Rhinophis fergusonianus Boulenger, 1896

Gower, David J., Narayanan, Surya, Deepak, V., Anvar, Muhammed & Das, Sandeep, 2023, Rediscovery and systematics of Rhinophis fergusonianus Boulenger, 1896 (Serpentes: Uropeltidae) from India’s Western Ghats, Zootaxa 5374 (1), pp. 74-92 : 76-86

Rhinophis fergusonianus Boulenger, 1896

Figs. 1–9 , Tables 1–2 , Appendix 1

Rhinophis sp. (Attappadi): Sampaio et al. (2023)

Holotype ( Figs. 2 , 3 ; Table 1 ). BMNH 1946.1.16.77 (formerly BMNH 95.7.29.1), female. Collected from Cardamom Hills by J.S. Sealy ( Boulenger 1896), sometime between 1888 and 18 July 1895 ( Ferguson 1895).

Referred specimens (n = 6) BNHS 3687 (field tag: MW1609 ), female. Collected from between Mannarkad and Attapaddi, west of the Annamooli check post, Palakkad district, Kerala, India (11.03° N, 76.513° E; elevation ca. 130 m: Fig. 2 ) by Oommen V. Oommen and field assistants on 22 August, 2000 . BNHS 3688 (field tag: MW1476 ), female, same collection data as holotype; specimen lacks head and anteriormost end of body; DNA-sequence data for this specimen were reported by Sampaio et al. (2023: reported as tissue sample MW1477 ) as “ Rhinophis sp. Attapaddi ” . BNHS 3689 and BNHS 3690 , (both female) collected from Chathalloor village , Mallapuram district, Kerala, India (11.235° N, 76.120° E; 62 m asl) by Sandeep Das and Muhammed Anvar on 28 August, 2022 . BNHS 3755 , a fresh roadkill male collected near the Aanamooli checkpost, Mannarkad district, Kerala, India (11.032° N, 76.522° E; ca. 140 m) by Siddharth Sasidharan and Muhammed Anvar on 6 October 2022 . BNHS 3691 , female collected from Thattekad , Ernakulam district, Kerala (10.129°N, 76.687°E; 66 m) by Sandeep Das and Muhammed Anvar on 4 December, 2022 .

Revised diagnosis. Rhinophis fergusonianus differs from all other species of Rhinophis except R. karinthandani , R. melanoleucus , and R. sanguineus Beddome , 186 3 in having 15 dorsal scale rows at (or just behind) midbody (versus 17 or 19 in other congeners). Differs from R. melanoleucus in having fewer ventrals (<198 versus> 217). Differs from R. sanguineus in lacking a vivid red venter in life, and in having more and larger dark patches on the venter. Differs from R. karinthandani in tending to have fewer ventrals (183–197 versus 189–205), in tending to have more subcaudals in females (5, 6 or 7 versus 5) and perhaps also in males (8 or 9 in only known male specimen versus 6–8), and also in tending to have larger dark blotches on the venter in irregular, staggered, transverse half-bands (forming an irregular zigzag) versus smaller dark spots. Some specimens of R. karinthandani and R. fergusonianus are superficially similar, though the former is known only from higher elevation localities (ca. 800–900 m) on the Wayanad Plateau, and the latter from lower elevations (<150 m) on the edges of the Palghat Gap ( Fig. 1 ).

Rhinophis fergusonianus also differs substantially from congeners in mitochondrial and nuclear DNA sequences. For example, R. fergusonianus differs from its closest known relatives ( Sampaio et al. 2023)—those Indian species with 15 dorsal scale rows at midbody—by uncorrected p-distances of> 4% (12s),> 2% (16s),> 10 % (nd4) and> 1 % (prlr) for the aligned sequences ( Table 2 ).

Redescription of holotype ( Fig. 2 ). See Table 1 for morphometric and meristic data. Specimen condition varied; anterior 110 mm reasonable, then increasingly soft with flaccid, somewhat macerated body anterior to firmer posteriormost ca. 12 mm of body and tail. Posteriormost left infralabial shield damaged. Outermost layer of scales missing for most of posterior half of specimen except for small, isolated patches and on tail. Two longitudinal incisions, at midbody and posteriorly; former 12.2 mm on right ventrolateral surface; latter 12.9 mm, slightly right of midventer, ca. 54 mm from tail tip, through which empty oviducts visible.

Head small, snout pointed. Rostral pointed, longer than wide, without dorsal crest; in lateral view with slightly convex ventral and dorsal margins; widest at level of anterior upper corner of first supralabials. Rostral many (>5) 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 shorter, wider than rostral. Nasals separated from each other by intervening rostral. External naris small, subcircular, slightly countersunk within small depression, located at anteroventral corner of nasal. Nasal contacts supralabials 1 and 2. Prefrontals in contact with each other along midline (left overlapping right) for longer distance than portion of rostral posterior to nasals. 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 small but visible, diameter slightly less than 0.25 times length of ocular, located near anteroventral corner of ocular (close to lower edge, closer to this edge than to anterior of ocular), bulging slightly from ocular surface. Paired parietals longer than wide, posteriorly broadly rounded, angle between posteromedial and posterolateral edges approximately 100–110°. Parietals in longer midline contact than between opposite prefrontals, left overlapping right. Each parietal contacts four scales other than head shields. No mental groove; mental wider than long, smaller than infralabials, contacting first infralabials but not first ventral; three pairs of infralabials, second infralabial slightly the largest, first and third subequal. First and second ventrals longer than wide, third approximately as long as wide, fourth and subsequent ventrals wider than long. Five maxillary and five mandibular teeth on each side.

Body subcylindrical. Body scales generally evenly sized on dorsum and along body except for those involved in dorsal scale-row reductions or level with the approximately 55 th ventral (see formula below).Midline ventral scales between mental and anal of even size though anteriormost ones gradually narrow. At midbody, exposed part of ventrals approximately 1.3 times wider than scales in first dorsal row. Ventrals 195. Dorsal scale rows 19 anteriorly, reducing to 17 by level with 36th ventral, then to 15 by the 103rd ventral and maintained thereafter until the vent. At level of 55th (left) and 57th (right) ventral the third dorsal-row scale is noticeably enlarged (“b” in square brackets in formula below) and the fourth-row scale notably small (“s”), without there being a change in numbers of dorsal scale rows:

4 + 5 (36) [b3, s4 (57)] - 4 (103) 19 --------------- 17 ------------------ 17 --------------- 15 4 + 5 (36) [b3, s4 (55)] 4 + 5 (93)

Dorsal scale rows (i.e., excluding subcaudals) 13 at level of first subcaudals. Head and body scales macroscopically smooth. Paired anal scales (right overlying left) considerably larger than posteriormost ventrals, subequal to largest subcaudals. Distal margin of each anal overlaps two small scales in addition to anteriormost subcaudals. Five subcaudals on each side, the last of which is undivided (unpaired). Low, short parallel ridges, towards posterior edges of some scales on tail, more so posteriorly, dorsally and laterally. Tail 'shield' large, forming tip of tail, visible from below and especially clearly from above, convex, approximately as long as wide in dorsal view, approximately as wide as depth of tail (at base of shield), approximately same length (in strict dorsal view) as distance between snout tip and back of parietals, base surrounded by 10 scales (including last subcaudal). Shield surface roughened, bearing narrow, discontinuous ridges (longer, more continuous stretches located anteriorly and laterally towards shield base), receding and somewhat converging towards tip. ‘Shield’ ridges evenly spaced, subparallel, approximately straight; shield apex without projections or ridges.

Plate in Boulenger (1896: reproduced here in Fig. 3 ) presents a good, reasonably accurate impression of the colour pattern. Upper surface of head same brown colour as anterior of body, except for slightly orange tip of rostral and off-white (slightly yellow-orange tinted pale beige) patches on parts of supralabials; almost all of last supralabial pale, first supralabial pale only along lip edge. Underside of head mostly pale brown, except most notably for off-white parts of infralabials (increasingly so posteriorly).

Dorsal surface (7–9 dorsalmost scale rows) of body brown with isolated small pale flecks, slightly paler and greyer brown towards anterior end of body; scales darker brown at bases. Ventral surface of body pale, off-white. Laterally and ventrolaterally with largely punctate darker (brownish) smaller and larger spots (generally smaller than a single scale in size), more regular and posteriorly. Midventrally with more or less complete brownish markings, paler and slightly greyer anteriorly, without any entirely immaculately pale ventral shields. Midventral darker markings more irregular on ca. 55 mm anteriomost part of specimen, where restricted largely to ventrals and first dorsal scale row; on subsequent ca. 60 mm darker markings present as acutely angled zigzag extending to second dorsal scale row; on posterior half of body zigzag wider, less complete, with small breaks.

Anal shields pale with brown speckles and blotches, more so on left. Subcaudals brown except for pale yellow-orange half of last subcaudal. Lateral surface of tail off-white, dorsal surface of tail dark brown, with irregular boundary between these two regions. Tail shield mostly dark brown; faded yellow(-orange) at base; irregular, inconspicuous, narrow pale markings at apex.

Variation among newly referred material. Variation in some meristic and morphometric characters is reported in Table 1 . Specimens are illustrated in Figs. 4–8 . BNHS 3687 and 3688 are notably smaller than the holotype, and BNHS 3689, 3690 and 3755 substantially larger in both girth and length. BNHS 3688 is incomplete, and BNHS 3755 is a damaged roadkill specimen. Among the new material, teeth were counted only in BNHS 3687, and counts for this specimen were the same as in the holotype of R. fergusonianus . Beyond quantitative differences in external morphology reported in Table 1 , the new specimens differ slightly from each other and from the holotype of R. fergusonianus in some aspects of scalation and colour pattern.

The third ventral is much wider than long in BNHS 3690. As in the holotype, the reduction from 17 to 15 dorsal scale rows occurs close to midbody in BNHS 3691 (197 ventrals):

4+5 (37), +5 (38), 3+4 (42), +4 (56), 3+4 (59) 3+4 (88), +3 (93), 3+4 (97)

19------------------------------------------------------------17----------------------------------------15

4+5 (34), [b3, s4 (56)] 3+4 (85)

In contrast, the reduction from 17 to 15 dorsal scale rows occurs anterior to the 75 th ventral in three specimens from localities on the northern edge of the Palghat (= Palakkad) Gap, as follows:

BNHS 3687 (186 ventrals)

4 + 5 (29) [b4, s5 (51)] 4 + 5 (69)

19 --------------- 17 ---------------------------- 17 --------------------------- 15

4 + 5 (30) 3 + 4 (50), +4 (51) - 4 or 5 or 6 (66–71)

BNHS 3689 (194 ventrals)

3+4 (31) 3+4 (73)

19------------------17-------------------15

4+5 (34) 4+5 (71)

BNHS 3690 (192 ventrals)

3+4 (36) 3+4 (67), +4 (68), 3+4 (69)

19----------------17--------------------------------------15

4+5 (35) 3+4 (68)

BNHS 3688 lacks its head and anterior end of the body. The preserved portion bears 156 ventrals so the missing portion likely was the length of approximately 30 or so ventrals. This specimen has 15 dorsal scale rows from midbody up to the vent, and a slightly different scale reduction anterior of midbody (where X = the anteriormost ventral on preserved part of specimen):

4 + 5 (X+16), +4 (X+17) 3 + 4 (X+40)

17 (X) -------------------------------- 17 ------------------ 15 ------------------------------------------------------------- 15

4 + 5 (X+12), +4 (X+13) 4 + 5 (X+33) +4 (X+36), 4 + 5 (X+38), +4 (X+41), 3 + 4 (X+42) The damaged roadkill BNHS 3755 ( Figs. 3 & 4 ) is the only known male specimen. Scalation data for this specimen indicates some degree of sexual dimorphism given that it has the fewest number of ventrals and highest number of subcaudals among the seven known specimens. This specimen also differs from the females in having more prominent (but still low), short parallel ridges on the subcaudals and nearby scales. BNHS 3755 is distinct in also having a darker venter than in the other known specimens. The sample size is very small, but there are hints of sexual dimorphism in numbers of ventrals (186–197 in the five females, 183 in the single male) as well as subcaudals (5–7 per side in females, 8–9 in the male). There is also perhaps a hint of geographic variation among the females (in addition to the scale-reduction character reported above), with the two specimens from South of the Palghat Gap having more ventrals (195, 197 versus 186–194) and fewer subcaudals (5 versus 6 or 7 per side) than the four and a half specimens from the northern edge of the Palghat Gap.

In colouration, the ventral surface of the body of BNHS 3687 has a more extensive anterior whitish region and is more spotted than blotched behind this. BNHS 3689 also has a somewhat more spotted than blotched venter. Dark spots on the venter are much denser in BNHS 3690, with much darker blotches towards the posterior end of the body. In both BNHS 3689 and 3690 the underside of the tail is more whitish with a slightly thicker dark line medially, extending from the vent to the base of the shield which is interrupted by a white patch just above the base. The amount and distribution of orange on the tail shield varies somewhat among the series.

A summary of DNA-sequence differences is presented in Table 2 , and a phylogenetic tree is presented in Fig. 9 . The newly generated 12s and 16s sequences for BNHS 3689 are similar to those previously published ( Sampaio et al. 2023) for BNHS 3688 (as “ Rhinophis sp. Attappadi”). The specimen from Thattekad, South of the Palghat Gap (BNHS 3691) is slightly more distinct, but differs from BNHS 3688 only in six, nine and one sites for the aligned sequences of 16s, nd4 and prlr, respectively. Thus, the available DNA-sequence data are consistent with the hypothesis that these specimens from three localities from both sides of the Palghat Gap are conspecific.

Colour in life. Most detailed observations were made for BNHS 3689. Rostrum paler and more orange-brown towards the tip. Infralabials and chin shields mostly pale orange. Dorsal surface of body brown with very few small, isolated, pale flecks mostly on the anterior half of the body; scales darker brown at bases. Ventral surface of body pale off-white anteriorly to the level of approximately the 60th ventral and posterior to the 170th ventral, between which the ventrals bear small, irregular dark spots mostly present towards the anterior edge of the scales. Irregular dark spots largely present only on the scales adjacent to the ventrals making a longitudinal line. Anal shields off-white with irregular dark spots. Underside of tail orange with a longitudinal dark line in the middle, extending from the base of the anal shield to the base of the tail where it is broader. Tail shield mostly dark, with a wide orange stripe on the edge of the base of the upper surface and larger, semicircular orange patches anterolaterally towards the shield base on each side.

Distribution, habitat, and conservation status. The newly documented specimens are from three localities. All lie below an elevation of 150 m, and are from both sides of the Palghat Gap ( Fig. 1 ), a known biogeographic barrier, especially for moist-habitat and higher-elevation animals (e.g., Gower et al. 2007; Vijayakumar et al. 2016). Perhaps Rhinophis fergusonianus has greater tolerance and/or preference for drier environments than its closest known relatives from the Wayanad Plateau and Nilgiri Hills, making the Palghat Gap a relatively ineffective barrier to this species. The upper elevational limit of R. fergusonianus is unclear but is 140 m based on the recent observations reported here. The imprecise type locality of the Cardamon Hills was seemingly considered to lie above 600 m (Meyer et al. 1908–1931), but the veracity of that locality for R. fergusonianus should probably be considered doubtful unless extant, higher-elevation populations can be found.

BNHS 3687 and BNHS 3688 were found during a very short period (<30 minutes) of digging in marshy soil below herbaceous ground cover on a steeply sloping roadside verge, in an area of low-intensity agriculture with sparse cashew trees. BNHS 3755 was found dead on the road. BNHS 3687 and 3688 were dug from marshy soil below herbaceous ground cover on a steeply sloping roadside verge, in an area of low-intensity agriculture with sparse cashew trees. BNHS 3689 and 3690 were found moving on the surface at 09:00 and 13:00 hrs in a mixed agricultural plot of coconut, arecanut and plantain. BNHS 3691 was found close to a teak plantation at 08:35, moving in leaf litter following light rains.

Rhinophis fergusonianus clearly tolerates at least low-intensity agriculture and human disturbance, and is known from multiple locations up to approximately 140 km apart. However, road traffic is clearly a threat at some locations, only three or four extant populations are known with certainty, and population sizes and trends are unknown. It seems unlikely that the species is particularly widespread and/or abundant at mid-elevations in the southern Western Ghats, areas that are generally better sampled for uropeltid snakes. Given the small amount of information currently available, we believe that R. fergusonianus currently qualifies as Data Deficient under IUCN Red List criteria ( IUCN Species Survival Commission 2012), though expect that a more informative assessment could soon be made now that the species is better characterised and its geographical distribution somewhat clarified.

Figures

FIGURE 1. Map showing localities for Indian species of the uropeltid snake genus Rhinophis in southwestern peninsular India. Smaller-scale map shows type localities for the six currently recognised Indian species; though note that the type locality for R. fergusonianus is the vague “Cardamom Hills”, delineated by the red frame based on Meyer et al. (1908–1931). The larger-scale, inset map shows verified localities for the four species occurring in the Nilgiri-Wayanad and Anamalai regions of the Western Ghats (see also Sampaio et al. 2020).

FIGURE 2. Holotype (BMNH 1946.1.16.77, formerly 95.7.29.1) and previously only known specimen of Rhinophis fergusonianus. Scale bar increments 1 mm. Photographs by H. Taylor.

FIGURE 3. Illustration of Rhinophis fergusonianus reproduced from Boulenger’s (1896) original description of the species.

FIGURE 4. BNHS 3687, one of the newly referred specimens of Rhinophis fergusonianus from the northern edge of the Palghat Gap. Scale bar increments 1 mm. Photographs by H. Taylor.

FIGURE 5. Three of the newly referred specimens of Rhinophis fergusonianus from localities on the northern edge of the Palghat Gap in approximately dorsal (left column) and ventral (right) views: A & B, BNHS 3689; C & D, BNHS 3690; E &F, BNHS 3755. Scale bars 10 mm. Photographs by S. Narayanan.

FIGURE 6. Three of the newly referred specimens of Rhinophis fergusonianus from localities on the northern edge of the Palghat Gap: A, BNHS 3690; B, BNHS 3689; C, BNHS 3755. From left to right, heads in dorsal, ventral, right lateral and left lateral views; tails in dorsal, ventral, left lateral and right lateral views. Scale bars 5 mm. Photographs by S. Narayanan.

FIGURE 7. BNHS 3691, a newly referred specimen of Rhinophis fergusonianus from South of the Palghat Gap, in dorsal (upper) and ventral (lower) views. Total length of specimen is ca.440 mm. Photographs by S. Das.

FIGURE 8. Rhinophis fergusonianus in life. Left, BNHS 3689 (ca. 420 mm total length); right, BNHS 3691 (ca. 440 mm). Photographs by S. Das.

FIGURE 9. Maximum Likelihood (ML) phylogeny for Rhinophis. Numbers on internal branches are Bayesian posterior probability/ML bootstrap support values. See Appendix 1 for data sources. Based on analysis of data available at https://data. nhm.ac.uk/dataset/gower_fergusonianus (https://doi.org/10.5519/95s2p6kx).