Eight new species of the genus Anaplecta Burmeister, 1838 (Blattodea: Blattoidea: Anaplectidae) from China based on molecular and morphological data

In this study, we examine 500 specimens of Anaplecta collected from China, of which 26 samples were used for COI sequencing. We confirm eight new species, i.e., Anaplecta corneola Deng & Che sp. nov., Anaplecta staminiformis Deng & Che sp. nov., Anaplecta arcuata Deng & Che sp. nov., Anaplecta strigata Deng & Che sp. nov., Anaplecta furcata Deng & Che sp. nov., Anaplecta cruciata Deng & Che sp. nov., Anaplecta nigra Deng & Che sp. nov. and Anaplecta bicolor Deng & Che sp. nov. based on morphological and molecular data using ABGD and GMYC analyses. The results of ABGD and GMYC were basically consistent with the morphospecies of Anaplecta. The intraspecific and interspecific genetic distances of Anaplecta ranged from 0 to 6.6% and 16.8% to 31.8%, respectively. We found the male genitalia of Anaplecta to exhibit intraspecific variation, especially in the phallomeres.

The description of Anaplecta relied mostly on the venation and color of the wings (Shelford 1906) and not on the male genitalia; but in many cases, the venation and color are unstable (Bruijning 1948). Roth (1990) further supported Bruijning's view and pointed out that Anaplecta needed to be revised and redescribed using male genitalia when available. However, members of Anaplecta exhibit an intraspecific variation in the male genitalia (W.B. Deng, personal observation): L2d concave or spiny as well as a different position of the spine on R1. Therefore, it is challenging to distinguish species of Anaplecta only based on morphological characters. To clarify the situation, we obtained 26 COI sequences of Anaplecta spp. in order to check the reliability of the morphological data in distinguishing species. Photographs of male individuals of the new species described are provided.

Morphological study
All specimens were collected in Hainan, Hunan, Fujian, Guangdong, Guangxi, Yunnan, Xizang, Anhui, Chongqing, Jiangsu and Sichuan in China. The morphological terminology used in this paper mainly follows McKittrick (1964), Roth (1990), Grandcolas (1996) and Li et al. (2018). The measurements are based on the specimens examined. The genitalia segments of the examined specimens were placed in a centrifuge tube containing 10% NaOH and soaked in hot water at 98°C for 10 minutes, rinsed with distilled water, and stored in glycerin for observation. All segments were observed in a glycerin jelly using a Motic K400 stereo microscope. Photographs of the genitalia and body were taken using a Leica M205A stereo microscope with a Leica DFC camera. All photos and images were edited with Adobe Photoshop CS6.

PCR amplification and sequencing
The legs and thoracic muscle were used for molecular studies; other body parts were stored in 95% ethanol as voucher specimens. A total of 26 specimens were used for COI sequencing in this study. All sequences are deposited at the National Center for Biotechnology Information GenBank (Table 1). The extraction procedure was according to the Hipure Tissue DNA Mini Kit. The total DNA was stored at -20°C. Primers for the amplifications are COI-F3 (5'-CAACYAATCATAAAGANATTG GAAC-3') and COI-R3 (5'-TAAACTTCTGGRTGACCAAARAATCA-3'). The amplification conditions were: initial denaturation at 98°C for 2 min, followed by 35 cycles of 10s at 98°C, 10s for 51°C, and 15s for 72°C, with final extension of 2 min at 72°C. Laboratory reagents were provided by TsingKe Co, Ltd., People's Republic of China. All voucher specimens are deposited in College of Plant Protection, Southwest University.

Sequence processing and phylogenetic analyses
A total of 31 COI sequences were analyzed (26 sequences representing species of Anaplecta, 3 sequences representing species of Periplaneta Burmeister, 1838, and 2 sequences representing a mantis outgroup downloaded from GenBank; no appropriate Anaplecta COI data of other countries could be found on GenBank) ( Table 1). All COI sequences were aligned using MEGA 7.0 and adjusted visually after translation into amino acid sequences. The genetic divergence value was quantified based on the Kimura 2-parameter (K2P) distance model (Kimura 1980) by MEGA 7. Maximum Likelihood (ML) analysis was implemented in RAxML 7.3.0 (Stamatakis et al. 2008) using a GTRGAMMA model with 1000 bootstrap replicates.
We also performed two molecular species delimitation methods, the Automatic Barcode Gap Discovery (ABGD: Puillandre et al. 2012) and the General Mixed Yule-coalescent (GMYC: Pons et al. 2006), in order to check the reliability of new species based on morphological data. The ABGD method was available at the web interface (https://bioinfo.mnhn.fr/abi/public/abgd/abgdweb.html) and used as a simple, quick and efficient method with the default settings, by using the Jukes-Cantor (JC69) and p distance model with relative gap width (X = 1.0). The GMYC method requires a fully-resolved ultrametric tree for this analysis to define species. Time-resolved gene trees were inferred in BEAST 1.8.1 (Drummond & Rambaut 2007) using the best models from PartitionFinder ver. 1.1.1 (Lanfear et al. 2012). The best-fitting models were as follows: COI_pos1, HKY + G; COI_pos2, TrNef + G; COI_pos3, F81 + I. The following settings were: rate variation modeled among branches using a strict clock model with the mean clock rate fixed to 1 and the Birth-Death speciation used as a tree prior. We elected to use the GMYC method to build the ultrametric gene tree using the SPLITS package (Ezard et al. 2009) in R (R Code Team 2013). The species delimited were compared to a one species null model using a likelihood ratio test.

Morphological delimitation of Anaplecta
On the basis of the morphology, including male genital characters, we were able to identify 11 morphospecies of Anaplecta among 500 samples examined from China (Fig. 1A).

Phylogenetic analysis based on COI and MOTUs estimation
In this study, we acquired 26 COI sequences of Anaplecta representing 11 morphospecies of Anaplecta, whose length, excluding primers, was 658 bp. All of the new sequences have been deposited in GenBank with accession numbers MT800285 to MT800310 (Table 1). The parts of COI we sequenced had a relatively high AT level (60.5%), with an average nucleotide composition of A = 29.5%, T = 31.2%, G = 16.9%, and C = 22.4%. ML analysis revealed that samples from the same morphospecies, including females, constitute monophyletic groups ( Fig. 1) with high support values.
We used two molecular species delimitation methods (ABGD, GMYC) in our study to delimit species of Anaplecta. The ABGD analysis for MOTUs detection was estimated with JC69 and P = 0.001, and the likelihoods of the null and GMYC models from COI analysis were 96.16 and 117.59, respectively. The same MOTUs were detected in both the ABGD and GMYC analyses ( Fig. 1B-C) for nine morphospecies, which were recovered as single MOTU in both methods. There was a minor difference in these two analyses: four samples of A. corneola Deng & Che sp. nov. were recovered as one single MOTU in the ABGD method, which is consistent with the result based on morphological data, while they were recovered as two MOTUs in GMYC ( Fig. 1B-C).

Establishment of eight new species after evaluating morphological and molecular data
The number of MOTUs recovered from our ABGD analyses (Fig. 1B) is nearly consistent with the number of morphological species (Fig. 1A) that we identified, and almost conform with the result from the GMYC analyses (Fig. 1C). The only discrepancy between the two molecular species delimitation methods were the four samples (JFL1, WY, ZQ and MS) from Hainan, Fujian, Guangdong and Hunan, which were recovered as two MOTUs in GMYC but one in ABGD. These four samples exhibit a high similarity in morphology but, after careful examination, we found a slight difference in R1 besides size ( between ZQ and JFL1, 6.6% between MS and JFL1), we considered the slight morphological differences between these samples ( Fig. 10F-I) as intraspecific variation, not as interspecific difference, conforming to the result of the morphological identification. Another discrepancy existed between the morphological identification and the molecular species delimitation in specimens from Mt. Diaoluo (DLS) and Mt. Limu (LMS), which exhibit some difference in L2d: a sample from DLS is bifurcated (Fig. 3L), while a sample from LMS is spiny (Fig. 3O); and R2: a sample from DLS is slender (Fig. 3M), while the other is flat (Fig. 3P). But according to the genetic distance between them (1.9%) and the absence of other morphological differences, the difference of L2d and R2 between the two morphospecies should be considered as an intraspecific variation. Therefore, we identified ten species of Anaplecta among the 500 samples using morphological and molecular data ( Fig. 1 The intraspecific and interspecific genetic distances of Anaplecta ranged from 0 to 6.6% and 16.8% to 31.8%, respectively (Supplementary File 1). Anaplecta -Hebard 1929: 27. -Bruijning 1948: 43. -Princis 1965: 367. -Roth 19901996: 304;2003: 36. -Beccaloni 2014.

Diagnosis
Sexual dimorphism indistinct. Size small. Body smooth. Color ranging from brownish yellow to dark brown. Eyes large, wide apart. Ocelli absent, or only indicated by two dim white spots. Antennae longer than body length. Clypeus distinct, swollen (except for Australian species A. calosoma Shelford, 1912). Tegmina and wings: usually fully developed, both extending over the end of abdomen (except for Australian species A. brachyptera Roth, 1990). Tegmina narrow, usually with three longitudinal or sublongitudinal sectors. Wings often infuscated, radial field and part of appendicular field darker than remaining part and with large appendicular field about 40% of wings; costal veins and radial veins sometimes obvious, sometimes indistinct; two weakly developed veins, one longitudinal along fold, the other somewhat oblique in appendicular field. Legs: front femur type B2. Pulvilli absent or apparent on the fourth proximal tarsomere only. Claws usually simple, rarely serrated, symmetrical. Arolia present. Abdomen: male: middle area of supra-anal plate with a cluster of pubescence; apical margin convex, with long setae. Subgenital plate generally symmetrical; styli cylindrical with setae; female: subgenital plate valvular. Genitalia: male: hook (L3) on left.

Distribution
North America, South America, Africa, Asia, Oceania.

Etymology
The specific name is derived from the Latin word 'corneolus', referring to the right phallomere with a horn-shaped structure.

Etymology
The specific name is derived from the Latin word 'staminiformis' and refers to L2vm being stamenshaped.

Diagnosis
Anaplecta arcuata Deng & Che sp. nov. resembles A. corneola Deng & Che sp. nov., but can be distinguished from it by the following characters: 1) L2vm slender with brush-like bone sheet in the former, simple sheet-like in the latter; 2) L3 with apex of uncinate part sharp in the former, enlarged and blunt in the latter. The right phallomere was not seen when examining the species, we are not sure whether it is degenerated or missing is uncertain and needs to be checked in future examinations.

Etymology
The specific name is derived from the Latin word 'arcuatus', meaning that the brush-like bone sheet (L2vm) looks like a bent bow-like arcuate structure. Paratypes CHINA • 1 ♂, 1 ♀; same collection data as for holotype; SWU.

Diagnosis
This species resembles A. malayensis, but can be distinguished from it by the markings on the disk of pronotum. The former has two chocolate-brown markings on the disk of pronotum, yellowish-brown in the middle, while the latter is chocolate-brown, without markings.

Etymology
The specific name is derived from the Latin word 'strigatus' and refers to the two chocolate-brown stripes on disk of pronotum.

Etymology
The specific name is derived from the Latin word 'furcatus' and refers to the shape of the left paraproct plate looking like a fork. Paratypes CHINA • 2 ♂♂; same collection data as for holotype; SWU.

Etymology
The specific name is derived from the Latin word 'cruciatus' and refers to the crossed bone fragments (L1 and L2v) of the left phallomere.

Diagnosis
This species resembles A. basalis but can be distinguished from it by: 1) subgenital plate distinctly asymmetrical with posterior margin strongly produced at middle, while slightly asymmetrical with posterior margin more or less produced in A. basalis; 2) supra-anal plate with paraprocts nearly identical, distinctly different for A. basalis.

Etymology
The specific name is derived from the Latin word 'niger' and refers to the black markings at the base of tegmina. Paratype CHINA • 1 ♀; same collection data as for holotype; SWU.
Male abdoMen and genitalia. Supra-anal plate with paraprocts almost identical, simple sheet-like (Fig. 8I). Subgenital plate fan-shaped, distinctly asymmetrical with posterior margin strongly produced at middle; two styli long, left stylus skewed to hind margin of subgenital plate, distance between two styli very short (Fig. 8J). Phallomere complex, L1 small with slender and curved filamentary structure, L2v and L2vm broad, L2d dentate and sheet-like, L3 robust with apex of the uncinate part tapering; R1 arc-shaped, R2 irregular bone sheet and weakly sclerotized, R3 short simple sheet-like ( Fig. 8K-M).

Anaplecta bicolor
Deng & Che sp. nov. urn:lsid:zoobank.org:act:9EAA3F86-9C7F-453E-BA98-651CAF105F17 Fig. 9 Diagnosis This species resembles A. basalis in disk of pronotum with anterior half crimson-brown and posterior half yellowish-white, but can be distinguished by the following characters: 1) tegmina without black markings in Anaplecta bicolor Deng & Che sp. nov., the base of tegmina with black markings in A. basalis; 2) phallomere with a cluster of setae in former, the latter without; 3) L2d and L2vm absent in former, present in latter.

Etymology
The specific name is derived from the Latin word 'bicolor' and refers to the pronotum with the anterior half crimson-brown and the posterior half yellowish white. Coloration. Body reddish brown, eyes black, antennae yellow ( Fig. 9A-B). Vertex, frons reddish brown; clypeus yellowish brown (Fig. 9D). Disk of pronotum with anterior half reddish brown, posterior half yellowish white, lateral borders of pronotum hyaline (Fig. 9C). Tegmina pale reddish brown, lateral margins of tegmina hyaline (Fig. 9E). Wings with costal field and appendicular field infuscated, other part pale gray, with veins yellow (Fig. 9H). Legs and cerci yellowish brown (Fig. 9B).
Male abdoMen and genitalia. Supra-anal plate with paraprocts almost identical, slender piece of bone sheet connected with each paraproct (Fig. 9I). Subgenital plate fan-shaped, asymmetrical, hind margin extending beyond end of left stylus, but almost equal to right; two styli very small, distance between them long (Fig. 9J). Phallomere complex, L1 small with slender and curved filamentary structure, L2v clubbed, L2d and L2vm absent, L3 robust, uncinate part with apex enlarged and sharp; R1 irregular and middle wide, R2 irregular bone sheet, weakly sclerotized, R3 short simple sheet-like. Moreover, cluster of setae in right phallomere (Fig. 9K-M).

Diagnosis
This species is widely distributed in southern China and has a yellowish-brown body ( Fig. 10A-B). This species can be distinguished from other species by the right paraproct with many short, stout spines on curly posterior margin and left paraproct sheet-like (Fig. 10E).

Diagnosis
This species is easily separated from other species by its pale-yellow body. It can be distinguished by two other characters: 1) disk of pronotum with anterior half crimson-brown, posterior half yellowishwhite; 2) base of tegmina with black markings (Fig. 10C-D).

Discussion
DNA-based analyses resolved the vast majority of samples of Anaplecta to a putative species, but there is a difference in species delimitation between the two molecular methods used. Four samples of A. corneola Deng & Che sp. nov. (ZQ, JFL1, WY, MS) were recovered as two MOTUs in GMYC, but only one in ABGD. In spite of the high morphological similarity of the four samples, the differences exhibited in the four samples of A. corneola Deng & Che sp. nov. were determined to be intraspecific variation after a critical examination on the R1. The genetic distances among three samples of A. corneola Deng & Che sp. nov. (ZQ, WY, MS) are 0.3%-2% (Supplementary File 1) which was distinctly lower than the interspecific distance (16.8% to 31.8%); only the genetic distances between sample JFL1 and the other three samples (ZQ, WY, MS) are larger (Supplementary File 1: 6.3%-6.6%), but still distinctly lower than the interspecific distance. And there is the fact that the R1 of the samples from ZQ and JFL1 does not exhibit any difference (Fig. 10F, I), although a larger genetic divergence exists between them (Supplementary File 1: 6.5%). Also, there is a conflict between the morphological identification and the molecular species delimitation in our study, mainly in specimens of A. staminiformis Deng & Che sp. nov. These were treated as two morphospecies, firstly on the basis of a difference of L2d (Fig. 3L, O). However, we suggest they should be considered as one species via DNA-based analyses because of minor genetic distances (Supplementary File 1: 1.9%). Accordingly, a comprehensive analysis, including morphological or molecular methods, may help to delimit the species of Anaplecta successfully.
Members of the genus Anaplecta are widely distributed all over the world. Our collecting data reveal that species of Anaplecta have a strong migration ability and adaptability. For example, A. omei shows a wide distribution in southern China (Chongqing City, Jiangsu Prov., Fujian Prov., Sichuan Prov., Guangxi Prov., Anhui Prov.). Similarly, samples of A. corneola Deng & Che sp. nov. were collected across four provinces of China which are geographically distant from each other, especially the sample JFL from Hainan Province (an isolated island in the South China Sea). Therefore, the gene flow might be hindered by natural barriers among geographically distant populations, which may account for the larger intraspecific genetic distance among these four samples.