Review of the Family Thanerocleridae (Coleoptera: Cleroidea) and the Description of Thanerosus gen. nov. from Cretaceous Amber Using Micro-CT Scanning

Simple Summary The use of new technologies applied to paleontological studies offers more detailed and attractive results each time. We describe a new genus and species of the relative rare beetle family Thanerocleridae (Cleroidea). Thanerosus antiquus gen. and sp. nov. is described as the fourth Cretaceous representative of Thanerocleridae. The set of characters observed in the new species suggests its basal position within the family and its relationship with the ancestral North American Zenodosus sanguineus, the single extant member of the subfamily Zenodosinae. The observance of some characters, including the mouthparts and details on the thorax’s ventral side, was only possible after the 3D reconstruction of the fossil following its X-ray micro-CT scanning. This fact reinforces the effectiveness of this non-destructive technique for evolutionary studies through the use of fossils. We speculate on a wide diversification and distribution of this predatory family originally connected with an ecological community of saproxylic insects in the Cretaceous that were followed by habitat change and specialization of life inside fruiting bodies of fungi. Consequently, the family has not been recorded from the Cenozoic fossil collections yet and currently shows limited diversification. Abstract The predaceous beetle family Thanerocleridae is one of the smallest families of Cleroidea. It comprises only 36 extant species widespread on all continents. Three more species have been described from Cretaceous ambers of Myanmar and France. The fourth fossil representative of Thanerocleridae is described herein. Thanerosus antiquus gen. and sp. nov. is based on one fossil specimen preserved in an amber piece from Upper Cretaceous Kachin amber. The holotype was imaged using an X-ray micro-CT system to obtain high-quality 3D images. A phylogenetic analysis based on 33 morphological characters supports the placement of the new genus at the basal position in a tree of Thanerocleridae, in the vicinity of extant Zenodosus Wolcott and three extinct Mesozoic genera with which the new fossil shares open procoxal and mesocoxal cavities and transverse procoxae. We offer here a key to all extant and extinct genera in the family together with a complete list of all valid thaneroclerid taxa.


Introduction
Thanerocleridae is a small family of cleroid beetles comprising 36 living species in ten extant genera [1][2][3][4] and three fossil genera and species described from Cretaceous ambers [5][6][7]. One fossil described from the Yixian Formation (Lower Cretaceous), Mathesius liaoningensis Kolibáč and Huang (Cleroidea incertae sedis), was denoted a presumptive This study is based on one fossil specimen found in a sample of Kachin amber (Myanmar). The amber from Myanmar, many times referred to as Burmese amber, has a long history of excavation. The active amber mines in Myanmar have increased in the last years [23,24]. This fact forced the use of more defining terms to refer to the general Burmese or Myanmar amber. The Kachin amber is commonly cited as being from an area near Noije Bum peak, around 20 km to the southwest of Danai (=Tanai) in Hukawng Valley, Kachin Province, northern Myanmar ( Figure 1). However, another less known area is also 35 km to the southeast of Hkamti [23,25]. The amber is preserved within Cretaceous flysch-type units. The surrounding matrix consists of fine-grained sedimentary rock, greyish to bluish-green in color, with fine fragments of volcaniclastic sediments [26]. These amber deposits are lower Cenomanian in age (98.79 ± 0.62 Ma), according to current dating using U-Pb zircon crystals obtained from the amber matrix [26]. There exists a current discussion about the true age of this amber, which could be slightly older due to the high roundness of the amber surface and the presence of marine faunal inclusions, which indicates a reworking process before the amber was buried in the surrounding rock matrix. Some other amber sources in Myanmar have recently been dated from ~110 to ~72 Ma [23,24]. As such, the refined understanding of amber deposits in Myanmar highlights the importance of distinguishing sources.

Methods of Observation
The holotype denoted as NIGP180154 is included in a polished oval sample of amber along with one syninclusion of a Diptera specimen (probably Tipulidae). The sample was further ground and polished to better observe the characters of the specimens. It was examined under a Leica MZ95 stereomicroscope (Leica Microsystems, Heerbrugg, Switzerland) and a Leica DME compound microscope (Leica Microsystems, Heerbrugg, Switzerland). Detailed photomicrographs of the fossil were created using an Olympus CX41 compound microscope (Olympus, Tokyo, Japan) and a Leica MS5 stereomicroscope (company Leica Microsystems, Heerbrugg, Switzerland), both equipped with a digital camera sCMEX-20 (Euromex Microscopen, Arnhem, The Netherlands), under incident

Methods of Observation
The holotype denoted as NIGP180154 is included in a polished oval sample of amber along with one syninclusion of a Diptera specimen (probably Tipulidae). The sample was further ground and polished to better observe the characters of the specimens. It was examined under a Leica MZ95 stereomicroscope (Leica Microsystems, Heerbrugg, Switzerland) and a Leica DME compound microscope (Leica Microsystems, Heerbrugg, Switzerland). Detailed photomicrographs of the fossil were created using an Olympus CX41 compound microscope (Olympus, Tokyo, Japan) and a Leica MS5 stereomicroscope (company Leica Microsystems, Heerbrugg, Switzerland), both equipped with a digital camera sCMEX-20 (Euromex Microscopen, Arnhem, The Netherlands), under incident light and using the software ImageFocusAlpha version 1.3.7.12967.20180920 (Euromex Microscopen, Arnhem, The Netherlands), finally merged with the software CombineZP.
The holotype was imaged at the Nanjing Institute of Geology and Palaeontology (NIGP) with a Zeiss Xradia 520Versa X-ray micro-CT system to obtain high-quality 3D images. Considering the comparatively small size of the fossil, a CCD-based 4× objective was used, providing isotropic voxel sizes of 2.3938 µm with the help of geometric magni-fication. Images were generated at an X-ray voltage of 50 kV. Four frames per projection were acquired with an integration time of 3.5 s for 2801 projections over 360 • . Acquired images were rendered and visualized using Avizo 8.1 (Thermo Fisher Scientific, Schwerte, Germany). Video presentations are included in Supplemental Video Files S1 and S2.
The final images were edited with Adobe CS6. All relevant structures were measured from the digitized images. The specific terminology for characters follows that of Kolibáč & Leschen [1]. We followed the family-group classification of Gimmel et al. [12]. The nomenclatural acts established herein are registered under ZooBank LSID: 9BC3609F-64F6-4AAB-AC8C-7683508BF73E.

Phylogenetic Analysis
An analysis of 33 morphological characters was used to find an approximate position of the fossil within Thanerocleridae and reveal the extant representatives' phylogeny. The following extant species of Cleroidea were used for a character analysis as outgroups: Acanthocnemus nigricans Hope (Acanthocnemidae), Tenebroides mauritanicus Linnaeus (Trogossitidae), Tilloidea transversalis Charpentier (Cleridae: Tillinae), and Clerus mutillarius Fabricius (Cleridae: Clerinae). Trees were rooted with Acanthocnemus nigricans, which was revealed as the basal taxon in the specialized molecular phylogeny of Cleroidea by Gimmel et al. [12]. The data matrix of morphological characters was assembled in WinClada 1.00.08 [27]. TNT 1.5 [28,29] was used for parsimony analysis using the implicit enumeration strategy. The characters were unordered, and all were switched as non-additive (see Supplemental Table S1 for a list of the characters and the matrix). The unknown character states were denoted by the question mark (?), the lacking characters by the dash (-). Two analyses were conducted: (1) with character state 7 (1) (mandible bidentate) exactly as mentioned in particular descriptions of Mesozenodosus insularis Tihelka et al., 2020 and Cretozenodosus fossilis Cai & Huang, 2018; (2) with modified character state 7 (0) (mandible unidentate) and state unknown 7 (?), respectively. Parsimony analysis was conducted with implied weighting (K = 12; see Goloboff et al. [30] and Smith [31]). Tree support was measured as Bremer support (TBR, 1000 steps) conducted in TNT 1.5. Etymology. The generic name is derived from the prefix 'Thanero-' following the family name, and '-sus', termination of the most basal extant representative of the family (Zenodosus Wolcott). It is masculine in gender. The genus is registered under Zoo-Bank LSID: 84DA2BE8-C471-44F9-A4AF-FEE0C6EE6F2E.
Remarks. Thanerosus gen. nov. can be referred to Thanerocleridae based on the general body shape, head prognathous and its base as wide as pronotum; prothorax with complete lateral carina, the base of pronotum bordered; all coxae narrowly separated; metacoxae extending laterally to meet elytra, metanepisternum elongate; protarsomeres 1-4 wide with tarsi compact, tarsal claws simple; abdomen with five ventrites [5,9,13]. The family has some other apomorphic characters that are difficult to observe in fossil specimens, even after 3D reconstruction, such as tegmen without the median strut and lateral struts or four malpighian glands [9].
The recent genus Zenodosus and all three fossil thaneroclerids described to date are placed in Zenodosinae based on the procoxal cavities open and the procoxae transverse. By contrast, species in Thaneroclerinae possess the procoxal cavities externally closed and procoxae more or less spherical [9]. The shape of both maxillary and labial terminal palpomeres in Thanerosus gen. nov. also corresponds with zenodosines, whereas depressions along notosternal are known only in the tribe Isoclerini. Description. Body length dorsally 2.24 mm; length of elytron 1.41 mm, the width of elytron at base 0.85 mm, a maximum width of pronotum 0.7 mm, length of pronotum 0.66 mm. Body 1.59 times as long as wide.
Legs: slender and long, setose. Trochanters triangular. Femora widest near middle, weakly grooved for the reception of tibiae. Tibiae is widest apically, each with two short

Thanerosus antiquus gen. and sp. nov. Peris & Kolibáč Figures 2-4
Holotype. Holotype NIGP180154, adult specimens, sex unknown. The type specimen is deposited in the Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences, Nanjing, China. The holotype lost the apical section of the left antenna before being embedded in resin.
Etymology. The specific epithet is designated after the Latin antiquus, meaning old or ancient. It is masculine in gender. The species is registered under Zoo-Bank LSID: 13AB22E2-8519-4C5A-BF09-0926A7889884.
Diagnosis. As for the genus.
Description. Body length dorsally 2.24 mm; length of elytron 1.41 mm, the width of elytron at base 0.85 mm, a maximum width of pronotum 0.7 mm, length of pronotum 0.66 mm. Body 1.59 times as long as wide.
Coloration and structure: body elongate and subcylindrical ( Figure 2A); winged; generally black and dark brown ( Figure 4A). The body surface is densely clothed with long, erect hairs, more conspicuous dorsally. Body densely punctate, interspaces smaller than the diameter of punctures in the head and pro-and mesosternum, interspaces larger than the diameter of punctures in pronotum and elytra. Metaventrite, abdominal ventrites and legs without sculpture. Elytral surface with punctation regular but disordered. Pubescence of ventral surface scarce, conspicuous only in legs.

Discussion and Conclusions
4.1. Classification of Thanerosus antiquus gen. and sp. nov.
Thanerocleridae had been originally classified within the Cleridae until Kolibáč [9] raised the taxon to family rank, supporting the monophyly of the family in a list of characters that differ from Cleridae, most notably by having the minutely emarginate eyes, expanded protarsomeres 1-4, mostly five visible abdominal ventrites and aedeagus without median and lateral struts. Thanerosus antiquus gen. and sp. nov. are placed in Zenodosinae, as the other three fossil species of the family, based on the procoxal cavities open and procoxae transverse ( Figure 2C). The new genus differs from the other thanerocleride genera in the scutellum quadrate. It differs from Zenodosus in the loose but distinct antennal club and regular elytral sculpture, while the club is weak and sculpture irregular in Zenodosus [9]. Mesozenodosus insularis Tihelka, Cheng, Huang, Perrichot & Cai, the most ancient fossil of the family from French amber [7], has the eyes coarsely setose with interfacetal setae, distinct depression in apical antennomere, pronotom with lateral carina weakly developed, the base of pronotum narrowed, depression along notosternal suture absent and scutellum strongly transverse. According to the original description, the mandibles are bidentate in Mesozenodosus, but such a feature is unknown in the all clerid lineage and the figure of the mandible by Tihelka et al. [7] (p. 393, Figure 3C) shows the unidentate apex of the mandible. By contrast, Thanerosus antiquus gen. and sp. nov. has the eye glabrous, depression in the apical antennomere absent, lateral pronotal carina distinct, pronotum only slightly constricted towards the base, depression along notosternal suture present and scutellum quadrate ( Figure 2). The new species is also different from both previously described species from the Kachin amber. Archaeozenodosus bellus Yu & Kolibáč is more than two times bigger, clothed with short setae, pronotum widened anteriorly, depression along notosternal suture absent, procoxal cavity oval (but not spherical), scutellum transverse and broadly rounded apically, and elytral base about as broad as the base of pronotum [5]; Thanerosus antiquus gen. and sp. nov. is much smaller in length, clothed dorsally with long setae, pronotum widest at one-third from its base, depression along notosternal suture present, procoxal cavity transverse, scutellum quadrate and apically arrowhead-shaped, and elytral base distinctly wider than the base of pronotum. Cretozenodosus fossilis Cai & Huang lacks depression along notosternal suture and base of pronotum half the width of elytral bases [6] while depression along notosternal suture is present and the base of the pronotum is only slightly narrower than elytral bases in Thanerosus antiquus gen. and sp. nov. Moreover, the description of Cretozenodosus mentions bidentate mandibles; however, this structure is not figured and is highly improbable in Thanerocleridae.
Sister relation of the family Thanerocleridae and Cleridae is well-supported in our analysis (Bremer support > 1), as similarly occurred in previous studies [12]. Kolibáč [9] proposed Zenodosus as the most ancestral member of the family. A proof supporting the idea of Zenodosinae as the basal thaneroclerid group is that all the fossils described in the family so far, always from Cretaceous ambers, are different genera of the same subfamily. All four fossil species possess ancestral characters such as externally open procoxal cavities (12-0) and transverse or oval procoxae . Up to now, Melyridae, Mauroniscidae, Prionoceridae, Lophocateridae, Trogossitidae and Cleridae are the six cleroid families found in the Middle Jurassic of northeastern China [32]. Fossil descriptions and the calibrated molecular clock suggested that the clerid lineage was fully developed and well-differentiated from other cleroid lineages in the middle Jurassic [15,19]. In the molecular analyses, Gimmel et al. [12] showed Thanerocleridae as a sister to Cleridae, while Kolibáč et al. [19] considered it a sister to Chaetosomatidae and estimated the split event of both clades using Bayesian tip dating analysis to the Lower Jurassic. Finally, Cai et al. [18] estimated the Cleridae-Thanerocleridae split approximately in the mid-Jurassic (Chaetosomatidae was not included in the analysis). Although Jurassic thaneroclerid fossils are unknown, we already know that Thanerocleridae was well diversified and distributed by the mid-Cretaceous. Strange enough, the family has not been recorded from the big collections of Cenozoic ambers, revealing a possible loss of diversity, paleogeographic constrictions or different natural history of the thaneroclerids since the Upper Cretaceous. The latter reason may be related to the specialization of the thaneroclerids on life in fungi. At the same time, ancestral members of the clerid lineage are predatory and live under bark or in galleries of wood-boring insects similar to Zenodosus sanguineus, the major part of extant thaneroclerids (the species-richest genera Isoclerus and Neoclerus) hunt for prey in a tree or freely growing fungi. Probably, the four extinct thaneroclerids, including Thanerosus antiquus gen. and sp. nov., found in European and Asian Cretaceous ambers, had the ancestral biological pattern connected with the ecological community of saproxylic insects, like many other groups of amber bearing beetles from the Cretaceous ambers [33]. It is possible that Cretaceous thaneroclerids were living under bark or on logs where the likelihood of sinking into the resin was higher than for species dwelling in fungal fruiting bodies [34]. Although characters in the cleroid mouthparts together with molecular studies suggest that primitive cleroids were mainly fungus-feeding [12,19], as similarly suggested for other groups in Coleoptera [35], and predatory shifts and flower-feeding in Cleroidea occurred presumably later, predation on insects living in fungi is probably derived in the clerid lineage from common hunting on bark surface or inside galleries of wood borer larvae as we can observe in the ancestral Zenodosus sanguineus. Zenodosus and Thanerosus share the open mesocoxal cavities (14-0) (mesepimeron touches mesocoxa) while all other thaneroclerids but Ababa Casey (character unknown in Mesozenodosus) have the cavities closed by projections of meso-and metaventrite which character state is considered apomorphic (14-1). That is the reason for the sister relation between Zenodosus and Thanerosus shown in Figure 4A. However, the analysis with unresolved relations among five 'zenodosine' taxa ( Figure 4B) is more realistic because the relations figured in the previous analysis are, excepting mesocoxal cavities open/closed, based on unknown character states (?) in the fossil taxa and dubious observation of bidentate mandible in Cretozenodosus and Mesozenodosus (see above). No single taxon in the whole clerid lineage (Chaetosomatidae, former Metaxinidae, Thanerocleridae, Cleridae) has the bidentate mandible with two apical teeth situated side by side (or in the horizontal axis) as in other cleroids and the major part of cucujoid beetles. Moreover, as already noted above, the figure of the Mesozenodosus mandible shows the unidentate mandible [7].
Up to now, Melyridae, Mauroniscidae, Prionoceridae, Lophocateridae, Trogossitidae and Cleridae are the six cleroid families found in the Middle Jurassic of northeastern China [32]. Fossil descriptions and the calibrated molecular clock suggested that the clerid lineage was fully developed and well-differentiated from other cleroid lineages in the middle Jurassic [15,19]. In the molecular analyses, Gimmel et al. [12] showed Thanerocleridae as a sister to Cleridae, while Kolibáč et al. [19] considered it a sister to Chaetosomatidae and estimated the split event of both clades using Bayesian tip dating analysis to the Lower Jurassic. Finally, Cai et al. [18] estimated the Cleridae-Thanerocleridae split approximately in the mid-Jurassic (Chaetosomatidae was not included in the analysis). Although Jurassic thaneroclerid fossils are unknown, we already know that Thanerocleridae was well diversified and distributed by the mid-Cretaceous. Strange enough, the family has not been recorded from the big collections of Cenozoic ambers, revealing a possible loss of diversity, paleogeographic constrictions or different natural history of the thaneroclerids since the Upper Cretaceous. The latter reason may be related to the specialization of the thaneroclerids on life in fungi. At the same time, ancestral members of the clerid lineage are predatory and live under bark or in galleries of wood-boring insects similar to Zenodosus sanguineus, the major part of extant thaneroclerids (the species-richest genera Isoclerus and Neoclerus) hunt for prey in a tree or freely growing fungi. Probably, the four extinct thaneroclerids, including Thanerosus antiquus gen. and sp. nov., found in European and Asian Cretaceous ambers, had the ancestral biological pattern connected with the ecological community of saproxylic insects, like many other groups of amber bearing beetles from the Cretaceous ambers [33]. It is possible that Cretaceous thaneroclerids were living under bark or on logs where the likelihood of sinking into the resin was higher than for species dwelling in fungal fruiting bodies [34]. Although characters in the cleroid mouthparts together with molecular studies suggest that primitive cleroids were mainly fungus-feeding [12,19], as similarly suggested for other groups in Coleoptera [35], and predatory shifts and flower-feeding in Cleroidea occurred presumably later, predation on insects living in fungi is probably derived in the clerid lineage from common hunting on bark surface or inside galleries of wood borer larvae as we can observe in the ancestral Zenodosus sanguineus.

Key to the Extinct and Extant Genera of Thanerocleridae
Kolibáč [9] proposed a key to classify the higher taxa in the family. An improved version of the key was needed after the recent description of additional extant and extinct taxa in the family. The genus Cleridopsis Champion (type of genus: C. latimanus Champion by monotypy) from Central America (Guatemala, Panama) was described within Cryptophagidae [36] (pp. 60, 94-95); however, the original description and illustrations [36] (pp. 94-95, Pl. III: Figures 10, 10a) perfectly determine its familial membership. Corporaal [37] had not noted this fact, but Crowson [38] (p. 310) recognized it and rightly affiliated Cleridopsis with Thanerocleridae. Kolibáč [9] (p. 338, footnote) did not include the genus in his family revision because he had not studied the two only known C. latimanus specimens. A synonymization of Cleridopsis with Ababa from the same geographic region has been recently made by Opitz [22], who also restituted the generic rank for Ababa and Parathaneroclerus, although the taxa are extremely similar and rather congeneric. Therefore, Cleridopsis is not included in the following key, and both latter taxa are treated as genera.