A new Cretaceous psocodean family from the Charente-Maritime amber (France) (Insecta, Psocodea, Psocomorpha)

ABSTRACT Arcantipsocus courvillei n. gen., n. sp. is described from the Cretaceous amber of Archingeay (France). It is placed within the suborder Psocomorpha, and in the Mesozoic extinct family Arcantipsocidae n. fam. characterized by 14-segmented antenna; legs with tarsi 3-segmented; forewing setose with evanescent veins; pterostigma dark, thickened and setose; M 2-branched; areola postica free; nodulus present; hind wing with M bifurcated, without basi-radial cell; claws with a preapical tooth. A cladistic phylogeny for Psocomorpha is given including the new fossil taxon. The discovery of this new taxon demonstrates the necessity of a deep phylogenetic redefinition of the currently admitted major subdivisions of this suborder.


INTRODUCTION
Recent cladistic analyses reveal the paraphyletic nature of several orders of insects. Th e most signifi cant being the lice (order Phthiraptera), which is now included within the Psocoptera to form the order Psocodea (Yoshizawa & Johnson 2003a, b, 2006Johnson et al. 2004;Grimaldi & Engel 2005, 2006a.
We describe herein Arcantipsocus courvillei n. gen., n. sp., from the Archingeay Cretaceous amber of France. It is placed into the suborder Psocomorpha, and in the Cretaceous extinct family Arcantipsocidae n. fam.

SYSTEMATIC PALAEONTOLOGY
We follow in part the catalogue of Lienhard & Smithers (2002), and the works of Smithers (1972Smithers ( , 1990 and Mockford (1993) as essential tools for the systematic of the order. We follow the nomenclature of wing venation and body structures of Smithers (1972), andLienhard (1998). Th e fossil was carefully prepared in Canada balsam medium, following the method described by Azar et al. (2003), in order to observe as many characters as possible. Th us the "absences" of structures are accurate, which is diff erent of structures that are "not visible" but may be present. ETYMOLOGY. -After "Arcanti" from "Arcantiatum" old name of Archingeay and "psocus"; gender masculine. , with two shoulders made of two smooth teeth each, the fi rst being in the inner middle and the second situated slightly before the tip, the apex formed of two smooth teeth, one of them being very small. Th orax 0.58 mm wide; mesothorax nearly triangular.
Forewing patterned and setose, 1.97 mm long and 0.57 mm wide (Fig. 3). Marginal setae crossing. Two rows of setae on veins. Apex slightly acuminate. Most veins evanescent except in their terminal parts. Pterostigma dark, thickened and setose, convex and not connected to Rs by a cross-vein. Sc diff use and evanescent. R1 simple reaching costal margin at 1.55 mm from wing base. Rs evanescent and hardly visible; fork of R2 + 3 and R4 + 5 1.41 mm distal of wing base; R2 + 3 and R4 + 5 strongly curved; reaching wing margin respectively at 1.7 and 1.88 mm from wing base. Hind wing hyaline, smaller than forewing, with setose margin, 1.55 mm long and 0.5 mm wide (Fig. 8). Sc not visible. R fused basally with M and Cu. R1 0.23 mm long, not reaching anterior wing margin. No basi-radial cell. Bifurcation of Rs into R2 + 3 and R4 + 5 1.1 mm from wing base. M bifurcated. Bifurcation of M into M1 and M2 1 mm from wing base; M1 0.45 mm long; M2 0.32 mm long. Cu1 reaching posterior wing margin at 1.16 mm. Remaining veins hidden.

DISCUSSION
Th e new fossil possesses characters shared by both suborders Psocomorpha and Troctomorpha, viz. adult with tarsi 3-segmented, hind wing with M vein 2-branched (probably a plesiomorphy as it is present in the Archipsyllidae and the Hemipsocidae Pearman, 1936, psocomorphan families having an inclusive position in Yoshizawa's phylogeny), and forewing with nodulus. According to the keys proposed by Mockford (1993) and Lienhard (1998), Arcantipsocus courvillei n. gen., n. sp. falls in the suborder Psocomorpha because of the characters "forewing with thickened and sclerotized pterostigma", and "no scales on wings and body". Th e character "thickened and sclerotized pterostigma" is considered as being apomorphic of Psoco morpha by Yoshizawa (2002) and Mockford (1967). But with antennae with 12 fl agellomeres, the new fossil has also features of the Troctomorpha. Th us its position is problematic and we have to discuss more precisely these characters.
Recent sometimes fewer (Mockford 1993). Th e presence of only 11 fl agellomeres is likely plesiomorphic, as the Archipsyllidae have 11 fl agellomeres and are in a very inclusive position in the Psocodea (Huang et al. 2008). Th us the number of fl agellomeres is not suffi cient to assign this fossil taxon to one of the two suborders since this character seems to be homoplastic.
If we neglect the character "thickened and sclerotized pterostigma" that is diagnostic of the Psocomorpha in the key to recent families of Smithers (1990), Arcantipsocus n. gen. falls in the troctomorphan Amphientometae family Compsocidae Mockford, 1967 for its hind wing vein M forked and the forewing with a nodulus. But all Compsocidae have hyaline and unsclerotized pterostigma, unlike Arcantipsocus n. gen.
If we consider this character, Arcantipsocus n. gen. falls near the psocomorphan family Bryopsocidae Mockford, 1984 after the following combination of characters: macropterous insect; legs with trimerous tarsi; absence of scales; complex wing venation, in contrast to some psocids with venation reduced to some parallel veins; sclerotized pterostigma; head not elongate; free areola postica with Cu1a and Cu1b separating near posterior margin; forewing margin and membrane setose; hind wing with some setae on margin in addition to setae between arms of radial fork. However, Arcantipsocus n. gen. diff ers from all Bryopsocidae by the following characters: antennae with 12 fl agellomeres instead of 11; hind wing without basi-radial cell; bifurcation of M in hind wing into M1 and M2 (this last character is shared by the majority of psocids belonging to the suborders Troctomorpha and Trogiomorpha).
Th erefore, the new fossil cannot be assigned to any of the known families. Because of the unique combination of characters mentioned above and the thickened and sclerotized pterostigma that is apomorphic for the Psocomorpha, we attribute Arcantipsocus n. gen. to a new extinct family within this suborder.

PHYLOGENY
Th ere are very few attempts of phylogenetic analyses of the Psocoptera. Smithers (1972) dedicated a large part of his work Th e Classifi cation and Phylogeny of Psocoptera to the study of the phylogenetic relationships within Psocoptera, but his work is typological. Smithers (1972) proposed dendrograms with several lineages based on homoplastic or plesiomorphic characters. Several years later he admitted that his phylogeny needed revision (Smithers 1991). Perrichot et al. (2003) presented a tentative of cladistic phylogeny for Trogio morpha, nevertheless this later phylogeny is incomplete because based on few characters.  proposed another phylo geny of the same group, based on molecular data. Yoshizawa (2002), Johnson et al. (2004 and Grimaldi & Engel (2006a), proposed molecular and morphological phylogenetic analyses; the last one concerned the Psocomorpha only.
Our new family Arcantipsocidae n. fam. should fall within the Psocomorpha Homilopsocidea (sensu Yoshizawa 2002), if we admit that the character "pterostigma thickened and sclerotized" is not homoplastic, and is a real apomorphy of the Psocomorpha, as proposed by Yoshizawa (2002). Th e status of this character is debatable as Yoshizawa (2002) noted that "although a thickened pterostigma is observed in Archipsocidae, it appears to be much thinner than in other families of Psocomorpha (...) diff erent degrees of thickness of the pterostigma may provide further evidence for the phylogenetic placement of Archipsocidae as the basalmost clade of Psocomorpha". Also it is shared by the Mesozoic family Archipsyllidae that is supposed to be a more inclusive group than all recent Psocodea (Huang et al. 2008). In order to verify the position of the Arcantipsocidae n. fam., we made an attempt of cladistic analysis based on all the 68 characters and 50 taxa used by Yoshizawa (2002)  together with a group of four families Philotarsidae, Pseudocaeciliidae, Calopsocidae, and Trichopsocidae but this clade of fi ve families is not supported by any clear synapomorphy. Th is later was inserted in the cladogram without however aff ecting the topology of the strict consensus tree obtained by Yoshizawa GEODIVERSITAS • 2009 • 31 (1) (2002). Traditionally Psocomorpha were regarded as including four infraorders (Psocetae, Homilopsocidea, Epipsocetae, and Caeciliusetae) but Yoshizawa (2002), based on a cladistic phylogeny added two more, Archipsocetae comprising Archipsocidae, and Hemipsocetae including Hemipsocidae, these two families were previously assigned to Homilopsocidea and Psocetae respectively. More morphological or/ and molecular characters are needed for future studies of phylogeny of Psocodea including the fossil taxa in order to precise the history and scenarios of evolution of this group.