INTRODUCTION

The genus Lycodon Boie is one of the most diverse genera of colubrid snakes, with 50 currently recognized species (Uetz et al., 2017). The members of this genus have a broad distribution from eastern Iran to southern China and Japan, southward to the Philippines as well as the Indo-Australian Archipelago (Lanza, 1999; Siler et al., 2013; Neang et al., 2014). Six species of Lycodon have been described within the last five years, namely Lycodon synaptor Vogel & David, 2010, L. gongshan Vogel & Luo, 2011, and L. liuchengchaoi Zhang, Jiang, Vogel & Rao, 2011 from China; L. davidi Vogel, Nguyen, Kingsada & Ziegler, 2012 from Laos; L. zoosvictoriae Neang, Hartmann, Hun, Souter & Furey, 2014 from Cambodia; and L. cavernicolus Grismer, Quah, Anuar, Muin, Wood & Nor, 2014 from Malaysia.

From Laos, ten species of Lycodon have been reported to date, comprising L. capucinus (Boie, 1827), L. davidi Vogel, Nguyen, Kingsada & Ziegler, 2012, L. fasciatus (Anderson, 1879), L. futsingensis (Pope, 1928), L. laoensis Günther, 1864, L. meridionalis (Bourret, 1935), L. rufozonatus Cantor, 1842, L. ruhstrati abditus Vogel, David, Pauwels, Sumontha, Norval, Hendrix, Vu & Ziegler, 2009, L. septentrionalis Günther, 1875, and L. subcinctus Boie, 1827 (Deuve, 1970; Vogel et al., 2012; Siler et al., 2013; Luu et al., 2013; Teynié et al., 2014).

Our recent field survey in the karst forest of Phou Hin Poun National Protected Area (NPA), Khammouane Province, central Laos led to the discovery of a snake, which could be identified as a member of the genus Lycodon based on the following characters: eye with a vertically elliptical pupil; nostril enlarged; robustly arched upper maxillary bone with an inward curve in the anterior part; anterior and posterior maxillary teeth interrupted by a diastema; dorsal scales smooth or weakly keeled, in 17 rows anteriorly and at midbody, and 15 rows posteriorly; ventral scales weakly notched (Lanza 1999; Grismer et al., 2014). Although only a single specimen was collected, it proved to be morphologically distinct. Our finding is corroborated by another record, a closely resembling specimen photographed in about 12.4 km distance. The morphological results are further supported by molecular analyses so that we describe the single adult male from Phou Hin Poun, Khammouane Province, central Laos as a new Lycodon species.

MATERIAL AND METHODS

Sampling: The field survey was conducted by Vinh Quang Luu and Thomas Calame in Phou Hin Poun NPA, Khammouane Province, central Laos in April 2016. The collected specimen was fixed in approximately 85% ethanol, and subsequently transferred to 70% ethanol for permanent storage. Liver tissue sample was preserved separately in 95% ethanol. The specimen from Phou Hin Poun NPA, central Laos was deposited in the collections of the Vietnam National University of Forestry (VNUF), Hanoi, Vietnam. Another specimen was photographed on 22 July 2016 by an arachnology team led by Peter Jaeger, in ca. 12.4 km distance from the type locality.

Morphological analysis: Measurements were taken following Vogel et al. (2009) with a digital caliper to the nearest 0.1 mm, except body and tail lengths. These measurements included: head length (HL, from snout tip to jaw angles); head width (HW, maximum head width at posterior margin of parietals); head height (HH, vertical height between upper and under sides of head were measured at HW); interorbital distance (IO, the distance between outer edges of supraoculars); eyenostril distance (EN, from anterior edge of the orbit to posterior edge of nostril); internarial distance (IN, horizontal diameter between nostrils); eye diameter (ED, horizontal diameter of the orbit); snout length (SnL, from the tip of rostral to the anterior edge of the orbit); snout-vent length (SVL, from tip of snout to the vent); tail length (TaL); ratio of tail length / total length (TaL/TL); total length (TL).

Scale counts were taken following Vogel et al. (2009). Ventral scales (VEN) were counted according to Dowling (1951); dorsal scale rows (DSR): number of dorsal scale rows at neck (ASR, at one head length behind head), number of dorsal scales at midbody (MSR), and number of dorsal scale rows before the vent (PSR, at one head length before the vent); supralabials (SL, counted on upper lips); infralabials (IL, counted on lower lips); loreals (Lor); loreal scale touching the orbit (yes or no); preoculars (PreOc); postoculars (PosOc); temporals (Temp, counted immediately behind postoculars and between posterior SL and parietals). Bilateral scale counts were given as left/right. Keel (keeled dorsal scale rows); PreVEN (preventral scales); VEN notched (present or absent); VEN keeled (present or absent); SC (subcaudal scales); numbers of pattern-units (like crossbars or vertebral blotches) are provided as number on body + number on tail.

For comparisons, we referred to the data provided by Boulenger (1893), Smith (1943), Orlov & Ryabov (2004), Neang et al. (2014), and Grismer et al. (2014); studied specimens are listed in the Appendix.

Museum abbreviations are as follows: CAS-California Academy of Sciences; GP-Specimens in the collection of Peng Guo; LSUMZ-The Louisiana State University Museum of Natural Science; MNHN-Muséum National d'Histoire Naturelle, Paris, France; VNUF-Vietnam National University of Forestry.

Molecular data and phylogenetic analyses: The mitochondrial cytochrome b gene was employed in this study, because it has been widely used in previous molecular analyses of Lycodon (e.g., Guo et al., 2013, Siler et al., 2013). We included six new sequences from samples collected in Laos and Vietnam (Table 1). Other sequences of related species were obtained from GenBank. Three species, Ahaetulla prasina, Boiga cynodon, and Dispholidus typus, were assigned as outgroups based on their phylogenetic relationships to the genus Lycodon (Guo et al., 2013, Siler et al., 2013) (Table 1).

We used the protocols of Le et al. (2006) for DNA extraction, amplification, and sequencing. A fragment of the mitochondrial cytochrome b was amplified using the primer pair L14910/H16064 (Burbrink et al., 2000). After sequences were aligned by Clustal X v2 (Thompson et al., 1997), data were analyzed using maximum parsimony (MP) and maximum likelihood (ML) as implemented in PAUP*4.0b10 (Swofford, 2001) and Bayesian analysis (BA) as implemented in MrBayes v3.2 (Ronquist et al., 2012). Settings for these analyses followed Le et al. (2006), except that the number of generations in the Bayesian analysis was increased to 1´10⁷ and the number of bootstrap replicates in ML to 1000. The optimal model for nucleotide evolution was set to TrN+I+G for ML and combined Bayesian analyses as selected by Modeltest v3.7 (Posada & Crandall, 1998). The cutoff point for the burn-in function was set to 19 in the Bayesian analysis, as -lnL scores reached stationarity after 19,000 generations in both runs. Nodal support was evaluated using Bootstrap replication (BP) as estimated in PAUP and posterior probability (PP) in MrBayes v3.2. BP ≥⃒ 70% and PP ≥⃒ 95% are regarded as strong support for a clade. Uncorrected pairwise divergences were calculated in PAUP*4.0b10 (Table 2).

Table 1.

Lycodon samples used in the molecular analyses (for abbreviations see Material and methods); * = listed as L. aulicus in Siler et al. (2013); ** = listed as L. fasciatus in genbank but as L. cf. fasciatus in the CAS catalogue.

Table 2.

Uncorrected (“p”) distance matrix showing percentage pairwise genetic divergence (cytochrome b) between new and closely related species.

RESULTS

Molecular data, Phylogenetic analysis: The final matrix consisted of 1100 aligned characters, of which 440 were parsimony informative. The alignment did not contain gaps. Maximum parsimony analysis of the dataset recovered two most parsimonious trees with 1649 steps (CI = 0.46; RI = 0.65). In the ML analysis, the score of the single best tree found was 8329.12 after 2678 arrangements were tried. The topology derived from the Bayesian analysis (Fig. 1) was similar to that in Guo et al. (2013), but nodes of the phylogeny received lower statistical support. The new species was recovered in a clade together with other species, which were previously placed in the genus Dinodon (see Siler et al., 2013; Guo et al., 2013). This clade was strongly supported by both MP and Bayesian analyses (BP = 86%, PP = 97%) (Fig. 1). The new species is most closely related to a clade containing ‘L. flavozonatus’, L. futsingensis, L. meridionalis in terms of genetic distance based on cytochrome b, and is diverged at least about 9.0–9.2% from the latter species (Table 2).

TAXONOMIC ACCOUNT

Fig. 1.

Bayesian cladogram based on the partial cytochrome b gene. Numbers above and below branches are bootstrap values of MP/ ML analyses (>50%) and Bayesian posterior probabilities, respectively. Asterisk denotes 100% value.

Table 3.

Measurements (in mm) and morphological characters of the holotype of Lycodon banksi sp. nov. (measurements in mm; for other abbreviations see material and methods; * tail tip lost).

Fig. 2.

Adult male holotype of Lycodon banksi sp. nov. (VNUF R.2015.20) in life. (A) dorsolateral view. (B) Head in dorsolateral view. (C) Head in dorsal view. Photos: V. Q. Luu.

Fig. 3.

Different head views of the adult male holotype of Lycodon banksi sp. nov. (VNUF R.2015.20). Photos V. Q. Luu.

Fig. 4.

Lateral head view of the adult male holotype of Lycodon banksi sp. nov. (VNUF R.2015.20). Drawing by T. Ziegler.

Fig. 5.

Additional record of Lycodon banksi sp. nov. in life: (A- B) dorsolateral views. (C) Head in dorsal view. Photos A-C A. Sahu. (D) Dorsal view. Photo P. Jaeger.

DISCUSSION

In our phylogenetic analyses, Lycodon banksi is placed in a clade with other species previously considered to be members of the genus Dinodon. In addition, the specimen of L. meridionalis from Bac Kan Province, Vietnam was nested in the same clade with ‘L. flavozonatus’ from Guangdong and Guangxi provinces in southern China. The genetic distance between the Vietnamese and Chinese samples is approximately 1.4–2.4% (2.6% between two Chinese samples). Morphological features of the specimens from Bac Kan, Ninh Binh, and Thanh Hoa provinces were consistent with those in the descriptions of L. meridionalis by Bourret (1935) and Orlov & Ryabov (2004) in the following characters: snout-vent length in males reaching 1295 mm; dorsal head with yellow-black marble markings; transverse bands on body 86–115; ventral scales 234–245; cloacal plate single; belly pattern uniform yellow with dark spots posteriorly (Table 4). Therefore, based on the molecular data, we herein initially assign two specimens (‘L. flavozonatus’ GP 1939, 2279) from China to L. meridionalis, although this placement needs to be confirmed by further morphological studies. Superficially, the new species is similar to L meridionalis in dorsal pattern. However, they are clearly distinguishable in other morphological features, e.g., dorsal scalation, dorsal head, and belly patterns. Although the new species has no clear sister species according to the phylogenetic analyses, it is most closely related to L. meridionalis in terms of genetic distance, but distinctly differing from the latter (ca. 9%).

Fig. 6.

Map showing the type locality of Lycodon banksi sp. nov. in Khammouane Province (black dots) and other records of the genus Lycodon occurring in Laos (after Deuve, 1970; Vogel et al., 2012; Siler et al., 2013; Luu et al., 2013; Teynié et al., 2014).

Fig. 7.

Habitat of Lycodon banksi sp. nov. at the type locality. Photo V. Q. Luu.

The new species seems to be a nocturnal and terrestrial snake, and rare. It is a karst-dweller in the northern Truong Son Range, as are L. davidi and L. ruhstrati abditus (Vogel et al., 2012; Luu et al., 2013). L. banksi has a unique pattern of dark grey and yellow colouration on its dorsal surface, which offers a perfect camouflage among litter on the forest ground and on karst surface. Moreover, the species has indistinct banding on the dorsal surface in comparison with its congeners of the L. ruhstrati and L. fasciatus groups which are clearly banded.

The discovery of this new species increases the number of Lycodon species known from Laos to eleven. In the same area, we recently discovered two new bent-toed geckos (Cyrtodactylus jaegeri Luu, Calame, Bonkowski, Nguyen & Ziegler, C. soudthichaki Luu, Calame, Nguyen, Bonkowski & Ziegler) and two new species of the genus Gekko (G. thakhekensis Luu, Calame, Nguyen, Le, Bonkowski & Ziegler, G. bonkowskii Luu, Calame, Nguyen, Le, Bonkowski & Ziegler) (Luu et al., 2014a, b, 2015a, b). These recent discoveries together with the new Lycodon species provide strong evidence that the northern Truong Son Range, and especially the extensive limestone karst formations in Khammouane Province, central Laos represent a hotspot of endemic biodiversity. The finding of this new Lycodon species with its unique colour pattern and morphology suggests that the species diversity within the genus Lycodon might be far greater than commonly assumed. Our study further underlines the need to combine detailed morphological and molecular data to delineate species complexes and better understand kinship. Unfortunately, it was not possible to obtain molecular data from all investigated reference species (e.g., Lycodon fasciatus MNHN 1928.69 from Xieng Khoang Province, northern Laos) due to formalinfixed state and long-term storage. Therefore, further field studies are needed to accurately assess the diversity of Lycodon in the poorly studied karst forest systems of central Laos.