Introduction
The present paper continues a series of articles with descriptions of Diptera from the Late Jurassic Shar Teg site (e.g. Kalugina 1992, Lukashevich 2009). The Upper Jurassic lacustrine deposits of Shar Teg Beds outcrop at Ulan Malgait Mt., 4–5 km west of Shar Teg Mt., 100 km ESE of Altai Somon, Gobi-Altai Aimag, SW Mongolia. The fossil assemblage of Shar Teg includes a diverse and abundant complex of flora and fauna (Gubin and Sinitza 1996).
About 600 identifiable dipteran fossils are known among 3000 fossil insects collected at Shar Teg. Up to now, members of two culicomorph families are described from this locality, Dixidae (Lukashevich 1996) and Chaoboridae (Lukashevich in press). The representatives of Culicidae and Ceratopogonidae (unknown in Jurassic beds), Simuliidae (rare Jurassic finds) and Thaumaleidae (one fossil from Transbaikalia, J3–K1) are not found in Shar Teg.
The Mesozoic records of Chironomidae are numerous, and usually it is the aquatic immatures that are dominant (Kalugina and Kovalev 1985, Jell and Duncan 1986, Kalugina 1993). In Shar Teg, the Chironomidae is one of the most numerically abundant groups: about fifty impressions of adults and twenty pupae and empty pupal exuviae have been collected (undoubted larvae are absent), but due to poor or fragmentary preservation most adults have not been determined even to subfamily. Therefore only several specimens are described herein (pupae will be described later). This chironomid assemblage “very much resembles the stranded corpses of adults and pupae left beside a falling stream, or on the wave swept shore of a lake” (P.S. Cranston, pers. comm.).
The adults of nearly all extant chironomid midges have reduced mouthparts and so their common name is “non-biting midges”. However the presence of toothed mandibles in a chironomid midge was recognized first by Downes and Colless (1967), and now they are described in two recent genera of Podonominae, Archaeochlus Brundin, 1966 and Austrochlus Cranston, 2002 known only from Australia and southern Africa (Cranston et al. 1987, Cranston et al. 2002). Their mandibles closely resemble those of many insectivorous predatory Ceratopogonidae such as Probezzia Kieffer, 1906; however, until females are observed feeding, the question will remain unresolved. A culicomorphan of uncertain affinity with a long proboscis is described from the Upper Triassic Cow Branch Formation (Late Carnian) of Virginia, USA (Blagoderov et al. 2007). Recently, functional mandibulate mouthparts are reported in females (and sometimes even in males) of several extinct genera of Chironomidae from Early Cretaceous Lebanese amber (Azar et al. 2008).
New chironomids with biting mouthparts from Shar Teg are described herein. These fossils are housed in the Borissiak Paleontological Institute, Russian Academy of Sciences, Moscow (PIN). Photographs were made using a Leica MZ 9.5 stereomicroscope with a Leica DFC420 digital camera, with further correction using Adobe Photoshop® CS 9.0 software. Measurements were made with an ocular micrometer in a Leica stereomicroscope. Morphological terminology and measurements mainly follow Sæther (1980). Vein nomenclature is after Wootton and Ennos (1989), followed by Shcherbakov et al. (1995): the chironomid veins traditionally named MCu and An are in fact bM3+4 (tb of Kalugina) and CuP, respectively. For further details regarding the mentioned fossil localities, see Rasnitsyn and Quicke (2002).
Discussion
The four new species of Chironomidae described in this paper are characterized by the elongate proboscis with well-developed (probably sclerotized) mandibles and/or maxillae. To date, a well-developed piercing proboscis has been described only in two recent and no less than four extinct genera of the Chironomidae (Azar et al. 2008). Among the recent Chironomidae, strongly elongate mouthparts are known also in some Orthocladiinae, namely, in both sexes of the North American species Pseudorthocladius macrostomus Soponis, 1980 and Rhinocladius Edwards, 1931 (all three species; distributed in South America and Australia). Their proboscis, superficially resembling that of a mosquito, is formed entirely of the extremely elongated labellae, devoid of stylets and presumably used for sipping nectar, not for piercing (Edwards 1931, Freeman 1961, Soponis 1980). Possibly, a poorly described species Camptocladius nigripectus Bigot 1888 has a similar type of the proboscis (Edwards 1931).
The proboscises of Cretaenne rasnicyni sp. n. and ?Podonomius robustus sp. n. are much longer than in any other fossil Chironomidae described to date (possibly except for an undeterminable culicomorphan from the Triassic Cow Branch Formation; Blagoderov et al. 2007). Among the Chironomoidea, similar strongly elongate mouthparts are known in a number of recent species in many genera of Ceratopogonidae belonging to different lineages of this family, such as Culicoides Latreille, 1809, Echinohelea Macfie, 1940, Atrichopogon Kieffer, 1906, Forcipomyia Meigen, 1818, Leptoconops Skuse, 1889, as well as in extinct species such as the Lower Cretaceous Protoculicoides skalskii Szadziewski, Arillo, 1998, Podonomius punctus Borkent, 2000, and the Upper Cretaceous Culicoides filipalpus Remm, 1976 (e.g. Borkent 2000, Borkent et al. 2009). In nearly all Ceratopogonidae with such mouthparts, females are either insectivorous predators, or blood-suckers on vertebrates, or haemolymph-suckers on insects. However, haemolymph-sucking is restricted to Forcipomyiinae and considered derived feeding mode which appeared in the Cenozoic (Borkent 2000). Nectar-feeding ceratopogonids usually have the stylets more or less reduced, but Forcipomyia (Forcipomyia) brevipennis (Macquart, 1826) considered nectarophagous retains the sclerotized, distinctly toothed mandibles subequal in size to piercing mandibles of its insectivorous congeners (Glukhova 1981). By analogy with Ceratopogonidae we assume that the females of the new chironomid species were entomophagous or haematophagous but secondary nectarophagy cannot be excluded. It is impossible to argue for one of these feeding types, because the fine details of the mouthparts are not discernible in our fossils.
In the general appearance (very long and strong proboscis, body size, shape and proportions; pattern and degree of sclerotization, e.g. strongly sclerotized abdomen and legs) ?Podonomius robustus differs from “typical” Chironomidae as well as from other species assigned to the genus Podonomius andresembles some “robust” Ceratopogonidae, especially many Palpomyiini, but this advanced tribe is unknown from the Mesozoic (Szadziewski 1996). Unfortunately, the posterior part of the wing is not visible in the holotype of ?Podonomius robustus, as well as in the holotype of Podonomius blepharis sp. n., and the presence of a forked M1+2 (characteristic of Ceratopogonidae) cannot be excluded. Thus, the position of these two species seems to be somewhat uncertain. Similar female wings bare of macrotrichia, with well-developed single radial cell, costal ratio more than 0.9 and vein C produced beyond R4+5 and almost reaching wing tip, are known in several Cretaceous species of the ceratopogonid genus Protoculicoides Boesel, 1937, such as Podonomius schleei (Szadziewski, 1996) and Podonomius unus Borkent, 2000 from Lebanese amber (Borkent 2000).
However, in the holotype of ?Podonomius robustus the partly visible transverse vein under r-m is undoubtedly coloured, that is not recorded for the basal part of vein M2 in ceratopogonid wing, but typical for bM3+4 in podonomine wing (Figs 2h–i). In addition, the venation pattern of the anterior part of the wing is more similar to Podonominae than to Ceratopogonidae (vein R1 long, cell r1 long and not narrow) and the vertex of Podonomius blepharis as well as ?Podonomius robustus possesses a coronal suture, which is a feature of the Chironomidae, absent in Ceratopogonidae (Sæther 2000). Moreover, ?Podonomius robustus has a well-developed elongate notum of gonapophysis IX, whereas its absence has been considered as a synapomorphy of the Ceratopogonidae (Sæther 2000), and only some early lineages of Ceratopogonidae have a differently-shaped, short squat notum (vaginal apodeme in Borkent et al. 1987). So we exclude ceratopogonid affinity for both discussed species in spite of the incomplete state of preservation and the strong resemblance to insectivorous predatory or bloodsucking ceratopogonids in the general appearance and consider them as members of Podonominae.
The subfamily Podonominae is shown to be a dominant one in specimen abundance and diversity in the Jurassic deposits of Siberia (Kalugina and Kovalev 1985). No relevant data are available for the other regions. Kalugina stressed the difficulties in differentiation of the Mesozoic Podonominae and Tanypodinae and suggested that often it was possible to classify new taxa with certainty only as members of Tanypodoinae (Tanypodinae + Aphroteniinae + Podonominae) (Kalugina and Kovalev 1985: 82). However, she assigned new genera to subfamilies and explained her choice at every turn. At the same time, Kalugina assumed that some taxa she described in the Podonominae might actually belong to the Tanypodinae, considering that these two subfamilies were less clearly distinguished morphologically in the Jurassic.
Recently, Veltz et al. (2007) have concluded that subfamily identification of Jurassic Chironomidae is impossible and all Podonominae described by Kalugina are Chironomidaeincertae sedis. Their only argument was rather methodical: Veltz and coauthors found it impossible to assign the different life stages to the same species as well as different species known only as pupae, to the same genus. Veltz and coauthors did not discuss the arguments proposed by Kalugina, e.g. that the larvae with translucent thoracic horns and pupae with translucent male genitalia were found among the numerous Jurassic impressions of Oryctochlus Kalugina, 1985, so the association of larvae with pupae was made with certainty and those of pupae and imago, with some doubts. Kalugina compared every life stage of Oryctochlus with those of recent Trichotanypus Kieffer, 1906 and drew a conclusion about an undoubted affinity of these two genera of Podonominae (e.g. in pupae of both genera, segment VIII is deeply emarginated posteriorly, which is remarkably similar to segment IX in shape and segment IX with 3 lateral setae, 2 of which are close together in mid-section). According to the time-calibrated molecular data (Cranston et al. 2010), Trichotanypus is one of the oldest genera of the subfamily, which split from Parochlini in the Early Cretaceous. The French authors did not discuss substantially any genus described by Kalugina. However, they considered that Podonominae may not be recorded in the Mesozoic (Azar et al. 2008), but that hardly can be true. A transfer of Libanochlites Brundin, 1976 (К1, Lebanese amber) from Podonominae to Tanypodinae made by these authors based on their new data was not supported by other specialists (Cranston et al. 2010).