Resolving pitfalls in pit viper systematics – A multi- criteria approach to species delimitation in pit vipers (Reptilia, Viperidae, Craspedocephalus) of Peninsular India reveals cryptic diversity

Asian pit vipers belonging to the genus Craspedocephalus are a complex group of vipers, distributed in South and Southeast Asia. Their taxonomy is unresolved in many lineages across their distributional range. Here, we reassess the taxonomy and systematics of pit vipers of the genus Craspedocephalus in Peninsular India based on extensive field sampling, in particular in the Western Ghats. We build and expand on the previous findings of genetic relatedness between the peninsular Indian lineages with the Sundaic clade (C. puniceus complex) with greater evidence, based on additional taxa sequenced herein. We reconstruct the phylogeny of the group using three mitochondrial genes and delineated lineages using coalescent species delimitation methods. We then used multiple criteria including genetic divergence and separation in morphological and geographic space to designate taxonomic units. Our work revealed the presence of a South Asian radiation of the clade Craspedocephalus, with a few Sundaic members. Our study reveals the systematic relationships of four Peninsular Indian species of Craspedocephalus, including Peltopelor macrolepis and C. strigatus, sequenced here for the first time, that are classified or confirmed as members of Craspedocephalus. Hence, we place the genus Peltopelor in the synonymy of Craspedocephalus. Using our multi-criteria approach, we delimit four new cryptic evolutionary lineages within the Western Ghats escarpment of Peninsular India. These cryptic lineages belong to the C. malabaricus, C. gramineus and C. macrolepis complexes and are geographically and/or ecologically (in terms of habitat association) distinct from their sister lineages across their distributional range, while others are separated in morphological space. Our new phylogenetic tree and delimitation analysis thus reveals the presence of multiple clades with several cryptic lineages separated by geographical barriers or habitat association.


Introduction
Pit vipers are a group of venomous snakes that comprise several genera of vipers with a facial pit; they are terrestrial and arboreal in habit and predominantly found in wet, humid forests and montane habitats as well as in a few dry forest habitats in the Oriental realm. Pit vipers of the genus Craspedocephalus Khul & Hasselt, 1822 are a taxonomically complex group of venomous snakes and comprise nine recognized species distributed in South and Southeast Asia, mainly in forested and mountainous tracts (Wallach et al. 2014). These species were, till recently, largely assigned to the genus Trimeresurus and continue to be referred to as such in many accounts (David et al. 2011). In Peninsular India, the species level taxonomy of this genus has not been refined since the early 20 th century (Rao 1917).
Some of the first members of the group to be named were from the Indian peninsula such as Coluber gramineus (now Craspedocephalus gramineus) by Shaw (1802). Subsequently, many such nominate descriptions were made across Peninsular India (see Smith, 1943). Three more taxa -Trimeresurus strigatus Gray, 1842, Trigonocephalus (Cophias) malabaricus Jerdon, 1854 and Trimeresurus macrolepis Beddome, 1862 -were described. Subjective synonyms and generic reallocations have been provided for these taxa that are currently placed under the genus Craspedocephalus (David et al. 2011, Wallach et al. 2014. Of these, C. gramineus has, in particular, had a long and convoluted taxonomic and nomenclatural history; in the past it was considered as a single, widespread and variable species from Peninsular India through the Indochina subregion to insular Southeast Asia (see Pope & Pope, 1933).
In a pioneering attempt to reconstruct a molecular phylogeny of the snakes of the Trimeresurus radiation, Malhotra & Thorpe (2004) recovered a tree revealing close clustering of southern Indian and Sri Lankan taxa with some Indo-Malayan ones (see their 'Clade E'). They also established molecular relationships of two Peninsular Indian congeners, C. gramineus and C. malabaricus. Malhotra & Thorpe (2004) allocated most southern Indian taxa to the genus Trimeresurus, with one lineage (macrolepis) assigned to the genus Peltopelor Günther, 1864, based mainly on scalation and hemipenial characters that were jointly diagnostic of the mitochondrial clades (Malhotra & Thorpe, 2004). This was further reinforced by osteological characterisation of the skull (Guo et al. 2009(Guo et al. , 2010.
The following taxa recognized as valid (Wallach et al. 2014) are relevant to the present work: C. gramineus (Shaw, 1802) inhabiting the hills of Peninsular India; and C. strigatus, C. malabaricus and P. macrolepis that are endemic to the Western Ghats (Wallach et al. 2014). However, the phylogenetic relationships of P. macrolepis and C. strigatus are unknown due to the absence of genetic data (Malhotra & Thorpe, 2004, Guo et al. 2010). These taxa not only inhabit multiple discrete ecoregions and habitat patches but also show considerable morpho-logical variation among populations (Whitaker & Captain 2008, Kanagavel et al. 2012.
Though previous genetic studies on the Trimeresurus radiation have included Peninsular Indian taxa, a dedicated study with comprehensive geographic sampling and a robust phylogenetic analysis is still lacking. Prompted by the morphological variations and informed by biogeographic processes, this study approaches the systematics of Peninsular Indian species using a multi-criteria approach. Here, we assess and delimit species boundaries in pit vipers of Peninsular India to reassess species diversity within this group and implement formal nomenclatural changes.

Taxon Sampling and Area of Study
Sampling was carried out in Peninsular India, including in the Western Ghats (WG) and the Eastern Ghats (EG), and in Northeast India as part of Indo-Chinese sub-region. Field surveys and collections were carried out in all possible habitat types, across environmental gradients and at different elevations, as well as type locality locations to maximize the sampling of both rare and range-restricted species, and to capture intra-species variability. Snakes were captured alive and tissue and scale samples were collected by non-intrusive methods and stored in 95% ethanol; snakes were released unharmed after sample collection and measurement. We also collected samples from dead specimens encountered on roads; voucher specimens and samples were collected and preserved in 95% alcohol. The specimens were deposited at Bombay Natural History Society (BNHS). The fieldwork was carried out with the prior written permission from the state forest departments of Maharashtra, Kerala, Karnataka, Tamil Nadu, Odisha, Andhra Pradesh, Arunachal Pradesh and Meghalaya.

DNA Extraction and Amplification
Total genomic DNA was extracted from all available tissue samples stored in 95% absolute alcohol. Genomic DNA was extracted using commercially available DNEasy extraction kit (QIAgen) as well as Phenol-Chloroform-Isoamyl alcohol protocol (Sambrook et al. 1989). To elucidate the phylogenetic relationships between members of the clade, sections of mitochondrial genes including 16S rRNA, Cytochrome-b (cyt b), and NADH dehydrogenase subunit 4 (ND4) were amplified with previously published primers (see Supplementary File 1: Table S1). The PCR conditions and preparation protocol were the same as in earlier studies; however, modifica-tions of annealing temperature were applied in a few PCR reactions. PCR amplified products were purified using QIAquick PCR purification kit (Qiagen). The sequencing of purified products was carried out commercially from Amnion Biosciences Pvt. Ltd., Eurofins Biotech Pvt. Ltd. and Centre for Cellular and Molecular Platforms (C-CAMP, NCBS), Bangalore, India.

Sequence Alignment
Sequences were edited and visually corrected using MEGA v5.2 (Tamura et al. 2011). The individual consensus sequences were derived from forward and reverse complements after checking for base miscalls. The sequence alignment was accomplished in MUSCLE (Edgar 2004) algorithm implemented in MEGA v5.2. The protein-coding genes cyt b and ND4 were checked for presence of indels and noncoding sequences to detect possible pseudogene amplification and premature stop codons by translating DNA to protein sequence. The 16S rRNA sequences were aligned against GenBank data submitted by earlier studies (Malhotra & Thorpe 2004) and visually edited; ambiguously aligned regions and gaps were cropped out of the analysis due to the presence of secondary structures in the sequence (Kjer 1995, Parkinson 1999, Gutell 2004. A matrix of 2154 bp was created by combining all three mitochondrial gene datasets. The gaps present in the dataset were treated as missing data. Available sequences for other species of the Trimeresurus radiation were downloaded from GenBank (See Supplementary File 2: Table S2).

Phylogenetic Analysis
We reconstructed the phylogeny of pit vipers to understand the relationships of pit vipers of Peninsular India including endemic species (C. strigatus and C. macrolepis) from the Western Ghats, which had not been included in earlier work (Malhotra & Thorpe 2004). The 2154 bp matrix was analysed with codon partitions where different codon positions of cyt b and ND4 genes were considered as different partitions using PARTITIONFINDER v1.1.1 (Lanfear et al. 2012). A Maximum Likelihood (ML) tree was reconstructed with non-parametric bootstrapping in RAXMLGUI v1.3 (Silvestro et al. 2012, Stamatakis 2006. The RAXML GUI platform implements only GTR substitution model. Therefore, the analysis was carried out with 1000 bootstrap replicates and GTR-GAMMA model was applied to every partition in this analysis. To root the tree, we used Protobothrops mucrosquamatus, Protobothrops jerdonii and Protobothrops kaulbacki as outgroups. The resulting tree was visualized in FIGTREE v1.4.3 (http://tree.bio.ed.ac.uk/software/figtree).

Morphological Measurements and Pholidosis
Morphological data were collected from all sampled specimens. Pholidosis analysis was carried out in the field as well as in the laboratory on preserved specimens (See Supplementary File 3: Table S3). Morphometric measure-ments were carried out with Mitutoyo Vernier calipers (to the nearest to 0.1 mm). We also examined specimens that were morphologically different in scalation or colour as potential ecomorphs. Sampling was carried out in every possible habitat type and along the environmental gradient to scrutinize the possible morphological space shared by individuals with respect to distance and isolation. We recorded the body colouration of individuals belonging to each population and examined both overall colour from head to tail as well as change in colour, if any occured. Colour variation of ventral scales along with any blotches and speckles present on the ventral shields were also recorded.

Geographical Distribution
We recorded geographic locations for each individual sample collected with the help of a GPS (WGS84) to identify isolation and geographic barriers between lineages if any.

Species Delimitation
For identification of putative species and delimiting species boundaries between clades, we used a multi-criteria approach including phylogeny, coalescent species delimitation (bPTP), genetic distance, morphological characters, and geographic and ecological isolation (Shanker et al., 2017). First, a broad level phylogeny was reconstructed to determine the position of the Peninsular Indian lineages of Craspedocephalus. The results showed the presence of a few potentially new lineages distributed in the Western Ghats and Peninsular India. We tested the congruence of these lineages with PTP estimates using three concatenated mitochondrial loci (cyt b, ND4 and 16S) (Zhang et al. 2013). We used the online-based server for bPTP and ran 500000 MCMC iterations to get convergence (http://species.h-its.org/ptp). From the bPTP output, the identified lineages on the tree were evaluated carefully using other criteria prior to elevation of evolutionary significant lineages as a putative species.
We calculated genetic p-distance between clades for three mitochondrial loci (cyt b ND4, and 16S rRNA) separately in MEGA, and designated deep and shallow divergent lineages. There is shallow genetic divergence (e.g 1.9 % at cyt b, 1.5% at 16S, 2.2 % at ND4) between a few closely-related sister species (e.g. Trimeresurus andersoni vs. Tri. purpureomaculatus and Tri. erythrurus) of pit vipers with clear morphological differentiation, which were used as a benchmark for comparison. We considered distances of > 7 % at cyt b to be deeply divergent and such lineages were treated as independently evolving and designated as species on the basis of depth of genetic divergence alone. Other lineages were further examined (using additional individuals) along the other axes of morphology and geography to determine whether they should be designated as species. Shallow divergent lineages were designated as species only if they differed in morphology and were geographically isolsated.

Phylogenetics
The mitochondrial tree topology derived from Maximum Likelihood analysis of the combined dataset is similar to the tree topology obtained in earlier studies on Trimeresurus (Malhotra & Thorpe 2004. The tree has two separate clades (supported by 100 % bootstrap value) (Fig. 1, 2). One clade (clade A) consists predominantly of Peninsular Indian and Sri Lankan taxa as well as two species from Southeast Asia and is referred to as the 'Peninsular Indian group'. The other clade (clade B) in the tree predominantly comprises Trimeresurus species (e.g. Tri. salazar and Tri. erythrurus) found in the drier parts of Peninsular India (Gowri Shanker, personal communication) and other genera of pit vipers (e.g. Popeia and Viridovipera) found in the foothills of Himalayas, Indo-Burma sub-region, South China and Southeast Asia.
All Peninsular Indian taxa are clustered within a single clade along with two species from Southeast Asia. Based on our results, we refer to this clade henceforth as the genus Craspedocephalus. Peltopelor is nested within the Craspedocephalus clade and since Craspedocephalus Khul & Van Hasselt, 1822 has nomenclatural priority over Peltopelor Günther, 1864, the generic name for all Peninular Indian taxa sampled here should continue to be Craspedocephalus.
The two species from Southeast Asia, C. borneensis and C. puniceus (clade I) are sister to the Peninsular Indian and Sri Lankan group (clade II) consisting of C. macrolepis, C. strigatus, C. malabaricus and C. gramineus from Peninsular India and C. trigonocephalus from Sri Lanka. Within clade II, the Craspedocephalus macrolepis species complex is sister to the other taxa distributed in Peninsular India and Sri Lanka (clade III). Within clade III, clades IV and V share a sister relationship. Clade IV comprises the C. malabaricus species complex, which is distributed throughout the Western Ghats and is represented by three separate lineages. Two of these lineages are distributed in the southern Western Ghats; the other is in the northern and central Western Ghats, separated by the Palghat Gap from the two southern lineages. Clade V consists of C. strigatus and a clade comprising C. trigonocephalus and the C. gramineus species complex. The C. gramineus group has two separate evolutionary lineages, which are distributed across the Indian peninsula.

Species Delimitation
We delimited species boundaries of the Craspedeocephalus populations in Peninsular India (Fig. 2, See Supplementary File 4: Fig. S1). Based on a combination of criteria (genetic, geographical and morphological), we inferred a total of seven lineages within the C. gramineus, C. malabaricus and C. macrolepis species complexes in Peninsular India (Figs. 1, 2; See Supplementary File 5: Ta-ble S4). Craspedocephalus malabaricus comprises three evolutionary lineages whereas C. gramineus and C macrolepis comprise two separate lineages each. In addition, an eighth lineage, C. strigatus, is part of this clade. The three new lineages in the C. malabaricus species complex have low levels of morphological divergence but strong genetic differences due to geographic isolation (Fig. 2, 3). The northern WG group (L5) is isolated from the southern WG group by a geographic barrier, the Palghat gap, which is evident in the genetic divergence between the north-south clades. For the lineages L3 and L4 from the southern WG, the boundary can only be broadly identified as the Periyar Plateau. There was no difference in the quantitative morphological characters between any of the lineages, but the body colour pattern of the southern WG groups (L3 & L4) differentiates them from the northern population (L5). In addition, there are morphological differences between the females of the southern (L4) and the northern (L5) group. The southern group has a conspicuous post-ocular streak, which is barely visible (or sometimes completely absent) in the northern population. In addition, the southern population has strongly keeled serrated scales on the head compared to the northern population.
Similarly, there is a shallow genetic break, morphological separation and geographical isolation between the populations of C. macrolepis. The lineages are divided by the Shencottah gap as a physical barrier. There is a higher Figure 2. A Maximum Likelihood tree generated from three mitochondrial genes shows the Peninsular Indian, Sri Lankan (I) and Southeast Asian (II) clades. In the Peninsular Indian group (I), the new lineages (L1-L7) are marked with different green bars, which represent the different criteria used to delimit the species boundary. P represents the putative species predicted by bPTP analysis, while GD, M and G represent genetic distance, morphology and geographic isolation. All clades are supported with >70% parametric bootstrap value (see Fig. 1). ventral shield count in the southern lineage (L7) than the northern lineage (L6) (Fig. 2). The shallow genetic divergence (3.7% cyt b and 0.7-1.0% 16S rRNA) between lineages L6 and L7 is the lowest genetic break for the lineages inferred here, but is supported by other criteria including bootstrap support on the ML tree, geographical isolation, and morphological data.

Craspedocephalus macrolepis complex
This species complex can be distinguished from regional congeners as follows: head scales large, shield-like (vs. as small as body scales in the C. gramineus, C. malabaricus complexes and C. strigatus); midbody dorsal scale rows < 19 (vs. 21 or more in the C. gramineus, C. malabaricus complexes and C. strigatus); last rows of costal scales abutting ventrals on either side smaller than rest of the dorsal scale rows (vs. larger in all other congeners throughout the range of the genus); 2 nd supralabial bordering the anterior margin of loreal pit (vs. not so in C. strigatus); supraoculars undivided (vs. frequently divided in the C. malabaricus complex); subcaudals of the tail tip bluish (vs. blackish in the C. malabaricus complex).

Description.
A medium sized pit viper (recorded till 920mm by Ganesh et al. 2008) with a prehensile tail of length up to 136 mm; a distinct triangular head with large shield-like scales, with 1-3 cephalic scales between the undivided supraoculars with five scales surrounding the supraoculars, up to two scales between the internasal scales and the tip of the rostral scale visible from above; nasal scale entire, sub-rectangular and encompasses the nostril completely with 2-3 scales between the internasals; 1-2 scales between the nasal and the anterior part of the loreal pit; 8 supralabials and 9-12 infralabials on both sides; 6-8 scales including the last infralabial from the first ventral scale; three preocular scales of which the second and third scale from above constitute to form the posterior shields of the loreal pit; one or two post oculars; temporals smooth; 14-17 NSR (dorsal scale rows at neck), 13-19 MSR (dorsal scale rows at midbody), and 10 PSR (dorsal scale rows before vent), all keeled; Last row of scales on both sides bordered with a slightly larger row of scales, separating the ventral scales from the body; Ventrals 133-140, Anal scale undivided, followed by divided subcaudals 50-57.

Colour in life.
Specimens in life are in a uniform dark green throughout the dorsal surface, often with a black post ocular stripe that extends to 2 scale rows; the postocular stripe continues to a white to creamy white lateral stripe that stops at the vent; uniform creamy to light green ventrals with hints of light blue, that continues to the mandibular portion that is either yellow, light green, light blue or creamy white in colour; tail after the vent continues with dark green, feeble blue bands at the end of the tail, the terminal scale black or greyish with feeble white bands.

Colour in preservative.
Specimens in preservative show a range of variation in colour depending on the preservative; range from dark to light green on the head and dorsum with hints of or blue black bordering the scales; tail tip banded with grey and white bands, tail darker blue bands with bluish green.

Habitat.
A typically arboreal (rarely terrestrial) species (Ganesh et al. 2010) that is found in high elevation shola forests (tropical montane stunted rainforests) and forests bordering high elevation grasslands. Due to anthropogenic changes to the landscape, this species is also sometimes found in cultivated landscapes such as tea estates and cardamom plantations, found at elevations from about 1100 m asl to 2600 m asl.

Distribution. Endemic to the southern Western Ghats.
The revised distribution restricts this species to the high elevation landscapes of its range, the northern most limit being south of the Palghat gap, in the Nelliampathi hills, Anaimalai, with Palni hills being the eastern-most end of its distribution, extending southwards across High Wavy's or Meghamalai, Kottaimalai Ranges in the Srivilliputhur-Periyar landscape, to the Sivagiri-Devarmalai range, ending north of the Shencottah gap. Description. Holotype in good condition, dissected, with a slender, cylindrical body of snout to vent length (SVL) 263mm and a prehensile tail of length (TL) 57 mm; dorsal scales keeled with anterior dorsal scale rows (DSR) 17, mid body scale rows (MSR) 15 and posterior scale rows (PSR) 10; head prominent, of length 20.5mm, clearly distinguished from the neck with large, smooth dorsal shields on the head; rostral scale triangular with the upper side roughly half the size of the lower side with the tip visible from above; supraoculars of length 5.46 mm and frontal separated by pair of scales on both sides and nasal scales separated by three scales from above; five scales other than the pre and post ocular scales bordering the supraoculars on both sides, with three scales between the posterior edge of the supraoculars; canthus rostralis distinct with 3/3 canthal scale; three preoculars, a postocular and a thin elongated crescent shaped subocular, in contact with a small scale, encompassed by the third and fourth supralabial scale; eye with a distinct elliptical pupil, vertical diameter of the eye 2.94 mm and horizontal diameter 3.52mm; temporal scales smooth; aperture of the nostril completely covered by the nasal scale, undivided and subrectangular; nasal scale bordering the first supralabial; loreal pit present in contact with the second supralabial with two scales between the nasal and the second supralabial; eight supralabials and 12 infralabials, with six scales between the last supralabial, including the last supralabial till the start of the ventral scales; 1 st , 2 nd and 3 rd infralabial scale in contact with the first pair of genials; a gap of four scales including the posterior genials followed by 150 ventrals, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales; anal scale undivided, followed by 59 divided subcaudals scales; terminal scale on the tail larger than the previous scale, blunt at the tip.
Variation. The paratype, of SVL 504mm and TL 145mm, is discoloured with a uniform dark greenish to black colour throughout the specimen in its current preservation state. It differs from the holotype with respect to pholidosis by having 14 DSR, 14 MSR, 11 PSR, 150 ventrals and 64 subcaudals; a distinct head of length 27.5 mm with supraoculars that are separated by one scale from above; two canthal scales on the canthus rostralis, and 3 cephalic scales from above; 11 infralabials on both sides.

Colour in life.
Dark to verdant green on the dorsal surface of the head that fades into a lighter green throughout the body dorsally up to the tail and along the lateral sides of the body including the head; a prominent, 2-scale wide, white lateral stripe runs from the creamy white mandibular region; small hints and patches of blue visible on the head and dorsal scales, with the tail tip banded with dark blue or sometimes fully covered with dark blue or black; the last 11 scale rows of the tail banded with black to dark blue and white to yellow; creamy white to yellow mentum fades into a light creamy green throughout the ventrals, sometimes separated by the white ventral stripe from the lateral part of the body.

Colour in preservative.
Greyish green on the head dorsum that fades into a bluish green shade throughout the body; lateral parts of the head light green to bluish green; mentum creamy yellow that blends with creamy green ventrals, separated by the white lateral stripe; tail tip with grey and white bands, tail bluish green with darker blue bands.

Habitat.
A typically arboreal (rarely semi-terrestrial) species , Seshadri, 2012 that is found in high elevation shola forests (tropical montane stunted rainforests) bordering the high elevation grasslands. Due to anthropogenic changes to the landscape, this species is also sometimes found in cultivated landscapes such as tea estates and coffee plantations. Found at elevations from about 1200 m asl to 1868 m asl.
Distribution. Endemic to the southern Western Ghats, south of the Shencottah gap. Geographically separated from its sister species, C. macrolepis that inhabits the ranges to the north of the Shencottah gap. Found throughout the high elevations of the Agasthyamalai hills.

Craspedocephalus malabaricus complex
This species complex can be distinguished from regional congeners as follows: head scales as small as body scales (vs. large, shield-like in the C. macrolepis complex); midbody dorsal scale rows 21-23 (vs. <19 in the C. macrolepis complex); last rows of costal scales abutting ventrals on either sides larger than rest of the dorsal scale rows (vs. smaller in the C. macrolepis complex); 2 nd supralabial bordering the anterior margin of loreal pit (vs. not so in C. strigatus); supraoculars often divided or scale sutures with an indented margin (vs. undivided in the remaining, regional congeners); subcaudals of tail tip blackish (vs. bluish in the C. gramineus and C. macrolepis complexes). Lineages within Craspedocephalus mal-abaricus group appear to be morphologically cryptic, and do not differ in scalation or external morphological characters, but show geographical separation and are deeply genetically divergent.

Craspedocephalus malabaricus (Jerdon, 1854)
Figures 10-12 Trigonocephalus (Cophias) malabaricus Jerdon, 1854 Trigonocephalus (Cophias) wardii Jerdon, 1854 Lachesis coorgensis Rao, 1917 Trimeresurus malabaricus -Smith, 1943 Lachesis anamallensis (non Günther, 1864) - Wall, 1919Wall, , 1924 Rao, 1917. Jerdon (1854 in his original descriptions of Trimeresurus malabaricus and Tri. wardii did not mention any locality names or geographic province. Our perusal of historical collection catalogues (Boulenger, 1896;Das et al. 1998;Sclater, 1891;Theobald, 1868 in two potential repositories associated with Jerdon's herpetological specimens (the Natural History Museum London and the Zoological Survey of India Kolkata) all reveal that only three precisely-named places have ever been associated with these nomina: Anamalais, Nilgiris and Wayanad (also see Beddome, 1862; Boulenger, 1890). The specific epithets, mostly toponyms, also suggest the same: C. malabaricus: from 'Malabar' i.e. north Kerala and Coimbatore / Nilgiris, in the Western Ghats; C. anamallensis: from Anamalai hills of the Southern Western Ghats; Lachesis coorgensis: from 'Coorg' (=Kodagu) a part of Malnad region in Central Western Ghats. The only remaining nomen in its synonymy is Tri. wardii of Jerdon, 1854. Our perusal of historical literature reveal that this eponym could only be named after Samuel Neville Ward (18 th June 1813 -31 st January 1897). Samuel M. Ward was with the Madras Civil Service (1832-63) and had a final appointment as Judge of Koyambatur (=Coimbatore). He had reportedly been corresponding with naturalists such as Charles Darwin, Edward Blyth, as well as T. C. Jerdon, who described Tri. wardii. Ward had been reported to have visited Sirsi, in North Canara for his work on Indian wildlife, mostly butterflies and birds (Pittie, 2016 and references therein). Beddome (1862) explicitly stated that Tri. wardii is from the Nilgiris. The earliest reference of this group from Agasthyamalai is that of Ferguson (1895), who used Günther's name (C. anamallensis) to refer to it. Thus historical collections prior to typifications (till Rao, 1917) of the C. malabaricus complex apparently happened only from Anaimalai, Nilgiris and Wayanad.
Type. Syntypes lost (after Smith, 1943; also see Wallach et al. (2014:188), 'Holotype' lost (after Das et al. 1998). Hence, as per Art. 75.3 of the Code (ICZN 1999) since we are revising the taxonomy of this group, we designate as neotype the holotype of its junior synonym  Rao, 1917. The selected neotype (read below) is chosen to clarify the identity of this nominal taxon, a well-preserved adult originating from a precise locality, in conformity with the original description of the conferred nomen and housed in a permanent national repository/museum. Thus Craspedocephalus malabaricus and Lachesis coorgensis are now objective synonyms.  Description. Neotype in good condition, small lesion near the nape, possibly caused while collecting and euthanizing the individual; specimen with a slender, cylindrical body of snout to vent length (SVL) 481 mm and a prehensile tail of length (TL) 73 mm; dorsal scales keeled with anterior dorsal scale rows (DSR) 21, mid body scale rows (MSR) 21 and posterior scale rows (PSR) 13; head prominent, of length 24.3 mm, clearly distinguished from the neck with small, mildly keeled scales on the head; rostral scale trapezoid, with the lower side roughly more than twice the size of the upper side with the tip visible from above; supraoculars divided, separated by eight cephalic scales between both supraoculars at its posterior border; seven scales bordering each supraocular. Canthus rostralis distinct with four canthal scales; three preoculars, two postoculars and a thin elongated crescent shaped subocular; eye with a distinct elliptical pupil, vertical diameter of the eye 3.31 mm and horizontal diameter 3.68 mm; strongly keeled temporal scales and cephalic scales in the posterior sides above the mandibular joint; aperture of the nostril completely covered by the nasal scale, undivided and sub-pentagonal shaped, in contact with the first two canthal scales and the 1 st and 2 nd supralabial; loreal pit present in contact with the second supralabial and the 2 nd and 3 rd preoculars; nine supralabials and 12 infralabials, with eight scales between the last supralabial, including the last supralabial up to the start of the ventral scales; 1 st , 2 nd and 3 rd infralabial scale in contact with the first pair of genials; a gap of four scales including the posterior genials followed by 148 ventrals, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales; anal scale undivided, followed by 38 divided sub-caudals scales; Terminal scale on the tail larger than the previous scale, blunt at the tip.

Lachesis coorgensis
Variation (n=20). The referred materials are of SVL up to 670 mm and TL up to 126 mm with colours varying from dark brownish red to light green throughout the specimens in its current preservation state, differs from the holotype with respect to pholidosis by having 19 to 22 DSR, 19 to 23 MSR, 13 to15 PSR, 145-149 ventrals and 52-54 subcaudals; head distinct with supralabials ranging between 8-10 and infralabials ranging between 10-13; one to three preocular scales, one to two scales (some specimens showing an absence of these scales) between the 3 rd supralabial scale and the suboculars, seven to nine cephalic scales and seven to eight scales surrounding the supraoculars from the dorsum.

Colour in life.
A highly variable and polymorphic species, with respect to colour, specimens can be found in a variety of colour morphs -greenish blue-cyan, bright yellow, green, rufous brown, bright orange and red coloured specimens have been encountered during this study; head characterized with a thick dark brown to black postocular stripe till the nape, labials sometimes marked with blotches and a highly variable pattern above the head, sometimes fully dark, some individuals with no markings at all, body with alternating zig-zag saddle shaped markings with the last rows of scales on the tail banded with different colour; these markings vary from brick reddish, dark brown to black, sometimes intermixed with spots of other colours such as green, yellow and blue; the base colour of the body varies from light brown in juveniles, light cream, orange, yellow, brick red, bluish green and sometimes morphs mottled with all or some of the aforementioned colours; ventrals sometimes vary from the colour of the dorsum, complementing the variety of vibrant dorsal colours, but often are coloured similar to the dorsum; colour change is also observed to be seasonal (Kanagavel et al. 2012); juveniles brown with dark brown to black markings, neonates and younger juveniles possess a tail 'lure' that is often different from the body's colouration. Mandibular region and the ventrals in alternating light green, blue, yellow to creamy yellow with speckles, separated from the dorsal scales with a longitudinal lighter irregular stripe, two prominent, labial stripes from the eye and the loreal pit, up to the edge of the lower end of the supralabial region.

Colour in preservative.
Fairly faded specimen with a light brown dorsum, ventrals and tail; scales bordered with dark brown; temporal stripe visible on the right side of the specimen in dark brown, left side with no visible temporal stripe; body with barely distinguishable dark brown saddle shaped markings throughout the body; tail with dark brown stripes from the vent up to the tip.
Habitat. An arboreal species, commonly found on bushes and in undergrowth in forests and near streams in evergreen forests to moist deciduous and lowland riparian forests. Due to anthropogenic changes to the landscape, this species is also abundantly found in agricultural landscapes such as coffee plantations, from 100-1800m MSL.
Distribution. Restricted to the central and northern Western Ghats from Mahabaleshwar -Koyna in Maharashtra to the Nilgiris and Elivalmalai hills, north of the Palghat Gap. Known to occur sympatrically with C. occidentalis comb. nov. in Nilgiris and Coorg, with C. gramineus in Amboli and may have some overlap with C. strigatus in the mid to high elevations of the western slopes of the Nilgiris (at the upper limit of its altitudinal distribution).
Description. Lectotype in a generally good condition, entire with a cylindrical body of SVL 505mm and a prehensile tail of TL 87mm; dorsal scales mildly keeled with DSR 21, MSR 21 and PSR 15; head of length 29mm prominent and clearly distinguished from the neck with strongly imbricate small scales; tip of the rostral scale visible from above, with the upper end roughly half the size as the lower; divided supraoculars with nine cephalic scales between both the supraoculars; nine scales surrounding each divided pair of supraoculars on both the sides with nine scales between the posterior border of the supraoculars; distinct canthus rostralis with four scales on the canthal ridge; two preoculars and two postoculars, an elongated cresent shapend subocular; strongly keeled temporals and keels continue to be present in other head scales behind the oculars except the supralabials towards the posterior; eye with a distinct elliptical pupil of vertical diameter 3.3 mm and a horizontal diameter of 3.98 mm; nostril aperture completely covered by the nasal scale, undivided and pentagonal-sub rectangular in shape, in contact with the first three canthal scales, first and second supralabial; nine supralabials and eleven infralabials, with eight scales between the edge of the mouth and the first ventral scale; 1 st , 2 nd and 3 rd infralabial scale in contact with the first pair of genials; a gap of six scales in between the first genial and ventrals; 157 ventrals separated laterally from the body scales by a row of slightly broader dorsal scales; anal scale undivided followed by 55 divided caudals; terminal scale rounded and blunt at the tip, slightly larger than the previous scale.
Variation (n=10). The following characters vary within the specimens of the examined type series. Variations in pholidosis between the specimens were: supralabials 9-10 and infralabials 11-13, preoculars 2-3, ventrals 144-145 and subcaudals 50-62, about 8-12 scales between the edge of the mouth and the ventral scales; 21-22 scale rows around the neck; the post ocular stripe sometimes extends to 2 rows of scales.

Colour in life.
Black dorsal head scales with the anterior scales with hints of light green and posterior head scales bordered with yellow, up to the postocular eye stripe, that extends untill the nape; light bluish green on the lateral parts of the head that fades into a creamy yellow to white underside, from the mandibular region up to the ventrals; ventral scales creamy yellow scales alternating with light greenish yellow scales, consecutively larger gaps between the lighter scales filled with the greenish yellow scales towards the tail -these correspond to the alternating between creamy yellow and green scales in the column that separates the ventrals and dorsal scales; caudal scales yellow, bordered and often blotched with black scales; black blotches throughout the dorsum with a gap of 3-4 scale rows.

Colour in preservative.
Head dorsum almost completely covered black to dark brown and scales bordered with light faded green; black/ dark brown postocular stripe about 2 scales wide, a preocular/ temporal stripe that continues to the loreal pit and ends at the supralabials be-low; black markings on the labials below the suboculars; body in light faded green with black saddle shaped markings, the markings centered with faded brown marbled markings; the row dorsal of scales that meet the ventrals alternate between the dorsal light faded green and dark  brown/ black markings with a gap of two to three scales in between them; ventrals plain light creamish yellow; subcaudals in black with yellow blotches.
Habitat. Similar to C. malabaricus, an arboreal species, commonly found on bushes and in the undergrowth in forests and near streams in moist evergreen forests to deciduous and lowland riparian forests. Due to anthropogenic changes to the landscape, this species is also often found in cultivated landscapes such as coffee plantations, from 100-1800 m asl.
Distribution. Endemic to the southern Western Ghats, south of the Palghat Gap and north of the Shencottah Gap.
We recorded this species in the Nelliamapthy, Anamalai and Palni hills, Cardamom hills, and northern sections of the Periyar plateau. Craspedocephalus anamallensis has been recorded to be sympatric with C. macrolepis at the highest limit of its elevational range (see Malhotra & Davis, 1991). Etymology. Toponym, named after its distribution in the far south of the Western Ghats, in the southern parts of the 'Travancore' hill ranges.

Remarks.
As previously elaborated and clarified (see C. malabaricus account), historical collections and typifications (both prevailing and previously synonymized treatments) did not involve the population circumscribed here as a distinct lineage. Therefore, this innominate population, previously cited (see Inger et al. 1984;  with the second supralabial with two scales between the nasal and the second supralabial: nine supralabials and 11 infralabials, with nine scales between the last supralabial, including the last supralabial till the start of the ventral scales; 1 st , 2 nd and 3 rd infralabial scale in contact with the first pair of genials; a gap of three scales including the posterior genials followed by 147 ventrals, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales; anal scale undivided, followed by 55 to 56 divided subcaudals scales; terminal scale on the tail larger than the previous scale, blunt at the tip.
Variation. The paratypes have SVL upto 282 mm and TL 65 mm, and differ from the holotype with respect to pholidosis by having 20 to 21 DSR, 21 MSR, 14 to 15 PSR, 157 ventrals and 55 subcaudals; three to four canthal scales on the canthus rostralis, and seven to nine cephalic scales from above; 10 to 12 infralabials on both sides.

Colour in life.
Head dorsum almost covered with dark brown to purplish brown colour with scales bordered with light yellowish green; postocular and preocular stripe almost indistinguishable from the head dorsum colour, separated with a faded yellowish green stripe; dark brown postocular stripe about 3 scales wide, a preocular/ temporal stripe that continues to the loreal pit and ends at the supralabials below; black border markings on the labials below the suboculars and anterior supralabials; body in light faded green and brown marbled scales with 28 brown saddle shaped markings, the markings centered with faded brown marbled markings; the row dorsal of scales that meet the ventrals alternate between the dorsal marked with faded yellow and dark brown with a gap of two to three scales in between them; ventrals plain light creamish yellow; subcaudals in black with yellow blotches; tail with 13 yellowish green bands on dark brown to black; eyes silverish with a tinge of yellow, rufous red blotches throughout the eye, concentrated towards the middle, perpendicular to the pupil, almost forming a cross.  Colour in preservative. Similar to colouration in life with brighter colours faded dull and yellows throughout the body bleached to a creamish white colour; pupils bluish white, dilated.
Habitat. Similar to other members of the C. malabaricus group, an arboreal species, commonly found on bushes and in the undergrowth in forests and near streams in moist evergreen forests to deciduous and lowland riparian forests, from 100-1800 m asl.

Distribution.
Recorded mostly from Agasthyamalai with a single isolated record from Devarmalai hills, across the Shencottah gap. This species may occur sympatrically with C. peltopelor sp. nov. in the highest elevations of Agasthyamalai.

Craspedocephalus gramineus complex
This species complex can be distinguished from regional congeners as follows: head scales as small as body scales (vs. large, shield-like in the C. macrolepis complex); midbody dorsal scalerows 21-23 (vs. <19 in C. macrolepis complex); last rows of costal scales abutting ventrals on either sides larger than rest of the dorsal scale rows (vs. smaller in the C. macrolepis complex); 2 nd supralabial bordering the anterior margin of loreal pit (vs. not so in C. strigatus); supraoculars undivided (vs. often divided or with indented scale in the C. malabaricus complex); subcaudals of tail tip bluish (vs. blackish in the C. malabaricus complex).

Figure 19
Coluber gramineus (1796). These nomina are connected by an objective synonymy and therefore share the same type specimen (see Russell, 1796) and type locality, Vizagapatam. Notwithstanding these, Lacépède (1804) Vogel et al. 2014 for its current status). Stejneger (1907) also followed Boulenger (1890). This diluted and over-circumscribed concept of 'Tri. gramineus' (sic) encompassing almost all green Trimeresurus species, gave rise to the false notion that it is a very widespread and variable species.
Remarks. Coluber viridis Bechestein, 1802 and Vipera viridis Daudin, 1803 are both objective junior synonyms of Coluber viridis Shaw, 1802, all being described on the basis of the very same type specimen described and illustrated by Russell (1796) as "Boodroo Pam". Trimeresurus viridis Lacépède, 1804 is a nomen oblitum superseded by Trimeresurus insularis Kramer, 1977 that is now in prevailing usage for this Indonesian species (Wallach et al. 2014). Two female paratype specimens from the Eastern Ghats and one male specimen from Matheran, northern Western Ghats of a former subjective junior synonym (fide Wallach et al. 2014) Trimeresurus occidentalis Pope & Pope, 1933 is herein referred to Tri. gramineus (Shaw, 1802) with respect to its phylogenetic and morphological affinities.

Lineage diagnosis.
A cryptic lineage (L1) belonging to the C. gramineus complex, it is distinguished from C. occidentalis (L2) as follows: lower ventral scale count 158-179 (vs.142-154). It is genetically divergent from C. occidentalis by 8.1 % and 1.0 % at cyt b and 16S respectively. This lineage is far more widespread than its parapatric sister taxon C. occidentalis, occuring almost throughout peninsular India from Odisha in the east, southern Gujarat to the north-west and as far south as the Srivilliputhur hills.

Description of referred material (n=11).
A species with a slender, cylindrical body of snout to vent length (SVL) up to 679 mm and a prehensile tail; dorsal scales keeled with anterior dorsal scale rows (DSR) 19-21, mid body scale rows (MSR) 19-21 and posterior scale rows (PSR) 15; head prominent, clearly distinguished from the neck with small, juxtaposed scales on the head; rostral scale sub triangular with the upper side roughly one fourth the size of the lower side with the tip visible from above; supraoculars separated by 8-11 scales, between the posterior edge of the supraocular scales; canthus rostralis distinct with three to four canthal scales on the ridge; two to three preoculars, two to three postoculars and a thin elongated crescent shaped subocular; eye with a distinct elliptical pupil; temporal scales mildly keeled; aperture of the nostril completely covered by the nasal scale, undivided and subrectangular; nasal scale bordering the first supralabial; loreal pit present in contact with the second supralabial with two scales between the nasal and the second supralabial; 10-12 supralabials and 11-13 infralabials, with 9-13 scales between the last supralabial, including the last supralabial up to the start of the ventral scales; ventrals 158-179, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales anal scale undivided, divided subcaudals 54-67; terminal scale on the tail larger than the previous scale, blunt at the tip.

Colour in life.
Head and dorsum colour from verdant green to leaf green, sometimes bluish green with black, alternating saddle shaped markings on the dorsum; preoc-ular/ temporal stripe in black, sometimes green fading into black; ventrals with bright yellow, creamish yellow to a dirty white colour; the region where the ventrals meet the dorsum alternating with the ventral colour once every 2-3 scales.

Habitat.
A highly arboreal species found in dry scrub, dry deciduous, semi evergreen and lowland riparian forests from 100-1600 m asl.
Distribution. Endemic to Peninsular India. Found in the Eastern Ghats (Simlipal hills, Chota Nagpur plateau southwards upto Sirumalai hills) and hills of Central India (Pachmahri, Seoni hill ranges) as well as the Western Ghats from Surat Dangs, southwards till the Shencottah gap. (Pope & Pope, 1933) comb. nov.

Lachesis graminea -Wall, 1919 part
Craspedocephalus occidentalis (Pope & Pope, 1933) Taxonomic History. Pope & Pope (1933), in an effort to fix the status of Southeast Asian taxa of superficially similar green Trimeresurus species, worked out the systematics of the Indian group. Unfortunately, they were mistaken in allocating the nomen Coluber gramineus Shaw, 1802 to the East Asian species (see David et al. 2011). Therefore, thinking that the western population inhabiting India required a new nomen, they thus erected Trimeresurus occidentalis. This taxon was described based on a series of specimens (holotype and paratypes) inhabiting both the Western Ghats and the Eastern Ghats.
Remarks. Among the four paratypes attributed to this taxon in its original description (Pope & Pope, 1933), one male from Brahmagiri, Western Ghats is currently attributable to this species based on the phylogenetic position of topotypical samples. The distribution of C. occidentalis, as defined here, encompasses the provenance of the holotype (Mudumalai, Wayanad, and the Nilgiri hills in general). We refer three other paratypes, two females originating from the Cuddapa hills and the Shevaroys in the Eastern Ghats and one male from Matheran to C. gramineus (Shaw, 1802)    loreal pit present in contact with the second supralabial with two scales between the nasal and the second supralabial; nine supralabials and 10 infralabials, with seven scales between the last supralabial, including the last supralabial up to the start of the first ventral scale; 1 st and 2 nd infralabial scale in contact with the first pair of genials; a gap of six scales including the posterior genials followed by 155 ventrals, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales; anal scale undivided, followed by 60 divided subcaudals scales; terminal scale on the tail larger than the previous scale, blunt at the tip.

Colour in life.
Head and dorsum colour varies from bright green to faded dull green, with no markings on the dorsum; preocular/ temporal stripe absent; ventrals in creamish yellow to dirty white colour; the region where the ventrals meet the dorsum alternating with the ventral colour once every 2-3 scales; the colour of the dorsum and ventrals divides exactly at the pre/postocular stripe area, with an absence of a dark stripe (as opposed to C. gramineus).

Colour in preservative.
Head and dorsum in light faded green; ventrals creamish to dirty white in colour; some of the dorsum scales with a black interscalar colour, more evident in the front half of the body; the colour of the dorsum and ventrals divides exactly at the pre/postocular stripe area, with an absence of a dark stripe (as opposed to Tri. gramineus).

Habitat.
A highly arboreal species found in moist evergreen forests from 1000-1900m MSL, sympatric with C. malabaricus throughout parts of its range.  thal scales; two to three preoculars, two to three postocular and a thin elongated crescent shaped subocular; eye with a distinct elliptical pupil; temporal scales smooth; aperture of the nostril completely covered by the nasal scale, undivided and subrectangular; nasal scale bordering the first supralabial; loreal pit present in contact with the second supralabial with two scales between the nasal and the second supralabial; nine to 10 supralabials and 10 to 12 infralabials, with six to eight scales between the last supralabial, including the last supralabial till the start of the ventral scales; 1 st , 2 nd and 3 rd infralabial scale in contact with the first pair of genials; a gap of three scales including the posterior genials followed by 134 to 142 ventrals, laterally separated from the dorsal scale rows by a slightly broader row of dorsal scales; anal scale undivided, followed by 38 to 44 divided subcaudals scales; terminal scale on the tail larger than the previous scale, blunt at the tip. Colour in life. Bronze to light brown dorsum blotched with a stark, continuous alternating saddle-shaped pattern in dark brown to black, strikingly similar to the markings on Vipera berus or Gloydius himalayanus; preocular/temporal stripe in dark brown; post ocular stripe in dirty brown continuing towards the loreal pit and the infralabials; another stripe below the subocular stripe fades into the infralabials followed by another blotch towards the end of the infralabials; base colour of the infralabials and ventrals being light creamish to white in colour, of-ten dotted with rufous spots in the supralabials the region where the dorsal scales meet the ventrals in altenating dark brown colour and light brown/bronze colour scales; dorsal bronze scales are dotted with darker brown; the nape is characterized with a prominent horse shoe shaped marking hence earning its common name; in juveniles, the bronze colour is replaced with light brown.
Colour in preservative. Brown and cream markings dorsally and laterally, occasionally interrupted with silvery/ grey scales, tail mostly brown and cream coloured with lighter more vivid cream coloured scales making triangular markings vertically along it (four in total); tip of tail fading to a light cream colour.

Habitat.
A highly terrestrial species found in montane evergreen, moist evergreen and high elevation grasslands.
Distribution. Endemic to the Nilgiri Massif north of the Palghat Gap, distributed in very high elevations (1700-2600 m asl) (also see Wall, 1919;Bhupathy & Nixon, 2011). This species has the smallest geographic range of all Craspedocephalus in Peninsular India (Ganesh & Chandramouli, 2018).

Discussion
The systematics of pit vipers of Peninsular India has been in flux, given their variation in morphology, their distribution in different habitat types and uncertainty of species delimitation criteria (Freitas et al. 2020). We carried out this study to delineate independent evolutionary lineages and evaluate congruence between genetic and morphological variation in the genus Craspedocephalus in Peninsular India, using a stepwise multi-criteria approach to delimit species boundaries in this group (Shanker et al. 2017). This study designates four new species in addition to four currently recognized species and confirms the systematic position of two endemic species (C. strigatus and C. macrolepis) belonging to Peninsular India. There are now a total of eight lineages found in the Western Ghats and Peninsular India (Fig. 26). The phylogeny generated using three concatenated mitochondrial fragments of almost all the populations and species of Craspedocephalus found in Peninsular India is similar to the tree topology published in earlier studies (Malhotra & Thorpe 2004, Alencar et al. 2016, including studies which used nuclear as well as mitochondrial genes . While a combination of nuclear and mitochondrial genes may provide stronger phylogenetic support for deeper relationships of pit vipers within the clade as well as with other clades, there is strong support for the species level divergences that we report here. Our study further supports the systematic position of C. strigatus within the Craspedocephalus clade. We also found that C. macrolepis is nested within Craspedocephalus. The chromatically variable and widely distributed taxon, C. malabaricus, was recovered with three divergent lineages, two of which are distributed in the Southern Western Ghats and are sister to each other, with high genetic divergence from a northern Western Ghats lineage. Considering the type locality of C. malabaricus as the Western Ghats (all west coast forests in Peninsular India) (Jerdon 1854, Günther 1864), the other two new lineages are recognized as distinct species on the basis of genet-ic distance, morphological dissimilarity, and geographic barriers between them.
This delineation takes into account the geographic ranges of each of the lineages within the C. malabaricus species complex: C. malabaricus (Jerdon, 1854), C. anamallensis (Günther, 1864) and C. travancoricus sp. nov. whose distributions coincide with segments of the Ghats separated by significant gaps. However, there were morphological differences (i.e. strongly keeled dorsal scales) between Jerdon's (1854) Tri. malabaricus and Günther's specimen, suggesting that our lineage L3 (Tri. anamallensis) is closer to Günther's description. Lineage L5 corresponds to C. malabaricus on the basis of morphological similarity, while Lineage L4 is similar to Jerdon's C. mala baricus but is geographically disjunct from the former. This historically untypified population, so far cited under the chresonymy of Tri. malabaricus sensu lato, is here named as C. travancoricus sp. nov.. We found that all species within the C. malabaricus group show colour polymorphism and are without much sexual dimorphism.
Similarly, the C. gramineus species complex comprises two highly divergent lineages (7 % at ND4 and 8 % at cyt b) with significant morphological variation and strong habitat preference in those lineages (authors' personal observation). Morphologically, they differ in ventral scales and supralabials. Lineage 1 (C. gramineus) is distributed in the drier and low-mid elevations of hill ranges (till ca. 1000 m asl) of Peninsular India while Lineage 2 (C. occidentalis) is distributed in the wet forests of the Western Ghats. These findings are supported by earlier work (Pope & Pope, 1933).
The endemic C. macrolepis was classified as a distinct species with significant morphological differentiation and was found to be nested within the Craspedocephalus clade of Peninsular India, as suggested by earlier studies based on morphology alone (Beddome 1862, Guo et al. 2010, David et al. 2011. Genetic support was too low to consider the genus 'Peltopelor' as different from Craspedocephalus despite differences in hemipenial and other morphological characters (Malhotra & Thorpe 2004, Guo et al. 2010). As Craspedocephalus Khul & Van Hasselt, 1822 has nomenclatural priority over Peltopelor Günther, 1864, the genus name for all Peninsular Indian taxa sampled here remain Craspedocephalus. We found two distinct lineages of C. macrolepis that were morphologically divergent, geographically separated and with shallow genetic divergence from each other in the southern Western Ghats, that were elevated to species status. In addition, C. strigatus was confirmed to be a member of the Indian subcontinent group as earlier suggested by Malhotra & Thorpe (2004) rather than Protobothrops as suggested by Kraus et al. (1996).
Our work suggests that there are (at least) eight independent evolutionary lineages in Peninsular India, supported by genetics, morphology, and geography. Besides naming two new lineages, both from the southernmost part of the Western Ghats, this study also clarifies the taxonomic status, distribution and phylogeny of Peninsular Indian Craspedocephalus involving both recognized and synonymized taxa. Of the eight species of Craspedoceph-alus belonging to Peninsular India, seven are restricted to the wet forests of the Western Ghats. The southern Western Ghats stands out in terms of species richness for Craspedocephalus in Peninsular India with four endemic species, which is evident in other taxa including frogs  . New lineages (green bars) are depicted within C. gramineus, C. malabaricus and the C. macrolepis species complex along with the other species from the Peninsular India. The clade comprising C. macrolepis and C. peltopelor sp. nov. is sister to the clade comprising all other members of Peninsular Indian Craspedocephalus (II). The Western Ghats C. malabaricus complex includes C. malabaricus, C. anamallensis and C. travancoricus (IV). C. trigonocephalus from Sri Lanka is nested within the Western Ghats sub-clade (III) along with C. strigatus, C. gramineus and C. occidentalis (V). Clade-I comprises species from Southeast Asia.
The Peninsular Indian clade comprises almost all taxa distributed across South Asia except Tropidolaemus huttoni. The phylogenetic relationship of Tro. huttoni with Trimeresurus is unknown. Tropidolaemus huttoni is currently known from two neonate museum specimens, which are underdeveloped sexually. Based on its morphological similarity to Tropidolaemus wagleri, David & Vogel (1998) placed Tro. huttoni in this genus.
This work underscores the indispensable role of genetic studies that enable us to fully document cryptic diversity including in lineages that have remained taxonomically 'stable' for over a century (since Rao, 1917). This study also highlights the profound and untested effect of geographic barriers on this group of snakes. Though there were several synonyms and typifications within, for example, the C. malabaricus group between 1854 and 1917, these were primarily driven by morphological variations and not based on geographic rationale. That we uncovered geographically concordant sister lineages within the genus Craspedocephalus in the Western Ghats exemplifies the effectiveness of allopatry as a prezygotic reproductive barrier, one that does not instigate morphological changes in lineages, resulting in a high degree of crypsis, but at the same time leading to deep genetic divergence. This pattern is also observed in vine snakes (Ahaetulla -Mallik et al 2020) and to a degree, in cat-eyed snakes (Boiga - Ganesh et al. 2021). A comparative study with the inclusion of nuclear genes in the future may explain and shed light on gene-flow between these populations over their distribution ranges and across barriers.