The taxonomic status of the kukri snake Oligodon arenarius Vassilieva, 2015 with a redescription of Oligodon macrurus (Angel, 1927) (Squamata, Serpentes, Colubridae)

We investigated the taxonomic status of the recently described kukri snake Oligodon arenarius Vassilieva, 2015 and the morpho­ logically similar Oligodon macrurus (Angel, 1927), two species endemic to the southern coast of Vietnam. Based on phylogenetic analyses using three mitochondrial genes (12S–16S rRNA, cytochrome b ), we recovered O. arenarius and O. macrurus in a clade within the O. cyclurus-taeniatus species group, agreeing with previous intrageneric classifications. Genetic distances between O. arenarius and O. macrurus are extremely low (less than 0.5% based on 12S–16S) and render O. arenarius paraphyletic. All preserved specimens of O. arenarius and O. macrurus convey little to no differences in color pattern, hemipenial morphology and osteological features; the latter of which is based on three dimensional micro computer tomography (µCT) scans of one specimen per species. Contrasting these results, univariate and multivariate analyses revealed significant differences in relative tail length, and the number of ventral and subcaudal scales between both species. Although the molecular and morphological datasets present conflicting results, integrating the evidence leads us to synonymize O. arenarius with O. macrurus . We provide a formal redescription of O. macrurus , designate a neotype specimen to avoid future taxonomic confusion, and provide the first detailed osteological description of this species. Oligodon macrurus sensu stricto is endemic to coastal dunefields and adjacent forest habitats in southern Vietnam, where ongoing human development, tourism and road mortality pose significant threats to its conservation. Consequently, we suggest that O. macrurus should be listed as “Vulnerable” based on the assessment criteria of the International Union for Conservation of Nature (IUCN).


Introduction
The kukri snakes of the genus Oligodon Fitzinger, 1826 are a speciesrich group of colubrids comprised of 89 rec ognized species (Nguyen et al. 2022;Uetz et al. 2022;Yushchenko and Lee et al. 2023).Although the genus is taxonomically diverse, most members are infrequently collected in the field and are known from only a few spec imens or their namebearing types (Supsup and Carestia 2020;Lee et al. 2021).As a result, most lack information regarding their basic biology, morphological variation, phylogenetic position and conservation status.Rough ly 24 species of Oligodon are known from Vietnam.Of these, eight are considered endemic, and ten have been described in the last decade (David et al. 2008a(David et al. , 2008b(David et al. , 2012;;Vassilieva et al. 2013;Vassilieva 2015;Nguyen et al. 2016Nguyen et al. , 2017Nguyen et al. , 2022;;Nguyen et al. 2020).Yushchenko and Lee et al. (2023) recently reassessed the taxonomy of one newly described Vietnamese species, Oligodon cattienensis Vassilieva et al., 2013, resulting in its syn onymziation with the more wideranging and morpholog ically variable Oligodon cinereus Günther, 1864.In addi tion, Nguyen et al. (2022) described another new species, Oligodon tuani Nguyen et al., 2022 from the Langbian Plateau of southern Vietnam.These studies indicate our knowledge of Oligodon in this country is far from com plete, and taxonomic revisions are still needed to under stand the species boundaries of many of its members.In this study, we reassess the status of two poorly docu mented kukri snakes known only from the coastal dune fields of southern Vietnam: the Angel's kukri snake Oligodon macrurus Angel, 1927 and the recently described dune kukri snake Oligodon arenarius Vassilieva, 2015.
The first species, O. macrurus, was described as Simotes violaceus macrurus based on a single specimen col lected in 1925 by M. Pierre Chevey from "Pointe Lagan", now the sand dunes surrounding Mui La Gan, Binh Thu an Province, Vietnam.The type specimen was deposit ed in the Museum Nationale d'Histoire Naturelle, Paris (MNHN), but according to Nguyen et al. (2009), it has not been traced in the collection and is thereby consid ered lost (Uetz et al. 2019, confirmed therein by I. Ineich March 2019).Smith (1943) elevated this taxon to species status as Oligodon macrurus and reported a second speci men collected 250 kilometers north of the type locality in what is now Khanh Hoa Province.Based on morphology, Smith (1943) included O. macrurus within the informally defined Oligodon cyclurus species-group specifically due to the presence of forked (bilobed, henceforth) hemipenes without obvious spines or "papillae" (lobular appendages found on the retracted hemipenes when insitu).Roughly 70 years later, Geissler et al. (2011) rediscovered this spe cies from Mui Ne, Binh Thuan Province, approximately 100 kilometers west of the type locality and presented ad ditional morphological data on three specimens.Vassilieva (2015) described the second species, O. arenarius, based on five specimens (two males and three females) from the coastal dunes surrounding Binh Chau-Phuoc Buu Nature Reserve in Ba Ria-Vung Tau Province.Although the type series of O. arenarius share similar color and head scalation features with O. macrurus, Vassilieva (2015) noted several differences between them.A key dissimilarity was the unusual structure of the hemipenes in the male holotype of O. arenarius (ZMMU Re14503), which was unilobed without spines or obvi ous 'papillae'.The two species also had markedly differ ent numbers of ventral scales, subcaudal scales and rela tive tail lengths, further supporting their differentiation as separate entities.The unique hemipenes of O. arenarius prevented Vassilieva (2015) from allocating it into any previous species grouping in Oligodon, but she surmised the O. cinereus and O. dorsalis species groups might be close relatives.Later, Nguyen et al. (2021) collected an other male O. arenarius from the type locality and found that its hemipenial morphology was dramatically differ ent from the description of Vassilieva (2015), bearing a long and deeply bilobed organ compared to the short and unilobed structure of the original male holotype.There fore, these authors (Nguyen et al. 2021) recommended placing O. arenarius into the O. cyclurus species group, which shares the presence of a deeply bilobed hemipenes without spines or 'papillae'.
Over the past decade, some of us (PVY, NAP, HMP, PG) have collected additional kukri snake specimens from the dunefields of southern Vietnam that putatively resem ble the original descriptions of O. macrurus and O. arenarius.Neither species has been sampled in a molecular phylogenetic assessment, and the descriptions of the O. arenarius hemipenes from Vassilieva (2015) and Nguy en et al. (2021) present two conflicting hypotheses.These circumstances prompted us to investigate the taxonomy of these snakes more carefully.We developed three core objectives for our study.First, we aimed to assess the phylogenetic position of both species to confirm whether their closest relatives are members of the morphologically defined O. cyclurus species group.Second, we wished to resolve the hemipenial morphology of O. arenarius to see if the results of Vassilieva (2015) or Nguyen et al. (2021) were correct.Third, we wished to assess the species boundaries between O. arenarius and O. macrurus under an integrative taxonomic approach to determine if they should be recognized as separate species.By examination of all available specimens of O. arenarius and O. macrurus, plus our newly collected material, and phylogenetic analyses of 16 samples using 3,033 bp of mtDNA, we re covered poor support for the species status of O. arenarius and relegate this taxon to the synonymy of O. macrurus.

Material and Methods
Taxon sampling and species delimitation by us.Data from an additional O. arenarius and four O. macrurus (including the lost holotype) originate from lit erature sources (Angel 1927;Geissler et al. 2011;Nguy en et al. 2021), increasing our dataset to a total of 24 specimens.One other specimen of O. arenarius (ZMMU NAP10091) was excluded from the morphological data set because it was discovered dead on road (DOR) in too poor of a condition to examine for morphological fea tures.All O. arenarius are from the vicinity of the type locality in the Binh Chau-Phuoc Buu Nature Reserve, Ba Ria-Vung Tau Province (locality 2; Fig. 1).All ex amined O. macrurus originate from seven localities: vi cinity of Ho Chi Minh City (locality 1; Fig. 1); near Phan Thiet city, Binh Thuan Province (locality 3; Fig. 1.); Mui Ne, Binh Thuan Province (locality 4; Fig. 1); type locali ty, Mui La Gan, Binh Thuan Province (locality 5; Fig. 1); Nui Chua National Park, Ninh Thuan Province (locality 6; Fig. 1); Nha Trang, Khanh Hoa Province (locality 7; Fig. 1); Hon Tre Island, Khanh Hoa Provice (locality 8; Fig. 1).More details are provided in the redescription of O. macrurus.For molecular analyses, we sampled tissues from nine O. arenarius and seven O. macrurus (16 total samples).We also used 47 sequences of other Oligodon species publicly available from GenBank (see Table 2 for sources of each sequence), with the exception of Oligodon melaneus Wall, 1909(Das et al. 2022).For outgroups, we chose one sequence each of the snake spe cies Oreocryptophis porphyraceus (Cantor, 1839) (sub family Colubrinae) and Hebius vibakari (Boie, 1826) (subfamily Natricinae) due to their use in previous phy logenies of Oligodon published in the literature (Nguyen et al. 2020).
For all aspects of species concepts and delimitation, we follow the General Lineage Concept (De Queiroz 2007), where a species represents a single evolutionarily inde pendent lineage following a separate trajectory compared to its congeners.Specifically, our definition of evolution ary independence is measured by substantial differentia tion in genetic divergence, discrete or statistical differenc es in morphological features, and evidence of monophyly from basic genetic markers, satisfying the criteria of an integrative taxonomic approach (Padial et al. 2010).The concept of the subspecies rank, and its potential appli cability towards the taxa studied here, is addressed later (see Discussion).Lastly, we note that we shall eventually consider O. arenarius to be a synonym of O. macrurus, however we will refer to it by its full species combina tion until the results and subsequent sections.

Molecular analyses
Total genomic DNA was extracted from muscle tis sue samples preserved in 95% ethanol using a Qiagen DNAeasy Blood and Tissue Kit following manufacture protocols.We performed polymerase chain reactions (PCRs) to amplify two fragments of mitochondrial DNA (mtDNA): the first fragment including partial sequences of 12S ribosomal RNA (rRNA), transfer RNA (tRNA)-Valine and 16S rRNA genes (total length 1942 bp), and the second fragment including complete sequences of the cytochrome b (cyt b) gene (1091 bp).Primers used for both PCRs and sequencing are summarized in Table 1.PCR protocol for 12S-16S rRNA roughly follows Green et al. (2010).For both primer pairs of 12S-16S rRNA, we used the following PCR protocol: (1) initial denaturation step at 94°C for 5 min; (2) 35 cycles of denaturation at 94°C for 1 min, annealing at 55°C for 1 min and exten sion at 72°C for 1 min; (3) final extension at 72°C for 10 min; and (4) cooling step at 4°C for storage.For cyt b se quences, we used a modified PCR protocol from Chen et al. (2014) with touchdown: (1) initial denaturation step at 94°C for 5 min; (2) 10 cycles of denaturation at 94°C for 1 min, annealing for 1 min with temperature decreasing from 50°C to 45°C (with cooldown at 0.5 °C per each cycle) and extension at 72°C for 1 min; (3) 24 cycles of denaturation at 94°C for 1 min, annealing at 45°C for 1 min and extension at 72°C for 1 min; (4) final extension at 72°C for 10 min; and (5) cooling step at 4°C for stor age.All PCR products were sequenced in both directions by the "Evrogen" company at the Institute of Bioorganic Chemistry, Russian Academy of Science (Moscow, Rus sia).Obtained sequences were deposited in GenBank under accession numbers OP752557-OP752577 and OP764580-OP764588 (Table 2).
To examine the position of O. macrurus and O. arenarius in a mtDNA phylogeny of Oligodon, we aligned our newly obtained sequences with other available Oligodon sequences from GenBank (summarized in Table 2).The final sequence alignment contained 65 specimens, includ ing 63 samples of 28 Oligodon species and two outgroup sequences of O. porphyraceus and H. vibakari, which were used to root the tree.Nucleotide sequences were ini tially aligned in MAFFT v.7 (Katoh and Standley 2013) with default parameters, and subsequently checked by eye in BioEdit 7.0.5.2 (Hall 1999) and slightly adjusted for translation when appropriate.We used the IQTREE web server (http://iqtree.cibiv.univie.ac.at ;Trifinopoulos et al. 2016) to estimate optimal evolutionary models for our phylogenetic analyses.MODELTEST v.3.6 (Posada and Crandall 1998) was applied to the alignment to esti mate optimal evolutionary models for the dataset analy ses under the Akaike Information Criterion (AIC).Mean uncorrected genetic distances (pdistances) were calcu lated in MEGA 7.0 (Kumar et al. 2016).Bayesian infer ence (BI) and Maximum Likelihood (ML) approaches were used to generate phylogenetic trees.The best-fitting model based on the AIC for both BI and ML analyses for 12S-16S rRNA fragments and the second codon parti tion of cyt b was the GTR+G+I model of DNA evolution.For the remaining partitions of cyt b, the AIC suggest ed the GTR+G model for the first codon partition, and the HKY+G+I model for the third codon partition.ML analyses were conducted using the IQTREE web server.Confidence in nodal topology was estimated by non-para metric bootstrapping (ML BS) with 100 pseudoreplicates (Felsenstein 1985).BI analyses were conducted in Mr Bayes 3.2.2(Ronquist et al. 2012), and ran with Metrop oliscoupled Markov chain Monte Carlo (MCMCMC) analyses with one cold chain and three heated chains for one million generations sampled every 1000 gener ations; two independent MCMCMC run iterations were performed, and 100 trees were discarded as burnin.The convergence of the runs was checked by examining like lihood plots in TRACER v1.6 (Rambaut et al. 2020) and the effective sample sizes (ESS) were all above 200.Nod al support was assessed by calculating posterior proba bilities (BI PP).Low ML BS and BI PP values (les than 50% and less than 0.90, respectively) were regarded as unresolved nodes (fide Huelsenbeck and Hillis 1993).

Morphological analysis
We included all examined O. arenarius (n = 9) and O. macrurus (n = 11), along with literature data from four specimens in our morphological dataset (n = 24).In all examined specimens, sex was determined by a ventral in sertion near the tail base to detect the presence or absence of hemipenes.Color characters and patterns were record ed during examination of specimens in preservation.For some specimens these data were taken from digital im ages that depicted their coloration in life.Morphological characters and morphometric ratios considered to be of taxonomic importance in Oligodon were used for species descriptions and followed a number of recent taxonomic treatments (David et al. 2008a(David et al. , 2008b(David et al. , 2011(David et al. , 2012;;Vas silieva et al. 2013;Vassilieva 2015;Nguyen et al. 2016Nguyen et al. , 2017Nguyen et al. , 2022;;Nguyen et al. 2020;Qian et al. 2021;Mirza et al. 2021;Amarasinghe et al. 2021;Bandara et al. 2022;Das et al. 2022).Body measurements such as Snoutvent length (SVL), tail length (TailL) and total length (TotalL) were taken using a flexible ruler.All other measurements were taken under a Leica dissecting microscope using Mitutoyo digital slidecalipers to the nearest 0.1 mm.Specimens were not examined by a single observer (JLL measured NHMUK specimens, Tan Van Nguyen mea sured DTU specimens, Morris Flecks and PG measured ZFMK specimens, PVY measured all other material).Dorsal scales were counted anteriorly at one head length behind the head, at midbody (halfway between the ante rior terminus of the head and the vent), and posteriorly at one head length anterior to the cloacal plate (given as anteriormidbodyposterior in the description); ventral scales were counted according to Dowling (1951); the terminal scute was not included in the number of subcau dal scales; counts for head scales are given in left/right order.The number of total body scales is the sum of the number of ventral scales, the cloacal plate (considered a single scale regardless of whether it is single or divided), and the number of subcaudal scales.Head scale suture angle descriptions are adapted from Kaiser et al. (2019).When possible, hemipenes were everted from freshly pre served male specimens using the protocols outlined by Jiang (2010).Information about structure and size of the hemipenis was taken from preserved specimens both by examination of partially everted and insitu organs, and from literature data.Terminology for hemipenial struc tures follows Dowling and Savage (1960).We use the term myoectases (coined from Wagner 1975) instead of the term "papillae" to describe appendages observed on the retracted hemipenes insitu in some Oligodon spe cies.These structures tend to transform into awnshaped lobular protrusions when the organ is partially or fully everted.
The micro Computer Tomograph (µCT) scans of O. arenarius and O. macrurus were performed at the Pale ontological Institute of the Russian Academy of Scienc es (RAS) using a SkyScan 1172 Computer Tomography scanner (Bruker microCT, Kontich, Belgium).Scans were generated from one adult specimen of O. arenarius (ZMMU Re14502, paratype) and one adult specimen of O. macrurus (ZMMU Re13857), both males.The heads of each specimen were scanned for 150 minutes at a vox el size of 5.3 μm and recorded data at every 0.2° rotation for 180° with an Aluminum 1.0 mm filter.The source voltage for the scan was 60 kV and the source current was 165 μA.For each specimen, a total of 1021 transmis sion images were reconstructed into a 1160 × 1160 matrix of 1000 TIFF slices using Nrecon 1.6.10.4 reconstruction software.Volume rendering and further processing was performed using Avizo 8.1.Cranial osteological descrip tions were based on three dimensional (3D) volume ren ders of each specimen retrieved from Avizo.Terminology used in osteological descriptions are adapted from Cun dall and Irish (2008).
The following linear measurements taken (all in mm), along with acronyms for morphological characters used in Tables 4-8 and statistical analyses are listed below: snout to vent length (SVL), measured from the tip of the snout to the vent; tail length (TailL), measured from the vent to the tip of the tail; total length (TotalL), sum of SVL and TailL; relative tail length to total length (TailLR) calculated as tail length to total length ratio; head length (HeadL) from the tip of the snout to the posterior mar gin of the rictus; head width (HeadW) measured at the widest part of the head immediately posterior to the eye; snout length (SnoutL), distance between the tip of the snout and anterior edge of eye; eye diameter (EyeD), maximal horizontal length of the eye; frontal scale length (FrontalL); frontal scale width (FrontalW); interorbital distance (IOD), the shortest distance between the eyes; and internarial distance (IND), distance between the nos trils.Additional morphological characters examined in clude the number of maxillary teeth (MT), counted based on examining the maxillary bone through µCTscans or from examining the bone by pulling back the gum lay er to count the teeth and its sockets; anterior scale rows (ASR), the number of scale rows at one head length be hind the head; midbody scale rows (MSR), the number of scale rows at midbody (halfway between the posterior portion of the head and the vent); posterior scale rows (PSR), the number of dorsal scale rows at one head length prior to the vent; dorsal scale row formula (DSR), a given acronym summarizing the three dorsal scale row counts (i.e., ASRMSRPSR); ventral scales (VEN), the number of elongated belly scales starting at the first dorsal scale row to the vent, excluding the cloacal plate; subcaudal scales (SC), the number of paired subcaudal scales ex cluding the terminal scute; total body scales (TOTAL), Table 1.A list of mitochondrial DNA primers used in this study.

Gene
Primer name Sequence (5'-3') Reference  the sum of ventral and subcaudal scales and the cloacal plate (included as one scale regardless of whether the plate is single or divided); subcaudal ratio (SCR), name ly the ratio between the number of subcaudal scales and the number of total body scales given as a decimal val ue; cloacal plate (CP), the number of terminal ventral scales immediately anterior to vent (given as single for one scale, and divided for two scales); condition of nasal scale (NASAL), given as either vertically divided, entire, or partially divided; condition of loreal scale (LOREAL), given as present or absent; supralabials (SL), the number of scales on upper lip; number of supralabials in contact with the eye (SLEye); infralabials (IL), the number of scales on lower lip; infralabials contacting each other (ILcontact), the number of pairs of infralabial scales in contact; infralabials contacting the anterior chin shields (ILCS), the number of infralabial scales contacting the anterior chin shields; number of preocular scales (PrO); number of presubocular scales (PrsO); number of postoc ular scales (PtO); number of anterior temporals (Ate), the number of temporal scales in contact with the postocular scales; number of posterior temporals (Pte), the number of temporal scales immediately contacting the anterior temporal scales.Symmetric characters are given in left/ right order.We quantified hemipenis shape into two char acter states: (1) unilobed, a single organ with no lobes at the apex; (2) bilobed, organ containing two lobes at its apex.Hemipenis length was quantified relative to the number of subcaudal scales when insitu and when par tially everted.

12S-16S Oligo12S2LM
For statistical analyses, we made adaptations to a workflow popularized by Chan and Grismer (2021) to determine whether O. arenarius and O. macrurus differ morphologically from one another.All statistical anal yses were conducted in R version 4.1.2(R Core Team 2022).Univariate analyses were ran separately on males, females and on a pooled dataset with both sexes to re duce the effects of sexual dimorphism; however, we only ran multivariate statistics on the pooled datasets because we lacked sufficient sample sizes for males/females.We used Shapiro-Wilks Tests and Levene's Tests to test for normality and heteroscedasticity, respectively.We used Mann-Whitney U Tests (also known as Wilcoxon signedrank tests) using the pairwise.wilcox.test()com mand and Student's T-Tests using the t.test() command from the stats package to compare quantitative differenc es between species.Tests for sexual dimorphism within species were performed using Mann-Whitney U Tests.For all univariate statistics, differences between charac ters were determined to be statistically significant when pvalues were less than 0.05.After univariate analyses, we logtransformed the pooled dataset and conduct ed a Principal Components Analysis (PCA) using the prcomp() command in R on the remaining residuals to determine whether there was visible structure in the mor phological variation observed between species.We ran Mann-Whitney U Tests on PC1 and PC2 factors using the same procedures described above to determine if the PC centroids of the two taxa were also significantly different.Following this, we ran the nontransformed portion of the dataset under a Discriminant Function Analysis (DFA; also known as a Linear Discriminant Analysis, LDA) us ing the MASS Package in R (Venables and Ripley 2002) to establish if the two species could be separated unam biguously based on the morphological data.The predict() command in MASS was used to calculate the posterior probability of a priori membership for each individual.All data in both the PCA and DFA were scaled to standard deviation prior to analysis to help eliminate the effects of covariance and ensure data was normally distributed.Re sults from multivariate analyses were visualized using the ggplot2 package (Wickham 2016).The following mor phological characters were used for univariate and mul tivariate analyses (abbreviations in parentheses): TailL/ TotalL ratio (TAILR), midbody dorsal scale rows (MSR), ventrals (VEN), subcaudals (SC), total body scales (TO TAL), subcaudal ratio (SCR), supralabials (SL), number of supralabials in contact with the eye (SLE), infralabials (IL), number of infralabials in contact with the anterior chin shields (ILCS), number of postocular scales (PtO), and number of posterior temporal scales (PT).
The uncorrected pdistances for the 16S rRNA gene fragment among and within examined Oligodon species are presented in Table 3. Intraspecific distances varied significantly and ranged from p = 0% in a number of ex

Morphological analyses
Four characters (VEN, SC, TOTAL, SCR) exhibited normal distributions in all datasets.Relative tail length (TAILR) was the only normally distributed character in the male and female datasets.All characters were hetero scedastic, except for VEN, which exhibited a nonhetero scedastic distribution in the male and combined datasets.Univariate analyses for both male, female and combined datasets showed significant differentiation in the number of body scales and relative tail length between O. arenarius and O. macrurus (Table 4).Strong sexual dimor phism in body scales and relative tail length were also noted within each species, and when both species were pooled together into one dataset to increase sampling size ( Table 5).PCA plots revealed that the two species are modestly separated in morphospace, although consider able overlap exists (Fig. 3A).PC1 accounted for 33.4% of the total variance and loaded most heavily for subcaudals and relative tail length (TAILR and SCR).PC2 accounted for 20.2% of the total variance and loaded most heavily for supralabials (SL and SLE) and midbody scale rows (MSR).PC3 accounted for 14.1% of the total variance and loaded most heavily for the number of ventral scales (VEN).More details on the PCA results can be found in Table 6.Contrary to the PCA, the results from the DFA show that neither species exhibit any overlap with one another in the density plots, and were clearly separat ed (Fig. 3B).The combined univariate and multivariate analyses show that O. arenarius from O. macrurus differ in body scalation and tail length, as previously mentioned by Vassilieva (2015).
Besides body scalation, two other morphological traits are considered distinct in O. arenarius and O. macrurus, specifically the hemipenes and the condition of the loreal scale (absent in O. arenarius vs. usually present in O. macrurus).However, Nguyen et al. (2021) reported a specimen of O. arenarius with a loreal scale on one side of the head.Our examination of addition al specimens confirms that the loreal scale condition is variable within O. arenarius, with three of the nine specimens we examined bearing the scale on at least one side of the head.We also everted the hemipenes of freshly preserved O. arenarius and O. macrurus speci mens and found that the organs more closely resemble the structuring described by Nguyen et al. (2021), rath er than Vassilieva (2015) (Fig. 4).The hemipenes in O. macrurus specimens also match those in Smith (1943), who described the retracted organ of this species insitu.We could not find any putative differences in ornamen tation or overall shape between the organs of O. arenarius and O. macrurus.In both species the hemipenes are deeply bilobed, with the point of bifurcation starting around one-third of the organ's length.No spinous ca lyces were found along the lobes or at the base of the hemipenes, nor myoectases.It is likely the eversion pro cess attempted by Vassilieva (2015) failed to fully evert the organ, which resulted in the small unilobed looking hemipenes figured in the original description of O. arenarius (Fig. 4D).
The skulls of O. macrurus and O. arenarius are quite similar in cranial shape, bone structure, and number of teeth (Fig. 7).Nevertheless, some differences can be re vealed.Between the two specimens examined, the neu rocranium in O. macrurus is flat and elongated, whereas it is rounder and stouter in O. arenarius.The nasals in O. macrurus have a more humped profile and are fused an teriorly forming the sharp tip of the snout; in O. arenarius the nasals are less humped and blunter at the tip.The frontals in O. macrurus are well separated from each other and slightly concave laterally in dorsal view, but in O. arenarius are more ossified and less concave.The parietal segment of the skull in O. macrurus is longer, thus making the whole skull more elongated.The postor bitals in O. macrurus are also longer and thinner distally than O. arenarius.The total number of teeth in O. macrurus is higher: 12/12 maxillary teeth (vs.11/10), 15/17 mandibular teeth (vs.14/14), 12/10 pterygoid teeth (vs.10/9).However, our analyses is still based on only two µCTscans (one model per each species).Since intra specific and interspecific variation within and between Oligodon taxa have not been investigated thoroughly, all the differences noted probably represent individual vari ation, and we interpret these differences with caution.We therefore consider the overall external morpholog ical differences between O. arenarius and O. macrurus to be negligible.
Rostral distinctly enlarged and truncate laterally, wider than high and triangular in dorsal view, partially sepa rating internasals; posterior scale suture of rostral with internasals "deepV" shaped, vertex of rostral rising far onto the dorsal surface of the head inline with nostrils as a narrow obtuse angle (~97º); internasals subrectangular, longer than wide, internasal suture shorter than prefron tal suture, anterior border with rostral and nasal concave; prefrontals subpentagonal, longer than wide, wider than internasals; frontal subpentagonal and shield shaped; length of frontal longer than prefrontals and internasals; anterior suture of frontal bordering prefrontals straight but somewhat indented; eyes placed posterior relative to the anterior edge of the frontal; angle formed by the sutures producing the posterior vertex of the frontal nar rowly obtuse (~94º); supraoculars subrectangular, longer than wide, narrower anteriorly than posteriorly; length of frontal longer than supraoculars; parietals subpentagonal, slightly longer than wide, width of each scale wider than length of parietal suture; length of parietal scale slightly longer than length of frontal; parietal suture shorter than length of frontal; anterior parietal angle formed by the sutures between the parietal/frontal and the suture be tween the supraocular/parietal an obtuse angle (~122º) with the lateral ray of the angle pointing posterolateral ly; nasal scale subrectangular, longer than wide and fully divided; loreal scale present (1/1), also subrectangular, slightly longer than wide, around half the size of nasal; supralabials 8/8, with the 4 th and 5 th scales in contact with the orbit; 7 th supralabials largest, 1 st supralabial smallest; preoculars 1/1; presuboculars (1/1), smaller and less wide than preocular; postoculars 2/2, uppermost postocular larger in size on left side, bottommost postocular on right side slightly wider; temporal scale formula 1+2, upper most posterior temporal longer with 6/6 scales surround ing scale; infralabials 9/9, first pair contacting eachother; 4/4 infralabials contacting the first pair of chin shields; 5 th infralabial largest, 2 nd infralabial smallest; mental sub triangular, wider than long; small mental groove present starting below the mental scale where the first pair of infralabials contact, then terminating at the level of the posterior chin shields; anterior pair of chin shields longer than the posterior pair; anterior chin shields slightly wider than posterior chin shields.
Dorsal scale rows 171715, smooth throughout; re duction from 17 to 15 scale rows occurring on 87 th ventral on either side; ventral scales 142, subcaudals 87, cloa cal plate divided (total body scales 230); subcaudal ratio 38.0%.Maxillary teeth not counted (see General descrip tion and variation for details on dentition).The hemi penes were partially everted in an unilobed state, but not fully prepared.The base of the organ is relatively naked with a few flounces present and the sulcus spermaticus partially visible.
In preservative, dorsum light brown with small dark brown or black reticulations formed by dark edges along the dorsal scales; dorsal reticulations concentrated ante riorly and at midbody, whereas the posterior portion of the dorsum is mostly immaculate; vertebrally a series of crossbars, beginning anteriorly as a pair of dark brown spots and continuing as narrow black bars around one or onehalf a scale wide, slowly fading in size by midbody and continuing up to the tail as small subrectangular spots.The head is brown dorsally with a poorly defined graybrown ocular bar edged posteriorly with black, extending across the eyes then meeting at the prefrontals and anteri or portion of the frontal; a grayish brown and blackedged temporal streak present on each side of the head starting from the medial portion of each parietal through the pos terior temporals and supralabials before dissipating at the gular region.The frontal and parietals have small dark vermiculations present on the scales, along with a dark brown irregularlyshaped spot on the posterior vertex of the frontal and on the suture of the parietals.Along the nape there is a large gray-brown Vshaped nuchal chev ron beginning at the posterior end of the parietals around six dorsal scales in length before forking at the nape as a dark brown streak terminating along the flanks of the first two dorsal scale rows.The remainder of the head, includ ing the labial and ventral regions, are beige and mostly immaculate.Ventral surface light brown to beige, immac ulate without any spotting or makings.Dorsal surface of tail light brown, immaculate laterally and vertebrally between two dark brown longitudinal stripes originating from the dorsum and extending posteriorly from the clo aca as a pair of stripes extending across the tail to its tip.Ventral surface of tail beige and immaculate.
General description and variation.All examined spec imens agree with the original description of the lost holo type and the newly designated neotype (ZFMK 88885).A summary of morphological data for all presently known specimens of O. macrurus is presented in Table 7.In all specimens, body elongated and semicylindrical, slight ly robust anteriorly and at midbody; head ovoid, slightly distinct from neck; snout narrowing in dorsal view, de pressed towards the rostral in dorsolateral view; snout tip terminating past lower jaw; eyes moderatelysized with a round pupil; nostrils pointed laterally; mouth flat, curving slightly posteriorly; tail long in males, consistent in di ameter until the posterior half before tapering gradually to a sharp terminal scute; in females, tail still elongate but much shorter than males, tapering gradually starting anteriorly from the cloaca to the terminal scute.
Dorsal scale rows 171715 (rarely 181715 in two specimens, or 171615 in two specimens), smooth throughout; reduction from 17 to 15 scale rows occurring on ventrals 71-98; ventral scales 131-169 (131-152 in males, 139-169 in females); subcaudals 36-94 (60-94 in males, 36-48 in females); total body scales 178-243 (191-243 in males, 178-214 in females); subcaudal ratio Table 7. Summary of examined morphological characters for all known specimens of O. arenarius and O. macrurus, with the third column including all specimens under Oligodon macrurus sensu stricto (herein redescribed including O. arenarius as a synonym).All characters include data for both sexes except TailLR, VEN, SC, TOTAL and SC, which are separated based on male (m) and female (f) material.Abbreviations are listed in the materials and methods.Cells denoted with a "/" symbol were unavailable or not examined.

Character
(2/3) (2/3) 20.0%-38.7%(29.3%-38.7% in males, 20.0%-25.0% in females).Maxillary teeth 9-12, all bladelike, with the posterior two or three greatly enlarged.Vassilieva (2015) provided lower tooth counts ranging from 6-9 for the type series of O. arenarius, however our examination of the maxilla using µCTscanning revealed that the num ber of teeth is likely higher by at least three.These dif ferences can reflect the difficulties of counting dislodged teeth on the maxilla, especially when the gum layer can obscure views of the upper jaw when examined careless ly.Significant sexual dimorphism was observed in the number of ventrals, subcaudals, TailLR and subcaudal ratio (Table 5).In general, the tails of male specimens are much longer and less tapered in female O. macrurus and are some of the longest tails out of any members of the genus Oligodon.The hemipenes are deeply bilobed in both retracted and everted specimens examined, in situ extending from the 25 th to 29 th subcaudal (fide.Smith 1943 and our own data); myoectases absent on retracted organ.When partially everted the hemipenis may appear unilobed with small apical flounces.Fully everted, the or gan is obliquely flounced until the point of bifurcation, which occurs at the first fourth or third of the organ; lobes calyculate and without spines, becoming smoother pos teriorly; lobes terminating as blunt apices; base of organ nude without ornamentation; sulcus spermaticus bifur cate (forked), extending straight from the base and then dividing at the point of bifurcation and continuing along the lobes to the apices.Most O. macrurus specimens exhibit a similar color pattern, but some intraspecific variation is present.The dorsum inlife is ochre brown, light brown or orangebrown, and all specimens have small darkbrown retic ulations across the body formed by dark edges along the dorsal scales; dorsum usually darker vertebrally, oc casionally with two darker lines creating a small weak lydistinguished and dusky pair of longitudinal stripes, especially distinct on the tail.The head is marked with a dark ocular bar that extends across the eyes and meet ing at the prefrontals and anterior portion of the fron tal, a temporal streak on each side of the head starting from the parietals to the supralabials and flanks, and a Vshaped nuchal chevron.The ventral surface is plain white or beige and without any spots or blotches.Three main color variants can be observed, well correlated with geographic location.The first is a "north" form observed in specimens from Khanh Hoa Province (NHMUK 1938.8.7.39, 1969.1855-56, ZMMU Re13857).In these individuals, the ocular bar is indistinguishable or barely distinguishable from the rest of the head, the temporal streak is faint (and reduced to the lower half of its length in NHMUK 1938.8.7.39 and NHMUK 1969.1855), and the narrow Vshaped nuchal collar is only 2-4 dorsal scales in length and of equal width.There are no dark markings on the frontal scale, and the dorsum has small dark brown crossbars starting posterior to the nuchal col lar fading by midbody to small indistinguished reticula tions.The second "middle" form (Figs 5, 6A) is found in specimens from Binh Thuan Province (ZFMK 88885 [neotype], UNS0500105002, the lost holotype and ZMMU Re14502), where the ocular bar is dark brown and fairly distinct (faint in ZMMU Re14502, but this is probably due to preservation conditions), the temporal stripe is also distinct (divided on two parts in ZMMU Re14502), and Vshaped nuchal collar is more elongated (4-6 scale rows long) and widest medially before nar rowing laterally, including the posterior portion of the parietals.The frontal spot is present and usually distinct (with exception of the lost holotype), and the remainder of the dorsum has several paravertebral pairs of narrow spotshaped crossbars on the anterior half of the body.The "south" form (Figs 6B-D) occurs in specimens from Ba Ria-Vung Tau (ZMMU Re11561, Re14503-04 and VNMN 04724) and Ho Chi Minh City (NHMUK 1969(NHMUK .1854)),where the ocular bar is distinct and dark brown (faint in NHMUK 1969.1854, but again probably due to conditions of preservation), the temporal stripe is also usually distinct (faint in NHMUK 1969(faint in NHMUK .1854)and the Vshaped nuchal collar is 4-5 scales long but wide both medially and on the flanks and fused with mark ings on the frontal.The dorsum has paravertebral pairs of darkbrown spots present mostly anteriorly (ZMMU Re11561, R14503, NHMUK 1969.1854) or throughout the whole body (ZMMU Re14504 and VNMN 04724).Oblique lateral streaking on the dorsum are also dis played on a few specimens (with exception of NHMUK 1969.1854).The specimen ZMMU Re15136 (from Ba Ria-Vung Tau) stands out for its saturated orange color in life and black speckling evenly distributed along the body.However, it seems to share more in common with the "south" form.The presence of more or less distinct vertebral stripe, especially on the tail, seems to be the common feature for all three groups with exception of ZMMU Re15136 and, apparently, the lost holotype.It should be noted that the color present in the "south" form includes specimens previously referred to O. arenarius, as well as specimens identified as O. macrurus.

Osteological description. The description of the skull of
Oligodon macrurus is based on 3D reconstructed µCT scans of two specimens: ZMMU Re13857 and ZMMU Re14502 (one of the paratypes of O. arenarius); (Fig. 7).The skull of O. macrurus is short, rounded, and well os sified.The snout is composed of the premaxilla, nasals, septomaxillae and vomers.The circumorbital bones in clude the prefrontals and postorbitals.Premaxilla single, flat and curved anteriorly, representing the front tip of the snout; deeply wedged in the space between and beneath the septomaxillae and the nasals.Nasals spatulateshaped and sharpened at the tip with an S-shaped profile; left and right articulated nasals form median septum between the nasal cavities covering them dorsally; ventrally nasals form a posterior process lying between the anterior edge of the frontals.Septomaxillae paired and platelike, in contact medially, forming the floor of nasal cavity, par tially fused with vomers; conchal processes of septomax illae well defined; each septomaxilla in contact with nasal septum medially and form a posterior process contacting anterior edge of frontals through the prokinetic joint.Vomers toothless, positioned beneath and behind the two septomaxillae and forming a pair of spherical fenestrae, in which lies the vomeronasal organs, opened by paired orifices into the buccal cavity.Prefrontals on either side of the head, block shaped and obliquely positioned form ing the anterior edge of the orbit; dorsally prefrontals in broad contact with the anterolateral surface of the fron tals, ventrally in loose contact with the maxillae.Left and right postorbitals articulate with the anterolateral surfaces of the parietal and form the dorsoposterior boundary of each orbit.The braincase is composed of compactly ossified bones consisting of the frontals, parietal, basisphenoid, basioc cipital, prootics, supraoccipital and exoccipital; partially fused to each other forming the complete enclosure of the brain.Frontals well separate.Parietals elliptical shaped and the largest cranial element, fused together to form a single bone that dorsally roofs the braincase, bearing no elaborated crests; laterally parietal extends far down contacting the basisphenoid and the prootics.Basisphe noid and parasphenoid fused with each other, forming the posterior snout and anterior floor of braincase.Basioccip ital forms the floor of the posterior portion of the brain cavity, and completes the foramen magnum creating a large and raised occipital condyle.Left and right prootics large, subrectangular ventrally and dome contoured dor sally, partly fused with the parietal and forming the ante rior walls of each internal otic capsule; prootics form the anterior half of each fenestra ovalis and the posterolateral wall of the braincase.Supraoccipitals fused together to form a single bone, externally roofing the posterior brain cavity, internally expanding to form the posterior part of each otic capsule.Exoccipital forming the posterolater al wall of the braincase and part of its roof; exoccipital fused with the opisthotics, surrounding the jugular fora men and extending forward to form the posterior border of the fenestra ovalis, the entire oval foramen magnum, and a small ventral portion of the occipital condyle along with most of the basioccipital.Stapes slender, rod like, proximally enlarged and form a footplate fitting into the fenestra ovalis and distally connect to the inner surface of the quadrate at about mid length level.
The palatomaxillary arches consists of the palatine, pterygoid, ectopterygoid and maxilla.Palatines long and narrow, in contact with the prefrontal process of the max illa laterally and pterygoid posteriorly; 7 small sized pal atine teeth.Pterygoids long and slightly bent bones, nar rower anteriorly, flattened posteriorly, and extend from the posterior palatines to the posterior mandible.Each pterygoid bears 10-12/9-10 teeth.Ectopterygoids flat, bifurcate anteriorly, notched posteriorly and connect the maxillae to the pterygoids.Left and right maxillae com paratively straight and posteriorly broadened due to the dorsal ridge, connected to the flattened ventral surface of the ectopterygoid by a mesial process; the maxilla medi ally contacts the ventral surface of the prefrontal.Each maxilla almost has no edentulous region anteriorly, and bears 10 to 12 teeth, with the posterior 2-3 enlarged and bladelike.This number of maxillary teeth agrees well with the known data for O. macrurus and differs from the data represented in the original description of O. arenarius (Vassilieva 2015).
The suspensorium contains the quadrate and supra temporal, connecting to the mandibles by elastic tissue.Supratemporals narrow, flattened, dermal elements, connected to the proximal end of quadrates and the pos terolateral part of braincase by fibrous connective tissue; each is long, straight, slightly bent toward the braincase, and overlay the exoccipital and prootic.Quadrates long, widely flattened, concave dorsally with a fenestra found posterolaterally; proximal end contacting the posterolat eral edge of each supratemporal; distally articulated sur face of each quadrate narrow, extended transversely and directed backward.Mandibles long and connected to each other anteriorly by an elastic ligament; each composed of the compound, angular, splenial and dentary.The com pound is strongly concave dorsally, narrow distally, and massive and laterally flattened proximally; prearticular crest higher than subarticular crest.Angular and splenial both triangular shaped elements that fuse at their broad est point of contact.Dentaries somewhat dorsally curved, bearing sockets for closely set 14-15/14-17 small teeth that decrease in size posteriorly.
Distribution and natural history.To date O. macrurus is reliably known from five provinces in southern Viet nam (Khanh Hoa, Ninh Thuan, Binh Thuan, Ba Ria-Vung Tau and possibly the vicinity of Ho Chi Minh City), where it is only found in coastal ecosystems associated with the Mui Ne dunefields and sandy coastlines to its north and south (Fig. 1).The locality of Ho Chi Minh City (based on specimen NHMUK 1969.1854) has no precise information associated with it, and requires ad ditional confirmation; although, suitable habitat may ex ist within the region.Vassilieva (2015) described aspects of behavior and habitat of O. arenarius.One specimen was found feeding on a frog Microhyla pulchra Hallow ell, 1861 (Vassilieva 2015).During the collection of new material, we found specimens crossing roads at night and on the crawl in habitat during daytime searches, includ ing a few that were recovered DOR.Based on collection records with dates, this species appears to be surface active yearround, although there is a slight increase in records between the months of November to March, cor responding with the end of southern Vietnam's monsoon season.Habitats where O. macrurus were found include low-sloped littoral dunefields with short vegetation and ecotones with sandy clearings along lowland dipterocarp forests adjacent to dune habitats.
Conservation status.This species is now known specif ically from seven localities across the coast of southern Vietnam.A few of these sites are found within nature re serves and other preserved tracts of land, however human development and increased tourism around these areas could pose a significant threat to this species.Geissler et al. ( 2011) collected two specimens as roadkill indicating that road mortality could be a potential hazard in some locations.Additionally, Vassilieva (2015) noted that the type locality of the synonym O. arenarius was highly disturbed, with surrounding areas where specimens were collected consisting of hotels and residential properties.The protection and management of dunefield habitats across the known localities of O. macrurus should be of conservation priority.Based on the assessment criteria adopted by the International Union for Conservation of Nature (IUCN), we suggest that O. macrurus should be listed as "Vulnerable" on the IUCN Red List, due to the risks associated with habitat destruction and disturbance.Additional research understanding the population dy namics and ecology of this species would greatly improve conservation efforts.

Discussion
Using an integrative taxonomic approach, we found that the recently described kukri snake O. arenarius represents a junior synonym of the species O. macrurus.Our study also provides new data on the distribution, taxonomy and phylogenetic position of O. macrurus and confirms its confinement to the coastal dunefields and sandy regions of southern Vietnam.The weak genetic divergence (less than 1.0% pairwise distance) and nonmonophyly of O. arenarius certainly justifies its demotion from species rank.However, the significant differences in body sca lation and relative tail length may lead some readers to question whether O. arenarius should be considered a subspecies of O. macrurus instead of a junior synonym.In herpetology, the contemporary application, criteria and usage of subspecies is controversial (Hawlitschek et al. 2012;Torstrom et al. 2014;Kindler and Fritz 2018;De Queiroz 2020;Hillis 2020;Burbrink et al. 2022) and a discussion of its utility is beyond the scope of this pa per.Recent designations of subspecies generally follow the suggestions of Mayr and Ashlock (1991) and Braby et al. (2012), who define subspecies as a distinct metapop ulations that are phenotypically diagnosible and are on a trajectory towards evolutionary independence.Following these standards, we offer the following justifications for treating O. arenarius as a synonym rather than a subspe cies: (1) Our phylogeny demonstrates that O. arenarius is paraphyletic with respect to O. macrurus, whereas other snake taxa recently described as subspecies with weak genetic divergence still display clear monophyly based on traditional genetic markers (Hawlitschek et al. 2012;Göçmen et al. 2017;Hofmann et al. 2018;Kindler and Fritz 2018).(2) When subspecies are nonmonophyletic, as is the case for North American night snakes (Hypsiglena) and the Southeast Asian pit viper subspecies of Trimeresurus (Popeia) sabahi Regenass & Kramer, 1981, they usually display allopatric or parapatric distributions with different habitat preferences (Mulcahy 2008;Mulcahy et al. 2017) and are still exhibit morphological diagnosibili ty (Vogel et al. 2004).Oligodon arenarius so far is known from one population at the southern end of the range of O. macrurus and occupies a very similar habitat and eco logical role.Furthermore, there exists one record of O. macrurus directly south of O. arenarius in the vicinity of Ho Chi Minh City, albeit with imprecise locality infor mation.Pending its rediscovery in this province, it would eliminate any sense of allopatry between the two taxa.(3) All other morphological features examined between O. arenarius and O. macrurus, besides body scalation and tail length, show broad overlap (color pattern, dentition, and hemipenial morphology).Regardless of whether O. arenarius is treated as a synonym or subspecies, the over all diagnosis of O. macrurus would not be affected, as its combination of morphological characteristics readily sep arate it from all other members of the genus, especially its uniquely elongate tail found in males (Table 8).
Our study attempted to examine all material available for O. macrurus, but still has limitations related to sample size and geographic scope.In total, we examined 24 spec imens of O. macrurus and O. arenarius, yet it is still dif ficult to conduct analyses related to geographic variation with such a sample size, which could reveal clinal patterns in the number of body scales and body size (sensu Ash ton 2001; Lee et al. 2016).Both traits are known to vary within snakes, and are often influenced by environmental pressures, directional selection, and phenotypic plasticity (Kelley et al. 1997;Krause et al. 2003;Fornasiero et al. 2007;MartínezFreiría et al. 2009).On one hand, these factors could explain why the O. arenarius population has such a divergent number of body scales (178-195 total scales) compared to other O. macrurus populations (203-243 total scales).On the other hand, the addition of more specimens may cause the differences in body scala tion to disappear.Besides morphology, results recovered in our molecular analyses could be influenced by mito chondrial introgression, which is a plausible explaination for the nonmonophyly of O. arenarius and O. macrurus in the mtDNA phylogeny.Finer scale population genom ics using rapidlyevolving markers such as nuclear SNPs or microsatellites, along with denser sampling between the southernmost localities of O. macrurus, might present different patterns of genetic variation in this species.If there is evidence that O. arenarius is headed on a trajecto ry towards evolutionary independence, then designating the taxon as a subspecies could be considered.However, we are currently convinced that it is most appropriate to consider O. arenarius a synonym of O. macrurus, given the evidence at hand.
As mentioned previously, members of Oligodon have traditionally been partitioned into several informal group ings based on hemipenial morphology, scalation, and den tition (Wall 1923;Smith 1943;Leviton, 1962;David et al. 2008aDavid et al. , 2008b;;Green et al. 2010;Neang et al. 2012;Vassilieva et al. 2013;Vassilieva 2015).The role of the hemipenis in delimiting clades within Oligodon was par tially confirmed by the phylogenetic analyses of Green et al. ( 2010), although the sampling in their study was quite limited.Among the species with available data on Table 8.Morphological comparisons between Oligodon macrurus sensu stricto and the other species of Oligodon native to Vietnam.Characters highlighted in bold are considered diagnostic.Abbrevia tions for characters can be found in the materials and methods.Data for other species are based on the following literature sources: Pope (1935), Smith (1943), Wagner (1975Wagner ( , 1976)), David et al. (2008aDavid et al. ( , 2008bDavid et al. ( , 2011David et al. ( , 2012)), Orlov et al. (2010), Vassilieva et al. (2013), Pham et al. (2014), Vassilieva (2015), Nguyen et al. (2016Nguyen et al. ( , 2017Nguyen et al. ( , 2022)), Nguyen et al. (2020), Lalbiakzuala and Lalremsanga (2020), Lee (2022), Yushchenko and Lee et al. (2023).cyclurus-taeniatus species grouphave bilobed hemipenes without spinous calyces or myoectaces (papillae).Smith (1943), describing the hemipenial morphology of male specimen NHMUK 1938.8.7.39, showed that O. macrurus has a deeply bilobed hemipenis without spines or myo ectases, and suggested this species should be part of his informal O. cyclurus species group (interchangeable with the O. cyclurus-taeniatus species group).Our observa tions of the same specimen confirm the presence of deep ly bilobed hemipenes.In some later works (Pauwels et al. 2017;Sumontha et al. 2017), O. macrurus was considered as part of the O. cinereus group, however such an alloca tion was not confirmed by any data and was an accidental mistake (P.David, pers comm).Our phylogenetic analysis places O. macrurus in one clade with the members of the "taeniatus-cyclurus-complex" (sensu Green et al. 2010;Nguyen et al. 2020), thereby confirming the earlier hy pothesis of Smith (1943).By this notion, in combination with the discoveries pointed out by Nguyen et al. (2021) on what was O. arenarius, we are confident allocating O. macrurus into the O. cyclurus-taeniatus species group.So far, all recent reports of O. macrurus have been from dunefields and adjacent sandy habitats surround ing the southern coast of Vietnam.Based on geological evidence, most of the dunefields in southern Vietnam were formed during the last interglacial period of the late Pleistocene, with deposits continuing throughout the Ho locene as a result of climate change and alterating mon soon conditions in the South China Sea (MurrayWallace et al. 2002;QuangMing et al. 2010;Tamura et al. 2020).These findings substantiate Geissler et al. (2011), who suggested that the lack of sandadapted reptiles (namely, burrowing skinks and sandswimming lizards) indicated that these coastal regions were comparatively young in age.It is possible that the formation of these dunefields over the course of the Quarternary period may have influ enced the phylogeography and evolutionary history of O. macrurus, however these conjectures remain speculative until more detailed studies can be conducted.A few rep tile species, such as Dibamus deharvengi Ineich, 1999, Dixonius aaronbaueri Ngo & Ziegler, 2009, Dixonius vietnamensis Das, 2004, and Leiolepis ngovantrii Gris mer & Grismer, 2010 also appear endemic to the sandy coastal habitats of southern Vietnam like O. macrurus.The continued conservation of these animals could be jeopardized in the future by increased tourist and devel opment activities.We recommend the pursuit of addition al studies and continued herpetological surveys in these regions to understand and conserve these animals.

Figure 2 .
Figure 2. Bayesian inference tree of Oligodon derived from the analysis of 3,019 bp of 12S rRNA-16S rRNA and cyt b mitochon drial DNA gene sequences.For voucher specimen information and GenBank accession numbers see Table 1.Numbers at tree nodes correspond to BI PP/ML BS support values, respectively; an endash denotes no support.Outgroup taxa are not shown.Colors of clades and locality numbers correspond to those in Fig. 1.Photographs taken by Nikolay A. Poyarkov Jr.
phylogenetic analyses demonstrate that the genetic differentiation between O. arenarius and O. macrurus is much lower (p = 0.5%) than the pairwise distances ob served in other species within Oligodon.These results are more complex when our morphological data are incor porated (see above), which show significant differences in body scalation and relative tail length between O. arenarius and O. macrurus, but weak differences in other characters (i.e., color pattern, head scalation, hemipenial morphology and cranial osteology).Based on our species concept, delimitation criteria (see Materials and methods) and combined evidence, it is clear that the specific sta tus O. arenarius should be downgraded.However, two taxonomic acts could be considered.Either O. arenarius could be synonymized with O. macrurus as a subjective junior synonym, or O. arenarius could be relegated as a subspecies of O. macrurus.The first decision would eliminate the paraphyly present between the two taxa and consider a single morphologically variable species, while the second would conserve the recognition of a potential ly morphologically divergent population but maintain the paraphyly recovered in the phylogeny.We elect to choose the first decision and consider O. arenarius a morpho logically divergent population within the variable O. macrurus.Synonymizing the two taxa would not affect the diagnosis of O. macrurus amongst other congeners, and stabilizes the paraphyly demonstrated by our phylogenet ic results.Additional justifications for our decision are provided in the discussion section.Redescription of Oligodon macrurusThe addition of O. arenarius into the synonymy of O. macrurus, along with the paucity of known specimens, prompts us to provide a formal redescription of this spe cies.The type specimen of Simotes violaceus macrurus, collected from "Annam: Pointe Lagan (sur les dunes)" (now Cape Lagan or Mui La Gan, northeast Binh Thu an Province, Vietnam), was once part of the herpetolog ical collection of the Muséum National d'Histoire Na turelle (MNHN) in Paris, but was lost between 1943 to 2009(Nguyen et al. 2009; Geissler et al. 2011;Wallach et al. 2014;Uetz et al. 2019).Owing to the fact that the namebearing type no longer exists and there is no other type material by the original author, we elect to designate a neotype for this species in compliance with the Interna tional Code of Zoological Nomenclature (ICZN; Article 72.2 and Article 75)(Anonymous 1999).We choose spec imen ZFMK 88885 as the neotype for Oligodon macrurus(Angel, 1927) due to its close proximity to the original type locality and overall similarity with the lost holotype.As a result of this action, the type locality of O. macrurus is emended to the catalogued locality of this specimen in accordance with ICZN Article 76.3.The new type local ity for O. macrurus is now the "coastal dune area about 1 km from the coast line, Mui Ne, Binh Thuan Province, Vietnam (near 10°57'59.6″N,108°19'46.5″E)",based on ZFMK 88885 collected DOR by Peter Geissler on 27 April 2009.The neotype is on permanent loan to the her petological collection of the Zoologisches Forschungsmu seum Alexander Koenig (ZFMK) in Bonn, Germany.

Figure 3 .
Figure 3. Multivariate analyses of morphology between Oligodon arenarius (light blue) and O. macrurus (dark blue) including: A plot of Principal Components Analysis (PCA) comprised of data from both sexes; B density plot of Discriminant Function Analysis (DFA) comprised of data from both sexes.

Figure 4 .
Figure 4. Hemipenes of Oligodon arenarius.A sulcal and B asulcal side of the fully everted hemipenis of ZMMU Re16804; C partially everted hemipenis of the same specimen demon strating the overall similar shape to the partially everted organ of the holotype figured by Vassilieva (2015); D partially everted hemipenis of the holotype ZMMU Re14503 from the original description (Vassilieva 2015), copyright Magnolia Press, repro duced with permission from the copyright holder.Photographs A-C taken by Nikolay A. Poyarkov Jr.

Figure 5 .
Figure 5. Photographs of the preserved neotype of Oligodon macrurus (ZFMK 88885), an adult male specimen.A dorsal and B ventral views of the whole specimen, and C dorsal D ventral, E right lateral and F left lateral views of the head.Scale bars for A-B represent 10.0 mm, and scale bars for C-F represent 5.0 mm.All photographs taken by Morris Flecks, used with permission.

Figure 6 .
Figure 6.Living specimens of Oligodon macrurus sensu stricto from various locations in southern Vietnam.A adult male ZMMU Re-16807 from Mui Ne, Binh Thuan Province, representing the 'middle' color phase; B adult female VNMN 04724 (formerly ZMMU NAP03884) from Binh ChauPhuoc Buu Nature Reserve; C and D two adult males ZMMU Re16804 and ZMMU Re 16805 from Binh Chau-Phuoc Buu Nature Reserve, Ba Ria-Vung Tau Province, B-D formerly identified as Oligodon arenarius.Photographs taken by Hieu Minh Pham A and Nikolay A. Poyarkov Jr. B-D.
only select members of the O.

Table 2 .
List of sequences and corresponding voucher specimens of Oligodon and outgroup taxa used in this study.Note that the numbers (column one) included in this table do not match the numbers used in other tables or figures in the text.Acronyms not described in the materials and methods include the following: CHS, unknown field tag series; RAP, field tags of R. Alexander Pyron; RS, field tags of Ruchira Somaweera; NP, national park; NR, nature reserve; WS, wildlife sanctuary.

Table 3 .
Genetic differentiation of Oligodon.Uncorrected p-distances (given as percentages) between sequences of 12S-16S rRNA (below diagonal) and intraspecific genetic p-distances of Oligodon species (on diagonal) included in phylogenetic analyses.

Table 4 .
Resulting Pvalues from univariate morphological analyses of sexual dimorphism in Oligodon arenarius and Oligodon macrurus specimens.Characters in bold are considered statistically significant.Cells denoted with a "-" symbol were not examined.All characters were examined using Mann-Whitney U Tests.Abbreviations include "n" standing for number of specimens; "(m)" standing for males, and "(f)" standing for females.

Table 5 .
Resulting Pvalues from univariate morphological analyses comparing Oligodon arenarius and Oligodon macrurus speci mens.Characters in bold are considered statistically significant.Cells denoted with a "/" symbol were not examined.All characters were examined using Mann-Whitney U Tests, except for those denoted with an asterisk, which were performed using Student's TTests.Abbreviations include "n" standing for number of specimens; "(ar)" standing for O. arenarius, and "(ma)" standing for O. macrurus.

Table 6 .
Summary of PC scores and statistics for the Principal Components Analysis (PCA) consisting of Oligodon arenarius and Oligodon macrurus specimens.Abbreviations are listed in the materials and methods.Cells denoted with a "-" symbol were un available or not examined.