Cladistic analysis of Zethus Fabricius, 1804 (Hymenoptera, Vespidae): a new subgeneric classification

Zethus is the largest genus in Vespidae with over 270 species. It is currently divided into four subgenera: Z. (Zethus), Z. (Zethoides), Z. (Zethusculus) and Z. (Madecazethus). While the last three are restricted to certain biogeographic areas, the first is spread through western and eastern hemispheres. Studies have shown possible phylogenetic incongruence regarding this current division and even raised the possibility of paraphyly in the genus. To evaluate this classification, morphological pasimony analyses under implied weights was carried out, examining external morphology and male genitalia. Analyses showed paraphyly of the genus under various “k” values and paraphyly of two subgenera. Ischnocoelia and Ctenochilus are lowered to subgenera of Zethus. Zethus (Madecazethus) is no longer restricted to Malagasy species and now includes those of the African continent as well. Zethus (Zethastrum) subg. nov. is defined for Oriental representatives. Z. (Zethus) is subdivided into three subgenera: Z. (Zethus), Z. (Didymogastra) and Z. (Wettsteinia). Zethus (Zethusculus) and Z. (Zethoides) are monophyletic and hold their status as subgenera. Three species-groups are Incertae sedis: Z. albopictus, Z. pallidus and Z. spinosus.


Introduction
Zethus is the most species diverse genus in Vespidae and represents 279 out of the 363 species of Zethini, a group of Eumeninae whose status as tribe or subfamily is still in debate (Carpenter 1982;Pickett and Carpenter 2010;Hines et al. 2007;Hermes et al. 2014;Bank et al. 2017;Piekarski et al. 2018). The genus' distribution is mainly Gondwanian excluding the Australian region (Stange 1979). Still, a few species may reach into the Nearctic (Bohart and Stange 1965) or into Palearctic region in areas bordering the Afrotropical (Giordani Soika 1979) or the Oriental Regions (Tan et al. 2018).
The genus is currently divided into four subgenera: Z. (Zethus) Fabricius, 1804; Z. (Zethusculus) de Saussure, 1855; Z. (Zethoides) Fox, 1899;and Z. (Madecazethus) Giordani Soika, 1979. While the last subgenus is restricted to Madagascar and the second and third to the Neotropics and parts of the Nearctic, the nominotypical subgenus has species in the New World as well as the Old.
The problematic history of subgeneric divisions in Zethus, from de Saussure (1852,1855,1875) to Bohart and Stange (1965), has been summarized by Lopes and Noll (2018), whose study revealed possible incongruences in the current classification, where some subgenera, and even the genus Zethus itself, may be unnatural groups as they are currently construed. Their study, although focused on Z. (Zethoides), which was retrieved as monophyletic, revealed the following features that discord with the current taxonomy: Zethus is paraphyletic in relation to Ctenochilus; Z. (Zethusculus) is paraphyletic in relation to the remaining Zethus (contradicted by Golfetti et al. (2020), who recovered the monophyly of this subgenus); and Z. (Zethus) is paraphyletic in relation to Z. (Madecazethus) and Z. (Zethoides). Although obtaining these results, the authors refrained from making taxonomic changes since the taxon sampling in these groups was poor. Carpenter and Cumming (1985) had already noted not only the possibility of Ctenochilus being a derived lineage of Zethus, but also Ischnocoelia since these genera are diagnosed based mainly on characters of the palpi. The first genus presents modifications on the labial palpi, with reduced segmentation and dense long pilosity, while the second presents only the reduced segmentation, but of both maxillary and labial palpi. Further, Lopes and Noll (2018) speculated that Z. (Zethus) may actually be a number of different lineages that might deserve their own subgeneric status.
Thus, taking such inconsistencies into account, a phylogenetic analysis of Zethus was carried out, attempting to verify the validity of the genus as a natural group as well as of its subgenera and to propose taxonomic changes according to the results.

Taxon sampling
A total of 48 representatives of Zethus species were examined (Table 1). Since the genus has subdivisions of species-groups, which are comprised of morphologically similar species (Bohart and Stange 1965), the sampling targeted having one species of each of these groups in the case of the Neotropical representatives. Oriental and African Zethus were sampled as widely as possible and the Neotropical Z. discoelioides group had two species sampled due to its high diversity and large number of acknowledged undescribed species (Bohart and Stange 1965).
No representatives from the Neotropical Z. stangei and Z. magretti groups nor from the Oriental Z. trimaculatus group were examined.
Additionally, 24 terminals were used as outgroups representing twelve other Zethini genera. Ischnocoelia and Ctenochilus were sampled as widely as possible, the former by nine species and the later by all five species, while the remaining genera were represented by a single species. The two genera were better represented due to the possibility of being placed within Zethus. Protodiscoelius merula was used for rooting the tree.
The following institutions provided material for this study: Obtained material had their identification checked by RBL by using identification keys in Bohart and Stange (1965) for New World specimens and Giordani Soika (1958Soika ( , 1969Soika ( , 1979 for Oriental, Australian and Afrotropical specimens respectively. Any specimen belonging to a species described after these studies were compared to its original description.

Character circumscription
The analysis was based on morphological characters from both external morphology and male genitalia. For the examination of the latter, insects were relaxed (when pinned) for extraction of the genital capsule, which was submitted to one of the following processes for dissolution of soft tissues and clarification: a) immersion in 10% KOH solution for approximately 24 hours, followed by neutralization of the base with 10% acetic acid solution, rinsed and stored in glycerin; b) immersion in lactophenol heated in a Thermolyne HP-A1G18B hot plate in the 200 potency for approximaetly five minutes and, after cooling down, stored in glycerin.
The structures were examined through a Leica Stereoscope Microscope MZ16 and pictures taken with a MZ16 Leica Stereoscope with an attached DFC295 camera. Pictures were taken with the Leica Application Suite software and image stacking performed with Combine ZP or Helicon Focus.

Phylogenetic analysis
Following circumscription, characters were coded and inserted into a matrix using the Winclada software (Nixon 2015) and exported for TNT (Goloboff et al. 2016).
An implied weighting (Goloboff 1993) scheme was carried in order to downweight homoplastic characters and the script "setk" by Salvador Arias (based on one of the equations in Goloboff et al. 2008) was used to determine a reference value of k. The determination of the constant k has been controversial as to what value should be used and no objective method has yet been set for this matter. For this reason, a series of values were tested based on the provided reference by the script in order to test nodal stability (Giribet 2003) across different concavities of weights through a sensitivity analysis (Wheeler 1995). The script suggested k=14.067383 and from this point, seven k values of whole numbers below and seven above the reference were taken (8 ≤ k ≤ 21; k ∈ Z), as well as the minimum value of 1 and an extremely high value of 500.
New technology search algorithms were used with the default settings, except for the following differences: Maximum trees held in memory: 10000; random seed 0; number of hits 100; Ratchet 200 iterations, with up-down perturbation 15; Drift 20 iterations; tree fusing 10 rounds.
After the search was complete, the resulting tree was opened in Winclada and exported for editing in Adobe Illustrator CS5. Support was calculated through symmetric resampling (Goloboff et al. 2003) with 10.000 replications.

Results and discussion
A total of 168 characters were circumscribed, 32 being from the head, 69 the mesosoma, 46 the metasoma and 21 from the male genitalia (the character list is provided in Suppl. material 1, while the matrix is provided in the Suppl material 4 in ".tnt" format -both available in OSF, under https://doi.org/10.17605/OSF.IO/FJV32). Characters based on previous studies (Garcete-Barrett 2014; Hermes et al. 2014;Lopes and Noll 2018) were indicated as such in Suppl. material 1.
The script suggested a k = 14.067383, for which only one tree was recovered ( Fig. 1), which shows Zethus as paraphyletic in relation to Ctenochilus and Ischnocoelia. The first was shown previously by Lopes and Noll (2018), who refrained from making taxonomic changes due to poor sampling and both cases were suspected by Carpenter and Cumming (1985). In this manner, both genera should be synonymized under Zethus. The clade comprised of Zethus, Ischnocoelia and Ctenochilus (herein recognized as Zethus), is supported by two homoplastic synapomorphies: the presence of a posterior proepisternal lamella ( Fig. 2A) and developed stem of T2 (Fig. 2B). The only tested k value that did not recover this clade was k = 1, probably due the strong weighting function, which discards homoplasies.
Out of the four subgenera of Zethus, only Z. (Zethoides) and Z. (Zethusculus) were recovered as monophyletic, while Z. (Zethus) and Z. (Madecazethus) were recovered as paraphyletic, what agrees in part with the results of Lopes and Noll (2018). Taxonomic changes proposed herein are based on these results. The genus has a long history of synonyms and subgeneric divisions and if a representative of a lineage deserving of its own subgenus was once affiliated to one of these names, the group or synonym is attributed a new status.
First, although Z. (Zethus) is indeed paraphyletic in relation to Z. (Madecazethus), it is regarding only the African representatives of the nominotypical subgenus (dark blue, Fig. 1), diverging from Lopes and Noll (2018) where the Malagasy subgenus is more  closely related to some Neotropical lineages, while Z. (Zethusculus) occupies a more basal position on the tree topology. Also, vice-versa occurs, as Z. (Madecazethus) is paraphyletic in relation to Zethus, but only the African representatives. Thus, the Malagasy and African representatives belong to the earliest (extant) lineage of Zethus, and due to the relationship between them, this group, in its entirety, shall be included in Z. (Madecazethus). This lineage shares with most of the other zethines the laterally straight pronotal carina (Fig. 2C), while most other Zethus present some sinuosity (Fig. 2D, E).
The redefined Zethus (Madecazethus) (dark blue, Fig. 1) is sister group to the remaining Zethus. This larger clade including Oriental and Neotropical representatives of the genus along with Ischnocoelia and Ctenochilus is supported by the narrower disposition of the ocelli, forming an equilateral triangle (Fig. 2F), the presence of the aforementioned sinuosity of the lateral portion of the pronotal carina (Fig. 2D, E) and wider gonocoxite.
The second robust lineage found in Zethus is that of the Oriental species (light blue, Fig. 1). Although no terminal from the Z. trimaculatus group was examined, the similarity of the species, based on literature (Giordani Soika 1958, 1995, reveal them to belong to this lineage. No genus group name was ever used to refer to this group, and thus, a new subgenus is proposed, Z. (Zethastrum).
The following lineage is the clade where the former genus Ischnocoelia is placed (light red, Fig. 1), and it is sister group to the Neotropical clade (tones of dark red, yellow, green and brown, Fig. 1). Four characters set this relationship: the pronotal carina dorsally lamellar with a medial notch (Fig. 2G) and with its lateral portion strongly sinuous (Fig. 2E) and the T1 strongly constricted apically. The presence of the dorsal suture in the mesepisternum ( Fig. 2E) with only few reversions and is potentially a synapomorphy but its optimization is ambiguous due to outgroup sampling. Still, Ischnocoelia is well supported and will be lowered to subgenus. Following recommendation of the International Code of Zoological Nomenclature for species-group names, an adaptation of the gender of the specific epithets will also have to be done and for the same reason.
All Neotropical species of Zethus, along with Ctenochilus, belong to one single clade (although there are representatives from the Nearctic, these are more derived and the stem lineages, therefore the origin of the group, Neotropical), supported by six homoplastic conditions: presence of a pretegular carina, which is incomplete (  In summary, Zethus presents now a complete Gondwanian distribution, in a manner that the lineages are related as African + (Oriental + (Australian + Neotropical)). This finding would suggest that the Zethus lineage probably already existed around 160 million years ago, with its lineages following the same separation sequence as Gondwana. However, Perrard et al. (2017) state that the major groups of Vespidae originated in the early Cretaceous (approx. 100 mya), using, among others, an Eumenine fossil of 89.6 million-years-old that calibrated the divergence of Zethus-Odynerus to that date. The fossil Symmorphus senex, used in calibration, belongs to a genus that is not among the stem lineage in Eumenine (Piekarski et al. 2018), meaning that Zethini, and therefore probably Zethus, could have derived much earlier and follow the vicariance pattern described above.
On the other hand, should we take into account the dating provided by Perrard et al. (2017), Zethus, a derived group, would have evolved later than 89 mya. In this sense, this would be a case of pseudocongruence, where the distribution pattern of species does not reflect the geographical history, despite appearences. Amorim et al. (2018) refer to this as a pseudogondwanian distribution, and explains as faunal exchange between Laurasia and Africa, the Nearctic and Neotropics and Oriental and Australasia Regions followed by the extinction of the representatives from the Holarctic. Since both hypotheses seem viable, more studies have to be carried out in order to come to a conclusion for the earlier biogeography of Zethus lineages.
Although there is poor support to properly establish a relationship between some Neotropical lineages, these lineages themselves are recovered at least moderately supported. The first identified lineage to diverge is Ctenochilus, the second genus to be included in Zethus (Fig. 1). All five species are included in the clade and will be placed in the subgenus Z. (Ctenochilus). This subgenus is sister group to a clade containing all the Neotropical Zethus, which is supported by three homoplasies: short galea; developed lamella of T2 ( Fig. 2L) and presence of gradulus in S3 (Fig. 2M).
The Neotropical Zethus clade reveals further paraphyly in the nominotypical subgenus, as Z. (Zethoides) and Z. (Zethusculus) are nested in this clade but derived from different points in the topology. The first clade to appear among the Neotropical group, which consists of two well defined lineages, is supported by the absence of the transverse interantennal carina and the outer margin of the tegula raised in its entirety (Fig. 2N). The first group is the already established Z. (Zethoides), whose monophyly had been recovered by Lopes and Noll (2018) and therefore, will matintain its status as subgenus.
Sister group to Z. (Zethoides) is a clade that consists of representatives from six species-groups: Z. cubensis, Z. heydeni (represented by Z. cerceroides), Z. montezuma, Z. sichelianus (represented by Z. cylindricus), Z. strigosus (represented by Z. adonis) and Z. sulcatus (represented by Z. harlequinus) groups ( Fig. 1). Dalla Torre (1904) described Wettsteinia sicheliana (= Z. sichelianus) and W. brasiliensis (= Z. adonis), but Brèthes (1906) rejected it as a distinct genus from Zethus. Since this was a distinct name for an assemblage that included representatives from two species-groups that are in the clade, the name will be carried to subgenus as Z. (Wettsteinia).
The second large clade of Neotropical Zethus consists of the remaining Z. (Zethusculus) and Z. (Zethus) and is supported by a wider disposition of the ocelli as an isosceles triangle (Fig. 2O), the absence of the apical propodeal lamella and the gradulus of T3 medially interrupted (Fig. 2P). Only three robust lineages can be identified within it (Z. (Didymogastra), Z. (Zethusculus) and Z. (Zethus)-see below) and there are some assemblages that lack strong support to establish a subgenus. Therefore, three speciesgroups will be dealt with as Incertae sedis: Z. pallidus (represented by Z. productus), Z. spinosus and Z. albopictus groups. The latter has controversial group taxonomy as it was once assigned to the Z. cubensis group only to be relocated into a group of its own (Stange, 2003). While there is moderate support for Z. albopictus + Z. (Zethusculus), only the male has been examined. A closer inspection of the female is desired and if more characters appear to corroborate this relationship, the inclusion of Z. albopictus in Z. (Zethusculus) could be supported.
The clade comprised of Z. fuscus and Z. smithii defines another new subgenus, which belongs to the Z. fuscus and Z. hilarianus species-groups, respectively. Perty (1833) described the genus Didymogastra. De Saussure (1855, 1875) later used the name as division of Zethus to allocate species which had an elongated stem of T2 and Dalla Torre (1892) later raised the status to subgenus. Since the name has been long since used, and the analysis supports this lineage as a subgenus, its status should be revised.
Despite Z. (Zethusculus) being recovered as paraphyletic by Lopes and Noll (2018), this subgenus was recovered here as monophyletic and with high support (Fig. 1). New characters now help support the monophyly of the subgenus, with lesser emphasis given to the aspect of sutures and grooves in the mesopleura and wider sampling of other genera of Zethini. Thus, Z. (Zethusculus) will keep its status as subgenus.
Finally, the last clade which consists of representatives from seven speciesgroups: Z. chalybeus (represented by Z. wagneri), Z. coeruleopennis, Z. discoelioides, Z. infundibuliformis, Z. prominens, Z. spinipes and Z. wileyi groups. The latter is a novelty, as upon the description of this monotypic group, Lopes et al. (2017) regarded it as being closely related to Z. hilarianus group and thus, would be expected to be part of Z. (Didymogastra). Since this large assemblage includes the type species of the genus, Z. coeruleopennis, the species in this clade will be kept in the subgenus Z. (Zethus). Although not examined, according to descriptions (Bohart and Stange 1965;Porter 2008;Carpenter 2011), Z. magretti, Z. stangei and Z. josefi fit best this subgenus.
The topology described above presented presented high nodal stability, except for the extreme values of 1 and 500 for k (the topologies for each k value can be seen in Suppl. material 2 -available in OSF, under https://doi.org/10.17605/OSF.IO/FJV32) used in the sensibility analysis, relationships and subgenera (Fig. 3 Type species: Vespa coeruleopennis (Fabricius, 1798). Description. Frons projected over antennal socket. Gena angled. Occipital carina ventrally complete, with small branch or completely absent. Pronotal carina short or lamellar. Proepisternum with posterior margin lamellar ( Fig. 2A), rarely blunt. Metanotum bent, rarely flat. Mid tibia with one or two spurs. Propodeum without apical rim. Propodeum with or without apical lamella, orifice slit like. Valvula projected, with superior margin free, infero-posterior angle straight. T1 variable. T2 with evident stem (Fig. 2C), occasionally absent.
Observations. The large diversity included in Zethus reflects diagnostic features, of which few encompass the entirety of the genus, only the projected frons, angled gena, absence of the apical rim in the propodeum, orifice slit like and traits of the valvula being constant within the group. Also, the absence of the apical rim in propodeum is constant throughout the entire genus while it is present in most of the other Zethini. The subgenera are more defined in more detail.
Distribution. Cosmopolitan (specified under each subgenus) Included species. 298 (listed under each subgenus in Suppl. material 3) Identification key to subgenera of Zethus Fabricius 1 Pronotal carina short, not lamellar, and at most slightly sinuous laterally (Fig. 2C, D)  T1 with lateral margins touching ventrally (Fig. 7B). Male antennae rolled (Fig. 12D); females with clypeus microstriate (Fig. 12B); tegula with outer margin raised and only medially interrupted (Fig. 12C)  T3 posterior margin projected, with apical lamella tapered laterally with indented secondary lamella present (Fig. 10B, C); S3 with apical lamella reduced and a medial lamellar lobe projecting subapically from sclerite (Fig.  10C)  A synopsis of each taxon in the key will follow, in alphabetical order, with a taxonomic catalog, description, observations, distribution and included species number. A list of species in each subgenus is provided in Suppl. material 3 (available in OSF, under https://doi.org/10.17605/OSF.IO/FJV32). The description will consist of a core description, where general characters are cited and the synapomorphies, which will be in bold. Unique synapomorphies will be in bold italic.
Observations. Although here the very long stem of T2 is marked as a synapomorphy, it is not constant throughout the entire subgenus. Some representatives from the hilarianus group have a medium-sized stem. Part of the hilarianus group (former Z. smithii group) has been revised and phylogenetically analysed by Lopes et al. (2015).

S5 and S6 in males with transverse carina. Volselar crest axe-like.
Observations. The ICZN stablishes that a species' epithet should agree with the gender of the genus. Since Ischnocoelia is female and Zethus, male, there is a mandatory spelling change (Art. 34.2, ICNZ) for names that are female. While most of the specific epithets could simply be converted into masculine form due to the new com-   Bequaert, 1928: 151 (cat.). Bohart and Stange 1965, 40: 11 (note (Fig. 7B) and narrow expansion. Anterior margin of S1 variable in shape. Stem of T2 developed. Apical lamella of T2 present, reduced in Malagasy species. Lamellae of T3 and S3 developed and simple, without lateral indent. S5 and 6 of male withour carinas. Aedeagus with apically flattened apodemes. Observations. Although Z. madecassus and Z. seyrigi were the only representatives of Z. (Madecazethus) they are not sister groups and thus, should not be placed in a species group of their own. A wider analysis of the African Zethus is needed to verify the validity of the already proposed pubescens group (represented by Z. empeyi, Z. rhodani and Z. rotschildianus), since it may also be paraphyletic.
Observations. Only three of four species-groups from the Oriental region were sampled: Z. quadridentatus group (represented by Z. varipunctatus); Z. luzonensis group (represented by Z. fulgens); and Z. dolosus group (all the other Oriental Zethus). Although the present analysis shows the basal longitudinal carina and weak apical constriction of T1 as synapomorphies for the subgenus and constant throut the clade, the sampling may not be as optimal as presumed, leaving out species with different traits than what was portrayed. Wang et al. (2019) recently published three new species and a key to the Oriental species. The constriction of T1 is occasionally used in their key, highlighting the existence of specimens with a stronger constriction. However, this is not illustrated, making it impossible to determine to what degree the sclerite narrows. Still, clearly strong constrictions are illustrated for the newly described species Z. asperipunctatus Wang and Li and Z. nullimarginatus Wang and Li. The basal longitudinal carine of T1 was not used in the key, but in the two aforementioned species, the structure was not observed.
The Z. dolosus species group appears to be paraphyletic in relation to the Z. quadridentatus group. A thorough analysis is needed of Z. (Zethastrum) including representatives from the trimaculatus species group to verify the validity of the assemblages.
Although Gusenleitner (2007) described Z. fulgens comparing it to Z. nigerrimus, a species in the trimaculatus group, the type was examined, and it fits more properly the Z. luzonensis group. Selis (2017) recently revised this group, but left Z. fulgens out, just as Wang et al. (2019) did not key it in couplets along with the group. However, in its description (Gusenleitner, 2007), more emphasis is given to punctation and color and little attention is given to the pilosity, an outstanding and diagnostic trait of the group and so, it is no surprise Selis did not consider examining this species.
Description. Interantennal longitudinal and transversal carinas absent. Male flagellum rolled (Fig. 12D). Female clypeus microstriate (Fig. 12B), short and without apical teeth. Male mandible 3-toothed. Galea short. Labial palpi 4-segmented with first palpomere curved. Occipital carina ventrally complete or as small branch. Vertex usually raised (flat only in the imperfectus group). Genal margin mostly evenly convex (sinuous in the imperfectus group). Pronotal carina lamellar with lateral portion strongly sinuous. Pre-tegular carina incomplete. Welts and notalulices absent. Tegula with outer margin usually angled, raised except for medial interruption (Fig. 12C). Epicnemial carina complete. Mesepimeron without carina. Mid tibia with one apical spur. Metanotum anteriorly margined with lateral carina extending at most to anterior half of the sclerite. Spines in posterior tibia scattered. Apical angle of marginal cell variable. Dorsal aperture of propodeum only present in imperfectus group. Propodeum  with lateral carinas and without submedian and sublateral carinas. Apical lamella of propodeum only present in imperfectus group. T1 with short stem and medium expansion. S1 with elliptical anterior margin. S2 with apicall lamellae. T3 with apical lamella only in imperfectus group. S3 without lamellae only in mexicanus group. S5 and 6 without carinas. Anterior margin of ventral lobe of aedeagus reflexed outward. Cuspis with apical tuft.
Observations. While Lopes and Noll (2018) recovered this subgenus as paraphyletic, the results here present support the wider analysis of the group (Golfetti et al. 2020) where it is recovered as monophyletic.

Nomen dubium
These taxa of doubtful application lack evidence that allows a proper affiliation to species-groups or subgenera, thus considering them incertae sedis as well. Three of the taxa are already cited in Bohart and Stange (1965): Z. assimilis (Brèthes), Z. holmbergii Brèthes and Z. medius Brèthes. Three species are added to this list: Z. didymogastra Spinola, Z. rufipes Fox and Z. scandens Zavattari.
Zethus didymogastra Spinola, 1841 -After its description, de Saussure (1852) synonymized it under Z. binodis Fabricius, only to go back and validate Z. didymogastra again after two decades (de Saussure, 1875). The last mention of this species was by Cameron (1912) when he synonymized it once again also under Z.binodis. Nevertheless, the taxonomic catalog of Z. binodis in Bohart and Stange (1965) does not bring the synonym nor is the species brought up anywhere else in their study.
Zethus rufipes Fox, 1899 -This taxon makes a rather curious case, where Bohart and Stange (1962) desginated a lectotype, but never mentioned it in their great revision of 1965. The fact that this species was omitted in Bohart and Stange (1965) was noted by Carpenter and Vecht (1991). Zethus scandens Zavattari, 1913 -This species was described by Zavattari and only again cited by Rasmussen and Asenjo's (2009) catalog for Peruvian fauna. No remarks or lists were made by any other authors, not even Bohart and Stange (1965).

Concluding remarks
This is the first broad phylogenetic analysis of Zethus and corroborates an already predicted paraphyly of the genus in relation to Ischnocoelia and Ctenochilus. Still, the study suggests an expansion regarding the subgeneric classification, as nine robust lineages were recovered and labeled as subgenera. Despite the fact that the clades comprising each subgenus are at least moderately supported, their relationships are not, and there are also three species-groups (albopictus, pallidus and spinosus) that require deeper investigation in order to assign them to a subgenus or create one of their own. Thus, this study reveals an incongruence with traditional classification, proposes a new one and forms a base for more specific approaches on the genus.