Phylogenetic analyses of the leafhopper tribe Chiasmini Distant, 1908 and delimination of species of the genus Exitianus Ball, 1929 (Hemiptera: Cicadellidae: Deltocephalinae: Chiasmini) in China based on molecular data

. Previous phylogenetic analyses of the grass-specialist leafhopper tribe Chiasmini have resolved relationships among genera but have included few representatives of individual genera. Here the phylogeny of 20 Chinese species belonging to 8 chiasmine genera was investigated by combining DNA sequence data from two mitochondrial genes ( COI , 16S ) and two nuclear genes ( H3 , 28S ). In both maximum likelihood (ML) and Bayesian inference (BI) analyses, relationships among genera were largely consistent with prior analyses, with most members of the tribe placed into two sister clades: ( Exitianus + Nephotettix ) and the remaining ﬁ ve sampled genera. To examine morphology-based species de ﬁ nitions in the taxonomically dif ﬁ cult genus Exitianus Ball, 1929, one mitochondrial gene ( COI ) and one nuclear gene ( ITS2 ) were used to infer the phylogenetic relationships and status of two common and widespread species and compare the performance of different molecular species-delimitation methods. These analyses divide the included populations into two well-supported clades corresponding to current morphological species concepts but some inconsistencies occurred under the jMOTU, ABGD and bPTP methods depending on the which gene and analytical parameter values were selected. Considering the variable results yielded by methods employing single loci, the BPP method, which combines data from multiple loci, may be more reliable in Exitianus .


Introduction
The grass-specialist leafhopper tribe Chiasmini Distant, 1908 (Hemiptera: Cicadellidae: Deltocephalinae) includes 324 species in 21 genera worldwide.Some of these species are important agricultural pests because of their ability to suck plant sap and spread plant pathogens.Adults and nymphs feed on phloem sap, resulting in inhibition of plant growth and development, yellowing of leaves and withering of the whole plant.Some species of Nephotettix Matsumura, 1902 transmit rice common dwarf disease, verticillium wilt disease and yellow dwarf disease, which can cause substantial economic and yield losses (Ruan 1985).Species of Exitianus Ball, 1929 are important vectors of corn stunt (Carloni et al. 2011), Napier grass stunt disease (Arocha et al. 2009), and Bermuda grass white leaf phytoplasma (Salehi et al. 2009).
The taxonomic status and scope of Chiasmini has changed substantially since the tribe was established by Distant (1908) as Chiasmusaria Distant, 1908, one of the eleven tribes of Jassidae Amyot & Audinet-Serville, 1843.Emeljanov (1962) established Doraturini Emeljanov, 1962, which was assigned to Deltocephalinae Fieber, 1869.Hamilton (1975) treated Chiasmusaria as a synonym of Eupelicini Sahlberg, 1871 (Eupelecinae Sahlberg, 1871) and assigned Doraturina Sahlberg, 1871 as a subtribe of Aphrodini Haupt, 1927(Aphrodinae Haupt, 1927).Oman, Knight & Nielson (1990) included both Chiasmusini Distant, 1908 andDoraturini Emeljanov, 1962 in Deltocephalinae.Zahniser & Hicks (2007) considered Chiasmusini and Doraturini to be synonyms, with Chiasmini having priority.Zahniser (2008Zahniser ( , 2010) ) accurately defi ned the scope of the tribe using phylogenetic criteria and identifi ed morphological synapomorphies uniting the group.Kuoh (1966) was the fi rst to study this tribe in China, recording one genus and three species.Zhang (1990) recorded fi ve species in two genera.Duan et al. (2009) reviewed the subgenus Leofa (Prasutagus) Distant, 1918 and described two new species from China.Duan & Zhang (2012a) recorded four species of Doratura Sahlberg, 1871 from China for the fi rst time.Duan & Zhang (2012b) redescribed the genus Doraturopsis Lindberg,1935 and described the new genus Zahniserius Duan & Zhang, 2012 and species Zahniserius cylindricus Duan & Zhang, 2012from China. Duan & Zhang (2012c) reviewed six species of the genus A conurella Ribaut, 1948 from China including two new records and two new species.Duan & Zhang (2012d) reviewed the genus Gurawa Distant, 1908 and described a new species Gurawa truncata Duan & Zhang, 2012.At the same time, the genus Chiasmus Mulsant & Rey, 1855 was also recorded from China for the fi rst time.Duan & Zhang (2013a) established the fi rst records for the genus Aconura Lethierry, 1876 and species Aconura ochrargentea Emeljanov, 1972 from China.Duan & Zhang (2013b) reviewed the genus Exitianus from China.Duan & Zhang (2014) reviewed the species of Nephotettix from the Chinese mainland, describing and illustrating variation in structure of the male pygofer spines and aedeagus.At present, a total of 11 genera and 39 species of Chiasmini have been recorded in China, but 13 species records are doubtful due to lack of specimens (see checklist in Appendix 1).
The traditional classifi cation of genera and species of Chiasmini was mainly based on external morphological characteristics.Singh-Pruthi (1925) fi rst showed that the morphology of the male genitalia is useful for classifi cation, thus, the identifi cation of species has increasingly become dependent on examination of the male terminalia.More recently, combinations of morphological and molecular data have been used to estimate phylogenetic relationships and delimit species.Ross (1968) published a phylogeny of Exitianus based on an intuitive assessment of morphological variation in this genus.Fang et al. (1993) used 16S rDNA gene sequences to study the phylogeny of New World Deltocephaluslike leafhopper genera, including Exitianus exitiosus Uhler, 1880 of Chiasmini.Kamitani (1999) included several genera now placed in Chiasmini in his phylogeny of Deltocephalini Fieber, 1869 and Paralimnini Distant, 1908.Comprehensive phylogenetic analyses of Deltocephalinae based on morphology (Zahniser & Dietrich 2008) and combined molecular and morphological data (Zahniser & Dietrich 2010, 2013) grouped Chiasmini and most other grass-specializing tribes of Deltocephalinae into a single large clade, and supported the monophyly of Chiasmini.However, a subsequent, more detailed analysis of relationships within the tribe based on DNA sequences placed the genera of Chiasmini in two large lineages but failed to support the monophyly of the tribe as a whole (Zahniser & Dietrich 2015).A few subsequent phylogenetic analyses have included very few species of Chiasmini (Song et al. 2017;Gao et al. 2021;Wu et al. 2022;Yan et al. 2022).Most recently, Cao et al. (2022) analyzed a dataset comprising 730 terminal taxa and > 160 000 nucleotide positions obtained through anchored hybrid enrichment to produce detailed, comprehensive phylogenetic estimates for Deltocephalinae.This analysis strongly supported Chiasmini as monophyletic, in contrast to previous analyses in which the two main lineages of this group (one comprising the macropterous genera Nephotettix and Exitianus, the other comprising the remaining, mostly brachypterous genera) did not consistently group together.Stenometopiini Baker, 1923 was recovered as sister group of Chiasmini in the coalescent gene tree analysis but concatenated maximum likelihood analyses suggested that the sister group of Chiasmini includes Stenometopiini plus three other grass-specialist tribes.
Relationships among major lineages of Chiasmini are well understood, thanks to these prior analyses, but relationships among species within single genera remain poorly explored.The species-level taxonomy continues to be based on very few characters of the male genitalia and many species have been recognized based on relatively minor differences.Thus, more detailed phylogenetic studies, including multiple species within genera and multiple populations within species, are needed to elucidate relationships among species and test the stability of morphological characters traditionally used to distinguish species.Few Chinese species of Chiasmini have been incorporated into prior phylogenetic analyses.Phylogenetic analyses, coupled with molecular-based species delimitation methods, are needed to provide a more reliable species-level classifi cation of this group (Fujisawa et al. 2013;Huang et al. 2013;Zhang et al. 2013;Solís-Lemus et al. 2015;Yang 2015).
Among the most widespread and abundant tropical and temperate species of grassland leafhoppers are the moderately large fully winged forms comprising the genus Exitianus.Exitianus was described by Ball (1929) for Cicadula obscurinervis Stål.Ross (1968) published a phylogeny of Exitianus based on an intuitive assessment of morphological variation in this genus and revised the genus.Fang et al. (1993) used 16S rDNA gene sequences to study the phylogeny of New World Deltocephalus-like leafhopper genera, including Exitianus exitiosus Uhler, 1880.Subsequently, Emeljanov (1999) moved Exitianus from Athysanini Van Duzee, 1892 into Doraturini.Duan & Zhang (2013b) reviewed for the fi rst time the species of Exitianus from China, and showed that the two widespread Chinese species are more variable in the male genitalia than suggested by Ross.Worldwide, this genus contains 43 species of which 6 species occur in Asia but only a few species appear to be relatively widespread.
In this study, the phylogenetic relationships of 8 genera and 20 species of Chiasmini collected in China are inferred based on two mitochondrial genes (COI, 16S) and two nuclear genes (H3, 28S).In addition, two widespread and morphologically similar species, Exitianus indicus Distant, 1908 andE. nanus (Distant, 1908), were selected for more detailed study using molecular species delimitation methods to determine whether the current morphology-based defi nitions of these species is reliable.Based on one mitochondrial gene (COI) and one nuclear gene (ITS2), three single-gene species delimitation methods (jMOTU, ABGD, bPTP) and one multi-gene species division method (BPP) were tested to determine whether the inferred species composition matches current morphological species concepts.The goals of this study are: (i) analyse the phylogenetic relationships of Chinese genera and species of this tribe; (ii) evaluate the validity of the current classifi cation; and (iii) use molecular methods to test the validity of current species concepts.

Taxa sampling and species identifi cation
Leafhopper samples included in this study were collected between 2010 and 2019 from various locations across China.20 representative species belonging to 8 genera of Chiasmini were selected for the phylogenetic study of this tribe.41 specimens of two species of Exitianus sampled from across their known range in China were selected for DNA barcoding and molecular delimitation of species.Specimens were collected directly into 95% or 100% ethanol and stored at -80°C prior to study.Identifi cation of each individual was based on examination of external morphology and male genitalia of the adult using an Olympus SZX10 stereoscopic microscope (Olympus Corporation, Tokyo, Japan) using keys provided by Duan & Zhang (2014).All specimens used in this study are deposited at Northwest A&F University, Yangling, China.The taxa included in the phylogenetic analyses listed in Table 1, and maps of collecting localities are shown in Fig. 1.All Exitianus specimens used for molecular identifi cation    and phylogenetic analyses are listed in Table 2 and geographical distributions are illustrated in Fig. 2. Whenever possible, specimens representing different intraspecifi c morphological variants were selected.
Because the most reliable morphological characters for distinguishing species are found in males, only male specimens were selected.Voucher specimens are deposited in the insect collection of the Anhui Agricultural University.

DNA extraction, PCR amplifi cation and sequencing
Total genomic DNA was extracted from abdominal muscle of individual, non-parasitized adult male specimens using an EasyPure Genomic DNA Kit (EE101; Transgen, Beijing, China) following the manufacturer's standard protocol, except that 20 μl proteinase K was mixed with 100 μl buffer for overnight lysis at 56°C and the fi nal elution volume was 60 μl due to small specimen size.After DNA extraction, the DNA solution was stored at -20°C for subsequent molecular experiments.The abdominal exoskeleton of each extracted individual was stored in glycerin in a micro vial as a morphological voucher specimen.
Standard PCR methods were used to amplify partial sequences of two mitochondrial genes (COI, 16S) and three nuclear genes (H3, ITS2, 28S).Primer sequences are shown in Table 3.The amount of template DNA was adjusted according to the DNA concentration and varied between 2 and 3 μl (Folmer et al. 1994;Simon et al. 1994;Colgan et al. 1998;Ji et al. 2003;Dietrich et al. 2001), combined with 12.5 μl of 2 Taq MasterMix, 1μl each of forward and reverse primer, and ddH 2 O added to make a total volume of 25 μl for each reaction.
The PCR conditions differed according to the gene and the specifi c primers, especially the annealing temperature, which was the most critical factor infl uencing product quality.Thermal cycling conditions for each gene were as follows: an initial denaturing step at 94°C for 5 min; 35 cycles of denaturing at 94°C for 1 min; annealing at 52, 54, 55, 56 and 52°C for 1 min for COI, 16S, H3, ITS2 and 28S respectively; an extension at 72°C for 1 min; and a fi nal extension step of 72°C for 10 min, and ending with incubation at 12°C.The PCR products were examined using 1% agarose gel electrophoresis with ethidium bromide stain to ensure the products were the target size.DNA products were subsequently sequenced in both directions by Qingke Biotech (Xi'an) Co., Ltd, using the original PCR primers.

Data analyses
To obtain single consensus sequences, chromatograms, including sense and antisense strands, were analyzed and assembled using Seqman software (Swindell & Plasterer 1997).In order to ensure that the correct target gene fragment was obtained, the Basic Local Alignment Search Tool (BLAST) was used to check all sequences against the NCBI database (Altschul et al. 1990).Concurrently, for the encoding gene fragments, MEGA 7 was used to translate the assembled contigs into amino acids to ensure that stop codons and pseudogenes did not exist ( Kumar et al. 2016).Conserved sites (C), variable sites (V), parsimony-informative sites (PI), and the average nucleotide composition for each region were calculated by MEGA 7 (Kumar et al. 2016).Genetic diversity parameters including the haplotype number (H), haplotype diversity (Hd), and nucleotide diversity (Pi) were calculated by DNASP ver.5.0 for our subsequent experiments (Librado & Rozas 2009).Before combining multiple genes to build trees, scatter plots of transitions/transversions of each gene were made by using DAMBE ver.5.0 software (Xia 2013) to test for substitution saturation.A heuristic search in PAUP ver.4.0b10 (Swofford 2002) was used to test for partition homogeneity of the six gene sequences, prior to combining them for analyses.

Phylogenetic analyses of Chiasmini Distant, 1908
Based on the analysis of Zahniser & Dietrich (2015)  within the tribe.Each of the two coding protein gene fragments (COI and H3) was aligned separately using the MAFFT plugin and G-INS-I algorithm in PhyloSuite, with gaps and ambiguous sites removed using Gblocks under default settings in PhyloSuite ver.1.2.2 (Zhang et al. 2020).The ribosomal gene fragment (16S and 28S) was aligned using the Q-INS-I method on the MAFFT ver.7 alignment server (Katoh & Standley 2013).Alignments of individual genes were concatenated to generate one 24taxa data set using PhyloSuite ( Zhang et al. 2020).Phylogenetic reconstruction was conducted using Maximum likelihood (ML) and Bayesian inference (BI).The most suitable substitution models and partition scheme were determined for the combined dataset using PartitionFinder ver.2.1.1 (Lanfear et al. 2017).The best-fi tting model was selected for each partition with the model search 'all' and 'mrbayes' for ML and BI analyses, 'greedy' search algorithm, 'linked' to estimate branch lengths and using the Bayesian information criterion (BIC) (Table 4).
Maximum likelihood analysis was conducted using the IQ-TREE plugin in PhyloSuite ver.1.2.2 ( Zhang et al. 2020).Ultra-fast Boot (UFB) algorithm with 1000 repetitions (Minh et al. 2013) and the SH-aLRT test was used to assess branch support.Bayesian inference analysis was performed using the MrBayes plugin in PhyloSuite (Zhang et al. 2020).Two simultaneous runs of 5 000 000 generations were conducted for the matrix and trees were sampled every 1000 generations.Convergence and stability were evaluated in Tracer ver.1.7 ensuring effective sample size > 200 for all parameters (Rambaut et al. 2018).With the fi rst 25% of trees discarded as burn-in, Bayesian posterior probabilities were calculated for a 50% majority rule consensus tree of the remaining trees.Trees were visualized using FigTree ver.1.4.3 (Rambaut 2016).

Phylogenetic analyses of Exitianus Ball, 1929
Based on the sister-group relationship of Exitianus and Nephotettix recovered by the analysis of Zahniser & Dietrich (2015), Nephotettix virescens Distant, 1908 and Nephotettix nigropictus Stål, 1870 were selected as the outgroup for analysis of Chinese populations of Exitianus (Table 2).The Exitianus phylogenetic tree was reconstructed by combining one mitochondrial gene (COI) and one nuclear gene (ITS2).Phylogenetic reconstruction was conducted using BI analysis and ML analysis.The analyses were run the same as described above for the broader analysis if the tribe, except that two simultaneous runs of 1 000 000 generations were conducted for the matrix in Bayesian inference analysis.The most suitable substitution models and partition scheme for the combined dataset are listed in Table 5.

Species delimitation
Three independent single loci species delimitation methods without a priori taxonomic information were performed: jMOTU (Jones et al. 2011), ABGD (Puillandre et al. 2012), and bPTP (Zhang et al. 2013).In addition, a species delimitation method that requires priori taxonomic information was used, and multiple loci were used to verify the above results using Bayesian coalescent method in the software BPP (Yang 2015).jMOTU According to the differences in genetic distance of sequences, the jMOTU divides sequences into different molecular operational taxonomic units (MOTUs).Whenever the genetic distance difference of the sequence is less than the specifi ed threshold, it will be assigned into the same taxon.Before running jMOTU, we confi gured JAVA and set the paths of formatdb and megablast.The FASTA sequence was loaded into the software.Then based on the empirical sequence cut off values , the shortest sequence matching percentage (95%), and the lowest BLAST value (97%) were used for cluster analysis.The number of MOTUs was plotted against the threshold values to determine the number of species inferred from each value of the different distance thresholds (Jones et al. 2011).

ABGD
Based on the assumption of a barcoding gap (i.e., intraspecifi c divergences are smaller than interspecifi c divergences), the ABGD procedure fi rst sorts the data into hypothetical species, and then computes recursively based on previous groupings to obtain an optimal number of partitions.The graphical web version was used (http://www.abi.snv.jussieu.fr/public/abgd/abgdweb.html).The FASTA sequences were uploaded and the analysis used the following parameters: prior intraspecifi c divergence values (Pmin = 0.001 and Pmax = 0.1), relative gap width (X = 1.0) and the Kimura 2-P (K80) distance model.

bPTP
For the bPTP analysis, branch lengths represent the number of substitutions, not time, eliminating the problems associated with requiring a calibrated tree when a priori information on divergence time is not available.An ML tree for each gene was reconstructed using the IQ-TREE plugin in PhyloSuite ver.1.2.2 (Zhang et al. 2020).The bPTP analysis was carried out on the bPTP server (https://species.h-its.org/)with the ML tree as input, specifying an outgroup, considering 100 000 MCMC generations, a thinning value of 100, and a burn-in of 10%.

BPP
For the BPP analysis, the BI tree constructed from the above two gene datasets (COI, ITS2) was selected as the guide tree for BPP analysis.The data for N. virescens and N. nigropictus for the BPP analysis was included because the statistical power of BPP can be increased when closely related outgroups are included (Rannala & Yang 2013).The program BPP 3.3.was used to validate the delimitation results generated by the above methods (Yang 2015).The analysis was run for 1 000 000 generations, with a sample frequency of 1000.The fi rst 25% of trees were removed as burn-in.The convergence and stability were evaluated in Tracer ver.1.7 ensuring an effective sample size > 200 for all parameters (Rambaut et al. 2018).Species delimitation in BPP requires a priori estimation of two evolutionarily signifi cant parameters: ancestral population size (θ) and degree of divergence among species (τ).According to Yang's procedure, we used the following combinations: .All BPP analyses were run for 500 000 generations with sampling every fi ve generations after discarding an initial burn-in of 20 000 generations.For verifi cation, every analysis was run twice to check for convergence between runs and agreement on the posterior probability of the species delimitation models.

Data analyses
Trimmed alignments of four genes (COI, 16S, H3, 28S) used for phylogenetic analysis of the Chinese Chiasmini were 612 bp, 487 bp, 276 bp, and 625 bp, respectively.For the analysis of Exitianus, one mitochondrial gene (COI) and one nuclear gene (ITS2) were used.Trimmed alignments for two genes were 615 bp and 517 bp, respectively.C, V, PI, average nucleotide composition, haplotype number, Hd, and Pi are listed in Table 6.Both included species are represented by multiple haplotypes for at least one locus.
Tests for substitutional saturation indicated that none of four genes have reached saturation and are appropriate for use in Chiasmini phylogenetic analysis.The ILD tests indicated that the different loci have phylogenetic signal suffi ciently homogeneous to allow four gene fragments to be combined into a single concatenated alignment for analyses.Similar results were obtained for the two genes (COI and ITS2) used for analysis of Exitianus.

Phylogenetic analyses of Chiasmini Distant, 1908
The phylogenies obtained from ML and BI analyses (Figs 3-4) are highly similar, differing only in the relationship among three species of Doratura that received low branch support in both analyses ; most other branches had moderate to strong support by ML bootstrap and Bayesian posterior probabilities.All genera for which more than one representative was included are recovered as monophyletic and the ingroup topology is as follows: Doraturopsis + ((Exitianus + Nephotettix) + ((Zahniserius + (Chiasmus + Gurawa)) + (Doratura + Aconurella))).Within Chiasmini, the genus Doraturopsis was consistently sister to all other genera included Chiasmini.The ML and BI analyses grouped the remaining genera into two sister clades.In clade I, Exitianus and Nephotettix are sister groups with strong support (BS = 100, PP = 1) and relationships among species within each genus are well resolved with moderate to strong support (BS > 84, PP > 0.88).Clade II comprising the remaining genera also has strong support (BS = 92, PP = 1).The clade (Zahniserius + (Chiasmus + Gurawa)) is sister to Doratura and Aconurella.

Molecular species delimitation analyses of Exitianus Ball, 1929
The ML and BI analysis grouped the ingroup taxa into two identical monophyletic clades with maximum support (BS = 100, PP = 1).All analyses consistently support the monophyly of the two species E. indicus and E. nanus (Fig. 5).

Species delimitation jMOTU
For the COI dataset, sequence cut off values of 2-24, yielded three MOTUs with all individuals of E. nanus included in one MOTU, and E. indicus divided into two MOTUs.When the sequence cut off values are 25-30, only two MOTUs are recognized, corresponding to the morphological defi nitions of the two species.For the ITS2 dataset, the MOTUs are identical to the morphological species when the sequence cut off values are .

ABGD
For COI, the distance-based method ABGD recovers two molecular operational taxonomic units (MOTUs) corresponding to the morphologically defi ned species when the prior intraspecifi c divergence (P value) is 0.001624 to 0.100000.For ITS2, when the prior intraspecifi c divergence ranges from 0.001274 to 0.002069, seven MOTUs are recognized, including two for E. indicus and fi ve for E. nanus.When the prior intraspecifi c divergence is 0.002069, the ITS2 dataset places all individuals of the two species into a single MOTU.This indicates that this gene cannot distinguish the species defi ned according to traditional morphological criteria using this method (Figs 6-7).

bPTP
The bPTP analysis gives two results based on Bayesian and ML support, respectively.The ML method result corresponds to the PTP analysis result, and the Bayesian method corresponds to the bPTP analysis result.Both methods give the same results for the COI gene recognizing two MOTUs corresponding to the two morphologically defi ned species.ITS2 also yields two MOTUs based on the ML method, consistent with the morphological species.Based on Bayesian method, ITS2 generated excessive partitions.The ML results are shown for bPTP .

BPP
Phylogenetic trees from two-gene (COI, ITS2) analysis were used as inputs to create the BPP guide tree.The analysis was run twice.All results recognize two species with strong support (Table 7, Fig. 5).

Discussion
Aside from recent phylogenomic analyses that used data from large numbers of genes, most molecular phylogenetic studies of insects, including those focused on leafhoppers and other Auchenorrhyncha have been based on sequence fragments of the 16S rDNA, ND1, H3, 28S rDNA and COII genes.In general, these analyses have yielded phylogenies that agree well overall with both morphological taxon concepts and phylogenomic results (Fang et al. 1993;Zahniser & Dietrich 2010, 2013, 2015;Cao et al. 2022).The nuclear gene regions evolve more slowly and are, therefore, more helpful in solving the relationships among genera and tribes, while the mitochondrial genes evolve more quickly and better refl ect the relationships between species.
In this study, combining data from four genes consistently resolved relationships among genera and species of Chiasmini.For the genera Aconurella and Nephotettix, represented by multiple species, the interspecifi c relationships are well resolved.Nevertheless, the relationships among some genera within the tribe are inconsistent with those of previous studies (Zahniser & Dietrich 2015;Cao et al. 2022), with the main area of disagreement confi ned to the relationship of Doratura +Aconurella and Gurawa + Chiasmus, and the position of Doraturopsis.Because the present phylogenetic analysis included species of Chiasmini primarily from China, further analyses with a broader sample of taxa from other geographic regions, and data from additional gene regions, are needed to resolve relationships among genera and species and confi rm previous morphological species concepts.
Owing to high intraspecifi c variation in morphological characters traditionally used for species identifi cation and interspecifi c similarity, delimitation of Exitianus species using morphology alone has been challenging.Previous taxonomic studies of Exitianus focused only on morphology-based classifi cation, without assessing species boundaries using other types of data (Duan & Zhang 2013b).In our study, populations of two common and widespread species of this genus showed very low genetic variability, suggesting that gene fl ow among Chinese populations is high.This is not surprising given that Exitianus species are fully winged and disperse readily.The different molecular species delimitation methods yielded results mostly consistent with current morphology-based species defi nitions.These  Table 7. Posterior probabilities for the number of delimited species using different priors for model parameters in Bayesian phylogenetics and phylogeography on concatenated data sets of mitochondria markers and all genetic markers.
results are consistent with those of our previous studies on the chiasmine genera and Aconurella ( Gao et al. 2021;Yan et al. 2022).The validity of morphology-based species concepts in Chiasmini was also confi rmed for several Palearctic species of Doratura using data from courtship calls (Tishechkin 2011).
In general, we found that the different molecular species delimitation methods yield similar results, but differences sometimes occur depending on the particular parameter settings selected.For methods such as jMOTU incorporating a user-specifi ed cut-off for genetic divergence, more stringent cut-offs yield fewer MOTUs.Results can also vary depending on the particular gene region selected, as we observed between COI and ITS2 for some methods.This accentuates the need to consider multiple sources of evidence when attempting to delimit species.Our results demonstrate that species delimitation analyses based on multiple loci give a more credible and consistent result than methods using a single locus and are more congruent with morphological differences (Gao et al. 2021;Yan et al. 2022).

Conclusions
For the ML and BI trees constructed from 8 genera and 20 species of Chinese Chiasmini combined with four gene fragments (COI, 16S, H3, 28S), a phylogenetic relationship of (Doraturopsis + ((Exitianus + Nephotettix) + ((Zahniserius + (Chiasmus + Gurawa)) + (Doratura + Aconurella))) was obtained.For the molecular species delimitation analyses of Chinese Exitianus, the ML result and BI result both show that E. indicus and E. nanus are well-supported independent clades that are highly genetically divergent from each other but with low intraspecifi c genetic variability.
Based on the COI gene, the single-locus species delimitation methods jMOTU, ABGD and bPTP divided the included Exitianus populations into two species, consistent with their defi nition as morphological species.For ITS2 gene, jMOTU recognized a two species result while ABGD grouped populations into a single species.The multi-locus BPP method consistently delimited the two species as previously defi ned based on morphology and appears more robust.However, this needs to be further tested by including a broader sample of Exitianus populations occurring outside of China.

Fig. 3 .
Fig. 3. ML bootstrap consensus phylogenetic tree for Chiasmini Distant, 1908 based on concatenated sequence data from two mitochondrial genes (COI, 16S) and two nuclear genes (H3, 28S).Bootstrap support of nodes is indicated below the branches.Note the proposed taxonomic arrangements shown by vertical coloured lines.

Fig. 4 .
Fig. 4. BI consensus phylogenetic tree for Chiasmini Distant, 1908 based on concatenated sequence data from two mitochondrial genes (COI, 16S) and two nuclear genes (H3, 28S).Posterior probabilities of nodes are indicated below the branches.Note the proposed taxonomic arrangements shown by vertical coloured lines.

Fig. 5 .
Fig. 5. ML bootstrap and BI consensus phylogenetic tree for Exitianus Ball, 1929 based on two-gene data set (COI and ITS2).Bootstrap support and posterior probabilities of nodes are indicated above the branches.The right vertical bars indicate the putative species using BPP.Morphological species are uniquely coloured.

Fig. 6 .
Fig. 6.Phylogenetic tree for Exitianus Ball, 1929 based on mitochondrial COI.Bootstrap support and posterior probabilities of nodes are indicated above the branches.The right vertical bars indicate the number of putative species using various methods as indicated at the top.Morphological species are uniquely coloured.

Fig. 7 .
Fig. 7. Phylogenetic tree for Exitianus Ball, 1929 based on ribosome ITS2.Bootstrap support and posterior probabilities of nodes are indicated above the branches.The right vertical bars indicate the number of putative species using various methods as indicated at the top.Morphological species are uniquely coloured.

Table 1 .
Specimens ofChiasmini Distant, 1908used in this study (all from China), with morphological identifi cations, collecting localities and GenBank accession numbers.

Table 2 .
Specimens of Exitianus Ball, 1929used in this study (all from China), with morphological identifi cations, collecting localities and GenBank accession numbers.

Table 3 .
Primer sequences for PCR amplifi cation and sequencing.

Table 5 .
Best partitioning scheme and models for two genes selected to construct the phylogenetic tree of Exitianus Ball, 1929 by the PartitionFinder.

Table 4 .
Best partitioning scheme and models for four genes selected to construct the phylogenetic tree of Chiasmini Distant, 1908 by the PartitionFinder.