A new species of Melitaea from Israel, with notes on taxonomy, cytogenetics, phylogeography and interspecific hybridization in the Melitaea persea complex (Lepidoptera, Nymphalidae)

Abstract Specimens with intermediate morphology are often considered to be the result of ongoing interspecific hybridization; however, this conclusion is difficult to prove without analysis of chromosomal and/or molecular markers. In the butterfly genus Melitaea, such an intermediacy can be detected in male genitalia, and is more or less regularly observed in localities where two closely related, presumably parental species are found in sympatry. Here I analyze a high altitude Melitaea population from Mt. Hermon in north Israel and show that its male genitalia are clearly differentiated from those found in phenotypically similar M. persea and M. didyma, but in some aspects intermediate between them. This hybrid-like population is unique because, although M. didyma is present on Mt. Hermon, the true, low-altitude M. persea has never been reported from Israel. Cytogenetic analysis revealed no apomorphic chromosomal characters to distinguish the Mt. Hermon population from other known taxa of the M. persea and M. didyma species groups. At the same time, DNA barcode-based phylogeographic study showed that this population is ancient. It was estimated to originate 1–1.6 million years ago in the Levantine refugium from a common ancestor with M. persea. Generally, the data obtained are incompatible with interpretation of the studied population as a taxon conspecific with M. persea or M. didyma, or a swarm of recent hybrids between M. persea and M. didyma, although the possibility of ancient homoploid hybrid speciation cannot be ruled out. I also argue that the name Melitaea montium assigned to butterflies from north Lebanon cannot be applied to the studied taxon from Mt. Hermon. Here I describe this morphologically and ecologically distinct entity as a new species Melitaea acentria sp. n., and compare it with other taxa of the M. persea complex.


Introduction
Butterflies of the genus Melitaea Fabricius, 1807 are distributed throughout the warm and temperate part of the Palaearctic region and occupy a wide range of habitat types, including meadows, grasslands, steppe, alpine biotopes, arid mountains and deserts (Tuzov and Churkin 2000). This group was revised by Higgins (1941Higgins ( , 1955 and more recently by Oorschot and Coutsis (2014) who used analysis of male genitalia as a main tool to document taxonomic structure of the genus. Despite these revisions, a large number of unresolved taxonomic questions persist among Melitaea, where specieslevel boundaries remain poorly defined. For example, DNA-barcode analysis revealed multiple deeply diverged lineages with properties of phylogenetic and partially biological species within Melitaea didyma (Esper, 1779) sensu lato, a widely distributed and common Melitaea species (Pazhenkova et al. 2015, Pazhenkova and.
Recent progress in improving our knowledge of relationships in Melitaea was made by using chromosomal (de Lesse 1960, Larsen 1975, Lukhtanov and Kuznetsova 1989, Hesselbarth et al. 1995 and molecular markers (Zimmermann et al. 1999, Long et al. 2014). In particular, molecular studies have helped to resolve some of the issues related to the composition of species groups within Melitaea Zimmermann 2000, Leneveu et al. 2009). However, with few exceptions (Kuznetsov et al. 2014, Toth et al. 2014, Pazhenkova et al. 2015, Pazhenkova and Lukhtanov 2016, molecular markers have not been used for analysis of taxonomic structure of Melitaea on level of closely related species or on intraspecific level. One of the most serious problems of the Melitaea taxonomy is the presence of so called "intermediates" (Oorschot and Coutsis 2014). The closely related sympatric species of the genus Melitaea can be distinguished by male genitalia structure; however, specimens with intermediate genitalia can be more or less regularly found in nature. Most likely, these intermediates represent results of recent interspecific hybridization (Oorschot and Coutsis 2014), but such a conclusion is difficult to prove without analysis of genetic markers. The majority of these intermediates are concentrated in south-west Asia where the widely distributed species M. persea Kollar, 1849 contacts with M. didyma (in Turkey and Armenia), M. interrupta Kolenati, 1848 (in the Russian Caucasus, Azerbaijan, Armenia, east Turkey, Iran and Turkmenistan), M. gina Higgins, 1941 (in Iran) and M. mixta Evans, 1912 (in Afghanistan and Pakistan) (Oorschot and Coutsis 2014).
While analyzing specimens of the genus Melitaea collected in 2013-2016 in Israel as a part of the Israeli butterflies DNA barcoding survey project, I encountered a series of distinctive samples, collected in June 2013 at high altitude of Mt. Hermon by Asya Novikova (the Hebrew University of Jerusalem). These specimens were preliminarily identified as M. persea montium Belter, 1934, a name described from north Lebanon (Belter 1934) and recently established to be a synonym of M. didyma (Oorschot and Coutsis 2014, pages 17-18). Analysis of their male genitalia revealed them to be clearly different from phenotypically most similar M. persea and M. didyma, but in some aspects intermediate between them. A subsequent search and collecting in 2013, 2014 and 2016 resulted in a number of additional specimens from Mt. Hermon and demonstrated that this population was sympatric and partially syntopic with phenotypically similar M. didyma liliputana Oberthür, 1909, M. deserticola Oberthür, 1909and M. trivia syriaca Rebel, 1905 as well as with phenotypically differentiated M. cinxia Linnaeus, 1758, M. telona Fruhstorfer, 1908and M. collina Lederer, 1861.
In an effort to analyze the origin of these unusual Israeli specimens and to determine their taxonomic status, their karyotype and morphology were studied and compared to those of M. persea and M. didyma. In addition, legs were sampled from all species and major populations in the M. didyma and M. persea groups (except for the extremely rare and local M. eberti Koçak, 1980 from N. Iran), and sequence data from the DNA barcode region of COI were obtained. The results of the M. didyma DNA barcode survey have already been published (Pazhenkova et al. 2015, Pazhenkova and. Herein I present the results of the M. persea DNA barcode analysis, and describe the distinctive Israeli Melitaea as a new species, Melitaea acentria sp. n.

Samples
Specimens examined are deposited in the Zoological Institute of the Russian Academy of Sciences, St. Petersburg, Russia and in the McGuire Center for Lepidoptera and Biodiversity (MGCL), Florida Museum of Natural History, University of Florida, Gainesville, Florida, USA. Photographs of all specimens used in the analysis, as well as collecting data, are available on the Barcode of Life Data System (BOLD) at http:// www.boldsystems.org/. Localities where specimens of the M. persea group were collected are shown in Figure 1.

Morphological analysis
For genitalia preparation, abdomens removed from adults were soaked in hot (90°C) 10% KOH for 3-10 min. Then they were transferred to water, the genitalia were carefully extracted and examined under a stereo-microscope using a pair of prepara-  tion needles or a needle and a watchmaker's tweezer. Once cleansed of all unwanted elements they were transferred and stored in tubes with glycerine. Cleansed genitalia armatures were handled, studied and photographed while immersed in glycerine, free from pressure due to mounting, and therefore free from the ensuing distortion. Genitalia photographs were taken with a Leica M205C binocular microscope equipped with a Leica DFC495 digital camera, and processed using the Leica Application Suite, version 4.5.0 software.

Molecular methods and DNA barcode-based phylogeographic study
The barcode analysis involved 96 COI sequences (Appendix 1: Table 1) including 53 samples of the species close to M. persea (M. persea, M. acentria sp. n. and M. higginsi) and 13 samples of M. casta that was previously recovered as a sister group to M. persea (Leneveu et al. 2009). It also involved samples of the phenotypically similar species M. didyma liliputana (7 samples), M. deserticola (14 samples) and M. trivia syriaca (9 samples) collected in Israel, Jordan and Syria. Nine M. trivia syriaca samples were selected as an outgroup.
Sequences were aligned using the BioEdit software (Hall 1999) and edited manually. Phylogenetic hypotheses were inferred using Bayesian inference as described previously (Vershinina andLukhtanov 2010, Lukhtanov et al. 2016a, b). Briefly, the Bayesian analysis was performed using the program MrBayes 3.2 (Ronquist et al. 2012) with default settings as suggested by Mesquite (Maddison and Maddison 2015): burn-in=0.25, nst=6 (GTR + I + G). Two runs of 10,000,000 generations with four chains (one cold and three heated) were performed. The consensus of the obtained trees was visualised using FigTree 1.3.1 (http://tree.bio.ed.ac.uk/software/figtree/). I used two criteria to evaluate the level of DNA barcode divergence between taxa and haplogroups. First, I calculated the number of fixed DNA substitutions, i.e. the number of invariable differences in the studied COI fragment. Second, I calculated the minimal uncorrected COI p-distance between taxa and haplogroups. For this calculation, two genetically closest samples from each taxon pair were selected, and the distance between them was calculated using both fixed and non-fixed substitutions.

Chromosomal analysis
Karyotypes were obtained from fresh adult males and processed as previously de scribed (Lukhtanov et al. 2014, 2015a, Vishnevskaya et al. 2016. Briefly, gonads were removed from the abdomen and placed into freshly prepared fixative (3:1; 96% ethanol and glacial acetic acid) directly after capturing the butterfly in the field. Testes were stored in the fixative for 1 month at +4°C. Then the gonads were stained in 2% acetic orcein for 7-10 days at +18-20°C. Haploid chromosome numbers (n) were counted in meiotic prometaphase, metaphase I (MI) and metaphase II (MII).
Males (Fig. 2a-c). Forewing length 16-19 mm. Forewing is roundish. Upperside: ground color orange-red; the wing markings are small and delicate when compared to those in M. didyma and M. persea. Forewings with very narrow black marginal border fused with internervural marginal black spots. Submargimal series formed by black triangular spots on the forewings and by fine lunules on the hindwing. Forewing postdiscal series formed by small 1-3 black spots. Forewing discal series is complete or nearly complete, formed by black spots of variable size, the first four spots near costa are often enlarged. Hindwing discal series reduced or absent. Basal marking of the fore-and hindwings is delicate. Black basal suffusion is developed only near the base of hindwings. Fringe is white, checkered by black dots.
Underside: forewing ground color orange-red except for the apical part which is yellowish. Black markings delicate, reduced as compared with those of the upperside of the wing. Hindwing ground colour yellowish-white with two orange-red fascias. The red-orange submarginal fascia shows segmentation as the yellowish-white ground color spreads along the nervures. The orange-red macules of this fascia are bordered by back lunules from the outer side. From the inner side these macules are edged by black scales and additionally bordered by black lunules, giving the appearance that the proximal border of the submarginal fascia is doubly edged. Fringe white, checkered by black dots.
Females (Figs 2d,3). Forewing length 17-20 mm. Forewing is roundish. Ground color of the upperside is slightly lighter and black markings heavier than in males. Costal area of the wing apex yellow-orange. Underside of the forewings as in males but black markings are heavier and there are additional yellowish maculae between discal and postdiscal spots. Underside of the hindwings as in males. Fringe white, checkered by black dots.
Karyotype. The genus Melitaea is known for its relatively low interspecific chromosome number variation (Pazhenkova and Lukhtanov 2016). However, in certain cases, the chromosome numbers are key to distinguish between closely related Melitaea species. For example, karyotype differences in combination with information about parapatric distribution were the main argument for the non-conspecificity of M. didyma and M. latonigena Eversmann, 1847 (Lukhtanov and Kuznetsova 1989). Therefore, chromosomal analysis is highly desirable in any taxonomic study of Melitaea. Here I conducted the chromosomal analysis of the high altitude population from Mt. Hermon (M. acentria sp. n.). The haploid chromosome number n=27 was found  in prometaphase I, MI and MII cells of three studied individuals (2016-006, CCDB-25458_C11; 2016-008, CCDB-25458_D01; 2016-009, CCDB-25458_D02) (Fig. 4). The MI karyotype contained one chromosome bivalent that was significantly larger than the rest of the bivalents.
The same chromosome number (n=27) was previously reported for M. persea from Iran (de Lesse 1960). A karyotype characterized by n=27 including one large chromosome element was also found in M. didyma neera Fischer de Waldheim, 1840 from the North Caucasus (Russia), although in some other studied populations of M. didyma n=28 was found (de Lesse 1960, Lukhtanov andKuznetsova 1989). The chromosome number n=27 was also mentioned for "M. didyma libanotica" from Lebanon (Larsen 1975), but the vouchers for this chromosomal analysis were larvae, and in my opinion their identification was not certain. They could represent M. didyma liliputana, but also M. acentria as well as M. persea (but certainly not M. deserticola in which n=29 was found and not M. trivia in which n=31 was found, Larsen 1975). Finally, n=27 was reported for "M. montium" from Lebanon (de Lesse 1960), but in the last case the identity of the studied samples was also not clear because the identification was not supported by genitalia analysis.
Thus, no fixed karyotype difference is known to exist between M. acentria and M. persea as well as between M. acentria and M. didyma. Therefore we cannot use the available chromosomal data for delimitation between these species.
Male genitalia structure. M. didyma from Israel (Mt. Hermon) and M. persea from Iran and Azerbaijan were analyzed and were found to possess typical characters described previously (Higgins 1941, Oorschot andCoutsis 2014).
In M. persea all the main structures (ring-wall, tegumen, saccus, valvae) are elongated ( Fig. 5a, b), longest in the genus Melitaea (Oorschot and Coutsis 2014). The valva is elongated from lateral view (Fig. 6a) and the valval distal process is massive (Fig. 6b). The dorsum of the valval distal process lies nearly in line with the remainder of the valval dorsum (Fig. 6a). The ventrum of the valval distal process possesses a keel bearing strong teeth (Fig. 6b). The saccus is bifurcate, with long, distally pointed branches (Fig. 5a, b). The aedeagus is curved, with a pronounced dorso-lateral ridge (Fig. 7a). The lateral sclerotized element of the tegumen is massive and its distal half is shaped like a smoker's pipe (Fig. 7b).
In M. didyma liliputana from Mt. Hermon all the main structures (ring-wall, tegumen, saccus, valvae) are significantly shorter than in M. persea (Fig. 5e, f). The valva is trapezoidal from lateral view (Fig. 6e). The valval distal process is delicate (Fig. 6f) and the dorsum of the valval distal process forms a clear angle with the remainder of the valval dorsum (Fig. 6e). The ventrum of the valval distal process is smooth, without a keel and/or teeth (Fig. 6f). The saccus is bifurcate, with short, distally rounded branches (Fig. 5e, f). The aedeagus is curved, without a pronounced dorso-lateral ridge (Fig. 7e). The lateral sclerotized element of the tegumen is delicate and its distal half is T-or Γ-shaped (Fig. 7f).
In M. acentria genitalia are clearly different from both M. persea and M. didyma, but at the same time are intermediate in some aspects. All the main structures (ringwall, tegumen, saccus, valvae) are similar to those in M. persea but shorter (however, longer than in M. didyma) (Fig. 5c,d). The valva is cylindrical from lateral view (Fig.  6c). The valval distal process is intermediate in its shape between M. persea and M. didyma (Fig. 6c, d). Its dorsal and ventral borders are roughly parallel from lateral view (Fig. 6c). The dorsum of the valval distal process forms a clear angle with the remainder of the valval dorsum (similarly to M. didyma) (Fig. 6c). At the same time, the ventrum of the valval distal process possesses a keel bearing teeth (similarly to M. persea) (Fig. 6d). However, this keel and teeth are smaller and more delicate than  in M. persea (Fig. 6d). The saccus is bifurcate, with relatively long, distally pointed branches; however, these branches are shorter than in M. persea, but longer than in M. didyma liliputana, where they are almost absent (Fig. 5c, d). The aedeagus is curved, with a dorso-lateral ridge (Fig. 7c); thus the aedeagus of M. acentria is not intermediate between M. persea and M. didyma, but similar to M. persea. The lateral sclerotized element of the tegumen is massive and its distal half is shaped like a smoker's pipe (Fig. 7d). This type of male genitalia was found in all seven studied samples including two samples (25453_E08 and 25458_C09) that were characterized by the mitochondrial haplogroup P2 (Figs 8 and 9). The analysis recovered the M. persea group (M. acentria + M. persea +M. higginsi) as a strongly supported monophyletic clade sister to M. casta (Fig. 9). This clade was divided into five lineages.
The first lineage (haplogroup P1) includes a huge range of M. persea populations from Daghestan (Russia) in the north to Shiraz province (Iran) in the south, including samples from Shiraz in SW Iran, which represents the type locality of M. persea. Across this range, M. persea shows various degrees of localized morphological diversification, and from this territory several taxa, currently attributed to M. persea, were described:    Belter, 1934;M. pesea hafiz Higgins, 1941; M. hafiz darius Gross &Ebert, 1975 andM. jitka D.Weiss &Major 2000. The taxonomy of these taxa was studied in more detail by Oorschot and Coutsis (2014), who found that they are closely related and should be considered no more than synonyms of M. persea persea. My DNA barcode results are consistent with this conclusion (Fig. 9).
The haplogroup P1 includes also a female sample 17966_F12 possessing intermediate morphological characters between M. interrupta and M. persea (Fig. 10b). In this specimen wing upperside is similar to that in M. interrupta, whereas the wing underside is without black scales along the veins which are typical for M. interrupta (Fig. 10h), but with orange-red submarginal spots edged by black scales typical for M. persea (Fig. 2e-h, 10a). This sample was collected at the same place with three typical M. interrupta males (samples17966_F09, 17966_F10 and 17966_F11) possessing typical M. interrupta phenotype (Fig. 10h) and COI haplotypes (GenBank # KT874702, KT874740 and KT874741), which were very different from those of M. persea (see Fig. 4 in Pazhenkova and Lukhtanov 2016). It is thus probable that the female 17966_F12 is a result of a more or less recent hybridization between M. interrupta and M. persea. Thus, it likely represents a first molecular evidence for sporadic interspecific hybridization in Melitaea.
The second lineage (haplogroup P2) is represented by three specimens from north Lebanon originally identified as M. persea (Wahlberg et al. 2005) and by three samples of M. acentria from Mt. Hermon: two males (25453_E08 and 25458_C09) that were indistinguishable in their genitalia from M. acentria of the haplogroup A and a single female (25458_E08). This lineage was found to be closest to P1 (Melitaea persea persea). It differed from P1 by 7 fixed DNA substitutions in the studied 658 bp fragment of the mitochondrial COI gene. The minimal uncorrected COI p-distance between the representatives of these two haplogroups was calculated using both fixed and non-fixed substitutions and was found to be 2.0 %.
The third lineage (haplogroup P3) includes samples from NE Iran (M. persea paphlagonia). It differed from P1 (M. persea persea) by 10 fixed DNA substitutions in the studied 658 bp fragment of the mitochondrial COI gene. The minimal uncorrected COI p-distance between these two haplogroups was found to be 2.3 %. They were also distinct in wing pattern: on the upper surface all the markings were well developed and the first four spots of the discal series were nearly fused to form a prominent costal bar (Fig. 10c, d). The male genitalia of M. persea paphlagonia were similar to those found in M. persea persea (Higgins 1941). This lineage was not recognized as a taxon by Oorschot and Coutsis (2014). However, it was recognized as a distinct subspecies by Higgins (1941), and my DNA barcode results corroborate this conclusion. The level of COI differentiation between M. persea paphlagonia and M. persea persea (10 fixed DNA substitutions) was found to be equal to that found between M. persea persea and M. higginsi (10 fixed DNA substitutions).
The forth lineage (haplogroup A), one of the most diverged lineages, is represented by samples from Mt. Hermon (M. acentria). It differed from P1 (M. persea persea) by 11 fixed nucleotide substitutions in the studied 658 bp fragment of the mitochondrial COI gene. The minimal uncorrected COI p-distance between these two haplogroups was found to be 2.4 %.
The fifth lineage (haplogroup H) includes samples of M. higginsi (Fig. 10e, f). This taxon is very rare in collections, and I have been lucky to find two specimens in the McGuire Center. It differed from P1 (M. persea persea) by 10 fixed DNA substitutions in the studied 658 bp fragment of the mitochondrial COI gene. The minimal uncorrected COI p-distance between these two haplogroups was found to be 2.4 %. This taxon is similar to M. persea with respect to male genitalia structure (Oorschot and Coutsis 2014), but quite different in wing pattern. Particularly, in males the hindwing uppersurface is without black spots which are always present in M. persea, and in both sexes hindwing underside veins are scaled with black, similar to M. interrupta and different from M. persea. My DNA barcode results confirm the distinctness of this high altitude very local Afghani taxon. They also confirm that this taxon is a member of the M. persea species group as suggested by Oorschot and Coutsis (2014), and not related to the Mongolian M. didymina Staudinger, 1895 as was supposed by Sakai (1978), as well as not related to M. didyma as was supposed by Kolesnichenko and Churkin (2004).
Diagnosis. Butterfly wing pattern and male genitalia morphology, as well as DNA barcodes certaintly indicate that Melitaea acentria belongs to the M. persea species complex. After Oorschot and Coutsis (2014) this complex includes three closely related species: M. persea, M. eberti and M. higginsi. Male genitalia of these three species were analysed by Oorschot and Coutsis (2014) and were found to be virtually indistinguishable. Melitaea acentria differs from these most closely related species by several characters in male genitalia. In M. acentria main genitalia structures (ring-wall, tegumen, saccus, valvae) are significantly shorter. The valva is cylindrical from lateral view, not elongated (Fig. 6c). The valval distal process is intermediate in its form between M. persea and M. didyma (Fig. 6c, d). The dorsum of the valval distal process forms a clear angle with the remainder of the valval dorsum (in similar way as in M. didyma) (Fig.  6c), but very different from M. persea. The keel and teeth of the valval distal process are smaller and more delicate than in M. persea (Fig. 6d). On average, in M. acentria the ground color of the wing upperside is more orange-red ( Fig. 2a-d). In other species of the M. persea complex it is yellowish-orange ( Fig. 2e-g, 10a, c-f). However, this character is not constant (e.g. see M. persea with orange-red wing color on Fig. 2h). The great majority of M. acentria samples significantly differ from all other taxa by their DNA barcodes; however, probably due to mitochondrial introgression, a minor part of the samples cluster with the haplogroup P2 of M. persea.
Melitaea acentria significantly differs from the distantly related but phenotypically similar species M. didyma, M. deserticola and M. trivia by DNA barcodes and male genitalia structures. Particularly, it differs from M. didyma by the ventrum of the valval distal process possessing a keel bearing teeth and by the elongated shape of the ringwall, tegumen, saccus and valvae. Melitaea acentria mostly differs from M. didyma by the hindwing underside with submarginal macules that are edged by black scales and then bordered by black lunules, giving the impression that the proximal border of the submarginal fascia is doubly edged; M. acentria shares this character with M. persea. In M. didyma submarginal macules of the hindwing underside are usually not edged by black scales and simply bordered by black strokes (Fig. 10g). However, elements of the black scaling of the submarginal macules can be found in few M. didyma samples, and sometimes this black scaling is strongly reduced in species of the M. persea complex.
Distribution. Melitaea acentria is known to occur at high altitudes (1730-2060 m above the sea level) of Mt. Hermon (Fig. 11 (Kent et al. 2013). Adults of M. acentria were found to fly in open grassy (Fig. 12) and stony (Fig. 13) areas of the upper part of the xero-montane open forest belt (1750-1850 m) (Fig. 14) and of the subalpine mountain steppe belt (1850-2060 m) (Fig. 15). Butterflies were observed from 3 May to 3 July. On the 3 rd of May 2016 they were abundant at altitudes from 1780 to 1900 m, therefore I conclude that they can start to fly at the end of April and continue to fly at least until mid-July.    The building with red roof is the winter café shown as 3 on Figure 11. Photo by V. Lukhtanov. Etymology. The name acentria is a noun of the feminine gender. This name originates from the Greek prefix "a" that means "not" and from the Latin word "centrum" (centre) derived from the Greek "κέντρον" (kentron, a sharp point). Acentria is the Internet nickname of Asya Novikova who collected the samples initiated this research. This name indicates also the peripheral position of the new species within the distribution range of the M. persea species complex.

Hypothesized evolutionary history of Melitaea acentria
Melitaea acentria was recovered as a diphyletic group with respect to COI barcodes being represented by two haplogroups A and P2. The major haplogroup A (22 samples of 25 studied) represents one of the most differentiated and thus most ancient mitochondrial lineages within the M. persea complex. The minor haplogroup P2 (3 samples of 25 studied) is also differentiated, but is more similar to the haplogroup P1 found in the core part of the M. persea species range.
To estimate the age of the haplogroup A (and the Israeli lineage as a whole) I used two calibration points: a lower rate of 1.5% uncorrected pairwise distance per million years estimated using a variety of invertebrates (Quek et al. 2004) for COI, and a faster rate of 2.3% uncorrected pairwise distance per million years for the entire mitochondrial genome of various arthropod taxa (Brower 1994). Using these points and the value 2.4% as the minimal uncorrected COI p-distance between the haplogroups A and P1, the haplogroup A can be estimated to originate approximately 1-1.6 million years ago from a common ancestor with the haplogroup P1 of M. persea, a species currently distributed throughout the whole Middle East. The Israeli lineage represented by haplogroup A evolved in isolation in the Levantine refugium and, most likely, relatively recently experienced episodes of hybridization with M. persea (haplogroup P2) resulting in mitochondrial introgression observed in the samples 25453_E08, 25458_C09 and 25458_E08. Thus, the haplogroup P2 of M. acentria seems to be a footprint of this introgression. Despite this supposed sporadic hybridization, the population from Mt. Hermon preserves clear diagnostic characters in male genitalia.
Melitaea acentria possesses male genitalia which are different from those found in both M. persea and M. didyma, but in some aspects intermediate between them. Such an intermediacy can theoretically be interpreted as a consequence of (i) an ancient hybridization resulting in homoploid hybrid speciation or (ii) a more recent hybridization resulting in the formation of a swarm of recently obtained hybrids (Lukhtanov et al. 2015b). In the first case (homoploid hybrid speciation), a new reproductively isolated, sexually reproducing species arises through hybridization and combination of parts of the parental genomes, but without an increase in ploidy (Rieseberg et al. 1995, Rieseberg 1997, Coyne and Orr 2004, Lai et al. 2005. In the second case (formation of hybrid swarms), interspecific hybridization results in a number of individuals which are not reproductively isolated from their parents and, thus, do not represent a new species (Lukhtanov et al. 2015b). The second scenario caused by occasional hybridization probably occurs in Melitaea as demonstated by the sample 17966_F12 with an intermediate morphology and, most likely, introgressed mitochondria.
Oorschot and Coutsis (2014) treated the Lebanese samples with mixed M. persea -M. didyma genitalia type as intermediates (i.e. hybrids) between M. persea and M. didyma. Such an interpretation seems to be logical for Lebanon where M. persea in its more or less typical form has been reported (Higgins 1941, Larsen 1974, Racheli 1980. However, this interpretation is inappropriate for the Israeli population (i.e. for M. acentria). First, typical M. persea has never been reported from Israel, which is one of the best studied territories in the world with respect to the butterfly fauna. Second and most importantly, the M. acentria samples posses very divergent COI haplotypes which can be attributed neither to M. persea and nor to M. didyma. Thus, M. acentria is not a swarm of recently obtained hybrids, but an old, well-established, morphologically and ecologically differentiated lineage with clear properties of phylogenetic and biological species.
At the same time, the hypothesis that M. acentria is a result of ancient homoploid hybrid speciation can not be ruled out. This highly speculative hypothesis should be tested in future through full genome molecular and chromosomal studies. While such a mode of speciation is widely accepted in plants (e.g., Soltis 2013), it has only relatively recently been thoroughly investigated in animals, including butterflies (Gompert et al. 2006, Mavárez et al. 2006, Mallet 2007, Kunte et al. 2011, Dupuis and Sperling 2015, Lukhtanov et al. 2015b).

Why Melitaea montium Belter, 1934 cannot be used as a valid name?
The identity of the taxon described under the name Melitaea montium Belter, 1934 has never been clear. Belter (1934) reported a difference between M. montium and M. didyma in the shape of the male genitalia valva. Although this supposed difference looks very distinct in Belter's drawing (Fig. 13 in Belter 1934) in fact it can hardly be traced. The real difference between these taxa is in the form of the tegumen, the distal process and, especially, of the saccus; however these structures were not shown on the very schematic drawings from Belter's paper. Thus, the genitalia description and figures provide little information on the identity of the taxon described as M. montium (see a more detailed discussion on this topic in the monograph by Oorshot and Coutsis (2014)).
The same can be said about wing pattern. In fact, Belter was the first author who described two types of hindwing underside in Melitaea: (i) with the proximal double (see Fig. 2b-h; 10a, c, d, f) and (ii) with the proximal single black border of the submarginal fascia (see Fig. 10b, g, h), and mentioned that both types existed in M. montium. These two types were later referred to as types "a" and "b" (Hesselbarth et al. 1995). None of these two types is species-specific, although the type (i) is much more common in the M. persea complex, and the type (ii) is more common in the M. didyma complex (Hesselbarth et al. 1995, Larsen 1974. After Belter, the name Melitaea montium was used in literature for the Middle Eastern (Higgins 1941, Hesselbarth et al. 1995, Tshikolovets 2011, Lebanese (Larsen 1974, Gross and Ebert 1975, Racheli 1980 and Israeli (Benyamini 2002) butterflies close or supposedly identical to M. persea. It was also used as a synonym of M. didyma (Oorshot and Coutsis 2014).
I should note that the identity of butterflies in these publications has never been clear, except for the monograph by Oorshot and Coutsis (2014) since at least three different groups of populations close to M. persea are recorded from the Middle East: 1) the populations close (but probably not identical) to true M. persea (Higgins 1941, Larsen 1974, Racheli 1980, Oorshot and Coutsis 2014, 2) M. acentria from Israel (this study), and 3) intermediates (hybrids) between M. persea and M. didyma from north Lebanon (Oorshot and Coutsis 2014). These three groups could be identified on basis of male genitalia characters; however, until the work of Oorshot and Coutsis (2014), the genitalia of these butterflies were not carefully studied. Higgins (1941) provided only schematic genitalia drawings that were good enough to exclude M. didyma from consideration, but not detailed enough to distinguish between M. persea and M. persea-M. didyma intermediates, and the consequent authors did not provide genitalia drawings at all (see the monograph of Oorshot and Coutsis (2014) for a more detailed analysis of the previous taxonomic interpretations).
The Gordian knot of this taxonomic and nomenclatural uncertainty was cut by Oorshot and Coutsis (2014)  I should also note that, despite the valid lectotype designation resulting in this synonymy, the name M. montium could theoretically be preserved for a valid taxon under the plenary power of the International Commission on Zoological Nomenclature through a neotype designation. Such a possibility exists for the cases in which the existing name-bearing type of a nominal species-group taxon is not in taxonomic accord with the prevailing usage of names and stability or universality is threatened thereby (Article 75.6, http://iczn.org/iczn/index.jsp). However, in case of M. montium the article 75.6 can hardly be applied because de facto the prevailing usage cannot be calculated. After Belter's publication there were few cases when this name was used, and in each case the identity of the butterflies called M. montium was unclear.
In this situation I see no other way than following the latest comprehensive revision (Oorshot and Coutsis 2014) that established the synonymy: M. didyma = M. montium based on lectotype designation and analysis.

Conclusion
The Melitaea persea species complex consists of the following taxa: The identity and taxonomic status of the M. persea-similar samples from north Lebanon, Jordan, Iraq, Pakistan, and Afghanistan remain still unclear. The populations from Lebanon characterized by the mitochondrial haplogroup P2 (Fig. 9) could actually represent (i) a distinct subspecies of M. persea, (ii) an undescribed subspecies of M. acentria, or even (iii) an undescribed species. Further morphological, molecular and chromosomal studies are required to select between these hypotheses.
The financial support for all molecular, chromosomal and morphological studies was provided by the grant N 14-14-00541 from the Russian Science Foundation to the Zoological Institute of the Russian Academy of Sciences. The travels to Israel and field studies of V. Lukhtanov were supported by RFBR grants 15-04-01581 and 15-29-02533. The work was partially performed using equipment of the 'Chromas' Core Facility, the Centre for Molecular and Cell Technologies and the Department of Entomo logy of St. Petersburg State University. A part of this equipment was purchased with support of the St. Petersburg University grant 1.40.490.2017.