A new species of Rhinophis Hemprich, 1820 (Serpentes: Uropeltidae) from the Wayanad region of peninsular India

Corresponding author. � d.gower@nhm.ac.uk; https://orcid.org/0000-0002-1725-8863 Abstract A new species of the shieldtail snake genus Rhinophis is described based on a type series of seven recently collected specimens from the Wayanad region of the Western Ghats of peninsular India. Rhinophis melanoleucus sp. nov. is diagnosed based on a combination of 15 dorsal scale rows at (or just behind) midbody, more than 215 ventral scales and a long rostral. The new species also has a distinctive (mostly black and white) colouration. A new key to the identification of Indian

Rhinophis is currently represented by 20 nominal species, with 16 of these endemic to Sri Lanka, and only four currently recognized species occurring in (and endemic to) India. Of the four species found in India, three are restricted to the central and southern Western Ghats (R. sanguineus, Beddome, 1863, R. travancoricus Boulenger, 1893, R. fergusonianus Boulenger, 1896 and one species is known to occur in the southern part of the Eastern Ghats (R. goweri Aengals & Ganesh, 2013). Here we describe a new Indian species of Rhinophis on the basis of recently collected specimens from the Wayanad region of the Western Ghats.

Materials and methods
In addition to specimens of the new species deposited in the Bombay Natural History Society, Mumbai, India (BNHS), and the Western Ghats Regional Centre of the Zoological Survey of India, Kozhikode (ZSI/WGRC), we examined uropeltid material in the Natural History Museum, London, UK (BMNH), the Muséum national d'Histoire naturelle, Paris, France (MNHN), Muséum fur Naturkunde, Berlin, Germany (ZMB), Museìum d'Histoire naturelle, Geneva, Switzerland (MHNG), Museum of Comparative Zoology, Harvard University, USA (MCZ), California Academy of Sciences, San Francisco, USA (CAS), American Museum of Natural History, New York, USA (AMNH), and Smithsonian National Museum of Natural History, Washington, USA (USNM). Comparative material of Indian Rhinophis is reported in Appendix 1. Taxonomy and taxon spellings follow McDiarmid et al. (1999) and Pyron et al. (2016).
Ventral scale counts were recorded following Gower & Ablett (2006). Scale row reductions were recorded following Dowling (1951). All measures were taken with dial calipers to the nearest 0.1 mm, except for total length and circumference, which were taken to the nearest 1 mm using a ruler and a piece of thread plus ruler, respectively. Snout-vent length was calculated by subtracting tail length from total length. Bilateral measures were taken on the right side of each specimen, unless that side was damaged. Tooth counts were made from wet specimens, these are estimates because it is often not possible to see or feel presence or absence of teeth among gingivae without error. Sex was determined by observing everted hemipenes, or by examining urogenital systems in situ through small ventral incisions in the body wall. Following Gower et al. (2008), Maduwage (2011) andJins et al. (2018), we made an effort to examine non-traditional characters for uropeltid taxonomy.
Description of holotype (Figs. 1-3). See Table 1 for morphometric and meristic data. Good condition; midventral incision (for removal of tissue biopsy) 17 mm extending back from 125 mm from snout tip. Head small, snout pointed. Rostral pointed, longer than wide, without dorsal crest but with narrow, rounded dorsal ridge; in lateral view with slightly convex ventral and (more strongly) dorsal margins; widest at level of anterior upper corner of first supralabials. Rostral many (>12) times longer (in dorsal view) than rostral-frontal gap. Frontal irregularly hexagonal, longer than wide, lateral (ocular) margins slightly converging posteriorly; lateral (ocular) margin shortest, posterolateral edges longest. Frontal much shorter, wider than rostral. Nasals separated from each other by posterior half of rostral. External naris small, subcircular, slightly countersunk within small depression, located at anteroventral corner of nasal. Nasal contacts supralabials 1 and 2. Prefrontals only briefly in contact with each other along midline (left overlapping right), separating frontal from rostral [rostral post-nasal longer than prefrontal midline contact]. Prefrontals wider than long, shorter than frontal. Supralabials four; first smallest, making the least contribution to margin of mouth; fourth largest. Ocular contacts supralabials 3 and 4. Eye distinct, diameter approximately 0.2 times length of ocular, located near anteroventral corner of ocular (closer to lower than anterior edge), only very slightly (at most) bulging slightly from ocular surface, pupil subcircular. Paired parietals longer than wide, approximately as long as but much wider than frontal, posteriorly broadly rounded, angle between posteromedial and posterolateral edges approximately 90°. Parietals in brief midline contact (longer than midline contact between prefrontals), left overlapping right. Each parietal contacts four scales other than head shields. No mental groove; mental pentagonal, slightly wider than long, smaller than infralabials, contacting first infralabials but not first ventral; three pairs of infralabials, second largest, first smallest. First and second ventrals longer than wide, third approximately as long as wide, fourth and subsequent ventrals wider than long. Four (fourth slightly smaller than first three) maxillary and five (fourth and fifth slightly smaller than first three) mandibular teeth on each side. Teeth simple, pointed, distinctly retrorse, straight, evenly spaced. Body cylindrical. Head and body scales macroscopically smooth. Body scales generally evenly sized on dorsum and along body except for those involved in dorsal scale row reductions. Midline ventral scales between mental and anal of even size, though anteriormost ones gradually narrow. At midbody, exposed part of ventrals approximately 1.4 times wider than scales in first dorsal row. Ventrals 231. Dorsal scale rows 19 anteriorly, reducing to 17 by level with 41st ventral and to 15 rows by 125th ventral, maintained thereafter until close to the vent; Scale row reduction formula: Views of the head (left) and tail (right) of the holotype (BNHS 3534) of Rhinophis melanoleucus sp. nov., shown in dorsal (upper), lateral (middle) and ventral (lower) views. Not to same scale; see Table 1 for size of specimen. Photographs by VPC.
Dorsal scale rows (i.e., excluding subcaudals) 15 at level of first subcaudals. Paired anal scales (right overlying left) considerably larger than posteriormost ventrals and subcaudals. Distal margin of each anal overlaps four (left) and three (right) small scales in addition to anteriormost subcaudals. Six subcaudals on each side, the posteriormost undivided. Some scales at posterior end of specimen bear low, short parallel ridges, towards posterior edges-on last few ventrals, small scales overlapped by anals, lower two or three dorsal scale rows of posterior of body and of tail, and on subcaudals. Last ventrals and subcaudals each bear three or four ridges, last undivided subcaudal bears seven. Tail 'shield' large, forming tip of tail, visible from below and especially clearly from above, flattened to slightly concave on anterior end of upper surface, domed posteriorly, longer than wide in dorsal view, wider than depth of tail (at base of shield), larger than head (longer than distance between snout tip and back of fourth supralabial), base surrounded by 15 scales (including last subcaudal). Shield surface roughened, bearing narrow, discontinuous ridges (longer, more continuous stretches located laterally towards shield base), receding and somewhat converging towards tip; evenly spaced, subparallel, approximately straight; low projections but no ridges at (just dorsal to) shield apex.
Colour in alcohol (Figs. 1-2). Rostrum pale orange brown at tip, becoming greyish black posteriorly on the dorsal surface. Head shields greyish black. Supralabial scales greyish black with pale whitish yellow markings, especially towards posterior and/or lower margins. Dorsal body colour uniformly blackish. Ventral surface blackish overall, more greyish on anterior quarter of body. Lateral sides each with approximately 56 obvious, irregular yellowish-white blotches occupying three to five ventralmost dorsal scale rows, many of these blotches taper irregularly dorsally. Ventral scales mostly greyish black, a small number of ventrals with yellowish white marks, typically only on left or right. Subcaudals black. Tail shield black with elongate pale yellowish orange patches on each lateral side, continuous with pale (more whitish, less or not yellowish) markings on ventrolateral surface of tail.
Colour in life (Fig. 3). Dorsal surface uniformly glossy blackish and somewhat iridescent. Lateral and ventral pale blotches on body whitish, more purely so on mid and posterior of body than anteriorly (where whitish scales appear translucent so that darker spots beneath scales slightly visible). Head scales blackish except for paler, orangebrownish rostral and whitish lower and/or posterior parts of supralabials. Ventral surface of tail blackish except for whitish ventrolateral markings. Pale elongate markings on tail shield pale orange, paler and more whitish anteriorly than posteriorly.  Paratypes. All paratypes in good condition, BNHS 3535 and BNHS 3536 a little dehydrated with convoluted spines, the latter specimen's head slightly crushed. Meristic and morphometric data are provided in Table 1. The number, arrangement and overlapping of head shields are similar to the holotype with the following exceptions. In BNHS 3537 and BNHS 3538 (Fig. 6) the parietals are each wider than long, slightly shorter than the frontal, and with somewhat scalloped rather than broadly rounded posterior margins. The prefrontals in ZSI/WGRC/IR/V/3100, BNHS 3537 and BNHS 3538 make more substantial midline contact than in the holotype, so that the length of rostral behind the nasals is shorter than the rostral-frontal distance. Tooth counts same as in holotype, where examined (BNHS 3535, BNHS 3536, ZSI/WGRC/IR/V/3100). Rostral length from approximately 5 (ZSI/WGRC/IR/V/3100) to 14 (ZSI/WGRC/IR/V/3101) times distance between rostral and frontal. Eye diameter approximately 0.2-0.25 times length of ocular shield. Everted hemipenes of BNHS 3537 (Fig. 6) short (ca. 2.5 mm), stout (ca. 1.0 mm wide at base). Interpreted as unilobed but with substantial medial lobe-like process (lacking any sign of sulcus spermaticus) ca. 1.0 mm from tip of organ. Medial process shorter and more slender than lobe, with irregular longitudinal folds. Lobe subcylindrical, ornamented densely with small spines throughout; shallow sulcus spermaticus terminating at base of lobe with large flap and several globular folds.
Sexual dimorphism. There is no clear evidence of sexual dimorphism in number of ventral scales in the types, with the three females having 225-235 and three males having 218-236 (Table 1). Males tend to have proportionately slightly longer tails (2.6-3.5 % of total length) than females (2.3-2.8 %) and have more subcaudals (7,7 or 8,8) than females (6,6 or 6,7). Both females and males have ridges on scales on the underside of the tail and posteriormost part of the body, but these are more prominent in males.    Etymology. From the Ancient Greek mélas (black) and leukós (white), in reference to the unusual (for uropeltids) black and white colouration. For nomenclatural purposes, the species name melanoleucus is a noun in the genitive case.
Distribution, habitat, natural history and conservation status. Rhinophis melanoleucus sp. nov. is known only from the vicinity of Lakkidi in the Wayanad District of Kerala state, at approximately 750-850 m elevation in the evergreen hills of the Western Ghats. The habitat in the vicinity of the type locality is shown in Fig. 7. We suspect that the new species has a larger distribution, at least in the Wayanad region, but it is not widespread and/or frequently encountered enough to have been previously collected or reported. The new species is likely to qualify for Data Deficient status under IUCN Red List criteria, at least until new field surveys are undertaken and/or additional specimens from other localities can be found in other collections.
The holotype was found at 08:00, moving on the surface of a forest track alongside a stream and close to an adjacent tea plantation. Paratypes BNHS 3537 and BNHS 3538 were found at 15:00 and 09:00, respectively, the former dead on a paved road, and the latter on the ground surface in an abandoned coffee plantation. Referred specimen BNHS 3539 was dug from a depth of approximately 0.5 m during excavations for a road extension in mid-elevation wet-evergreen forest (rainfall approximately 5,000 mm per year) with trees including Cinnamomum malabatrum (Burm. f.) J.Presl, Meliosma simplicifolia (Roxb.), Actinodaphne malabarica Balakr. and Elaeocarpus tuberculatus Roxb. in addition to farmed coffee. Paratypes ZSI/WGRC/IR/V/3100 and ZSI/WGRC/IR/V/3101 were found at approximately 07.00 and 18.30 respectively, moving among grass on the side of a tarred road inside the Veterinary and Animal Sciences University campus, Pookode. Paratypes BNHS 3535 and BNHS 3536 were found dead on a tarred road between 07.00 and 08.00.
In a few days of temporary captivity, BNHS 3538 refused to feed on live earthworms provided. When handled, none of the individuals in the type series attempted to bite. They showed an inclination to burrow in soil and in the hand. At the localities reported here, Rhinophis melanoleucus sp. nov. occurs broadly sympatrically (within a radius of ca. 15 km) with other uropeltids including at least R. sanguineus, Uropeltis cf. nilgherriensis, Teretrurus hewstoni, Melanophidium bilineatum and M. wynaudense. Aengals & Ganesh (2013) provided a key to the identification of Indian species of Rhinophis. We provide a new key here that incorporates the new species and that takes into account that R. fergusonianus has 15 rather than 17 dorsal scale rows at midbody.

Discussion
Rhinophis fergusonianus has long been reported to have 17 dorsal scale rows at midbody (e.g., Boulenger 1896, Smith 1943, Pyron et al. 2016) but the holotype (and only known specimen) has been re-examined during this study and it has 15 rows from approximately (slightly after) midbody up to close to the vent (Appendix 2). At least some specimens of R. melanoleucus sp. nov. could be argued to have 17 midbody dorsal scale rows because the transition to 15 rows occurs approximately at midbody, but R. melanoleucus sp. nov. is more like R. sanguineus and R. fergu-sonianus in that all specimens have 15 rows by shortly behind midbody versus the condition in most other Rhinophis spp., which have 17 rows from anterior to midbody up to (close to) the vent. Understanding of the taxonomy and distribution of Indian Rhinophis is somewhat hampered by the minimal or very small sample sizes available for R. fergusonianus and R. goweri, and by lack of verification of some historical locality and scalation data. For example, Wall (1919) reported a collection of 40 R. sanguineus from Rockwood Estate in Wayanad with ventral scale numbers ranging from 214-218 in 16 females and 200-213 in 19 males, but the identification and scalation counts for these specimens (including more pronounced sexual dimorphism and higher ventral counts than we or others have recorded for R. sanguineus) cannot be verified because the whereabouts of these specimens is unknown. In another example, Hutton (1949) reported three R. sanguineus from High Wavy (Meghamalai), but no specimens were found here by Chandramouli & Ganesh (2010) and this locality is ca. 250 km south and on the other side of the Palghat Gap from all verified records of the species; again, the whereabouts of Hutton's specimens is unknown.