The earliest evidence of Omophroninae (Coleoptera: Carabidae) from mid-Cretaceous Kachin amber and the description of a larva of a new genus

) The earliest evidence of Omophroninae (Coleoptera: Carabidae) from mid-Cretaceous Kachin amber and the description of a larva of a new genus. Arthropod Systematics


Introduction
The fossil record of the megadiverse Carabidae (ca. 40.000 described spp.; e.g., Raupach et al. 2022) is relatively rich, with impression fossils dating back to the Triassic (Ponomarenko 1977;Liu et al. 2023). However, larvae are extremely scarce in any deposits, with a total of only five specimens from the Mesozoic (Ponomaren-ko 1985;Makarov 1995;Prokin et al. 2013;Zhao et al. 2019), four of them preserved as impression fossils from the Late Triassic and Jurassic, and one larva of Migadopinae embedded in mid-Cretaceous Burmite from Myanmar (Liu et al. 2023).
The discovery of a conspicuous larva embedded in Burmite, apparently belonging to Carabidae but with an unusual morphology, inspired us to carry out the present study. The single well-preserved specimen was examined using light microscopy and synchrotron µ-computed tomography (SRμCT). The identification of fossil beetle larvae can be an enormous challenge (e.g. Batelka et al., 2019; see also misidentification in Zippel et al. 2022). However, in this case an identification as a species of the small and very distinctive ground beetle subfamily Omophroninae appeared as a well-founded working hypothesis, suggested for instance by a very unusual shape of the antennae, a wedge-shaped head capsule, and an enlarged prothorax. Presently Omophroninae comprise one genus with slightly less than 80 extant species (Valainis 2010;Kavanaugh et al. 2021). It is very widely distributed on the northern hemisphere, where it even reaches the Arctic circle (Valainis 2010). It also occurs in Guatemala and on Hispaniola, in South Africa and Madagascar, and in Malaysia and on the Philippines, but is not recorded from South America and Australia (Valainis 2010(Valainis , 2016. The adults are fairly small but conspicuous beetles, with a rounded rather than elongated body and a yellowish coloration with green metallic markings (e.g. Valainis 2010; Arndt et al. 2016). Superficially they resemble ladybirds rather than "normal" carabids, for instance of the megadiverse Harpalinae (e.g. Arndt et al. 2016). They are characterized by a number of apomorphic features, including a short and transverse head, an unusual type of externally closed procoxal cavities, a very broad prosternal process covering the entire mesoventrite, and a protibial burrowing spur (Beutel 1991). Shores of lakes and small ponds are the typical environment, and larvae and adults burrow efficiently in the sandy substrate (Landry and Bousquet 1984;Arndt et al. 2016).
The primary aim of the present study is a detailed morphological documentation of the larva, using light microscopy, microphotography, and also synchroton µ-CT scanning and computerbased 3D reconstruction. The observed features are entered in a data matrix and analysed cladistically, and also interpreted with respect to the possible habitat and life style.

2.
Material and methods

Sample origin and depository
The specimen described herein is from the lowermost Cenomanian (Cretaceous) deposits of the Hukawng Valley in Myanmar (Kachin). The age of deposits has been confirmed as 98.79 ± 0.62 Ma by radiometric analysis of zircons (Shi et al. 2012). The rough amber piece was trimmed and polished. The described holotype is deposited in the Charles University, Faculty of Science, Department of Zoology collection, Prague (prefix PřFUK) and is available for study upon request addressed to J. Prokop.

Preparation, imaging and SRμCT data reconstruction
The specimen was examined by transmitted light microscopy using a Leica S9 stereomicroscope and Olympus BX40 microscope with UIS2 objectives. The habitus and detailed photographs of the holotype specimen were taken using an Olympus BX40 fitted with a Canon EOS 550D digital camera or Leica S9D fitted with a Canon EOS 90D. The original photographs were processed using Adobe Photoshop CS (Adobe Systems Incorporated, San Jose, CA, USA). Some images were prepared as a series of focal layers, and then combined using the focus stacking software Helicon Focus (Helicon Soft, Kharkiv, Ukraine) or Zerene Stacker (Zerene systems LLC, Richland, USA). Line drawings of the specimen were prepared using camera lucida equipment and based on photographs using the Clip Studio Paint (CELSYS, Inc., Tokyo, Japan) and Adobe Photoshop CS software. Where the parts of the specimen were not visible the shape was completed according to the volume renders of the segmented SRµCT data.
Along with traditional optical microscopy we used synchrotron radiation based micro-computed tomography (SRµCT) to reconstruct the 3D habitus of the specimen and discern otherwise hardly accessible integumental details of cephalic, thoracic and abdominal structures. Imaging of amber specimen was performed at the Imaging Beamline P05 (IBL) (Greving et al. 2014;Haibel et al. 2010;Wilde et al. 2016) operated by the Helmholtz-Zentrum-Geesthacht at the storage ring PE-TRA III (Deutsches Elektronen Synchrotron -DESY, Hamburg, Germany) using SRµCT. A photon energy of 18 keV and a sample to detector distance of 50 mm has been used for imaging. Projections were recorded using a commercial 50 MP CMOS camera system (Ximea GmbH, Münster, Germany) with an effective pixel size of 0.46 µm. For the tomographic scan 4001 projections at equal intervals between 0 and π have been recorded. Tomographic reconstruction has been done by applying a transport of intensity phase retrieval approach and using the filtered back projection algorithm (FBP) implemented in a custom reconstruction pipeline (Moosmann et al. 2014) using Matlab (The MathWorks, Inc., Natick, USA) and the Astra Toolbox (van Aarle et al. 2015(van Aarle et al. , 2016Palenstijn et al. 2011). For the processing raw projections were binned for further processing two times resulting in an effective pixel size of the reconstructed volume of 0.92 µm.
The resulting 32-bit TIFF image stack was cropped, converted to 8-bit TIFF images and exported using Dragonfly software (Object Research Systems (ORS) Inc, Montreal, Canada). Segmentation of the whole larva was performed in Amira 6.0 software (Visage Imaging GmbH, Berlin, Germany). Parts of the larval body were marked in every 20th slice, in the region of mouthparts and pretarsus the structures were marked in every second to 10th slice. The segmentation process was then completed using Biomedisa (Lösel et al. 2020). The semiautomatic segmentation of the mouthparts provided insufficient results, therefore some structures were then segmented manually in Amira software. Still, parts of the mouthparts are missing from the final volume renders because of the little to no contrast of the structures in the image slices. The segmented data were exported as TIFF image stack using the plugin "multiExport" (Engelkes et al. 2018) in Amira software and imported into VG-Studio Max 3.4 software (Volume Graphics GmbH, Heidelberg, Germany) to create the final volume renders of the specimen.

Morphology, morphological terminology and cladistic analysis
The main aim of this study is to document the general morphological configuration of the larva and structures that can be related with specific functions. This includes the shape of the head capsule, the condition of the antennae and mouthparts, the general configuration of the postcephalic body, and features of the legs and urogomphi. Chaetotaxy, which can be useful in a taxonomic context, is not in the main focus of our contribution. Some features are included (partly based on personal communication with K. Makarov). However, we did not attempt a full treatment of the chaetotaxy. As our specimen is not a first instar the interpretation of the pattern of setae, sensilla and pores would have been difficult. The morphological terminology of the specimen in this study follows Arndt (1993), Beutel (1993), and Lawrence and Ślipiński (2013).
Characters were entered in a matrix with Winclada (Nixon 1999) and parsimony analyses were carried out with NONA (ratchet, 1000 replicates) (Goloboff 1995). The branch support value (Bremer 1994) of Omophron + †Cretomophron was calculated with NONA. The taxon sampling was limited as the primary aim was to confirm the placement of the fossil larva. A full scale analysis of Carabidae would have been far beyond the scope of this contribution. Moreover, larval characters alone would certainly be insufficient to reveal the phylogenetic pattern in the megadiverse family. A solid phylogeny of Geadephaga based on transcriptomic data is presently not available (Vasilikopoulos et al. 2021).

Data resources
The raw scan data, original unedited photos, and reconstructions will be made available at Zenodo repository at https://doi.org/10.5281/zenodo.8151974.

Systematic palaeontology
Order  Measurements. Length of the inclusion from the tip of the right antenna to the tip of right urogomphus 7.2 mm.
Etymology. The specific epithet refers to the damaged (mutilated) mandibles. Coloration. Sclerotized areas such as thoracic tergites, coxae or parts of the head middle brown to dark brown.
Other parts with some degree of sclerotization like legs and abdominal tergites light brown. Membranous or semimembranous regions, e.g., pleural areas, cream-colored.
Setation. Body surface with a well-developed vestiture of long setae, especially inserted on the dorsal side of the head, on the tergites, and on the pleural areas of the abdominal segments, and urogomphi. Legs with pattern of long chaetae and long, rather thin spike-like setae.  3B); 1 st antennomere markedly shorter than others; 2 nd antennomere distinctly longer than in Omophron; apical 4 th antennomere narrower than proximal segments, cylindrical, about half as long as 3 rd antennomere, and distinctly turned outwards, with three apical setae (broken off) ( 1 st palpomere about half as wide as distal edge of stipes, slightly longer than broad; 2 nd palpomere cylindrical, elongate, narrower than 1 st but more than twice as long; 3 rd palpomere slightly narrower than 2 nd and shorter than 1 st ; apical palpomere apparently broken off. Galea two-segmented; proximal galeomere slender, elongate and slightly curved; distal galeomere narrower and less than half as long, apically rounded; lacinia elongate, spine-like, straight or nearly straight, slightly longer than basal ga- leomere. Labium (Figs 2A, 5B): Labial submentum fully integrated into ventral wall of head capsule, medially divided by ventral ecdysial line; mentum short, trapezoid, membranous; prementum small, roughly quadrangular; distinctly protracted but covered by distal part of enlarged nasale in dorsal view; ligula not clearly visible, possible much shorter than in Omophron; palps two-segmented; elongate 1 st palpomere nearly twice as long as prementum and very slightly curved, slightly narrowing distally; 2 nd palpomere distinctly shorter, cylindrical, apically rounded. Thorax: Slightly more than 1/3 of total body length (excl. urogomphi) (Figs 1B, 4). In lateral view appearing moderately compressed dorsoventrally (Fig. 3B). Segments distinctly larger and broader than those of abdomen. Prothorax about half as long as all three segments combined; anteriorly with distinct collar with densely set longitudinal riffles. Pronotum well-sclerotized, with dark brown tergal halves separated by distinct median ecdysial suture; distinctly widening posteriorly, almost twice as wide posteriorly than at anterior edge; lateral edges straight, evenly diverging posteriorly; posterolateral cor-ners not clearly visible. Meso-and metathorax similar except for longer hind legs. Mesonotum slightly wider than metanotum and more distinctly rounded laterally; both sclerotized and divided by median ecdysial suture; both distinctly concave anteriorly and very slightly convex posteriorly. Legs (Figs 2B, C): Six-segmented, strongly developed, robust, almost as long as the thorax (Fig. 3B).

List of characters for the systematic placement of †Cretomophron
Orientation of head: (0) subprognathous; (1) prognathous; (2) hyperprognathous; (3) almost at right angle to longitudinal body axis. The head is prognathous in the larva of †Cretomophron like in almost all groups of Adephaga (Beutel 1993) (Fig. 3B). It is hyperprognathous in larvae of Metrius, and almost at a right angle to the longitudinal body axis in Cicindelinae, forming a lid-like structure (Breyer 1989;Beutel 1992a;Arndt 1993;Arndt et al. 2016).
Shape of head in lateral view: (0) dorsal and ventral side more or less parallel-sided; (1) wedge shaped. Distinctly wedge-shaped in extant Omophroninae (Landry and Bousquet 1984;Beutel 1991;Arndt et al. 2016) and also in the larva of †Cretomophron (Fig. 3B). The dorsal and ventral side are usually parallel-sided in Adephaga or the dorsal side is more or less convex (e.g., Arndt 1993;Beutel 1993).
M. craniolacinialis: (0) present and attached to the base of the lacinia; (10) replaced by M. craniostipitalis. The muscle with a typical attachment on the lacinia is present in Cupedidae and various groups of Polyphaga, but missing in all groups of Adephaga with the noteworthy exception of Gyrinidae (Noars 1956;Beutel 1991Beutel , 1992aBeutel -c, 1993.
Pronotum: (0) shorter than meso-and metanotum combined; (1) as long as meso-and metanotum combined. The pronotum of larvae of Omophron and †Cretomophron is about as long the meso-and metanotum combined and rounded laterally or widening towards the posterior margin (Landry and Bousquet 1984) (Figs 1A, 4). It is usually more or less parallel-sided in Carabidae and less long than the combined posterior thoracic tergites (Arndt 1993: fig. 1; Arndt et al. 2016).

Results of the phylogenetic analysis
The analysis of our limited larval data set with 38 larval characters and 28 terminal taxa yielded only two minimum length trees with 95 steps (consistency index 0.68, retention index 0.85). It clearly confirms the placement of †Cretomophron as sister to the extant genus Omophron, with four unambiguous apomorphies shared by both taxa and a branch support value of 6. The monophyly of Adephaga, of Adephaga excl. Gyrinidae, of Dytiscoidea, Geadephaga, and Carabidae (Fig. 6) is in agreement with previous studies (e.g. Beutel et al. 2020).

Phylogeny
The larvae we examined can be unambiguously assigned to the species-rich coleopteran suborder Adephaga.  Lawrence et al. 2011). The campodeiform configuration of the larva, the pattern of sclerotization of the postcephalic body, and the presence of urogomphi and a distinct pygopod formed by abdominal segment X distinguish it clearly from Archostemata (Beutel and Hörnschemeyer 2002a, b). The pronouncedly prognathous head with protracted ventral mouthparts and a fused labrum are apomorphies placing it in Adephaga (e.g. Beutel 1993;Haas 1996, 2000). The presence of ten well-developed segments and the absence of tracheal or microtracheal gills indicate that the larva belongs to Geadephaga. Elongate urogomphi and various other characteristics differ from conditions found in the relict family Trachypachidae (Lindroth 1960;Beutel and Arndt 1995). An entire series of features supports a placement in the specialized basal grade carabid subfamily Omophroninae (Landry and Bousquet 1984;Arndt et al. 2016): this includes the distinct wedge shape of the head in lateral view, a large triangular nasale, the highly unusual elevated posture of the antenna, the large bidentate retinaculum, an enlarged prothorax and pronotum, and legs with a distinct vestiture of spines. An additional potential synapomorphy is the presence of distinct lateral projections of abdominal segments I-VIII, formed by setose epipleurites, still absent in first instars of Omophron described by Landry and Bousquet (1984) but present in later stages (R.G. Beutel, pers. obs.). Despite of the clear phylogenetic assignment, †Cretomophron differs in several features from its sister genus Omophron. The thorax is not distinctly hump-shaped (Fig. 3B) as in larvae of the extant genus, even though this is possibly an artefact of preservation. In contrast to Omophron larvae, the posterior tentorial grooves are not shifted to the hind margin of the ventral wall of the head capsule (Figs 2A, 5B), apparently a plesiomorphic condition. The 2 nd antennomere is distinctly longer than in Omophron and the ligula is possibly much shorter. Whereas the elongate lacinia of Omophron is curved, it appears straight and spine-like in †Cretomophron (Figs 2A, 5B). A distinct lobe-like projection is present on the apical region of the large trochanters of the fossil (Fig. 2C). This structure was not observed in larvae of Omophron (Landry and Bousquet 1984), but may have been overlooked. Abdominal segments III-VI of Cretomophron displays many setae arranged in transverse rows, whereas such a pattern is present on tergites I-V in Omophron (K. Makarov, pers. comm.; Fig. 4A,B). A large, triangular nasal projection as it is characteristic for Omophroninae (Beutel 1991), is also found in few other groups of Carabidae, as for instance in Elaphrini (Makarov 1994: fig. 51), very likely the result of parallel evolution. A bidentate retinaculum is also present in larvae of extant and extinct species of Migadopinae (Thompson 1979;Liu et al. 2023). However, the specific shape is different, and a close relationship between both small subfamilies appears unlikely (Beutel 1991(Beutel , 1992, even though a robust phylogeny of Carabidae with an extensive molecular data set and taxon sampling is still lacking (Maddison et al. 2009;Vasilikopoulos et al. 2021;Raupach et al. 2022). The elongate lacinia and ligula are features shared with Paussinae (e.g. Beutel et al. 1992a;Arndt et al. 2016). However, considering distant placements of Omophron and this specialized subfamily in a recent transcriptomic analysis (Vasilikopoulos et al. 2021), this is also rather due to homoplasy, or possibly a symplesiomorphy in the case of the lacinia.
The presented taxon sampling is too limited to resolve the phylogeny of Carabidae. Moreover, larval characters alone will not be sufficient to reconstruct the evolutionary history of this extremely species-rich family. A robust phylogeny will require a dense sample of taxa and an extensive molecular data set, i.e., transcriptomes or ultraconserved elements (UCE) (see e.g. Vasilikopoulos et al. 2021).

Habitat and life style
Larvae of Omophron dig burrows in sand or clay in the direct vicinity of fresh-or saline aquatic habitats, and leave them at night to hunt prey (Landry and Bousquet 1984;Brandmayr et al. 1998;Arndt et al. 2016;Brandmayr 2020). As the entire configuration of the body of the larva we describe here and also various specific parts are very similar to conditions observed in Omophron, we assume that they are also similar in their biology and habitat preference. However, the legs of †Cretomophron clearly differ from Omophron, bearing long chaetae and rather thin and long spike-like setae, suggesting that Cretomophron may have lived on beaches with finer sand. Other adaptations to burrowing in sandy substrates are the wedge-shaped head and the enlarged prothorax.
It is very likely that the larval instars of †Cretomophron were active predators like almost all carabid larvae. Even though the mandibles are not fully preserved, it is apparent that they were suitable for grasping agile prey. It is likely that small arthropods and insect larvae were detected by the antennae and the slender maxillae functioning like accessory ventral tactile organs. The prey was likely fixed between the large triangular nasale and the mesal mandibular edge, and its body wall then pierced by the sharp teeth of the bidentate retinaculum. Even though the preoral hypopharyngeal filter is not visible in the fossil (Landry and Bousquet 1984;Beutel 1991Beutel , 1992a), it is very likely that the larva ingested preorally liquefied food.

Conclusions
The fossil documents the presence of Omophroninae in the Cretaceous.
The placement of †Cretomophron in this small but very distinctive carabid subfamily is unambiguously confirmed. The morphology suggests burrowing and predaceous habits, similar to larvae (and adults) of the extant genus Omophron.

Declaration of Interest
The authors declare no conflict of interest.