Introduction

The Pseudopolycentropodidae is an extinct and relatively nonspeciose family considered as phylogenetically basal to recent Mecoptera (Handlirsch 1925, Bode 1953, Willmann 1989, Carpenter 1992, Novokshonov 2002). Novokshonov (1997a) provided a general description of Pseudopolycentropus latipennis Martynov, 1927, then the best known species, and stated that all structural details of Pseudopolycentropus latipennis appear typical of the entire family except forewing shape. Grimaldi and colleagues (2005) comprehensively reviewed pseudopolycentropodid scorpionflies; they included additional details of seven previously described species of Pseudopolycentropus and their spatiotemporal distributions, and described the new taxa of Pseudopolycentropus daohugouensis Zhang, 2005, Pseudopolycentropodes virginicus Grimaldi & Fraser, 2005, Parapolycentropus burmiticus Grimaldi & Rasnitsyn, 2005, and Pseudopolycentropus paraburmiticus Grimaldi & Rasnitsyn, 2005. This contribution considerably expanded knowledge of the Pseudopolycentropodidae, and provided inferences on the possible diets of these bizarre, mid-Mesozoic insects.

Currently, the Pseudopolycentropodidae consists of thirteen described species assigned to three genera from the mid-Triassic to the mid-Cretaceous (Whalley 1985, Papier et al. 1996, Ansorge 1996, 2003, Novokshonov 1997a, 1997b, Grimaldi et al. 2005, Ren et al. 2009, Ren et al. 2010a). Five of these described species were based on fossils of forewings only, and one was based on fore- and hindwings. Of the remaining, seven species were described from wings and bodies; only five of which have preserved mouthparts, summarized in Table 1. They are Pseudopolycentropus latipennis (mid Late Jurassic of Karatau, Kazakhstan); Pseudopolycentropus daohugouensis, Pseudopolycentropus janeannae Ren, Shih & Labandeira, 2010, Pseudopolycentropus novokshonovi Ren, Shih & Labandeira, 2010 (all from the late Middle Jurassic of Daohugou, Inner Mongolia, China), and Parapolycentropus burmiticus (late Early Cretaceous of Kachin Province, northern Myanmar).

These five scorpionfly taxa are highly significant because Pseudopolycentropus possessed distinctive, elongate tubular, or siphonate, proboscides for surface fluid feeding on exposed plant fluids, such as the pollination drops of seed plants (Ren et al. 2009, Labandeira 2010). By contrast, one known species of Parapolycentropus bore a proboscis housing styliform structures. Although this genus, like other Pseudopolycentropodidae, probably imbibed ovulate fluids of seed plants (Grimaldi et al. 2005; Labandeira et al. 2007; Ren et al. 2009), the presence of one or perhaps two unserrated, styliform structures ensheathed within a proboscis is anomalous. One explanation is that this species of Parapolycentropus may have fed in a manner similar to a technical ink drawing pen with a central elongate wire for assisting laminar flow when food tube diameters are very narrow. Several genera of long-proboscid clades of Diptera have similar food tube arrangements, in which a thin, centrally positioned, rod-like but flexible hypopharynx extends to almost the proboscis terminus (Nagatomi and Soroida 1985, e.g. Figs 409, 425, 437). Such styliform processes lack serrations or apical piercing structures, and would be inconsistent with a blood feeding diet mentioned by Grimaldi et al. (2005).

Ren and colleagues (2009, 2010a, 2010b, 2011) reported that three families of mid Mesozoic, Eurasian Mecoptera had particularly elongate siphonate proboscides. They were the Mesopsychidae (Lichnomesopsyche gloriae Ren, Labandeira & Shih 2010, Lichnomesopsyche daohugouensis Ren, Labandeira & Shih 2010, and Vitimopsyche kozlovi Ren, Labandeira & Shih 2010), Aneuretopsychidae (Jeholopsyche liaoningensis Ren, Shih & Labandeira 2011), and the afore-mentioned Pseudopolycentropodidae (Pseudopolycentropus janeannae and Pseudopolycentropus novokshonovi). Structural details of these and other long-proboscid scorpionfly taxa are listed in Table 1, indicating together with other insect and botanical evidence, that these taxa fed on gymnospermous ovulate fluids such as pollination drops and likely engaged in pollination mutualisms with their host plants (Ren et al. 2009, Labandeira 2010).

The long proboscides of Mecoptera, in addition to other clades such as nemestrinid flies, seem to have originated during a 15 million-year interval during the mid Jurassic from a 170 to 155 Ma (Ren et al. 2009), indicated in part by earlier, Late Triassic pseudopolycentropodids lacking proboscides. However, this burst of long-proboscid origination is preceded by two considerably antecedent lineages that bear small but long proboscides, both from the Permian of Russia: the relatively small neuropteran Permithonidae, Tschekardithonopsis, from Early Permian Chekarda of Perm Province (Labandeira 2010), and the even smaller, mecopteran Nedubroviidae from Late Permian Isady of Vologda Province (Bashkuev 2011). Although much poorly known than the aneuretopsychine Mecoptera of the later mid Mesozoic, both the Late Permian and mid Mesozoic long-proboscid assemblages considerably antedate the similar and independent associations of nectar-feeding by flies, moths, and beetles on angiosperms several tens of millions of years later (Grimaldi 1999, Labandeira et al. 2007, Labandeira 2010). All three scorpionfly families became extinct during the later Early Cretaceous to mid-Cretaceous, coincident with global gymnosperm-to-angiosperm turnover (Ren et al. 2009, Labandeira 2010).

Adding to this inventory of long-proboscid scorpionflies, we recently collected a well-preserved fossil pseudopolycentropodid from the Middle Jurassic Jiulongshan Formation in Daohugou Village, Ningcheng County, Inner Mongolia, China. Based on its unique combination of antennae, mouthparts, and wing venation, a new genus and species is erected herein.

Geological and paleobiological context

The Jiulongshan Formation is a lacustrine sequence that crops out near Daohugou Village, Shantou Township, Ningcheng County, in Inner Mongolia of northeastern China (41°19.532' N, 119°14.589' E) (Ren et al. 1995, Ren et al. 2002). The section at Daohugou is composed of grey tuffaceous sandstone and sandy mudstone. The paleoenvironment reconstructed for this locality was a volcanic region with montane streams and lakes (Ren et al. 2002). Daohugou has yielded an abundant and diverse insect fauna in addition to the Mecoptera, (Ren et al. 2010c) consisting of complete specimens of Ephemeroptera, Odonata, Plecoptera, Blattodea, Grylloblattida, Dermaptera, Orthoptera, Phasmatodea, Heteroptera, Sternorrhyncha, Neuroptera, Raphidioptera, Coleoptera, Hymenoptera, Diptera, Trichoptera and Lepidoptera. Apart from insects, Daohugou also has produced spiders, freshwater conchostracans, salamanders, feathered dinosaurs, pterosaurs, and mammals.

There is considerable evidence for a diverse local flora at Daohugou, which was important for plant interacting insects, such as pseudopolycentropodid scorpionflies. The Jiulongshan Formation has provided evidence for the floral composition of the surrounding forests. These forests were dominated by arborescent seed plants, principally Coniferopsida (Pityophyllum, Rhipidiocladus, Elatocladus, Schizolepis, Podozamites), Ginkgopsida (Ginkgoites, Ginkgo, Baiera, Czekanowskia, Phoenicopsis), Cycadopsida (Pseudoctenis, Zamites), and Bennettitopsida (Anomozamites) (Mi et al. 1996). Lower statured plants, including herbaceous ground cover, consisted of Lycopsida (Lycopodites, Selaginellites), Sphenopsida (Equisetum), and Filicopsida (Todites, Coniopteris) (Mi et al. 1996). These paleobotanical data were interpreted as indicating humid and warm-temperate climate (Tan and Ren 2002), although ecological amplitudes of specific habitats have yet to be determined. Verifiable 40Ar/39Ar and SHRIMP 238U/206Pb dating shows that the age of the volcanic rocks overlying the Daohugou fossil-bearing beds is ca. 164-165 Mya, and consequently the age of this fossil-bearing beds is slightly older than or equal to 165 Mya (Chen et al. 2004). Combined with the above-mentioned composition of insect fauna and conchostracans, the age of Daohugou biota is considered as Middle Jurassic (Zhang et al. 1987, Ren et al. 1995, 2002, Wang et al. 2000, Shen et al. 2003), equivalent to the Bathonian–Callovian boundary interval (Ogg et al. 2008).

Materials and methods

This study is based on a fossil specimen (CNU-MEC-NN-2010044 p/c), with part and counterpart, housed in the fossil insect collection of the Key Lab of Insect Evolution and Environmental Changes, College of Life Sciences, Capital Normal University, Beijing, China (CNUB; Dong Ren, Curator). The specimen was examined dry or under alcohol using a Leica M165 C dissecting microscope, and illustrated with the aid of a drawing tube attachment. Photographs of specimens were taken by Leica dfc500 and line drawings in Figure 1 were made by CorelDraw 12. The drawing in Figure 2 was done as a camera lucida tracing that subsequently was inked on polyester film and then reduced in size. Illumination for the drawing of this specimen consisted of three types of light and variation in light angle and origin for accentuation of morphological features, such as wing venation, that normally were difficult to observe and typically unavailable at lower magnification microscopes. The morphological terminology used herein is that of Grimaldi et al. (2005).

Systematic paleontology Family Pseudopolycentropodidae Handlirsch, 1925 Discussion

Three aspects of this discovery are significant for understanding the ecological roles of Pseudopolycentropodidae with plants in the local ecosystem at Daohugou. First is recognition of the distinct morphological features that separate Sinopolycentropus gen. n. from all other coexisting pseudopolycentropodid taxa. Second are the implications that the unique antennal and especially mouthpart modifications have for host-plant use. Last is the importance of rarity in understanding the pollinator associations in an increasingly well-documented, preangiospermous ecosystem from the Middle Jurassic.

Distinctiveness of Sinopolycentropus from other Pseudopolycentropodidae

This new, long-proboscid species is distinct from all other members of Pseudopolycentropodidae by several differentiating features. These differences are the base of the Sc vein merging with R vein; the R2+R3 vein forking earlier than that of the R4+R5 vein; and relatively short, moniliform, and robust antennae bearing hairs on their annulae. To date, all described pseudopolycentropodids have their R4+R5 vein forking earlier than that of their R2+R3 vein, except for Pseudopolycentropus triasicus Papier, Nel & Grauvogel-Stamm, 1996, which has its R4+R5 vein forking at variable levels vs. its R2+R3 vein (Papier et al. 1996).

The presence of small, distinctive, three-segmented labial palps is only shared with Parapolycentropus burmiticus (Grimaldi et al. 2005). We reinterpret the proboscis sheath and palps of Parapolycentropus burmiticus (Grimaldi et al. 2005) as labial, rather than maxillary, in origin (Ren et al. 2009), a tendency in other Antliophora, such as the Diptera (Chaudonneret 1990). However, the proboscis stylets in Parapolycentropus, absent in Sinopolycentropus, are likely derivatives of maxillary laciniae or galeae.

Typically, pseudopolycentropodids have filiform or moniliform antennae with a relatively high ratio of antenna length to unappendiculate body length. For example, Pseudopolycentropus daohugouensis was reported to have an antenna/body length ratio somewhat greater than 0.37 (2.8/7.5); Pseudopolycentropus janeannae with greater ratios of 0.5 (3.5/7) and 0.57 (4/7); and Pseudopolycentropus novokshonovi with an intermediate ratio of 0.43 (3/7). However, Parapolycentropus burmiticus has a distinctively different type of antennae, characterized by a funnel-shaped scape, a scoop-shaped pedicel, the basal five flagellomeres tapered in size and the apical eleven segments significantly more diminutive, forming an arista (Grimaldi et al. 2005), all of which amount to an antenna/body length ratio of 0.4 (1.2/3). By contrast, Sinopolycentropus rasnitsyni gen. et sp. n. possesses compact, moniliform, and robust antennae, but with a relatively smaller antenna/body length ratio of 0.36. Last, the presence of distinctive, short, three-segmented labial palps, laterally placed at the proboscis base, is a condition not found in other, palpless pseudopolycentropodids, except for Parapolycentropus (Grimaldi et al. 2005, Ren et al. 2009).

Significance of antennal and mouthpart modifications

Structural features of the antennae and mouthparts may be relevant for seeking conspecifics and host plants. The unique characteristics of antennal shape, structure and length in Sinopolycentropus rasnitsyni might be associated with the sensory detection of mates or food sources. The presence of prominent hairs encircling the annular area of each antennal article may be involved in detection of conspecific pheromonal cues or specific chemicals from particular plant hosts. Currently, we cannot draw any conclusions regarding these suggestions, pending scanning electron or high-resolution light microscopy of the antennal hair setal bases and other features. Of more importance are the structure, size, and shape of the proboscis, which indicate that it would have been used for access and imbibition of pollination drops or similar ovulate fluids from a variety of smaller gymnospermous fructifications (Ren et al. 2009, Labandeira 2010). The purpose of comparatively short labial palps, shared only with one other Mid-Cretaceous amber taxon with similar mouthparts and presumed feeding on ovular fluids, could have been for sensory detection of appropriate chemical cues from particular host plants. Such cues would have included fructification exudations or pollination-drop scents, perhaps similar to that of extant insects and their cycad and angiosperm hosts (Pellmyr and Thien 1986, Terry et al. 2004).

The rarity of pseudopolycentropodid pollinators

In the extensive collection of more than 250, 000 fossil insect specimens from Daohugou at Capital Normal University, we currently have collected seven specimens of Pseudopolycentropus janeannae, one specimen of Pseudopolycentropus novokshonovi, and a single specimen of Sinopolycentropus rasnitsyni. This suggests that these three species are extremely rare compared to other, co-occurring insect taxa. Significantly, approximately 20 specimens of small, nonglossate moths also have been found in this collection, of similar size and ecological relationships with plants as the pseudopolycentropodids. This comparison indicates that the ecologically equivalent pseudopolycentropodids are considerably rarer, and by inference, may have had more specialized associations with host plants than those of moths. These low abundances indicate that rarity, indeed, was an important feature of the Middle Jurassic Jiulongshan insect fauna and flora, contrary to the viewpoint that rarity was only a feature of angiosperm-dominated biotas from the mid Cretaceous and younger (Vermeij and Grosberg 2010). In addition to the Pseudopolycentropodidae, it also should be noted that two other mecopteran lineages with long-proboscid, siphonate mouthparts—the Mesopsychidae and Aneuretopsychidae—also had likely associations with gymnospermous seed plants, including pollination mutualisms (Ren et al. 2009, Labandeira 2010). The variability of proboscis lengths and widths, diameters of proboscis food canals, and other associated features of the mouthparts, antennae and wings of these three lineages, summarized in Table 1, indicate a similar variety of accessible plant hosts. Although the proboscis of Sinopolycentropus rasnitsyni is slightly longer than other pseudopolycentropodid species, nevertheless it is within the range of other confamilial taxa, indicating that the Pseudopolycentropodidae were rare, specialized associates of gymnospermous plants during the Middle Jurassic at Daohugou.