Description of a New Cobra (Naja Laurenti, 1768; Squamata, Elapidae) from China with Designation of a Neotype for Naja atra

Simple Summary Cobras (Naja Laurenti, 1768) are a group of well-known highly venomous snakes, which cause numerous cases of snakebites every year, especially in South Asia and Southern China. Taxonomic framework is essential for the medical treatment of snake bites and accurate antivenin development. However, the taxonomy of Asian cobras is still puzzling, especially for the widespread species Monocled Cobra (N. kaouthia Lesson, 1831). This study provided new materials and understanding for the taxonomy of this species by combining mitochondrial phylogenetic analysis and morphological comparisons based on samples from a vast area in Asia. The results showed that the Chinese population of N. kaouthia represents a new species. This study also provided new data for N. atra and designated a neotype for it. Furthermore, the subspecies N. naja polyocellata was resurrected and recognized as a full species, N. polyocellata comb. nov., and the subspecies N. sumatrana miolepis was also resurrected. This study highlighted the necessity to evaluate the effectiveness of cobra antivenin based on a comprehensive taxonomic framework. Abstract Taxonomic frameworks for medically important species such as cobras (genus Naja Laurenti, 1768; Squamata, Elapidae) are essential for the medical treatment of snake bites and accurate antivenin development. In this paper, we described the former N. kaouthia populations recorded from China as a new species and designated a neotype for N. atra-based morphological and mitochondrial phylogenetic analysis. The new species N. fuxi sp. nov. was morphologically diagnosed from N. kaouthia by (1) regular single narrow crossband present on the middle and posterior parts of the dorsum (3–15, 7.9 ± 2.7, n = 32) and the dorsal surface of the tail (1–6, 4.2 ± 1.1, n = 32) of both adults and juveniles, buff-colored with dark fringes on both edges, vs. South Asian populations (n = 39) and Southeast Asian populations (n = 35) without cross bands, with irregular cross bands or multiple light-colored crossbands pairs, or densely woven lines; (2) small scales between the posterior chin shields, usually three (40%) or two (37%), rarely four (13%), or one (10%) (n = 30) vs. mostly one (81%) and rarely two (19%) (n = 28); (3) ventrals 179–205 (195.4 ± 6.7, n = 33) vs. South Asian populations 179–199 (188.7 ± 5.9, n = 12); Southeast Asian populations 168–186 (177.8 ± 4.9, n = 18). Phylogenetically, the new species forms an independent sister clade to the clade including N. atra, N. kaouthia, N. oxiana and N. sagittifera. Furthermore, the subspecies N. naja polyocellata should be resurrected and recognized as a full species, N. polyocellata comb. nov., and the subspecies N. sumatrana miolepis should be resurrected.

To further understand these problems, we compiled a data set including a series of fresh cobra samples from China together with available cobra sequences and reconstructed a comprehensive mitochondrial phylogenetic tree. Together with morphological data based on 143 specimens and photos of 208 living individuals of nine taxa from Asia, we found the populations of Naja kaouthia from southwestern China represent a new species that we described herein. We also designated a neotype for N. atra for future taxonomic study on the genus.

Molecular Phylogenetic Analysis
Genomic DNA was extracted from muscle or liver tissues using QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). We sequenced three mitochondrial genes for analysis: cytochrome b (cyt b) (primers cited from Burbrink et al. [34]), NADH dehydrogenase subunit 4 (ND4) (primers cited from Arevalo et al. [35]) and cytochrome C oxidase 1 (COI) (primers cited from Che et al. [36]). PCR amplifications were performed in 25 µL reactions (12.5 µL I-5 2×High-Fidelity Master Mix, 10 µL ddH 2 O, 1 µL F-primers, 1 µL R-primers, 0.5 µL DNA template) under the following cycling conditions: initial denaturation for 2 min at 95 • C, 35 cycles with denaturation at 94 • C for 40 s, annealing at different temperatures (48.5 • C for cyt b and COI, 56 • C for ND4) for 25 s, extension at 72 • C for 15 s, and final extension for 2 min at 72 • C. PCR products were sequenced by Beijing Qingke New Industry Biotechnology Co., Ltd., Beijing, China. Raw trace files for sequences were edited in Genious 7 (Biomatters Limited, Auckland, New Zealand) before constructing alignments using MEGA 7 ( [37]). Due to differences in taxon sampling for each gene, we reconstructed separate alignments for phylogenetic analysis. The first one was based on a concatenated sequence alignment using cyt b and ND4, while the other was based solely on COI. Sequences were uploaded to GenBank (accession numbers see in Table 1). Available cyt b, ND4 and COI sequences of all 12 recognized Asian cobra species and 20 of 21 recognized African cobra species (except Naja nana [38]) were downloaded from Gen-Bank (Table 1). Our DNA dataset employed sequences of 86 specimens from GenBank and 20 specimens collected in this study. Ophiophagus hannah was chosen as an outgroup according to Li [39]. Optimal models of sequence evolution of nucleotide substitution were identified by Bayesian Information Criterion (BIC) using Partition finder 2.1.1 [40]. We performed maximum likelihood (ML) analysis using RaxML v8 [41]. The first ML analysis was implemented in RaxML v8 [41] following the GTRGAMMA model with 1000 fast bootstrap replicates to assess node support. We considered bootstrap proportions of 70% or greater to be strong support for existence of a clade following [42]. The best evolution models are shown in Supplementary Table S1. Bayesian inference phylogenetic trees were inferred using MrBayes 3.2 [43]. We used a random starting tree and four independent runs with a maximum of 20 million generations each, sampled every 1000. Runs were stopped when the average standard deviation of split frequencies had reached 0.001. The first 25% of each run were discarded as burn-in. Nodes with Bayesian posterior probabilities (BPP) of 0.95 and above were considered well supported [44]. Genetic distance between species with uncorrected p-distance model was estimated using MEGA 7.

Morphological Analysis
Measurements of head and head scales were taken with digital calipers and rounded to the nearest 0.1 mm; snout-vent length and tail length were taken with a measuring tape and rounded to the nearest 1 mm. Terminology and descriptions follow Slowinski (1994), Vogel et al., (1994) and Kuch et al. (2005) [45][46][47]. Morphometric and meristic characters are abbreviated as follows: total length (TL), from the tip of snout to the tip of tail; snoutvent length (SVL), from the tip of snout to anterior margin of cloaca; tail length (TaL), from posterior margin of cloaca to the tip of tail; ratio of tail length to total length (TaL/TL); head length (HL), from snout tip to the end of parietals suture; head width (HW) was measured at the widest part of the head on posterior side; head height (HH), at the maximal highest part of the head; interorbital distance (IOD); the dorsal scale rows (DSR) were counted at one head length behind the head, at midbody, and at one head length before the vent; ventral scales (VEN) were counted according to Dowling [48], half ventrals were counted as one. The enlarged shield(s) anterior to the first ventral were regarded as preventral(s); for subcaudals (SC), the first scale under the tail meeting its opposite was regarded as the first subcaudal scale, and the unpaired terminal scute was not included in the number of subcaudals; paired scales on the head were counted on both sides of the head and presented in left/right order; supralabials (SL); infralabials (IL) were considered scales and shields that are completely below a supralabial and border the gap between lips. The number of cross-bands on the body (BB) and on the tail (BT) were counted. Sex was determined by making a small incision caudal to the vent to visually inspect for the existence of hemipenes. Hemipenis terminology follows Dowling (1951), and the organs were prepared based on Jiang [49]. Fangs adjusted for spitting or not were judged according to Wüster and Thorpe [17] and Young et al. [50]. For morphological comparison, a total of 143 specimens from 12 taxa were examined and photos of 208 living cobras from Asia from institutional and personal collections and the Global Biodiversity Information Facility (GBIF) were analyzed (Supplementary Table S2). Morphological data for some congeners of the subgenus Naja were obtained from the literature (Supplementary Table S3). The statistical summaries of scale counts are presented as ranges (Mean ± Standard deviation, Sample size).  For genetic divergence within a clade, samples of "N. kaouthia" with large geographical distance had a small divergence. The samples of the Southeast Asian clade had minor genetic divergences (0.0% to 1.2% for cyt b) between geographical populations from Southern Myanmar, Southern Thailand, and Vietnam, whose largest geographical distance is more than 1300 km. In contrast, the insular species N. sagittifera, which originated in the Andaman Islands, had a larger genetic divergence (2.4-2.5% for cyt b) with the N. kaouthia Southeast Asian clade, which is about 300 km away. The genetic divergence between samples of the N. kaouthia Southwestern China clade were minor (0.0-0.3% for cyt b, 0.0-0.2 for COI) while they are maximally separated by about 500 km. The genetic distances between the samples of N. kaouthia South Asia clade (Rangpur, Bangladesh and Mizoram, India; geographical distance about 400 km) were 0.0-0.6 for COI. These minor genetic distances within clades of N. kaouthia were distinctly smaller than those between the Southwestern China clade and Southeastern Asian clade, which further indicated that the latter is interspecific. From the aspect of genetic distance, considerable divergence existed between different clades under the name N. kaouthia. The Southwestern China clade (F) was divergent from the South Asian clade (L, 2.0%-3.0% for COI) and Southeast Asian clade (A) (2.5%-2.9% for COI, 5.3%-6.7% for cyt b). These are comparable to the genetic distance between of N. kaouthia Southwestern China clade and N. atra (2.5%-3.2% for COI; 4.1%-5.0% for cyt b), N. kaouthia South Asian clade and N. atra (2.7%-3.4% for COI, 4.6%-6.3% for cyt b) (Tables 2 and 3). The latter two species were long regarded as separated species [25,31,53,54]. The genetic distances between the Southwestern China clade and Southeastern Asian clade of N. kaouthia are also comparable for some other known species pairs from Asia: N. siamensis and N. sumatrana (4.3%-5.0% for cyt b), N. siamensis and N. mandalayensis (3.2%-3.9% for cyt b), N. sumatrana and N. mandalayensis (3.2%-3.9% for cyt b). Considerable divergence also existed between the South Asian clade (L) and Southeast Asian clade (A) of N. kaouthia (1.8%-2.4% for COI).

Morphological Results and Taxonomic Conclusion
Comparisons of selected morphological characters and spitting behavior for the subgenus Naja are listed in Supplementary Table S3. The "Naja kaouthia" from Sichuan and Yunnan, Southwestern China is morphologically different from the N. kaouthia from South Asian and Southeast Asian populations by (1) regular single narrow crossband present on the middle and posterior parts of the dorsum (3-15, 7.9 ± 2.7, n = 32) and dorsal surface of tail (1-6, 4.2 ± 1.1, n = 32) of both adults and juveniles, buff-colored with dark fringes on both edges, vs. South Asian populations (n = 39) and Southeast Asian populations (n = 35) without cross bands or with irregular cross bands or multiple light-colored cross bands pairs or densely woven lines; (2) small scales between posterior chin shields were usually three (40%) or two (37%), rarely four (13%), or one (10%, n = 30) vs. one for the South Asian clade (n = 8) and mostly one (72%) and rarely two (28%) for the Southeastern Asian clade (n = 18); (3) ventrals 179-205 (195.4 ± 6.7, n = 33) vs. the South Asian clade 179-199 (188.7 ± 5.9, n = 12) and Southeast Asian clade 168-186 (177.8 ± 4.9, n = 18).   Table 3. Uncorrected P-distance of subgenus Naja based on COI genes. (Values in %). Table 1 In conclusion, the samples of "Naja kaouthia" from Southwestern China represent a new species, which is phylogenetically and morphologically different from N. kaouthia from South Asia and Southeastern Asia.  [   Remark. Naja atra Cantor, 1842 was described from Chusan (=Zhoushan 舟山, Zhejiang Province), China and no holotype has been designated [25,55]. According to Lin et al. [33], samples of Naja atra from vast areas across southern China (except two samples from Yunnan, which are allocated to the new species described in this paper) and Northern Vietnam form two clades with a small divergence (Vietnam + southern China + southwestern China; eastern + southeastern China) based on analysis of 12 microsatellite loci and 1117 bp of the mitochondrial cytochrome b gene. This supports the synonymisation of Naja tripudians var. scopinucha Cope, 1859 (type locality: Canton River), Naja tripudians var. larrata Steindachner, 1867 (type locality: Hongkong), Naja tripudians var. unicolor Peters, 1876 (substitute name), Naia tripudians var. fasciata Boulenger, 1896 (part b) (Kiukiang, Canton, Hainan) with Naja atra Cantor, 1842. To avoid future confusion with other Asiatic Naja species, we designate a neotype from the type locality Zhoushan Island, Zhejiang, China, and redescribe this species based on examined specimens from Southern China. The identification of these examined specimens of N. atra were further confirmed by their localities at or near the sampling localities of clade A & B of Lin et al. [33]. Descriptions are based on 30 specimens examined from Southern China and photos of six unvouchered living individuals from Fujian and Guangdong provinces, China (Supplementary Table S2). Remark. Naja atra Cantor, 1842 was described from Chusan (=Zhoushan 舟山, Zhejiang Province), China and no holotype has been designated [25,55]. According to Lin et al. [33], samples of Naja atra from vast areas across southern China (except two samples from Yunnan, which are allocated to the new species described in this paper) Neotype. CIB 12273, subadult female, collected from Zhoushan Island, Zhejiang Province, China, in July 1983.
Description of Neotype. ( Figure 4A)  Head scalation. Scales smooth. Rostral slightly visible in dorsal view. Nasals large, divided, nostril vertically oval, larger than half the eye, surrounded by prenasal anteriorly and postnasal for the remainder of the sides. Loreal scale absent. Preocular 1/1 (left/right), bordering widely with third supralabial, supraocular, prefrontal, and slightly with eye, internasals and postnasals. Internasal paired, widely in contact with each other. Prefrontal paired, in shape of right trapezoid, in broad contact with one another. Frontal moderate, longer than wide, shield-shaped. Parietals large and paired, largest length about 1.4 times of frontal length, each bordering frontal, upper two supraoculars, anterior and posterior temporals, and two small scales on dorsal head. Eyes rounded, moderate, vertical diameter equal to height of border of fourth supralabial. Postoculars 3/3, middle postocular larger than the two other ones, much smaller than preocular, the lowest widely in contact with fourth and fifth supralabials. Temporals 2+3/2+2; lower anterior temporals hexagonal, longer than high; lowest posterior temporal is largest on lateral head, shape irregular. Supralabials 7/7; first two smallest, bordering nasals; third and fourth higher than wide, bordering eye; the seventh longest. Mental triangular in front and ventral view of head. Infralabials 9/9; first to third in contact with anterior chin shields; the first pair contacting each other, and anterior edge of anterior chin shields; the fourth in contact with both anterior and posterior chin shields; the fifth small and poorly developed, barely seen in ventral view of head, leaving the fourth in broad contact with the sixth; the sixth in contact with posterior chin shields; the seventh to ninth elongated, not contacting chin shields. The two pairs of chin shields butterfly shaped. Only one elongated scale separating posterior chin shield.
Coloration in preservative. Dorsal and upper lateral head brown, lower head paler, sutures of the first to sixth supralabials and fourth to seventh supralabials partially edged with narrow deep brown borders. Dorsum and tail deep brown, 16 light brown crossbands present on body, covering one row of dorsal scales, with narrow dark fringes, crossbands indistinct anteriorly, gradually distinct on posterior body and tail. Seven light brown crossbands present on dorsal tail, with clear dark fringes. Most crossbands on body and tail bifurcate near ventrals, with a dark spot at the base of crossbands. Hood markings mostly pale brown, heart-shaped; with a distinct larger dark core patch and two smaller dark side patches, broadly in contact with pale brown edges of the side of the neck. A small dark patch present on borders of eighth ventral scale. Dorsal scales between hood marking and small dark patch pale brown. Ventral surfaces of head and neck mostly white, stained with light yellow; first broad band on ventrals dark, commencing at 14th ventral and covering almost five ventrals; three ventrals (19th-21st) after first dark band almost yellowish-cream, then becoming gradually brown, completely deep brown by the 24th ventral; ventrals after 24th mostly uniformly deep brown, a few with pale lateral edges due to light brown crossbands on dorsum partially intruding upon ventrals; cloacal shield and subcaudals light brown, paler at places where light brown crossbands on dorsal tail intrude.
Fang. Fangs firm, short, tip reaching middle of fourth supralabial; not modified for spitting, venom discharge orifice relatively large and elongated.
Hemipenes. (Figure 6A,B) Description based on five adult males (Supplementary Table S2). Hemipenis bilobed, slender, covered with dense small spines, smaller and sparser at proximal part, larger and denser on distal part; spines on distal part weakly connected at base, arranged in rows. Two pairs of distinctly developed lateral folds present on the trunk from sulcate view, nearly identical, highly raised and covered with dense spines. Lower lobes located at two-thirds of the trunk, near the base; upper lobes raise at twothirds of the trunk and reach the first third; a transverse constriction divides the lobes from asulcate view. Sulcus forked, centripetally spiral to the top at asulcate side, tip of sulcus not visible from sulcate view. Sulcus lip well developed and raised, covered with dense spines. Hemipenes reaching the 10th-12th subcaudal when not everted, forked at 6th-9th subcaudal; hemipenial retractor muscle reaching 29th-30th subcaudal, forked at 11th-14th subcaudal. Bifurcation ratio 59-67%, hemipenes length/snout-vent length 3.5-4.6%.
Distribution Remarks. Present samples of Naja kaouthia from the South Asian population and Southeast Asian population form sister clades. Since the type locality of N. kaouthia is the Bengal region [57], those specimens from northeastern India, Bangladesh and adjacent areas including Bhutan, Nepal, and southern Tibet Autonomous Region, China should be treated as N. kaouthia sensu stricto [5,[26][27][28][29]. The following diagnoses of N. kaouthia are based on data from South Asia, including data of 13 preserved specimens and 32 living individuals. Measurements and scale counts of three adult females and one adult male were cited from Lalremsanga et al. [58]. Data of living individuals were based on photos of from literatures [28,32,59,60], personal collections of Vogel Gernot and Vivek R Sharma, and GBIF.com) (Supplementary Table S2 (4) small scale between posterior chin shields mostly one (81%), rarely two (19%, n = 28); (5) dorsum of adults mostly brown, juveniles olive-brown, yellowish-brown, or deeper; (6) dorsal middle and posterior body and dorsal tail without crossbands or with irregular cross bands or multiple light-colored cross bands pairs or densely woven lines; (7) throat pale without dark mottling, ventrolateral throat spots usually distinct, usually followed by a broad dark band, the band occasionally light brown; (8) hood pattern usually a monocellate light brown or white circle, often with a large deep colored center and two lateral dots; in few cases, monocellate pattern absent; (9) populations from lowlands of Bangladesh and adjacent area with multiple light-colored crossbands pairs or dense woven lines on body, or few with one or two clear crossbands on neck after hood, populations from mountainous area in southern slope of Himalayan without crossbands on body except few with one or two crossbands after hood, populations from Thailand and peninsular Malaysia light brown, populations from southern Vietnam deep brown; (10) spitting venom.
Fangs. (Description based on examination of one adult female CIB 12285 from southern Tibet, China) Firm and short, not exceeding third supralabial; not modified for spitting, venom discharge orifice relatively large and elongated.
Variation. (Figure 7) The coloration of South Asian clade Naja kaouthia varies between geographical regions. The populations from the lowlands (Odisha, West Bengal and Tripura of India; Bangladesh) ( Figure 7A-E) are yellowish-brown or deep brown dorsally, usually with distinct multiple dense light-colored crossbands, or at least a distinct broad crossband on the neck behind the hood pattern; only one out of 16 individuals uniformly brown dorsally. The populations from southern slope of Himalayan Mountains (Mizoram, Assam, and Sikkim of India; Bhutan, Nepal) uniformly brown in adults and mostly olive-brown or deep brown in juveniles. Hood pattern usually a monocled light brown or white circle, often with a large deep colored center and two lateral dots; in few cases, monocled pattern absent ( Figure 7F) or connected with light colored patches on lateral neck ( Figure 7G). Adults and juveniles from lowlands in West Bengal and adjacent areas without multiple light-colored crossbands on body or solely with one or two clear crossbands on neck after hood markings, only small part of populations from mountainous area along southern slopes of Himalayas without multiple crossbands on body. lateral dots; in few cases, monocled pattern absent ( Figure 7F) or connected with colored patches on lateral neck ( Figure 7G). Adults and juveniles from lowlands in Bengal and adjacent areas without multiple light-colored crossbands on body or s with one or two clear crossbands on neck after hood markings, only small part of p lations from mountainous area along southern slopes of Himalayas without mul crossbands on body.       Holotype. Adult male CIB DL2018053147 ( Figure 4D)  (4) small scales between posterior chin shields mostly three (40%) or two (37%), rarely four (13%) or one (10%) (n = 30); (5) dorsum light brown in adults, deep brown or black in juveniles (n = 32); (6) throat pale usually without dark mottling, ventrolateral throat spots distinct, followed by a broad light brown band; (7) hood markings usually a pale oval marking with narrow dark inner and outer border (73%), sometimes irregular residual patterns of a monocle (20%), rarely indistinguishable (7%), no dark side spots in any individual (n = 30); (8) both adults and juveniles with clear regular single narrow buff cross bands with dark fringes on middle, posterior dorsum 3-15 (7.9 ± 2.7, n = 32) and dorsal tail 1-6 (4.2 ± 1.1, n = 32); (9) hemipenial spines on distal part well connected at base, forming calyculate fold; (10) fangs not modified for spitting, venom discharge orifice relatively large.
From other cobras from South Asia, the new species differs as follows: from Naja naja by (1) hood markings usually form a complete monocle, in a few specimens incomplete, rarely absent vs. spectacle-shaped; (2) throat pattern distinct, lateral spots encroach on lowest dorsal scale row vs. throat pattern indistinct, lateral spots encroach on second dorsal scale row [14]; (3) regular single narrow crossbands present on middle and posterior parts of the dorsum (3-15, 7.9 ± 2.7, n = 32), and dorsal surface of tail (1-6, 4.2 ± 1.1, n = 32) of both adults and juveniles, buff-colored with dark fringes on both edges vs. mostly without regular light-colored crossbands on body (83%), some with brown and yellow densely woven lines on body (17%) (n = 46).
The new species differs from Naja sagittifera by (1) more ventral scales 179-205 (195.4 ± 6.7, n = 33) vs. 175-176 (175.40 ± 0.55, n = ?) () for males, 183 for one female); (2) dorsum coloration light brown in adults, dark brown in juveniles, regular narrow crossbands with dark fringes on body and tail of adults as well as juveniles vs. juveniles dark with irregular or shark-fin-like outlines; (3) distribution confirmed in southwestern China and expected in adjacent Indochina vs. restricted to the Andaman Islands.
Differences from cobras from southeastern Asia: the new species differs from Naja mandalayensis by (1) hood markings usually a distinct pale oval marking with narrow dark inner and outer border vs. hood markings absent; (2) throat pale, followed by a broad light brown crossband vs. throat dark, followed by two or three dark broad crossbands; (3) fangs not adapted for spitting, venom discharge orifice large vs. adapted for spitting, discharge orifice smaller.
The new species differs from Naja siamensis by (1) more ventrals 179-205 (195.4 ± 6.7, n = 33) vs. 153-174 (? ± ?, n = 67); (2) hood markings usually monocellate, a pale oval marking with narrow dark inner and outer border vs. U-, V-, or H-shaped spectacle; (3) dorsum coloration in adults light brown, with single regular buff crossbands with dark fringes on body and tail vs. brightly contrasting black and white pattern in central plain of Thailand, also in Laos; uniformly blackish-brown or black in southeastern and western Thailand.
The new species differs from Naja sputatrix by (1) regular narrow crossbands present on body and tail of adults and juveniles with dark fringes vs. never any light dorsal crossbands (except, on rare occasions, a light band behind the hood); (2) throat area pale, lateral spots distinct vs. no clearly defined light throat area, or very dusky and indistinct; lateral spots often missing; (3) adults usually with clear hood markings vs. usually lacking; (4) venom discharge orifice on fang relatively large vs. relatively small.
Description of holotype. Adult male. Head broad, triangular, widest at middle parietals. Body dimensions: snout-vent length 800 mm; tail length 135 mm, about 14% of total length; head length 30.3 mm (snout to end of parietal suture); maximum head width 19.8 mm, about 65% of head length; maximum head height 12.8 mm, about 42% of head length; distance between eyes 14.2 mm, 47% of head length.
Head scalation. Scales smooth. Rostral nearly U-shaped, slightly visible in dorsal view. Nasals large, including one prenasal and one postnasal, nostril mainly surrounded by postnasal. External nares moderate, higher than wide, half of vertical diameter of eye. Preocular 1/1, bordering widely with third supralabial, supraocular, prefrontal, and slightly with eye, internasals and postnasals. Loreal scale absent. Internasal paired, widely contact with each other. Prefrontal paired, in shape of right trapezoid, widely contact each other. Frontal moderate, shield-shaped, size similar as prefrontal. Parietals paired, large, largest length about 1.6 times of frontal length, each bordering frontal, supraocular, upper postocular, anterior and posterior temporals, and two small scales on upper side of the head. Eyes rounded; moderate, vertical diameter equals height of fourth supralabial. Postoculars 3/3, similar in size, about half of preocular, the lowest widely in contact with the fourth and fifth supralabial. Temporals 2+3/2+3; lower anterior temporals hexagonal, longer than high; lowest posterior temporal largest, shape irregular. Supralabials 7/7; first two smallest, bordering nasals; third and fourth higher than wide, bordering eye; the seventh longest. Mental triangular in front and ventral view of head. Infralabials 9/9; the first to third in contact with anterior chin shields, the first contact with each other, and anterior edge of anterior chin shields; the fourth contacting both anterior and posterior chin shields; the fifth small and poorly developed, barely seen in ventral view of head, leaving the fourth in broad contact with the sixth; the sixth in contact with posterior chin shields; the seventh to ninth elongated, not contacting chin shields. The two pairs of chin shields form a butterfly shape. A total of three scales separating posterior chin shield, arranged in a "Λ"-shape; first of them much smaller than gulars behind, contacting both anterior chin shields; posterior two scales elongated, broadly contacting each other.
Coloration in preservative. Head pattern: Dorsal head uniformly light brown; lateral head gradually paler on supralabials; ventral head uniformly cream white. Dorsal pattern: Dorsum light brown; six buff crossbands present on middle to posterior body, width about length of one dorsal scale, with faint dark fringes, not bifurcate near ventrals. A brown monocellate marking with dark edges and dark oval core present on the hood, not reaching ventrals. Interstitial skin feebly lighter than dorsal scales. Dorsal tail also light brown with four crossbands similar with those on body ( Figure 8A). Ventral pattern ( Figure 8B): throat pale, ventrals 1st-13th cream white; first ventral crossband covering ventrals 14th-19th, faintly brown, the 19th ventral paler; a small dark patch present on ventrolateral neck at lateral edge of 9th-10th ventral and the first row of dorsal scales; ventrals 20th-24th cream; ventrals 25th and following uniformly pale brown except several ventrals more paler at position opposing crossbands on dorsum; coloration on middle and posterior venter similar with coloration on the first broad crossband; two small dark patches present on skin between lateral edge of eighth and ninth with first row of dorsal scales. Ventral tail uniformly pale.
Coloration in life. Similar to coloration in preservative, generally browner. Fangs. Fangs firm, short, not exceed third supralabial; not modified for spitting, venom discharge orifice relatively large and elongated.
Hemipenes. (Figure 6C,D). Description based on three adult males (Supplementary  Table S2). Hemipenis bilobed, covered with dense small spines, spines smaller and sparser at proximal part, larger and denser on distal part; spines on distal part well connected at base, forming calyculate fold. Two pairs of weak lateral folds present on the trunk from sulcate view, nearly identical, covered with dense spines. Lower lobes located at threefourths of the trunk; upper lobes raised at three-fourths of the hemipenes and reaching the base; a transverse constriction divide the lobes from asulcate view. Two pairs of longitudinal folds present above and below the constriction, feebly raised; longitudinal folds above reaching middle of the lobes and joined near the constriction; longitudinal folds below parallel and short, located at three-fourths of the trunk. Sulcus forked, centripetally spiral to the top, tip of sulcus not visible from sulcate view. Sulcus lip well developed and raised, covered with dense spines. Hemipenes reaching eleventh subcaudal when not extruded, hemipenial retractor muscle reaching 23rd subcaudal. Bifurcation ratio 64-71%, hemipenes length/snout-vent length 3.3-4.1%.
Etymology. The new species is named after Fuxi (伏羲), one of the human ancestors in Chinese mythologies, usually depicted in cultural relics as half-man and half-snake. The common postures of Fuxi resembles a cobra in reared status; this species is named for the impact that snake had in human culture. The specific nomen is a noun in apposition. To identify this species quickly when dealing with snakebites, we suggest "Brown Banded Cobra" as a common English name, for its unique coloration with a brown body and light crossbands on the body and tail. For the Chinese name, we suggest " 西南眼镜蛇" (Xī Nán Yǎn Jìng Shé) as it is distributed in the southwestern part of China.
Distribution and ecology. The Brown Banded Cobra is currently known from tropical and southern subtropical areas of southwestern China at elevations between 1000-1400 m. Specimens examined in this study are recorded from Renhe District and Miyi County, Panzhihua Prefecture, southwestern Sichuan Province, and Jiangcheng County, Simao District, Menglian County, Ximeng County, and Pingbian County, Yunan Province. This species was also reported from western Guangxi [30][31][32]. This species is expected to be found in adjacent areas including western Guangxi Zhuang Autonomous Region and southwestern Guizhou Province of China, northeastern Myanmar, northern Laos, northern Thailand, and northwestern Vietnam.
This species was found on gentle slopes of open bush or edges of the forest ( Figure 10G,H) during daytime. One individual was kept in captivity for about one year and lived well, being fed on mice and toads. According to Yang and Rao [61], the brown banded cobra has a wide spectrum of food including frogs, snakes, birds, and small mammals. Local people from Yunnan Province reported that this species sometimes sneaks into villages and preys on chicks. Local people from Menglian County observed five to dozens of individuals gathering in an abandoned termite nest during winter in Yunnan Province [61]. The Brown-Banded Cobra is venomous, and easily provoked. This species is the perpetrator that caused the largest number of snakebites in Xishuangbanna, Yunnan Province, based on an analysis of 126 snake bite cases caused from 2007 to 2014 [62].

The Taxonomy of Some Asian Cobras
The population of Naja naja (=N. polyocellata comb. nov.) from Sri Lanka: Naja naja, is widely distributed throughout the India subcontinent and Sri Lanka [22,25]. There are multiple color and pattern variations within the populations in the present geographical distribution of N. naja, which resulted in the description of five subspecies by Deraniyagala [8][9][10][11] from the Indian subcontinent, namely N. n. gangetica Deraniyagala, 1945 (Gangetic Plain), N. n. madrasiensis Deraniyagala, 1945 (Southern India), N. n. indusi Deraniyagala, 1960 (Northwestern India, Pakistan), N. n. bombaya Deraniyagala, 1961 (Maharashtra of western India), and N. n. karachiensis Deraniyagala, 1961 (a black form from southern Pakistan), and one subspecies from Sri Lanka, N. n. polyocellata Deraniyagala 1939. These subspecies were synonymized with N. naja by Wüster [16] and further discussed based upon a morphometric character analysis [14]. However, the insular population of Sri Lanka is different from continental populations by having 15 or more (sometimes up to 20) dark ventral bands vs. one to four bands [8,10,14]. Furthermore, the venom composition of the Sri Lanka population is different from that of the Indian populations in quality and quantity, whereas the antibody binding affinities towards venom of the Sri Lanka population is lower than that of the Indian populations [4]. The Sri Lanka population is divergent from continental populations from Pakistan and Nepal with distinct genetic distance (4.3-4.9% for cyt b), which is larger than those between some known specie pairs N. siamensis and N. mandalayensis (3.2-3.9% for cyt b), N. sumatrana and N. mandalayensis (3.2-3.9% for cyt b). These indicates that the subspecies N. naja polyocellata Deraniyagala, 1939 should be resurrected and recognized as a full species, N. polyocellata comb. nov. Deraniyagala, 1939. Further taxonomic research that combines molecular, morphological and ecological methods is needed for the N. naja complex.
The populations of Naja sumatrana (=N. s. miolepis) from Borneo and Palawan Islands: These populations were described as a subspecies Naia tripudians miolepis Boulenger, 1896 based on their juveniles having a unique "V"-shaped hood pattern [19]. It was synonymized with N. sumatrana by Wüster [15] based on a morphometric character analysis. The N. sumatrana sample from the Philippines (distributed in Palawan Islands) clustered into a clade with one sample from Malaysia ( Figure 2). Additionally, it is divergent from two samples from Indonesia and Malaysia with genetic distance (2.9-3.4% for cyt b). These support the resurrection of the subspecies N. sumatrana miolepis Boulenger, 1896. However, further studies on adult morphological comparisons and distribution boundary of N. sumatrana sumatrana and N. sumatrana miolepis are still needed to determine whether they are different species.
The population of southeastern Asian Naja kaouthia (= N. naja isanensis?): A creamcolored subspecies, N. kaouthia suphanensis, Nutaphand,1986 was described from central Thailand [63]. The subspecies was then synonymized with N. kaouthia by Wüster [13] based on a multivariate morphometric analysis between N. k. suphanensis and N. kaouthia from central Thailand, but no specimens from or near the type locality of N. kaouthia, the Bengal region [57], were included in the analysis. The southeastern Asian N. kaouthia population from Myanmar, Thailand, and Vietnam differs from N. kaouthia from the lowlands in Bengal and adjacent areas by usually no crossbands on body and a few (usually one or two) clear crossbands on neck after the hood marking (21%) (n = 24) vs. parallel light-colored crossband pairs or densely woven lines on body and tail, or at least with a clear crossband on the neck behind the hood markings (n = 24)). However, the coloration of southeastern Asian population is similar to populations of N. kaouthia from the southern slope of the Himalayan Mountains and the Southeast Asian N. kaouthia population do not spit venom (based on the observation of more than 100 individuals), whereas multiple cases (n > 16) were recorded in populations from eastern India and Nepal [59,64]. A cryptic species was proposed from the northeastern population of N. kaouthia by Ratnarathorn et al. [5]. This population is the phylogenetic sister to N. atra ("N. naja") and N. kaouthia from Thailand based on the Control Region with weak support (BPP 0.65) (Figure S1 of [5]). However, the northeastern population is sister to N. siamensis based on partial cyt b (about 603 bp) with much stronger support (BPP 0.986) ( Figure S2 of [5]). Due to only four operational taxonomic units (OTUs) (N. atra, N. siamensis, the northeastern population of N. kaouthia, and N. kaouthia) were included in the analysis, the phylogenetic position of the cryptic species is still doubtful. A brown spitting population of cobra with indistinct spectacle mark from northeastern Thailand has been described as N. naja isanensis Nutaphand, 1982 [65] and it was treated as a color variety of N. siamensis by Wüster and Thorpe [66]. This population is possibly phylogenetically the sister to N. siamensis, and morphologically resembles "N. kaouthia" from Thailand due to the brown coloration and indistinct spectacle mark, so, it is possible the nomen Naja naja isanensis represents the northeastern population of N. kaouthia in Ratnarathorn et al. [5]. Since the samples of N. kaouthia from Myanmar to Vietnam form a clade (A; 98/1.00) sister to insular species N. sagittifera and have minor genetic divergence within clade (0.0% to 1.2% for cyt b), these samples should represent the same species or subspecies. However, the name for this taxon remains unsolved; the status of subspecies N. k. suphanensis and N. n. isanensis needs to be further studied.

Geographical Variations of N. kaouthia
The coloration form of N. kaouthia varies among geographic regions. Those specimens from the southern slope of the southeastern Himalayas (Sikkim, Bhutan and southern Tibet, China) have typically monocellate hood markings and uniformly brown dorsum coloration. However, specimens from West Bengal and adjacent Odisha, India demonstrated clear crossbands of various forms. Similar situation is also found in N. atra, no mitochondrial gene differentiation was apparent among different ventral color morphs, the color morph is more related to the geographic populations [33]. The venom composition of N. kaouthia varies between the southern Himalayan population and lowland population from West Bengal [6]. These geographic variations might be related to ecological differences. To clarify these questions, biogeographic research for N. kaouthia based on comprehensive sampling in these areas is highly recommended to build a strong framework for snakebite medical treatment and antivenin development.

Spitting Behavior and Venom Discharge Orifice
The venom discharge orifice on fangs of N. atra, N. kaouthia and N. fuxi are relatively large and elongated. N. atra and N. kaouthia were listed as non-spitting species according to Wüster and Thorpe [16] and Young et al. [50]. However, there are cases reported for N. kaouthia spitting venom in eastern India and Nepal [59,64]. There is also a case for N. atra from southern China spitting venom with description for relative venom discharge orifice [56]. According to personal observation on N. fuxi sp. nov. by Liang Zhang, the cobra could spit venom as far as about 0.6 m. The venom discharge orifices of the Asian cobras in Clade I (except N. sumatrana miolepis, unknown) are small, while those for species in clade D are relatively larger [17,56,67] (this study). The Asian spitter clade J originated~2.5 Ma, which is contemporaneous with the beginning of the quaternary glaciation [51,68]. The less spitting-adapted species of clade D with larger venom discharge orifices, also orig-inated~2.5 Ma ( Figure S14 of [51]). This suggests that venom spiting behavior evolved independently twice in subgenus Naja and the latter one (clade D) might be relevant to climate change and distribution change of H. erectus during the quaternary glaciation.

Conclusions
Our results led to taxonomic revisions of some Asian cobras: (1) the former Chinese population of N. kaouthia represents a new species, N. fuxi sp. nov.; (2) the subspecies N. naja polyocellata was resurrected and recognized as a full species N. polyocellata comb. nov.; (3) the subspecies N. sumatrana miolepis was resurrected. This study highlights the necessity to evaluate effectiveness of cobra antivenin based on comprehensive taxonomic frameworks. However, there are still some open questions about the systematics of widespread species such as N. kaouthia. The taxonomy of Asian cobras still needs further revision. A wide international collaboration network is recommended to construct a solid taxonomic framework of these medically important species for the common well-being of Asian people.  [12,14,17,18,28,32,51,56,[58][59][60]64,69,70] are cited in the supplementary materials.

Data Availability Statement:
The data presented in this study are available on request from corresponding author.