Pronemouridae fam. nov. (Insecta: Plecoptera), the stem group of nemouridae and notonemouridae, from the Middle Jurassic of Inner Mongolia, China

Authors


Abstract

Abstract:  A new family, Pronemouridae fam. nov., with a new genus, Pronemoura gen. nov., and five new species of fossil stoneflies are described from Daohugou village (Middle Jurassic), Inner Mongolia, China: Pshii sp. nov., P. angustithorax sp. nov., P. longialata sp. nov., Pminuta sp. nov. and P. peculiaris sp. nov. Wing venation of pronemourinids presents some or considerable variety in the fore and hind wings of the same or different individuals. Pronemourinidae, retaining plesiomorphic characters (short, multisegmented cerci and CuA with forks), should be treated as the stem group of Nemouridae and Notonemouridae. Based on fossil data, we propose a model in which the extant plecopterid distribution began in the Early Cretaceous. Nemouridae and Notonemouridae must have occurred in Asia and begun to migrate from here by at least the Early Cretaceous.

N emouridae is one of the largest families of Plecoptera with approximately 450 living species, containing two subfamilies and eighteen genera distributed across the Nearctic, Palearctic and Oriental regions (Baumann 1975). Fossil stoneflies of the family Nemouridae are rare, because they prefer running water, an environment unfavourable for burial and fossilization. Hitherto, only seven species (including one indeterminable species) have been described, representing three genera. Five species assigned to Nemoura Latreille were described from Baltic amber (Eocene). The monotypic Nemourisca was erected based on a single nymph, Ndiligens (Sinitshenkova 1987); Dimoula was described based on an imago, Ddimi (Sinitshenkova 2005). They were both found in the eastern Transbaikalia (Lower Cretaceous). Zwick (2000) pointed out that Jurassic species of ‘Nemouridae’ should be treated as stem-group representatives of Nemouridae + Notonemouridae, at least based on the described genus Nemourisca.

Recently, numerous Plecoptera fossils (numerous imagoes and a few nymphs) have been found in Daohugou village, Ningcheng County, Inner Mongolia, China. Many well-preserved stonefly imagoes show typical characteristics of Nemouridae. Here, we carefully compare these new fossil materials with extant nemourids using characters of wing venation, body structure, cerci and legs and find that they also reveal some plesiomorphic characters that differ from nemourids, such as short multisegmented cerci and a CuA vein with forks. Thus, we establish a new family Pronemouridae with a new genus Pronemoura. At the same time, we consider that Dimoula should be referred to this new family based on its preserved characters. Being a nymph, Nemourisca shows few characters differing from nemourids but shares the same locality and geological age as Dimoula. Thus, we also assign Nemourisca provisionally to this new family. In addition, Pronemouridae is significant for understanding the origin and dispersal of Nemouridae and Notonemouridae and for phylogenetic relationships of the superfamily Nemouroidea.

Material and methods

All the fossils described herein were collected from Daohugou village, examined with a LEICA MZ12.5 stereo microscope and illustrated with the aid of a drawing tube. Line drawings were processed using CorelDraw 12. Photographs were taken using an EPSON PERFECTION 1650 flatbed scanner. We follow the systematics of Sinitshenkova (1987, 2002). All type materials are housed at the Key Lab of Insect Evolutionary and Environment Change, College of Life Sciences, Capital Normal University, Beijing, China.

Background to the Daohugou beds

Daohugou village is situated in southern Inner Mongolia, which is close to Lingyuan in the western Liaoning Province, north-eastern China. The lacustrine deposits at Daohugou village are composed of greyish or grey tuffaceous shales and mudstones with tuffaceous breccia at the bottom, and various tuff layers were interbedded in the shales and mudstones (He et al. 2004). It contains a diverse insect fauna composed of at least of fourteen orders and more than 50 genera and 200 species (Zhang 2002). Numerous and diverse fossils from this ecosystem consist of plants dominated by Ginkgopsida, Coniferopsida, Lycoposida, Filicopsida and Cycadopsida (Mi et al. 1996), a few conchostracans (Shen et al. 2003), salamanders (Wang 2004), pterosaurs (Czerkas and Ji 2002), haired Maniraptora (Zhang et al. 2002) and a mammal (Ji et al. 2006).

The age of the Daohugou beds is still under debate, ranging from the Early Middle Jurassic to Early Cretaceous. Even the same authors, as their scientific research progressed, provided different opinions at different times (Zhang 2005). Wang et al. (2005) always considered the fossil beds in the Daohugou area as the most basal deposit of the Early Cretaceous Jehol biota, representing the earliest fossil assemblages of this biota. They alleged that the stratigraphic sequence at the Daohugou fossil sites was overturned (Wang et al. 2005). However, many researchers who have investigated the Daohugou beds criticized this opinion. They did not find the overturned stratigraphic sequence of the Jehol biota near Daohugou, therefore, they did not agree with the veracity of the section and fossil checklist proposed by Wang et al. (2000) and He et al.’s (2004) conclusions about the post-Jurassic age of the Daohugou beds based on an unfounded assumption of an overturned stratigraphic sequence (Liu et al. 2004; Ji et al. 2005). Recently, the accurate Ar–Ar and SHRIMP U-Pb dating methods show that the age of intermediate-acid volcanic rocks overlying the Daohugou fossil-bearing beds (N41°18.979′, E119°14.318′) is about 164–165 Ma, and thus, the age of the Daohugou beds is at least 165 Ma (Chen et al. 2004). This conclusion, supported by palaeontological evidence from conchostracans and insects, provides the most conclusive evidence for a Middle Jurassic (Bathonian) age of the Daohugou beds (Gao and Ren 2006).

Systematic palaeontology

Order PLECOPTERA Latreille, 1802
Suborder NEMOURINA Newman, 1853
Superfamily NEMOUROIDEA, Newman, 1853
Family PRONEMOURIDAE Liu, Sinitshenkova and Ren, fam. nov. 

Type genus. Pronemoura Liu, Sinitshenkova and Ren, gen. nov.

Diagnosis.  Small stoneflies. Head large, its width slightly narrower or larger than prothorax. Prothorax transverse, almost rectangular, its width obviously narrower than mesothorax. Mesothorax and metathorax developed and similar in shape. In the wings, ‘X’ vein is present and formed primarily by the terminal portion of Sc, c-r, r-rs, rs-m and the base of the posterior Rs branch. Costal crossveins absent or two at most. M bifurcates basal to the Rs fork, CuA forks distally, CuP simple, short. Crossveins between M and CuA are almost as long as crossveins between CuA and CuP. Two curved anal veins present. Hind wings with or without broadened anal area. When folded, wings obviously surpass tip of abdomen. Legs long and slender, first and the third segments of tarsus long, second distinctly short. Cerci short, multi-segmented.

Included genera. Pronemoura gen. nov., Dimoula and Nemourisca in the Glushkovo Formation (Lower Cretaceous) of Eastern Transbaikalia.

Remarks.  Pronemouridae can be related to the five extant families in the superfamily Nemouridea based on its well-preserved morphological characters. The new family can be distinguishable from Taeniopterygidae by the short second tarsal segment and from the other four families by CuA with two branches.

In this new family, it is interesting to note that – judging from the wing venation – the ancient Jurassic Pronemoura that lacks the multiple crossveins is simpler than the Cretaceous Dimoura. We think these multiple costal crossveins should be considered a derived character, the same as the costal crossveins that occur in the family Taeniopterygidae (Ricker and Ross 1975).

Identification key to genera and species of imagoes belonging to the family Pronemouridae

  • 1 Multiple costal crossveins present on all wings and anal region of hind wing not broadened… (Dimoula dimi Sinitsh. 2005).
    •  Multiple costal crossveins absent on all wings and anal region of hind wing broadened… 2 (Pronemoura gen. nov.).
  • 2 Anterior branch of CuA shorter than CuA stem length and prothorax obviously wider than the half of the mesothorax… 3.
    •  Anterior branch of CuA longer than CuA stem length, prothorax almost equal to the half of the mesothorax width, m-cu short, straight, enters M at the level of its fork and reaches CuA proximad of the anterior branch of CuA… (P. angustithorax sp. nov.).
  • 3 Posterior branch of Rs strongly curved… 4.
    •  Posterior branch of Rs gently curved, RS forks distinctly after r-rs, CuP very short, reaches the posterior wing margin at the fourth of wing length… (P. peculiaris sp. nov.).
  • 4 Forewings notably longer than 10 mm, CuA forks are appreciably long, the posterior branch of Rs S-shaped… (P. longialata sp. nov.).
    •  Forewings shorter than 10 mm, CuA forks short, the posterior branch of Rs deeply curved, not S-shaped… 5.
  • 5 Antenna long, nearly as long as body, CuA forks short, anal region of hind wing not very broad… (P. minuta sp. nov.).
    •  Antenna short, shorter than combined length of the head and thorax, CuA forks are very short, anal region of hind wing very broad… (P. shii sp. nov.).

Genus PRONEMOURA Liu, Sinitshenkova and Ren, gen. nov.

Type species. Pronemoura shii Liu, Sinitshenkova and Ren, sp. nov.

Included species. P. shii sp. nov., P. longialata sp. nov., P. angustithorax sp. nov., P. peculiaris sp. nov. and P. minuta sp. nov.

Derivation of name.  From the prefix ‘pro-’ and the recent genus Nemoura.

Diagnosis.  Imago. Forewings slightly longer than hind wings, and when folded, wings exceed the body by almost half. Venation of left and right wings similar, anal region of hind wing broad with five longitudinal veins. Multiple crossveins are absent in costal area. c-r long or short, well developed. Rs forks distal to r-rs, and its posterior branch strongly arched at its base, the so-called X is well expressed. CuA forks close to or before the level of M fork. Legs long and slender, first and the third segments of tarsus long, second distinctly short. Cerci short with several segments.

Remarks.  The new genus differs from Dimoula by the absence of additional costal crossveins, the broadened anal area of hind wings, and a better expressed ‘X’ on the forewings.

Pronemoura shii Liu, Sinitshenkova and Ren, sp. nov.
Text-figures 1A, 2

Figure TEXT‐FIG. 1..

 A, Pronemoura shii sp. nov. Holotype, no. CNU, NMDHG3; B, Pangustithorax sp. nov. Holotype, no. CNU, NMDHG13; C, P. longialata sp. nov. Holotype, no. CNU, NMDHG8; D, Plongialata sp. nov. Paratype, no. CNU, NMDHG9. Scale bar represents 5 mm.

Figure TEXT‐FIG. 2..

Pronemoura shii sp. nov. Holotype, no. CNU, NMDHG3. Scale bar represents 2 mm.

Derivation of name.  The species name is dedicated to the insect taxonomist Shi Fuming, tutor of Mrs Liu, for his affection, care and guidance.

Holotype.  CNU, NMDHG3 (Text-fig. 1A), a well-preserved body, with fore and hind wings, fragments of legs. Four paratypes (imagoes) from the same locality: CNU, NMDHG4–7.

Diagnosis.  Head large, antenna short, shorter than combined length of head and thorax, wings large, ‘X’ well expressed, posterior branch of Rs deeply curved, not S-shaped, and the broadened anal area of hind wing with at least five longitudinal veins.

Description.  Male (Text-fig. 2). Body length 8.6 mm. Head large, its width slightly narrower than prothorax, prothorax about three-fourths width of mesothorax. Mesothorax as wide as metathorax.

Forewing. 9.5 mm long, Sc about two-thirds of the total wing length, enters R almost at level of r-rs, crossvein c-r very long, almost straight, enters C distal to the midlength of Rs branches. Base of Rs deeply curved, Rs forks distal to r-rs, which is oblique and short. Rs branches obviously curved. M forks almost at midlength of forewing, its branches slightly curved, rs-m long, enters Rs basal to r-rs and touches M basal to its fork. CuA with long branches, which are slightly shorter than CuA stem length, CuA forks obviously in the basal half of wing, and before M forks. CuP short, reaching the posterior wing margin basal to CuA forks, m-cu long, slightly S-shaped, enters M distal to rs-m and reaches CuA considerably basal to CuA fork. Not less than four crossveins between M and CuA, and not less than five between CuA and CuP. A2 with two branches, the anterior one strongly arched, the posterior one short and straight.

Hind wing. ‘X’ well expressed. Main venation is similar to forewing. CuA and CuP simple, long and almost straight. Anal area widened, with at least five longitudinal veins.

Abdomen nearly 1.5 times as long as thorax, with nearly equal width for all segments except the tenth segment, which is sclerotized. Segments 3–9 wide and short, segment 10 about two-thirds of anterior segment width and strongly sclerotized, forming a large, flat area before base of hypoproct, which is also strongly sclerotized. Paraprocts sclerotized, inflexed distally, their lobes inconspicuous. Cerci apparently short, not visible in segments.

Pronemoura angustithorax Liu, Ren and Shih, sp. nov.
Text-figures 1B, 3A–D

Figure TEXT‐FIG. 3..

Pronemoura angustithorax sp. nov. Holotype, no. CNU, NMDHG13. A, Fore tarsus. B, Middle tarsus. C, Hind tarsus. D, Body, ventral aspect. Scale bar represents 1 mm in A–C, 2 mm in D.

Derivation of name.  A combination of the Latin prefix angusti- (meaning ‘narrow’) and thorax, implying the narrow prothorax of this species.

Holotype.  CNU, NMDHG13 (Text-fig. 1B), well-preserved part and counterpart of an adult, with complete forewings, fragments of legs; abdomen and cerci not preserved.

Diagnosis.  Body large, antenna short, prothorax narrow, half width of mesothorax, posterior branch of Rs S-shaped, branches of CuA long, m-cu straight, shorter than rs-m, enters M at the level of its fork and connects obviously the anterior branch of CuA.

Description.  Head large, width slightly larger than prothorax. Prothorax transverse, almost rectangular, width about half of mesothorax. Mesothorax and metathorax developed and equally wide.

Forewing (Text-fig. 3D). 15.2 mm long, Sc nearly two-thirds of the total wing length, enters R at the level of r-rs. c-sc very short and enters C considerably far from Sc tip. The crossvein c-r curved, begins near the Sc tip and enters C clearly basal to the midlength of Rs branches. Base of Rs deeply curved, Rs forks clearly distal to r-rs, which is short, almost perpendicular to Rs. Rs branches obviously curved, posterior branch almost S-shaped. M forks distal to midlength of wing, its branches almost straight, rs-m long, enters Rs basal to r-rs and touches M slightly basal to its fork. CuA with very long forks, forks distinctly basal to M fork, and longer than CuA stem length, CuP long, reaches posterior wing margin obviously distal to CuA fork, m-cu short, straight, enters M at the level of its fork and connects obviously the anterior branch of CuA. Six crossveins between M and CuA, and eight crossveins between CuA and CuP. Anal veins not clearly preserved. Hind wing not preserved.

Legs (Text-fig. 3A–C). Femora sturdy, tibiae gracile, tarsus of front and middle legs well-preserved, the first and third segment are almost equal in length, second about half of the first one.

Pronemoura longialata Liu, Sinitshenkova and Ren, sp. nov.
Text-figures 1C–D, 4A–D, 5

Figure TEXT‐FIG. 4..

Pronemoura longialata sp. nov. Holotype, no. CNU, NMDHG8. A, Fore tarsus. B, Hind tarsus. C, Body, ventral aspect. D, genitalia and cerci. Scale bar represents 1 mm in A–B, D; 2 mm in C.

Figure TEXT‐FIG. 5..

Pronemoura longialata sp. nov. Paratype, no. CNU, NMDHG9. Scale bar represents 2 mm.

Derivation of name.  Named after the Latin ‘longi’ and ‘alatus’, referring to the long wings.

Type material.  Holotype (Text-fig. 1C). CNU, NMDHG 8, a well-preserved adult male, with forewings and part of hind wings, legs and cerci. And four paratypes (imagoes) from the same locality: CNU, NMDHG9-12, among which CNU, NMDHG9 (Text-fig. 1D) with well-preserved genitalia of an adult female.

Diagnosis.  Body large, antennae long, posterior branch of Rs slightly S-shaped, CuA with two short branches, clearly shorter than CuA stem, cerci short, about six segments; legs slender.

Description.  Male (Text-fig. 4A–D). Body length 10.9 mm, head long, ovoid because of its taphonomy, its width narrower than prothorax, prothorax slightly narrower than mesothorax. Mesothorax and metathorax well developed and equally wide.

Forewing (Text-fig. 4C). 12.5 mm long. Sc two-thirds of total wing length, Sc enters R at the level of r-rs. The crossvein c-r long, almost straight, enters C basal to midlength of Rs branches. r-rs oblique, enters Rs clearly before the Rs forks. Base of Rs not preserved. Anterior Rs branch slightly curved and posterior branch obviously curved, S-shaped. M forks distal to midlength of wing, its branches arched forward near the wing margin or almost straight, rs-m long, enters Rs basal to r-rs and touches M basal to its fork. CuA with short fork, clearly shorter than CuA stem, CuA branches basal to M forks and slightly basal to midlength of wing. CuP short, reaches posterior wing margin before CuA forks, m-cu long, slightly S-shaped, enters M basal to, or at the level of rs-m, reaches CuA considerably basal to CuA fork. Five crossveins between M and CuA and four crossveins between CuA and CuP. A2 with two branches, the anterior one strongly arched, the posterior short and straight.

Hind wing (Text-fig. 4C). ‘X’ well expressed, main venation similar to forewing. Most of hind wing not preserved.

Abdomen not well preserved, the last few segments are twisted, with almost all preserved from the lateral side. Cerci short, with six segments (Text-fig. 4D).

Legs (Text-fig. 4A–B). Femora not particularly sturdy, tibiae very slender, first and third segment of tarsus almost same length, second slightly shorter than half of first in fore legs, and about one-fourth of first in hind legs.

Female (Text-fig. 5). CNU, NMDHG9. Body 12.2 mm length, forewings 14.5 mm long. Venation of forewings almost the same as holotype, except r-rs. r-rs enters R distal to Sc tip, and connects Rs at the level of Rs forks.

Remarks.  We consider the paratype, CNU, NMDHG9 to be a female of this new species for the following reasons. The fossils show, respectively, male and female individuals, come from the same locality, and the forewings have very similar venation, and the body size is an appropriate match (the female is slightly bigger than the male).

Pronemoura minuta Liu, Sinitshenkova and Shih, sp. nov.
Text-figures 6A, C, 7A–B

Figure TEXT‐FIG. 6..

 A, Pronemoura minuta sp. nov. Holotype, no. CNU, NMDHG17. B, P. peculiaris sp. nov. a. Holotype, no. CNU, NMDHG14. C, P. minuta sp. nov. Paratype, no. CNU, NMDHG18. D, P. peculiaris sp. nov. Paratype, no. CNU, NMDHG15. Scale bars represent 5 mm.

Figure TEXT‐FIG. 7..

 A, Pronemoura minuta sp. nov. A. Holotype. no. CNU, NMDHG17. B, Paratype, no. CNU, NMDHG18. Scale bar represents 2 mm.

Derivation of name.  Named for its small size.

Type material.  Holotype (Text-fig. 6A). CNU, NMDHG17. a well-preserved adult body, with complete forewings, fragments of hind wings and legs; head not preserved. The paratype from the same locality: CNU, NMDHG18 (Text-fig. 6C), which is a part and counterpart of body, has head, thorax and forewing all well preserved.

Diagnosis.  Body small, antennae very long, the posterior branch of Rs deeply curved, not S-shaped, very short branches of CuA, both medial crossveins and mediocubital crossveins fewer than five, and narrow anal area of hind wing.

Description.  Body length about 7–8 mm. Head narrower than prothorax, antennae long, prothorax slightly narrower than mesothorax (Text-fig. 7B).

Forewing (Text-fig. 7A). 8–10 mm long, Sc two-thirds of the total wing length, enters R slightly basal to r-rs (or slightly distal to r-rs on the left wing of holotype), c-sc not short and enters C obviously basal to Sc tip. The crossvein c-r almost straight, quite long, enters C basal to midlength of Rs branches. Base of Rs not preserved, Rs forks slightly distal to r-rs, which is oblique and short. Anterior Rs branch straight with posterior branch obviously curved. M forks at midlength of wing, its branches almost straight or slightly curved, rs-m long, clearly curved, enters Rs distinctly basal to r-rs and touches M obviously basal to its fork. CuA with short fork, branching slightly basal to M fork, CuA fork half as long as CuA stem, CuP short, reaches the posterior wing margin distinctly basal to CuA fork, m-cu long, slightly S-shaped, enters M distal to or at the level of rs-m and reaches CuA considerably basal to CuA fork. No fewer than four crossveins between M and CuA, and no fewer than five between CuA and CuP. Anal veins not well preserved. The left forewing almost identical to the right one, except that rs-m and m-cua join in right forewing, but not in left.

Hind wing (Text-fig. 7A). ‘X’ in wing well expressed, the main venation similar to forewing. CuA and CuP simple, long and almost straight. Anal area widened. Hind wing of holotype expresses some variation of venation: Sc enters R at the level of r-rs, Rs forks slightly basal to r-rs and M forks obviously distal to Rs branching.

Legs (Text-fig. 7A). Femora sturdy, their width less than 2 times width of tibiae, tibiae gracile, tarsus well preserved, the first segment long, second shorter than half length of the first, third slightly shorter than first.

Pronemoura peculiaris Liu, Sinitshenkova and Ren, sp. nov.
Text-figures 6B, D, 8A–B

Figure TEXT‐FIG. 8..

 A, Pronemoura peculiaris sp. nov. A, Holotype, no. CNU, NMDHG14. B, Paratype, no. CNU, NMDHG15. Scale bar represents 2 mm.

Derivation of name.  Named for the posterior Rs branch of this species, which is different from other species in this new genus.

Type material.  Holotype (Text-fig. 6B). CNU, NMDHG14. well-preserved part and counterpart of an adult male, with complete forewings, fragments of hind wings and legs; head and prothorax are not preserved. Besides the holotype, two paratypes (imagoes) from the same locality: CNU, NMDHG15 (Text-fig. 6D) and 16, whose heads and thorax are well preserved.

Diagnosis.  Body small, head large, antennae long, forewings long, the posterior branch of Rs gently curved and two branches of CuA short.

Description.  Body length about 6.4 mm (Text-fig. 8B). Head slightly wider than prothorax. Antennae long, prothorax transverse, about three-fourths width of mesothorax.

Forewing (Text-fig. 8A). 10.2 mm long, Sc about two-thirds total wing length, enters R basal to r-rs, c-sc short and enters C obviously far from the Sc tip. Crossvein c-r short, enters C distinctly basal to midlength of Rs branches. Base of Rs deeply curved, Rs forks clearly distally of r-rs, which is oblique and short. The posterior Rs branches are gently curved. M forks distal to midlength of wing, its branches almost straight, rs-m long, enters Rs basal to r-rs and touches M basal to its fork. CuA with short forks, which are clearly shorter than CuA stem, CuA forks slightly basal to midlength of wing and before the M forks, CuP short, reaches posterior wing margin basal to CuA forks, m-cu long, curved, enters M at level of rs-m and reaches CuA considerably basal to CuA forks. Five or six crossveins between M and CuA, four or six crossveins between CuA and CuP. Anterior branch of A2 strongly arched, posterior branch not preserved.

Hind wing (Text-fig. 8A). ‘X’ in wings well expressed, main venation similar to forewing. Anal part of wing not preserved.

Legs (Text-fig. 8B). Femora sturdy, about two times width of tibiae, tibiae gracile, tarsus of hind legs well preserved, the first and third segments almost of same length, second about half of first.

Discussion

Variability of wing venation

Variability of wing venation is very common among stoneflies. Sinitshenkova (2005) recently described different venation of the fore and hind wings of the same individual in Dimoula. The same pattern is found in Pronemoura gen. nov., but this variability does not always occur in every species of the new genus. Some species are uniform in their left and right wings, for example, P. shii, P. angustithorax and P. peculiaris; other species are slightly different, such as P. longialata and P. minuta. On the other hand, there are also slight variations in different individuals (including male vs. female) or the same species. A detailed description of the variability of venation is very important in identifying fossil stoneflies. We must clearly elucidate which features of the venation are more important for identifying fossil specimens, which variations of the venation are only intraspecific characters within the same species and which are between species or genera.

Affinities of Pronemouridae and Nemouridea + Notonemouridae

When Baumann (1975) discussed the zoogeography of Nemouridae and Notonemouridae, he suggested that the prenemourids or stem nemourids were present while there was still a distributional connection between Laurasia and Gondwanaland. This view is corroborated by our discoveries of the new family Pronemouridae. Based on its well-preserved morphological characters, Pronemouridae should be treated as the stem group of Nemouridae + Notonemouridae. Of these two families, Pronemouridae is closer to the extant family Nemouridae based on the following characters: (1) The so-called nemourid X, the most distinctive character, is formed at the cord in the forewings; (2) the second tarsus segment is distinctly short, the first and the third ones long and equal; (3) Rs with two branches only; (4) many crossveins between M and CuA and between CuA and CuP; (5) small body, shorter than 15 mm. On the other hand, there is little to compare Pronemouridae explicitly with Notonemouridae based on the characters observable in fossils. Pronemouridae shows those characters of Notonemouridae, which generally occur in Nemouridae too. Considering character distribution in extant Nemouridae and Notonemouridae, we prefer Zwick’s (2000) opinion. He explicitly pointed out that Notonemouridae belongs to Nemouridae (s.l.) and that it appears to be heterogeneous; a paraphyletic assemblage of early nemourid lineages surviving on fragments of Gondwanaland where Nemouridae did not occur.

Palaeogeography

Extant Nemouridae are distributed throughout the Holarctic, and a few genera have invaded subtropical and tropical areas and reached the Orient. Notonemouridae is restricted to the Southern Hemisphere (Illies 1965). This new family Pronemouridae, as the probable stem group of Nemouridae and Notonemouridae, is found in the Eastern Transbaikalia (Upper Cretaceous, Sinitshenkova 2005) and Inner Mongolia (Middle Jurassic). Baumann (1975) assumed that Nemouridae radiated throughout the Northern Hemisphere, with evolutionary centres in south Central Asia and western North America. Based on Pronemouridae, we suggest Nemouridae might have evolved initially in Central Asia. The breakup of Pangea began in the Middle Jurassic (180 Ma), and North America and Eurasia were the one landmass until the end of the Mesozoic (70 Ma; Scotese 2004). In the Middle Jurassic, numerous pronemourids lived in Central Asia and might have migrated out from here. Nemouridae splits from Pronemouridae and formed two subfamilies (Amphinemurinae and Nemourinae) before the complete separation of Eurasia and North America, such that Amphinemura and Nemoura became widely distributed in Palearctic and Nearctic regions. After complete separation of Eurasia and North America, nemourids existed respectively in Eurasia and North America; later, as Baumann suggested, corresponding with the appearance of the Rocky and Himalayan Mountains (50 Ma), nemourids radiated from evolutionary centres in western North America and south Central Asia, respectively, leading to many endemic genera (Baumann 1975). The Bering Sea land bridge later connecting North America and north-east Asia allowed species exchange, such as in Podmosta weberi and Nemoura arctica.

Although Pronemouridae appears to represent the stem group of Nemouridae + Notonemouridae, it sheds little light on the problem of notonemourid origins and distribution. Based on their present distribution and palaeogeographic events, notonemourids should have immigrated to Gondwana from Eurasia beforehand. Because long-distance dispersal across oceans is near impossible for Plecoptera, notonemourids living in Gondwana must have been present before the breakup of this supercontinent. Afterwards, notonemourids survived in disjunct continents, and some groups became extinct, yielding present distributional patterns.

In addition, many modern groups (such as Leuctridae, Taeniopterygidae, Perlodidae and Chloroperlidae) began to occur in the Early Cretaceous. Stoneflies are unknown in the Late Cretaceous because of large-scale transgression. All fossil stoneflies found from the Palaeocene belong to living families (Sinitshenkova 2002). We also found Capniidae and Taeniopterygidae in Daohugou (Middle Jurassic) and found Perlidae, Taeniopterygidae and Neouridae in the Yixian Formation (Late Jurassic–Early Cretaceous), Liaoning, particularly Perlidae and Neouridae, which possess well-preserved genitalia similar to living species (Liu et al. 2008). We speculate that some extant families were present in the Jurassic, and the modern distribution of Plecoptera originates from the Early Cretaceous. Nemouridae and Notonemouridae were present by then and had begun to migrate by at least the Early Cretaceous.

Acknowledgements.  We are grateful to Dr Jason A. Dunlop and the anonymous referees for helpful reviews of the original manuscript. This research was supported by grants from the National Nature Science Foundation of China (nos 41002008, 40872022, 31071964), PhD Start-up Foundation of Shenyang Normal University (No. BS200804), Education Bureau of Liaoning Province (No. 2009A659), and the program of the Presidium of the Russian Academy of Sciences ‘The origin and evolution of the biosphere’.

Editor. Jason Dunlop

Ancillary