, loaned and imaged type specimens from different European museums, some of which had never been examined by the authors of the previous taxonomic treatments of the genus (Pschorn-Walcher 1955, Townes 1977, Van Achterberg 2006).
So far, four species were considered valid from central Europe: Helorus anomalipes (Panzer, 1798), H. ruficornis Förster, 1856, H. striolatus Cameron, 1906, and ( Townes 1977, Meyer 2001) which we specifically address in this study.First, H. ruficornis includes Helorus coruscus Haliday, 1857, which was synonymised by Townes (1977).The somewhat dubious statements of Townes (1977), his reference to the species treatment of Pschorn-Walcher (1955), as well as the assessment of Meyer (2001) suggesting two separate species, motivated us to find further evidence speaking either for or against the conspecificity of the two names.Second, H. striolatus includes Helorus meridionalis Pschorn-Walcher, 1955, which was also synonymised by Townes (1977).As in the first case, the treatment by Townes (1977) is not fully conclusive and the identities of these two names must be further scrutinised.Neither of these two taxonomic ambiguities were mentioned or addressed in the latest paper on European Helorus (Van Achterberg 2006).There are two additional documents (accessible, but not published in the sense of the International Code of Zoological Nomenclature) treating European Heloridae: Pripic-Schäper (2010), a privately published essay) and Mühlhäuser (2014), a BSc thesis at Lund University).Both indicate a separation of H. coruscus and H. ruficornis, but do not include formal taxonomic action.Neither provide any additional data on the status of H. striolatus and H. meridionalis.
The combination of the first available DNA barcode data from a considerable number of freshly-collected specimens with morphological data of voucher and type specimens allows for some new insights into the identity of the included species.New results and the DNA barcode data deposited in publicly available reference databases (bolgermany.de/ergebnisse/results; boldsystems.org)will make these few, but fascinating species accessible for future ecological and evolutionary studies, conservation, and wider recognition in science and public.

Specimens
We studied 49 ethanol-preserved specimens which were collected using Malaise traps, sweep netting and light traps at eight different sites in Belgium and Germany between June 2013 and August 2021.The specimens were databased with their metadata, identified to genus level, and preserved in 96% ethanol at 4°C or lower.Additional material came from museum collections, as specified in the respective material examined section.

Preparation, imaging and pterostigma index measurements
We point-mounted (after DNA extraction, see below) the specimens and tentatively identified and sexed them according to Van Achterberg (2006).We used a Keyence VHX-2000 digital microscope to produce multi-focus images that were stacked with the built-in programme.As reference for morphological terminology, we followed Van Achterberg (2006).We measured the pterostigma indices (pterostigma length:width) from images as shown in Fig. 1 from all species, except H. anomalipes.In addition to the pterostigma index, we measured the petiole index (petiole length:width in dorsal view) with an eyepiece micrometer mounted on a Leica M205 C stereo microscope and the flagellomere 1 and 2 indices of selected specimens.A table with all measurements is attached in Suppl.material 1.
In all morphological descriptions, we indicate values of the primary types in parentheses in the respective species treatments.

Molecular analysis
We extracted DNA from ethanol-preserved specimens at the Center for Molecular Biodiversity Research (ZMB) at the ZFMK in Bonn by following the procedures and protocols as described in Jafari et al. (2023).We amplified the mtDNA barcode region of the CO1 gene using the primers shown in Table 1.The sequencing was performed at BGI (Hong Kong).The barcodes fulfil all necessary criteria (in short: 1.Sequence length >= 500 bp, 2. bin quality = high, 3. ≤ 1% ambiguities/disagreements between the forward and reverse reads) to meet the defined GBOL gold standard (see Jafari et al. (2023) for more details).We added some outgroup sequences downloaded from BOLD (Cratomus sp.BOLD-ID: OPPEG1764-17, Exallonyx sp.BOLD-ID: BCHYM4222-14).We aligned the sequences using the built in MUSCLE alignment algorithm with a maximum of eight iterations (Edgar 2004) in GENEIOUS PRIME v.2022.1.1 (Biomatters Ltd.).The dataset can be accessed via bolgermany.de(German Barcode of Life Consortium et al. 2011) and on boldsystems.org(Ratnasingham and Hebert 2007).We list all respective BOLD-IDs of all specimens used in Suppl.material 2. Using IQ-TREE v.2.2.2.6 (Minh et al. 2020), we reconstructed a maximum likelihood tree and used ultrafast bootstrap to calculate branch support values (Hoang et al. 2018) without further specifications.For automated species delimitation, we applied the ASAP algorithm Table 1.

Primer
Primers used for the amplification of an abbreviated barcode region for Helorus spp.(625bp).
For the molecular characterisation of species, we analysed the distance matrix (Suppl.material 3) from the alignment provided in GENEIOUS to extract maximum intraspecific distances and minimum interspecific distances, stating the name of the closest species in parentheses.We generated a consensus sequence by aligning the sequences of each species separately in Geneious.

Remarks
The species treatments of H. anomalipes in Townes (1977) and Van Achterberg ( 2006) are still complete and valid and are, therefore, not repeated here.All three examined and DNA barcoded specimens herein are morphologically very similar and key to H. anomalipes using Van Achterberg (2006).The comparatively robust and sub-basally swollen petiole in combination with the absence of pronounced coarse reticulation on head and mesosoma (cf.Helorus nigripes, below) is very distinctive (Fig. 3).In the molecular analysis, the sequences form a distinct cluster.In summary, this species had already been well described and diagnosed and results herein corroborate this.New evidence on the identity of the European Helorus species (Hymenoptera, ...  Meyer (2001) in his list of Helorus species from Germany also stated that he observed morphological differences between the two species in question and, ergo, listed them both, yet without formal taxonomic action.The genetic differences and the molecular species delimitation methods we applied result in ambiguity (Fig. 2).Only by a combination of the results from the analyses of DNA barcode data with the morphological examination, including the primary types of both H. coruscus and H. ruficornis, are we able to take taxonomic action and formally re-instate H. coruscus from synonymy.

Helorus coruscus
The two species are not easy to diagnose.The specimens examined here match the characters given by Pschorn-Walcher (1955) and Pschorn-Walcher (1971) only in part.Pschorn-Walcher (1955) does not provide a formal diagnosis of H. coruscus, but complements his additions to the description with characters to differentiate H. coruscus from H. ruficornis in his key to the species.Most importantly, he gives differences in the pterostigma index that we can corroborate.The pterostigma of H. coruscus is more robust, index 1.9-2.5 (holotype 2.1) (n = 22) (with only two specimens with a pterostigma index > 2.3) (Fig. 4D), compared to the more elongated pterostigma of H. ruficornis (index 2.5-2.9 (holotype 2.6) (n = 6) (Fig. 5 and Fig. 7C).
The shape is comparable with the drawings of Pschorn-Walcher (1955) 1955).After examination of fresh material and types, we find no species-specific differences in body colour.Meyer (1969) imaged male genitalia of both species; these are obviously not from the (female) holotypes and, in addition, do not seem to differ distinctly.Therefore, we refrained from examining male genitalia ourselves.The morphological diagnostic characters we found to differentiate between H. coruscus and other species of Helorus, including H. ruficornis, are summarised in the identification key below.New evidence on the identity of the European Helorus species (Hymenoptera, ...

Remarks
We could only study historical material of this reportedly rare species, loaned from the RMNH.The specimen shows the distinct coarse reticulation on head and mesosoma (Fig. 6B) that was already described by Förster and also mentioned, described, and imaged by subsequent authors (Pschorn-Walcher 1955, Townes 1977, Van Achterberg 2006).Note that Pschorn-Walcher (1955)   primary type of H. nigripes, we agree with Townes (1977) that the statement describing the unique coarse sculpture by Förster leaves no doubt about the identity of the name.The absence of the primary type of H. nigripes is irrelevant.Therefore, H. nigripes is used here, maintaining the principle of priority.(1977).

Helorus ruficornis
The four specimens of H. ruficornis examined herein are very similar to the lectotype of H. ruficornis from NHMW.Note that the lectotype designation was done by Pschorn-Walcher (1955) by stating that it is a "type", which -according to the ICZN for publications prior to 1999 -is sufficient for a valid lectotype designation.We examined a second specimen, collected on 11/09/1860, which was examined and listed as a nontype specimen by Pschorn-Walcher (1955) though it bears a hand-written label by Förster and a label "ruficornis Förster, type".It is very similar to the H. ruficornis lectotype; however, the fore wings are damaged and we cannot measure its pterostigma index value.Buffington and Copeland (2016) gave some characters to differentiate H. ruficornis from H. striolatus (including pterostigma shape) and differentiation between these species is relatively easy (see also Van Achterberg (2006); and see below for remarks on the identity of H. striolatus and H. meridionalis).However, the species from which H. ruficornis is most difficult to distinguish is H. coruscus.For a discussion on the characters separating both and results from analysis of CO1 barcode data, see Fig. 2 and the treatment of H. coruscus above.Note that Buffington and Copeland (2016) did not contribute to the discussion of possible synonymy of H. ruficornis and H. coruscus.They also did not examine the type of H. ruficornis.Additionally, their values given for the petiole index (i.e.petiole length:width) of H. ruficornis (i.e."4x longer than wide" in the diagnosis and "> 5-6x longer than broad" in the re-description) are not in line with the values of the H. ruficornis specimens examined here, including the holotype (2.7-3.3 (holotype 2.7)).This might indicate that the central European H. ruficornis, including the primary type, are different from the Afrotropical material examined by Buffington and Copeland (2016).2015) described an additional species that is most similar to H. striolatus from Iran.We did not examine this species and type.Currently, arguments for separation between H. striolatus and H. meridionalis and for re-instating H. meridionalis are weak and we keep the synonymisation and place all our specimens under the name H. striolatus.Please note that the pin holding the holotype female of H. meridionalis bears a second (male) specimen (Fig. 8E).The male allotype mentioned by Pschorn-Walcher ( 1955) is on a separate pin and in a very poor condition (i.e.missing head, hind wings, one fore wing, and metasoma).The male specimen that is with the holotype is indeed a H. anomalipes specimen.It is unknown to us who put this specimen next to the H. meridionalis holotype.New evidence on the identity of the European Helorus species (Hymenoptera, ...

Analysis
All European species of the genus Helorus are represented in the studied material.No fresh material of H. nigripes was available to us for sequencing and the species is, hence, not included in the DNA barcode data and analyses.In total, 49 DNA barcode sequences were obtained and 67 specimens were examined morphologically.
The analyses of DNA barcode data imply the presence of three to four species in our material.The ASAP analysis resulted in two partitions with an identical score (1.5).Both partitions are implemented in Fig. 2 and ambigously labelled ASAP 1 and ASAP 2 .ASAP 1 infers four, ASAP 2 three putative species.The mPTP analysis is in line with the results of ASAP 2 .We ultimately based our species delimitations for the species treatments on an integration of both automated molecular species delimitation and morphological examination.We recognise four species within the DNA barcoded material (Fig. 2).
The four clusters/species differ from each other in their DNA barcode sequence.Helorus coruscus and H. ruficornis differ distinctly from the other two clusters/species, by a minimum of 12.5%.Helorus anomalipes and H. striolatus are clearly separated from each other by a minimum 12.3% difference.All 26 specimens of the H. striolatus cluster are genetically identical, although from two different locations.The difference between H. coruscus and H. ruficornis is smaller than the other interspecific differences in the dataset, with a minimum of 1.8% and a maximum of 2.4%.
For more details on the species delimitation, also including the fifth species H. nigripes, see the respective treatments and the identification key.

Discussion
Helorus is a species-poor genus, but even these can harbour persistent taxonomic problems.Matters are further complicated by the comparative rarity of all included species which leads to small series available for taxonomic studies, both in fresh collections as well as in museum collections, even for the central European region.The addition of the first large-scale DNA barcode sequence data and the examination of the type specimens was key to our contribution and we are confident that we were able to add some clarity and enable future studies on this group of wasps.In fact, helorids have been studied neither in an evolutionary, nor an ecological or conservation context.So far, they have shared the fate of many "Dark Taxa" (sensu Hausmann et al. ( 2020)) because they were simply not accessible or "on the map".We do not add any information on life history, most notably host use, so that a lot of fundamental information about the species is still missing (see Pschorn-Walcher (1955), Pschorn-Walcher (1971) and Townes (1977) for the current knowledge on host associations).We expect this study to be part of the first step towards a better understanding of these species.In this context, we also tried to make the key included here as simple as possible, in order to have users from ecology, citizen science or conservation feel invited and not repelled by excessive lists of characters that are difficult to assess by the untrained eye.
We are grateful to Theo Peeters and the curational staff at the RMNH for organising the gracious loan of the Helorus specimens that were included in the work of Van Achterberg (2006).
We are indebted to Tanja Schweizer, Dominique Zimmermann, Joseph Monks, Stefan Schmidt, Lars Ove Hansen, and Aidan O'Hanlon for making the types and other specimens accessible to us via loan or imaging.Especially Aidan O'Hanlon who provided invaluable help to locate the type specimen of H. coruscus.

Figure 1 .
Figure 1.Pterostigma of fore wing of a Helorus specimen.The yellow lines indicate the distances we used to measure the pterostigma index (pterostigma length:width).Scale bar 200 µm.

Figure 2 .
Figure 2. Maximum likelihood tree inferred with IQTree and the results of the molecular species delimitations.The first two ranks of ASAP have an identical score.The box colours show the results of the species delimitations that are either congruent (black) or incongruent (red) with our morphological examinations.The dotted lines connecting the outgroup are not to scale.Ultrafast-bootstrap support is shown on the branches.

Figure 5 .
Figure 5.Pterostigma index values in H. coruscus stat.rev.and H. ruficornis.The values of the primary types are indicated with dots next to the boxes.Graphs were created by using R, while the data points of the types were added using Inkscape.
wrote that the name H. rugosus Thomson should be preferred because Thomson's material was, in contrast to Förster's, still present.Pschorn-Walcher (1971) continued to use the name H. rugosus.Pschorn-Walcher (1955) designated a lectotype for H. rugosus, deposited at Stockholm Museum (SMNH).It is unclear ifTownes (1977) was aware of this designation and was looking for the same specimens Pschorn-Walcher (1955) examined when he briefly reports the type as "… lost, not found in Stockholm 1975".Despite the allegedly lost

Cameron, 1906
Pschorn-Walcher (1955)all other species by the combination of a slender petiole, a long pterostigma, and darker legs (seeVan Achterberg (2006)and the key below).In our analyses of molecular sequence data, specimens with these characters clearly form a separate species.In addition to our material, we examined the types of H. striolatus and H. meridionalis and found all specimens to be very similar.Helorus meridionalis is a synonym of H. striolatus which has been synonymised byTownes (1977), again without examination of the type (of H. meridionalis; cf. the case of H. coruscus above).Synonymisation was solely based on re-evaluation of the description byPschorn-Walcher (1955).Pschorn-Walcher (1955)did not include H. striolatus in his revision, because he had no material of this species at hand and did not examine the type.Therefore, he does not give any diagnostic characters to separate his new H. meridionalis from H. striolatus.Later,Townes (1977)examined the type of H. striolatus and found the description of H. meridionalis byPschorn-Walcher (1955)sufficient to synonymise the two.Prpic-Schäper ((2011), privately published essay) outlines some doubts on the validity of the synonymisation of the two species, based on differences in the first two flagellomere indices and the differences in type locality.The type of H. striolatus was collected in Pakistan, i.e. geographically separated from, for example, the H. meridionalis type from Italy or the specimens examined herein from Germany.Furthermore, the types of H. striolatus and H. meridionalis are both lacking their heads, which makes an examination of the flagellomere lengths impossible.The colouration of the legs is lighter in the H. striolatus type than in the H. meridionalis type, but these are only subtle colouration differences that are hardly useful for species delimitation.Van Achterberg (2006) also notes differences in H. striolatus specimens from Spain (quote "The specimens from Spain have the vertex strongly punctate and the pterostigma slightly more robust than other specimens") which could point to the existence of a separate species, maybe H. meridionalis, in southern Europe.As the primary types of both species have lost their heads, the first character cannot be evaluated further.The pterostigma is, as Van Achterberg (2006) described, slightly more robust in the H. meridionalis type than in the H. striolatus (pterostigma index 2.9 meridionalis type, 3.4 striolatus type).While the pterostigma index value of the H. meridionalis lectotype falls into the range that we find in our specimens of H. striolatus (2.8-3.3), the holotype of H. striolatus exhibits an extreme value of 3.4.From his note, that, specifically, specimens from Spain were different from other examined material (see above), we can conclude that the ones from Turkey did not differ from his central European material.The analysis of DNA barcode data herein is of limited significance because all included specimens originated from Germany; they currently do not point towards two species under the name H. striolatus.To investigate this further, more specimens, ideally with both morphological and molecular data, are needed from southern Europe, as well as from Pakistan and neighbouring countries and the Middle East.Izadizadeh et al. (