The first ‘Grylloblattida’ of the family Liomopteridae from the Middle Permian in the Onder Karoo, South Africa (Insecta: Polyneoptera)

Here we describe a new genus and four new species of the extinct ‘Grylloblattida’: Liomopteridae Sellards, 1909: Liomopterum connexus Cawood & Nel, n. sp., Liomopterum daenerys Cawood & Nel, n. sp., Colubrosopterum karooensis Cawood & Nel, n. gen., n. sp., and Paraliomopterum sp. The fossil wings were collected from a new Middle Permian locality near Sutherland, Northern Cape, South Africa, with the horizon close to the Ecca-Beaufort Group contact in the southern Karoo Basin.


INTRODUCTION
Although the extant order Grylloblattodea Walker, 1914 contains only the extant family Grylloblattidae Walker, 1914 with five genera, its alleged fossil record 'Grylloblattida' comprises at least 50 extinct families of winged insects (Bai et al. 2010). The Grylloblattidae is currently considered the sister group of the Mantophasmatodea Zompro, Klass, Kristensen & Adis, 2002(Wipfler et al. 2019. The exact relationships among the extant Grylloblattidae and the late Palaeozoic and Mesozoic taxa currently attributed to the 'Grylloblattida' remain unclear. Originally, Rasnitsyn (1976) grouped the Mesozoic family Blattogryllidae Rasnitsyn, 1976 with Grylloblattidae on the basis of a series of unpolarised and probably plesiomorphic characters (Bai et al. 2010). Later, various authors added new extinct families to the set 'Grylloblattida', again without phylogenetic argument. The only phylogenetic analysis of the 'Grylloblattida' (including the fossil taxa) was made without a real outgroup, on the basis of a priori polarisations of the characters (Storozhenko 1998). Thus the monophyly of the 'Grylloblattida' (including the fossil taxa) is not demonstrated and the group is not supported by any apomorphy. It could be paraphyletic or even polyphyletic. The removal of some 'grylloblattid' families and their transfer into the clade Paoliida illustrate this hypothesis (Prokop et al. 2014). The oldest fossils attributed to the 'Grylloblattida' date back to the late Carboniferous period (Aristov 2009;Wipfler et al. 2014). 'Grylloblattida' apparently formed a significant component of the global Permian insect palaeofauna in lowland habitats, as evidenced by the abundance and high diversity of fossilized specimens described in the literature (see Fossilworks database).
'Grylloblattidan' wing fossils were recently excavated at a new fossil locality near Sutherland in the Northern Cape, within the uppermost Waterford Formation, Ecca Group, Karoo Supergroup, and are probably of early Wordian Age. These 'grylloblattidan' wings represent the first recorded from the Middle Permian of Gondwana.

LocaLity
To date, about 30 wing fragments attributable to the 'Grylloblattida' have been recognised from the Onder Karoo locality near Sutherland. Six specimens with well-preserved venation were selected for study, attributed to three genera, all in the Liomopteridae Sellards, 1909. The Onder Karoo locality is approximately 40 km from Sutherland. Although it is currently mapped within the uppermost Waterford Formation, Ecca Group, it probably lies at the Ecca-Beaufort group contact (M. Day and B. S. Rubidge, ESI, University of the Witwatersrand; pers. comm.).

GeoLoGicaL settinG
The deposits at the Onder Karoo site are olive-grey, fine-grained and finely laminated mudrock and siltstone, indicating deposition in a low-energy setting such as a lake. The fossils are all impressions. Considering the regional geological context, including uranium-lead (U-Pb) dates obtained from volcanic ashes in the nearby Ouberg Pass (Lanci et al. 2013;Day et al. 2022), the fossils are probably Middle Permian (Wordian) in age, dating to approximately 268 Mya.

ProcessinG of materiaL
Standard excavation techniques using geological hammers and chisels were used to excavate the fossils. The material was bulk-collected and large numbers of specimens were collected for ecological studies. The material used for this research

MOTS CLÉS
Insecta, Polyneoptera, Permien moyen, 'Karoo Supergroup', Gondwana, genre nouveau, espèces nouvelles. is housed in the fossil collections of the Albany Museum, Makhanda, South Africa. Photographs of each specimen were taken using a Nikon D5000 camera mounted onto a Zeiss Discovery V8 microscope. Low-angle lighting was used to highlight surface relief of the impressions, with a polarised filter to reduce glare. Images were photographed at multiple lighting angles to recover as much detail as possible. Photographs were taken at the highest resolution possible and edited with GIMP v2.8. The specimens were identified to genuslevel using their wing venation, following the wing venation nomenclature of Carpenter (1950) and Storozhenko (1998 descriPtion Forewing with only extreme base and apex, and anal areas missing; wing 19.0 mm long, 7.5 mm wide; membranous and bare, apical part not strongly widened; double rows of cells absent; crossveins simple throughout. Anterior margin convex; costal area (between C and ScP) twice as wide as subcostal area, with transverse veinlets and few crossveins between them; ScP sclerotized, terminating at distal third of wing; few crossveins between ScP and R; RA weakly curved with six-seven short anterior branches; origin of RP proximal to basal third of wing; RP with two weak distal branches fused together distally; area between RA and RP broad, with a series of parallel transverse   rites Sharov, 1961, Fumopterum Kukalova, 1964, Liomopterella Sharov, 1961, and all of the other liomopterid genera except Abashevia Sharov, 1961, Khosara Martynov, 1937, Liomopterites Sharov, 1961, Liomopterum, Micropermula Storozhenko, 1992, Mioloptera Riek, 1973, and Mongolopermula Storozhenko, 1992.  HoLotyPe. -AM11312 (forewing impression), stored in the fossil collection at the Albany Museum, Makhanda, South Africa (Fig. 3).    diaGnosis. -Costal area three times as wide as subcostal area, RP weakly curved and simple; fork of M distal to that of the origin of RP; CuA 1 with at least six branches, area between CuA 2 and CuP broad, numerous S-shaped crossveins (Fig. 4).
HoLotyPe. -AM11389 (forewing impression), stored in the fossil collection at the Albany Museum, Makhanda, South Africa (Fig. 4). descriPtion A nearly complete forewing with only the apex and part of the anal area missing. Wing 22.0 mm long and 10.5 mm wide; anterior margin of wing convex; costal area with forking transverse veinlets and crossveins linking them, three times as wide as subcostal area; apex of ScP at distal third of wing; few crossveins present between ScP and R; RA weakly curved with several anterior branches; RA branches dichotomising; origin of RP proximal to basal quarter of wing; RP simple; area between RA and RP with S-shaped crossveins; M divided into two main branches, branching distal to base of RP; MA forking slightly more proximally than MP; MA with three branches; MP with two branches; area between M and CuA broad; CuA and CuP separating near wing base; CuA divided into CuA 1 and CuA 2 proximal to mid wing; CuA 1 with at least six branches, approximately parallel to posterior wing margin; area between CuA 1 and CuA 2 with a double row of cells; CuA 2 simple; area between CuA and CuP broad, with a double row of cells; CuP simple; numerous S-shaped crossveins; coloration is represented as spots (Fig. 4) notes (Table 2) This fossil belongs to a group of genera with RP reduced (i.e., with fewer than three branches and two forks), viz. Abashevia, Micropermula, Khosara, Mongolopermula, Mioloptera, Liomopterites, and Liomopterum. It differs from Liomopterum in having the fork of M in a distal position, well after the origin of RP. Micropermula is based on a hindwing that, unlike our fossil, has a simple MA. The forked MA also excludes Mongolopermula. Khosara has a CuA 1 with a simple fork, unlike our fossil. Liomopterites has a narrower area between CuA 2 and CuP. Our fossil shares with Mioloptera and Abashevia the rather distal position of CuA 2 . It differs from all of them except the upper Permian South African genus Mioloptera in the broader area between CuA 2 and CuP. Colubrosopterum Cawood & Nel, n. gen. differs from AM11265 (provisionally assigned to Paraliomopterum Sharov, 1961) in the simple RP, a character also seen in Liomopterum connexus Cawood & Nel, n. sp.. Nevertheless, the first fork of M in our fossil is comparatively in a very distal position, well after the base of RP, and this warrants its allocation to a new genus and species.   descriPtion Preserved part of wing 16.0 mm long, 9.0 mm wide; RA weakly curved and apparently simple; origin of RP proximal to basal third of wing; RP with three branches; area between RA and RP broad; M divided into two main branches proximal to origin of RP; MA simple; MP with two branches; area between M and CuA broad; CuA and CuP separating close to wing base; area between CuA and CuP narrow; CuA divided into CuA 1 and CuA 2 ; CuA 1 and CuA 2 branch proximally; CuA 1 with three branches; CuA 2 apparently simple (Fig. 5).
notes (Table 3) AM11265 differs strongly from the taxa described above in the following characters: RP has at least three long branches; CuA 1 has long branches; and the area between CuA and CuP is narrow. It corresponds to a different genus from the other species found at the site. It shows some similarities with the genus Expartalioma Aristov, 2004 in the branching of RP and CuA, and the narrow area between CuA and CuP. The main difference is the simple MA in our fossil while it has numerous branches in Expartalioma. Paraliomopterum could better fit with our fossil for the same characters as above including a simple MA (Storozhenko 1998;Aristov 2004Aristov , 2019. The difficulty is that the material currently attributed to Paraliomopterum consists of incomplete tegmina (Storozhenko 1998;Aristov 2004Aristov , 2019, thus their exact patterns of venation remain partially conjectural. Paraliomopterum rectum Aristov, 2004 was revised by Aristov (2019) who attributed to this species wings with three branches of MA and of MP and incomplete wings in which it is not possible to determine the number of branches of main veins. Paraliomopterum alium Aristov, 2004 is based on an incomplete wing in which the number of branches of main veins is unknown. Paraliomopterum paulum Storozhenko, 1991 is based on a specimen in which MA and MP are simple (but incomplete) (Storozhenko 1991(Storozhenko , 1998. Paraliomopterum karaungirense Storozhenko, 1991 is based on a hind wing, hardly comparable to the preceding forewings. We provisionally attribute our fossil to the genus Paraliomopterum and not create a new genus, because we prefer to avoid increasing the current taxonomic confusion.

CONCLUSION
We are aware that the assignment of these new species to genera is tentative because the limits and variations in the branching patterns of the main veins remain very poorly investigated among the fossil 'Grylloblattida'. The attribution of a wing to one genus rather than another is mainly a question of personal opinion because of the current lack of phylogenetic analysis of the 'Grylloblattida'. For instance, some genera are separated by authors on the basis of having four branches of CuA vs three, which remains a rather weak character. Nevertheless, with these discoveries, we demonstrate that the Middle Permian diversity of the South African 'Grylloblattida' was significant, with taxa closely related to forms known from the northern part of Pangea. It is not surprising to find many 'grylloblattidan' wing fragments assignable to the family Liomopteridae because it is the largest family within the 'Grylloblattida', consisting of over 45 genera (Storozhenko 1997;Aristov 2004; Fossilworks database site at http://fossilworks.org/bridge.pl). Liomopterids are also the most abundant and diverse 'grylloblattidans' at Permian localities globally (Aristov 2004).
These new taxa confirm that current knowledge of the Permian entomofaunas from the southern part of this supercontinent remains quite poor. Future discoveries in the Middle Permian of South Africa would add valuable knowledge to the entomofaunas of this period. Despite the incompleteness of these wings, the exquisite preservation of the structures and colour patterns show that the outcrop is very promising for this endeavour.