the Balkan Psorodonotus ( orthoptera : tettigoniidae ) : testing the existing taxa confirmed presence of three distinct species

A review of the Balkan representatives of the genus Psorodonotus was made with the aim of revealing the relationships between taxa and their systematic arrangement. For this purpose we used morphological (qualitative and quantitavive), acoustic (amplitude-temporal song parameters) and molecular (mtDnA CoI gene sequence) data. The analyses and comparisons with other taxa of Psorodonotus support the distinction of three species occurring on the Balkan Peninsula: P. fieberi, P. illyricus and P. macedonicus. Analysis of the CoI sequences suggested the following relationships: outgroups + (P. illyricus + (P. fieberi + (P. macedonicus + P. caucasicus))). All available information on the distribution of the Balkan taxa was synthesized and mapped. * Corresponding author. Authors’ contributions: The idea conceived by bÇ; data produced by Sk, DPC, jS, kGH, and bÇ; DPC and BÇ wrote the paper.


IntroductIon
Brunner von Wattenwyl (1861) established the genus Psorodonotus (originally a subgenus of Decticus) for P. pancici Brunner von Wattenwyl, 1861 (presently regarded a synonym of P. fieberi (Frivaldszky in Fieber, 1853) and the then known Pterolepis fieberi and Pterolepis venosa Fischer de Waldheim, 1839 (described under Peltastes Fischer de Waldheim, 1839, which was recognized by its author as a preoccupied name; Fischer de Waldheim, 1846).Later on, Brunner von Wattenwyl (1882) included in this genus also Pterolepis specularis Fischer de Waldheim, 1839 (also described under Peltastes).At present, the genus is thought to include 22 valid taxa (15 species) (Eades et al., 2014;Kaya et al., 2014), distributed in the mountainous regions of the Caucasus, Anatolia and the Balkan Peninsula.The highest diversity of species is concentrated in Eastern and Northeastern Anatolia and the Caucasus-Transcaucasus region.Three valid taxa are presently known from the Balkan Peninsula: Psorodonotus fieberi fieberi (Frivaldszky in Fieber, 1853), Psorodonotus fieberi illyricus Ebner, 1923 and Psorodonotus fieberi macedonicus Ramme, 1931.After their description, significant discrepancies existed between authors concerning the taxonomic status and systematics of the Balkan taxa.The taxa are regarded either as subspecies of two species (e.g., Ramme, 1931;Mikšić, 1965;Harz, 1969), separate species (Ramme, 1951) or subspecies of a single species (Heller, 1988;Eades et al., 2014).Most of the cited studies rely on morphological criteria.Heller (1988) used parameters of the male calling song in addition to general morphology to define the subspecies relationships of the three taxa, a view that is currently accepted (Eades et al., 2014).Recent studies on related taxa in Anatolia (Ünal, 2013;Kaya et al., 2013Kaya et al., , 2014;;Taylan et al., 2014) revealed new distinct lineages in the genus with morphological and acoustic differences of the same order, which showed significant genetic differences (Kaya et al., data in prep.).Thus, the aim of the present study is to test the hypotheses concerning the taxonomic status of the Balkan taxa and reveal their relationships using molecular, morphological and acoustic data.

MaterIal and Methods sampling
For the purposes of the present study we concentrated on collecting specimens from at least two mountain areas within the geographically outlined ranges of the known taxa (e.g., Mikšić, 1965;Harz, 1969).The material used for morphological and molecular studies is listed in Appendix 1.

Morphology
Specimens from the populations sampled were preserved in 70-96% ethanol in order to preserve original body shape.Upon preliminary identification (according to Ramme, 1951;Mikšić, 1965;Harz, 1969) specimens were arranged into respective taxonomic units (Appendix 1).We examined the morphological characters previously used for discriminating taxa.Respective morpho-structures were photographed, qualitatively examined and measured using a Leica MZ6/DC600 stereomicroscope equipped with a digital camera and Image J v. 1.36 software (http://rsb.info.nih.gov/ij/).
Previously, the Balkan representatives of Psorodonotus were distinguished mainly on their cercal morphology, titillator and ovipositor shape, and pronotum size (Ramme, 1951).Therefore, we conducted a geometric morphometrics analysis of all of these structures except the titillators.Illustrations of the pronotum and Kaya et al. (2013Kaya et al. ( , 2014) ) (Fig. 1).The intersections marked in the TPS files were aligned using a Procrustes superimposition and matrices of partial warp scores for each structure were generated using MorphoJ v.1.03d(Klingenberg, 2011).Partial warp scores were analyzed using Canonical variate (Cv) analysis using the same software.song Male songs were recorded using the following equipment: (1) electret condenser microphone Knowles BT-1759-000 (with a cut-off at about 45 kHz; W. Schulze, Friedrich-Alexander Universität Erlangen-Nürnberg, pers.comm.), equipped with a custom-made preamplifier, connected to a ZooM H2 handy recorder (Zoom Corporation) (96 kHz sampling rate); (2) Pettersson D500 external microphone (frequency range 1/2->100 kHz) connected to a ZooM H2 handy recorder; (3) UHeR 4200 IC tape recorder with microphone UHeR M645.

Bioacoustic terminology
Modified from Heller et al., 2004;kaya et al., 2014.Calling song: the song produced by an isolated male; phrase: a group of densely arranged song elements or syllables forming main discrete part of the song; syllable: sound produced during one cycle of movements of tegmina (opening -often quiet -and closing -typically loud); here we termed the main distinguishable part of the song a syllable, which may be a single repeated unit of the phrase used by Kaya et al. (2014); element: a distinguishable compact oscillographic component that can be practically measured; syllables repetition rate: reciprocal of the syllable period (unit Hz = 1/s); syllable period: the duration of a syllable and the following silent interval until next syllable starts; syllable duration: the time interval starting from beginning of one repeated unit to its end; impulse: a recognizable undivided transient train of sound waves produced by the stridulatory tooth striking the plectrum (the anal edge of the opposite tegmen).

song analysis
For manual and statistical comparisons of temporal song parameters between populations we selected the less variable and easily detectable part of the song: the second/main part of the syllable (see Results: Bioacoustics).As temperature (i.e., body temperature) is the most important factor affecting temporal song parameters in orthoptera (e.g., Martin et al., 2000;Arias et al., 2012) we produced a subjective manual calibration of the recording temperature for the recordings of P. illyricus (see Results).

Molecular sequences and phylogenetic analysis
For determining the phylogenetic relationships between taxa we used Cytochrome c oxidase subunit I (CoI) fragments of mitochondrial DNA (mtDNA).This region of mtDNA has revealed phylogenies of high resolution for bushcrickets (e.g., Ullrich et al., 2010) and also provided quality results for Psorodonotus (Kaya et al., in prep.).
The sequences were aligned manually in Sequencher v. 4.1 (Gene Codes Corporation) and checked manually by eye.DnaSP v. 5 (Librado & Rozas, 2009) was used to determine unique haplotypes.The haplotype matrix was prepared using MeGA v. 5 (Tamura et al., 2011).A maximum parsimony (MP) analysis was carried out with 100 random additions following the heuristic search approach and the tree bisection-reconnection (TBR) algorithm using PAUP v. 4.10b (Swofford, 2000).A 10,000 non-parametric bootstrap resampling was used to assess the branching confidenc-es (Felsenstein, 1985).The parameters and best fit model were estimated using jModelTest v. 0.1.1.(Guindon & Gascuel, 2003;Darriba et al., 2012).The selected model was implemented in maximum likelihood (ML) and Bayesian phylogenetic inference (BI) analyses.ML was performed using a heuristic search procedure with TBR branch swapping and as-is addition replicates using PAUP v 4.10b.Nonparametric bootstrapping (Felsenstein, 1985) was used to evaluate the support of nodes based on 1000 pseudoreplicates analyzed using ML.BI analysis was carried out in Mrbayes v. 3.1.2(Ronquist & Huelsenbeck, 2003;Ronquist et al., 2005) using four simulations of Markov chains, 2,000,000 generations and sampling every 100 generations.The software tool TRACER v. 1.5 (see Rambaut et al., 2014) was used to examine the parameters and determine the number of trees needed to reach stationarity.BI posterior branch probabilities were calculated using the majority rule consensus of the sampled trees, excluding the first 200 trees as burn-in.P. specularis, P. venosus and P. caucasicus were chosen as out groups.Pairwise genetic distances of species were calculated using MeGA.

Morphology
Preliminary identification of the populations sampled fitted the data from earlier sources very well (e.g., Ramme, 1951;Mikšić, 1965;Harz, 1969).Specimens also conform with the Psorodonotus caucasicus species group as defined by kaya et al. ( 2013), which may be characterized morphologically by the dorsal surface of the pronotum being distinctly rugose but less so than in the venosus species group (smooth or slightly wrinkled in the specularis group).For their general appearance see Fig. 2. Graphs of the results of the Cv analyses of matrices of partial warp scores recorded for male and female pronota, cerci and ovipositors are presented in Fig. 3.

Bioacoustics
Acoustically, the Psorodonotus caucasicus group is characterized by a song consisting of syllables (monosyllabic phrases if compared to the other two species groups of this genus) that are composed of two isolated elements (possibly resulting from a step-interrupted closing move- ment of the tegmina).This is the main difference between the P. caucasicus group and the P. specularis and P. venosus groups, in which two or more syllables are grouped into phrases (oscillographically consisting of three or more song elements) (Kaya et al., 2013(Kaya et al., , 2014)).
The syllable (Fig. 8) consists of a quiet (low amplitude) first element, which may be compact (P.fieberi, P. macedonicus) or separated into two parts (P.illyricus), and after a short silent interval, a second (main) element.The main element starts with a crescendo of dense impulses, in which the maximum amplitude is reached and then the impulses become sparser at the end of the element.The main element may finish abruptly (in P. illyricus) or be followed by a few isolated impulses or after-clicks (P.fieberi and P. macedonicus).There is a tendency for the number of after-clicks to increase with decrease in body temperature.Syllable length and syllables repetition rate depend on temperature.As all the recordings of P. illyricus obtained in this study do not seem to reflect the real body temperature (the animals were exposed either to sun or to the light from an incandescent lamp, see Table 1) the approximate body temperature during recordings was determined by comparing the data of Keuper et al. (1988), Weidemann (1993) and Jatho et al. (1994) and the temporal parameters in relation to the temperature change recorded in P. macedonicus.Weidemann (1993) records a 90-120 ms length of the main syllable element of P. illyricus at about 24°C, which we find the most reliable measurement.Furthermore, some differences in the syllables repetition rate were recorded (see Fig. 8, Table 2).

Molecular data and phylogeny
After alignment and trimming, the final length of CoI sequences was 1111 bp. of the 1111 aligned sites, 873 sites were constant, 238 were variable, and 178 were parsimony informative.We used 73 ingroup and 4 outgroup sequences, that represented 40 different haplotypes. of those 15 ingroup and 4 outgroup haplotypes were used in the phylogenetic analyses.As early analyses indicated for haplotypes of P. caucasicus clustering within the haplotypes of the Balkan populations, one haplotype per P. specularis and P. venosus was used as the outgroup in later analyses and MP, ML and BI analyses with this setting are presented.jModeltest suggested the Tamura-Nei model based on the AIC value.The MP, ML and BI analyses resulted in similar tree topologies (Fig. 9).Most of the nodes were constant in all trees with high bootstrap (MP and ML) or posterior probabilities (BI) supporting the following results: ( 1

discussion and taxonomic conclusions
So far, the Balkan populations of Psorodonotus have been treated in various systematic studies, either as three distinct species, three subspecies of two species, or subspecies of a single variable species (Ramme, 1931(Ramme, , 1951;;Mikšić, 1965;Harz, 1969;Heller, 1988).new recently accumulated data indicate a better view of speciation within Psorodonotus, which in allopatric taxa have resulted in weaker morphological and acoustic differentiation, though genetic differences frequently point to early lineage separation (kaya et al., in prep.).on the other hand, sympatric/ parapatric taxa are clearly distinct (Kaya et al., 2014).Present data revealed good morphological distinction between P. fieberi, P. macedonicus and P. illyricus, similar to or even bigger than that recorded within well separated lineages of the P. venosus (Kaya et al., 2013) and P. specularis (Kaya et al., 2014) species groups.Although the differences in song between the Balkan taxa are weaker (this paper) nevertheless they are within the interspecific range of distinction of the other lineages (see last citations).This is not surprising as divergence in song is usually slower in the case of allopatry (Heller, 2006;Çıplak et al., 2009).Most important support for the distinctiveness of the Balkan species comes from molecular data as all of the phylogenetic analyses confirmed the species are monophyletic.In addition, pairwise genetic divergence between species is high, about or above the mean for closely related allo-/parapatric orthopterans (e.g., Allegrucci et al., 2005;Shapiro et al., 2006) and even about the mean between-families distances recorded for some insect orders (Hebert et al., 2003).our data indicated that the significant genetic distances between selected taxa were largest between P. macedonicus and P. illyricus.Interestingly, in all trees P. caucasicus is a sister group of P. macedonicus.
our results do not support the opinion of Mikšić (1965) and Ingrisch & Pavićević (2012) that the populations in Montenegro are intermediate forms between P. illyricus and P. macedonicus.The specimens from Eastern Montenegro (bjelasica Mt. S of Tara River) identified here as P. macedonicus grouped within the latter together with specimens from Western Macedonia and northern Greece, while specimens from Northern Montenegro (N of Tara River) fit P. illyricus and group within it in the Cv analysis.According to Mikšić (1965) the lengths of the cerci of specimens from Montenegro differ and the titillator shape is intermediate.Though the apical parts of cerci were indeed longer in specimens from Northern Montenegro and Southern bosnia and Hercegovina (see Fig. 5C2, C3), the shape of their bases and the cercal spine clearly put these Fig. 6.Morphology of female ovipositor (A1, A2, B1, B2, C1, C2) and subgenital plate (A3, B3, B4, C3-C5) of Balkan Psorodonotus: A -P. fieberi (Bulgaria, Belassitsa Mt.); B -P. macedonicus (b1 -Republic of Macedonia, bistra Mt; b2 -Greece, Metsovon); C -P. illyricus (Croatia, n.P. Sjeverni velebit; note: in C5 the base of plate is covered by the seventh sternite).Scale = 5 mm.animals within P. illyricus.our small sample from this area cannot rule out the absence of intermediate forms.However all specimens studied could be related to a respective taxonomic unit.In terms of the male titillators, the specimens of P. illyricus and P. macedonicus in general differed with shorter and wider apical titillator arms in P. illyricus and more prominent with longer and stouter basal arms in P. illyricus (compare Fig. 7B with C).The gracile basal arms of the titillator of a specimen of P. illyricus from Bosnia and Herzegovina (Fig. 7C2; note the left apical arm is distorted) may be due to the weak sclerotization.Its geographic origin between Montenegro and Croatia indicates rather intraspecific variation than intermediate (i.e., hybrid) populations as stated by Mikšić (1965) and Ingrisch & Pavićević (2012) for specimens from Montenegro.In addition, differences in the shapes of the titillators of P. illyricus and P. macedonicus are similar to those among the latter taxa and other representatives of the caucasicus species group (P.ebneri, P. caucasicus) or even representatives of other species groups (e.g., P. venosus, P. specularis, etc.) (compare Ramme, 1951;Kaya et al., 2013Kaya et al., , 2014)).on the other hand, the shape of the titillator of P. fieberi (Fig. 7A) clearly distinguished this species.
Both the geometric morphometrics and molecular phylogeny presented in this study confirm the species status of the three taxa on the Balkan Peninsula: P. fieberi, P. illyricus and P. macedonicus.our results also indicate an interesting phylogeographic pattern in the P. caucasicus species group with P. caucasicus placed within the Balkan lineage.However, in order to reveal the evolutionary history of the latter a comprehensive phylogeographic study based on mitochondrial and nuclear DNA phylogenetic analyses (Kaya et al., in prep.) and ecological niche modelling is needed.

General description and diagnosis
Males usually brownish (brownish-violet), females mostly green (rarely females also brown) (Fig. 2A1, 2).Male pronotal (Fig. 4A1-3) disc is strongly rugose (with distinct dense dents), less widened in the metazona (but some exceptions are present).The metazona is short with the transverse furrow behind the middle of the pronotum (rarely at its middle) and the hind pronotal sulcus reaching the median keel slightly before the middle of the pronotum.The female pronotal (Fig. 4A4-6) disc also usually with short and narrow metazona (though may be similar to that of male), less rugose than in male but more than in other species.Male tegmina (Fig. 5A) usually reach the third tergite.Male cerci (Fig. 5A2, 3) longer than cerci of P. illyricus and shorter than those of P. macedonicus, apical part (from the tooth to the tip) over 2 times and up to 3.5 times longer than the basal part (from base to the tooth); base of cerci 3-4 times wider than the apex; cercal tooth similar in length but usually wider than that of P. macedonicus.Titillator (Fig. 7A) wider apically with strong teeth distributed in two or more rows; basal arms short, wide medially and narrowing laterally.Female subgenital plate (Fig. 6A3) elongated with a moderately deep incision being up to onethird of the length of the plate and forming narrow apical lobes.ovipositor (Fig. 6A1, 2) variable in curvature and length (23.5-32 mm;Harz, 1969) but is usually stouter than in other taxa, wide at its base and gently narrowing towards the tip.Average length of ovipositor is the shortest among these three taxa (Table 3).Hind femora in the specimens studied contrary to the statement by Mikšić (1965) did not differ significantly in length (but mean values lower than in other taxa) or in the ratio length: width from those of the other two taxa (Table 3; see also Harz 1969).In the West the known range of the species, mostly according to Mikšić (1965), borders the ranges of P. macedonicus and P. illyricus, approximately following the valleys of the vapa, lim, Drina, Ibar (Ibër), Sitnicë, Pchinja and vardar Rivers, but it is not known where this border connects with the Pchinja and Sitnicë valleys.The exact border of its range, e.g., the locality "Tara Mt." (Grebenshchikov, 1950;Mikšić, 1965) is not well defined.
This species inhabits open grass-herbaceous plant associations, usually with Juniperus communis alpina, the bushes of which it uses for protection.It occurs within the altitudinal range 1100-2300/2400 m alt.(usually between 1500 and 2200 m).

General description and diagnosis
Male usually brownish (brownish-violet), females mostly green (rarely are females also brown) (Fig. 2C1, 2).Male pronotal (Fig. 4B1-3) disc is less rugose than in P. fieberi, but similar to that of P. illyricus; metazona is strongly widened or sometimes elongated, shovel-like.The metazona is long with a transverse furrow (sulcus) before the middle of the pronotum (rarely at its middle).Female pronotal (Fig. 4B4-6) disc usually with a longer metazona than in P. fieberi and a less rugose surface; however, distinction by the female pronotum may be difficult.Male tegmina (Fig. 5B1, 2) similar in shape to those of P. fieberi or slightly elongated and reaching the third tergite.Male cerci (Fig. 5B3, 4) much longer than those of P. illyricus and longer or about the same length as those of P. fieberi but the basal part is very short, between over 5 to 6 times shorter than the apical part; base of cerci 2-3 times wider than the apical part; cercal tooth similar in length but narrower than that of P. fieberi and shorter than that of P. illyricus.Titillator (Fig. 7B) narrower apically bearing one row of teeth (rarely additional small teeth present on its dorsal surface); its basal part similar in length to that of P. fieberi but usually thinner, especially medially.Female subgenital plate (Fig. 6B3, 4) is shorter than that of P. fieberi, with wider apical lobes and incision reaching the middle of the sclerite.ovipositor (Fig. 6b1, 2) varies in curvature and length (23.8-27 mm;Harz, 1969) but on average is longer than that of P. fieberi (Table 3); it is usually more slender than that of P. fieberi and generally slightly stouter or similar to that of P. illyricus.Hind femora in the specimens studied slightly longer than in P. fieberi (mean values) and very similar in length to those of P. illyricus, but with greater ratio length: width (Table 3).
Song (Fig. 8B) consists of sparsely repeated syllables with a slightly higher repetition rate than recorded for P. fieberi (0.4-1 Hz at 23-28°C).The main element starts with a compact impulse series and usually ends (especially at low temperatures) with a sparse sequence of impulses, the last of which resemble after-clicks.

Distribution
Fig. 10, Appendix 2. occurs throughout the Pindos Mountain Range, mountains of the Pelagonia horst-anticlinorium and southern part of the Dinaric Mountain Range (Greece except Peloponnese, Republic of Macedonia, Albania, W Kosovo, S Montenegro and possibly partly Serbia).The borders with the ranges of P. fieberi and especially P. illyricus are not clear.The border with the latter possibly passes along the Drina -Tara River valleys, where contact between populations or a hybrid zone is suspected.The locality in Skopska Crna Gora Mt. (n R Macedonia) belongs to a geographically intermediate region between the Pindo-Dinaric Range and Macedonian-Thracian Massif, but as the specimens from there have not been included in a special study (e.g., by Mikšić, 1965), their taxonomic affinity needs to be confirmed.
This species inhabits open grass-herbaceous plant associations, usually with higher humidity than those inhabited by P. fieberi and is even recorded from peat bogs.This species is recorded within the altitudinal range (?1200) 1400-2500 (?2600) m alt., where it is common between 1500 and 2200 m.

General description and diagnosis
Either a greenish or brownish (brownish-violet) colour seems to be common in this taxon, while in different populations one of these colours may dominate (Fig. 2C1, 2).Male pronotal (Fig. 4C1-3) disc is the least rugose of the three taxa with the metazona sometimes almost smooth.The metazona is notably wider and longer than the prozona if compared to P. fieberi, the transverse furrow (if visible) is located before the middle of the pronotum, and the hind pronotal sulcus reaches the median keel well before the middle of the pronotum.Female pronotal (Fig. 4C4-6) disc with notably elongated metazona, very similar to that of P. macedonicus.Male tegmina (Fig. 5C1) wider and longer than in the other two taxa and usually covering the third tergite.Male cerci (Fig. 5C2-4) much shorter than in the other two species with a very wide basal part and a long tooth; basal part 2-2.5 (rarely up to 3.5) times shorter than the apical part; base of cerci 3-7 times wider than the apical part.Titillator (Fig. 7C) similar to that of P. macedonicus apically but the arms are usually shorter and wider; its basal arms very wide, much stouter than those of P. macedonicus (with some exceptions, e.g., see Fig. 7C2), and very long (longer than in the other taxa).Female subgenital plate (Fig. 6C3-5) resembles that of P. macedonicus but the excision may be wider and deeper, exceeding the middle of this sclerite.ovipositor (Fig. 6C1, 2) varies in curvature and length (26-31 mm;Harz, 1969) but is usually more gracile than in P. fieberi, elongated and less wide at its base, and on average the longest of the three taxa (Table 3).Hind femora in the specimens studied the longest of the three species, with lowest ratio length: width, but the values overlap with the other taxa (Table 3).
Song (Fig. 8C) consists of densely repeated syllables (syllables repetition rate highest for these three taxa) (1-3 Hz at 23-28°C).The main element is the longest of the taxa studied at a particular temperature and lacks afterclicks (though at low temperatures the last impulses may be separated by longer intervals as in P. macedonicus).

Distribution
Fig. 10, Appendix 2. occurs along the Dinaric Mountain Range from Istra Peninsula in S Slovenia and NW Croatia, through bosnia and Herzegovina to the central regions of Montenegro.It is possibly isolated from populations of P. macedonicus to the South/South-east by the Drina and Tara River valleys, where it is suspected populations come into contact and there may be a hybrid zone.Ebner (1923)  records a locality for P. illyricus in Bulgaria ("Bulgarien") but this may be the result of wrong label.
This species inhabits open grass/herbaceous plant associations in the colline to subalpine mountain zones.It seems less sensitive to drought and high temperature than the other two taxa.This species is recorded at altitudes between 800 and 1800 m. appenDiX 2. Locality list of the Balkan Psorodonotus from literature and unpublished data.
In the following list we present all published and unpublished localities of the Balkan Psorodonotus known to us.Data are arranged into respective taxonomic units after critical review.Doubtful records are shortly commented or related to "Incertae sedis".The localities that were traceable are geographically ref-erenced using decimal coordinates according to appropriate location and habitat and mapped (Fig. 10 in the text).The localities are arranged mostly in a north-south direction.For published data usually only original references (i.e.such reporting unpublished before information or treating the taxon in a different sense) are used given in brackets.
Song (Fig. 8A) consists of sparsely repeated syllables (0.3-0.6 Hz at 23-28°C) that contain a compact main element and well detached after-clicks.Distribution Fig. 10, Appendix 2. P. fieberi occurs in the Balkan (Stara Planina) Mountains, Macedonian-Thracian Massif (Rhodopean Mts s.l.) and the eastern arc of the Dinaric Mountain Range (Bulgaria, E Serbia, E Republic of Macedonia, ne Greece).The border of its range in the north is the Danube River valley, in the East -the lower parts of the Stara Planina Mts and Rhodope Mts, and in the South -the southern slopes of the Macedonian-Thracian Massif.

Fig. 9 .
Fig. 9. Phylogenetic tree based on the Bayesian analysis of 15 haplotypes of Balkan and four haplotypes of Anatolian Psorodonotus.MP/ML bootstrap proportions (values > 50) and BI internode posterior probabilities (values > 0.50) are shown on BIresolved internodes.

taBle 1 .
Recordings of the male song of Psorodonotus used in this study.

taBle 3 .
Measurements recorded for some of the morphological structures of Balkan Psorodonotus.Measurements are given in mm: first row -Min-Max values, in brackets -Average ± Standard deviation and number.
Question marks indicate uncertainty about the respective information.appenDiX1. Material of Psorodonotus used for geometric morphometrics and/or molecular phylogenetic analysis.