Pupae of the mega-diverse rove beetle tribe Staphylinini (Coleoptera, Staphylinidae): their traits and systematic significance

Abstract This paper is the first comprehensive work on the pupae of Staphylinidae. It is the first-ever attempt to employ the morphological characters of these pupae in phylogenetic analysis. The study shows that the external structures of Staphylinini pupae could be a useful, alternative source for assessing the relationships of certain taxa within a tribe. The paper also includes an illustrated key to the identification of pupae at the subtribe and generic levels (Abemus, Acylophorus, Astrapaeus, Atanygnathus, Bisnius, Creophilus, Emus, Erichsonius, Gabrius, Hesperus, Heterothops, Neobisnius, Ocypus, Ontholestes, Philonthus, Quedius, Quedionuchus, Rabigus, Staphylinus, and Tasgius) of the tribe Staphylinini, found in Europe. Based on current knowledge of the morphology of pupal stages of Staphylinini species, eight morphological pupal types are presented: Acylophorus, Astrapaeus, Atanygnathus, Erichsonius, Heterothops, Philonthus, Quedius and Staphylinus. The paper also comments on pupal habitat, phenology and morphology in the context of antipredator and environmental adaptations.


Introduction
Rove beetles (Staphylinidae) are the largest family of organisms and dominate all ground-based cryptic microhabitats in every habitable landscape of the globe. Among insects, hyper-diverse families like rove beetles are the most difficult to analyse phyloge-netically. They display an evolutionary radiation that took place 150-200 million years ago, since the fossil record indicates a notable diversity and abundance of Staphylinidae from at least the Late Jurassic . The overall pattern of rove beetle evolution is not well understood, and the phylogenetic system of this family is incomplete. Staphylinini, one of the largest tribes of rove beetles is an exception, however. It has recently been the focus of several phylogenetic studies involving adult morphological data, larval morphological data, the integration of adult morphologies of extinct and extant taxa, DNA sequences, as well as the integration of DNA sequences with adult morphology (e.g., Chani-Posse et al. 2018). Here we attempt to contribute a new set of data to the phylogeny of Staphylinini, relating to their pupae.
The difficulties of collecting and identifying pupae are due to their cryptic biology, and the need to link their morphology with the respective adults explains why little is known about the pupae and, in particular, why they have not been used for phylogenetic purposes.
In comparison with larvae or imagines, the pupae of Staphylinini are far poorer in morphological characters of diagnostic significance. The identification of pupae to species level is based on a small number of morphological characters revealed by morphometric analysis: the size and proportions of various body parts, the structure of the last abdominal segment, the structure and number of cuticular processes (including their range of variability), body microstructure and spiracular structure. Since rove beetles usually pupate in or near the habitats of their adults and larvae, the pupal biotope also provides a useful clue to their identification. Ecological data of this kind are especially helpful when comparing closely related species living in different habitats.
In view of the above, the idea arose to compile a summary of existing knowledge of Staphylinini pupae. This is the first such comprehensive review worldwide dealing with Staphylinidae pupae. The main body of the paper is an illustrated key to assist the identification of known pupae of European Staphylinini at the subtribal and generic level. We also attempt to shed light on the potential importance of pupal characters in constructing phylogenetic hypotheses. This is the first attempt at applying the morphological characters of pupae to phylogenetic analysis and testing the hitherto accepted systematics at the generic level. Table 1. Known pupal stages of Staphylinini species. Symbols: #-species considered in phylogenetic analysis. State of knowledge on morphology: very good-detailed and well-illustrated, descriptions reliable for diagnostics and sufficient for phylogenetic analysis; good-detailed enough descriptions with sufficient illustrations, reliable for diagnostics but not fully reliable for phylogenetic analysis; fair-moderately informative descriptions, maybe with sketchy illustrations (sometimes without), can be used for diagnostics but not for phylogenetic analysis; poor-hardly informative descriptions, mostly without any illustrations or no description with sketchy illustration, can be ambiguous even for diagnostic purposes. fair Byrne (1993) P. sanguinolentus (Grav.) poor Xambeu (1907Xambeu ( , 1910) P. sericans Grav.

Description of pupal morphology and key to subtribes and genera of the tribe Staphylinini
The diagnostic characters given in this paper were established generally on the basis of current knowledge of the pupal stage in Staphylinini. The key covers 8 subtribes and 20 genera (highlighted in Table 1 by an asterisk) of pupae found in Europe. Most of the data and all the drawings have been taken from papers by the present authors, published between 1996-2014. Some information relating to Abemus and in part to: Creophilus, Ocypus, Philonthus, Platydracus, Quedius and Staphylinus, has been taken from papers by other authors (e.g., Szujecki 1965;Dajoz and Caussanel 1968;Orth et al. 1976;Boháč 1982Boháč , 1987Schmidt 1994b;Moseley et al. 2006;Vorst and Heijerman 2015). The paper also includes new information on the pupa of Emus hirtus (L.) which was hitherto unknown, and photographs and notes supplementing existing descriptions of the pupae of nine genera, represented by species: Acylophorus wagenschieberi Kies., Creophilus maxillosus (L.), Gabrius appendiculatus Sharp, Ocypus fulvipennis (Er.), Quedius microps (Grav.), Rabigus tenuis (Fabr.) and Staphylinus erythropterus L. The photographs were taken with an Olympus DP72 digital camera mounted to an Olympus SZX16 compound microscope (Fig. 7) or with a VEGA3 TESCAN SEM (Figs 2a,2b,6b,13a,14a), and corrected using CorelDRAW Graphics Suite X6. Material that was here examined for the first time includes one pupa of C. maxillosus (male) and one exuvium of E. hirtus, obtained from the collection of the Zoological Museum of the University of Copenhagen, Denmark (NHMD, the Natural History Museum of Denmark). The pupae of these four species, as well as others previously described by the authors, are deposited in the collection of the Department of Zoology, Marie Curie Skłodowska University, Lublin, Poland.

Phylogenetic analysis
The phylogenetic analysis was carried out at the generic level. A data matrix was assembled in Nexus Data Editor for Windows v. 0.5.0 (Page 2001) that included 29 morphological characters of the well-known pupae from 22 species as the ingroup belonging to 20 genera of Staphylinini from 8 subtribes (Table 4). Each genus is represented by one species, except for Quedius (three species from the subgenera Distichalius, Raphirus, Microsaurus). As the pupae of the species from these three subgenera differ in certain morphological characters, they have been included in the data matrix. Some genera (Belonuchus, Cafius, Emus, Remus, Smilax, Triacrus, Hadropinus) have not been included because the available morphological data on their pupae are too fragmentary and superficial, and therefore deemed unreliable. The pupa of Hypnogyra angularis (Ganglbauer, 1895) from the tribe Xantholinini is added as an outgroup to the tribe Staphylinini Pietrykowska 2005a, Pietrykowska-Tudruj andStaniec 2006a, unpublished data). Inapplicable characters are assigned a gap value ('−') and treated as equivalent to missing data ('?'). The matrix was analysed in TNT (Goloboff et al. 2008) under settings as follows: the 'traditional search' option for the parsimony  (17) analysis -1000 replicates with tree bisection reconnection (TBR) branch swapping and saving 1000 trees per replicate, zero-length branches collapsed, all characters were treated as unordered and equally weighted.

Diagnostic description of pupae of the tribe Staphylinini
Pupa obtect. Body clearly slender, almost cylindrical and weakly sclerotised (e.g., Neobisnius), or moderately elongate, slender and moderately sclerotised (e.g., Erichsonius, Gabrius, Heterothops, some species of Philonthus), or moderately stocky and well sclerotised (e.g., Astrapaeus, Quedius) to extremely stocky and strongly sclerotised (e.g., Atanygnathus). Colour: almost white or pale yellow shortly after pupation; from dark yellow to reddish brown a few days after pupation; usually almost black just prior to emergence of imago. Head directed ventrally towards thorax, without any setiform projection or spines, rarely with a few protuberances. Labrum usually V-shaped, exceptionally Ushaped, with short, longitudinal groove running from its anterior margin. Mandibles elongate, usually pointing posteriorly, falcate or almost straight. Maxillae usually moderately long. Antennae curved, rest on knees of fore and mid legs; apex usually protruding beyond knee of mid tibia. Scutiform pronotum widest at the base, usually about as wide as long with 6-32 setiform projections, or a pair of micro spines or 8-26 protuberances, sometimes with no structures. Mesonotum separated from pronotum by a furrow, distinctly wider than long. Metanotum narrower than mesonotum with deeply bisinuate anterior margin. Elytra shortened. Wings protruding to ventral side. Apex of wings protruding at most beyond posterior margin of I (morphologically III), clearly visible abdominal segment. Tibiae and tarsi directed obliquely towards body middle. All tibiae, or only some of them with pointed protuberances. Hind tarsi at most reaching midpoint of V (morphologically VII), clearly visible abdominal segment.
Abdomen with 9 somewhat flattened tergites and 7 convex sternites visible. Abdominal tergite I wider than others and about twice as long as tergite II. Abdominal shape of three kinds: arcuate, with parallel sides or funnel-shaped. Sides of abdomen with: spines on segments II-VIII or II-VII, or setiform projections on segments III-VIII or VII-VIII. Rarely abdomen without any lateral cuticular projections. Last segment usually strongly protruding into two terminal, elongated prolongations, sometimes weakly protruding into two triangular prolongations, exceptionally without prolongations. Terminal sternite with well-marked sexual dimorphism. Gonotheca in female double, in male single. In female pupae, terminal sternite often with a pair of prolongations. Abdominal tergites I-IV with tuberculate, functional spiracles, the first pair usually situated more laterally, most often larger and protruding farther than the others; tergites V-VIII with externally visible, but apparently atrophied spiracles.

Comparison
The following crucial characters distinguish the pupae of the tribe Staphylinini from the tribe Xantholinini within the subfamily Staphylininae for which the pupae are known: abdominal segments divided laterally into ventral and dorsal sclerites (not grown into uniform rings); body with setiform projections, spines or protuberances, apart from the genus Astrapaeus which has no cuticular processes Pietrykowska 2005a, Pietrykowska-Tudruj andStaniec 2006a). The combination of characters distinguishing the pupae of Staphylinini within the family Staphylinidae, i.e., the subfamilies Aleocharinae, Omaliinae, Oxyporinae, Oxytelinae, Paederinae, Steninae and Tachyporinae for which the pupae are known, includes: i. exarate pupa; ii. no projections whatsoever on head; iii. short labium; iv. lack of short setae on dorsal and/or ventral part of abdominal sclerites; v. lack of setae on hind margin of prothorax.

Phylogenetic analysis
The parsimony analysis retrieved 100 most parsimonious trees. The 50% majority rules consensus tree showed the following: i) separation of Astrapaeus from a clade of all other Staphylinini; ii) a well-supported clade of Erichsonius+Heterothops+Atanygnath us+ Acylophorus; iii) a well-supported clade of Staphylinini propria represented here by the subtribes Philonthina and Staphylinina (Fig. 35).

Pupal morphology as an adaptation to environment
The occurrence of obtect pupae in the subfamily Staphylininae (including the tribe Staphylinini) is exceptional compared to other rove beetles and the majority of Coleoptera. The obtect pupa type with a compact body and usually heavily sclerotised cuticle  appears to be far more resistant to negative impacts like attack from predators, parasitoids or mechanical damage. Therefore, an obtect pupa is probably a defensive adaptation. Presence of a tough cuticle in a pupa reduces the need for fully-grown larvae to construct a protective pupal cocoon, such as has been reported in a few Staphylinini species (G. splendidulus (Grav.), O. murinus (L.), R. tenuis) (Staniec 2004b;Pietrykowska-Tudruj and Staniec 2007;Staniec and Pietrykowska-Tudruj 2008c). Pupal cocoons are often encountered in representatives of the Aleocharinae, which have free, delicate pupae with weakly sclerotised cuticle (Frank and Thomas 1984;Staniec et al. 2010;Zagaja et al. 2014).
Another form of defensive adaptation is cuticular processes, which occur on various parts of the body (pronotum, abdominal segments) in almost all known pupae of the Staphylinini. They can take several forms: flexible, often arcuate setiform projections (e.g., Gabrius, Philonthus, Staphylinus), stiff spines (e.g., Quedius, Erichsonius, Heterothops) and, more rarely, convex protuberances (e.g., Acylophorus). These structures, besides having a defensive function (Hinton 1955), appear to minimise direct contact between the pupal body surface and the surrounding substrate (e.g., soil, leaf litter, plant remains, decaying wood), which probably allows the spiracles to function more efficiently. The number, length and shape of setiform projections or spines are also usually associated with the degree of cuticular sclerotisation and pupal body size. Their presence is particularly important for pupae with a weakly sclerotised cuticle, less resistant to damage. For this reason, such pupae usually bear numerous long, flexible, arcuate (especially on the pronotum) setiform projections (e.g., G. splendidulus, N. villosulus (Steph.) and most Philonthus species). In contrast, pupae with highly sclerotised cuticle bear far fewer such structures on the pronotum (e.g., Quedius sp. and A. terminalis), and in a few cases, the entire body surface is devoid of them (e.g., A. wagenschieberi and A. ulmi).
The pupae of some species from very wet habitats (Atanygnathus terminalis, Acylophorus wagenschieberi) exhibit special adaptations to their environment in the structure and localisation of spiracles (Staniec 2005a, b). They pupate among unsubmerged peat mosses (Sphagnum). However, water levels frequently vary in the peat-bogs they inhabit. The first pair of spiracles is thus especially large and protrudes strongly from the body outline, which facilitates gas exchange even at high levels of moisture. In the case of A. terminalis, the next three active spiracles additionally bear peculiar phylliform lobes. These probably serve to accumulate a supply of air in case the peat mosses are suddenly flooded or act as an additional respiratory surface (Staniec 2005b).

Some comments on pupation
In the natural environment, larvae and adults of Staphylinini mostly live in the same microhabitats; pupation usually takes place there, too (Table 3). Only in the case of the above-mentioned species from wet microhabitats subject to flooding were pupae found in distinctly drier locations than those inhabited by mobile adults or larvae. These were usually unsubmerged layers of peat mosses, directly adjacent to higher-lying areas of bogs, 2-3 m from small bodies of standing water (Staniec 2005a, b). In the field, pupae of Staphylinini were found from spring to autumn (IV-X), although the pupae of most Quedius species were found in spring (IV-V). Among Quedius, only species confined to special microhabitats, such as tree hollows, crevices under tree bark or the vicinity of hornets' nests, pupated during the summer (VII) or autumn (X). These phenological observations are in broad agreement with laboratory breeding data. In the laboratory, however, specimens from the same species pupated earlier (mainly V-VI) than in nature (Table 3), probably because of the more stable and warmer conditions (T = 20-24 °C) there, where the pupal stage lasted from 7 to 13 days, depending on the species. Larger species, e.g., A. ulmi and S. erythropterus, took distinctly longer to pupate than smaller ones, e.g., R. tenuis, P. nigrita (Table 3) (Staniec and Pietrykowska-Tudruj 2008a, c;Pietrykowska-Tudruj and Staniec 2012;Pietrykowska-Tudruj et al. 2014b).
Since subtribal classification within the non-Staphylinini propria has undergone substantial changes in recent years. We discuss below the phylogenetic potential of the external pupal structures of some taxa in the light of such taxonomic revolutions.

Pupa of Astrapaeus
There are practically no cuticular structures on the pupa of A. ulmi. The cuticular surface is devoid of any visible processes or protuberances (not including those on the legs of all pupae of the Staphylinini and the tiny accessories on terminal prolongations), which makes this species unique among the known pupae of the Staphylinini. Phylogenetic research based on adult and larval morphology, including fossil taxa, suggests that the monotypic genus Astrapaeus is not related to the subtribe Quediina (its traditional placement) but is a member of a rather isolated and basal lineage within Staphylinini (Solodovnikov and Schomann 2009;Solodovnikov 2012;Brunke and Solodovnikov 2013;Solodovnikov et al. 2013;Pietrykowska-Tudruj et al. 2014b). Based on molecular and morphological evidence, Astrapaeus is now included in subtribe Cyrtoquediina Brunke et al. 2016. Within Cyrtoquediina, a subtribe comprising species with mostly isolated distributions in the Neotropical, Oriental or Palaearctic regions (e.g., Bolitogyrus, Cyrtoquedius, Parisanopus, Sedolinus), only Astrapaeus occurs in and is restricted to Europe. The pupal characters of Astrapaeus support the isolated position of A. ulmi (and potentially other Cyrtoquediina) within the tribe, and outside the Quediina sensu Brunke et al. 2016. However, given the lack of data on the pupae of other members of the subtribe Cyrtoquediina, it is it is difficult to tell which morphological features of Astrapaeus are representative of the subtribe versus just genus level.

Pupa of Erichsonius
Within Erichsonius, a genus including more than 160 species distributed over almost all the world, the pupal stage is known for just three: the Nearctic E. alumnus Frank and E. pusio (Horn) (Schmidt 1996) and the Palearctic E. cinerascens (Grav.) (Pie-trykowska-Tudruj and Staniec 2006c). Until the end of the 20 th century, this genus was placed the subtribe Philonthina. But many recent phylogenetic analyses of adults utilising morphological and molecular data have indicated that the original placement of Erichsonius was incorrect (e.g., Brunke and Solodovnikov 2013;Chani-Posse 2013;Brunke et al. 2016). Initially, the genus was withdrawn from Philonthina and allocated to the Anisolinina grade within Staphylinini propria. However, the latest analyses show that Erichsonius is monophyletic and forms a separate subtribe Erichsoniina (Brunke et al. 2016, Chani-Posse et al. 2018. The pupa of Erichsonius possesses a series of characters clearly distinguishing it from species classified among Staphylinini propria. They are: i) a lack of setiform projections on the pronotum; ii) the presence of protuberances on the pronotum; iii) cuticular processes on the abdominal segments in the form of spines. At the same time, these characters are shared with species of four genera of non-Staphylinini propria, i.e., Atanygnathus and Heterothops (all characters), Acylophorus (characters i and ii) and Quedius (characters i and iii). The results of our analyses suggest Erichsonius is distinguished from all other non-Staphylinini propria with known pupae by the number of protuberances on the pronotum. Pupae of Erichsonius have few protuberances (10 at most), whereas they are more numerous (more than 10) on the pupae of other taxa. Since the pupal stage is unknown in many other genera of Staphylinini and Erichsonius species, it is hard to assess the extent to which the number of protuberances is consistent within and unique to the genus. Given the present state of knowledge of pupae, we can regard it as unique to Erichsonius, and therefore evidence in favour of the recently erected subtribe Erichsoniina (Brunke et al. 2016).

Pupa of Heterothops
Within Heterothops, a globally distributed genus with 149 described species, the pupal stage is known only in H. praevius (Herman 2001;Pietrykowska-Tudruj and Staniec 2006c). This poorly defined genus was moved from the conventional subtribe Quediina and initially included in the large lineage Tanygnathinina sensu Solodovnikov and Schomann 2009; later it was placed in the subtribe Amblyopinina, containing fauna mainly from the Neotropical and Australian regions (Solodovnikov and Schomann 2009;Assing and Schülke 2012;Solodovnikov 2012).
Our analyses have demonstrated that the pupa of Heterothops has many characters in common with Atanygnathus. They are: processes on the head and pronotum, spines on abdominal segments II-VII, broad elytra, short hind leg tibiae (not reaching the lateral margin of the body), protuberances on the mid and hind legs, and long antennae. There are not many characters (not present in Staphylinini propria) shared between Heterothops (H) and Quedius (Q) (spines on abdominal segments and protuberances on the mid and hind tibiae), whereas there are many differences: head size (in proportion to the rest of the body) (small -H, large -Q), protuberances on the head and pronotum (present -H, absent -Q), antenna length (long -H, short -Q), width of elytra (wide -H, narrow -Q). In the light of current knowledge of Staphylinini pupae, one can state unequivocally that the morphology of Heterothops pupae supports the separation of this genus from the subtribe Quediina. There are several recent studies based on adult characters, or in combination with DNA that have confirmed its placement within the subtribe Amblyopinina (e.g., Brunke et al. 2016;Chani-Posse et al. 2018;Brunke et al. 2019). Among the 17 genera forming this group, only the pupa of Heterothops, the single taxon in this group which occurs beyond the southern hemisphere, is known (Brunke et al. 2016).
The present study has shown that the external structures of Staphylinini pupae could be a useful, alternative source of evidence for resolving the relationships of some higher taxa within the tribe. However, much more descriptive work is needed -mainly expanding the data matrix to include new species/genera and compiling new morphological data. Unfortunately, the pupae of many species of phylogenetic interest will probably remain unknown owing to the great difficulties with their collection and identification.