New Color-Patterned Species of Microtendipes Kieffer, 1913 (Diptera: Chironomidae) and a Deep Intraspecific Divergence of Species by DNA Barcodes

Simple Summary Non-biting midges are the most widely distributed, and frequently the most abundant, insect family in freshwater environments. Species delimitation concerning color patterns and the shape or distribution of thorax pigmentation, wing spots, abdomen pigmentation, and leg pigmentation are disputable and unstable in the family. This research focuses on a genus that shares the general appearance of the hypopygium, but with variations in coloration of the antennae, thorax, wings, and legs. In this study, we analyzed collected species along with public sequences, resulting in a preliminary DNA library including 21 morphospecies. DNA barcodes can successfully delimit Microtendipes species and showed deep intraspecific divergence in some species. We also confirmed that color patterns can be important diagnostic characters. As a result of this analysis, five species new to science are identified and described, and an updated key to male adults of known Microtendipes species from China is provided. Abstract The genus Microtendipes Kieffer (Diptera: Chironomidae) has a nearly worldwide distribution, comprising more than 60 species, which are further divided into two species groups based on larval stage. However, species delimitation and identification among the adults of this genus are controversial and uncertain. For instance, previous studies have provided many synonymies based on conspecific color pattern variations in Microtendipes species. Here, we used DNA barcode data to address Microtendipes species delimitation as well as to test whether color pattern variations can be diagnostic characters for interspecific identification. The 151 DNA barcodes used, 51 of which were contributed by our laboratory, represent 21 morphospecies. Species with specific color patterns could be accurately separated based on DNA barcodes. Consequently, the color patterns of adult males could be important diagnostic characters. The average intraspecific and interspecific sequence divergences were 2.8% and 12.5%, respectively, and several species exhibited deep intraspecific divergences higher than 5%. Molecular operational taxonomic units (OTUs) ranged from 21 to 73, based on methods including phylogenetic trees, the assemble species by automatic partitioning method, the Poisson tree process (PTP), and the general mixed Yule-coalescent (GMYC) method. As a result of these analyses, five new species were recognized (M. baishanzuensis sp. nov., M. bimaculatus sp. nov., M. nigrithorax sp. nov., M. robustus sp. nov., and M. wuyiensis sp. nov.).


Introduction
The family Chironomidae, informally known as non-biting midges, is one of the most abundant and species-rich insect families, with over 10,000 species worldwide [1,2]. It is the most widely distributed of all aquatic insect families, occurring in all zoogeographical regions of the world, including Antarctica [3]. Its members can inhabit different kinds of environments, ranging from undisturbed to human-impacted ecosystems, which makes them useful as bio-indicators of water quality or environmental changes [4].
Microtendipes Kieffer, 1915 (Diptera: Chironomidae) is a cosmopolitan genus of the tribe Chironomini in the subfamily Chironominae, comprising more than 60 described species globally [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. The immature stages of Microtendipes can occur in littoral and sublittoral sediments of large bodies of water, with a few species occurring in running water [20]. The genus was discovered by Kieffer in 1915, via the type species Tendipes abbreviatus Kieffer [=chloris (Meigen, 1818)]. The males of Microtendipes can be distinguished from all other Chironomini by one or two rows of stout and a proximally directed setae on the fore femur. However, finding diagnostic characters for species delimitation within the genus is still a Gordian knot. For example, when Towns (1945) discussed the color varieties of the species Microtendipes pedellus (De Geer, 1776), in the key, he primarily paid attention to the colors of the legs, thorax, and abdomen to delimit the varieties [21]. Additionally, Tang et al. (2017) proposed eleven synonymies with Microtendipes umbrosus Freeman, 1955, regarding color varieties as instances of conspecific variation [16]. However, they also doubted that such a morphologically defined but variable species with such a wide range might show geographically discrete populations or cryptic species, even if molecular data were considered.
For decades, mitochondrial DNA has been used as the molecular marker of choice to identify evolutionarily significant units and infer their phylogenetic relationships. Specifically, Hebert (2003) proposed an identification system based on a standardized fragment of the mitochondrial gene cytochrome c oxidase subunit I (COI), which could be used for quick identification and delimitation of species [22]. This method has been widely used and marked as one key character of species recognition, especially for animals [23][24][25][26][27]. It has also revealed cases of apparent cryptic speciation [28,29]. For instance, Hebert et al. (2004) noted ten species in skipper butterflies [30], and a recent publication by Sharkey (2021) presented 403 new species in Costa Rican braconid parasitoid wasps [31]. This technique countervails the previous taxonomic impediment that identification relies on microscopy, which requires substantial experience in sample preparation and taxonomic training. However, such identifications largely depend on MOTUs (molecular operational taxonomic units) defined by their mitochondrial nucleotide divergence, skipping the tedious step of individual morphological identification [32]. Furthermore, DNA barcoding (or metabarcoding) is popularly employed in community ecology to quantify animal diversity distributions and to assess biodiversity patterns from environmental samples [33]. However, DNA barcoding cannot name the unknown taxa; it only delimits them or confirms identification based on what is available in the library database. Therefore, building a barcode library requires the expertise of taxonomists who can name and describe species.
The purpose of our study is to (1) test the feasibility of COI barcodes for the quick identification of Microtendipes species and (2) evaluate whether color patterns are appropriate for species delimitation. Furthermore, as a result of this study, obtained or analyzed sequences will complement the DNA barcode reference library of Chironomidae.

Materials and Methods
The material examined in this study was collected using light traps; the specimens were preserved in 75% ethanol at 4 • C or −20 • C in a refrigerator before slide mounting. Specimens were slide-mounted in Euparal after genomic extraction following the procedure described by Saether (1969) [50]. Morphological terminology follows that of Saether (1980) [51], and the description follows Longton and Pinder (2007) [52]. The photographs of the specimens' habitus were obtained with a DV500 5MP digital camera mounted on a Chongqing Optec SZ680 (Chongqing Optec Instrument Co., Ltd., Chongqing, China) stereo microscope and ZEISS camera mounted on ZEISS stereomicroscope(Carl Zeiss AG, Jena, Germany). The photographs of the mounted specimens were obtained using a Leica DMLS compound microscope (Leica Camera AG, Wetzlar, Germany).
The type materials, including holotype and paratypes of the newly described species, were deposited in the collection of the College of Life Sciences, Taizhou University, Taizhou, China (TZU).
Tissues for total genomic DNA extraction were removed from the thorax, head, and three legs of adult specimens. The genomic extraction procedure followed that of Frohlich et al. (1999) [53] and Song et al. (2018) [54]. The standard barcode region of the 5 portion of the mitochondrial gene cytochrome c oxidase I (COI-5P) was amplified using the universal primers LCO1490 and HCO2198 [55]; PCR amplifications followed those in Song et al. (2018). PCR products were electrophoresed in 1.0% agarose gel, purified, and sequenced using an ABI 3730XL capillary sequencer (Beijing Genomics Institute Co., Ltd., Hangzhou, China). Raw sequences were edited in BioEdit 7.2.5 [56]. Sequences, trace-files, and metadata of the new species were uploaded to the Barcode of Life Data System (BOLD) [57].
In addition to our own data, Microtendipes COI barcodes, longer than 500 bp and without stop codons, were searched, and 952 sequences added to the dataset named "DNA barcodes of Microtendipes species (DS-MICROT)", DOI: dx.doi.org/10.5883/DS-MICROT in the Barcode of Life Data System (http://www.boldsystems.org/ (accessed on 12 December 2022)). To reduce computing time, a reduced dataset containing 151 sequences was generated (Sequence S1).
Alignment was performed in MEGA 7 [58] using ClustalW, then a neighbor-joining tree was constructed using the K2P substitution model, and 1000 bootstrap replicates and the "pairwise deletion" option for missing data were utilized. The pairwise distances were calculated using the Kimura 2-Parameter (K2P) substitution model in MEGA 7. A maximum-likelihood (ML) tree was constructed using IQ-TREE v2.1.3 [59]. Node supports were estimated using ultrafast bootstrapping with 1000 replicates. Bayesian inference analysis was carried out using Markov Chain Monte Carlo (MCMC) randomization in MrBayes v3.2.7 [60], with 10 million generations, and the first 25% of sampled trees were discarded as burn-ins. Trace files of BI analysis were inspected in Tracer 1.7 [61], and then the tree was visualized in FigTree v.1.4.2.
Assemble species by automatic Partitioning (ASAP) analysis was implemented on the website https://bioinfo.mnhn.fr/abi/public/asap/asapweb.html (accessed on 6 January 2023) (Puillandre et al. 2021 [62]) with the K2P model. The PTP analyses used a rooted phylogenetic input tree constructed with raxmlGUI version 1.3, using 1000 nonparametric replicates and the GTR + G + I nucleotide substitution model [63]. The Bayesian Poisson tree process (bPTP) analyses were run on a web server (http://species.h-its.org/ptp (accessed on 11 December 2022)) with 500,000 MCMC generations, a burn-in of 0.1, and other parameters as defaults. The general mixed Yule-coalescent (GMYC) method was applied using the splits package, with the guidelines available on Tomochika's web page (https://tmfujis.wordpress.com/2013/04/23/how-to-run-gmyc/ (accessed on 16 December 2022)). The input ultrametric tree for GMYC was constructed using BEAST 1.7 [64]. Settings were as follows: relaxed clock, MCMC chain using 100 million generations, TN93 substitution model, and Yule speciation model. Other parameters are available by request from the authors.

Barcode Analysis
The 151 aligned and reduced sequences ranged from 506 to 654 base pairs, among which 236 were variable sites (221 parsimony informative; Table 1). Most variable sites occurred in the third codon position. The average intraspecific divergence was 2.8%, with the maximum intraspecific divergence found in Microtendipes famiefeus Sasa, 1996 (9.2%, Table S1), which was apparently larger than the acknowledged 3% threshold in insect species. Sequences labeled as M. famiefeus formed three genetically divergent clades, which might indicate cryptic diversity or misidentifications. Similar conditions of intraspecific divergence larger than 5% were also observed in Microtendipes pedellus (De Geer, 1776), Microtendipes chloris (Meigen, 1818), and Microtendipes bimaculatus sp. nov. in this study.
For Microtendipes bimaculatus sp. nov., the intraspecific genetic divergence ranged from 0 to 11.6 % (Table S2), and a total of three well-separated barcode clades were found in NJ ( Figure S1), ML (Figure 1), and BI inferences ( Figure S2), forming genetically paraphyletic phylogenetic trees. Three OTUs were estimated by ASAP, GMYC, and other analyses. However, no structural differences and no clear ecological separation were detected in this species as far as we could observe. There are cases of molecular discordance in which morphospecies have commonly been found in insect groups. Phylogeny based on a single gene may not follow the species history because of incomplete lineage sorting and introgressive hybridization, such as nuclear copies of mitochondrial DNA (NUMTs) or endosymbiosis [61,62]. However, such cases as these have not yet been recorded in nonbiting midge species. Under special circumstances, such as geographical and demographic expansion, nuclear genomes will come into contact and fully recombine (in the absence of reproductive isolation), while divergent mitochondrial genes will be retained as drift, but this is no longer the case in large expanding populations [63]. Such high divergence among the mt DNA sequences of these morphologically indistinguishable sympatric and allopatric populations might represent more than one species, but in this study, we regard them as one and the same species and will continue to do so until more evidence can be found.
For the sequences labeled M. pedellus, the intraspecific pairwise distance ranged from 0 to 12.0% (Table S2), and a total of three separated barcode clades were found ( Figure 1). If all sequences labeled "Microtendipes pedellus grp" are included, there are five well-separated clusters, and intraspecific divergence increases by up to 15.7%. The species was originally established by (De Geer), and three variations were recorded: Microtendipes pedellus var. pedellus, Microtendipes pedellus var. aberran, and Microtendipes pedellus var. stygius. The main differences between these were color patterns, which had previously been regarded as conspecific variation by Towns (1945) [17]. We tried to recheck the specimens of M. pedellus and found some images from the BOLD system. As can be seen, at least two kinds of color patterns existed, especially regarding the leg colors-clade-1 of M. pedellus: image not available; clade-2 of M. pedellus: apical 1/3 of the fore femur is dark brown, and the basal 1/3 and apical part of the fore tibia is dark brown; clade-3 of M. pedellus: fore femur and tibia are brown. Therefore, such color patterns of thorax pigmentation, wing spots, abdomen pigmentation, and leg pigmentation should be regarded as interspecific For the sequences labeled M. pedellus, the intraspecific pairwise distance ranged from 0 to 12.0% (Table S2), and a total of three separated barcode clades were found ( Figure 1). If all sequences labeled "Microtendipes pedellus grp" are included, there are five well-separated clusters, and intraspecific divergence increases by up to 15.7%. The species was originally established by (De Geer), and three variations were recorded: Microtendipes pedellus var. pedellus, Microtendipes pedellus var. aberran, and Microtendipes pedellus var. stygius. The main differences between these were color patterns, which had previously been regarded as conspecific variation by Towns (1945) [17]. We tried to recheck the specimens of M. pedellus and found some images from the BOLD system. As can be seen, at least two kinds of color patterns existed, especially regarding the leg colors-clade-1 of M. pedellus: image not available; clade-2 of M. pedellus: apical 1/3 of the fore femur is dark brown, and the basal 1/3 and apical part of the fore tibia is dark brown; clade-3 of M. pedellus: fore femur and tibia are brown. Therefore, such color patterns of thorax For the sequences labeled M. chloris, the intraspecific pairwise distance ranged from 0 to 10% (Table S2), forming three well-separated clades in the phylogenetic trees ( Figure 1 and Figures S1-S5). In Tang et al. (2017) [13], the species M. chloris, previously identified by Sasa (1984) and Sasa and Kamimura (1987), was regarded as a new species [65,66]. As vouchers are not accessible to check, we assume such sequences have not been updated or that potential cryptic species may exist. Such cases are also found in Polypedilum and Tanytarsus species [50,67]. Several reasons might account for this: (1) Diagnostic characters might be unreliable due to intraspecific morphological variation in morphometric ratios and hypopygial structures caused by different temperature regimes and food quality [68].
(2) Artifacts created during the slide-mounting process can also obscure species-specific characteristics, such as shapes or length [69]. (3) Morphological differences are always presented in one or a few life stages, but not associated with other stages [70,71]. In the case of cryptic species in Micropsectra, for instance, what could not be observed in adults was distinct in pupal stages [70].
The mean interspecific divergence was 12.5%, with the maximum interspecific divergence found between M. rydalensis grp. and M. sp.1BD, up to 18.7% (Table S3). The minimum interspecific distance was found to be lower than 2.5%, between M. pedellus and M. chloris, and M. pedellus and M. brevitarsis Brundin, 1947 (Table S2), which formed a monophyletic clade. This may be a misidentification of specimens or a new synonym, as vouchers could not be examined.

Species Delimitation
In many cases, even a practicing taxonomist who thoroughly knows their group can hardly interpret intraspecific and interspecific distances. Due to different species with different population sizes and divergence times, a universal threshold that fits all taxa does not exist [72]. A value close to the 2% COI threshold was adopted for vertebrate birds [73], Ephemeropteran, Plecopteran, and Trichopteran [74,75], while a fixed 3% was adopted for lepidopteran insects [76]. For non-biting midges, 4-5% was adopted for Tanytarsus [67], and 5-8% for Polypedilum [50,77]. In this study, a barcode gap ( Figure 2) between 4 and 5% was observed in the K2P genetic distance histogram. Is this threshold appropriate for the Microtendipes delimitation? In the ASAP analysis, only using a 6.2% threshold k2p distance, the lowest score, 4.00 (the lower the score, the better the partition), yielded 26 OTUs ( Figure S3). Applying the prethreshold clustering method with hierarchical thresholds from 2% to 8% gave 9-38 OTUs ( Figure 3). Setting higher initial threshold values from 6% to 6.5% gave 21-26 OTUs. Consequently, a threshold of 6% might be more applicable for Microtendipes species. However, distance-based species delimitation ignores the evolutionary relationships within the species [78], and phylogeny-based methods apply the "phylogenetic species concept", which defines a species as the smallest resolvable separately evolving lineage or the smallest diagnosable cluster. Based on the NJ tree, ML tree, and BI tree, 151 DNA barcodes of 21 initially morphospecies were clustered into 28 clades. Most of the morphospecies formed a monophyly clade, and some did not, for instance, M. pedllus, M. bimaculatus, and M. chloris. Some species without large geographical barriers grouped into nested clades, with deep intraspecific divergences, such as Microtendipes nigrithorax sp. nov. and Microtendipes robustus sp. nov. The single-threshold general mixed Yule-coalescent calculations In the ASAP analysis, only using a 6.2% threshold k2p distance, the lowest score, 4.00 (the lower the score, the better the partition), yielded 26 OTUs ( Figure S3). Applying the prethreshold clustering method with hierarchical thresholds from 2% to 8% gave 9-38 OTUs ( Figure 3). Setting higher initial threshold values from 6% to 6.5% gave 21-26 OTUs. Consequently, a threshold of 6% might be more applicable for Microtendipes species. However, distance-based species delimitation ignores the evolutionary relationships within the species [78], and phylogeny-based methods apply the "phylogenetic species concept", which defines a species as the smallest resolvable separately evolving lineage or the smallest diagnosable cluster. Based on the NJ tree, ML tree, and BI tree, 151 DNA barcodes of 21 initially morphospecies were clustered into 28 clades. Most of the morphospecies formed a monophyly clade, and some did not, for instance, M. pedllus, M. bimaculatus, and M. chloris. Some species without large geographical barriers grouped into nested clades, with deep intraspecific divergences, such as Microtendipes nigrithorax sp. nov. and Microtendipes robustus sp. nov. The single-threshold general mixed Yule-coalescent calculations (ST-GMYC) yielded 30 entities ranging from 28 to 34 ( Figure S4). While more OTUs were estimated, using the bPTP method gave 47 and 34-73 species ( Figure S5). Our results suggest that the numbers of OTUs estimated by phylogeny-based approaches are more than that by the distance-based methods. The results of this molecular species delimitation provide strong support for m phological species of Microtendipes. Species here represented by different color patte never intersect with other species in the phylogenic trees. Different permutations a combinations, including antenna, wing, thorax, legs, and abdomen pigmentation dis bution patterns, may indicate rich species diversity and cryptic diversity. In this stu five new species formed seven separated clades and at least nine OTUs as estimated different analyses. For instance, M. bimaculatus sp. nov and M. nigrithorax sp. nov. w both estimated to form more than two OTUs, although specimens were collected fr geographically close areas. Nevertheless, morphological differences were not observ according to current evidence. In future studies, multiple genes or genomes are needed delimit and to discover the full diversity of Microtendipes.  The results of this molecular species delimitation provide strong support for morphological species of Microtendipes. Species here represented by different color patterns never intersect with other species in the phylogenic trees. Different permutations and combinations, including antenna, wing, thorax, legs, and abdomen pigmentation distribution patterns, may indicate rich species diversity and cryptic diversity. In this study, five new species formed seven separated clades and at least nine OTUs as estimated by different analyses. For instance, M. bimaculatus sp. nov and M. nigrithorax sp. nov. were both estimated to form more than two OTUs, although specimens were collected from geographically close areas. Nevertheless, morphological differences were not observed according to current evidence. In future studies, multiple genes or genomes are needed to delimit and to discover the full diversity of Microtendipes.  Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: antenna pale brown with most antenna plumage blackish; yellowish brown ground of thorax with dark lateral stripes; dark brown anepisternum II and postnotum; fore femur pale with light brown ring in the anterior part, and all knees slightly brown; and wings without faint markings. Superior volsella has one basal seta and five to seven, six setae in the middle, while the median volsella is poorly developed, with one to two clustered setae.

Taxonomy
Etymology. The new species is named after the reserve (Baishanzu) where the holotype was collected. The name is to be regarded as a noun in apposition.
Description. Male imago (n = 5).  Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: antenna pale brown with most antenna plumage blackish; yellowish brown ground of thorax with dark lateral stripes; dark brown anepisternum II and postnotum; fore femur pale with light brown ring in the anterior part, and all knees slightly brown; and wings without faint markings. Superior volsella has one basal seta and five to seven, six setae in the middle, while the median volsella is poorly developed, with one to two clustered setae.
Etymology. The new species is named after the reserve (Baishanzu) where the holotype was collected. The name is to be regarded as a noun in apposition.
Description. Male imago (n = 5).  Coloration ( Figure 4). Mature male adult mostly pale yellowish to light brown; antenna light brown, and antenna plumage blackish; ground of thorax yellowish brown with dark lateral stripes, anepisternum II, and postnotum dark brown; abdomen yellowish; wing without markings; legs with poorly defined pigmentation. P1: mostly pale, anterior part of femur with pale brown ring, tibia slightly pale brown, tarsus pale. P2 and P3: pale, except knees slightly brown.
Remarks. The species is highly similar to Microtendipes umbrosus Freeman, 1955 in the hypopygial structure: the anal point is parallel-sided in dorsal view; superior volsella is sickle-shaped, with one basal and 3-7, five dorsal setae; median volsella is poorly developed, consisting of small tubercles with one to two, two setae. The two species can be separated based on the following characteristics: M. baishanzuensis's wings have no markings, whereas M. umbrousus has a median transverse dark band on its wings; wing length (2.15-2.38, 2.28) in M. baishanzuensis is shorter than M. umbrosus (2.5-3.8 mm) (Freeman 1961: 720) [79]      Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: antenna and antenna plumage are dark blackish; ground of thorax is yellowish brown with small dark medial stripes, dark brown lateral stripes, and light brown postnotum; fore femur is pale with a brown ring in the distal 1/3, the basal 1/2 and apical 1/4 of fore tibia are dark brown, and all knees dark brown; and the wing has a median band around the vein RM and FCu. The superior volsella wing has one basal seta and four to six, five setae in the middle, and the median volsella is poorly developed, with one to two clustered setae; segments I-V are white, and segments VI-IX are brown.
Etymology. The new species is named based on the characteristics of the two dark brown segments of the tibia of the fore legs. The word "bimaculatus" is Latin, meaning "two dark brown spotted segments".
Description. Male imago (n = 9   Table 3.  Remarks. The species is similar to Microtendipes simantofegeus Sasa, Suzuki & Sakai, 1998, based on a faint wing marking, poorly developed median volsella, and superior volsella bearing one basal seta and 5-6 long setae in the middle [80]. It differs from the existing species based on the following characteristics: (1) the thorax patterns of M. simantofegeus include a pale ground color of the scutum, pale median stripes, dark brown lateral stripes along the midline, and brownish yellow coloration in the median and lateral areas, while in the new species, the median stripes, and lateral stripes dark brown; and (2) the ninth tergite setae of M. simantofegeus has twelve setae, while that of the new species has four setae. Distribution. The species is only known in oriental China (Fujian and Zhejiang Province).

Microtendipes nigrithorax
Remarks. The species is similar to Microtendipes simantofegeus Sasa, Suzuki & Sakai, 1998, based on a faint wing marking, poorly developed median volsella, and superior volsella bearing one basal seta and 5-6 long setae in the middle [80]. It differs from the existing species based on the following characteristics: (1) the thorax patterns of M. simantofegeus include a pale ground color of the scutum, pale median stripes, dark brown lateral stripes along the midline, and brownish yellow coloration in the median and lateral areas, while in the new species, the median stripes, and lateral stripes dark brown; and (2) the ninth tergite setae of M. simantofegeus has twelve setae, while that of the new species has four setae.  Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: antenna and antennal plumage uniformly brown; thorax dark brown without any scutal vittae, abdomen yellowish or light brown; wing without any markings, fore tibia dark brown and other portion yellowish brown; superior volsella thumb-shaped with one basal inner seta and 4-6, five distal setae along outer side.

Microtendipes nigrithorax
Etymology. The new species is named based on the characteristics of the color of the thorax. The word "nigr" is Latin meaning "black", referring to the black thorax. Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: antenna and antennal plumage uniformly brown; thorax dark brown without any scutal vittae, abdomen yellowish or light brown; wing without any markings, fore tibia dark brown and other portion yellowish brown; superior volsella thumb-shaped with one basal inner seta and 4-6, five distal setae along outer side.
Etymology. The new species is named based on the characteristics of the color of the thorax. The word "nigr" is Latin meaning "black", referring to the black thorax.
Description. Male imago (n= 6). Thorax. Acrostichals absent, dorsocentrals 7-12, 10, prealars 4-6, 5, scutellars 8-12, 10. Wing ( Figure 9A)  Table 4. Hypopygium ( Figure 9D Distribution. The species is only known in Sichuan Province of China. Remarks. The species is similar to M. shoukomaki Sasa 1989, based on its similar hypopygium, wings without spots or colors, and dark brown fore tibia. It differs from the latter in the following characteristics: (1) the color pattern of the thorax, which in M. shoukomaki includes a brown ground coloration of the scutum, dark brown stripes, brown scutellum, and dark brown postnotum, while the new species is uniformly blackish; (2) the patterns of the middle and hind legs, entirely yellow except brown tarsi V in M. shoukomaki, while the tibia is slightly darker than other segments in the new species; and (3) the anal point is parallel-sided in M. shoukomaki, while it tapers to the apex in the new species.   Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: most of the antenna and antennal plumage is brown to dark brown; the thorax is uniformly dark brown; the distal half of the fore femur is light brown, with 20-24 proximally directed setae in two rows; the tibia is brown with a dark brown apical part; and it has wings without setae. Regarding the superior volsella wing, there is one basal seta and 5-8, seven setae in the middle, while the median volsella is poorly developed, with 2-2 clustered setae; the gonostylus is strong and bulb-like; and the abdomen is pale except for the hypopygium.
Etymology. The new species is named based on the characteristics of its strong and bulb-like gonostylus. The word "robustus" is Latin, meaning "strong".
Description Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: most of the antenna and antennal plumage is brown to dark brown; the thorax is uniformly dark brown; the distal half of the fore femur is light brown, with 20-24 proximally directed setae in two rows; the tibia is brown with a dark brown apical part; and it has wings without setae. Regarding the superior volsella wing, there is one basal seta and 5-8, seven setae in the middle, while the median volsella is poorly developed, with 2-2 clustered setae; the gonostylus is strong and bulb-like; and the abdomen is pale except for the hypopygium.  Table 5. Hypopygium ( Figure 11D,E). Tergite IX with 6-9, eight setae medially, which are divided into two groups. Laterosternite IX with 1-2, two setae. Anal point straight and parallel-sided in dorsal view, 70- Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: most of the antenna and antennal plumage are brown to dark brown; the ground of the thorax is yellowish brown with dark brown lateral stripes, medial stripes, and postnotum; the fore femur and distal wing have a dark brown ring, and the fore tibia is uniformly dark brown; and the wings have faint clouds around the vein RM and FCu. The superior volsella wing has one Diagnostic characters. The male imago can be separated from the known species of Microtendipes Kieffer, 1915 by the following combination of features: most of the antenna and antennal plumage are brown to dark brown; the ground of the thorax is yellowish brown with dark brown lateral stripes, medial stripes, and postnotum; the fore femur and distal wing have a dark brown ring, and the fore tibia is uniformly dark brown; and the wings have faint clouds around the vein RM and FCu. The superior volsella wing has one basal seta and 4-6 setae in the middle, and the median volsella is poorly developed with 2-3 clustered setae. Segments I-V are white, and segments VI-VIII are brown.
Etymology. The new species is named after the reserve (Wuyi) where the holotype was collected. The name is to be regarded as a noun in apposition. Coloration ( Figure 12). Mature male adult mostly yellowish to light brown. Most of the antenna and antennal plumage are brown to dark brown. The ground of thorax is brown with media and lateral stripes, and the postnotum is dark brown. Abdomen segment I-V: pale; segments V-IX: yellowish brown. Wing is with light marking around vein RM and FCu. Legs: P1: distal part with dark brown ring of femur, and tibia uniformly dark brown; Ta I-III: pale; and Ta IV-V: yellowish brown. P2 and P3: pale except knees brown.
Legs ( Figure Table 6. Hypopygium ( Figure 13D,E). Tergite IX with 4-6, five setae medially, which are divided into two groups. Laterosternite IX with 2-4, three setae. Anal point weakly tapered toward pointed apex in dorsal view, 48-58, 53 µm long and 14-18, 16 µm wide at base, 5-5, 5 µm wide at apex; 10-12, 11 setae distributed on each side of the base of anal point.  Diagnostic characters. This species can be separated from the known Microtendipes by the following characteristics: most of the antenna and antennal plumage are dark brown, the wings do not have markings, the connection parts of the legs are dark brown, the abdomen has brownish pigmentation from segments III-VIII, the front femur has a  Diagnostic characters. This species can be separated from the known Microtendipes by the following characteristics: most of the antenna and antennal plumage are dark brown, the wings do not have markings, the connection parts of the legs are dark brown, the abdomen has brownish pigmentation from segments III-VIII, the front femur has a small tubercle, the superior volsella has a basal lobe bearing five setae, and there is one long lateral seta.
Distribution. The species is distributed in Oriental China (Zhejiang, Hainan, and Guizhou Province).
An updated key to the known males of Microtendipes from China. The following key is modified from Qi et al. (2014) 1 Hypopygium without median volsella ( Figure 9D) 2 -Hypopygium with median volsella (e.g., Figure 5D

Conclusions
DNA barcodes can successfully delimit Microtendipes species and showed deep intraspecific divergence in some species. Those specimens initially identified as species groups formed several separate clades in the phylogenetic analysis. This also indicates either the presence of cryptic species or that the genus requires major revision of all life stages using several nuclear genes to explain the highly divergent COI lineages. Furthermore, based on the results obtained from DNA barcoding, color pattern variations of the wings, legs, thorax, and abdomen should be regarded as interspecific differences and thus as important diagnostic characters for the species of Microtendipes.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/insects14030227/s1, Sequence S1: Reduce alignment of 151 Microtendipes barcode sequences. Figure S1: Neighbour-joining tree for species of Microtendipes based on K2P model, with bootstrap 1000 replicates; scale equals K2P genetic distance; Figure S2: Bayesian tree for species of Microtendipes based on DNA barcodes data, numbers at nodes represent posterior probabilities; Figure S3: number of subsets based on ASAP; Figure S4: results of the GMYC analysis; Figure S5: Maximum likelihood tree based on the PTP model; Table S1: Average intraspecific divergence among Microtendipes species.

Data Availability Statement:
The molecular data presented in this study are openly available in BOLD (DOI: dx.doi.org/10.5883/DS-MICROT) and GenBank (accession numbers: OQ174670-OQ174720).