Re-examination of the Chinese record of Opisthotropis maculosa (Squamata, Natricidae), resulting in the first national record of O. haihaensis and description of a new species

Abstract The taxonomic status of the previous record of Opisthotropis maculosa Stuart & Chuaynkern, 2007 from Guangdong and Guangxi, southern China, is revised based on the comparison of morphological and molecular data collected from the Chinese specimens and the holotype of O. maculosa from Thailand and O. haihaensis Ziegler, Pham, Nguyen, Nguyen, Wang, Wang, Stuart & Le, 2019 from Vietnam. Results reveal that the population from Shiwandashan Nature Reserve in southern Guangxi, China belongs to O. haihaensis, and represents the first national record for China; the populations from western Guangdong and southeastern Guangxi are described as a new species, Opisthotropis hungtaisp. nov. We suggest that O. maculosa should be removed from the Chinese herpetofauna checklist. The new national record of O. haihaensis and the description of the new species bring the total number of Opisthotropis to 13 in China.

Opisthotropis maculosa was originally described based on a single male specimen from northern Thailand (Fig. 1, site 6). Subsequently, it was reported based on morphological identifications in Guangdong (Fig. 1, sites 1, 2) and Guangxi (Fig. 1, sites 3, 4) in southern China (Yang et al. 2011), and northern Vietnam (Fig. 1, site 5) (Nguyen et al. 2018). Although some minor morphological differences among Chinese, Vietnamese and Thai populations were acknowledged in these publications, molecular data only became available recently, which lead to the resolution of the taxonomic statuses of the Vietnamese and Chinese records of O. maculosa. The Vietnamese record of O. maculosa was described as a distinct species, O. haihaensis Ziegler, Pham, Nguyen, Nguyen, Wang, Wang, Stuart & Le, 2019, by comparing the molecular and morphological data with the holotype of O. maculosa (Ziegler et al. 2019). In addition, their results also pointed out that the population of O. maculosa from Heishiding Nature Reserve in Guangdong, southern China may represent another distinct lineage.
In the present study, the Opisthotropis specimens from Guangdong and Guangxi, southern China previously recorded as O. maculosa, were re-examined using an integra- tive taxonomic approach, by combining results from both morphological and molecular analyses. In particular, morphological comparisons among the Chinese 'O. maculosa', the true O. maculosa from Thailand and the recently described O. haihaensis from Vietnam were undertaken in detail. The results demonstrate that the populations from southeastern Guangxi and western Guangdong represent a distinct taxon, which is described as a new species; the population from southern Guangxi is identified as O. haihaensis.

Morphometrics
Morphological examinations were performed on the holotype of Opisthotropis haihaensis, specimens reported as O. maculosa by Yang et al. (2011), and several other newly collected specimens. The collection information is given in the taxonomy accounts below. All specimens were fixed in 10 % buffered formalin and later transferred to 70 % ethanol for preservation, and deposited in the Museum of Biology, Sun Yat-sen University (SYS), Kadoorie Farm and Botanic Garden (KFBG), and Institute of Ecology and Biological Resources, Hanoi, Vietnam (IEBR).
Measurements followed Wang et al. (2017) and Ziegler et al. (2019) and were taken with digital calipers to the nearest 0.1 mm. These measurements were as follows:

TL
total length (from tip of snout to tip of tail); SVL snout-vent length (from tip of snout to posterior margin of cloacal plate); TaL tail length (from posterior margin of cloacal plate to tip of tail).
Maxillary teeth counts (MT) were determined by subequal teeth or sockets on right upper maxilla, and sex was determined by dissection or by the presence/absence of everted hemipenis.

Phylogenetic analyses
The mitochondrial cytochrome b (CYTB) gene was used for molecular analyses. Two new samples from Mt. Wuhuang, southeastern Guangxi and Shiwandashan Nature Reserve, southwestern Guangxi, were included in our study. DNA extraction, PCR amplification and sequencing followed the protocol employed by Wang et al. (2017). In addition, 31 Opisthotropis and two outgroup sequences (following Ziegler et al. 2019) were attained from GenBank for the phylogenetic analysis (Table 1).
Amino acid sequences for the CYTB gene of all samples were first aligned using Clustal W with default parameters. After checking the alignment to make sure that there was no stop or error codon, the amino acid sequence dataset was transformed to a nucleotide sequence dataset. We then applied JModelTest v2.1.2 on the nucleotide sequence dataset under Akaike and Bayesian information criteria to determine the bestfit nucleotide substitution model. The dataset was analyzed using maximum likelihood (ML) in RaxmlGUI 1.3 (Silvestro and Michalak 2012) and Bayesian inference (BI) in MrBayes 3.2 (Ronquist et al. 2012) with the GTR + I + G model. For ML analysis, a bootstrap consensus tree inferred from 1000 replicates was used to represent the evolutionary history of the taxa analyzed. Branches reproduced in less than 50% of bootstrap replicates were collapsed. For BI analysis, two independent runs with four Markov Chain Monte Carlo simulations were performed for ten million iterations and sampled for every 1000 th iteration. The first 25% of samples were discarded as burn-in. Convergence of the Markov Chain Monte Carlo simulations was assessed using Tracer v.1.4 (http://tree.bio.ed.ac.uk/software/tracer/). We also calculated pairwise sequence divergence based on uncorrected p-distance using MEGA 6.06 (Tamura et al. 2013).

Results
The CYTB nucleotide sequence matrix contained 1059 characters without insertiondeletions. The MP and BI analyses produced essentially identical topologies, which were integrated in Fig. 2. Major nodes of the tree were sufficiently supported, with Bayesian posterior probabilities (BPP) > 0.90 and bootstrap supports (BS) for maximum likelihood analysis > 80. Uncorrected p-distances among Opisthotropis species based on the CYTB gene are shown in Table 2. Within clade G (BPP 1.00 and BS 99), the Opisthotropis sample (SYS r000537) from Shiwandashan Nature Reserve, southern Guangxi, was placed with the holotype of O. haihaensis from northeastern Vietnam, with high node support values (BPP 1.00 and BS 100) and moderate genetic distance (p-distance 4.6%). The detailed morphological examination suggests that they represent individuals of the same species. Thus, we herein revise the identification of the specimen as O. haihaensis, and report it as a new national record for China.
Besides, the Opisthotropis samples from Mt. Wuhuang, southeastern Guangxi and Heishiding Nature Reserve, western Guangdong, were reconstructed as a monophyletic clade with strong nodal supports (BPP 1.00 and BS 100) and small genetic distance (p-distance 2.8%). The populations should be considered as a distinct taxon, which is sister to O. haihaensis. These specimens show almost no morphological differences from those collected at Dawuling Forestry Station, western Guangdong, which is located in the same mountain belt as Heishiding Nature Reserve. Therefore, we describe these specimens as a new species, Opisthotropis hungtai sp. nov.  Etymology. According to the original description, the specific name "haihaensis" refers to the type locality of this species, Haiha District (Quang Ninh Province) in Vietnam. As this species is currently reported in China, we suggest its Chinese name "Hai He Hou Leng She (海河后棱蛇)", derived from its scientific name.
Coloration in life (SYS r000537). Eye black; scales on dorsal surface of head glossy black with scattered yellow flecking; chin shields yellow with brownish black mottling; body and tail glossy black with iridescence above, with single yellow spot on each scale, yellow spots becoming larger on sides of body; ventrals yellow with brownish black lateral margins and scattered brown flecks; subcaudals yellow with brownish black anterior and lateral margins in both specimens.
Coloration in preservation (SYS r000537). Ground color of upper head and body surface dark brown, that of venter yellowish-beige. Dorsal scales each with light blotch in the center. Dorsal tail scales likewise with light central blotches. Dorsal head surface in part with indistinct light mottling. Anterior supralabials with large light mottling. Infralabials, chin shields and smaller throat scales anterior to ventrals yellowish-beige with dark brown mottling per scale. Belly with few, scattered dark flecks on sides. Outermost edges of light ventrals brown. Ground color of subcaudals brown with transversally enlarged light blotches at each scale end.
Distribution and habits. Opisthotropis haihaensis is currently known from its type locality, the forest near Tai Chi Village (ca 950 m a.s.l.), Quang Ninh, northern Vietnam, and Shiwandashan Nature Reserve (ca 500 m a.s.l.), southwestern Guangxi, southern China. The straight-line distance between the two localities is approximately 150 kilometers, indicating that the distribution area of this species is the mountain region on the border between China and Vietnam.
The holotype was found at night in a small rocky stream at 21:30h. The surrounding habitat was secondary evergreen forest consisting of small hardwoods, bamboo, and shrubs. The air temperature was 24-29 °C and the relative humidity was 65-88%. The holotype revealed to be an adult female, as dissection showed up to 16.5 mm long eggs and the oviducts were folded, indicating that eggs had already been laid (Ziegler et al. 2019). Besides, the other specimen, SYS r000537, was collected from a rocky stream (about 8 m wide and 0.3 m deep at the collecting site) running through wellpreserved, dense deciduous forests. The collected individual was spotted swimming at night and swiftly hiding under stones when disturbed.  ocular, one or two postocular(s), (5) temporals 1+1, (6) supralabials seven, the fourth and fifth in contact with eye; (6) maxillary teeth 16-18, (7) anterior pair of chin shields longer than or equal to posterior pair; (8) ventrals 170-189 (+ 2 preventrals) in males, 168-175 (+ 2 preventrals) in females, (9) subcaudals 76-98 in males, 69-84 in females, (9) nasal cleft pointing to the second supralabial, (10) body scale in 15-15-15 rows, (11) body scales smooth, tail scales smooth or indistinctly keeled, (12) chin shields yellow with brownish black mottling, and (13) body and tail dorsum dark, each with a light spot per scale.
Comparisons. Opisthotropis hungtai sp. nov. is compared with O. maculosa and O. haihaensis, which share a very similar appearance. Measurements, scalation and body proportions of O. haihaensis and Opisthotropis hungtai sp. nov. are listed in Table 3.
Opisthotropis hungtai sp. nov. differs from O. maculosa by prefrontal not touching supraocular (vs. prefrontal touching supraocular in O. maculosa), by frontal touching preocular (vs. frontal not touching preocular in O. maculosa), by fourth and fifth supralabials in contact with eye (vs. fourth supralabial in contact with eye in O. maculosa), by anterior pair of chin shields longer than or equal to posterior pair (vs. anterior pair of chin shields shorter than posterior pair in O. maculosa), by a higher number of subcaudals, 76-98 in males (vs. 67 in the single male holotype of O. maculosa), and by chin shields yellow with brownish black mottling (vs. immaculate in O. maculosa).
Description of holotype. Body cylindrical, slender, round to oval in cross section; TL 501.2 mm (SVL 401.6 mm, TaL 99.6 mm); tail thin and pointed, TaL 20% of TL; head small, indistinct from neck; right upper maxilla with 16 subequal teeth or sockets, teeth small, curved, without diastema; rostral nearly flattened, small, slightly less than twice as broad as deep, barely visible from above; two internasals, crescent-shaped, in contact with each other medially behind the rostral, not in contact with loreal, posteriorly in contact with prefrontal; a single prefrontal, in contact with loreal and preocular laterally, with frontal posteriorly, not in contact with supraocular; a single frontal, hexagonal, in contact with supraocular laterally, with two parietals posteriorly; parietals large, in contact with each other medially; nasal directed dorsally, polygonal, in contact with first and second supralabials ventrally, with loreal and prefrontal posteriorly, with internasal dorsally, with rostral anteriorly; nostril horizontally oval, in the upper part of nasal; a short vertical cleft below the nostril and dividing nasal into anterior and posterior parts, pointing to middle of upper edge of second supralabial; a single loreal, trapezoid, not entering the orbit, in contact with second and third supralabials laterally; a single supraocular, much longer than wide, obliquely set; a single preocular, higher than wide, in contact with frontal; a single postocular; a single anterior temporal, significantly elongate, in broad contact with the elongated sixth supralabial; a single posterior temporal, pentagonal; supralabials 7/7, the sixth one significantly elongate, the last one much shorter than the adjacent preceding supralabial; fourth and fifth supralabials entering orbit; infralabials 7/7, the first one in contact with its fellow behind the mental; two pairs of chin shields; anterior chin shields larger, in contact with each other medially, and in contact with the first four infralabials on both sides; posterior chin shields smaller, in contact with each other; dorsal scales in 15-15-15 rows; dorsal scales of body smooth throughout; dorsal scales of tail weakly keeled; ventrals 170; cloacal plate divided; subcaudals 76, paired.
Coloration of holotype in life. Eye black; scales on dorsal surface of head glossy dark brown with scattered yellow flecking; chin shields yellow with brownish black mottling at each margin; body and tail glossy dark brown with single yellow spot on each scale, yellow spots becoming larger on sides of body; ventrals yellow with brownish black lateral margins and few scattered brown flecks; subcaudals yellow with brownish black anterior and lateral margins.
Coloration of holotype in preservative. Ground color of upper head and body surface dark brown (Fig. 4B), that of venter yellowish-beige. Dorsal scales each with light yellow blotch in the center. Dorsal blotches almost equal in size. Blotches becoming wider towards body sides; largest at outermost dorsal scale row, where the light blotches stretch towards the posterior scale end. Dorsal tail scales likewise with light central blotches. Dorsal head surface in part with indistinct light mottling that becomes more obvious on temporals. All supralabials with a light blotch. Infralabials, chin shields and smaller throat scales anterior to ventrals light yellow with brown mottling/blotches per scale. Belly with few, scattered dark flecks. Outermost edges of light ventrals brown. Ground color of subcaudals light yellow with black anterior and lateral margins.
Variations. Measurements, body proportions and scale counts are listed in Table  3. All paratype specimens are very similar to the holotype in appearance (Fig. 6) except: more maxillary teeth, ventrals and subcaudals, and relatively longer tail length in specimens KFBG 2002.01, SYS r001515, 2017 from Dawuling Forestry Station and SYS r000538 from Mt. Wuhuang; in the three female specimens from the same locality (Heishiding Nature Reserve) as the holotype, there are 17 maxillary teeth (vs. 16 maxillary teeth) and fewer subcaudals, 56 (broken tail) in SYS r000720, 69 in SYS r001350, 70 in SYS r001525 (vs. 76 in the male holotype).
The specimen from Mt. Wuhuang was collected in a rocky stream. Besides, specimens from Heishiding Nature Reserve were found in pelitic gutterways along the dirt path, and specimens from Dawuling Forestry Station were collected in a pelitic stream. The collection sites were all surrounded by well-preserved, dense deciduous forest.

Discussion
As a representative snake group of the Oriental Realm, the mountain Keelback genus Opisthotropis receives more attention for its important role as an environmental indicator. Mountain Keelbacks are generally adapted to rocky forest streams (Wang et al. 2017a, b). However, species delimitations in this genus are still poorly resolved. The true diversity of Opisthotropis was underestimated, which is fatal for appropriate conservation of these habitat specialists. According to the integrative taxonomic approach, i.e., combining detailed morphological and molecular analyses, used in this study, the record of O. maculosa should be removed from the Chinese herpetofauna. So far, the true O. maculosa is restricted to northern Thailand, with only a single individual recorded. Opisthotropis haihaensis occurs in the mountain regions along the China-Vietnam border with only two female specimens recorded up to now, and O. hungtai is known only from the hilly regions between Guangxi and Guangdong of southern China. Extended surveys are urgently needed over the broad region from southern China to northern Thailand to investigate the distribution and the population status of these three species. Due to their beautiful color patterns, the snakes are in high demand in the animal trade market (Ziegler et al. 2019), and they must be evaluated for inclusion in one of the conservation categories in the IUCN Red List of Threatened Species.
The discovery of Opisthotropis hungtai sp. nov. brings the total number of species of Opisthotropis to 24. Nevertheless, with regard to recent phylogenetic results (Ren et al. 2017;Wang et al. 2017;Ziegler et al. 2019; this study), the relationships of clades within this genus still remain largely unresolved. As the mitochondrial CYTB gene is unable to generate significant support values, further work employing multilocus nuclear-gene and matrilineal mtDNA genealogy is recommended to decipher this puzzle. In addition, the similar appearance of O. maculosa, O. haihaensis and O. hungtai, together with their distant genetic divergence, indicates cryptic speciation in the genus Opisthotropis. The non-monophyletic relationships between O. maculosa and the clade composed of O. hungtai and O. haihaensis in our phylogenetic tree indicate that identical or similar phenotypes have evolved independently.