A preliminary phylogenetic analysis of the New World Helopini (Coleoptera, Tenebrionidae, Tenebrioninae) indicates the need for profound rearrangements of the classification

Abstract Helopini is a diverse tribe in the subfamily Tenebrioninae with a worldwide distribution. The New World helopine species have not been reviewed recently and several doubts emerge regarding their generic assignment as well as the naturalness of the tribe and subordinate taxa. To assess these questions, a preliminary cladistic analysis was conducted with emphasis on sampling the genera distributed in the New World, but including representatives from other regions. The parsimony analysis includes 30 ingroup species from America, Europe and Asia of the subtribes Helopina and Cylindrinotina, plus three outgroups, and 67 morphological characters. Construction of the matrix resulted in the discovery of morphological character states not previously reported for the tribe, particularly from the genitalia of New World species. A consensus of the 12 most parsimonious trees supports the monophyly of the tribe based on a unique combination of characters, including one synapomorphy. None of the subtribes or the genera of the New World represented by more than one species (Helops Fabricius, Nautes Pascoe and Tarpela Bates) were recovered as monophyletic. Helopina was recovered as paraphyletic in relation to Cylindrinotina. One Nearctic species of Helops and one Palearctic species of Tarpela (subtribe Helopina) were more closely related to species of Cylindrinotina. A relatively derived clade, mainly composed by Neotropical species, was found; it includes seven species of Tarpela, seven species of Nautes, and three species of Helops, two Nearctic and one Neotropical. Our results reveal the need to deeply re-evaluate the current classification of the tribe and subordinated taxa, but a broader taxon sampling and further character exploration is needed in order to fully recognize monophyletic groups at different taxonomic levels (from subtribes to genera).


Introduction
The tribe Helopini Latreille, 1802 currently contains two subtribes (Cylindrinotina and Helopina), 42 genera, and 686 species (Gebien 1943, Blackwelder 1946, Nabozhenko and Löbl 2008. A significant part of this richness is concentrated in the Palearctic Region, for which a recent catalogue is available (Nabozhenko and Löbl 2008) and where taxonomic work has been relatively constant. In contrast, only four genera are recognized for the New World, three of which are Holarctic: Helops Fabricius, 1775;Tarpela Bates, 1870;Odocnemis Allard, 1876, and one is exclusively Neotropical: Nautes Pascoe, 1876. Odocnemis is currently classified in the subtribe Cylindrinotina, Helops and Tarpela in the subtribe Helopina, and Nautes has not been classified in a subtribe because it is not included in the catalogue of Nabozhenko and Löbl (2008) as it is not present in the Palearctic region.
Helops, the type genus of the tribe, was described by Fabricius (1775) based on a few cephalic structures, such as the maxillary and labial palps, the labium, and the antennae of a European species, H. caeruleus (Linnaeus, 1758). In the following centuries more than two hundred Palearctic species were added to this genus, but subsequent regional taxonomic revisions transferred most of them to other genera, leaving Helops with nine species in the region (Reitter 1922, Nabozhenko andLöbl 2008). With one exception, no such revisions have taken place for the American component of the tribe, currently composed of 150 species. In the first synoptic work for the family in North America, Horn (1870) listed 23 species of Helops and Stenotrichus rufipes LeConte, 1851, which was placed in Amphidorini, but later synonymized with Helops (Bouchard et al. 2005). Allard (1876Allard ( , 1877, author of the only world revision of the tribe, recognized Helops opacus LeConte, 1859 and reassigned the remaining species among the following genera: Diastixus Allard, 1876 andCoscinoptilix Allard, 1877 with exclusively American distribution, and Stenomax Allard, 1876, Nesotes Allard, 1876and Catomus Allard, 1876, with Palearctic distributions. The Stenomax subgenus Omaleis Allard, 1877, which included three species from California, was recently synonymized with Odocnemis Allard, 1876by Nabozhenko (2001a. Allard included three other genera for the continent: Hegemona Laporte de Castelnau, 1840, Nautes, and Tarpela, which were described from Neotropical species. Hegemona was later transferred to Stenochiinae (Doyen 1987). Twenty-six species of Nautes are Neotropical (Blackwelder 1946, Papp 1961, Steiner 2006 while Tarpela currently contains three Nearctic species (Gebien 1943, Papp 1961, 51 Neotropical species (Blackwelder 1946) and 15 species from Asia, mainly from Japan (Nabozhenko and Löbl 2008).
In the monumental Biologia Centrali-Americana, Champion (1887Champion ( , 1893 described approximately half of the current Helopini species known from North and Central America. Even though he was aware of the heterogeneity of the group, he synonymized Allard's five genera with Helops. In his opinion, retaining Allard's names for the species originally placed in Helops would have required him to propose generic names for the species in Nautes and Tarpela. Unlike Helops, the genera Tarpela and Nautes have more detailed taxonomic descriptions and were thought to be closely related (Bates 1870). The configuration of the prosternum and mesosternum were the main characters proposed to differentiate the two genera (Bates 1870). Champion (1887) considered these characters to be inconsistent, changing Allard's classification by transferring two species from Nautes to Helops and Tarpela: N. farctus (LeConte, 1858) and N. eximia (Bates, 1870), respectively. More recently, Doyen (1988) described two Mexican species of the tribe: Helops scintillatus and H. noguerai, but had problems assigning them to this genus because they shared characters with some species currently placed in Nautes.
In short, this diverse tribe includes two subtribes and multiple genera with worldwide distributions (Gebien 1943) and with different and conflicting circumscriptions, at least in the Holarctic and Neotropical components, considering from three (Champion 1887(Champion , 1893 to seven genera (Allard 1877). For the reasons detailed above, an evaluation of the recent classification seems necessary. A phylogenetic approach including all taxa is at this moment unrealistic, but a well design taxon sampling could shed light upon the naturalness of the genera and provide a basis for further research strategies aiming to translate phylogenetic hypotheses into natural classifications. The goals of this work are to explore and codify the morphological variation observed within the Neotropical helopines, for the first time test the monophyly of the subtribe Helopina and of three of the four genera present in the New World (two genera belonging in subtribe Helopina plus Nautes that is currently unassigned), and highlight issues in the current classification to provide guidance for future studies.

Phylogenetic data: characters
Two hundred eighty-one specimens were examined with an Olympus SZH10 stereomicroscope (magnification: 17.5-350×) equipped with an ocular graticule for length measurements, and a drawing tube. Morphological characters were measured as follows: width of the head was measured across the vertex, length of the last antennomere in the female was measured along its longest edge; width was measured across its widest point; length of pronotum was measured along the midline from its anterior edge to its posterior edge; width was measured across its widest point. Puncture density follows modified conventions used by Paulsen (2005) and Smith et al. (2011): either confluent (separated by one or less than a puncture diameter), moderate (separated by 2-3 puncture diameters), or sparse (separated by 4 or more puncture diameters). Nomenclature and interpretation of female genital tract follows Tschinkel and Doyen (1980) and Doyen (1994).
Thirty-two characters correspond to external morphology; characters used in generic descriptions (Pascoe 1866, Bates 1870 or in previous phylogenetic studies (Doyen and Tschinkel 1982) were included (Figs 1,2). The remaining 35 are based on male and female genitalia. Internal characters  were coded according to previous works (Antoine 1947, Doyen 1994, Flores 1996, Nabozhenko 2001a, Aalbu 2005, Rosas et al. 2011) independent of the fact that some were used to investigate other families as they are considered to be useful in Tenebrionidae as well (Rosas et al. 2011). Two characters (35, 43) plus two character states (67: 1, 2) were used for the first time. Female genitalia were dissected, cleared and stained following Tschinkel and Doyen (1980), replacing NaOH with KOH. Photographs were taken using a Leica microscope equipped with a camera Leica Z16 APO A. The imaging software used was Leica Application Suite 2.8.1.
In total, 44 binary and 23 multistate characters were coded and treated as nonadditive. Individual consistency and retention indices (ci, ri) are provided for all characters from the consensus tree (synapomorphies have a value of 1 for both indices). Four additional characters (listed and explained at the end of the character list) were initially explored but removed from the final analysis due to their high homoplasy, assessed by a character removal methodology (see below).

Phylogenetic analysis
The matrix was compiled using WinClada (Nixon 2002). Heuristic searches were conducted through NONA (Goloboff 1999) with multiple Tree Bisection and Reconnection (TBR) using 1,000 initial Wagner trees (mult * 1000), holding 20 trees per replication (hold/20) and expanding the memory for a final TBR to completion with up to 10,000 trees (max * 10000). The cladograms were rooted with Uloma mexicana. All most parsimonious trees (MPTs) found were collected, and ambiguously supported branches were collapsed in WinClada. Identical trees were then removed and a consensus was calculated using the option "Strict" in WinClada.
A simple sequential character removal analysis (modified after Davis et al. 1993) was carried out as implemented in WinClada (Nixon 2002), using the same search parameters as explained. The length of the resulting 71 consensus trees (one for each matrix resulting from the progressive removal of the 71 characters) was compared to determine the influence of each character in the topology of the consensus of the MPTs. In this way, four characters (listed above) were detected to particularly introduce conflict in the analysis due to high homoplasy values and were removed from the matrix. When removing each of these characters, the length of the consensus decreased by more than 30 steps and the resolution of the topology greatly improved. The final 67-character matrix (character listed and explained above) was then analyzed with the parameters described in the previous paragraph. These characters are mapped onto the consensus only if their optimization was not ambiguous and if they were present among all the MPTs. This was assessed using the option "Map Common synapomorphies" on the sub-menu "Synapomophies" menu "Optimize" of TNT (Goloboff et al. 2003). The consensus was used to map homoplasy at the level of characters in WinClada; a metafile was created and the tree was edited using Corel Draw X6 (Corel Corporation 2012).
To evaluate statistical branch support, a bootstrap analysis was conducted with NONA (Goloboff 1999) through WinClada (Nixon 2002). For this analysis 1,000 replicates were conducted for each using 100 initial trees holding 20 trees and expanding the memory up to 1,000 trees (mult*100 hold/20 max*1000). Frequencies were calculated on the consensus of the 67-character matrix and only values above 50% are shown.

Results
The 67-character matrix ( Table 2) yielded 12 most parsimonious trees with 301 steps (length = L), a consistency index (ci) of 0.29, and a retention index (ri) of 0.59. The strict consensus (L=314; ci=0.28; ri=0.56) is presented in Figure 7. Six out of seven characters retrieved as synapomorphies are from internal morphology. Four synapomorphies correspond to the female genitalia: vagina strongly curved in the apex (character 40: state 1), more than one spermathecal tube (41:1), smooth texture of spermathecal tube (48:0), and terminal position of the accessory gland (49:2). Two synapomorphies correspond to the male genitalia: distribution of evident setae on the parameres (58:1), and presence of a dorsal projection or keel on the parameres (64:0). One synapomorphy corresponds to external morphology: the filiform shape of the antennae: (3:0). Although only six clades had bootstrap values over 50%, most clades are supported by a unique combination of at least two characters.
From the sampled Palearctic Helopina, only Helops rossii Germar, 1817 and H. insignis Lucas, 1846 constitute a clade that is sister to the rest of the tribe, and is supported by the pubescent ventrites with homogeneous punctures (32:1).
A large clade of mostly Neotropical species from the genera Helops, Nautes and Tarpela, plus two Nearctic species of Helops, was recovered with support from three characters: clypeus slightly depressed (1:0), more than one spermathecal tube (41:1) (retrieved as synapomorphy), and an accessory gland terminal to the spermathecal tubes (49:2) (retrieved as synapomorphy). Helops punctipennis LeConte, 1870 andH. rufipes (LeConte, 1851), both Neartic, are supported as sister to this mostly Neotropical clade by two internal characters: the deep notch of the eighth sternite (52:0) and the lobate shape of the median lobe (66:1). The earlier divergent lineage within this mostly Neotropical species is a clade formed by Tarpela contigua Champion, 1887 and Helops perforatus Horn, 1880, supported by two internal characters: paraproct three or more times longer than the coxite (39:0) and the presence of a ventral projection at the parameres apex (63:0). Helops inanis Allard, 1877 and Tarpela depressa Champion, 1887 form a grade with respect to the remaining mostly Neotropical species. There is then a polytomy that includes Tarpela torrida Champion 1887, a clade with three Tarpela species including the type (T. browni Bates, 1870), and another clade that is mostly composed of Nautes species and includes all the sampled species of this genus, Table 2. Matrix for the cladistics analysis of the tribe Helopini (Coleoptera, Tenebrioninae, Tenebrionini); "-" represents inapplicable character states, "?" represents not observed data.

Taxonomic implications
Although supported by our results, the monophyly of the tribe still requires a more rigorous test including a wider sample of species from more tribes including species from other closely related tribes (e.g. Triboliini, Blaptini). The only synapomorphy supporting the tribe, the filiform shape of the antennae, could be an artifact of our sampling, as the antenna have also been reported as moniliform or gradually clavate  within the tribe (Aalbu et al. 2002). Based on our examination of many additional species, we know of no Helopini with moniliform or submoniliform antennae, nevertheless gradually clavate antennae are present in some species, such as Nautes antennatus Champion, 1887, N. varians Champion, 1887, Helops durangoensis Champion, 1887 The fact that Cylindrinotina is nested within Helopina implies that there is no justification for the recognition of two subtribes: either no subtribes should be recognized or more subtribes should be recognized. A denser sampling of Palearctic species   (L= 314; ci=0.28; ri=0.56). Characters are mapped onto the consensus only if their optimization is not ambiguous and if they are present among all the MPTs. The consensus is used to map homoplasy at the level of characters. Black rectangles represent single, non-homoplasious character state transformations, and white rectangles represent multiple, homoplasious character state transformations. The number depicted above each rectangle represents the character and the number below the rectangle represents the character state. The bigger number below the branches corresponds to Bootstrap values over 50%. The combination of characters for each terminal is not shown. Three important synapomorphies are illustrated in the cladogram: the filiform antennae (3:0), as the single synapomorphy of the tribe, and the number of spermathecal tubes (41:1) plus the terminal position of the accessory gland (49:2), as the synapomorphies supporting a mostly Neotropical clade. These character states are reported for the first time for the tribe. Two shades of gray in the cladogram indicate the subtribe to which the terminals belong (except Nautes). Colors in the terminals indicate their geographic distribution. Below seven terminals the former classification (genus or subgenus) is shown. An asterisk indicates the type species included in the analysis: Tarpela browni and Nautes fervidus. could help reveal which of these alternatives is better supported. According to the current sampling, it is possible that the Palearctic Helops remain as an independent earlier divergent lineage within the tribe, including the type species (H. caeruleus), which is morphologically similar to the sampled Paleartic species. If this was the case, Helops would have to be re-circumscribed to include only the Palearctic species and new generic names would be necessary for the New World lineages.
Further earlier divergent lineages may be revealed as sampling of Entomogonus, Raiboscelis and Probaticus is improved, as well as other genera not included in our sampling (e.g. Catomus Allard, 1876, Hedyphanes Fischer von Waldheim, 1820, and Nesotes Allard, 1876. The unresolved position of Helops cisteloides Germar, 1824 indicates the possibility that other New World lineages could be identified as sampling is increased. If subtribes are to be recognized, Cylindrinotina would need to be expanded to include Asian species of Tarpela (as T. cordicollis) and Neartic Helops (as H.aereus). This subtribe would also have to include several Holartic genera (besides Odocnemis). The Holarctic region has an intricate history (Brown and Lomolino 1998), with dispersion of groups taking place in several moments of the Tertiary (Sanmartín et al. 2001). The geographic heterogeneity of the cylindrinotine clade shows the importance of using a phylogenetic approach in which the morphological diversity of the taxa is represented, regardless of their present geographic distribution.
The polyphyletic nature of Helops and Tarpela render Champion's classification (1887, 1893) and those of previous authors like Horn (1870) artificial. In contrast to Champion's conservative classification, Allard's classification (1876, 1877) was more natural in the sense that he recognized several lineages in the New World, some of them with Holarctic distribution. Allard's placement of H. aereus in cylindrinotine is supported by our results; nevertheless our results suggest that it should be classified in Nalassus, not in Stenomax. However, further analyses including more genera from the subtribe are necessary before taxonomic changes are made. This is also the case of the Asiatic Tarpela cordicollis, which was classified in a different subgenus (Lamperos) by Allard (1877). Allard (1876) proposed the genus Lamperos to comprise some Tarpela species from North America and Japan, but later reduced it to subgenus (Allard, 1877), including T. cordicollis. Aside from this species of Tarpela, all the others, including the type (T. browni) are placed in a different lineage formed mostly by Neotropical species. This lineage, nevertheless, also includes species of Nautes.
The paraphyletic nature of Tarpela with respect to Nautes, could imply different outcomes as a wider taxon sampling (including more Nearctic species of Helops, Nautes and mainly Tarpela) and character (e.g. from DNA or fine structures revealed using SEM) is considered. Either several lineages could be recognized as different genera or all the species could be lumped in a single larger genus (Nautes due to nomenclatural priority, or if applicable, a conserved name Tarpela). Even if Nautes was supported as a different genus, taxonomic rearrangements seem to be likely. According to the current sampling, Helops farctus and Tarpela aerifera would need to be reassigned to Nautes.

Morphology
Female genitalia have been used as a source of characters to study the relationships among suprageneric taxa in Tenebrionidae (Tschinkel and Doyen 1980, Doyen and Tschinkel 1982, Doyen 1994. Nabozhenko (2006) recognized four morphological patterns for the female genitalia that he associated to lineages from Helopina and Cylindrinotina, two patterns within each subtribe. In our sampling we only observed two of these patterns (Fig. 4A, B), but we also observed two patterns not previously reported for the tribe (Fig. 4C, D). Nevertheless, one of these patterns (Fig. 4C) was previously described for species belonging to Pimeliinae (Doyen 1994). These two patterns newly reported for Tenebrioninae were only seen in the mostly Neotropical clade. Most of the members of this clade share the pattern previously reported for Pimeliinae (Fig. 4C) and the pattern that we report here for the first time (Fig. 4D) was present only in the earlier divergent group of this clade (Helops perforatus-Tarpela contigua) as well as in T. depressa. Nabozhenko (2001bNabozhenko ( , 2002aNabozhenko ( , 2002bNabozhenko ( , 2005 describes the morphological patterns for the female genitalia tubes of the helopiod type as follows: basal spermathecal duct distinct; spermatheca consisting of two ducts of different length, without additional reservoirs and short processes; basal duct about as long as duct between place of running of gland and branching of spermatheca (Fig. 4A). The female genital tubes of the nalassoid type consist of a short and simple spermatheca, without lateral processes, reservoirs, and branching; gland short, about as long as spermatheca (Nabozhenko 2001b(Nabozhenko , 2002a(Nabozhenko , 2002b Fig. 4B). The pattern shared with some Pimeliinae (Doyen 1994) consists of several spermathecal tubes close to each other or united at the base as a fascicle, always originating near or at the vagina apex, hence without a basal spermathecal duct (Fig. 4C). The newly documented pattern presents several spermathecal tubes distant from each other (Fig. 4D). In both cases, the accessory gland emerges from the common duct (if it is present), always in a terminal position with respect to the spermathecal tubes (Doyen 1994).
Due to its high variation, male genitalia have also been used to explore the relationships among species and higher taxonomic groups (e.g. Tschinkel 1982, Aalbu 2005). As in the case of the female genitalia, Nabozhenko (2006) also recognized four morphological patterns for the male genitalia in lineages of the subtribes Helopina and Cylindrinotina. In contrast to the female genitalia, the morphological patterns found among the sampled species fit three of the previously described patterns by Nabozhenko (2006), only with what we consider a minor variation in the catomoid type. The patterns that we recognize correspond to Nabozhenko's helopioid, nalassoid and catomoid types. According to Nabozhenko (2001bNabozhenko ( , 2002aNabozhenko ( , 2005 the helopiod male genitalia type in the broad sense (Fig. 5B) has, among other characters: heavily sclerotized parameres, covered with elongate punctures; baculiform sclerites of spiculum gastrale approximate, not curved outwards in dorsal view. The nalassoid male genitalia type (Figs 5A, E, and 6C, F) is characterized by: an aedeagus weakly sclerotized, semitransparent; parameres elongate, produced apically into compressed keel (Nabozhenko 2001b(Nabozhenko , 2002a. The catomoid male genitalia type is only present in the mostly Neotropical clade and is characterized by: penis with two or three apices, rounded in apical part; phallobase very long in comparison with short parameres; parameres with elongate aspirate punctuation and inconspicuous short hairs (Fig. 6B, E) (Nabozhenko 2006). The variation we found for all the species with respect to the catomoid aedeagus type is a lobate apical part of the penis and a shorter basal piece (relative to the length of the parameres) (Fig. 5F).
The recognition of the female and male genitalia types is translated into several homology hypotheses reflected in the matrix as characters 33 to 67 and their corresponding character states (see the list of phylogenetic data: characters above).
Although widely used as a taxonomic character, the keel on the parameres (64) has been reported as not always present through the subtribe (Nabozhenko 2001a). Nevertheless, this could be an artifact of the observation tools, as small keels can be detected when using a scanning electron microscope (SEM) (results not shown). For this reason we prefer to code this condition as "not evident" (see character 64) (in contrast to lacking). This is the same for the "absence" of setae on the parameres, here coded as "not evident" (see character 57).
Other diagnostic or traditionally used characters of the clypeus, antennae, prosternum, wings and tarsi were homoplastic but generally informative, contributing to the overall resolution of the tree. Only four characters from the original matrix introduced high levels of conflict, resulting in a lack of resolution in the consensus. These characters were all continuous and without a more refine codification, e.g., using statistical or morphometric tools, they only obscured the relationships posed by the remaining characters. On the other extreme, the shape of the antenna, generally considered to be a homoplastic character, was recovered as synapomorphic for the tribe. However, this synapomorphy needs to be tested with a broader taxon sampling.

Conclusions
Although supported by our results, the monophyly of the tribe still requires a more rigorous test in terms of the taxon sampling from related tribes.
None of the subtribes or the analyzed subordinate genera of Helopini sampled by more than one species was corroborated as monophyletic. A wider taxon sampling is required to circumscribe them in a natural way.
Helops and Tarpela are polyphyletic, while Nautes is paraphyletic, and hence it is expected that further taxon and character sampling in a cladistic context will provide evidence for further splitting of Helops and Tarpela and a re-circumscription of Nautes including some Helops and Tarpela.
Our results show that in order to achieve a natural classification of Helopini, sampling of taxa should not be based on geographic distribution, although there might be some geographically correlated lineages. This approach has shown that there is a derived New World clade that is mainly composed by Neotropical species. Future efforts should also concentrate on increased sampling within this clade, to reveal other lineages or to corroborate the current ones, so that taxonomic changes can be concordantly proposed.