The Pyrenean species of Chelidura (Dermaptera, Forficulidae)

The Pyrenees are inhabited by scattered populations of earwigs of the genus Chelidura Latreille, 1825. There is some controversy about the specific assignment of these populations: while most authors assign them to C. pyrenaica (Gené, 1832), other consider that C. aptera (Mégerlé, 1825) is also present in the Pyrenees. The main objective of this work was to revise the identity and synonyms of Pyrenean Chelidura. Specimens from recent fieldwork and collections (MNCN-CSIC) were used for morphological and molecular studies (cytochrome oxidase 1). All Pyrenean specimens shared similar cox1 sequences, very divergent from those of Alpine C. aptera. As a consequence, the variability observed in male cerci morphology from the Pyrenees, ranging from long and slightly curved to short and very curved, corresponded to C. pyrenaica, and the presence of C. aptera in the Pyrenees can be rejected. As previously suggested by Maccagno (1933) and Fontana et al. (2021), the revision of the synonymic list uncovered the misplacement of the name F. simplex Germar, 1825 under the synonymy of C. aptera, while it rather represents a synonym of C. pyrenaica (syn. nov.). Forficula simplex has nomenclatural priority over C. pyrenaica, however both names meet the requirements of the article 23.9.1 of the International Code of Zoological Nomenclature to retain the prevailing usage of C. pyrenaica (nomen protectum) over F. simplex (nomen oblitum). Additionally, we discuss the taxonomic status of Chelidura arverna David & Van Herrewege, 1973 stat. nov. from the French Massif Central.


Introduction
One of the most characteristic genera of Dermaptera in the high elevations of the European Mountains is Chelidura Latreille, 1825, represented by robust large-sized species often found in the upper limit of the coniferous forests. After the recent revision by Kirstová et al. (2020) who reconsidered the status of Mesochelidura Verhoeff, 1902 andChelidurella Verhoeff, 1902, previously synonymized with Chelidura by Steinmann (1993), the genus Chelidura includes 13 species confined to mountains of the Palaearctic region (Kočárek 2004;Kirstová et al. 2020). However, the genus is still in need of a rigorous taxonomic revision to determine the status of the Asian species (Kirstová et al. 2020).
Chelidura pyrenaica and C. aptera are two European species with ecological similarities. Both species are found in mountains at relatively high elevations, between 1000 and 2500 m of altitude (Albouy and Caussanel 1990). Most reports and descriptions indicate that C. aptera and C. pyrenaica are easily differentiated morphologically by the shape of male cerci (see Finot 1890;Azam 1901;Chopard 1922;Albouy and Caussanel 1990). According to those authors, males of C. aptera have long, relatively thin and slightly curved cerci, while males of C. pyrenaica have short, broad and very curved cerci. In both species, cerci of females are short, thin and practically straight, with a slight curvature at the apex. However, Dohrn (1867), followed by Brunner von Wattenwyl (1882), Maccagno (1933) and Fontana et al (2021) considered that long, thin and slightly curved cerci together with short, broad and very curved cerci were part of the intraspecific variability of each taxon, and questioned the presence of C. aptera in the Pyrenees (or the presence of C. pyrenaica in the Alps). To complicate matters, as already noted by Maccagno (1933), the earliest descriptions of Pyrenean specimens of Chelidura as a differentiated taxon correspond to Forficula simplex Germar, 1825, which was described based on long cerci Pyrenean specimens (Germar 1825). Forficula simplex was subsequently included in the synonymy of C. aptera by Dohrn (1867), followed by Bolívar (1876), Brunner von Wattenwyll (1882), Finot (1890), Kirby (1904), , Sakai (1973), Harz and Kaltenbach (1976), and Herrera-Mesa (1999) among others, or treated as a variety of C. aptera (Dubrony 1878).
The known distribution range of Chelidura in the Pyrenees is quite limited, with very few records in Andorra and the Spanish (Lapeira and Pascual 1980) and French slopes (Albouy and Caussanel 1990) (see "Species accounts" section). During field surveys aimed to document the persistence of the species in some of the classical localities, we were surprised to find consistently specimens with long cerci (referred to as C. aptera in the literature; e.g. Lapeira and Pascual 1980;Albouy and Caussanel 1990), and short cerci (referred to as C. pyrenaica; op. cit.) coexisting at the same localities. These observations, together with the lack of consensus on the presence of C. aptera in the Pyrenees (see references above), prompt-ed us to carry out a study to determine the correct identification of long and short cerci specimens of Pyrenean Chelidura. For this purpose, we obtained cytochrome oxidase 1 (cox1) partial sequences of a few Pyrenean specimens, a representative of each cerci morphology, and we also raised under controlled conditions, a series of nymphs collected from the same clutch till they metamorphose. The results of these analyses revealed that both long and short cerci males corresponded to a single taxonomic entity.
With this main aim, the specific objectives of this work are: (i) to confirm the taxonomic identification of Chelidura specimens with long and short cerci present in the Pyrenees, (ii) discuss the taxonomic entity of the subspecies C. pyrenaica arverna from the French Massif Central, and (iii) provide a species account including all known localities and synonymies of Pyrenean Chelidura.

DNA extraction and amplification
Total DNA was obtained from six specimens (Table 1). DNA was extracted from one leg, using the DNeasy Blood and Tissue Isolation Kit (Qiagen, Hilden, Germany), following the manufacturer's instructions, and then stored at 4 °C until further processed. The polymerase chain reaction (PCR) consisted of, with occasional minor variations, 18.8 μL of distilled water, 2.5 μL of 10 × PCR buffer, 1 μL of dNTP mix (10 mM), 0.5 μL of MgCl2 (50 mM), 0.5 μL of each primer (10 μM), 0.2 μL of DNA polymerase (5u/μL) and 1 μL of DNA template, consisting of a final reaction volume of 25 μL. The universal pair of primers LCO1490 and HCO2198 (Folmer et al. 1994) were used to amplify a fragment of cox1, with the following PCR cycling profile: initial denaturation at 96 °C for 5 min, followed by 40 cycles at 94 °C for 30 s, 42 °C for 45 s and 72 °C for 1 min, and a final extension step at 72 °C for 5 min. After the amplification, 4 μL of the reaction was analyzed by electrophoresis on a 1% agarose gel. Samples with single bands were sent to the company Macrogen Inc. (Macrogen Europe, Madrid, Spain) for sequencing in both directions.

Phylogenetic analyses and species concept
The cox1 data set included four Pyrenean specimens (with diverse cerci morphology), two specimens from the Italian Alps, one specimen of C. p. arverna from Kirstová et al. (2020), and six additional specimens of C. aptera from Fontana et al. (2021) (Table 1). One additional specimen of Chelidurella vignai Galvagni, 1995 and another of Chelidurella thaleri Harz, 1980(from Kirstová et al. 2020; Table 1) were used as closely related outgroups. We also included one specimen of Mesochelidura occidentalis Fernandes, 1973 and another of Anechura bipunctata (Fabricius, 1781) (from Kirstová et al. 2020; Table 1) as distant outgroups to root the phylogenetic analyses. Gen Bank accession numbers for the newly sequenced specimens are provided in Table 1.
The obtained cox1 partial sequences were aligned with MAFFT v.7 (Katoh et al. 2019) using default parameters. Uncorrected (p) pairwise genetic distances were estimated using PAUP* v.4.0a (Swofford 2002). The best substitution model obtained using Partition-Finder2 (Lanfear et al. 2016)   Evolutionary (taxonomic) units within Chelidura were defined using the evolutionary species concept as discussed in Sánchez-Vialas et al. (2020). The evolutionary species concept considers species as "a single lineage of ancestral descendant populations of organisms that maintain its identity from other such lineages and which has its own evolutionary tendencies and historical fate" (Wiley 1978(Wiley , 1981Wiley and Mayden 2000).

Results
Based on the phylogenetic analyses, studied specimens of Chelidura compose three well-supported clades (posterior probabilities = 1) (Fig. 2). A clade includes the Pyrenean specimens (PP = 1), a second clade includes the Massif Central specimen, and the third clade includes the Alpine specimens (Valle d'Aosta -Biella -Sondrio -Valais: Grand Saint Bernard) (PP = 1). The Alpine samples are geographically structured in two main subclades, one including samples from Valle d'Aosta and Col du Grand Saint Bernard (PP = 0.99), the second from Biella and Sondrio (PP = 0.94). Uncorrected "p" distances between different groups based on cox1 partial sequences are summarized in Table 2.
Pyrenean specimens (Andorra and Girona) form a monophyletic group of poorly differentiated sequences (uncorrected p distance ranging from 0 to 0.03) ( Fig. 2; Table 2). These samples include male specimens from Tossa d'Alp (Girona) and La Rabassa (Andorra) with typical short cerci (MNCN_Ent 296014, 296016) and specimens from Tossa d'Alp (Girona) with very long cerci (MNCN_ Ent 296013). The sister taxon relationship of the Alpine clade with respect to the Pyrenean and Massif Central clades is poorly resolved (PP = 0.72) forming a possible polytomy with respect to Chelidurella. Genetic distance between the specimens of the Pyrenean and the Alpine clades is very large (uncorrected p distance = 0.19-0.22) ( Table 2). It is almost as high as those found among different genera of Forficulidae (Table 2), suggesting that nucleotide changes in cox1 might be already saturated at that level.
Male specimens included in the Pyrenean clade (Girona and Andorra) present large variability in the shape of the cerci. Cerci range from long, almost straight convergent cerci (Figs 3A, 4A), to very curved, broad, short cerci (Fig. 3D). Short cerci present the maximum curvature at the middle, forming an angle of approximately 90°; cerci are wider at the base, strongly narrowed at the area of greatest curvature (Fig. 3D, E), and maintaining a more or less constant width up to the apex (Fig. 3D). Long cerci are sub-cylindrical, slightly curved at their maximum width near the base, progressively narrowed towards the apex, acuminate at the end (Fig. 3A). Cerci may present a ridge on the inner margin ( Fig. 3B-F) or not (Fig. 3A). This ridge, when present, arises after the point of greatest curvature of the cerci and can continue to the apex of the cerci (Fig. 3D, E) or ending earlier, resembling a broad tooth (Fig. 3F). Intermediate specimens between these extreme shapes also occur (Figs 3B, C, F, 4C, D). In the same way, the diameter of the cerci is variable, including specimens with thick cerci compared to others with finer cerci. In all the individuals studied, cerci diameter expands over half or more of the width of the last segment. Males raised under controlled conditions from a single group of last instar nymphs from   Tossa d'Alp present long, short and intermediate cerci (Fig. 4A, C, D). The specimens studied from the Alpine clade present long cerci with little curvature, cylindrical apically, progressively narrowed towards the apex and the inner margins without teeth or with one tooth. The diameter of the cerci of those specimens studied is generally smaller than that of the specimens of the Pyrenean clade. However, our sample is not representative of the variability already reported for the Alpine clade (Burr 1912;Amiet 1961;David and Van Herrewege 1973;Sakai 1973;Caussanel and Albouy 1987;Albouy and Caussanel 1990;Herrera-Mesa 1999). Fontana et al. (2021) indicated that cerci variability in C. aptera is larger than previously considered, pending of a detailed geographic analysis.
The Pyrenean clade is sister to the single sequence representing the Massif Central clade (PP = 0.87). The genetic distance between Pyrenean and Massif Central populations is quite large (uncorrected p distance = 0.14-0.15). Among the large series of Pyrenean specimens studied we did not find the cerci morphology described for C. p. arverna by , Kirstová et al. (2020) and Fontana et al (2021). Cerci of the specimen included in the Massif Central clade, from Chalmazel (France), are more robust, relatively wider and less curved than cerci of the specimens of the Pyrenean clade. David and Van Herrewege (1973) confirmed that the morphology of the Pyrenean populations and those of the Massif Central (morphometric traits and male cerci), differ statistically.
Male genitalia from specimens of the Pyrenees, Alpine and Massif Central clades, including the lectotype of C. pyrenaica, the neotype of C. aptera and the holotype of C. p. arverna, were studied in detail and photographed by Fontana (1999) and Fontana et al. (2021). Maccagno (1933) also provided an illustration of the male genitalia of Pyrenean specimens corresponding to C. pyrenaica. The genitalia of the male specimens of C. pyrenaica we examined (Virgen de Ares, Lleida) match the description presented by Maccagno (1933) and Fontana (1999); variability is however large, including size of parameres. They differ from those of the Alpine specimens, by showing thinner parameres with almost parallel margins (shorter and curved on the external margin in Alpine specimens). Male genitalia of typical C. pyrenaica and the holotype of C. p. arverna do not differ significantly, although C. p. arverna seems to present a more arcuate vesicle.

Notes on Natural History
Chelidura pyrenaica is found in mountain slopes, between 1000 and 2500 m, usually in pastures in areas covered by flat stones, near the forest edge or in open areas (Fig. 5D, E) (Borelli 1905;Chopard 1922;Chopard 1951;David and Van Herrewege 1973;Albouy and Caussanel 1990). The geologic substrates of the area are diverse and complex, dominated by schists and limestone (see for a general overview Dendaletche 1982). Adult specimens are usually found under stones, bark of fallen trees and clods of earth in summer and fall (Azam 1901;Xambeu 1903;Chopard 1922;Albouy and Caussanel 1990). Xambeu (1903Xambeu ( , 1907 and Chopard (1951) mentioned that C. pyrenaica can be also found in spring. Mating takes place in April or May, in galleries that earwigs dig under their shelters. Females lay the eggs grouped in a shallow excavation at the end of one of these galleries and take care of them until the larvae hatch (Xambeu 1903;Caussanel and Albouy 1987;Albouy and Caussanel 1990). We observed female attending eggs in the second half of June in Andorra (Fig. 5A). Upon disturbance females hide in small burrows, but usually return soon to the egg mass and start moving the eggs to the burrow, holding them one by one with their mandibles. Herter (1943) indicates that females may lay 40-45 eggs per clutch. Diverse nymphal stages were observed in the second half of June in Andorra and in the first half of July in Coll d'Ares (Girona) also attended by females (Fig. 5B, C).

Discussion
There is a strict correspondence between mtDNA clades and geographic areas, with all samples from the Pyrenees included in a well-supported clade, sister to the Massif Central specimen, and those, in turn, related to the Alpine specimens. Sequences of the specimens from Tossa d'Alp (with short and long cerci respectively) are closer to each other than to the short cerci specimens from Andorra, therefore, at the molecular level, specimens with short and long cerci from the Pyrenees correspond to a single taxon. Results from the nymphs raised under controlled conditions, with adult males including long (see for example MNCN_Ent 296013; Fig 4A), short, and intermediate cerci (MNCN_Ent 269481; Fig. 4C) also support that cerci variability corresponds to a single taxon.
Populations of Chelidura from the French Massif Central have been treated as a differentiated subspecies, C. pyrenaica arverna Kirstová et al. 2020). The large genetic distance observed between Pyrenean and Massif Central populations of Chelidura suggests that that they have been isolated for long time. Lasting isolation between the Pyrenean and Central Massif populations is also supported by the morphological differentiation observed in male cerci. Reciprocal monophyly, long isolation reflected by large sequence divergence, and cerci morphological differentiation at morphometric level, suggest that C. arverna likely represents a separate taxonomic unit with respect to C. pyrenaica as previously suggested by Fontana et al. (2021). Using the evolutionary species concept (Wiley 1978(Wiley , 1981Wiley and Mayden 2000), there is little doubt that the Massif Central populations of Chelidura can be considered to represent an independent species: Chelidura arverna David and Van Herrewege, 1973 stat. nov. Intraspecific variability of morphological structures, as pygidium or cerci, is well known in earwigs (Srivastava 1970;Simpson and Mayer 1990;Tomkins and Simmons 1996;García-París 2017;Kirstová et al. 2020). Many species of Dermaptera show large variability in the size and shape of male cerci (Dohrn 1867;Diakonov 1925;Ollason 1970;Srivastava 1970;Mourier 1986;Simpson and Mayer 1990;González-Miguens et al. 2020;García-París et al. in press). The level of variability found in Pyrenean Chelidura is apparently higher than the levels of variability found by these previous authors in other taxa (Fontana et al. 2021 ; Fig. 3). This large variability in male sexual characters might be a consequence of strong sexual selection (Kawano 2006;Brown 2007). Alternatively, the large variability observed could be a consequence of the absence of directional selective pressures as Kirstová et al. (2020) mentioned as a possible explanation for the variability of the shape of the pygidium in some species of Chelidurella.
Most of the confusion derived from the early synonymization of Forficula simplex Germar, 1825, described based on Pyrenean specimens displaying long cerci (Germar, 1825), with C. aptera (Dohrn 1867;Bolívar 1876;Brunner von Wattenwyll 1882;Finot 1890;Kirby 1904;Sakai 1973;Harz and Kaltenbach 1976;Herrera-Mesa 1999). The treatment of F. simplex as a synonym of C. aptera carried two consequences, first the early inclusion of the Pyrenees within the geographic range of C. aptera, leading to the current confusion, and second, the unnecessary descriptions of Forficula pyrenaica Gené, 1832 and Forficula dilatata Burmeister, 1838, based on specimens also from the Pyrenees. According to the principle of priority (article 23 of International Code of Zoological Nomenclature, ICZN 1999), the name Forficula simplex Germar, 1825 has nomenclatural priority over Forficula pyrenaica Gené, 1832. However, both names meet the provisions of the article 23.9.1 (ICZN, 1999) to retain prevailing usage in order to assure nomenclatural stability. In consequence, we propose the reversion of precedence of F. simplex with respect to F. pyrenaica. Forficula simplex Germar, 1825 has not been used as a valid species name after 1899, and Forficula pyrenaica Gené, 1832 has been used in in at least 25 works, published by at least 10 authors in the immediately preceding 50 years and encompassing a span of not less than 10 years (see for example David and Van Herrewege 1973;Sakai 1973;Harz and Kaltenbach 1976;Lapeira and Pascual 1980;Vancassel and Foraste 1980;Shah 1984;Vancassel 1984;Caussanel and Albouy 1987;Dauphin 1987;Pascual 1988;Steinmann 1989Steinmann , 1993Albouy and Caussanel 1990;Dendaletche 1982;Coutin 1983;Plate 1987;Haas 1995;Herrera-Mesa 1996Guillet and Vancassel 2001;Klass 2001;Barrientos 2004;Dusoulier 2004;Matzke and Klass 2005;Costa 2006;Leraut 2007;Kirstová et al. 2020;Fontana et al. 2021). Therefore, and according to the Article 23.9.2. of the International Code of Zoological Nomenclature (ICZN 1999), the name Forficula pyrenaica Gené, 1832 is considered a nomen protectum with nomenclatural precedence over the name Forficula simplex Germar, 1825, a nomen oblitum.