Circularva reichardti Shcherbakov et Ponomarenko, 2023

Shcherbakov, D. E. & Ponomarenko, A. G., 2023, The first known fossil hygropetric beetle larva related to Myxophaga (Coleoptera) from the Permian of European Russia, Russian Entomological Journal 32 (3), pp. 261-270 : 265-268

Circularva reichardti Shcherbakov et Ponomarenko ,

sp.n.

Figs 1–13.

MATERIAL. Holotype PIN 199 /379 (part and counterpart), mature larva; spoil-heaps of Kuzminovskii Mine 65 km NNW Orenburg,

abandoned Kargala [Kargaly] copper mines, Orenburg Province, SE

European Russia;Guadalupian(Middle Permian),Wordian (Urzhumian),

Bolshaya Kinel Formation, Estemmenosuchus uralensis tetrapod zone

[ Golubev, 2000; Sennikov, Golubev, 2017]. Fig. 13. Circularva reichardti gen. et sp.n., dorsal habitus;

DESCRIPTION. Body 3.3 mm long, 3.0 mm wide. Dor- structures of ventral side in dashed line.

sum moderately convex at life, flattened at burial (flat medi- Рис. 13. Circularva reichardti gen. et sp.n., общий вид ally and sloping at sides; tergites VII–IX retaining original сверху; структуры вентральной стороны пунктиром.

arched shape but tilted towards bedding plane; preceding tergites flattened, slightly telescoped medially and drawn apart

of tergites I–IX; gills visible on some posterior abdominal seglaterally; dorsal head surface less sloped than at life). Head

ments (those of anterior segments possibly superimposed onto transverse (1:1.7), slightly convex in central part, with a pair

abdomen; gill visible caudad of segment X belongs to segof shallow depressions at anterior margin, lateral angles acute.

ment IX); gills no less than 0.5 mm long, annulate proximally, Antennae not visible. Thoracic segments rather short along

smoothly curved (probably not segmented); plastron mesh and midline (pronotum 0.7 as long as meso- + metanotum there);

spiracular openings not visible. Tergite X as wide as head, meso- and metathoracic paranota much longer at lateral mar-

2.5 times as wide as long, medially not separated from tergite gin than at base, whereas prothoracic and segment I paranota

IX. Body somewhat darkened; gills pale.

slightly shorter at lateral margin than at base. Setigerous tu-

ETYMOLOGY. After Hans Reichardt, Brazilian entomolbercles along hind margin of thoracic and abdominal tergites

ogist, an expert on Myxophaga .

preserved as pits on counterpart. Thoracic and abdominal sternites wide (~1/2 and 2/3 segment width, respectively), their posterior margins straight. Leg bases indicated by depressions Discussion at sides of thoracic sternites; legs not visible, apparently short.

Abdominal segments with small rounded epipleurites (clearly Schizophoroids and Torridincolidae . Phoroschizidae visible on left side, distorted and curved on right side) and (= Schizophoridae ), Catiniidae , Asiocoleidae , and Rhombomuch wider hypopleurites. Spiracular gills originating sublat- coleidae possess so-called schiza, the ridge or tongue on the erally from emarginations developed near posterolateral angles underside of the elytron interlocking with the body, interpreted

Figs 5–12. Circularva reichardti gen. et sp.n.: 5–6 — habitus (arrows, spiracular gills); 7 — head (on crosshair of arrows, centre of group of stemmata); 8 — meso- and metanotum and anterior abdominal tergites (arrows, setigerous tubercles); 9 — left side of abdomen (arrows, spiracular gills of segments VI and VII); 10 — base of spiracular gill of segment VI (annulations visible); 11 — left side, meso- (p2) and metapleuron (p3); 12 — same, enlarged, abdominal epipleurites (asterisks) and hypopleurites (numbered). Positive (5, 7, 9–12) and negative impression (6, 8). Polarized light (6); SEM (BSE: 11, 12). Scale bar: 1 mm (5), 0.5 mm (9), 0.2 mm (10).

Рис. 5–12. Circularva reichardti gen. et sp.n.:5–6— общий вид (стрелки– дыхальцевые жабры); 7 — голова(на перекрестье стрелок центр группы стемм); 8 — средне- и Заднеспинка и первые тергиты брюшка (стрелки – щетинконосные бугорки); 9 — леваЯ сторона брюшка (стрелки – дыхальцевые жабры сегментов VI и VII); 10 — основание дыхальцевой жабры сегмента VI (видна кольчатость); 11 — леваЯ сторона, меЗо- (p2) и метаплеврит (p3); 12 — то же, увеличено, брюшные Эпиплевриты (ЗвеЗдочки) и гипоплевриты(пронумерованы).ПрЯмой (5, 7, 9–12) и обратный отпечаток (6, 8). ПолЯриЗованный свет (6); СЭМ (BSE: 11, 12). Длина масштабной линейки: 1 мм (5), 0.5 мм (9), 0.2 мм (10).

266 First fossil hygropetric beetle larva from the Permian of Russia as an adaptation for storing the air under the elytra, and evidence of aquatic lifestyle, despite the absence of swimming adaptations in these families [Ponomarenko, 1968, 2003b, 2011, 2013]. Such an interlocking tongue (sometimes situated more caudally) is also found in Triaplidae (end-Permian to Triassic Haliploidea), Timarchopsinae (= Necronectinae, Mesozoic Coptoclavidae, Dytiscoidea ), and many other extant beetle families of three suborders: Myxophaga (except Hydroscaphidae ),Adephaga ( Haliplidae , Meruidae , Aspidytidae , Noteridae , Hygrobiidae , Amphizoidae , some Dytiscidae , some Carabidae ), and Polyphaga (some Hydrophiloidea, some Scirtoidea, some Byrrhoidea (including Dryopoidea), some Buprestoidea, some Curculionoidea, etc.) [Ponomarenko, 1968, 1977; Lawrence et al., 2011; Lawrence, Ślipiński, 2013]. Isolated smooth elytra with schiza may belong to any suborder, so genera based on such fossils are referred to the form family Schizocoleidae [Ponomarenko, 1968, 2011].

Among Myxophaga , Torridincolidae are most diverse morphologically and taxonomically and approach extinct schizophoroids most closely [Lawrence, Reichardt, 1991]. Phoroschizidae (= Schizophoridae ; Middle Permian–Cretaceous) are similar to Torridincolidae in many adult characters: head often prognathous and retracted into prothorax, 11-segmented antennae without club, elytra with interlocking tongue, hindwing with oblongum, 5-segmented tarsi, five abdominal ventrites, prosternal process well-developed (often broad, apically convex, in contact with mesoventrite), and metanepisternum participating in bordering mesocoxal cavities (archostematan feature retained also in some Adephaga and Polyphaga ) [Lawrence et al., 2011]. The elytra of Deleveinae [ Bilton, Mlambo, 2023] and Phoroschizidae lack serial punctures, while the elytra of Torridincolinae with 9–13 striae [ Perkins, Bergsten, 2019] more closely resemble those of Rhombocoleidae and Asiocoleidae (indicating that the superfamily rank given to these fossil groups could be exaggerated). Torridincolidae are more derived than Phoroschizidae in the very short antennae with modified scape and pedicel, very broad prosternal process, metaventrite without transverse suture, proximal tarsomeres very short, andabdominalventritesII–IVshort.Torridincolidsare 1.0– 2.9 mm long, whereas described schizophorids are larger, at least 5 mm long. However, there are Permian schizocoleid elytra of minute size, e.g. Late Permian Pseudochrysomelites ovum 1.5–1.8 mm long[Ponomarenko,2004,2013]belongingtothebeetles2.0– 2.4mm long, almost surely Phoroschizidae (only a few finds of the earliest Adephaga and Polyphaga are known from the Permian, see below). Even smaller schizocoleid elytra up to 0.82 mm long (corresponding to beetles about 1.1 mm long) are found in the Permian-Triassic boundary beds [Ponomarenko, 2015].On the other hand, Catiniidae (Triassic–Cretaceous) are similar rather to Hydroscaphidae in the prosternal process very short and pointed.

Affinities of Circularva gen.n. Body shapes reminiscent of Oniscus isopods have evolved many times in various arthropod groups, including crustaceans, millipedes, and insects. All Isopoda like other Malacostraca have their trunk clearly divided into a large thorax containing at least six free large segments, and a shorter and narrower abdomen, consisting of no more than five smaller segments [Schram, 1986]. In onisciform Diplopoda, trunk tergite II is often enlarged, more rarely either only tergite III or both tergites III and IV, while tergite I (collum) is usually relatively small [Golovatch, 2003]. On the contrary, Circularva gen.n. shows insect tagmosis: a three-segmented thorax, clearly differentiated from a ten-segmented abdomen, with thoracic tergites longer than the abdominal tergites (though only moderately, as e.g. in larvae of Delevea Reichardt, 1976 , non-torridincolid Myxophaga , Amphizoidae , and many other beetles). The insect nature of Circularva gen.n. is confirmed by its wide, well sclerotized thoracic and abdominal sternites separated by straight intersegmental boundaries,and the leg bases at the sides of thoracic sternites. In millipedes, the sternites are small and the leg bases are close to the mid-ventral line, while in isopods, the sternal sclerotizations are less developed, paired, with curved boundaries, and the legs have the first free segment long, attached laterally and directed medially.

Somewhat enlarged lateral lobes (paranota) of meso- and metanotum indicate that Circularva gen.n. is a pterygote immature (or wingless adult). These lobes, which do not protrude caudad to form wingpads, and the stemmata instead of compound eyes, allow to distinguish the fossil from onisciform nymphs of Palaeodictyoptera, Blattodea, and Hemiptera . Moreover, the new fossil is distinct from Blattodea in the shape of pronotum, and from Hemiptera in the abdominal tergites VIII–IX not deeply U-shaped, segment X large, and anal opening not dorsal. Therefore, this fossil is a holometabolan larva.

Among Holometabola, onisciform, sometimes disc-like larvae are met with in various groups of Coleoptera , as well as in some advanced families of two orders not known from the Permian — Diptera (e.g. Platypezidae ) and Lepidoptera (e.g. Lycaenidae ). Two characters of the ventral side confirm that Circularva gen.n. is a beetle larva. (a) Two rows of sclerites in the lateroventral area of the abdomen, epipleurites and hypopleurites, looking like e.g. in Carabidae and resembling the embryonal configuration [ Kobayashi et al., 2013]. (b) The thoracic metapleura are slightly longer than the mesopleura (as e.g. in Torridincolinae [ Perkins, Bergsten, 2019] and Elmidae [ Shepard, 2020]); beetles are the only posteromotoric holometabolans.

Onisciform beetle larvae are known in Myxophaga ( Torridincolidae ), Adephaga ( Amphizoidae , Cychrini ), and various groups of Polyphaga ( Silphidae , Scydmaeninae , Scirtidae , Byrrhidae , Elmidae , Psephenidae , Brachypsectridae , Lampyridae , Cerylonidae , Discolomatidae , Endomychidae , Coccinellidae , Corylophidae , Tenebrionidae , Chrysomelidae ). In Circularva gen.n. the head is large, entirely visible from above, longer than the pronotum, like in some Torridincolidae (e.g. Delevea [ Endrödy-Younga, 1997, pl. 7B]), Amphizoidae , and Brachypsectridae , whereas in onisciform larvae of the other Polyphaga the head, sometimes small and retractable, is more or less concealed by (or at least shorter than) the pronotum in dorsal view. The fossil shows neither urogomphi nor branched lateral processes and spine-like terminal segment, in contrast to amphizoid or brachypsectrid larvae. Instead, it possesses filiform sublateral abdominal gills, like in Torridincolidae .Such detail as proximal gill annulations allow to assume that these fossil gills were designed like in living torridincolids, i.e. with spiracular openings and plastron mesh, although neither of the last two characters is visible.

Several other features shared by Circularva gen.n. and some Torridincolidae genera support the assumption of its relationship with this family. In the large diamond-shaped head, contiguous abdominal paranota, and dorsolateral position of spiracular gills, the new genus is most similar to Torridincola Steffan, 1964 [Steffan, 1964, fig. 9], but in the latter the anterior pronotal margin is only slightly sinuate, and the segment X is transformed into paired anal flaps scarcely visible from above. In the short thorax and setigerous tubercles on tergites, the fossil is similar to Delevea , but in the latter the body is elongate, and the pronotum is longer than mesonotum. The legs of the fossil were apparently short, not visible from above, like in Satonius Endrödy-Younga, 1997 and Ytu Reichardt, 1973 , but in these latter abdominal paranota of successive segments are widely separated, festoon-like, bearing spiracular gills on their apices.

However, still other features of Circularva gen.n. are more primitive than in all known larvae of Myxophaga .

(1) Stemmata on each side 6, vs. 5 in Hydroscaphidae , find of Circularva gen.n. agrees with supposition that Schizopho- 4 in Sphaeriusidae , 4 or 3 in Torridincolidae [ Beutel et al., 1999; romorpha, the earliest Adephaga [Ponomarenko, 1977; Erwin, Vanin, Costa, 2001]. Six is the primitive number of stemmata, 1979], and possibly also the earliest Polyphaga were aquatic or found in most families of Adephaga and several groups of riparian [ Ponomarenko, 1983] and had megalopteran-like larvae. Polyphaga , including most Hydrophilidae The fossil is peculiar Placing the new genus into the system of Coleoptera . The in the dorsolateral position of stemmata, in contrast to lateral in long, sublateral, proximally annulate spiracular gills developed Myxophaga ; the area of stemmata is flat (on a distinct elevation on the abdominal segments is a derived character shared by in Torridincolidae ). the new genus only with Torridincolidae . The features shared

(2) The abdominal segment X is large, not concealed by the by Circularva gen.n. with one or another torridincolid genus segment IX, whereas myxophagans have the segment X dimin- (disc-like body, diamond-shaped head, short legs, contiguous ished, rarely visible from above (see e.g. [Böving, Craighead, abdominal paranota, setigerous tubercles on tergites) also sup- 1931, pl. 9]), sometimes split into 2–3 lobes.The well-developed port this relationship, but can only be regarded as shared traits. segment X is presumably an ancestral condition in the larvae of However,the other characters (six stemmata in laterodorsal posi- Coleoptera [Lawrence, Ślipiński, 2013]. This difference could tion, large segment X, nine pairs of spiracular gills) demonstrate be considered very important if it were not for the coexistence that the fossil is more primitive than all known torridincolids. of both states of the larval segment X (well developed or re- If we consider larval spiracular gills of specific structure as duced) within certain families. Both of these states are found a unique synapomorphy of Torridincolidae , then we can assign not only in such a large and diverse family as Hydrophilidae Circularva gen.n. to this family. The differences in the number [ Archangelsky, 1998], but even in such a small and homoge- of stemmata and gills and degree of segment X development do neous family as Haliplidae —segment X reduced in Peltodytes not contradict this placement, so far as these three characters are Régimbart, 1879, singled out into the subfamily Peltodytinae variable within such a polymorphic family as Hydrophilidae , [ Böving, Craighead, 1931], but large and visible from above and two latter in Haliplidae as well. If so, then the present-day in other genera [Lawrence, Ślipiński, 2013; Makarov, Prokin, torridincolids are relics of a more diverse family that left nearly 2015]. The semicircular shape of the segment X is unknown in no trace in the fossil record due to taphonomically unfavourable other beetle larvae. mode of life.

(3) Nine pairs of spiracular gills, vs.eight in Torridincolidae . However, tracheal gills could have evolved into spiracular The number of abdominal segments bearing tracheal gills varies gills of the torridincolid type already in some schizophowithin Haliplidae (8 or 9) and Hydrophilidae (7 or 8). roids, and be inherited by torridincolids, while modified or

(4) Two rows of sclerites in the lateroventral area of the lost in remaining Myxophaga . Then torridincolid-type spiabdomen, a primitive condition, not recorded in Myxophaga . racular gills represent a synapomorphy of all myxophagan

Megalopteran-like beetle larvae. From the same Permian families, subsequently lost in all of them but torridincolids. fossil site as Circularva gen.n., one more aquatic larva was If so, similarities of the new genus to Kargalarva in retaining reported. This large (24 mm long), elongate larva with 10-seg- 9 pairs of tracheal gills and the well-developed segment X may mented abdomen bearing tracheal gills on the segments I–IX was be interpreted as evidence to assign Circularva gen.n. to a group originally assigned to Megaloptera and to the genus Permosialis more primitive than Torridincolidae .

Martynov, 1928 [Sharov, 1953], but later this genus, based on In the case, the new genus probably belongs to Phorosadults, was transferred to Miomoptera, believed to be terrestrial chizidae (or some other, possibly unknown, extinct group). insects [ Riek, 1976], and the larva was re-interpreted as that of Schizophoromorphs, including the small ones, are quite com- Coleoptera , namely Gyrinidae [ Beutel, Roughley, 1988]. Vari- mon in the fossil assemblage of Kargala. The estimated length ous similar larvae from the Middle Triassic Voltzia Sandstone, 2.0– 2.4 mm of the Late Permian beetle Pseudochrysomelites Vosges (Early Anisian,~245 Ma) were attributed to Megaloptera ovum ( Ponomarenko, 2004) , described in Schizocoleidae based [Marchal-Papier, 1998]. Later, all these larvae were assigned on five isolated elytra [ Ponomarenko, 2004, 2008], corresponds to the coleopteran infraorder Schizophoromorpha , the above- well to the 3 mm long Circularva larva (the adults are somewhat mentioned larva from Kargala was redescribed, named as Kar- smaller than the mature larvae in torridincolids).These smallest galarva permosialis Prokin, Ponomarenko et Kirejtshuk, 2019 schizophoroids (up to now known only from isolated elytra) and tentatively associated with Rhombocoleidae [Prokin et al., were even smaller than the largest Torridincolidae (2.9 mm).

2013, 2015, 2019], and another larva of this type was described Regardless of which of these two families the new genus is from the Late Triassic of Germany [ Prokin, Bashkuev, 2021]. placed in, it deserves separation at the subfamily level due to Among living beetles, megalopteran-like larvae (i.e. elongate important differences from both torridincolid and putative fossil with paired tracheal gills on abdominal segments) are known in schizophoromorph larvae. Among fossil aquatic beetles, three the most primitive Adephaga ( Gyrinidae , some Haliplidae and families were based on larvae: Coptoclavidae , Parahygrobiidae , Dytiscidae ) and some primitive Polyphaga (some Hydrophilidae , and Colymbotethidae , all Dytiscoidea [Ponomarenko, 1977, e.g. Hydrochara Berthold, 1827 , Berosus Leach, 1817 ). 1994]. However, classification of Coleoptera relies generally on

At first glance the onisciform larvae of Torridincolidae and adult characters.Therefore, we consider establishing a subfamily Circularva gen.n. look very dissimilar to the megalopteran-like for Circularva gen.n. to be premature and tentatively assign it larvae of primitive hydradephagans, some hydrophilids, and to Phoroschizidae (= Schizophoridae ).

presumably also schizophoromorphs.However, the gap is partly There is little chance for the Circularva adult to be found and bridged by the larvae of the most primitive torridincolid genus recognized among beetle fossils. For living insects, larvae may Delevea and the early instar larvae of derived torridincolids be associated with adults on direct (rearing, DNA sequences) or [ Reichardt, 1973; Beutel et al., 1999]—more elongated, less indirect (common occurrence) evidence.For compression fossils, flattened, with proportionally greater, leg-like tracheal (spiracu- only the latter is applicable, and only for rather common taxa of lar) gills on the abdomen, and thus more megalopteran-like. limnic (amphibiotic and aquatic) insects spending (most of) their Moreover, Kargalarva and Circularva gen.n. share with the nymphal/larval and adult life in/near the water bodies produclarvae of Gyrinidae and some Haliplidae two important primitive ing fossiliferous sediments. In compression fossil assemblages characters: well-developed abdominal segment X visible from dominated by limnic insects, rheophiles are very rare—only one above, and abdomen with 9 pairs of tracheal gills. Therefore, the specimen of C. reichardti sp.n. from about one thousand total 268 First fossil hygropetric beetle larva from the Permian of Russia insects collected at Kargala, so the chance of finding another such larva or conspecific adult is very small.Adult Torridincolidae are not as distinctive as their larvae in characters usually preserved in compression fossils, so even if found, a Circularva adult may pass unnoticed among Phoroschizidae or (if represented with an isolated elytron) Schizocoleidae .

Plastron respiration. Whether Circularva gen.n. belongs to Torridincolidae or their ancestors, one may conclude that by the Middle Permian the beetles of schizophoroid lineage ( Schizophoromorpha = Myxophaga s.l.) already had diverse aquatic larvae, not only resembling those of Megaloptera and Gyrinidae , but also highly specialized rheophilic plastron-bearers, adapted for life in hygropetric habitats.

Plastrons are widespread in insects, arachnids, and even in plants subject to drowning in their wetland habitats [Thorpe, 1950; Heckman, 1983; Pedersen, Colmer, 2012]. Insect plastrons were in existence in the Mesozoic, at least in those extant families that now possess plastrons among aquatic Heteroptera, Coleoptera , and pupae of Diptera [ Brown, 1987; Crosskey, 1991; Hasiotis, 2000; Lukashevich, 2012; Criscione, Grimaldi, 2017]. A covering of modified microtrichia for trapping an air film is found on the underside of elytra in froghopper-like Dysmorphoptilidae from the Permian-Triassic boundary beds [Shcherbakov, 2022]. The spiracular gills of Circularva gen.n. are even earlier fossil plastron evidence, though indirect.

On the age of coleopteran suborders. Until recently, the subordersAdephaga, Polyphaga and Myxophaga had no reliable pre-Triassic records, though it was suggested that some isolated elytra of advanced structure from the Late Permian could belong to archaic polyphagan and adephagan taxa [ Ponomarenko,2002]. Now the earliest finds of adult Myxophaga [Kirejtshuk et al., 2023] and Adephaga ( Triaplidae , rather than Gyrinidae [ Volkov, 2013; Kirejtshuk, Prokin, 2018; Yan et al., 2018; Ponomarenko, 2021], Haliplidae [ Ponomarenko, Prokin, 2015], and Trachypachidae [ Ponomarenko, Volkov, 2013]) are described from the so-called Permian-Triassic boundary beds (Babiy Kamen of the Kuznetsk Basin; Anakit and Untuun of the Tunguska Basin), dated basal Triassic or rather end-Permian [Shcherbakov et al., 2021; Shcherbakov, 2022]. Beetles possibly belonging to Polyphaga are now recorded ~10 Ma earlier than the Permian/ Triassic boundary, in the Middle Permian (late Capitanian, ~261 Ma) of China [ Yan, Strelnikova, 2022], while elytra characteristic only of Polyphaga are found in the latest Permian (Changhsingian, ~253 Ma) of European Russia [Ponomarenko, 2003a]. The discovery of an even older torridincolid-type larva from European Russia may indicate that Myxophaga also appeared in the Permian.

Acknowledgements. The authors express their deep gratitude to Alexey Bashkuev, who found this unique fossil, to Roman Rakitov for the excellent SEM micrographs, to Eugeny Karasev and Julia Shuvalova (all PIN) for assisting with microscopy, to participants of the 5th All-Russia Symposium on Amphibiotic and Aquatic Insects, Borok , 15–17 October 2013 for constructive discussion of taxonomic position of the fossil, to Daniela Takiya ( Universidade Federal do Rio de Janeiro ) for donating torridincolid specimens, to Sergei Golovatch ( Severtsov Institute for Problems of Ecology and Evolution RAS, Moscow), Alexander Prokin ( Papanin Institute for Biology of Inland Waters RAS, Borok ), Bernhard van Vondel ( Natuurhistorisch Museum Rotterdam ), Alexander Kirejtshuk ( Zoological Institute RAS, St. Petersburg) and Kirill Makarov (Moscow State Pedagogical University ) for constructive comments, to Max Barclay ( Natural History Museum , London ) for critical reading of the manuscript, and to an anonymous reviewer for valuable comments on the first version of this paper. The study was supported by the Russian Science Foundation (project 21- 14-00284) .

Competing interests. The authors declare no competing interests.