Phylogenetics, Molecular Species Delimitation and Geometric Morphometrics of All Reddish-Brown Species in the Genus Neotriplax Lewis, 1887 (Coleoptera: Erotylidae: Tritomini)

Simple Summary Due to the instability of interspecific morphological characteristics and the significant convergence phenomenon, it is difficult to identify most species of Neotriplax based on morphology. Through integrated research on the reddish-brown species of Neotriplax, which feature many phenotypic similarities, and by comprehensively utilizing morphology, molecular phylogeny, and geometric morphometrics, the relationship among all these species was reconstructed. The taxonomic status within the genus was partly clarified, with three new species being identified: N. qinghaiensis sp. nov., N. maoershanensis sp. nov., and N. guangxiensis sp. nov. Abstract To date, five species of reddish-brown Neotriplax have been described, but their highly similar body color and other phenotypic traits make accurate taxonomy challenging. To clarify species-level taxonomy and validate potential new species, the cytochrome oxidase subunit I (COI) was used for phylogenetic analysis and the geometric morphometrics of elytron, pronotum, and hind wing were employed to distinguish all reddish-brown Neotriplax species. Phylogenetic results using maximum likelihood and Bayesian analyses of COI sequences aligned well with the current taxonomy of the Neotriplax species group. Significant K2P divergences, with no overlap between intra- and interspecific genetic distances, were obtained in Neotriplax species. The automatic barcode gap discovery (ABGD), assemble species by automatic partitioning (ASAP), and generalized mixed Yule coalescent (GMYC) approaches concurred, dividing the similar species into eight molecular operational taxonomic units (MOTUs). Geometric morphometric analysis using pronotum, elytron, hind wing shape and wing vein patterns also validated the classification of all eight species. By integrating these analytical approaches with morphological evidence, we successfully delineated the reddish-brown species of Neotriplax into eight species with three new species: N. qinghaiensis sp. nov., N. maoershanensis sp. nov., and N. guangxiensis sp. nov. Furthermore, we documented the first record of N. lewisii in China. This study underscores the utility of an integrative taxonomy approach in species delimitation within Neotriplax and serves as a reference for the taxonomic revision of other morphologically challenging beetles through integrative taxonomy.


Introduction
Erotylinae is the largest subfamily in Erotylidae, encompassing five tribes and more than 2600 species [1][2][3].The adults and larvae feed on fungi, and some of them are Insects 2024, 15, 508 3 of 25 currently deposited at the Museum of Hebei University (MHBU), the Institute of Zoology, the Chinese Academy of Sciences (IZAS), and the Hebei Agricultural University (HEBAU).

Morphological Comparison and Terminology
Habitus photographs of specimens were captured using the Olympus E-M5II camera (OLYMPUS, Beijing, China) for stacked photography.The detailed photographs of mouthparts and genitalia were illustrated using an α7RIII SONY camera (SONY, Tokyo, Japan).Final plates were enhanced using Adobe Photoshop CS6.0 (Adobe Systems Inc., San José, CA, USA).Morphological terminology follows Lawrence et al. [48,49].

Genomic DNA Extraction, Sequence Amplification and Data Analysis
A DNeasy Blood & Tissue kit (TIANGEN, Beijing, China) was used to extract the isolated genomic DNA.DNA quality and concentration were measured on the Nanodrop2000 (ND-2000) spectrophotometer (Thermo fisher, USA) and assessed through electrophoresis in a 1% agarose gel.
A fragment of the COI gene has been proven to be a standard DNA barcode region for animal species identification and the detecting of molecular operational taxonomic units (MOTUs) [50], which can be widely used in different fields (for example, species delimitation [51] and cryptic diversity discovering [52]).The primers used to amplify mt DNA COI were synthesized by the Biomed company (Beijing, China) and are listed in Table S2 [53,54].The COI was amplified using the 40 µL system: 2×Taq PCR Mix 20 µL, forward and reverse primer 1 µL, template DNA 2 µL, ddH 2 O 16µL.The reaction procedure at the system of COI was pre-denaturation at 94 • C for 4 mins, denaturation at 94 • C for 30 s, annealing at 52 • C for 1 min, extension at 72 • C for 1 min, 35 cycles and extension at 72 • C for 7 min.The sequencing results were spliced and corrected using Chromas v2.6.6.A Nucleotide BLAST on the NCBI (https://www.ncbi.nlm.nih.gov/)(accessed on 1 Jun 2023) was used to test the accuracy of sequence.
Clustal_X [55] was used to conduct DNA alignment from the amino acid alignment of COI.Dacne picta (Erotylidae: Dacnini) was selected as outgroup [56].The maximun likelihood (ML) method was conducted with IQ-Tree v1.6.8 in Phylosuite to construct the phylogenetic tree [57,58].The TIM2 + F + I + G4 model was selected by ModelFinder.A model of nucleotide evolution was selected using the Akaike Information Criterion (AIC) [59].Branch support was evaluated using the ultra-fast bootstrapping method with 1000 replicates [60].Phylogenetic analysis was conducted using MrBayes v3.2, run with a total of 2 million generations, sampled every 100 generations, and cut off with 20% of the sampled trees [61].Two Markov chain Monte Carlo (MCMC) chains were employed on DNA sequence alignment with the GTR + F + I + G4 model.FigTree v1.4.3 [62] was used to view and illustrate the inferred phylogenetic trees.A Kimura 2-parameter (K 2 P) genetic distance model was used to calculate the intra-and interspecific genetic divergences [63,64].One way analysis of variance (ANOVA) and LSD multiple comparison were used to detect pairwise differences.

Geometric Morphometrics Analysis
An Olympus E-M5II camera (OLYMPUS, Beijing, China) was utilized to capture twodimensional images of the pronotum, elytron and wing.To ensure accurate geometric morphometric analysis, it was crucial to maintain consistent placement position and shooting orientation throughout the photography process.The specimens used for geometric morphometrics are shown in Table S1.
Geometric morphometric information was obtained from the left contour of the pronotum, elytra, and wing.The curve of the pronotum was taken from the midpoint of the anterior margin to the midpoint of the posterior margin.The curve of the elytron was taken from the boundary between the anterior margin of the elytron and the apex of the scutellar shield to the apex of the laterior margin of the elytron.The curves for the pronotum or elytron were resampled by length after 50 and 100 semi-landmarks (SLM), respectively, to ensure that their morphological information could be extracted completely (Figure 1).The curve of the wing shape was taken from the costal margin humeral angle, punctuated in a clockwise direction, and it ended at the apex of the inner margin, which was resampled by length after 100 SLM.All curves were digitized by the software TpsUtil v1.46 and Tps-Dig v2.12 [69][70][71].For the wing vein, the intersection points and endpoints of each pulse were selected as the landmark, creating a total of 19 landmarks (Figure 2).

Geometric Morphometrics Analysis
An Olympus E-M5II camera (OLYMPUS, Beijing, China) was utilized to capture twodimensional images of the pronotum, elytron and wing.To ensure accurate geometric morphometric analysis, it was crucial to maintain consistent placement position and shooting orientation throughout the photography process.The specimens used for geometric morphometrics are shown in Table S1.
Geometric morphometric information was obtained from the left contour of the pronotum, elytra, and wing.The curve of the pronotum was taken from the midpoint of the anterior margin to the midpoint of the posterior margin.The curve of the elytron was taken from the boundary between the anterior margin of the elytron and the apex of the scutellar shield to the apex of the laterior margin of the elytron.The curves for the pronotum or elytron were resampled by length after 50 and 100 semi-landmarks (SLM), respectively, to ensure that their morphological information could be extracted completely (Figure 1).The curve of the wing shape was taken from the costal margin humeral angle, punctuated in a clockwise direction, and it ended at the apex of the inner margin, which was resampled by length after 100 SLM.All curves were digitized by the software TpsUtil v1.46 and Tps-Dig v2.12 [69][70][71].For the wing vein, the intersection points and endpoints of each pulse were selected as the landmark, creating a total of 19 landmarks (Figure 2).TpsRelw v1.74 (with the sliders file included) was used to save the aligned file [72].The shape differences among these taxa were inferred using principal component analysis (PCA) and canonical variate analysis (CVA) in MorphoJ v1.07 [73].MorphoJ v1.07 [73] was used to calculate the Procrustes distance and Mahalanobis distance between the morphological data of each group to evaluate the degree of difference between the average morphologies of each group.TpsRelw v1.74 (with the sliders file included) was used to save the aligned file [72].The shape differences among these taxa were inferred using principal component analysis (PCA) and canonical variate analysis (CVA) in MorphoJ v1.07 [73].MorphoJ v1.07 [73] was used to calculate the Procrustes distance and Mahalanobis distance between the morphological data of each group to evaluate the degree of difference between the average morphologies of each group.

Phylogenetics of the Genus of Neotriplax
Phylogenetic analyses were performed on the mt DNA COI.A total of 39 COI sequences representing nine species (including the outgroup Dacne picta) were included in the final dataset.The COI alignment (585 bp; 33.4% T, 19.7% C, 30.6% A, 16.3% G) included 346 conserved sites, 239 variable sites, and 183 parsimony informative sites.The results of the phylogenetic analyses based on BI and ML reached the same topology, and almost all the nodes were highly supported (Figure 3).The phylogenetic analyses of Neotriplax revealed that all groups were defined as monophyletic with high ultrafast bootstrap support (BS) and posterior probabilities (PP), and it can be seen that these species can be well divided.All the species were divided into eight clades (8 species), including five known species and three new species, N. qinghaiensis sp.nov., N. maoershanensis sp.nov.and N. guangxiensis sp.nov.The first clade (N.miwai-N.lewisii) was the sister to all the other species.The remaining taxa constituted three clades.N. rubens occurred as the sister group with the remaining other taxa.The second clade contained N. maoershanensis sp.nov., occurring as sister group with N. qinghaiensis sp.nov.and N. minima.The crown clade consists of N. arisana and N. guangxiensis sp.nov.The phylogenetic analyses of Neotriplax revealed that all groups were defined as monophyletic with high ultrafast bootstrap support (BS) and posterior probabilities (PP), and it can be seen that these species can be well divided.All the species were divided into eight clades (8 species), including five known species and three new species, N. qinghaiensis sp.nov., N. maoershanensis sp.nov.and N. guangxiensis sp.nov.The first clade (N.miwai-N.lewisii) was the sister to all the other species.The remaining taxa constituted three clades.N. rubens occurred as the sister group with the remaining other taxa.The second clade contained N. maoershanensis sp.nov., occurring as sister group with N. qinghaiensis sp.nov.and N. minima.The crown clade consists of N. arisana and N. guangxiensis sp.nov.

Genetic Distances and Species Delimitation
The Kimura-2-parameter (K 2 P) distances between different species of Neotriplax ranged from 0.1256 to 0.2084, with the smallest distance being between N. guangxiensis sp.nov.and N. arisana and the largest distance being between N. qinghaiensis sp.nov.and N. miwai.All K 2 P distances between the identified species were greater than 0.1256 (Table 1).The results showed that the average inter-GD ranged from 0.0034 (N.lewisii) to 0.0333 (N.arisana), with no overlap with the genetic distances of interspecies, and the difference was obvious (Table 1).
The result of the ABGD was generated by DNA barcode, and the initial partition and recursive partition were obtained through K 2 P model analysis, as shown in Figure S1.Refer to Figure 3 for species division results.It was divided into eight molecular operational taxonomic units (MOTUs), which was consistent with morphology and phylogeny.The result of ASAP showed that species were divided into eight MOTUs based on the ASAPscore (Treshold dist.= 0.081), which was consistent with those obtained by the ABGD method.Both the results based on distance distribution supported the classification status of these new and known reddish-brown species in Neotriplax.The GMYC analysis based on DNA barcoding yielded eight ML clusters (without outgroup), a confidence interval of 3-10, supporting eight morphospecies, and the results were consistent with the topology of ML and BI (Figure S2).For the analysis of bPTP, these specimens formed 15 MOTUs, and N. guangxiensis sp.nov.and N. maoershanensis sp.nov.formed two MOTUs, respectively.N. rubens formed three MOTUs and N. arisana formed four MOTUs (Figure 3).However, all of these showed an indistinct boundary in ABGD, ASAP and GMYC.
Therefore, significant genetic differentiation had occurred between the species of Neotriplax, and DNA barcoding could better distinguish different species in the genus.

Geometric Morphometrics Analyses of Pronotum across Groups
The PCA of the pronotum showed that the first two components accounted for 67.789% of the total variation in Neotriplax (the first principal component, PC1: 51.036%; the second principal component, PC2: 16.753%) (Figure 4a).The overall shape change trend of the pronotum can be obtained from the shape variation map obtained by the PCA (Figure 4b,c).PC1 changed in the positive direction.The anterior border and posterior angle had a trend of extending outwards, while the anterior angle and posterior border had a trend of shrinking inwards.Overall, the aspect ratio of the pronotum increased; PC2 changed in the positive direction.The lateral and posterior border had a trend of shrinking inwards, while the anterior border and the posterior angle had a trend of extending outwards.Overall, the aspect ratio of the pronotum decreased.The results showed that all taxa clustered together with less morphological variation (Figure 5a).The vector space constructed by the two canonical variates (CV1, CV2) with the largest proportion was employed to test the canonical variate of the pronotum contour shape variation (Figure 5b).N. qinghaiensis sp.nov.and N. minima were clearly separated from other taxa without overlapping.Both N. arisana and N. guangxiensis sp.nov.showed a minor overlap with the remaining taxa but can still be distinguished with others.However, there are many overlaps between N. rubens, N. lewisii, N. maoershanensis sp.nov., and N. miwai which cannot be distinguished.The vector space constructed by the two canonical variates (CV1, CV2) with the largest proportion was employed to test the canonical variate of the pronotum contour shape variation (Figure 5b).N. qinghaiensis sp.nov.and N. minima were clearly separated from other taxa without overlapping.Both N. arisana and N. guangxiensis sp.nov.showed a minor overlap with the remaining taxa but can still be distinguished with others.However, there are many overlaps between N. rubens, N. lewisii, N. maoershanensis sp.nov., and N. miwai which cannot be distinguished.In terms of statistical test parameters (Table S3), the Mahalanobis distance of N. maoershanensis sp.nov.and N. miwai was the smallest, N. miwai and N. minima was the largest, at 6.1333 and 18.2966, respectively, and the difference was extremely significant (p < 0.01).N. qinghaiensis sp.nov.and N. lewisii had the smallest Procrustes distance and N. minima had the largest Procrustes distance, at 0.0206 and 0.0756, respectively, and the difference was extremely significant (p < 0.01).

Geometric Morphometrics Analyses of Elytron across Group
The PCA of elytron showed that the first two components accounted for 74.530% of the total shape variation in the elytron of Neotriplax (PC1:63.858%;PC2:10.671%)(Figure 6a).The change trend of the contours of the elytron can be obtained from the shape variation map obtained by the PCA of the elytron (Figure 6b,c).PC1 changes in the positive direction.The scutellum, anterior, and posterior border of the elytron had a trend of shrinking inwards, while the lateral border had a trend of extending outwards.The shape change trend was not large, and the aspect ratio decreased; PC2 changed in the positive direction.The scutellum and lateral border of the elytron had a trend of extending outwards, while the anterior and posterior border had a trend of shrinking inwards.The shape variation was not obvious.The results showed that all taxa clustered together with less morphological variation (Figure 7a).Based on the canonical variates of CV1 and CV2, all species were clearly separated from other taxa without overlapping.(Figure 7b).
In terms of statistical test parameters (Table S4), the Mahalanobis distance between N. qinghaiensis sp.nov.and N. guangxiensis sp.nov.was the smallest, and the Mahalanobis distance between N. minima and N. rubens was the largest, at 13.4801 and 31.9947,respectively, and the difference was extremely significant (p < 0.01).N. rubens and N. maoershanensis sp.nov.had the smallest Procrustes distance, and N. qinghaiensis sp.nov.and N. minima had the largest Procrustes distance, at 0.0084 and 0.0561, respectively, and the difference was extremely significant (p < 0.01).In terms of statistical test parameters (Table S3), the Mahalanobis distance of N. maoershanensis sp.nov.and N. miwai was the smallest, N. miwai and N. minima was the largest, at 6.1333 and 18.2966, respectively, and the difference was extremely significant (p < 0.01).N. qinghaiensis sp.nov.and N. lewisii had the smallest Procrustes distance and N. minima had the largest Procrustes distance, at 0.0206 and 0.0756, respectively, and the difference was extremely significant (p < 0.01).

Geometric Morphometrics Analyses of Elytron across Group
The PCA of elytron showed that the first two components accounted for 74.530% of the total shape variation in the elytron of Neotriplax (PC1:63.858%;PC2:10.671%)(Figure 6a).The change trend of the contours of the elytron can be obtained from the shape variation map obtained by the PCA of the elytron (Figure 6b,c).PC1 changes in the positive direction.The scutellum, anterior, and posterior border of the elytron had a trend of shrinking inwards, while the lateral border had a trend of extending outwards.The shape change trend was not large, and the aspect ratio decreased; PC2 changed in the positive direction.The scutellum and lateral border of the elytron had a trend of extending outwards, while the anterior and posterior border had a trend of shrinking inwards.The shape variation was not obvious.The results showed that all taxa clustered together with less morphological variation (Figure 7a).Based on the canonical variates of CV1 and CV2, all species were clearly separated from other taxa without overlapping.(Figure 7b).
In terms of statistical test parameters (Table S4), the Mahalanobis distance between N. qinghaiensis sp.nov.and N. guangxiensis sp.nov.was the smallest, and the Mahalanobis distance between N. minima and N. rubens was the largest, at 13.4801 and 31.9947,respectively, and the difference was extremely significant (p < 0.01).N. rubens and N. maoershanensis sp.nov.had the smallest Procrustes distance, and N. qinghaiensis sp.nov.and N. minima had the largest Procrustes distance, at 0.0084 and 0.0561, respectively, and the difference was extremely significant (p < 0.01).

Geometric Morphometrics Analyses of Wing Shape and Wing Vein across Group
The PCA of the wing shape and wing vein showed that the first two components accounted for 81.432% of the total shape variation in the elytron of Neotriplax (PC1: 64.364%; PC2: 17.068%) (Figure 8a).The change trend of the contours of the wing shape and the landmarks of the wing vein (Figure 8b,c) can be obtained from the shape variation map obtained by the PCA.PC1 changed in the positive direction.Points 1-15, 18, 19 tended to move in the opposite direction to the base of the wing, while points 16, 17 tended to move towards the base of the wing.The costal margin humeral angle and the outer margin of the hind wing had a trend of extending outwards, and the costal margin half and inner margin had a trend of shrinking inwards.PC2 changed in the positive direction.Points 1-19 tended to move in the opposite direction to the base of the wing.The outer margin had a trend of shrinking inwards, while the costal and inner margin, as well as the apical angle, tended to expand outwards.The results showed that all taxa clustered together with less morphological variation (Figure 9a).PC2 changed in the positive direction.Points 1-19 tended to move in the opposite direction to the base of the wing.The outer margin had a trend of shrinking inwards, while the costal and inner margin, as well as the apical angle, tended to expand outwards.The results showed that all taxa clustered together with less morphological variation (Figure 9a).Based on the canonical variate of CV1 and CV2, all species were clearly separated from other taxa without overlapping (Figure 9b).In terms of statistical test parameters (Table S5), the Mahalanobis distance between N. maoershanensis sp.nov.and N. rubens was the smallest, and the Mahalanobis distance between N. minima and N. guangxiensis sp.nov.was the largest, at 14.8569 and 43.7562, respectively, and the difference was extremely significant (p < 0.01).N. guangxiensis sp.nov.and N. maoershanensis sp.nov.had the smallest Procrustes distance, and N. miwai and N.arisana had the largest Procrustes distance, at 0.0189 and 0.0815, respectively.
Taking into account all three results, all eight morphospecies have a large degree of shape variation and can be effectively differentiated.

Taxonomy of Neotriplax
Through morphological research, geometric morphometric analysis, genetic distance, species delimitation, and phylogenetic construction, the taxonomic status of reddish-brown Neotriplax were clarified, and N. guangxiensis sp.nov., N. maoershanensis sp.nov.and N. qinghaiensis sp.nov.were determined as new species.Our taxonomic study confirmed the accuracy of diagnostic characteristics that were useful on a species level; for similar species of this genus, valid and objective diagnostic characteristics include the presence or absence of coxal lines, the shape and aspect ratio of the pronotum, the shape of the mentum and submentum, the shape of the prosternal process, and the shape of the scutellar shield.
3. Body with the sides less rounded; maxillary terminal palpomere transverse subtriangular, with the sides distinctly rounded…N.atrata.
-Body with the sides a little more rounded; maxillary terminal palpomere triangular, with the inner margin evenly but very slightly arched and the apical angle rather acute…N.delkskampi.
4. General color black, antennae deeply reddish-brown, labial yellowish-brown; elytron humerus with red patches…N.biplagiata.Based on the canonical variate of CV1 and CV2, all species were clearly separated from other taxa without overlapping (Figure 9b).
In terms of statistical test parameters (Table S5), the Mahalanobis distance between N. maoershanensis sp.nov.and N. rubens was the smallest, and the Mahalanobis distance between N. minima and N. guangxiensis sp.nov.was the largest, at 14.8569 and 43.7562, respectively, and the difference was extremely significant (p < 0.01).N. guangxiensis sp.nov.and N. maoershanensis sp.nov.had the smallest Procrustes distance, and N. miwai and N.arisana had the largest Procrustes distance, at 0.0189 and 0.0815, respectively.
Taking into account all three results, all eight morphospecies have a large degree of shape variation and can be effectively differentiated.

Taxonomy of Neotriplax
Through morphological research, geometric morphometric analysis, genetic distance, species delimitation, and phylogenetic construction, the taxonomic status of reddish-brown Neotriplax were clarified, and N. guangxiensis sp.nov., N. maoershanensis sp.nov.and N. qinghaiensis sp.nov.were determined as new species.Our taxonomic study confirmed the accuracy of diagnostic characteristics that were useful on a species level; for similar species of this genus, valid and objective diagnostic characteristics include the presence or absence of coxal lines, the shape and aspect ratio of the pronotum, the shape of the mentum and submentum, the shape of the prosternal process, and the shape of the scutellar shield.
-Body with the sides a little more rounded; maxillary terminal palpomere triangular, with the inner margin evenly but very slightly arched and the apical angle rather acute. ..N. delkskampi.

7.
Posterior border of pronotum strongly narrowed backwards, with a narrow lobe in middle. ..N. minima.
-Posterior border of pronotum slightly narrowed backwards, with a broad lobe in middle. ..8.
-Mentum with pentagon in bump, and middle area triangularly depressed. ..10.Comparative notes.All specimens are similar to the type specimen, but considerable interspecific variability of N. lewisii was observed in the body size (Figure 10).
Mesoventrite transverse, smooth and with punctures.Metaventrite coarsely punctured at sides and indistinctly in middle.Postmesocoxal lines absent.Abdomen with dense and fine punctures.Postmetacoxal lines absent.
Legs long and fine, femora slightly emarginated at the inside of apex, tibiae slightly transverse at the apex, but not triangular, the 1-3 segments of tarsus gradually widening to apex.
Male genitalia median lobe curved, apex narrow and about 1.1× as long as median strut (Figure 13a).
Remarks.The body shape and color of this species are similar to N. lewisii.This new species can be identified by its slightly narrower body, with prosternal lines, postmesocoxal lines, and postmetacoxal lines absent.N. lewisii has a broader body, with only postmesocoxal lines absent, while prosternal lines and postmetacoxal lines are present.Distribution: China (Qinghai).Etymology.The specific name is derived from the type locality, Qinghai Province, China.Pronotum transverse, nearly 1.96x as wide as long, convex dorsally, finely punctured.Anterior margin with narrow and complete marginal border, lateral margins with distinctly marginal border.Anterior angles blunt tip, posterior angles almost rectangular.Prosternum with coarse punctures in middle, anterior margin slightly emarginated with narrow marginal border, the middle of base slightly emarginated.Prosternal lines absent.Scutellar shield large, nearly semicircle, arc at the apex, without punctures.Elytra long, 1.5x as long as wide, more convex dorsally than the pronotum, widest at the base 1/3.
Mesoventrite transverse, smooth and with punctures.Metaventrite coarsely punctured at sides and indistinctly in middle.Postmesocoxal lines absent.Abdomen with dense and fine punctures.Postmetacoxal lines absent.
Legs long and fine, femora slightly emarginated at the inside of apex, tibiae slightly transverse at the apex, but not triangular, the 1-3 segments of tarsus gradually widening to apex.
Pronotum nearly trapezoidal, 2× as wide as long, with fine and uniform punctures and denser than head.Anterior margin shallowly bisinuate, lateral margins straight in basal half, converges forward in middle, with distinct and complete marginal border, basal margin with complete marginal border, weakly bisinuate.Anterior angles acute, protruded and blunt, posterior angles blunt, each with one pore.Prosternum with fine and sparse punctures, anterior border produced to short point in middle, with distinct marginal border.Prosternal process with depression at the apical emargination and with one pore at each side.Prosternal lines short, converge slightly forward.Scutellar shield nearly semicircle, round posteriorly.Elytra long and wide, 1.3× as wide as long, with eight striae, internals extremely fine punctures, widest at the base 1/6.Pronotum nearly trapezoidal, 2× as wide as long, with fine and uniform punctures and denser than head.Anterior margin shallowly bisinuate, lateral margins straight in basal half, converges forward in middle, with distinct and complete marginal border, basal margin with complete marginal border, weakly bisinuate.Anterior angles acute, protruded and blunt, posterior angles blunt, each with one pore.Prosternum with fine and sparse punctures, anterior border produced to short point in middle, with distinct marginal border.Prosternal process with depression at the apical emargination and with one pore at each side.Prosternal lines short, converge slightly forward.Scutellar shield nearly semicircle, round posteriorly.Elytra long and wide, 1.3× as wide as long, with eight striae, internals extremely fine punctures, widest at the base 1/6.
Mesoventrite transverse, each side with one shallow depression, with coarse punctures.Metaventrite with coarse and shallow punctures.Postmesocoxal lines extending to basal 1/2 of metaventrite.Abdomen with dense punctures, postmetacoxal lines long, and extending to basal 3/4 of ventrite 1.
Male genitalia (Figure 13b) median lobe weakly curved, gradually narrow from base to apex; median strut straight, 0.9× as long as median lobe.
Body length: 6.8-7.5 mm; width: 3.7-4.5 mm.Remarks.This new species is similar to N. arisana in terms of body color and shape.It is distinguished by its narrow triangular mentum, narrower submentum, and yellow maxillary and labial palpomeres.In contrast, N. arisana has a pentagonal mentum, broader submentum, and black maxillary and labial palpomeres.
Distribution: China (Guangxi).Etymology.The specific name is derived from the type locality.
Pronotum slightly trapezoidal, convex dorsally, 1.8x as wide as long, with fine and densely punctures.Anterior margin with narrow and completed marginal border, shallowly forward projection in middle; lateral margins broadly rounded, converges forward, with distinct marginal border.Basal border weakly sinuate.Anterior angles acute, blunt tip, posterior angles blunt.Prosternum impunctate laterally, with sparse punctures medially, prosternal process with one indistinctly pore at each side.Prosternal lines long, converge forward.Scutellar shield large, nearly heart-shaped, angulate posteriorly, with fine punctures.Elytra long, 1.3× as long as wide, with eight striae, internals extremely fine punctures.
Mesoventrite transverse, with coarse punctures.Metaventrite with dense and coarse punctures on each side of base and without punctures in middle.Postmesocoxal lines long, extending to basal 1/2 of metaventrite.Abdomen with dense and fine punctures.Postmetacoxal lines extending to basal 1/2 of ventrite 1.
Male genitalia (Figure 13c) median lobe curved, nearly equal in length to median lobe.Female genitalia (Figure 14c) stylus narrow from base to end; female spermatheca nearly oval.
Body length: 5.0-5.2mm; width: 2.6-3.1 mm.Remarks.This species is similar to N. arisana in terms of body color and shape.It is distinguished by the lateral margin of the pronotum and is strongly curved from the base 1/2 to top.The anterior angle is protruded and blunt; the mentum is upturned on both sides, triangular in bump, and the submentum is broad.In contrast to the new species, N. arisana has its lateral margin of the pronotum straight.The mentum is flat on both sides, pentagon in bump, submentum width centering.
Etymology.The specific name is derived from the type locality, Guangxi Province, China.

Discussion
An integrative taxonomy analysis was firstly provided by comprehensively utilizing morphology, molecular phylogeny, and geometric morphometrics, which supported the relationship between all known and new reddish-brown species by integrative taxonomy.Including the three new species described here, Neotriplax now comprise 12 species.The results have increased the biodiversity of this genus within China and worldwide.
Currently, the application of integrative taxonomy in the insect species delimitation has achieved remarkable results, such as in the species delimitation of Bactrocera dorsalis, Machilis, Encyrtus sasakii, and other insects [74][75][76].This comprehensive taxonomy approach has been proven to provide a reference for distinguishing closely related species, which deepen the understanding and knowledge of the boundaries between species.

Phylogeny of Neotriplax
This study provides the first insight into the phylogenetic relationships within the genus Neotriplax.In the phylogenetic tree constructed using the ML and BI methods, all these species are divided into eight clades, and all of them are monophyletic.Our study supports the current taxonomic status of reddish-brown species in Neotriplax in terms of phylogeny, which is consistent with the results of the morphological research [7,8,13].Among them, N. arisana and N. guangxiensis sp.nov.have a close phylogenetic relationship, and they also resemble each other morphologically.
DNA barcoding has been proven to be a powerful method for rapid species identification [77], and our study further validates its application in Neotriplax.These similar species showed an obvious barcoding gap.For the first time, we analyzed the inter-and intraspecific genetic distance threshold of Neotriplax, considered as a crucial basis for species delimitated [78][79][80][81], revealing that the minimum inter-GD was significantly larger than the maximum intra-GD and the threshold of genetic distance was set at 0.1256 can be used to distinguish between species.

Molecular Species Delimitation
The four approaches of molecular species delimitation showed two different results.The methods ABGD, ASAP, and GMYC unanimously recognized eight MOTUs, aligning with the morphospecies.However, the bPTP approach revealed a discrepancy, identifying a total of 15 MOTUs.Obviously, our results indicate that the bPTP method yields more putative species than suggested by other methods and morphology-based classification.This method has previously shown a tendency to over-split species [82].The bPTP method relies on ML trees; species delimited with a well-defined morphology may be more accurate than with GMYC [28,83].Therefore, a comprehensive consideration using multi-gene and multi-method approaches is needed.These results proved that morphological and molecular species definition can be effectively combined to improve the accuracy of species classification and provide data support for molecular species definition and integrated taxonomic research on this genus.

Geometric Morphometrics
Geometric morphometrics can be used for the rapid identification of insects at lower taxonomic levels [84,85].First, the pronotum and elytron, which have rich morphological diversity and powerful functions, are considered continuous characteristics [86][87][88].The geometric morphometrics of the pronotum and elytra have been well applied in the identification of related genera and species of Coleoptera, the classification of geographical populations, and the study of biological evolution [88].However, in Neotriplax, the continuous characteristics of the pronotum did not distinguish all similar species, indicating that the shape variation in the pronotum is not applicable in the geometric morphological analysis of Neotriplax.This finding is consistent with the previous results, suggesting the shape variation in the pronotum is not reliable for the classification of lower-level taxa [89].
Subsequently, this study further selected the hind wings for geometric morphological analysis.Hind wing characteristics, such as wing shape and wing veins, are important in taxonomic and phylogenetic analyses [90][91][92].Combining the shape variations in the wing shape and wing veins, we can distinguish closely related species, which also supports the accuracy of the status of these species.This also shows that the shape variations in the hind wing can be used as a taxonomic basis to distinguish Neotriplax species, ensuring that all these species are correctly classified.
Since genitalia can reflect information about insect mating behavior, evolutionary history, and other aspects, they can also be used as an auxiliary basis in some specific cases [93,94].However, due to the limitation of the number of unisexual samples, it has increased the difficulty of comparative morphological studies.In future work, we need to further collect samples and conduct comparative morphological analysis of genitalia.
Geometric morphometrics provides a valuable reference for the classification of the genus.Moreover, it significantly improves the accuracy of identifying closely related and morphologically similar species.By comprehensively analyzing quantitative morphological and molecular data, more reliable classification results can be obtained through integrative taxonomic studies.

Integrated Taxonomy
In the process of species identification, adopting different classification methods often leads to differences in classification results.This is primarily due to the different perspectives and focuses of different classification methods [95].
Morphology is the foundation of species diversity research, providing support for the discovery, description, and identification of species [96].The characteristics of geometric morphometrics lie in its ability to eliminate factors such as the size, color, and so on, focusing only on the difference of shape variations.Moreover, it can quantitatively analyze continuous traits, which offers great advantages in its application in the study of species delimitation among closely related species [46,89].However, the phenotype of species is influenced by both genetic and environmental factors, which increases the uncertainty in the identification [95,97].
Molecular analysis not only provides strong technical support for species delimitation but also offer a new technical reference for the discovery of new species and cryptic species [98].However, while molecular techniques exhibit unique advantages in species delimitation, primarily focusing on revealing genetic differences within and between species, there are still potential limitations, and relying solely on molecular data for species delimitation may lead to underestimation or overestimation of species diversity [81].
Integrated taxonomy aims to provide more comprehensive and accurate species identification results by comprehensively utilizing multiple taxonomy methods, providing a more reliable foundation for biological research [99].By integrating multiple taxonomy evidence and techniques [100][101][102], species delimitation can be performed more accurately, which can provide a more reliable basis for the protection and utilization of biodiversity.

Figure 1 .
Figure 1.Description of the curves (a) and semi-landmarks (SLM) (b) of the pronotum and elytron used in the geometric morphometric analysis (Neotriplax miwai).The blue dots stand for starting and ending points.

Figure 1 . 26 Figure 2 .
Figure 1.Description of the curves (a) and semi-landmarks (SLM) (b) of the pronotum and elytron used in the geometric morphometric analysis (Neotriplax miwai).The blue dots stand for starting and ending points.Insects 2024, 15, x FOR PEER REVIEW 5 of 26

Figure 2 .
Figure 2. Description of the curves of the wing shape (A(a)) and semi-landmarks (SLM).(B(a)); Distribution of veins (A(b)) and 19 landmarks (B(b)) in the wing used in the geometric morphometric analysis (Neotriplax arisana).The blue dots stand for starting and ending points.

Figure 3 .
Figure 3. Phylogenetic relationships among the genus of Neotriplax.Shown here is the phylogeny inferred from the COI gene using ML and BI.The number on the left represents ultrafast bootstrap support (BS, %) and the number on the right represents the posterior probabilities (PP).

Figure 3 .
Figure 3. Phylogenetic relationships among the genus of Neotriplax.Shown here is the phylogeny inferred from the COI gene using ML and BI.The number on the left represents ultrafast bootstrap support (BS, %) and the number on the right represents the posterior probabilities (PP).

Figure 4 .
Figure 4. Shape variation in pronotum in Neotriplax.(a) The proportion of the total variation explained by each principal component based on the contour of the pronotum.(b)Variation in the pronotum along PC1.(c) Variation in the pronotum along PC2.

Figure 4 .
Figure 4. Shape variation in pronotum in Neotriplax.(a) The proportion of the total variation explained by each principal component based on the contour of the pronotum.(b)Variation in the pronotum along PC1.(c) Variation in the pronotum along PC2.

Figure 5 .
Figure 5.The pronotum morphological variations in Neotriplax based on PCA (a) and CVA (b).The 90% equal frequency ellipses containing approximately 90% of the data points.

Figure 5 .
Figure 5.The pronotum morphological variations in Neotriplax based on PCA (a) and CVA (b).The 90% equal frequency ellipses containing approximately 90% of the data points.

Figure 6 .
Figure 6.Shape variation trend of elytron in Neotriplax.(a) The proportion of the total variation explained by each principal component based on the contour of the elytron.(b) Variation in the elytron along PC1.(c) Variation in the elytron along PC2.

Figure 6 . 26 Figure 6 .
Figure 6.Shape variation trend of elytron in Neotriplax.(a) The proportion of the total variation explained by each principal component based on the contour of the elytron.(b) Variation in the elytron along PC1.(c) Variation in the elytron along PC2.

Figure 7 .
Figure 7.The elytron morphological variations in Neotriplax based on PCA (a) and CVA (b).The 90% equal frequency ellipses containing approximately 90% of the data points.

26 Figure 8 .
Figure 8. Shape variation trend of hind wing shape and wing vein in Neotriplax.(a) The proportion of the total variation explained by each principal component based on contour of wing shape and wing vein.(b)Variation in the wing shape and wing vein along PC1.(c) Variation in the wing shape and wing vein along PC2.

Figure 8 .
Figure 8. Shape variation trend of hind wing shape and wing vein in Neotriplax.(a) The proportion of the total variation explained by each principal component based on contour of wing shape and wing vein.(b)Variation in the wing shape and wing vein along PC1.(c) Variation in the wing shape and wing vein along PC2.

Figure 9 .
Figure 9.The wing shape and wing vein morphological variations in Neotriplax based on PCA (a) and CVA (b).The 90% equal frequency ellipses containing approximately 90% of the data points.

Figure 9 .
Figure 9.The wing shape and wing vein morphological variations in Neotriplax based on PCA (a) and CVA (b).The 90% equal frequency ellipses containing approximately 90% of the data points.

Table 1 .
The Kimura 2-parameter (K 2 P) distances matrix between the taxa of Neotriplax based on the COI gene (diagonal is the intra-GD).