Introduction

The Commentry Basin (Allier, France; Upper Carboniferous) is famous for the numerous and well-preserved fossil insects discovered in this locality. Because this collection is one of the longest studied, it includes numerous type taxa, some being type species of orders. This deposit is 20My younger than the oldest deposit providing winged insects (Brauckmann et al. 1994), which ensures a strong significance of the fossils in the phylogenetic analyses of high rank in Pterygota.

Here I redescribe the first taxon described from this locality, Protophasma dumasii Brongniart 1879, type species of the Protophasmatidae Brongniart 1885. Despite its historical significance, this taxon has escaped a correct interpretation since its discovery, mainly because of the poor state of preservation of the material. Thanks to the definition of the Archaeorthoptera Béthoux & Nel, 2002 (including Orthoptera and extinct taxa such as the orders Titanoptera and Caloneurodea), a new homologization of the venation pattern of the Protophasmatidae is proposed. This family turns out to be of unexpected importance in the resolution of the relationships of the Orthoptera with the ‘dictyopterid’ orders.

Systematic paleontology

I follow the nomenclature of wing venation of the ‘archaeorthopterid’ insects proposed by Béthoux & Nel (2001, 2002). These authors did not specify the nomenclature of the anal veins. Thus, I use the nomenclature of the anal area proposed by Haas & Kukalová-Peck (2001), emended in a conservative way, without further hypothesis on the nature of the veins of the areas of AA and AP, i.e., the composite nature of AA1+2 and AA3+4 stems, sensu Haas & Kukalová-Peck (2001) is considered as not clearly demonstrated.

Composition

Type genus.— Protophasma Brongniart, 1879.

Diagnosis.— Forewings: M simple, at least basal of its mid-length; CuA + CuPa with numerous branches covering large part of posterior wing margin, but not organized in clear pattern of ramification; occurrence of well-developed AP area densely covered by veins and cross-veins. Hindwings: some branches of RP reaching RA; M forked very distally; contact between CuA and CuP limited to a simple point; CuA slightly arched towards CuP at this point and distally diverging from CuP; CuP branched distal of its mid-length; area between CuP and first AA vein broad, with secondary convex veins between them; two convex veins in AA area; AP area densely covered by veins and cross-veins; posterior margin nearly straight at level of anal fan. Both wing pairs: cross-veins reticulated except in AP area.

Discussion

The Protophasmatidae can be attributed to the Archaeorthoptera after the occurrence in the forewings of a fusion of CuA (from M + CuA) with the anterior branch of CuP (CuPa) into a composite common stem. Nevertheless, most of the venational features of this family can be considered as plesiomorphies within Archaeorthoptera or cannot be reliably polarized (see below). A possible apomorphy of this family would be ‘in hindwings, CuP branched distal of its mid-length’.

Composition

Type species.— Protophasma dumasii Brongniart, 1879

Diagnosis.— That of the family.

Phylogenetic implications

The homologization of the forewing area labeled ‘AP’ must be discussed. The relief of the veins is altered and there is no clear evidence that veins in this area were concave, as expected from veins of the posterior sector. Nevertheless, their organization strongly differs from veins in the AA area, and is very similar to that occurring in the undisputable AP area of the hindwings. Thus, I interpret this area as that of AP. Kukalová-Peck (1991) considered that convex veins could be branches of AP in the ‘polyneopterous’ orders. This contradicts her ground-plan pattern of Pterygota. Furthermore, these veins do not show the typical organization shown by the AP veins in Protophasma. Thus, at least for ‘polyneopterous’ orders, I propose that the veins interpreted as AP by Kukalová-Peck are AA veins, after the organization of true AP veins in Protophasma.

The occurrence of a well-developed AP area in forewings is unknown in all other representatives of the Archaeorthoptera and, actually, in the whole of Neoptera. Polarized after the putative sister-groups of Neoptera (e.g., the ‘paleopterous’ orders), this character must be considered as a plesiomorphy within Pterygota.

Protophasma shows some other features unknown in Archaeorthoptera. Among them, the most prominent is the absence of fusion of CuA with (a branch of) CuP in the hindwings, also absent in Blattodea. Because this last order is a putative sister group of the Archaeorthoptera, this character must be considered as a plesiomorphy within Archaeorthoptera. It must be noted that both veins CuA and CuP are connected near the wing base, which could be considered as a ‘transitional’ state toward the more complete fusion occurring in other Archaeorthoptera.

In the hindwings, the M, long, simple and not clearly differentiated into MA and MP, is a character unknown in other Archaeorthoptera, in which it is branched near its origin. Nevertheless, this structure can be found in some Blattodea (Schneider 1984). Possibly, this structure could be a plesiomorphy within the Archaeorthoptera.

The character state in hindwings, ‘CuP branched distal of its mid-length’ is unknown in all the ‘polyneopterous’ orders. In Blattodea, CuP is simple (in both fore and hindwings), but branched near the wing base in Archaeorthoptera (in both fore and hindwings). Thus, the state ‘CuP branched distal of its mid-length’ cannot be polarized. Its apomorphic status in Protophasmatidae is debatable: it may be a state ‘transitional’ between Blattodea and Archaeorthoptera.

Furthermore, the hindwing ‘anojugal lobe’ of the Protophasmatidae is similar to that defined for Blattodea and Isoptera by Haas & Kukalová-Peck (2001), viz. 2 simple convex veins occurring in the AA area, separated from the AP area by the ‘anal fold’, along which the hindwing is folded, as in the holotype specimen of Protophasma dumasii (see Figs 2.2, 3). Thus, the AA area is very reduced: it is the ‘partial anojugal lobe’ sensu Haas & Kukalová-Peck (2001). In ‘Orthoneoptera’ sensu Haas & Kukalová-Peck (2001) (a group that comprises at least all the representatives of the Archaeorthoptera), the AA area is well developed, contributing to the ‘full anojugal lobe’. Following Haas & Kukalová-Peck (2001) but emendating them, I consider that the ‘partial anojugal lobe’ is present in ‘Blattoneoptera’, ‘Hemineoptera’, ‘Endoneoptera’ but also in basal ‘Orthoneoptera’ (i.e., Protophasmatidae) and must then be considered as plesiomorphic within Neoptera.

In conclusion, the hindwings of the Protophasmatidae retain many plesiomorphic features linkable with the hindwing venation pattern of Blattodea, despite the fact that their forewings already show the typical and complex medio-cubital pattern of the Archaeorthoptera.

Conclusion

Following the present work, the Protophasmatidae are assigned to the Archaeorthoptera but must be considered as belonging to the stem group of this supraordinal clade, and potential sister-group of the whole clade. It will be necessary to include Protophasma in future attempts at phylogenetic analyses of the Archaeorthoptera (in prep.). The Protophasmatidae could be considered as a model of a ‘link’ between the Archaeorthoptera and the ‘dictyopterid’ orders.

The Protophasmatidae were contemporaneous with much more derived Archaeorthoptera (e.g., Orthoptera and Caloneurodea — an extinct Paleozoic order). Thus, by its numerous putative plesiomorphies, this family seems to have been a ‘relict’ taxa during the Upper Carboniferous. Because the record of Lower Carboniferous insects is very sparse, such remains are of great help in phylogenetic attempts.