We used sequences of three mitochondrial (Cytochrome c oxidase subunit I (COI), NADH dehydrogenase subunit 2 (ND2), and a segment of 16S rDNA + tRNAval + 12S rDNA (VAL)) and one nuclear (Internal transcribed spacer 1 and 2 with 5.8S rDNA in between (ITS)) gene segments to obtain a robust and dated phylogeny along with automated species delimitation tests. We also downloaded sequences of these gene segments available from GenBank (Kaya 2018). The data matrix comprises sequences from 20 populations representing all putative species/subspecies in the species group, except P. varicornis from Lebanon, as it is currently known only from type material (Kaya 2018; Sevgili et al. 2018; Cigliano et al. 2024).

The specimens of the P. zonatus species group stored at MEVBIL (Molecular Evolution and Biogeography Laboratory, Department of Biology, Akdeniz University, Antalya, Turkey) in 98% ethanol at –20°C were used to obtain DNA sequences. The muscle tissue from the femur of the hind leg was used to isolate the total DNA with a ThermoFisher Scientific Inc., following the manufacturer’s protocol. The amount of dsDNA was measured by Qubit 4 Fluorometer (ThermoFisher Scientific Inc.) and used for the PCR reactions. The following primer couples were used: (1) the forward 5’ GGRGGATTTGGAAATTGACTW-GTTCC 3’ and the reverse 5’ TCCAATGCA­CTAATCTGCCATATTA 3’ for a >1200 bp segment of COI (Simon et al. 1994), (2) the forward 5’-AATT­A­AGCTAATGGGTTCATACCC-3’ and the reverse 5’-G­G­CTGAGCARTAGGCGATAAACT GTAAA-3’ for a >1000 bp segment of ND2 (Simon et al. 2006); (3) the forward 5’ ATGTTTTTGTTAAACAGGCG 3’ and the reverse 5’ AAGGTGGATTTGATAGTAAT 3’ for a >800 bp segment of VAL (Simon et al. 1994), and (4) the forward 5’ GGAAGTAAAA GTCGTAACAAGG 3’and the reverse 5’-TCCTCCGCTTATTGATATGC 3’ for a >900 bp segment of ITS1-5.8S-ITS2 (White et al. 1990). Sanger sequencing was executed via Macrogen Europe (Macrogen Inc.) using 23ABI3730XL DNA analyser after the purification of the amplicons. SEQUENCHER v.4.1.4 (Gene-Codes Corp.) was used to optimize the forward and reverse sequences of each gene segment and to obtain consensus sequences per specimen. Sequences newly produced and downloaded from GenBank (Kaya 2018) were used to create the matrix using MEGA v.7 (Kumar et al. 2016). In addition to sequences belonging to the P. zonatus species group, three sequences were added to the matrix as outgroups (Poecilimon inflatus, Poecilimon luschani, Isophya major). Multiple alignments were attained by MAFFT v.7 online version (https://mafft.cbrc.jp/alignment/server) using appropriate settings for each gene. All sequences were checked for nuclear mitochondrial copies (NUMTs) according to parameters defined by Kaya and Çıplak (2018).

Phylogenetic relationships were estimated via maximum likelihood (ML) and Bayesian (BI) phylogeny inferences applied to the mitochondrial dataset comprising concatenated sequences of COI+ND2+VAL. The phylogenetic signals in the nuclear ITS dataset were assessed by establishing a haplotype network, due to low sequence variation. Unique haplotypes per gene segment of COI, ND2, VAL and ITS were detected using the online tool FABOX (Villesen 2007) and were used for phylogenetic analyses and network establishment. The ML analysis was executed via IQTREE v.2.2. (Minh et al. 2020) with ultrafast bootstrapping (Hoang et al. 2018) and 1000 replicates. The substitution models used in the analysis were estimated using the implemented algorithm ModelFinder (Kalyaanamoorthy et al. 2017) with site partitions for the protein-coding genes COI and ND2 (Chernomor et al. 2016). The BI analysis was performed with MrBayes v.3.2.2 (Ronquist et al. 2012) by applying four chains, 500000 generations choosing the sampling frequency as 1000 and initial burn-in frequency as 0.25. The substitution model implemented in BI analysis was calculated by PARTITIONFINDER v.2.1.1 (Lanfear et al. 2017; also uses PhyML; Guindon et al. 2010) with the greedy algorithm (Lanfear et al. 2012). The haplotype network was established by PopART v 1.7 (Leigh and Bryant 2015) with the median-joining option (Bandelt et al. 1999).

The divergence times within the group were estimated by conducting a BEAST v.2.6.7 (Bouckaert et al. 2019) analysis using the mitochondrial dataset. BEAST was run with the following settings; site model parameters uniformly linked across all genes, the General Time Reversible (GTR) sequence evolution model, Yule tree model and the main nodes constrained according to ML/BI trees. The substitution rates reported in Çıplak et al. (2022) for the Tettigoniidae family were used to calibrate the molecular clock; 0.0187, 0.018 and 0.0141 substitutions/sites/million years (s/s/myr) for COI, ND2 and VAL, respectively. The analysis was performed with 50000000 generations, sampling each 5000 trees and initial burn-in frequency as 0.25. The majority rule chronogram was obtained using the TREEANNOTATOR implemented in BEAST. The effective sample size (ESS) parameters for both BEAST and MRBAYES analyses were assessed using TRACER v.1.7 (Rambaut et al. 2018), and all final trees were visualized using FIGTREE v.1.4.4 (http://tree.bio.ed.ac.uk/software/figtree).

For a comprehensive exploration of the historical biogeography of the species group, we utilized RASP (Reconstruct Ancestral State in Phylogenies) (Yu et al. 2020) to infer ancestral area. The analysis was accomplished using the time-calibrated tree obtained by excluding outgroups and distantly related taxa P. azizsancar and P. tauricola. Sequences were categorized into geographical regions, including Caucasus (A), East Anatolia (B), Mid Taurus (C), West Taurus High Altitudes (D), West Taurus Low Altitudes (E), and Balkans (F) based on the taxa’s distribution (Figs 1, 3). Range combinations were constrained for F with only A, B, and C, based on their phylogenetic relationships. For time stratification, 3 dispersal events between B–C and 10 between D–E were allowed. Biogeographical model comparisons were conducted using the BioGeoBEARS package (Matzke, 2013a, 2013b, 2014; Massana et al. 2015) implemented in R environment (R Core Team 2020). The best-explaining model was determined by the highest weighted Akaike Information Criterion (AIC) value.

Figure 1. 

Ranges of the members of the P. zonatus species group. Ranges shown by shaded color-coded areas for each species with the number of sampling locality. Sampled localities are marked by color-coded circles; for additional details about sampled localities in each species see Table 1. The question mark (?) shows the uncertain distribution areas of the P. zonatus.

To assess the species diversity within the P. zonatus group, three distinct delimitation tests were performed using the concatenated mitochondrial dataset. Initially, the Assemble Species by Automatic Partitioning (ASAP) method (Puillandre et al. 2021), which utilizes hierarchical clustering, was performed through a web-based tool (https://bioinfo.mnhn.fr/abi/public/asap). Later, using the ML tree, the Bayesian Poisson tree processes (bPTP) (Zhang et al. 2013) was employed with 100000 Markov Chain Monte Carlo (MCMC) generations via online server https://species.h-its.org. Lastly, using an ultrametric and bifurcating time-calibrated tree obtained by BEAST, the Generalized Mixed Yule Coalescent Approach (GMYC) was executed, selecting single and multiple thresholds separately through the online tool https://species.h-its.org/gmyc. To determine the optimal threshold model, we compared the maximum likelihood value of the selected models with the likelihood value of the null model, selecting the one closest to the null model. To identify apomorphic and/or synapomorphic characters (mutations) as indicative of putative species, we utilized a parsimony analysis using PAUP 4.0b10 (Swofford, 2002) with the ‘pset opt = DELTRAN’ optimization model. The analysis was performed with the commands “describetrees 1/ plot = phylogram root = outgroup labelnode = yes apolist = yes;” and the ML tree served as the guide for the analysis. Differentiation between the species suggested by species delimitation tests was assessed by calculating pairwise differences for each of COI, ND2, VAL, and ITS datasets, using MEGA v.7. (Kumar et al. 2016). Additionally, intra-group genetic clusters within the species group excluding distantly related species P. tauricola and P. azizsancar were delineated through the application of Discriminant Analysis of Principal Components (DAPC) (Jombart et al. 2010) using the ‘adegenet’ (Jombart 2008; Jombart and Ahmed 2011) and ‘seqinr’ (Charif and Lobry 2007) packages in the R environment. The determina­tion of the optimal number of genetic clusters was accomplished through the Bayesian Information Criterion (BIC), with the optimal number of principal components estimated using a-score optimization-spline interpolation (optim.a.score). The resulting clusters and their respective membership probabilities were visually represented through the ‘scatter’ function. The DAPC analysis was conducted for each of COI, ND2, VAL, and ITS datasets separately, including all sequences.

Morphological structures were photographed, qualitatively examined, and measured using a digital camera attached to a Leica MZ6/DC600 stereomicroscope and ImageJ v.1.36 (http://rsb.info.nih.gov/ij). The structures frequently referred to in previous studies were examined qualitatively and illustrated when necessary (further rich illustrative material can be found in Kaya (2018) and Sevgili et al. (2018)). We concentrated mainly on diagnostic characteristics in both sexes (cercus + subgenital plate in males and anal tergum + cercus + subgenital plate + ovipositor in females).

The male calling songs were recorded in the laboratory with animals caged separately using the following digital recording equipment: Fostex FR-2 (sampling at 192 kHz, 24-bit mode), equipped with a G.R.A.S. 1/4” 40BF microphone connected to a 1/4–1/2” AB Preamplifier and 12AK Power Module; Tascam DR-680MKII (sampling at 192 kHz, 24 bit) equipped with a Pettersson D500 microphone. Sound analyses were done using the software Audacity 3.3.2 (https://www.audacityteam.org) on a PC. As a considerable amount of song data is given in Sevgili et al. (2018) and Kaya (2018), here songs of the species/population not, or insufficiently presented in these publications, were discussed.

Song terminology: Calling song: song produced by an isolated male. Syllable: the sound produced by one complete up (opening) and down (closing) stroke of the tegmina (reduced forewings). Complex syllables: the sound produced by a group of complete and/or partial up- and down-strokes of the tegmina, i.e., a first-order assemblage of syllables and clicks. Syllable period: time period measured from the first impulse of a simple or a complex syllable to the first impulse of the next syllable (reciprocal value: syllable repetition rate SRR). After-clicks: single or grouped impulses that follow the main syllable. Impulse: a simple, undivided, transient train of sound waves (here: the highly dampened sound impulse arising as the impact of one tooth of a stridulatory file).

In total, sequences were obtained from 20 localities, encompassing eight putative morphospecies (Fig. 1; ­Table 1) for COI, ND2, VAL, and ITS (for detailed information about sequences and localities see Tables 1, 2). The dataset comprised a total of 3059 base pairs for all fragments, with the identification of 1258 polymorphic sites. No haplotype sharing between species/subspecies is found in mitochondrial genes. However, in nuclear ITS data, haplotype sharing was observed between the populations from Ziyarettepe, Uzunkarıştepe (belonging to P. denizliensis kizildagi subsp. nov. and P. datca montana subsp. nov., respectively) and Fethiye (P. ciplaki) (see populations 5, 6 and 8 in Table 1).

The substitution models selected for ML analysis were as follows: TIM3e+G4 (COI^1st), F81+F+I (COI^2nd), TN+F+G4 (COI^3rd), HKY+F+G4 (ND2^1st&2nd), TIM2+F+G4 (ND2^3rd) and HKY+F+I+G4 (VAL). Meanwhile, the substitution models chosen for BI analysis were SYM+G (COI^1st), F81 + I (COI^2nd), GTR+I+G (COI^3rd& ND2^3rd), HKY+G (ND2^1st&2nd) and HKY+I+G (VAL).

The ML and BI analyses applied to the mitochondrial data matrix, which included 44 ingroup + 3 outgroup concatenated sequences (COI+ND2+VAL), produced topologically similar trees (Figs 2 and 3 [right panel]). Monophyly of the P. (s. str.) zonatus group was confirmed by 95/0.97 bootstrap (BS)/posterior probability (PP) supports and the ancestral node dated to 6.64 myr ago by the BEAST analysis. The P. (s. str.) zonatus group consists of two main clades, Clade I includes P. tauricola + P. azizsancar, and Clade II, all others. The ancestral age of Clade I is dated to 2.54 myr ago and that of Clade II to 3.95 myr ago. The node of the Clade II is dichotomic consisting of two subclades named as Clade IIA and Clade IIB. The Clade IIA comprises 14 sequences from South-western Anatolian populations with an ancestral age of 1.81 myr and Clade IIB – the remaining 21 sequences from the Balkans, North + East Anatolia, and Southern Mediterranean Taurus with an ancestral age of 3.23 myr. The Clade IIA consists of five infraclades corresponding to distinct species at least by one species delimitation tests. The nodal ages for the clades proposed as species are 0.7 myr or earlier, corresponding to the end of the Mid Pleistocene Transition. Clade IIB includes six infraclades each of which is suggested as a distinct species by the delimitation tests. Splitting of the species level clades fall into the Mid Pleistocene Transition around 1.1–0.7 myr ago. All of the internal nodes from the basal node to each candidate species’ node in the species group received >90/0.90 BS/PP supports.

Figure 2. 

The phylogenetic tree obtained by application of maximum likelihood (ML) and Bayesian inference (BI) analyses (Node values: bootstrap (BS)/posterior probability (PP)) to the data matrix comprising 44 ingroup + 3 outgroup concatenated sequences of COI+ND2+VAL. Haplotype names established by species/subspecies names and locality (see Table 1 for details). The bar on the right side marks hypothesized species boundaries based on the ASAP, bPTP and GMYC species delimitations tests (for details see text) and the taxonomical composition proposed in this study.

The clusters suggested by the haplotype network produced from 86 sequences of ITS (Fig. S1) are largely coupling with the phylogenetic units of the mitochondrial tree. Each of P. tauricola, P. azizsancar, P. variicercis, P. vodnensis, P. anisozonatus sp. nov., and P. datca datca occurred as distinct clusters. Sequences from P. zonatus, P. parazonatus sp. nov., and P. isozonatus constitute a separate cluster while those from P. boncukdagensis sp. nov., P. denizliensis denizliensis stat. nov., and P. denizliensis kizildagi subsp. nov. another one. Two sequences of P. ciplaki occur within two different clusters (Fig. 3).

Figure 3. 

Right panel The chronogram showing radiation time of P. zonatus group obtained by a BEAST analysis applied to 44 ingroup + 3 outgroup concatenated sequences of COI+ND2+VAL; (Node values: values shown above the branches indicate time (myr), while values below the branches indicate HPD intervals.) The haplotype names are according to Table 1 and Figure 2. Left panel Ancestral area reconstruction (most likely state) of the P. zonatus group based on the DEC model. The coloured circles at the tree nodes correspond to the areas ancestral for the clade. Colours correspond to generalized geographic areas mapped in the upper left corner (see the text for details)

Model evaluation of RASP suggested the DEC model to best fit the data of the P. zonatus subgroup (AIC = 43.31). The DEC model (Fig. 3–left panel) suggested the following modes of the speciation: (i) The total range of the group as the most likely area of origin of the ancestral stock of Clade II and a vicariant splitting between Clade IIA and IIB, (ii) an autochthonous radiation in southwest Anatolia with two habitat shifts, first from highlands to lowlands and the second secondarily returning to highlands, (iii) dispersal from Anatolia (Southern Taurus + East Anatolia + Caucasus) to Balkans to found P. vodnensis, (iv) autochthonous radiation both in Southern Taurus (speciation of P. isozonatus and P. anisozonatus sp. nov.) and in East Anatolia (speciation of P. zonatus and P. parazonatus sp. nov.), and (v) extinction in the Caucasus.

Delimitation tests ASAP, GMYC, and bPTP suggested 12, 13, and 16 candidate species, respectively, in the P. zonatus group (Fig. 2). Consistent with the existing taxonomical schema, the species status of P. tauricola and P. azizsancar were verified. The taxonomical schema suggested by all or at least two delimitation tests for the Clade IIA is as follows; (i) P. ciplaki ciplaki and P. salmani constitute a single species (as P. ciplaki, the other as synonym; see below for details), (ii) each of P. ciplaki denizliensis and P. zonatus datca occur as distinct species (P. denizliensis stat. nov. and P. datca stat. nov.; see below), (iii) the population from Tuzlabeli Pass, Boncuk Mts. in Muğla Province of Turkey (occur as sister group to P. ciplaki + P. datca) was suggested as a separate species by bPTP and GMYC (proposed as P. boncukdagensis sp. nov.; see below), (iv) the population from Ziyarettepe (belongs to Kızıldağ Mts.) in Antalya Province occurs as sister group to P. denizliensis and was suggested as distinct species by ASAP and GMYC, but not by bPTP (proposed as P. denizliensis kizildagi; see below), and (v) the samples from each of the Uzunkarıştepe Mt. and the Bakırlıdağ Mt. belonging to the Beydağları Mts. in Antalya Province were suggested as separate species by bPTP, but in P. datca by ASAP and GMYC (regarding their allopatry and phenotypic differences with the nominate species P. datca, proposed as P. datca montana ­subsp. nov.; see below). The taxonomical schema suggested by delimitation tests for the Clade IIB is more consistent and the species status of P. variicercis and P. vodnensis were verified by all delimitation tests. The Munzur Mts. population from Erzincan-Tunceli identified as P. zonatus zonatus (see Sevgili et al. 2018) occurred as a distinct species with relationships of P. variicercis + (P. zonatus + P. parazonatus) (for taxonomy see below). P. isozonatus Kaya was split into two distinct, yet closely related, sister taxa (proposed as P. isozonatus and P. anisozonatus; see below for details).

The DAPC analysis applied to the single gene datasets of COI, ND2, VAL and ITS comprising sequences of the P. zonatus subgroup (or Clade II in the phylogenetic tree) suggested 8 (the first two datasets) and 9 (the last two datasets) clusters, with a BIC value of 112.827972, 145.11361, 60.460393 and 118.3611 respectively (Fig. 4). P. parazonatus, P. ciplaki, and P. vodnensis are not represented in COI, P. datca montana in ND2 and VAL, and P. denizliensis kizildagi in COI, ND2 and ITS. The taxon contents of the clusters of different datasets are partly inconsistent. Each of the COI-DAPC clusters corresponds to a putative/candidate taxon (species/subspecies) with exception of the splitting of P. isozonatus into two independent units as C4 (the sequences from Karaman Province) and C5 (the sequences from Konya Province). Most of the ND2-DAPC clusters also correspond to putative/candidate taxa with two exceptions; the C2 consists of P. ciplaki + P. datca datca and the C5 of P. denizliensis denizliensis + P. boncukdagensis, both as close clusters. Of the 9 clusters suggested by VAL-DAPC analysis, six correspond to putative/candidate taxa while three do not; C3 includes P. boncukdagensis + P. denizliensis kizildagi, C6 P. zonatus + P. parazonatus and C9 P. ciplaki + P. ­datca datca. Although some sequences of P. ciplaki clustered with P. datca datca, the others occurred as an independent cluster. The ITS-DAPC analysis resulted in nine clusters, six of which correspond to single taxa while three clusters to two or more; C4 includes P. ciplaki + P. datca montana + P. denizliensis, C6 P. zonatus + P. isozonatus (some sequences of P. zonatus constitute a separate cluster C8) and C7 P. boncukdagensis + P. datca datca. Although each matrix lacks some taxa, of the putative/candidate species/subspecies (see taxonomy section below) P. variicercis, P. vodnensis and P. anisozonatus were identified by all datasets, P. zonatus by COI and ND2, P. parazonatus by ND2 and ITS, P. isozonatus by VAL, P. ciplaki by ITS and VAL, and P. datca datca and P. datca montana by COI. The case for P. boncukdagensis is significantly contradictory as none of the DAPC suggested it as a unique cluster and it occurs with P. denizliensis kizildagi in VAL-DAPC, with P. denizliensis denizliensis in ND2-DAPC and with P. datca datca in ITS.

Figure 4. 

The discrete genetic clusters identified by DAPC (determined by the highest BIC values) analyses applied to the single gene datasets of COI, ND2, VAL and ITS comprising sequences of the P. zonatus subgroup (or Clade II), with the different clusters shown as different colors along with the membership of each genetic cluster based on the species codes (DAM: P. da. montana, DAD: P. da. datca, DED: P. de. denizliensis, DEK: P. de. kizildagi, IZ: P. isozonatus (Krm: Karaman, Kon: Konya), ZO: P. zonatus, AZ: P. anisozonatus, VA: P. variicercis, CI: P. ciplaki, VO: P. vodnensis, BO: P. boncukdagensis, PZ: P. parazonatus).

Genetic divergences of the species/subspecies suggested by delimitations tests were estimated by calculating pairwise distances for each of four single gene matrices. The pairwise distance patterns obtained from these four genes were also inconsistent (Table S2). The pDist values obtained from sequences of COI are 0,009<0,134 for 62 out of 66 pairs, ND2 0,012<0,21 for 86 out of 91 pairs, VAL 0,004<0,175 for 84 out of 91 pairs, and ITS 0,001<0,076 for 81 out of 91 pairs (Table S2).

The three species delimitation tests applied to the data set including 44 mitochondrial sequences suggested 12–16 species within the P. zonatus group (Fig. 2) (samples of P. varicornis were not available for inclusion in the genetic data set). Genetic data confirmed the species status of two species in the P. tauricola subgroup and these two species are also distinguishable, both from each other and other members of the group. Similarly, the species identity of P. variicercis and P. vodnensis, each distributed in a discrete geographic fragment, were also warranted by genetic data. However, genetic data suggested a taxonomical rearrangement for the remaining species of the P. zonatus group. Giving priority to phylogenetic uniqueness and regarding diagnosability support from phenotypic features (morphology, song, or distribution as a reference of ecological features), the following taxonomical composition is proposed for the group. Phylogenetic units consistently suggested by three species delimitation tests are considered identical species. P. boncukdagensis is the only new species which identity was supported by two delimitation tests, while not by the third, possibly because of a limited number of unique sequences. This decision was made because of prominent phenotypic diagnosability, allopatric occurrence, and particularly because of it’s splitting from the closest clade before the Mid Pleistocene Transition. We proposed two polytypic species, regarding the results of species delimitation tests. Although each of them was suggested as two distinct species at least by two delimitation tests and they are somewhat diagnosable by phenotype, and even if they are allopatric with their sister subspecies in distribution (see below), we avoid defining them as species, because their splitting time from closest clade is relatively recent, after the Mid Pleistocene Transition and as the number of available sequences was limited. Thus, they were proposed as subspecies (P. datca montana and P. denizliensis kizildagi) to indicate their uniqueness. Although morphology suggests P. varicornis in the P. zonatus group, it was left out of the two species subgroups, as its samples were not available to include in phylogenetic studies. The new taxonomical content suggested for the P. zonatus species group is as follows (ordered according to branching order in the phylogenetic tree).

Poecilimon (s. str.) zonatus species group

1 P. varicornis (De Haan, 1843)

Subgroup P. tauricola (Clade I)

2 P. tauricola Ramme, 1951

3 P. azizsancar Sevgili, 2018

Subgroup P. zonatus (Clade II)

4 P. denizliensis Kaya, 2018 stat. nov.

P. denizliensis denizliensis Kaya, 2018

P. denizliensis kizildagi subsp. nov.

5 P. boncukdagensis sp. nov.

6 P. ciplaki Kaya, 2018

P. salmani Sevgili, 2018 syn. nov.

7 P. datca Sevgili, Sirin, Heller & Lemonnier-­ Da­r­cemont, 2018 stat. nov.

P. datca datca Sevgili, Sirin, Heller & Lemon­nier-Darcemont, 2018

P. datca montana subsp. nov.

8 P. vodnensis Karaman, 1958

9 P. variicercis Miram, 1838

10 P. parazonatus sp. nov.

11 P. zonatus Bolivar, 1899

12 P. anisozonatus sp. nov.

13 P. isozonatus Kaya, 2018

A huge amount of phenotypic data, morphological and acoustic, is presented in Sevgili et al. (2018) and Kaya (2018), but phenotype-based taxonomy given in these studies is largely rejected by DNA-based phylogeny for the majority of putative species, indicating that phenotype is not sufficiently reliable to define intragroup diversity (see Discussion section below). Thus, here we presented only a key to species for diagnosing the phylogenetic units or species suggested by delimitation tests. After examining the material available and the published data we found that males are distinguishable, though by minor differences in many taxa, while females are not. Among all, only three species are distinguishable by females; P. tauricola and P. azizsancar by the ratio of hind femur/ovipositor ≥2.1 (both species cannot be distinguished from each other), and the cercus extending much beyond the end of epiproct, P. varicornis by the relatively long ovipositor (ratio hind femur/ovipositor ≤1.72) and 2< >1.75 in others. Thus, the key to species is prepared only for males.

Figure 5. 

Male cercus. A1 P. varicornis redrawn from Sevgili et. al. 2018; A2 P. varicornis from OSF2 (Cigliano et. al. 2024); B P. azizsancar; C P. tauricola; D P. denizliensis denizliensis subsp. nov.; E P. denizliensis kizildagi subsp. nov.; F P. boncukdagensis sp. nov.; G P. ciplaki; H P. datca datca; I P. datca montana subsp. nov.; J P. vodnensis; K P. variicercis; L P. zonatus; M P. parazonatus sp. nov.; N P. isozonatus; O P. anisozonatus sp. nov. Terminology used in the key implemented in the Figures A and B.

Figure 6. 

Male subgenital plate. A P. varicornis redrawn from Sevgili et. al. (2018); B P. azizsancar; C P. tauricola; D P. denizliensis denizliensis subsp. nov.; E P. denizliensis kizildagi subsp. nov.; F P. boncukdagensis sp. nov.; G P. ciplaki; H P. datca datca; I P. datca montana subsp. nov.; J P. vodnensis; K P. variicercis; L P. zonatus; M P. parazonatus sp. nov.; N P. isozonatus; O P. anisozonatus sp. nov.. Terminology used in the key implemented in Figure O.

Along with morphology, we also examined the male calling song to seek diagnostic characters. A considerable amount of song data from the P. zonatus group has already been published by Sevgili et al. (2018) and Kaya (2018). The male calling songs are roughly similar across all members of the group in respect to features in song patterns, consisting of irregularly produced tick-like syllables encompassing different number and arrangement of impulses (Figs 10–14; see also Figs 30–33 in Sevgili et al. 2018; Fig. 3 in Kaya 2018). The syllable characteristics, such as syllable duration and impulse number and arrangement per syllable, as well as the production of after-clicks following the main syllable (P. variicercis, P. anisozonatus), or even a higher order of distinct syllables (P. azizsancar, P. ciplaki and P. vodnensis) are the characters to be used in diagnosing species (Table 3). Other temporal parameters, such as syllable period, syllable intervals, or duty cycle can also differentiate species though they may be highly variable depending on the age and physiological condition of the animals or other factors and thus may be misleading.

Currently, male calling songs of 11 species in the group were available for examination (song recordings of P. parazonatus sp. nov. and the Lebanese species, P. varicornis were unavailable). Data relating to the above-mentioned three characters (i- isolated impulses following the syllable, ii- syllable duration, and iii- the impulse number per syllable) are examined and presented together with previously published data in Table 3. However, from present and previous data, we concluded that the presence/absence of the isolated impulses are not good diagnostic characters because of intra-specific variation (Table 3). The same conclusion is valid for the syllable duration and the impulse number per syllable between some species because of intra-species/phylogenetic unit variation (e.g., in P. isozonatus). For example, P. isozonatus from Konya-Taşkent and Niğde-Çamardı, both belonging to the same genetic species, are different by the syllable duration and the impulse number per syllable (Table 3). On the other hand, there are differences between the new and previously published data for some species. For example, the syllable duration for the Muğla-Fethiye population (P. ciplaki) calculated here is 145.3 (119–175) ms due to the regular appearance of a second complex syllable group following the short first syllable, which was only measured and reported as 18 (15–19) ms, for P. salmani, by Sevgili et al. (2018) and 49 (42–72) ms, for P. ciplaki ciplaki, by Kaya (2018). These variations may either be natural due to species song production (e.g., depend on the age of the animal or its response to the acoustic environment) or may be due to technical reasons (properties of the recording equipment used and/or distance to the sound source) or personal interpretation while processed in different studies.

For instance, a statistical correction may be required regarding recording temperature, and without such a normalization, using these differences in syllable duration interpretation for the diagnosability of the song may be misleading. Additionally, members of the P. zonatus group are duetting animals, and, in such species, song-producing background may be more complicated (Kaya et al. 2018). Thus, elucidating song characteristics of P. zonatus group is a task to be done using inclusive data by standardized methodology.

Apart from the above-mentioned handicaps, song characteristics provide data to diagnose some of the phylogenetically unique clades. The new species, P. anisozonatus sp. nov. and P. isozonatus, are two sister species with the lowest pairwise genetic difference in the group, but both can be well distinguished by syllable duration of male song and even larger differentiation in the male songs is observed between the genetically sister species P. ciplaki and P. datca (see below). On the other hand, songs of the closely related taxa P. denizliensis (recordings from Honaz) and P. boncukdagensis sp. nov. from Tuzlabeli Pass are hardly distinguishable (Table 3).

3.8 . Poecilimon (Poecilimon) denizliensis Kaya, 2018, stat. nov.

https://orthoptera.speciesfile.org/otus/847297/overview

Poecilimon ciplaki denizliensis Kaya, 2018: 93.

Remarks.

Kaya (2018) described P. ciplaki as a new species, including two new subspecies, P. ciplaki ­ciplaki and P. ciplaki denizliensis. Although sequences from their type locality occurred within Clade IIA together with other sequences from Southwest Anatolia, Clade IIA was suggested as 3 to 6 distinct species by delimitation tests. Considering the new taxonomical composition suggested by species delimitation tests, and genetic and phenotypic differences between these phylogenetic units, P. ciplaki denizliensis Kaya, 2018 was elevated to species level as P. denizliensis stat. nov., leaving P. ciplaki as a monotypic species. The population from Honaz Mt., Denizli, was reported as the only type locality of this species in the original description by Kaya (2018). The sequence from Ziyarettepe belongs to Kızıldağ Mts. in the northwest corner of Antalya and constitutes a sister branch to those from the type locality of P. denizliensis. These two populations were suggested as species by bPTP while each as a separate distinct species by ASAP and GMYC, both sharing the last common ancestor 0.73 myr ago, just at the end of the Mid Pleistocene Transition. Considering differences between these two populations in male cercus (compare D and E in Fig. 5), absence of shared haplotypes and unique mutations per ancestral node of each, we proposed each as a separate subspecies – the nominate subspecies P. denizliensis denizliensis Kaya (see population 4 in Table 1) and the new subspecies Poecilimon denizliensis kizildagi Uluar, Chobanov & Çıplak, subsp. nov. (see population 5 in Table 1). The mutations at ancestral node of the species and each of the subspecies, detected in the concatenated matrix of COI+ND2+VAL by applying a PAUP analysis (File S1), are also proposed as further diagnostic characters for ach taxon: the positions 105 (C→T), 240 (A→T), 400 (T→C), 522 (A→G), 2409 (A→G), 2497 (A→T) are unique to P. denizliensis clade including both subspecies; 1208 (A→C, 1317 (G→A), 2044 (T →C) and 2113 (C→G) to P. denizliensis denizliensis, and 75 (T→C), 579 (A→C), 2147 (T→C) and 3027 (T→C) to P. denizliensis kizildagi subsp. nov.

Kaya (2018) provided rich illustrative material describing the morphology of this species (see E in Figs 5, 6, 9–12 in Kaya 2018), but provided no data for the song. The male song consists of one type of syllable. After-clicks were not observed. The syllable lasts 10–14 ms (average of 12 ms) and contains 15–18 impulses (average of 17). It starts with dense crescending impulses, which are followed by a few high-amplitude sparse impulses. The male song of P. denizliensis denizliensis is exemplified in Figure 7 and song measurements are provided in Table 3. The song resembles that of P. boncukdagensis to a great extent and based on this character it may be difficult to differentiate both taxa. From our recordings the song of P. denizliensis has a higher repetition rate and a syllable with higher number of impulses within the first crescending part (compare Figs 7 and 10, and Table 3). Yet, the latter characters may be due to individual characteristics and could be masked if more data is used.

Figure 7. 

Oscillographic representation of the male song of P. denizliensis stat. nov. – male from Honaz, 1972 m. Song shown at different speeds. A A frame of 1 min, recording at 27.1°C; B single syllable (frame of 100 ms; 27.1°C); C single syllable (frame of 100 ms; 26.7°C).

Material examined.

See populations 4 and 5 in Table 1.

3.9 . Poecilimon (Poecilimon) boncukdagensis Uluar, Chobanov & Çıplak, sp. nov.

https://zoobank.org/6D63A8C5-555A-4FD8-8B4B-ABE765451376

Description.

Holotype, male. Head. Fastigium of vertex equal or slightly narrower than half of scapus. Thorax. Pronotum short, slightly constricted in the middle, median sulcus located after the middle, cylindrical in prozona and distinctly raised in metazona, caudal margin of the disc concave, median carina occurs as a weak yellowish line; paranotal margin almost straight along prozona and oblique along metazona. Tegmina short, extending beyond the posterior margin of pronotum and reach to half of the second abdominal tergite; stridulatory vein almost totally covered by pronotum; stridulatory file with ca. 60 teeth. Male terminalia. Cercus cylindrical, gradually tapering toward apex, prominently incurved at apical half, incurved roughly as L-shaped, with a cylindrical distal branch almost as long as proximal branch and with 4–5 distinguishable denticles on external margin of distal branch and 3–4 denticles along the tip. Subgenital plate as wide as or slightly wider than long, with a quadrangularly concave caudal margin. Song. Male song consists of 11.3 (9–16) syllable duration (ms) and 14.65 (10–21) impulse number per syllable with occasional after-clicks following the main syllable at 20–30 ms. The peak frequency spectrum lies between 35 and 50 kHz. Thus, it is very similar to the song of P. denizliensis (see under the latter). Male song is exemplified in Figure 9 and song measurements are provided in Table 3. — Female. Similar to male in general. Pronotum slightly raised in metazona, tegmina well visible beyond hind margin of pronotum. Coloration. General coloration black with a creamish pattern; vertex black or with black dots on a creamish brown background, antennae black with regular white rings as in the group. Thorax. Disc of pronotum with black patterns or spots on a creamish brown background at the beginning of prozona, black in the middle and reddish brown in metazona; ventral half of the paranota creamish and dorsal half with black pattern; tegmina with typical black/light (marble or yellow) pattern; all legs are black dorsally. Abdomen. Abdominal terga black in front 3/4 and light in the remaining part, the black bands do not extend to subsequent tergum laterally. Female terminalia. Subgenital plate triangular, ovipositor typical of the group.

Diagnosis.

The new species, P. boncukdagensis sp. nov., shows sister group relationships with P. ciplaki + P. datca. P. boncukdagensis sp. nov. was suggested as a separate identical species by bPTP and GMYC while was placed within P ciplaki by ASAP delimitation tests. However, the new species and P. ciplaki well differ from each other by male cercus. Cercus is weakly incurved, with rounded apex, denticles constitute a single row along the tip in P. ciplaki, while strongly incurved, L-shaped and with truncate apex, denticles constituting two lines, one along the tip and the other along external margin in P. boncukdagensis. Additionally, male subgenital plate is as wide as long or slightly wider than long in the new species, while it is longer than wide in P. ciplaki. The new species and P. datca are not monophyletic and no delimitation test suggested it belongs within P. datca, but the new species is rather similar to P. datca montana especially by the male cercus. The new species differs by the distal branch of cercus as long as proximal branch (longer than the half-length of the proximal branch), while it is at most as long as the half-length of the proximal branch in P. datca datca. Additionally, the distal branch of cercus is black in the new species while dark but not black in P. datca. Apart from the male cercus, P. datca and P. boncukdagensis sp. nov. can also be distinguished by the male calling song; a syllable consists of 1–4 and 10–21 impulses in the first and second species respectively (Table 3). Along with these phenotypic characters, there are 8 mutations, detected in the concatenated matrix of COI+ND2+VAL by applying a PAUP analysis (File S1), specific to the ancestral node of P. boncukdagensis sp. nov., which we considered further diagnostic characters of the species, at the position 873 (T→C), 1322 (A→G), 1331 (C→T), 2039 (T→C), 2063 (A→G), 2498 (A→G), 2499 (G→T), 2500 (T→G) and 2897 (G→A).

Derivatio nominis.

The name of the new species is established by the name of range area “Boncuk Dağları” Mts., located between Muğla and Denizli Provinces of Turkey.

Remarks.

Currently the new species, P. boncukdagensis sp. nov., is known only from the type locality Tuzla Pass of Boncuk Mts., but this altitudinal chain is isolated by lowlands from surrounding highlands. Regarding this statement, the record of P. zonatus zonatus from Sandras Mt. in Muğla Province (close to Tuzlabeli) by Sevgili et al. (2018) may refer to P. boncukdagensis but needs confirmation. The male song consists of 11.3 (9–16) ms of 14.65 (10–21) impulses and with occasional after-clicks following the main syllable at 20–30 ms. The peak frequency spectrum lies between 35 and 50 kHz. Male song is exemplified in Figure 9 and song measurements are provided in Table 3.

Material examined.

See population 10 in Table 1. Type material: TURKEY, Muğla, Boncuk Mts. Tuzlabeli Pass, 36.87161N, 28.16340E, 1401 m, 30.05.2021 (leg. B. Çıplak and Ö. Yahyaoğlu). Two males (including holotype) and 2 females, Turkey, Muğla, Boncuk Mts. Tuzlabeli Pass, 36.87161N, 28.16340E, 1401 m, 30.05.2021 (leg. B. Çıplak and Ö. Yahyaoğlu) (in alcohol in MEVBIL); 1 male, 4 females, Turkey, Muğla, Fethiye, Tuzlabeli Pass, 1650 m, 29.07.1997 (leg. B. Çıplak) (in AUZM). For descriptive structures see Figure 8, and for calling song Figure 9. The Mgt population given under the P. ciplaki in Kaya, 2018: p. 87, Fig. 1. represents this new taxon.

Figure 8. 

Diagnostic structures in male and female of P. boncukdagensis sp. nov. The upper panel show female structures (A pronotum from above, B pronotum from lateral view, C ovipositor from lateral view, D subgenital plate, and E epiproct and cercus), and the lower panel shows male structures (F pronotum from above, G pronotum from lateral view, H coloration of 2. and 3. abdominal terga, I abdominal terminalia from above, J subgenital plate, and K cercus).

Figure 9. 

Oscillographic representation of the male song of P. boncukdagensis sp. nov. – male from Tuzlabeli Pass, 1401 m. Song shown at different speeds. A A frame of 1 min, recording at 23.6°C; B single syllable (frame of 100 ms; 27.1°C); C single syllable (frame of 100 ms; 26.7°C).

3.10 . Poecilimon (Poecilimon) ciplaki Kaya, 2018

https://orthoptera.speciesfile.org/otus/847296/overview

Poecilimon ciplaki Kaya, 2018: 92; Poecilimon salmani Sevgili, 2018 in Sevgili et al. 2018: 37, syn. n. (https://orthoptera.speciesfile.org/otus/847293/overview).

Remarks.

Samples collected from highlands in the west of Antalya Province of Turkey (namely Tahtalıdağ, Ovacık Village, Bakırlıdağ and Uzunkarıştepe), lowlands (namely Fethiye, Dalaman, Ortaca and Marmaris) and the highland (Tuzlabeli-Boncuk Mts.) in south of Muğla Province, and highlands in east/southeast of Denizli Province (Honaz Mt.) were differently identified by Kaya (2018) and Sevgili et al. (2018). Kaya (2018) listed these localities under the new species, P. ciplaki, as two subspecies. However, Sevgili et al. (2018) identified them as three different taxa; (i) the highland populations from Beydağları Mts. (namely Tahtalıdağ Mt. and Ovacık Village) as P. zonatus zonatus Bolivar, (ii) the lowland Ortaca population as the new species, P. salmani Sevgili, and (iii) the Marmaris population as P. zonatus datca Sevgili, Sirin, Heller & Lemonnier-Darcemont. Sequences from these localities constitute a single clade in the phylogenetic tree (Clade IIA). Species delimitations tests suggested 3 to 6 distinct species in the clade and we classified them as four species, two of them polytypic. One of these species, consistently suggested identical by all species delimitation tests, consists of sequences from the lowland plain of Fethiye, Dalaman and Ortaca, which corresponds to the type localities of P. ciplaki ciplaki Kaya and P. salmani Sevgili.

Proposing the population in the lowlands of Fethiye, Dalaman and Ortaca as an identical species suggests P. ciplaki Kaya and P. salmani Sevgili as a single species and requires a nomenclatural change. Kaya (2018) and Sevgili et al. (2018) were published in the same year. The online version of Kaya (2018) appeared on 12 April 2018 and the published version on 15 April 2018. Sevgili et al. (2018) appeared on 3 May 2018. According to the priority rule of the International Commission on Zoological Nomenclature (ICZN), Kaya (2018) has priority and thus the new taxon proposed by Sevgili et al. (2018), namely P. salmani Sevgili, constitutes a junior synonym of P. ciplaki Kaya (Kaya 2018). It should be noted that the type localities of P. salmani and P. ciplaki are very close to each other along the same lowland. Rich illustrative material describing the morphology and song of these populations can be found in Kaya (2018) and Sevgili et al. (2018). However, phenotypic characters should be used with caution, as the phenotypic units defined in the previous studies do not correspond to genetic units (see Discussion section below). This species can be easily distinguished from all others in Clade IIA by the male cercus weakly incurved, with rounded tip and with denticles only at the tip. In addition to typical male cercus, there are unique mutations, detected in the concatenated matrix of COI+ND2+VAL by applying a PAUP analysis (File S1), at the position 456 (T→C), 810 (C→T) 915 (A→T) and 2431 (G→A) defining P. ciplaki as a unique clade.

Song.

Male song exemplified in Sevgili et al. (2018) shows a compact syllable with occasional appearance of an after-click. Our recordings (Fig. 10) revealed much higher complexity with the main part usually followed by two types of impulse groups – one or two high-energy short clicks (with 2–3 impulses), the first or both of which followed by a group of low-energy sparse impulses counting ca. 3–12. The whole complex syllable may thus last up to ca. 200 ms. As the presence and arrangement of distinct syllables varies, the reason why Sevgili et al. (2018) did not notice the same patterrn may be either a result of using short recordings of young males and/or of the properties of the recording equipment. This song structure is a good diagnostic character of the species by which it can be easily distinguished from all other members of the group. Its genetically sister taxon, P. datca stat. nov., is well differentiated by its very simple song of isolated short syllables of 1 to 4 impulses lasting 7–15 ms (Sevgili et al. 2018 and Table 3).

Figure 10. 

Oscillographic representation of the male song of P. ciplaki – male from Muğla, Fethiye, Dalaman. Song shown at different speeds. A A frame of 1 min, recording at 23.6°C; B single syllable (frame of 100 ms; 22.9°C); C single syllable (frame of 100 ms; 26.7°C).

Distribution.

Regarding the above-listed localities (Kaya 2018; Sevgili et al. 2018) and those given in Table 1, this species occurs in the lowlands of Fethiye, Dalaman, and Ortaca in the south of Muğla provinces of Turkey.

Material examined.

See population 6 in Table 1. Regarding new taxonomic rearrangements made here, the distribution of P. ciplaki Kaya requires to be redefined, especially for published localities. Holotypes and paratypes of P. ciplaki (material examined): TURKEY: Muğla, Fethiye, road to Dalaman, N:36.75000 E:28.90000, 258 m, 14.V.2011, 7 males (including holotype), 7 females (leg. S. Kaya, Z. Boztepe and Ö. Pekter) (MEVBIL); Holotypes and paratypes of P. salmani (other records): TURKEY: Muğla, Dalyan, İztuzu, 36°46'.490"N, 28°39'.575"E, 239 m, 27.05.2002, 10 males, 7 females (leg. H. Sevgili and Y. Durmuş) (in Hacettepe University Zoological Museum (­HUZOM), Ankara, Turkey); Ortaca, Dalyan, 36°46'.77"N, 28°38'.22"E, 350 m 13.05.2016,13 males, 5 females (leg. H. Sevgili) (in alcohol in ODUZOOL, HSC – Ordu University, Zooloji, Zoology Laboratory, Turkey).

3.12 Poecilimon (Poecilimon) vodnensis Karaman, 1958

https://orthoptera.speciesfile.org/otus/847285/overview

Poecilimon vodnensis Karaman, 1958: 39; Poecilimon vodnensis Karaman, 1958 in Harz, 1969: 145; Poecilimon vodnensis Karaman, 1958 in Chobanov & Mihajlova, 2010: 92; Poecilimon vodnensis Karaman, 1958 in Lemonnier-Darcemont & Darcemont, 2016; Poecilimon (Poecilimon) vodnensis Karaman, 1958 in Sevgili et al. 2018: 47.

Remarks.

Detailed morphological and acoustic descriptions, with rich illustrative material, were given by Karaman (1958) and Sevgili et al. (2018). Our song recordings of male-female duets provided in addition to the data by Sevgili et al. (2018), reveal two types of male song (either of single syllable or groups of two syllables) (Fig. 11). The main, first, or single, syllable has 2–4 impulses (rarely up to 7) and lasts 5 to 16 ms. The second syllable was measured with 3 to 9 impulses and lasted 7 to 14 ms. Together the two-syllable groups had a duration of 34–82 ms. Male song of two syllable types is represented within the two subgroups of P. zonatus group and was reported for P. azizsancar and P. variicercis (Sevgili et al. 2018), as well as is common for the Poecilimon jonicus group (e.g., Borissov et al. 2020, 2023).

Figure 11. 

Oscillographic representation of the male and female songs of P. vodnensis from Bonche vill., Mariovo, North Macedonia, 1000 m alt. Song recording at T = 25.5°C. A A frame of 1 min. with the male song above and female song below; B example of a male calling (above) and female responding (below) syllables (female song of a single main syllable followed by after-clicks); C example of a male calling (above) and female responding (below) syllables; single syllable (female syllable of two main impulses followed by a quiet buzz); D example of a bi-partite male song (two type of syllables). Frames of B–D correspond to 100 ms.

The female song consists of a first part of one or, more frequently, two impulses, and of a second part consisting of a few impulses with lower energy or a quiet ‘buzz’ (compare Fig. 11 – B and C). The first part of two impulses lasts 5–11 ms, while the whole song if the second part is present lasts ca. 40–60 ms. The latency times for the female response measured from the beginning of the male song was 27–39 ms (the latency lasts longer when there is one impulse in the first part of the female response). Table 3 provides additional characteristics of the species song.

Distribution.

This species is the only representative of the group in the Balkans, known only from the type locality, Vodno Mt., and a few closely situated locations in the Mariovo region of North Macedonia (Fig. 1, Table 1, and Sevgili et al. (2018)).

Material examined.

See population 11 in Table 1.

3.14 Poecilimon (Poecilimon) parazonatus Uluar, Chobanov & Çıplak, sp. nov.

https://zoobank.org/7D6BB4A9-C25F-417F-902C-F2C68D8D3402

Description.

Holotype, male. Head. Fastigium of vertex equal or slightly wider than half of the scapus. Thorax. The pronotum short, slightly constricted in the middle, median sulcus located after the middle, cylindrical in prozona and distinctly raised in metazona, caudal margin of the disc concave, medial carina occurs as a yellowish line or absent, disk bordered by large light lines divergent in anterior and posterior margins constituting roughly as “) (“ shape; paranotal margin almost straight along prozona and oblique along metazona. Tegmina short, extend beyond the posterior margin of pronotum, stridulatory vein hardly visible under the pronotum; the stridulatory file with 58 teeth. Male terminalia. Cercus cylindrical, gradually tapering toward apex, the curvature is more prominent apically, incurved roughly as L-shaped, with a robust, but slightly tapered apex and 2–3 hardly distinguishable denticles apically. The subgenital plate is as long as wide or slightly wider than long, distal margin is almost truncated.

Song.

Male song is not known.

Female.

Similar to the male in general. Thorax. Pronotum distinguishably raised in metazona, tegmina slightly extended beyond the hind margin of pronotum. Female terminalia. Subgenital plate triangular, ovipositor typical of the group.

Coloration.

General coloration black with a creamish pattern; vertex black or with black dots on a creamish brown background, antennae black with regular white rings as in the group. Disc of pronotum with black patterns or spots on a creamish brown background at the beginning of prozona, black in the middle and reddish brown in metazona; paranota with black spots on a creamish brown background; tegmina with typical black/light (marble or brown) pattern; all legs are black dorsally. Abdominal terga black in front 2/3 and light in the remaining part, the black and light bands extend into each other showing a population-specific pattern.

Diagnosis.

The three infraclades in Clade IIB, each of which was consistently suggested as distinct species by all species delimitation tests, show P. variicercis + (P. parazonatus sp. nov. + P. zonatus) relationships on the phylogenetic tree. However, they are very similar in the traditionally used structures/characters (Fig. 12). The distinct cercus, with almost blunt apex and indistinguishable denticles located at the tip, is the most prominent character distinguishing P. parazonatus sp. nov. from the other two related species. The black coloration of abdominal terga is more distinct in P. parazonatus sp. nov., but it is not reliable as it may differ in young and elder individuals, or may be locality-specific. Apart from phenotype, there are unambiguous mutations specific to the ancestral node of this species, which we considered as diagnostic characters of the species. These positions are (see File S1); 1961 (A→G), 1259 (T→C), 1331 (A→T), and 1424 (T→ C). The first is unique to P. parazonatus sp. nov., and the other three unambiguous mutations are not unique to this clade, but none of them is shared with the two closest relative species, P. variicercis and P. zonatus.

Figure 12. 

Diagnostic structures in male and female of P. parazonatus sp. nov. The upper panel shows female structures (A pronotum from above, B pronotum from lateral view, C ovipositor from lateral view, D subgenital plate, and E epiproct and cercus), and the lower panel shows male structures (F pronotum from above, G pronotum from lateral view, H coloration of 2. and 3. abdominal terga, I abdominal terminalia from above, J subgenital plate, and K cercus)

Derivatio nominis.

The name of the new species is constituted by considering the phylogenetic position of P. parazonatus sp. nov. with P. zonatus on the phylogenetic tree as P. variicercis + (P. parazonatus sp. nov. + P. zonatus).

Remarks.

The geographic section bordered by the two main branches of Euphrates, namely Murat and Karasu rivers, is an isolated section especially for the species preferring high-altitude habitats (Uluar et al. 2021). Previously Sevgili et al. (2018) identified samples from some localities belonging to this geographic section (from Tunceli and Erzincan Provinces of Turkey) as P. zonatus. However, none of these was from Pülümür Valley, the type locality of P. parazonatus sp. nov. The reports by Sevgili et al. (2018) from this geographic section possibly refer to the new species, yet they require confirmation by genetic data, as phenotype is not reliable enough for identification (see Discussion section below). Studying songs may provide further diagnostic characteristics.

Material examined.

See population 13 in Table 1. Type locality: TURKEY, Tunceli, Pülümür, Erzincan-Tunceli road, 39°31.7454’N, 39°53.5074'E, 1697 m, 08.07.2009, 1 male (holotype), 2 females (leg. M. Korkmaz) (all in alcohol in MEVBIL). For descriptive structures see Figure 12. The TP population given under the P. zonatus in Kaya, 2018: p. 87, Fig. 1. represents this new taxon.

3.16 Poecilimon (Poecilimon) anisozonatus Uluar, Chobanov & Çıplak, sp. nov.

https://zoobank.org/00F2BBE3-FEB8-40BA-8385-56554C2F4E96

Description.

Holotype, male. Head. Fastigium of vertex equal or slightly wider than half of scapus. Thorax. Pronotum short, slightly constricted in the middle, median sulcus located after the middle, cylindrical in prozona and somewhat raised in metazona, caudal margin of the disc concave, medial carina occurs as a yellowish line, disk bordered by large light lines slightly divergent in anterior and posterior margins constituting roughly as “)(“ shape; paranotal margin almost straight along prozona and oblique along metazona. Tegmina short, extending beyond the posterior margin of pronotum, stridulatory vein not totally covered by pronotum; stridulatory file with ca. 55 teeth. Male terminalia. Cercus cylindrical, gradually tapering toward apex, curvature is more prominent apically, incurved roughly as L-shaped, with a flattened apex and with 4–5 distinguishable denticles on external margin. Subgenital plate wider than long, with a wide roughly quadrangular median processes apically, distal margin is quadrangularly concave.

Song.

Male song consists of short (9–14 ms) syllables of 10–13 impulses that are usually followed by one to four after-clicks, the complex syllable lasting 27–34 ms at ca. 28°C. The peak frequency spectrum lies between 35 and 50 kHz. Male song is exemplified in Figure 14, and song measurements are provided in Table 3.

Female.

Similar to males in general. Pronotum slightly raised in metazona, tegmina well visible beyond the hind margin of pronotum. Subgenital plate triangular, ovipositor typical of the group.

Coloration.

General coloration black with a creamish pattern; vertex black or with black dots on a creamish brown background, antennae black with regular white rings as in the group. Disc of pronotum with black patterns or spots on a creamish brown background at the beginning of prozona, black in the middle and reddish brown in metazona; paranota with black spots on a creamish brown background; tegmina with typical black/light (marble or brown) pattern; all legs are black dorsally. Abdominal terga black in front 1/2 and light in the remaining part, the black bands laterally extend to subsequent tergum and light bands remain in the middle showing a population-specific pattern.

Diagnosis.

The new species, P. anisozonatus sp. nov., shows sister group relationship with P. isozonatus, and each of them was consistently suggested as a distinct species by all species delimitation tests. The pair of P. isozonatus / P. anisozonatus sp. nov. can easily be distinguished by traditionally used phenotypic characters (Table 3 and Fig. 13). The typical cercus with flattened apex and prominent denticles located externally (no denticle on the internal side of apex), and male subgenital plate wider than long, are the most prominent characters distinguishing P. anisozonatus sp. nov. from P. isozonatus. Apart from the cercus and subgenital plate of the male, P. anisozonatus sp. nov. can be distinguished from P. isozonatus also by its male song – the syllable duration being 30.57 (27–34) ms in the first, while 13.6 (10–19) ms (Konya-Taşkent) or 6.46 (2–8) ms (Niğde-Çamardı) in the second (Table 3). Along with these phenotypic characters, there are six mutations, detected in the mitochondrial concatenated matrix by applying a PAUP analysis (File S1), specific to the ancestral node of P. anisozonatus sp. nov., which we considered as a diagnostic character of this species; 1730 (C→T), 1799 (A→G), 1917 (A→C), 1928 (T→C), 2118 (C→T), 2223 (C→T).

Figure 13. 

Diagnostic structures in male and female of P. anisozonatus sp. nov. The upper panel shows female structures (A pronotum from above, B pronotum from lateral view, C ovipositor from lateral view, D subgenital plate, and E epiproct and cercus), and the lower panel shows male structures (F pronotum from above, G pronotum from lateral view, H coloration of 2. and 3. abdominal terga, I abdominal terminalia from above, J subgenital plate, and K cercus)

Figure 14. 

Oscillographic representation of the male song of P. anisozonatus sp. nov. – male from Hadim-Gündoğmuş, 1900 m. Song recording at T = 27.8°C. Song shown at different speeds. A A frame of 1 min; B single syllable without after-click (frame of 100 ms); C single syllable with an after-click (frame of 100 ms).

Derivatio nominis.

The name of the new species is constituted to express its close relation, but the clear distinction, from P. isozonatus.

Remarks.

Currently, the new species, P. anisozonatus sp. nov., is known only from the type locality, adjacent to that of P. isozonatus, but separated by a lowland valley. Although the samples constituting the type specimens of the new species were reported as P. isozonatus and these two species show sister group relationship, signs from genetic data suggest that they are two independent evolutionary and reproductive units and there are considerable phenotypic differences, especially in male calling songs, supporting their distinctiveness.

Material examined.

See population 17 in Table 1. Type locality: TURKEY, Antalya, Gündoğmuş, road to Hadim-Konya, 36.88333N, 32.11667E, 1887 m, 15.6.2014. 5 males (including holotype), 2 females (leg. S. Kaya and D. Chobanov) (all in alcohol at MEVBIL). For descriptive structures see Figure 13. The KnH population given under the P. isozonatus in Kaya, 2018: p. 87, Fig. 1. represents this new taxon.

Traditionally systematics and taxonomy were defined as two different research enterprises. Systematics was defined as “the scientific study of the kinds and diversity of organisms and of any and all relationships among them” and taxonomy as “the theoretical study of classification, including its bases, principles, procedures and rules” (Simpson 1961). Developments in DNA-based systematics merged both research disciplines. The case of the P. zonatus group presented here constitutes a hallmark example highlighting that taxonomy remains incomplete or even incorrect, without systematics. The species taxonomy of the P. zonatus group before this paper was established by phenotype (morphology and male calling song) and based on the traditional approach in two parallel and almost simultaneously published studies (Kaya 2018; Sevgili et al. 2018). Here, we merge data from the latter studies and performed a phylogenetic study. As a result, the taxonomy of the group was changed extensively when subjected to a DNA-based systematic approach. Along with the proposed new nomenclatural changes, the systematic approach applied required redefinition of the reproductive units, their distribution range, and phylogenetic relationships, and consequently a more functional diversity definition for the group.

Another aspect in the present study highlighting the importance of linking taxonomy to systematics is related to the species concepts. Although the species concepts are diverse and controversial in applying to the taxonomy of a lineage (Coyne and Orr 2004), the present study indicates that species definition based solely on morphology (or any other phenotype) may lead to the guardianship of the phenotype over the genotype, a case conflicting with the nature of organisms. It is a simple knowledge that parents do not inherit their phenotypes to their offspring, but they inherit their genes, and their phenotypes re-occur according to this genetic inheritance. Thus, genotype includes direct signs to define the evolutionary history of a species lineage, while those of a phenotype are indirect. However, the general taxonomical practice, as for the P. zonatus group, gives priority to phenotype. The unified species concept by De Queiroz (2007) suggests a solution for this handicap, defining species by the primary property, the phylogenetic uniqueness or independently evolving lineages, and then delimiting them by phenotype. Application of this concept to the P. zonatus group allowed us to detect hidden species, or combine wrongly defined ones, as a consequence of a proper definition of intra-group diversity.

Following the order in this concept enabled us to determine further aspects related to the P. zonatus group. The rate of genetic and phenotypic divergence among the phylogenetic units in the group seems to be different. For example, P. isozonatus and P. anisozonatus constitute two sister clades with relatively low pairwise genetic distance within the group, but they are well distinguishable by cercus, subgenital plate and calling song of the male. Contrary to the amount of genetic divergence, P. anisozonatus is more similar to some other species, e.g. to the distantly related species P. variicercis by the wider than long male subgenital plate. The similar structure of male cercus (denticles located at the tip) in P. variicercis and P. ciplaki each belonging to separate subclades, constitute another such example. Again, variation per species in several phenotypic characters (e.g. size measurements for pronotum, tegmina, hind femur etc. or the number of stridulatory teeth and duration of a syllable) per species in the P. zonatus subgroup mostly overlap between species (see Sevgili et al. 2018; Kaya 2018), but similarity/dissimilarity in these characters do not corroborate with the amount of genetic difference. Possibly these cases also suggest that the phenotypic characteristics of the group contain a considerable amount of homoplasy. This is the reason that species in the group are diagnosable by minor phenotypical differences, which led to a controversial taxonomy. For example, Sevgili et al. (2018) collected all of P. zonatus, P. isozonatus, P. anisozonatus, P. denizliensis and P. datca under a single species, P. zonatus. There are further similar examples indicating that the amounts of genetic and phenotypic differentiation do not corroborate, but all together indicate that the rate and the route of divergence in the phenotype and genotype were different after a split from the ancestral stocks, which resulted in current lineages or species/subspecies. Regarding the consistent and inconsistent cases of phenotypes and genotypes, we arrived at the following conclusions. First, the phenotype, especially morphology, is unproductive in defining intragroup diversity within the P. zonatus group. Second, members of the group exhibit a conserved morphology. Third, because of this conserved phenotypic evolution, genetic data constitutes a reliable character source to define intra-lineage phylogeny and diversity.

The rectification of the group allowed us to determine intragroup hidden diversity, yet raised new questions to be answered, particularly regarding the potential reasons behind the conserved phenotype. A song consisting of tick-like, irregularly producing syllables is common in the group (Figs 8–11; Kaya 2018; Sevgili et al. 2018), and syllable duration or impulse number in some cases are similar in distantly related species (see Table 3). Calling song has a crucial role in the mating system (Heller et al. 2015) and divergence in song may be harmful unless there is no selection pressure imposing it. Allopatry or the absence of related species in sympatry was suggested as a reason leading to conserved song features in other tettigoniids (e.g. Çıplak et al. 2009; Uluar et al. 2023). This may be the case also for the P. zonatus group, as all species are allopatric in distribution and possibly evolved through an allopatric speciation process including vicariance and dispersal events.

The reasons leading to a conserved morphology may also be identified in the radiation history of the lineage. Except for the lowland P. ciplaki and one population of P. datca, all species/populations of the group are cold-adapted mountainous species occurring on particular altitudes and/or mountain chains. Distribution, and thus ecological preference of P. datca seems exceptional as it is represented both by highland and lowland populations. The Bakırlıdağ and Uzunkarıştepe populations (P. datca montana) occur at about 1500 m, while Marmaris-Datça populations (P. datca datca) at about 50–300 m elevation (Table 1, Fig. 1). Although these populations have diverged to some degree, their haplotypes constitute a single haploclade, and two delimitation tests suggested them as a single species. Thus, it seems that P. ciplaki and P. datca datca evolved after a habitat shift (see below). Given that all other members of the P. zonatus group are mountainous, conserved ancestral habitat preference seems to be the general pattern. Due to conserved habitat preference or tracking similar ecological conditions/niches, these animals were subjected to similar selection pressures, and as a consequence of such an evolutionary process, they exhibit similar or slightly diverged morphology (Ackerly 2009; Cadotte et al. 2013; Uluar et al. 2022, 2023).

In conclusion, merging taxonomy to systematics by using sequence data in examining the P. zonatus group allowed for a better definition of intra-group diversity/phylogeny, and led to extensive nomenclatural rectification in examining the P. zonatus group through DNA-based sequence data. This approach allowed us to detect three new species named as P. anisozonatus sp. nov., P. parazonatus sp. nov. and P. boncukdagensis sp. nov., two new subspecies as P. datca montana subsp. nov. and P. denizliensis kizildagi subsp. nov., to synonymize P. salmani Sevgili with P. ciplaki Kaya, and to elevate two subspecies to species level, P. denizliensis stat. nov. and P. datca stat. nov. As of now, the P. zonatus group consists of 13 species, two of which are polytypic. Although P. varicornis was not available to include in our phylogenetic analysis, the relationships among the remaining species in the group are ((P. tauricola + P. azizsancar) + ((P. denizliensis + (P. boncukdagensis + (P. ciplaki + P. datca))) + ((P. vodnensis + (((P. variicercis + (P. zonatus +P. parazonatus)) + (P. isozonatus + P. anisozonatus)))). This phylogenetic pattern supports the P. tauricola and P. zonatus subgroups as proposed by Sevgili et al. (2018), but the monophyly of both subgroups still requires confirmation (see below).

The evolutionary narrative of the P. zonatus group can be evaluated from four distinct perspectives. The first pertains to the monophyly of the species group. The common ancestor of the species group is dated approximately 7 myr ago, a date significantly older than that reported by Borisssov et al. (2023). However, the dataset utilized by Borisssov et al. (2023) only includes sequences representing three species of the group, all from P. zonatus subgroup and with no representive from P. tauricola subgroup. The ancestral age of the P. tauricola + P. zonatus subgroups may appear older as they are two distant lineages. However, an old ancestral age raises further inquiries. Did the species group remain undiversified for approximately 3 million years following the first splitting, or did a significant extinction event occur in this interval? One plausible explanation for these questions could be that the P. tauricola and P. zonatus subgroups do not form a monophyletic clade, and the current chronograms may indicate a misleading intermediate time. No question remains if these two sub-lineages are not monophyletic and if there are other unknown basal-internal branches within the lineage, particularly following the common ancestor, that have not been included in current phylogenies. Although existing studies report these two groups as monophyletic sister clades, unpublished data (Uluar et al.: unpublished) suggest that these two subgroups may not form a monophyletic lineage, and present monophyly is potentially due to the low taxonomic coverage of the current datasets. More extensive datasets may demonstrate that these two lineages are polyphyletic or paraphyletic, necessitating a redefinition of the species group boundaries.

The second aspect regarding the evolution of the group pertains to the intra-group diversification period. Clade I, or the P. tauricola subgroup, diverged into two species approximately at the beginning of the Pleistocene. However, the diversification of Clade II, or the P. zonatus subgroup, indicates a pattern that can be explained in terms of diversification time and climatic events (or potential evolutionary drivers) that occurred during this period. Although the ancestral node of Clade II corresponds to the early Pliocene, the ages of the ultimate common ancestors of the nodes representing species taxa fall within the Pleistocene, particularly around the Mid Pleistocene Transition (Köhler and van de Wall 2020). This temporal correlation not only suggests that major climatic cycles are the main evolutionary drivers but also confirms that this is a general pattern, especially for Anatolian forms with similar habitat preferences, particularly cold-tolerant flightless tettigoniids (Kaya and Çıplak 2016, 2017; Uluar et al. 2023; Çıplak et al. 2024; Ortego et al. 2024; and references therein). Similarly, corresponding of the ages of the proximate node of the lineages proposed as species/subspecies taxa within the last four glacial periods indicates that the last four intensive glaciation periods were the main determinants of intra-clade diversification. Habitat preference of current populations, occurrence in mountainous habitats, is consistent with this pattern. In this context, RASP analysis suggests that two habitat shifts occurred in the history of Clade IIA, first from high to low altitudes that led foundation of P. ciplaki and P. datca datca, and second, secondarily returning to highlands, which led the foundation of P. datca montana. It is noteworthy that the time of both habitat shifts corresponds to the Pleistocene again (Kaya et al. 2015).

The third aspect to be discussed regarding the evolution of the group is its phylogeography. The species comprising Clade I, namely P. tauricola and P. azizsancar, are local species distributed along the Anatolian Diagonal (Fig. 1). Although both species are allopatric, they exhibit remnants of some ancestral stocks dispersed along the whole Anatolian Diagonal, a geographic entity considered as both a dispersal and diversification corridor (Çıplak et al. 1993; Kantor et al. 2023; Korkmaz et al. 2014; Mutun 2016). The recent description of P. azizsancar as new species (Sevgili et al. 2018) from this corridor indicates the presence possibility of further new species on certain peaks along this line. The phylogeography of Clade II harbours several details. Firstly, the distribution area of Clade II can be delineated into five subfragments: The Balkans, Caucasus, Southwest Anatolia, Eastern Anatolia, and Southern Taurus Mountains (Fig. 3). Regarding the phylogeny of lineages distributed in these fragments and the results of RASP analysis, the following relationship of the fragments can be assumed; Southwest Anatolia + (Balkans + (Southern Taurus Mountains + (Eastern Anatolia +Caucasus))) (see Fig. 3). Based on this statement it can be suggested that one of the two earliest derivatives of common ancestor of Clade II dispersed in Southwest Anatolia and radiated here autochthonously while the other occupied remaining parts of Anatolia. The Balkan lineage represented by P. vodnensis was established by an ancestral stock dispersed from Anatolia. Considering the node ages indicated by the chronogram and the results of RASP analysis, it is conceivable to assume that dispersal occurred in Pliocene when terrestrial corridor existed and the ancestral stock split when terrestrial connections terminated at the beginning of Pleistocene when the Sakarya Strait developed, as observed for many other lineages (Kaya and Çıplak 2016, 2017; Chobanov et al. 2017; Uluar et al. 2023). These statements can be explained by two possible phylogeographic scenarios. The first proposes that the dispersal between the Balkans and Eastern Anatolia occurred through the North of Anatolia. Currently, there is no representative of the lineage in the broad area between the Balkans and the Eastern Anatolian Region, especially in the northern part of Anatolia, thus this scenario requires a possible extinction in this vacant area. The second possible scenario proposes that the transition between the Balkans and Anatolia occurred via the “Taurus Way”, through the southern altitudinal chains (Çıplak 2008; Kaya and Çıplak 2017). Although this scenario does not require the presence of representatives of this group in Central and Northern Anatolia, it indicates a potential extinction of Clade IIB in the Aegean Region of Anatolia. However, the RASP analysis suggested extinction only for the Caucasian sublineage and not for others. The current data are insufficient for deciding the validity of one of these scenarios, but they allow us to conclude that mountain ranges are one of the main determinants in the formation and survival of the biodiversity of Anatolia (Atalay 2006; Çıplak 2003; Şekercioğlu et al. 2011).

The fourth aspect to be discussed concerning the evolution of the group is related to the dynamics of speciation. The phylogenetic relationships revealed by the analyses in our study, the species/subspecies composition proposed according to species delimitation tests, and the relationships of these taxa in the context of genetic similarity/dissimilarity estimated by pDist and DAPC are noteworthy for revealing intriguing patterns or addressing questions. When viewed through the lens of the general model of speciation (splitting, autonomous differentiation, and acquisition of reproductive isolation) (Coyne and Orr 2004), it is expected that the rates of genetic differentiation would be parallel/proportional to the speciation rate. However, the genetic differentiation patterns of some phylogenetic units considered as species/subspecies vary according to genes. For instance, each species of the P. anisozonatus / P. isozonatus and P. zonatus / P. parazonatus pairs exhibit considerable differences from each other in ND2 and ITS genes but not in VAL (see DAPC cluster in Figure 4 and pDist values in Table S2). Similarly, while P. datca datca / P. datca montana form sister clades, they form the most distant COI-DAPC cluster and independent ITS-DAPC clusters. Additionally, while P. boncukdagensis is phylogenetically sister to the P. datca + P. ciplaki clade, it clusters with P. denizliensis denizliensis in ND2-DAPC, with P. denizliensis kizildagi in VAL-DAPC, and with P. datca datca in ­ITS-DAPC. There are other clustering patterns inconsistent with phylogenetic relationships, especially among phylogenetic units belonging to Clade IIA (see Fig. 4 and Table S1). Some statements regarding these situations may help in providing an explanation on these inconsistencies. Within the P. zonatus species group, currently, there is no sympatric or even parapatric pair of phylogenetic units. All these incongruous situations are observed among phylogenetic units that do not overlap (thus we assume full allopatry) but are geographically adjacent in distribution. Moreover, for all three mitochondrial gene segments, there is no mitochondrial haplotype sharing among phylogenetic units showing inconsistent similarity pattern. Particularly, these incongruous examples are widespread among phylogenetic units distributed in a narrow area of approximately 150 × 250 km in Southwest Anatolia but are rare in other parts of the range area. In light of these observations, a possible explanation is that the gene pool of several current phylogenetic units is not inherited from a single independently evolved preceding unit, rather inherits the legacy of more than one. The fact that Clade IIA, which is diversified in a narrow area in Southwest Anatolia, similarly to the Poecilimon luschani complex (Ortego et al. 2024), exhibits the pattern of microgeographic speciation and historical genetic mixing. This explanation along with the phylogeography of the group described in the preceding paragraph, also indicate to scenario proposed for the Poecilimon bosphoricus species group (Çıplak et al. 2024), where genetic admixture occurred during secondary contacts of temporally isolated and partially differentiated populations due to range shifts driven by climatic cycles. However, the isolation and diversification periods were likely different for Clade IIA of P. zonatus radiated in the south of the Anatolian refugium compared to the P. bosphoricus species group radiated in the north of the Anatolian refugium (Çıplak et al. 2024). Possibly this is the reason why the southern clades do not share haplotypes, contrary to the northern ones.