Introduction

Tillyard (1917b) described Mesopanorpa wianamattensis from the Middle Triassic Ashfield Formation of the Wianamatta Group in New South Wales as an aberrant Mecoptera and erected the family Mesopanorpidae. Later (Tillyard 1918) he changed the genus name to Mesopanorpodes to avoid homonymy with Mesopanorpa Handlirsch, 1906. Subsequently (Tillyard 1919b: 611–612) he decided that Mesopanorpodes was closely related to Mesopsyche and the allied genera Aristopsyche, Triassopsyche, and Neuropsyche (the three latter have eventually been synonymized under Mesopsyche by Riek 1956), all described from the Upper Triassic Blackstone Formation, Queensland (Tillyard 1917a; Fig. 1E), and transferred Mesopanorpodes into the order Paratrichoptera.

Confusingly, this remark passed completely unnoticed by all the subsequent authors who discussed the composition and taxonomy of the Mesopanorpodidae. Riek (1953) placed into Mesopanorpodidae two new species of Mesopanorpodes, Mesopanorpodes belmontensis and Mesopanorpodes robustus (Fig. 1C–D), and the genus Prochoristella Riek, 1953 with six species from the Upper Permian of New South Wales (Belmont fossil beds). Since then, Mesopanorpodes and Prochoristella have been considered together either as forming a separate family Mesopanorpodidae (Riek 1955, 1974, Jell and Duncan 1986, Willmann 1978, 1989, Carpenter 1992, Van Dijk and Geertsema 1999, Hong et al. 2002, Hong and Guo 2003, Hong 2007, Sun et al. 2007a, b), or as two closely related genera within Permochoristidae (Martynova 1962, Novokshonov 1997b, Novokshonov et al. 2004).

In fact, species of Prochoristella (with the possible exception of Prochoristella belli Riek, 1953, known only from a hindwing) display characters typical of Permochoristidae, that is, first of all, a deep fork on MA vein and an oblique base of CuA, either fused with MP for a distance, or forming cubito-median Y-vein. Prochoristella and Mesopanorpodes are similar only in the number of RS+MA and MP endings being reduced to four. However, this is characteristic of many independent lineages, e.g. Permotanyderidae + Liassophilidae, Nannochoristidae, or Bittacidae, and cannot be accepted as a reliable apomorphy. I consider Prochoristella, including its possible junior synonyms (or at least closely related genera) Afristella Riek, 1974, Mesotanyderus Riek, 1955, and Austrochoristella Willmann, 1989, as belonging to Permochoristinae (Permochoristidae) and not related to Mesopanorpodes.

Equally unacceptable is the assignment to Mesopanorpodidae of a number of species from the mid-Triassic Tongchuan Fm of Shaanxi Province, China, described by Hong You-Chong and co-authors (Hong et al. 2002, Hong and Guo 2003, Hong 2007). Erdosia pectinata Hong, Guo & Wang, 2002 can be placed with certainty into Parachoristidae, as indicated by the typical pectinate RS and 5- or 6-branched MP. “Mesopanorpodes” shaanxiensis Hong, Guo & Wang, 2002 is based on a highly incomplete hindwing and may belong to either Parachoristidae, Permochoristidae, or Mesopsychidae. The systematic position of Ladinochorista lata Hong, 2007 is uncertain because of the poor preservation of the holotype. Several other taxa described in Mesopanorpodidae do not even belong to Mecoptera. Triassochoristites jinsuoguanensis Hong & Guo 2003 has already been recognized as a member of the dipteran family Vladipteridae (Blagoderov et al. 2007). Forcinerva tongchuanensis Hong and Guo 2003 belongs to the order Miomoptera (as determined by A.P. Rasnitsyn). Allochorista erdosensis Hong, 2007 and Longifurcula hejiafangensis Hong, 2007 cannot be attributed to any particular order based on the published photos and drawings. It should be noted that many descriptions of fossil insects from the Tongchuan Formation, and particularly Mecoptera, are based on inadequately preserved specimens and contain inaccurate drawings and fantastic venational interpretations, in which obvious artifacts are interpreted as characters of specific or even generic rank and hindwings are often misinterpreted as forewings. Thus, these taxa should be considered as of unknown systematic position until revised in detail.

The Jurassic genera Itaphlebia Sukatcheva, 1985 and Chrysopanorpa Ren, 1995 were originally placed in Mesopanorpodidae as well. Novokshonov (1997a, b) synonymized these two genera and transferred them to Nannochoristidae. Itaphlebia appears to be the most common and widespread nannochoristid during the Jurassic. Its diagnosis, composition, and distribution, as well as the overall composition of the Mesozoic Nannochoristidae have been discussed recently by Liu et al. (2010).

Hong (2007) referred Itaphlebia again to Mesopanorpodidae and described Itaphlebia tongchuanensis from the Tongchuan Formation of Shaanxi. However, this species is not related to Itaphlebia, being instead rather similar to Mesopsyche, where it apparently belongs.

Stylopanorpodes eurypterus, Netropanorpodes decorosus, and Netropanorpodes sentosus Sun, Ren & Shih, 2007 (Sun et al. 2007b) from the mid-Jurassic Jiulongshan Formation of Inner Mongolia, China, originally placed in Mesopanorpodidae, are venationally similar to the species of Itaphlebia occurring in the same deposits (Sun et al. 2007c, Liu et al. 2010) and belong to Nannochoristidae as well.

Mesopanorpodes mostovskii Novokshonov et al., 2004 from the basal Triassic of Northern Russia has been transferred to the related family Nedubroviidae (Bashkuev 2011).

Therefore, with the above-mentioned taxa discarded, Mesopanorpodidae remain composed of only three species of Mesopanorpodes from the Upper Permian and Middle Triassic of Australia. But are they actually congeneric?

Re-examination of available photographs of Mesopanorpodes wianamattensis reveals that its CuA is distinctly bent distal to the origin of MP, and the base of CuA is shifted distad and inclined backwards, like in Mesopsyche (Fig. 1A–B). Thus, Mesopanorpodes appears venationally indistinguishable from Mesopsyche. Here I provisionally retain Mesopanorpodes as a separate genus until the complete revision of Mesopsyche. Unlike Mesopsyche wianamattensis and Mesopsyche, both species described by Riek have the base of CuA short and transverse, with the intercubital space not expanded, which is presumed to be a plesiomorphic condition. These species are included here in the new genus Permopsyche gen. n.

In this paper, two new species of Permopsyche, Permopsyche issadensis sp. n. (type species)and Permopsyche rasnitsyni sp. n., are described from the Upper Permian of European Russia. These are the oldest currently known records of Mesopsychidae. The new material includes a number of well-preserved fore- and hindwings, which allow establishing yet another difference between Permopsyche and Mesopsyche, the position of the hindwing CuA vein with respect to the membrane. In Permopsyche issadensis CuA is concave (while the R1 is convex), as is typical of Permochoristidae, while in Mesopsyche spp., including the Permian species, as well as the other Mesozoic mesopsychid genera, both CuA and R1 are convex, like in most post-Paleozoic Mecoptera.

The earliest record of Mesopsyche, and the only one known so far to be definitely pre-Triassic, Mesopsyche incompleta sp. n., is described from the uppermost Permian of the European Russia.

Bittacopanorpa javorskii Zalessky, 1936 from the transitional Permian–Triassic sequence of the Kuznetsk Basin (Babiy Kamen’ locality, Maltsevo Formation), known only from a wing base (Fig. 8G–H), originally placed in Neorthophlebiidae and tentatively transferred to Permotipulidae (Bechly and Schweigert 2000), can be identified as a hindwing with venation almost identical to that of the above-mentioned Mesopsyche incompleta sp. n. (Fig. 8B).

Material and methods

The study is based on examination of ca. 60 fossil specimens from the collection of Borissiak Paleontological Institute, Russian Academy of Sciences, Moscow (PIN). The holotype of Bittacopanorpa javorskii Zalessky is preserved in the collection of the Central Research Geological Prospecting Museum, Saint Petersburg (CRGPM).

Fifty-three isolated wings of Permopsyche were collected from a lens of fluvio-lacustrine deposits within the Sukhona River section, in Vologda Province, north-central Russia (Isady locality). It corresponds to the Kalikino Member of the Poldarsa Formation, and is dated the latest Severodvinian (correlated with the Lower Wuchiapingian: Golubev in press).

Five specimens of Mesopsyche were collected from the Balymotikha locality, an outcrop of lacustrine deposits of Vyazniki Permian–Triassic sequence, at the town of Vyazniki, Vladimir Province, central Russia. It is dated the latest Vyatkian (Late Changhsingian), somewhat below the Permian–Triassic boundary (Newell et al. 2010).

The fossils were examined with a Leica M165C stereomicroscope and photographed using an attached Leica DFC 425 digital camera. Images were digitally processed with Helicon Focus v. 5.1 and Adobe Photoshop CS3 graphic software. Line drawings were made using Inkscape v. 0.48 vector graphics editor.

On the line drawings the wings are oriented with their apices to the right, while some photographs are reversed left-right for better comparison.

The wing venation terminology follows Novokshonov (2002).

Systematic paleontology Discussion

The systematic position of Mesopsychidae and the entire “suborder” Paratrichoptera was discussed by Novokshonov and Sukatcheva (2001). Recent studies of complete mesopsychid fossils from the Middle Jurassic and Lower Cretaceous of China revealed the long siphonate mouthparts indicating fluid feeding on reproductive organs of gymnosperm plants (Ren et al. 2009, 2010). Mesopsychidae, together with mid-Mesozoic Aneuretopsychidae, Pseudopolycentropodidae, and now the Late Permian to Early Triassic Nedubroviidae (Bashkuev 2011), form a distinct long-proboscid clade within Mecoptera, the Aneuretopsychina.

The finds described herein reveal main trends in the evolution of Mesopsychidae after their separation from the permochoristid-like ancestor. These are:

– Change in the forewing CuA base inclination and position: from proximal and inclined forwards (typical of Permochoristidae) to inclined backwards, and probably also shifted distad.

– Change in position of the hindwing CuA: from concave (Permopsyche) to convex (most Mesopsychidae).

– Loss of fusion of CuP and A1 veins in the hindwing: from quite basally fused in Permopsyche, through a rudimentary basal fusion in Mesopsyche (as indicated by Novokshonov (1997c) in Mesopsyche shcherbakovi), to becoming secondary free along their whole length in the mid-Jurassic and early Cretaceous genera.

– Increase in the average wing size, from 5–6 mm in the Late Permian to 20–30 mm since the Late Triassic.

The Permian Mesopsychidae occurred together with Nedubroviidae (in Isady, Balymotikha, and probably Belmont; Bashkuev 2011), which already have typical long siphonate mouthparts, and the possession of those in early mesopsychids (and even in the ancestral permochoristid lineage) can be assumed as well.

New finds also cast light on the stratigraphic distribution of the Aneuretopsychina scorpionflies, which appear to be the most common group of Mecoptera during the latest Permian – early Middle Triassic time interval (viz. Mesopsychidae, Nedubroviidae, and later also Pseudopolycentropodidae). They initially replaced Permochoristidae in the end of Permian, and were in turn progressively supplanted during the Triassic – Early Jurassic by Parachoristidae and Orthophlebiidae, which gave rise to the modern mecopteran lineages. In the mid-Jurassic – Lower Cretaceous insect assemblages, Aneuretopsychina scorpionflies are rather rarely found, constituting only a minor component of the mecopteran faunas.