Discovery of a deeply divergent new lineage of vine snake (Colubridae: Ahaetuliinae: Proahaetulla gen. nov.) from the southern Western Ghats of Peninsular India with a revised key for Ahaetuliinae

The Western Ghats are well known as a biodiversity hotspot, but the full extent of its snake diversity is yet to be uncovered. Here, we describe a new genus and species of vine snake Proahaetulla antiqua gen. et sp. nov., from the Agasthyamalai hills in the southern Western Ghats. It was found to be a member of the Ahaetuliinae clade, which currently comprises the arboreal snake genera Ahaetulla, Dryophiops, Dendrelaphis and Chrysopelea, distributed in South and Southeast Asia. Proahaetulla shows a sister relationship with all currently known taxa belonging to the genus Ahaetulla, and shares ancestry with Dryophiops. In addition to its phylogenetic position and significant genetic divergence, this new taxon is also different in morphology from members of Ahaetuliinae in a combination of characters, having 12–13 partially serrated keels on the dorsal scale rows, 20 maxillary teeth and 3 postocular scales. Divergence dating reveals that the new genus is ancient, dating back to the Mid-Oligocene, and is one of the oldest persisting monotypic lineages of snakes in the Western Ghats. This discovery adds to the growing list of ancient lineages endemic to the Agasthyamalai hills and underscores the biogeographic significance of this isolated massif in the southern Western Ghats.


Introduction
The Western Ghats (WG) of Peninsular India is a global biodiversity hotspot with a high diversity of snakes [1][2]. While the mountain range has been explored since the colonial period, the systematics of snake fauna remains poorly known. There have been intensive explorations of WG biodiversity in the last decade revealing many new genera and species of trees [3][4][5][6], a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 the standard deviation of the split of frequencies 0.01. We diagnosed the state of convergence of chains and ensured that the effective sample size (ESS) was above 200 for each parameter in Tracer v1.6 [43]. Additionally, log likelihood scores were plotted against other parameter values from each run against generation time. A final consensus tree and clade posterior probabilities were summarized after 25% of burn-in of the total number of trees sampled. The resulting ML and BI trees topologies were visualized in FigTree v1.4.2 [44].
Divergence dating. The divergence time was estimated with the DNA dataset used for Bayesian inference and maximum likelihood analysis and based on four fossil and biogeographic reference calibration points in BEAST v1.8.2 [45,46]. We used the Birth-Death process for the tree prior, which is a continuous-time Markov process of speciation and extinction through time from an initial lineage to the birth of a new lineage [47]. We used an uncorrelated relaxed clock model with lognormal distribution [48] and GTRGAMMAI model for all the partitions in BEAST [40,49]. The substitution models and tree topology were linked and clock model was unlinked among partitions (Table 1).
Three external fossil records [50][51][52][53][54] and a secondary calibration [49] were placed on the nodes after carefully considering the calibrations as suggested in an earlier study [49] (S3 Table). This Bayesian analysis was carried out twice using MCMC algorithm and 50 million generations sampled every 10000 generations and implemented in the online-based server CIPRES Science Gateway [42]. The convergence and effective estimated sample size (ESS) of the posterior probability distribution were estimated for all parameters in Tracer v1.6.2 [43]. 25% of the trees were discarded and the final divergence estimate tree was derived using a maximum clade credibility tree and node height kept as median in TreeAnnotator 1.8.2 [45,46]. Trees were visualized in FigTree 1.4.2 [44].

Morphological data
Snout to vent length, tail length and total body length were measured by marking with a string and using a measuring tape. Ventral scales were counted according to Dowling's scheme [55]. Subcaudals were counted on one side excluding the terminal scale. Measurements were taken using Mitutoyo dial calipers. Meristic characters that were collected from the specimens included: pre-ventrals, ventrals (V), subcaudals (SC), supra-labials-right (SLr), supra-labialsleft (SLl), largest supra-labial, supra-labials in contact with the eye-right, (SL2r), supra-labials in contact with the eye-left (SL2l), loreal-right (Lr), loreal-left (Ll), nasal-right, nasal-left, pre-subocular-right (PRSOr), pre-subocular-left (PRSOl), pre-ocular-right, pre-ocular-left, infra-labials-right (ILr), infra-labials-left (ILl), post-ocular-right (POr), post-ocular-left (POl), gulars-right, gulars-left, scales around the body (N-after neck, M-at midbody and Tbefore vent), temporals-right, temporals-left, sub-ocular-right, sub-ocular-left, pre-frontal  and pre-ocular contact, cloacal plate, supra-labial scale division if any and the nature of ventrals (angulated, keeled, notched, no ventral keels). The mensural characters that were collected from the specimens are as follows: snout to vent length, tail length, total length, head length, nostril to eye length, vertical eye diameter, horizontal eye diameter, eye to snout length, pre-frontal length, frontal length, snout to vent length divided by the tail length, snout to vent length divided by head length and the relative tail length. Coloration characteristics that were considered are as follows: dorsal coloration, inter-scalar skin colour, cross bars across the body, iris colouration, ventral colour and ventral stripes. Hemipenis was everted and examined in situ. The upper jaws of the specimens were carefully dissected for counting the maxillary teeth. Empty sockets when present were counted as proxies for teeth and not as diastema.
Morphological analyses. A principle component analysis (PCA) was performed separately using all the characters to assess the major loading components. These contributing morphological characters were plotted using a multivariate analysis (V, SC, SL, SL2, L, PRSO, IL, PO, M and keels) that exhibited variability. We tested for differences in morphological space by comparing taxa representing all genera of the subfamily Ahaetuliinae with our primary and secondary datasets (S2 Table). The statistical analysis was carried out using PAST v3.14 for Macintosh [56].

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix "http://zoobank.org/". The LSID for this publication is: urn:lsid:zoobank.org:pub: 777BD774-53BE-432D-AF2F-3948B8E1E8D4. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS.

Phylogenetic relationship of Ahaetuliinae
The tree topology recovered from a Bayesian analysis for Ahaetuliinae (comprising currently known genera) was similar to the result obtained by an earlier study [23]. We found that Proahaetulla antiqua gen. et sp. nov. is nested within the Ahaetuliinae clade and shares ancestry (PP 1.0 & ML bootstrap >70%) with the rest of the members of genus Ahaetulla (Fig 1A). The clade comprising Ahaetulla and Proahaetulla antiqua gen. et sp. nov. was recovered as sister to the Dryophiops clade but with low bootstrap support (< 70%); however, the relationship between Ahaetulla and Dryophiops is strongly supported in a previous study [23] (Fig 2).
There is 14.1-17.4% (on Cytb), 16.0-18.1% (on ND4) and 5.2-6.5% (on 16S) genetic divergence between the new lineage Proahaetulla gen nov. and Ahaetulla ( Table 2). This distance is comparable with the genetic divergence value between two other related sympatric genera Chrysopelea and Dendrelaphis (13.7-22.0% on Cytb, 15.3-20.1% on ND4 and 4.1-8.2% on 16S). Moreover, the genetic distance between Proahaetulla gen nov. and Ahaetulla (on 16S) is in the same range as the distance between Dryophiops and Ahaetulla, which are also distributed sympatrically. Given the genetic divergence between these established genera in the clade Ahaetuliinae, there is a strong case to treat Proahaetulla as a genus based on genetic distance alone. We also carried out a PCA including taxa representing all genera of the subfamily Ahaetuliinae ( Fig 3B and S4 Table). Here, PC1 and PC2 explained 70.1% and 27.3% of the variance; the number of sub-caudal scales had a higher loading (0.81) on PC1 while ventral scales had a higher loading (0.81) on PC2. There is overlap in the morphological space occupied by The other genera (Dendrelaphis from India, Chrysopelea, Dryophiops and Proahaetulla) occupy a relatively small morphological space compared to Ahaetulla. However, Dendrelaphis has 45 species across its range, and it is expected that the morphological space occupied by this genus would be much larger if all the taxa were included in the analysis.

Time of divergence
Chrysopelea occupies a non-overlapping position with respect to the other genera due to the variation in a few morphological characters (V 201-236, C 106-138 and M 17), but this does not represent the full range of variation in the genus. The morphological space of Ahaetulla overlaps with Dryophiops, Dendrelaphis and Proahaeatulla gen. nov. due to the presence Etymology: The generic epithet Proahaetulla gen. nov. stems from the generic nomen Ahaetulla indicating the early divergence of the lineage from the rest of the Ahaetuliinae members. Gender feminine. The specific epithet antiqua is Latin for 'antique' or old, a term alluding to the evolutionary age or antiquity of this new taxon.

Diagnosis
Lineage diagnosis.      oculars absent; temporals 1+3 (anterior + posterior) on the right and 2+3 (anterior + posterior) on the left; prefrontal scale in contact with the preoculars; two preventrals ( Fig 2B-2D). Colour in life: Body uniform bright green ( Fig 2B); rostral, infralabials and mid body on the underside creamish yellow to light green; creamish yellow ventral stripe along the notched ventral keels; slight discoloration in the preocular; when threatened, the inter-scalar skin is revealed, colored with a consecutive series of black and white bars that converge towards the head; an occasional light blue coloration at the proximal end of scales; concentration of black speckles both in the anterior and posterior end of the horizontal pupil and a slight discoloration around the pupil; tongue greenish yellow.
Colour in ethanol: Body uniform yellowish green to bright green; rostral, infralabials and the underside creamish white; creamish yellow to white ventral stripe along the notched ventral keels; slight discoloration in the preocular; the inter-scales are colored with a consecutive series of black and white stripes that slant towards the head; light blue coloration in the angle of the jaw and patches on the body (due to the preservative); yellow clouded with white eyes with black speckles; concentration of black speckles both in the anterior and posterior end of the dilated pupil; tongue yellowish ( Fig 4A-4E).
Maxillary arch dentition: Maxillary bone mildly arched with a curve towards the anterior end of the prediastemal tooth set ( Fig 2D); a total of 20 teeth that rise perpendicular to the maxilla and curve inward posteriorly; prediastemal teeth 13 and postdiastemal teeth 7, observable gradual tooth size increase in the prediastemal tooth set with a smaller tooth suffixing the largest tooth of the prediastemal set. A comparatively small diastema, 2-2.5 tooth sockets wide, suffixed with a set of 5 smaller teeth; a pair of large grooved teeth at the end.
Hemipenal description: Organ fully everted, examined in-situ. Organ short, thick, heavily flounced and ornamented with spines ( Fig 4C). Pedicel slim, barely visible, 8.31 mm long, 6.41 mm wide, extending till 6 th subcaudal scale; lobe very wide and large, head not quite bilobed, flat; lobed head crown with small spines radiating towards sides; lobed body and sides with very large cursive spines, some spines half the size of the pedicel; sulcus spermaticus barely visible in sulcate view, being hidden by the protruding lobe spines.
Variation shown by paratype: Agreeing with the holotype is most respects, and exhibiting the following intraspecific variations: total length 1189 mm; snout to vent length 764 mm and tail length 425 mm; relative tail length 0.36; head length 22.4 mm; horizontal diameter of the eye 4.6 mm; vertical diameter of the eye 3.5 mm; distance from nostril to eye 5.9 mm; distance from snout tip to eye 8 mm; dorsal scale rows 13:13:13 (after neck, at midbody and before vent); ventrals 207; subcaudals 165, divided; temporals 3+3 (anterior + posterior) on the both side; three post-oculars (both left and right); supralabials 7-8, 4 th the largest, 6 th in contact with the eye; infralabials 9, 5 th and 6 th contacting posterior genials.
Comparisons: Proahaetulla gen. nov. differs from the genera in Ahaetuliinae with the combination of the following characters. Dendrelaphis-rostral lacking a protuberance; pupil circular; loreal region not strongly concave; ventral scales mildly notched; dorsum never completely verdant green; Chrysopelea-rostrum lacking a protuberance; pupil circular; loreal region not strongly concave; dorsum never verdant green; ventrals strongly notched; Dryophiops-dorsum never verdant green; dorsal scales smooth; posterior temporals not greater than two; loreals one on each side (Fig 1B-1F) Table).
Relationships: The new lineage is a sister taxon to the genus Ahaetulla, as recovered in our phylogeny. Our tree topology varies slightly in comparison to previous reconstructions, likely due to the use of the different nucleotide substitution models for the analyses (Bayesian analysis with mixed substitution models vs. ML analysis with GTRG model). In addition, we were unable to include other available RAG1 sequences from the GenBank as they could not be aligned with our generated RAG1 sequences. Regardless of the variation in the topology within the Ahaetulla clade, there is very strong support for the node of interest (the relationship between Proahaetulla gen. nov. and Ahaetulla) in both the analyses. The generic status of Proahaetulla gen. nov. is further supported by the date of divergence (26.57 Ma) from its MRCA, indicating that it is the oldest lineage in the group, besides also differing from Ahaetulla and other genera in morphological characters.
Distribution: The new taxon was found only in the far south of the Western Ghats. It occurs in Agasthyamalai hills, where it was recorded in the high elevation wet forests (> 1200 m asl) at Agasthyamalai peak and in Pandimotta (Fig 5). The new taxon's range is likely to encompass other high elevation regions of Agasthyamalai. This taxon also broadly overlaps in its latitudinal distributional range with A. cf. dispar and A. cf. nasuta.
Natural history: The holotype was collected from a tree branch inside a forest patch at an elevation of 1640 msl, on the way to Agasthyar peak. It was found resting coiled on a tree at ca. 1640 h about 2.5 meters above the forest floor. The paratype was sighted at night around 2200 h, sleeping on a shrub at a height of about 2 m from the ground, inside a dense forest patch at an elevation of 1224 msl in Pandimotta ca. 25 km north of the type locality. Members of Ahaetuliinae are mostly arboreal snakes with a few outliers such as Ahaetulla perroteti and A.dispar with environmental adaptations and a body plan suited for a life in open montane grasslands. The body morphology of this genus is similar to arboreal Ahaetulla and Dryophiops. Since both the specimens were found inside thickly wooded forests resting and showing activity on trees and shrubs, we speculate that this taxon is also adapted to a completely arboreal lifestyle.

Discussion
The Colubrid subfamily Ahaetuliinae is a diverse group of predominantly, arboreal snakes, distributed widely in the tropical regions of the Southeast Asia [21][22][23]. The species are currently grouped under four genera: Ahaetulla Link 1807, Chrysopelea Boie 1826, Dendrelaphis Boulenger, 1890 and Dryophiops Boulenger, 1896 [23]. The first three genera are widely distributed in Peninsular India, with high diversity in the Western Ghats region [17,18,25,28]. In this work, we add a new endemic genus and species Proahaetulla antiqua gen. et sp. nov. for the Western Ghats.
Phylogenetically, the new genus was recovered, with strong nodal support, as sister to the genus Ahaetulla. The phylogenetic position and divergence dates reveal new insights into the historical biogeography of snakes of the subfamily Ahaetuliinae. Divergence dates show that the ancestors of Proahaetulla gen. nov. and Ahaetulla diverged during the Oligocene/ Early Miocene boundary and the discovery adds one of the oldest persisting monotypic lineages of snakes to the Western Ghats.
Distribution records suggest that the new lineage is potentially a narrow endemic occurring in the high elevations of the southernmost massif, the Agasthyamalai in the Western Ghats. Its geographic range roughly mirrors the restricted distributional range that many taxa have within the Agasthyamalai region [57,58,59]. For example, a deeply divergent monotypic lineage of agamid lizard Microauris aurantolabium is also known only from the high elevations of the Agasthyamalai [57] and a number of old lineages of frogs are also confined to Agasthyamalai [19]. The southern regions of the Western Ghats have long been recognized as rainforest refugia [60], though the exact extent of the refugia remains uncertain. However, local endemism in massifs such as Agasthyamalai suggests the possibility of multiple micro-refugia across the different massifs in the southern Western Ghats. This study provides further evidence that the Western Ghats served as a refugium in the past, including for the common ancestors of this clade comprising the genera Ahaetulla and Proahaetulla.
Our delimitation of this new lineage is in keeping with the recent consensus on the need for an integrative taxonomy, invoking multiple lines of evidence [60][61][62][63]. The present new lineage is distinct in its phylogenetic position, is deeply divergent (26.55 Ma, in being a sister lineage of all sampled Ahaetulla, and also differs in morphology from other genera in the same subfamily. Traditional methods of morphological analysis using PCA revealed large overlaps between the genera Chrysopelea, Dendrelaphis, Ahaetulla, Dryophiops making it difficult to separate any of these genera with quantitative external morphological characters. Considering the morphological synapomorphies and similarities between Chrysopelea and Dendrelaphis, it is not surprising that the new genus superficially resembles Ahaetulla. With the addition of Proahaetulla, a number of unique patterns of character evolution can be highlighted (Table 3). The dorsal keeled scales that are shared among Chrysopelea, Dendrelaphis, Dryophiops, and Proahaetulla gen. nov. are lost in Ahaetulla's ancestor. Pupil shape shows early evidence of divergence in Ahaetuliinae, with round pupils (Chrysopelea, Dendrelaphis) and horizontal pupils (Ahaetulla, Dryophiops and Proahaetulla gen. nov.). Apical pits are present in Chrysopelea, Dendrelaphis, and Dryophiops, but are lost in the common ancestor of Ahaetulla and Proaheatulla.
In addition, certain morphological characters shared by the clade composed of the three genera-Dryophiops, Ahaetulla and Proahaetulla gen. nov.-reveal interesting patterns of early divergence of characters during the ancestral split between this clade and its sister clade (Chrysopelea, Dendrelaphis) of Ahaetuliinae. The members of the clade, Dryophiops, Ahaetulla and Proahaetulla gen. nov., all possess a laterally compressed, elongated body plan, elongated sharp snout, large eyes with unique horizontal pupils paired with a well developed canthus rostralis for specialized binocular vision, which we hypothesize to have been acquired after the split between the two major clades of Ahaetuliinae.
We also highlight a unique convergence of character pattern divergence between two unrelated clades, Ahaetuliinae in Asia and the African clade comprising Thelotornis, Dispholidus and Thrasops occurring within Colubridae. As in the Asian clade, one of the genera in the African group (Thelotornis) has horizontal pupils (resembling Ahaetulla, Proaheatulla and Dryophiops), while the others (Dispholidus and Thrasops) have round pupils (resembling Dendrelaphis and Chrysopelea). In addition to these, there are several other genera referred to as 'vine snakes' across the world-Thamnodynastes, Oxybelis, Philodryas, Xyelodontophis, Uromacer and Langaha. It is noteworthy, that though unrelated, this group of snakes exhibits a vast array of uncanny morphological, behavioral and ecological convergences. These visual convergences led Boulenger to index these snake groups together in his catalogue of ophidians [64].
An additional finding from this study, the topology of our tree suggests that the genus Ahaetulla originated in Peninsular India. Ahaetulla is currently distributed along the Indian subcontinent and is found along South and Southeast Asia, including South China, Myanmar, Thailand Indonesia, Malaysia and western part of the Philippines island archipelago. The Genus Dryophiops occurs sympatrically with Ahaetulla throughout its range in SE Asia including Indonesia, Malaysia and the western part of the Philippines archipelago. Our study suggests the split of Proahaetulla gen. nov. with the rest of Ahaetulla in Peninsular India thus indicating the geographic origin of the genus Ahaetulla [24].
Proahaetulla gen. nov. is the first deeply divergent colubrid snake genus reported in recent decades from the southern Western Ghats in the Indian peninsula. Despite new discoveries of frogs in recent years [59,82], a new endemic genus of colubroid snake from the Western Ghats is a surprise, as the last such descriptions were at least a century ago, if not more-Dieurostus Berg, 1901; Rhabdops Boulenger, 1893 and Xylophis Beddome, 1878 -making Proahaetulla gen. nov. a once-in-a century find. Many new species of snakes have been described recently from Peninsular India across several genera (in Rhabdops [11]; Uropeltidae [15,[83][84]; Xylophis [14]; Dendrelaphis [17,18]; Lycodon [85]; Boiga [86]; and Calliophis [16]). Several of these such as the new Xylophis, Dendrelaphis, Boiga and Calliophis are taxa which are longknown and previously sampled by researchers, but were misclassified or assigned to other similar genera. However, the present new taxon Proahaetulla antiqua gen. et sp. nov. is a completely new finding, which does not appear to have been previously encountered by the scientific community to the best of our knowledge.