Introduction

The Tettigarctidae, known as hairy cicada, is the most primitive group of the superfamily Cicadoidea, comprising only two relict modern species within a sole genus (Tettigarcta tomentosa White, 1845 and T. crinita Distant, 1883) restricted to the mountains of Tasmania and South-East Australia (Moulds 1990; Shcherbakov 2009). The oldest record of Tettigarctidae is from the terminal Triassic of England (Whalley 1983; Shcherbakov and Popov 2002). The Mesozoic fossils of Tettigarctidae are relatively rich in Eurasia, Australia, Africa and South America (see Wang and Zhang 2009; Li et al. 2012; Chen et al. 2014). However, only few Cenozoic records were reported; from the Paleocene of France (Meuniera haupti; Piton 1936), the Eocene of Scotland (Eotettigarcta scotica; Zeuner 1944) and Germany (Tettigarctidae gen. et sp. indet.; Wappler 2003), and the Lower Miocene of New Zealand (Paratettigarcta zealandica; Kaulfuss and Moulds 2015).

Because of high diversity but poor preservation, the Mesozoic tettigarctids have a controversial taxonomic history and their evolutionary history remains poorly understood. Cicadoprosbole Becker-Migdisova, 1947 was transferred from Prosbolidae to Tettigarctidae by Becker-Migdisova (1949), and then separated at the family level by Evans (1956). Some researchers treated Cicadoprosbolidae as independent family (e.g., Hamilton 1990, 1996), but others considered it as a subfamily of Tettigarctidae (e.g., Boulard and Nel 1990; Nel et al. 1998; Menon 2005). Because diverse Mesozoic forms fill the morphological gap, it is reasonable to synonymize these two taxa (Shcherbakov 2009). The Tettigarctidae sensu lato was revised in detail by Shcherbakov (2009), and it was divided into two subfamilies, Cicadoprosbolinae with tribes Cicadoprosbolini, Architettigini and Turutanoviini, and Tettigarctinae with tribes Protabanini, Meunierini, and Tettigarctini, respectively. However, recent discoveries of Mesozoic tettigarctids with well-preserved body structures and complete forewings suggested that the current taxonomic system, just on basis of forewing venation, might not effectively reflect evolutionary relationships (Wang 2009; Li et al. 2012; Chen et al. 2014; this study). Based on two whole-bodied adults and one nymph from the Lower Cretaceous Crato Formation of northeastern Brazil, Hamilton (1990) described the tettigarctid genus Architettix (Architettigini) with some body and wing characters distinctly different from known extant and extinct tettigarctids: tarsi cicadellid-like and forewing with M3+4 unbranched. Hamilton (1990, 1996) considered that the genus shows a mixture of cicadoid and cicadellid characters, suggesting these two cicadomorph groups are closely related. Recently, some fossil tettigarctids were collected from the upper Middle—lower Upper Jurassic Daohugou fossil-bearing strata of Inner Mongolia, China (Fig. 1). These new tettigarctids possess an unusual forewing with M3+4 unbranched as in Architettigini, but their tarsus and other body structures are similar to Shuraboprosbole spp. (Turutanoviini) from the contemporaneous strata at Daohugou Village.

Fig. 1.

Distribution of the localities and strata of Architettigini, Turutano viini, and Sanmai gen. nov. 1, northeastern Brazil (Hamilton 1990); 2, England (Whalley 1985); 3, Kazakhstan (Shcherbakov and Popov 2002; Shcherbakov 2009); 4, Tajikistan (Becker-Migdisova 1949); 5, Ust'-Baley, southern Siberia (Shcherbakov 1985); 6, Myangad, western Mongolia (Shcherbakov 1986); 7, northeastern China (Wang and Zhang 2009; this study).

Although tettigarctids were rich in the Mesozoic, only one specimen of Turutanoviini with both forewing and hindwing preserved was reported to date (Shcherbakov and Popov 2002; Shcherbakov 2009). A new genus Sanmai with three new species are described herein based on six fossil specimens with well-preserved body structures, forewing and hindwing venations, making it the hitherto best known extinct tettigarctid taxon, and providing some new information about the evolutionary history and paleobiogeography of the Mesozoic Tettigarctidae. Additionally, we discuss herein the special disruptive coloration preserved on these new tettigarctids, which is remarkably different from that of known extinct and living tettigarctids as well as other cicadomorphs discovered in the same fossil layers of Daohugou.

Institutional abbreviations.—STMN, Shandong Tianyu Museum of Nature, Pingyi, Shandong Province, China.

Other abbreviations.—A, anal vein; CuA, anterior branch of the cubitus vein; cua-cup, crossvein connecting the anterior and posterior branches of the cubitus vein; CuP, posterior branch of the cubitus vein; im, crossvein connecting the branches of the media veins; ir, crossvein connecting the anterior and posterior branch of the radial vein; M, media vein; m-cua, crossvein connecting the fourth branch of the media vein and the anterior branch of the cubitus vein; RA, anterior branch of the radial vein; RP, posterior branch of the radial vein; r-m, crossvein connecting the posterior branch of the radial vein and the first branch of the media vein; ScP, posterior branch of the subcosta vein; u, ulnar cell (terminology after Moulds 2005).

Material and methods

The new Mesozoic tettigarctids described herein were collected from the Daohugou fossil-bearing strata of northeastern China, and are deposited in STMN.

Daohugou has yielded abundant and diverse fossil plants and animals, and is now considered to be one of the most important insect Lagerstätten (Rasnitsyn et al. 2006). The hitherto known Tettigarctidae is much more diverse in Daohugou than in any other faunae, with eight species described and attributed to four genera and some available fossils awaiting description (Wang 2009; Wang and Zhang 2009; Li et al. 2012; Chen et al. 2014). Generally, the fossil-bearing beds at Daohugou were placed within the Jiulongshan Formation of Bathonian—Callovian (late Middle Jurassic) (e.g., Wang and Zhang 2009; Chen et al. 2014; Liu et al. 2014). However, recent isotopic dating results indicated that the Daohugou beds were deposited in the geological age of 164–158 Ma (Liu et al. 2006, 2012; L. Wang et al. 2013), Callovian—Oxfordian (latest Middle—earliest Late Jurassic) according to the updated International Chronostratigraphic Chart (Cohen et al. 2013; Wang et al. 2015).

There is no consensus on the interpretation of vein nomenclature in Cicadomorpha (Wang and Zhang 2009). Nel et al. (2012) proposed a new interpretation of wing venation pattern for all Paraneoptera, assuming that CuA gets fused with M+R stem at wing base and connected with CuP by a specialized crossvein cua-cup after its departure from M+R, which is remarkably different from the traditional interpretations. The venational terminologies used herein follow Nel et al. (2012), including that the first longitudinal vein on the clavus is tentatively treated as A1 rather than Pcu, that is treated as open problem in Nel et al. (2012).

The fossil tettigarctids were examined dry or under alcohol, with details observed and microphotographed under a stereomicroscope (ZeissSteREO Discovery V8). Photographs were taken using a NikonD800 digital camera. Line drawings were prepared with image-editing software. All measurements were made using software ImageJ 1.42q (Wayne Rasband; National Institute of Health, USA). The following standards were used for measurements: whole length measured from the apex of the vertex to the apex of the tegmen; length of tegmen measured from the base to the apex; width of tegmen measured at the widest part from costal margin to posterior margin; length of clavus measured from the base of tegmen to the ending of vein CuP; width of clavus measured at its widest part; length of costal area measured from the base of tegmen to the ending of vein ScP; width of costal area measured at its widest part.

Systematic palaeontology

Fig. 2.

Hairy cicada Sanmai kongi sp. nov. from the upper Middle—lower Upper Jurassic Daohugou beds. A. Holotype STMN48-1800a. Photograph under alcohol (A₁), explanatory drawing (A₂). Hind leg (A₃). Enlargement of apical teeth set of hind tibia (A₄). Photomicrograph of ovipositor (A₅). B. Paratype STMN48-1801. Photograph (B₁), photomicrograph of male genitalia (B₂). Abbreviations: A, anal vein; CuA, anterior branch of the cubitus vein; CuP, posterior branch of the cubitus vein; M, media vein; RA, anterior branch of the radial vein; RP, posterior branch of the radial vein; ScP, posterior branch of the subcosta vein; u, ulnar cell.

Fig. 3.

Hairy cicada Sanmai mengi sp. nov., holotype (STMN48-1802) from the upper Middle—lower Upper Jurassic Daohugou beds. Photograph (A), explanatory drawing (B).

Fig. 4.

Hairy cicada Sanmai xuni sp. nov. from the upper Middle—lower Upper Jurassic Daohugou beds. A. Ho lotype STMN48-1803. Photograph (A₁), explanatory drawing (A₂), enlargement of head (A₃). B. Paratype STMN48-1804. Photograph (B₁), explanatory drawing of hind wing (B₂, horizontal mirror), photomicrograph of ovipositor and pygofer (B₃). C. Paratype STMN48-1805. Photograph (C₁); photomicrograph of antenna, showing segments of flagellum (arrowheads) (C₂); photomicrograph of part of rostrum (C₃). Abbreviations: CuA, anterior branch of the cubitus vein; M, media vein; RA, anterior branch of the radial vein; RP, posterior branch of the radial vein.

Discussion

Evolutionary implications.—The new genus seems to be closely related to the tribes Turutanoviini and Architettigini of Cicadoprosbolinae, and the three taxa seem to form a monophyletic group based on the following forewing characters; most apical cells extremely long and narrow, and nearly longitudinal, RA at most possessing three terminations, and distal section of CuA2 running along wing margin. Considering many dissimilarities of Sanmai gen. nov. from the tribe Architettigini, the M3+4 fork might be lost in parallel in the new genus and Architettigini. The alternative scenario, however, can not be excluded that Sanmai gen. nov. is a transitional form between widespread Turutanoviini and specialized Architettigini, since it possesses common body structures as Shuraboprosbole spp. of Turutanoviini but shares a simple M3+4 vein on forewing with Architettigini. Sanmai gen. nov. is provisionally assigned into the tribe Turutanoviini herein because the new genus and the genera of Turutanoviini share many common morphological characters besides the number of terminations of M vein.

Up to now, three genera (i.e., Turutanovia, Paraprosbole, Shuraboprosbole), with rich fossil record from the Lower Jurassic to the Upper Cretaceous of Eurasia, have been attributed to Turutanoviini (Fig. 1). Shuraboprosbole is very similar to Paraprosbole, but differs from the latter in possessing a larger forewing with costal margin strongly arched near wing base, and stem R and M+CuA bifurcating nearly at the same level. Some tettigarctid fossils recently discovered from the Daohugou Lagerstätte have filled the morphological gap between Shuraboprosbole and Paraprosbole, and some of these genus-level diagnostic differences are even intra-specific in Daohugou specimens, such as the relative bifurcating position of stem R and M+CuA in forewing of Shuraboprosbole daohugouensis (Wang and Zhang 2009). These two genera do not possess distinguishable morphological characters. Therefore, the genus Paraprosbole should be synonymized with Shuraboprosbole. Although the records of Mesozoic tettigarctids are relatively rich, only fossil specimens of two genera, Shaanxiarcta (forewing unknown) and Turutanovia, preserve information of hindwing to date. The hindwing of Sanmai gen. nov. is distinctly different from that of these two genera as well as extant Tettigarcta by simple M₁₊₂ and bifurcated M₃₊₄, but very similar to that of Shuraboprosbole media (unpublished specimen; see SOM: fig. S1, Supplementary Online Material available at  http://app.pan.pl/SOM/app61-Chen_etal_SOM.pdf). If Sanmai gen. nov. was a transitional form between Turutanoviini and Architettigini we can infer, based on morphological and palaeobiogeographical data, that tettigarctids with M threebranched on forewing (Architettix and Sanmai) probably originated from some Shuraboprosbole-related tettigarctids in Eurasia (Laurasia) in the Middle Jurassic, and then spread to South America (Gondwana) before the expanding Tethys Ocean became a geographic barrier for these flying insects.

Compared to living tettigarctids, Architettigini (Architettix) possesses some distinct body structures, such as broadly exposed mesonotum, long hind tibia, and cicadellidlike tarsi. Architettix possesses elongate hind basitarsomere with rows of plantar setae like modern cicadellids (leafhoppers). Hamilton (1990, 1996) considered structures of “cicadellid- like tarsi” as synapomorphies linking the Membracoidea and Cicadoidea. In modern Clypeata, only leafhoppers have the elongate hind basitarsomere with rows of plantar setae, and this morphological character is considered important for high level classification of leafhoppers (Davis 1975). The present study shows that the closely related Sanmai has common tarsi similar to modern Cicadoidea (Figs. 2, 3; Emeljanov 1987). Recent molecular and fossil evidences have confirmed that modern Clypeata comprises two monophyletic clades, divided into Myerslopoidea + Membracoidea and Cicadoidea + Cercopoidea, respectively (Shcherbakov and Popov 2002; Cryan and Urban 2012; Wang et al. 2012). Hence, the similarity of the tarsi between Cicadellidae and Architettigini is undoubtedly the result of convergent evolution. The specialized body structures of Architettigini are probably due to predation pressure from some early birds and insectivorous pterosaurs reported from the same formation of northeastern Brazil (Kellner 2002; Naish et al. 2007; Unwin and Martill 2007; Sayao et al. 2011; Carvalho et al. 2015). Architettigini, with long hind tibia and well-developed laterotibial spines, must be adapted to jumping (Shcherbakov 2009) as modern cicadellids. Meanwhile, the huge and broadly exposed mesonotum probably suggests an improvement of flight ability, like modern singing cicadas (Cicadidae), since the forewing is connected to flight muscles in mesonotum (Yao 2012). The adults of modern tettigarctids are crepuscular (hiding under loose bark during daylight hours; Moulds 1990; Shcherbakov 2009), but some structurally diverse Mesozoic tettigarctids (e.g., Architettix) with better locomotor capacity were likely diurnal.

Coloration pattern.—The disruptive coloration of forewing is prominent on some specimens of Sanmai gen. nov. (Figs. 2A, 4A, D), with light and irregular speckles and longitudinal stripes bold contrasting to dark membrane, and the type is new to Tettigarctidae. The mottled forewings of the relict modern Tettigarcta spp. blend with the lichen-covered eucalypt bark (Moulds 1990). The transverse alternating dark and light stripes on wings, as a common disruptive coloration type for fossil cicadas (Wang et al. 2010), were reported in some Mesozoic tettigarctid specimens, e.g., Protabanus (Hong 1982), Liassocicada (Nel 1996), and Shuraboprosbole (Wang and Zhang 2009). Most forewing membranes of Mesozoic tettigarctids, however, are almost monochrome (e.g., Sunotettigarcta hirsuta) or colorless and transparent (e.g., Turutanovia karatavica). The disruptive coloration of Sanmai gen. nov. might be an effective camouflage which enabled Sanmai. to reduce the detectability of arboreal and insectivorous predators from the same Lagerstätte, such as ancestral Avialae, gliding mammals, and rhamphorhynchoid pterosaurs.

The special disruptive coloration type seems to be an autapomorphy of Sanmai. In specimens STMN48-1800 of S. kongi sp. nov., STMN48-1803 and STMN48-1804 of S. xuni sp. nov., the coloration pattern of forewing is well preserved and similar in general: some speckles in various sizes and shapes on prenodal area and some longitudinal stripes on postnodal area. The color impression of the remaining specimens of S. kongi sp. nov. and S. xuni sp. nov. as well as the holotype of S. mengi sp. nov. is obscure or even absent, likely weakened or erased by diagenetic process. Hence, considering taphonomic factors, the presence or absence of forewing coloration is not a useful distinguishable character for Sanmai.

The Daohugou palaeolake was a low-energy preservational environment. Most well-preserved tettigarctid specimens collected from Daohugou area are preserved in the lateral position with all four wings folded (Wang and Zhang 2009; Li et al. 2012; B. Wang et al. 2013; Chen et al. 2014; this study). The large body with a greatly inflated frontoclypeus and a long rostrum indicates that the Mesozoic tettigarctids were arboreal and xylem feeding (Wang and Zhang 2009). Therefore, it seems that when the early tettigarctids stayed on the branch to suck xylem fluids, their folded and rooflike wings covered bodies just as extant Cicadoidea, and so the forewing color pattern was essential for tettigarctids to hide themselves. For Sanmai, most of the highly contrasting speckles and longitudinal stripes extend to the edge of the forewings. This disruptive coloration can effectively break up the body outline as well as surface (Cuthill et al. 2005; Schaefer and Stobbe 2006), and enabled Sanmai to hide in faint light under the cover of thick forest (see SOM: fig. S2). Disruptive camouflage is also frequent in Palaeontinidae, another common large-sized arboreal cicadomorph group of Daohugou Biota. However, unlike Sanmai, the disruptive wing coloration of palaeontinids is transverse alternating dark and light stripes or light spots on dark membrane (Shcherbakov and Popov 2002; Wang 2009).

To date, eight species within four genera (Shuraboprosbole, Sunotettigarcta, Tianyuprosbole, and Sanmai gen. nov.) attributed to the family Tettigarctidae have been recorded from the Daohougou beds (Wang and Zhang 2009; Li et al. 2012; Chen et al. 2014; this study). The forewing membrane in Sunotettigarcta and Tianyuprosbole is mostly dark and lacks highly contrasting coloration patterns, and is probably crypsis via background-matching. The postnodal area of Tianyuprosbole, ornamented with some thick ruffles, might be leaf mimesis (Chen et al. 2014). These well-preserved Daohugou specimens indicate that early tettigarctids likely have evolved different mechanisms of camouflage to avoid detection by potential predators. Besides, the different styles of forewing coloration of Daohugou tettigarctids suggest that although all of them were likely arboreal, they might have occupied slightly different ecotopes to reduce inter-specific competition.

Conclusions

We report the new genus Sanmai Chen, Zhang, and B. Wang with three new species (Sanmai kongi Chen, Zhang, and B. Wang, S. mengi Chen, Zhang, and B. Wang, and S. xuni Chen, Zhang, and B. Wang) from the Middle—Upper Jurassic of northeastern China (Daohugou). The new genus might be a transitional form between Turutanoviini and specialized Architettigini which suggests that tettigarctids with vein M three-branched on forewing probably originated from some Shuraboprosbole-related tettigarctids in Laurasia in the Middle Jurassic, and then spread to Gondwana before significant extension of the Tethys Sea. The disruptive coloration of forewing, prominent on some specimens of Sanmai, represents a novel type to Tettigarctidae, which can effectively break up the body outline as well as surface, and so likely enabled Sanmai to reduce the detectability of potential predators. The high divergence of forewing coloration preserved in Daohugou tettigarctids suggests that although most of early tettigarctids were arboreal (Wang 2009), they might have occupied slightly different ecotopes.