A new genus and species and a revised phylogeny of Stereomerini (Coleoptera, Scarabaeidae, Aphodiinae), with notes on assumedly termitophilic aphodiines

A new genus and species of Stereomerini, Cheleion malayanum gen. n. and sp. n., are described based on a single specimen from the Malacca peninsula in Malaysia. It is the fi rst stereomerine found on the mainland of the Asian continent. A key to genera of Stereomerini and a phylogenetic estimate of the stereomerines using 24 species representing 7 tribes of Aphodiinae based on 53 morphological characters. Th e phylogenetic analysis places the new species within the tribe Stereomerini as a sister to the monophyletic group including Adebrattia, Austaloxenella, Bruneixenus, Pseudostereomera and Stereomera. A survey is given of the various taxa of Aphodiinae that are termite and/or ant inquilines, or have been presumed to be so based on morphology, and their phylogeny is discussed.


Introduction
Th e aphodiine tribe Stereomerini was erected by Howden and Storey in 1992 for four genera of minute, rare and supposedly termite-associated aphodiines Stereomera Arrow, Termitaxis Krikken, Australoxenella Howden and Storey, and Bruneixenus Howden and Storey (Howden and Storey 1992). In 1995 Bordat and Howden excluded Termitaxis from the tribe and described 3 new genera and species: Adebrattia depressa Bordat and Howden, Danielssonia minuta Bordat and Howden and Pseudostereomera mirabilis Bordat and Howden (Bordat and Howden 1995). Since then one new genus (Daintreeola Howden and Storey) and 11 new species of the genera Daintreeola, Bruneixenus and Australoxenella have been described (Storey and Howden 1996;Howden and Storey 2000). Th e number of genera of Stereomerini is presently 7 and the number of species 18, all from insular south east Asia and Australia.
Synapomorphies for the Stereomerini include a grooved head as well as a particular pattern of carinae and grooves on the pronotum with three pronotal carinae concentrated in median third and one transverse groove, plus convex body shape, wide epipleura and reduced eyes. Characters that may be either apomorphic or plesiomorphic are a more or less lanceolate posterior prosternal process, clypeus without dentation and fl attened tibiae (Forshage 2002).
In the collections of the Swedish Museum of Natural History (NHRS) we found a remarkable specimen clearly belonging to Stereomerini but representing a new undescribed taxon, the fi rst one from a locality on the Eurasian continent (the Malaysian Peninsula). As it has not been possible to accommodate the new taxon in one of the existing genera we felt impelled to describe a new genus as well. In order to test this, we performed a phylogenetic analysis of all the genera of Stereomerini. As the relationships between various aberrant tribes within the Aphodiinae, and between those tribes and the poorly delineated Eupariini, we included a wealth of other genera in the analysis.

Methods
Images of the new species were obtained with a digital Olympus DP70 camera mounted on an Olympus SZX12 stereo microscope using the imaging software AnalySIS 5.0 to montage section images with diff erent focus. Micrographs were taken with a Zeiss Su-pra35VP scanning electron microscope on low vacuum. Th e low vacuum scanning electron microscopy does not require the specimen to be prepared in the traditional manner including critical point drying and gold-coating and is thus particularly suitable for type material. Character coding of the new taxon and other taxa at hand were done from a combination of stacked digital images and direct observation in stereo microscope. Taxa not at hand were coded from scorings and images found in literature. Th ese include Australoxenella humptydooensis Howden and Storey, Bruneixenus squamosus Howden and Storey, Cartwrigthia intertribalis Islas, Ascharhyparus peregrinus (Hinton), Termitodiellus es-akii Nomura and Stereomera pusilla Arrow and Termitaxis holmgreni Krikken coded from Howden and Storey (1992), Daintreeola grovei Howden and Storey coded from Howden and Storey (2000) and Termitoderus ultimus Krikken coded from Krikken (2008a).

Character selection
Several of the characters have been used in previous phylogenetic analyses (Howden and Storey 1992;Howden 1995, Forshage 2002). In an attempt to further defi ne the tribe Stereomerini based on morphological characters, we have added several characters (see Table 2) on the head (characters 4, 5, 10, 11), pronotum (characters 15, 18, 20, 22, 23, 24), legs (characters 34, 35), wings (characters 37, 38, 39, 44, 45), abdomen (characters 46, 53). We did not want to dissect the single type specimen of the new taxon so we have not been able to compare mouthparts and genitalia which are potential valuable sources of additional characters.

Taxon selection
For the phylogenetic analysis we included representatives from all genera of the tribe Stereomerini, as this was our primary group of interest. In addition we wanted to include representatives of the tribes believed to be closely related to the Stereomerini based on morphology as well as some taxa of unclear taxonomic status and phylogenetic position. In addition to the 8 Stereomerini species, we thus coded 5 species of Rhyparini (Aschnarhyparus peregrinus, Termitodiellus esakii, Rhyparus suturalis Schmidt, Cartwrightia intertribalis and Sybacodes simplicicollis Fairmaire), 1 species of Termitoderini (Termitoderus ultimus Krikken), 1 species of Termitotrogini (Termitotrox consobrinus Reichensperger), 2 species of Corythoderini (Corythoderus loripes Klug and Neochaetopisthes heimi Wasmann) as well as Termitaxis holmgreni (incertae sedis). As these aberrant taxa are of uncertain relationships with Eupariini -possibly diff erently to diff erent parts of Eupariini -we wanted to include a fair representation of the diversity of that tribe. Th us we included a representative of an aberrant genus, morphologically similar to some of the truly or assumedly termitophilous tribes (Notocaulus sculpturatus Boheman), plus the largest genera of the tribe (Ataenius scabrelloides Petrovitz and Saprosites laeviceps Harold), as well as the type genus of the tribe (Euparia castanea LePeletier and Serville). As outgroup taxa we chose Aphodiini, since they are the major group of Aphodiinae that we expect to be safely outside the group of interest here; we included one more or less typical and one aberrant representative of Aphodiini (Aphodius elegans Allibert and Sybax impressicollis Boheman).

A note on the type specimen of Termitotrox
Th e type specimen of the type species of the type genus of Termitotrogini, Termitotrox consobrinus Reichensperger 1915, was long considered lost (Paulian 1985), but was recently rediscovered by author MF in a cabinet at the Lund museum, where it was left forgotten many years ago by B-O Landin who was acting as a courier between the Gothenburg museum where it was originally deposited and professor Balthasar in Prague who had borrowed it. Th e specimen is now again available at the Natural History Museum in Gothenburg.

Phylogenetic estimate
Fifty three morphological characters (Table 2) were coded for 24 species from 7 tribes of Aphodiinae (Table 3). Fourteen of the species were coded from specimens, whereas 10 were coded from literature. Character 10 is ordered, while all other multistate characters are unordered. Uniform weights were applied. Th e phylogenetic analyses were performed in Paup* Portable version 4.0b10 for Unix (Swoff ord 2002). Character evolution was explored using MacClade 4.08 (Maddison and Maddison 2005). Heuristic searches produced three equally parsimonious trees with 195 steps. Ensemble consistency index is 0.345 and retention index is 0.589.
Pronotum large and transverse, sides evenly curved, anterior edge bisinuate, posterior edge with a broad median protrusion. Disc of pronotum medially with 5 furrows, converging towards the middle in an hourglass pattern, midfurrow shallower than lateral furrows. Anteromedial disc with distinctly raised tubercle, posteromedial disc and posterolateral sides with slightly lower tubercles. Lateral of the furrows are large elliptical depressions, delineated by furrows.
Hindbody narrower than pronotum, elytra only slightly longer than pronotum plus head.
Elytra posteriorly tapering, apically rounded, evenly, not abruptly (declivous) posteriorly. Each elytron with 6 longitudinal ridges before the lateral edge, elytral interstices alternating in height, elevated but discontinuous, consisting of series of longitudinal setae-bearing tubercles. Even interstices lower, rugose. Epipleura broadly infl exed, posterior two thirds of lateral edge slightly recurved to allow free movement of hindlegs.
Ventral surface. Prosternal process strongly elevated and apex expanded, transversely split anterior to procoxae, hastate posterior to procoxae. Mesosternum narrow with alutaceous and punctured surface. Metasternum triangular widest in front and tapering posteriorly, groove along midline, surface nitid and punctured. Abdomen with 5 segments visible ventrally.
Ventrites apparently fused. Legs short with broad femora and tibiae, tarsi 4-segmented and short. Fore tibiae moderately wide, with fi nely serrated outer edge and one strong apical lateral tooth, tarsus inserted well before apex. Mid and hind tibiae broad with fi nely serrated outer edge and concave apex; each with at least two inconspicuous tibial spurs. Etymology. Th e fi rst specimen of the species was collected in Malaysia. Diagnostic characters. Structures such as the hastate posterior prosternal process, pattern on the pronotum with a strong anteromedial knob as well as bulbous areas medially and laterally on each side at the posterior end of the pronotum, the very long antennae and the strongly tuberculate and rugose surface of the body will easily separate this species from other known Stereomerini.
Description. Length 1.8 mm, width at broadest point 0.9 mm. Chestnut brown, whole body rather densely covered with strong puncture; strongly convex. Morphology as in generic description and as in Figures 1-6.

Phylogenetic estimate
Th e three resulting most parsimonious trees ( Figure 7A-C) diff er only in the internal relationships within the Steremerini genera Australoxenella, Bruneixenus, Pseudostereomera, Stereomera and Adebrattia (as shown in the consensus tree, Figure 7D). Th e rest of the Stereomerini species, namely the new taxon, Danielssonia and Daintreeola appear in the same positions in all three trees ( Figure 7A-C). None of the three confi gurations corresponds very well with Bordat and Howden's (1995) cladogram (we do not compare with Storey and Howden as they only included three representatives of A 100μm B Stereomerini in their 1992 analysis) in which the sister species Adebrattia-Bruneixenus forms a sister group to Danielssonia-Australoxenella-(Stereomera-Pseudostereomera). One reason for this may be that neither the new taxon nor Daintreeola were included in the Bordat and Howden (1995) analysis. Our results rather points to a close relationship between Adebrattia and Pseudostereomera and between Australoxenella and Bruneixenus. Stereomera was placed as a sister group to the Adebrattia-Pseudostereomera in 2 of the 3 trees and as sister group to Australoxenella-Bruneixenus in the third tree.
Concerning the relationships outside Stereomerini, our results point to a close relationship between Stereomerini and the genus Termitaxis, that was formerly included in Stereomerini but is now considered incertae sedis (Bordat and Howden 1995), and the Termitotrogini (Termitotrox). Corythoderini appears monophyletic, as does a core of Rhyparini (if excluding Sybacodes, the position of which has been ambiguous, and Cartwrightia, previously classifi ed in Eupariini and recently transferred to Rhyparini). All representatives of the aberrant tribes form a monophyletic group together, while the more typical Eupariini genera form a grade basal to this group.

History of classifi cation of inquiline Aphodiinae
Th ere are several taxa in Aphodiinae which have inquiline or inquiline-like morphologies (listed in Table 4). Th ey are usually small, have integumental bulbs and ridges, particularly longitudinal dorsal ridges; sometimes contracted body shapes; short, broad extremities, sometimes less so; and occasionally hair tufts or infl ated abdomen. Th eir classifi cation has included some moments of confusion.
In the fi rst tribal division of Aphodiinae (Schmidt 1910), all the termitophilic or termitophile-like species were classifi ed as Rhyparini and Corythoderini (only later were Termitotrox considered to be part of or closely related to Aphodiinae).
Termitoderini was added explicitly as a probable sister group to Corythoderini (Tangelder and Krikken 1982), and Stereomerini in a similar way as the sister group to Rhyparini (Howden and Storey 1992). Th e boundaries between all these tribes have been somewhat confused due to the uncertain status of some genera, and to the citing of all recent or problematic taxa as Termitoderini in Dellacasa's catalogue (Dellacasa 1987-88).
Termitoderini was erected as monotypic, but Dellacasa listed six genera there (including several from Rhyparini), which may have been a mere mistake, or a conscious reclassifi cation with no arguments presented in the series of corrigenda to the cata- logue. Th e genera that actually belonged to Ceratocanthidae were deleted ("dropped") but no further information given on the others (Dellacasa 1991).
A number of more or less aberrant genera have been added to the Rhyparini in recent years (Howden 1995(Howden , 2003Howden and Storey 2000;Makhan 2006;Pittino 2006), with Skelley (2007) making an eff ort to straighten up the classifi cation and listing several papers describing new species or otherwise forwarding knowledge which are not listed here. Krikken (2008a, b) recently reconsidered and described new species in Termitoderini and Termitotrogini.
In addition to these, a number of ant inquilines have always been present in the Eupariini, including the type genus Euparia, and two species which are termite inquilines. At one point, Stebnicka erected the tribe Lomanoxiini for the most aberrant Neotropical ant inquilines (Stebnicka 1999), but that was synonymized into Eupariini (Skelley and Howden 2003). All these Neotropical forms were recently reviewed and pictured in Stebnicka (2009).    Th e classifi cation, and above all the relationships between the inquiline tribes, remains unsure because of the more or less far-reaching morphological specializations, including both obvious strongly derived specializations in habitus and integument, as well as reductions in characters traditionally used for distinguishing aphodiine tribes. Many of these specializations often overlap between the inquiline tribes, thus obscuring relationships between each other and between them and the large aphodiine tribe Eupariini, which is speciose in the tropics and characterized mainly by plesiomorphies visavis most other tribes of Aphodiinae.
Often, the appearance of tibial ridges and position of tibial apical spurs are used as diagnostic characters between Eupariini and Aphodiini. Th ese characters are often useless in inquiline-type aphodiines with highly modifi ed, short tibiae and reduced spurs. Th e abdominal characters traditionally used (fusing of sternites, pygidial furrow, etc.) often unite several but not all inquiline-type taxa with most Eupariini. Perhaps the dissection-requiring mouthparts and aedeagus, which both have provided characters considered important for aphodiine classifi cation (Dellacasa et al. 2001 and elsewhere), but were not studied here, may provide useful information in this respect.

Status of knowledge of biology of alleged inquilines
Strangely enough, it is only some of these inquiline-looking taxa that are actually found in association with social insects. Actually associated with termites are Termitotrogini, Corythoderini, Termitoderini, several small genera in Rhyparini (Termitodius, Termitodiellus, et al.), plus incertae sedis Termitaxis. One genus of similar morphology (Cartwrightia) and many more genera with more normal appearances are actually associated with ants. In several of the inquiline-type taxa, inquiline lifestyles have never been demonstrated; Stereomerini, a major part of Rhyparini (including Rhyparus), plus Notocaulus and Sybacodes (and Sybax). Instead, these are often found in forest litter, Notocaulus and Sybacodes also in dung. Of course, sifting forest litter and various trapping methods (light traps, Malaise traps, fl ight intercept traps) are far more common collecting methods than actually breaking into termite mounds. Th us, these taxa might be inquilines not yet encountered in their true habitat, or they may not live with social insects.
Typologies of social insect inquilines are from Wasmann (1894Wasmann ( , 1903Wasmann ( , 1918, who also studied many of these aphodiines, like Kolbe (1909). Balthasar (1963) summarized the knowledge. Later, Wilson (1971) and Kistner (1979Kistner ( , 1982 provided general discussion of inquline types. Th e basic division is between symphilic inquilines (those actually cared for by the hosts), and inquilines which are synoeketes (ignored by the hosts) or synechthrans (treated with hostility by the hosts). Morphologically, a symphilic lifestyle is often indicated by particular phenomena as possession of specialized tufts of glandular setae (trichomes) or enlarged abdomen (physogastry). Often, symphilic organisms display some kind of general morphological mimicking of their hosts. Synechtrans on the other hand are usually characterized with a "defensive" morphology, often very compact body (sometimes referred to as "limuloid" body shape, after horseshoe crabs (Limulus)), with reinforcing ridges and the capacity to withdraw protruding appendages into folds or grooves.
Among the aphodiines, Corythoderini, Termitoderini and Termitaxis clearly correspond to the "symphilic" type (and are indeed found with termites). Th e clearest examples corresponding to the "synechthran" type are presented by Termitotrogini and Stereomerini (the one found with termites, the other not recognized). Th e Rhyparini are probably best considered synechthran in general morphology, but less so than Termitotrogini and Stereomerini, and also have symphilic-type trichomes. Th e typical Rhyparini are not found with termites, and could be described as weakly synechthran in morphology. Within Rhyparini, only the more or less aberrant taxa are actually found among termites, often being intermediate between the weaker synechthran-type and symphilic-type morphologies.

Th e rationale for classifi cation of inquiline tribes
It is unsatisfactory that this cluster of characters is regarded as an indication of a particular lifestyle in so many taxa where that lifestyle has not been observed. Some of them are commonly collected under circumstances giving no support whatsoever for that assumption. In aphodiine classifi cation, various selections from this cluster are utilized to diagnose several aberrant tribes, while still being assumed to be intimately connected with a particular lifestyle and therefore largely homoplastic, and so not necessarily indicating a relationship between these diff erent tribes while still keeping them together one by one. Th is is a problematic line of reasoning. A Darwinist framework still rests on interpreting similarities as the result of shared ancestry unless in confl ict with other similarities. In this data set, the number of characters is small and the taxon sampling, particularly in various groups of possibly related Eupariini, is limited. For this reason, our results are uncertain and do not form the basis for any confi dence in suggesting a revised tribal classifi cation. Indeed the consistency is relatively low, but not extremely low, and there is a small number of most parsimonious trees, indicating that there is a fairly strong signal in the data. Th e Stereomerini, the core Rhyparini (excluding Sybacodes and Cartwrightia), and the Corythoderini, are all retained as monophyletic groups without confl ict. All the representatives of aberrant taxa form a monophyletic group together.
Again, this is not a strong test, but the possibility that their similarities are to a signifi cant extent due to shared ancestry rather than just shared lifestyle must be considered, especially since this particular lifestyle has in fact not been observed in so many of them. Th e aberrant or termitophilous tribes have been suggested to possibly be related to each other in a few works (Nikolajev 1993;Forshage 2002). Nikolajev suggests they could be related to each other and to Aulonocnemini. Th e phylogenetic status of Aulonocnemini is not well investigated, and it too might prove to be near or in Eupariini.
Th e few representatives of Eupariini included, on the other hand, admittedly covering very little of the diversity of the group but nevertheless representing its three most principal genera, did not form a clade together, but instead a grade basal to a single clade of all the "termitophilic" tribes. Eupariini is historically delineated in contrast with Aphodiini, and most of its diagnostic characters are defi ned in that contrast. As the phylogeny of Aphodiinae remains uncertain, the polarity of most of these characters is not known. Regardless of whether the character states found in Eupariini are plesiomorphic or apomorphic visavis Aphodiini, they are either shared with or clearly plesiomorphic visavis Psammodiini and most minor tribes. In fact it seems impossible to circumscribe Eupariini as a monophyletic group. Eupariini may perhaps be a paraphyletic grade of basal groups of Aphodiini, or it may be a huge possibly monophyletic complex, but in both cases very possibly with Psammodiini and several smaller tribes intermixed.
A more robust phylogenetic analysis including a wealth of eupariine taxa, several of the minor tribes, and many more characters than used here, will highlight the diffi cuties of keeping Eupariini in its present form, and suggest a revised tribal classifi cation. Whether this will result in the synonymy of several tribes into a solid Eupariini sensu lato, or identify monophyletic lineages that will allow for even further splitting into more tribes, will be up to the judgment of that revisor.