A new species of Rhinophis Hemprich, 1820 (Serpentes: Uropeltidae) from the central hills of Sri Lanka Author Mendis Wickramasinghe, L. J. Author Vidanapathirana, Dulan Ranga Author Gehan Rajeev, M. D. Author Gower, David J. text Zootaxa 2017 4263 1 153 164 journal article 33095 10.11646/zootaxa.4263.1.7 707b0d75-86cc-4ddf-a776-608137fc4745 1175-5326 572597 7CE2F6FE-792B-4997-9714-BD02BF230406 Rhinophis roshanpererai sp. nov. Figs. 1–5 ; Table 1 Holotype . NMSL 2016.08 .0 1 NH ( Figs. 2–3 ; Table 1 ), adult male, 207.9 mm SVL, Galkanda , Beragala , Badulla District , Uva Province , Sri Lanka ( 6° 45’ 07.98” N , 80° 57’ 20.23” E , elevation 940 m ). Collected by L.J. Mendis Wickramasinghe , Dulan Ranga Vidanapathirana , and M. D. Gehan Rajeev , 10 May 2010 . FIGURE 2. Dorsolateral view in life of the holotype of Rhinophis roshanpererai sp. nov. , NMSL 2016.08.0 1 NH, 214 mm (total length). Paratypes. DWC 2016.05.0 3, adult female, 205.2 mm SVL ( Fig. 4 A); DWC 2016.05 .0 4, adult female, 218.2 mm SVL ( Fig. 4 B–5). Collection data as for holotype. Diagnosis. A Rhinophis restricted to the Central Highlands of Sri Lanka with 17 dorsal scale rows at midbody, more than 160 and fewer than 175 ventral scales, a small tail shield with spines, three or four of which prominent, and lacking yellowish markings laterally or dorsally. Identification. The new uropeltid species is referred to Rhinophis because it has an eye that lies within an ocular scale (not so in Platyplectrurus Günther, 1868 ), has a clearly discrete tail shield, lacks a mental groove (present in Melanophidium Günther, 1864 ), lacks supra- or postoculars or temporals (at least one of which is present in Brachyophidium Wall, 1921 , Platyplectrurus, Plectrurus Duméril, 1851 , and Teretrurus Beddome, 1886 ), lacks midline contact between the nasals (present in Brachyophidium, Melanophidium, Platyplectrurus, Plectrurus, Pseudoplectrurus Boulenger, 1890 , Teretrurus, and Uropeltis ), and it has midbody dorsal scales in 17 rows ( 15 in Brachyophidium , Melanophidium , Platyplectrurus , Plectrurus, Pseudoplectrurus, Teretrurus ). Rhinophis roshanpererai sp. nov. differs from all four Indian species of Rhinophis by having a very small tail shield with spines (versus relatively much larger tail shield without spines). It differs further in having a ventral count of 168 or 169 (versus more than 200 in R. goweri Aengals and Ganesh, 2013 , fewer than 150 in R. travancoricus Boulenger, 1893 , and more than 170 in R. fergusonianus Boulenger, 1896 ); and in having 17 midbody dorsal scale rows (versus 15 in R. sanguineus Beddome, 1863 ). TABLE 1. Meristic and morphometric data for 52 characters of holotype (*) and paratypes (**) of Rhinophis roshanpererai sp. nov. Specimen numbers have NMSL prefix for holotype and DWC prefix for paratypes. Bilateral scale counts given in left, right order. Character numbers in parentheses, as follows: (1) supralabials, (2) infralabials, (3) diameter of eye, (4) maximum length of ocular, (5) length ocular/eye, (6) distance between eyes, (7) distance between eye-naris, (8) length of frontal, (9) width of frontal, (10) head length = distance between snout tip-posterior edge 4th supralabial, (11) total length/HL, (12) head width level with corner of mouth, (13) distance between nares, (14) maximum longitudinal length of prefrontal scales, (15) length of parietal scale = distance between posterior tip of parietal and posterior end of suture between frontal-ocular, (16) width of parietal scale = distance between posterior tip of ocular and posterior tip of frontal, (17) length of midline suture between parietals, (18) distance rostral-frontal, (19) distance between tip of snout and posterior edge of rostral, (20) distance between tip of snout and posterior limit of midline suture between parietals, (21) maximum width of rostral, viewed ventrally, (22) distance between snout tip-naris, (23) distance between snout tip-eye, (24) midbody width, (25) midbody depth, (26) midbody circumference, (27) exposed width of first scale row at midbody, (28) exposed width of ventral scales at midbody, (29) number of ventral scale level with which dorsal scale row is ‘lost’ (i.e. location of change between 19 and 17 dorsal scale rows), (30) dorsal scale rows behind head, at midbody and one head length anterior to vent (excluding ventrals), (31) exposed length of first scale row at midbody, (32) width of ventral/first dorsal scale rows, (33) ventral scales, (34) vertebral scales, (35) snout– vent length, (36) width at vent, (37) disposition of two anal scales (l/r) or (r/l), (38) number of small scales overlapped by anals, excluding first subcaudals, (39) subcaudal scales, (40) tail length, (41) total length, (42) tail length as % of total length, (43) total length/midbody width, (44) maximum width of tail shield, (45) maximum diameter of tail shield, (46) number of scales (including subcaudals) surrounding base of tail shield, (47) dorsal base to tip of tail shield (upper spines), (48) dorsal base to tip of tail shield (lower spines), (49) lateral base to tip of tail shield (upper), (50) lateral base to tip of tail shield (lower), (51) ventral base to tip of tail shield (upper), (52) ventral base to tip of tail shield (lower).

2016.08.0 1 NH*	2016.05.0 3 **	2016.05.0 4 **	2016.08.0 1 NH*	2016.05.0 3 **	2016.05.0 4 **
(1)	4	4	4	(27)	1.4	1.5	1.6
(2)	3	3	3	(28)	2.2	1.9	2.0
(3)	0.5	0.5	0.6	(29)	30, 30	30, 30	30, 30
(4)	1.7	1.7	1.8	(30)	19, 17, 17	19, 17, 17	19, 17, 17
(5)	3.2	3.2	3.3	(31)	1.3	1.1	1.4
(6)	2.9	3.0	2.8	(32)	1.5	1.3	1.3
(7)	2.1	2.1	2.2	(33)	169	168	169
(8)	2.1	2.1	2.3	(34)	176	175	174
(9)	1.6	1.6	1.8	(35)	207.9	205.2	218.2
(10)	6.2	6.0	6.3	(36)	4.0	4.1	4.2
(11)	34.8	35.3	35.6	(37)	r/l	r/l	r/l
(12)	4.4	4.4	4.6	(38)	3, 3	3, 3	3, 3
(13)	1.4	1.4	1.5	(39)	7, 7	7, 7	7, 6
(14)	1.6	1.5	1.6	(40)	6.1	5.8	5.8
(15)	2.2	2.1	2.3	(41)	214.0	211.0	224.0
(16)	1.8	1.8	2.1	(42)	2.8	2.8	2.6
(17)	0.5	0.4	0.4	(43)	30.0	29.8	30.9
(18)	0.6	0.4	0.6	(44)	1.6	1.4	1.6
(19)	2.0	2.1	2.3	(45)	2.3	2.1	2.4
(20)	5.4	5.2	5.7	(46)	8	9	8
(21)	1.3	1.1	1.3	(47)	1.4	1.5	1.3
(22)	1.1	1.1	1.1	(48)	2.0	2.2	2.1
(23)	3.3	3.1	3.3	(49)	0.8	0.8	0.9
(24)	7.1	7.1	7.2	(50)	0.9	1.0	1.0
(25)	6.6	6.4	6.2	(51)	1.0	1.2	1.3
(26)	22.2	24.0	25.4	(52)	0.5	0.4	0.5

FIGURE 3. Holotype of Rhinophis roshanpererai sp. nov. , NMSL 2016.08.0 1 NH, scale bars in centimeters. FIGURE 4. Head and tail ends of the two paratypes of Rhinophis roshanpererai sp. nov. (A) DWC 2016.05.0 3, (B) DWC 2016.05.0 4. Scale bar 2 cm. FIGURE 5. Rhinophis roshanpererai sp. nov. , paratype DWC 2016.05.0 4, posterodorsal view of tail in preserved specimen showing three (two lower, one upper) rather than four major spines on the tail shield (bases of these major spines are surrounded by pale areas on shield). Among Sri Lankan congeners, Rhinophis roshanpererai sp. nov. differs from R. saffragamus ( Kelaart, 1853 ) in not having a large and flat tail shield or midline contact between the opposite nasal shields, and by having dorsal scales in 17 rather than 19 rows at midbody. The new species differs from R. dorsimaculatus Deraniyagala, 1941 , R. homolepis (Hemprich, 1820) , R. lineatus , R. oxyrynchus (Schneider, 1801), R. porrectus Wall, 1921 , R. punctatus Müller, 1832 and R. zigzag by having fewer than 175 ventral scales (versus more than 180), and by having a very small tail shield with spines (versus relatively much larger tail shield without spines). Rhinophis roshanpererai sp. nov. differs from R. phillipsi ( Nicholls, 1929 ) in having fewer than 190 ventrals and in lacking yellow lines on the dorsum. Rhinophis roshanpererai sp. nov. resembles R. melanogaster in having a small tail shield, but the new species has a shield surface with four (or three) notably prominent spines, one pair above the other (versus two slightly larger spines ventrally); absence of yellowish lines laterally (versus present); perhaps more ventral scales (168–169 versus 152–166); and a distinct geographical distribution (central highlands of Badulla District vs Knuckles Range, Matale and Kandy Districts). The ventral scale count in R . roshanpererai sp. nov. is similar to or overlapping with those for R. blythii Kelaart 1853 , R. drummondhayi Wall, 1921 , R. philippinus (Cuvier, 1829) , and R. tricolorata Deraniyagala, 1975 , but the new species differs from these four Sri Lankan congeners by having a smaller tail shield with spines, three or four of which are prominent (versus large tail shield without notable spines). The new species differs from R. erangaviraji by having more than 165 ventrals (versus fewer than 155), a smaller tail shield, and by lacking substantial yellow areas on the lateral surface of the body and tail. Description of holotype . See Table 1 for morphometric and meristic data. A preserved specimen in good condition; 20 mm long left of ventral incision into coelom extending anteriorly from 10 mm anterior to vent; outer layer of scales loose and missing in parts; a few flank scales more profoundly damaged on left at approximately midbody. Head small, snout pointed ( Figs. 2–3 ). Rostral pointed, longer than wide, without dorsal crest; widest at level of anterior superior corner of first supralabials. Rostral several times longer (in dorsal view) than rostralfrontal gap ( Fig. 3 ). Frontal irregularly hexagonal, longer than wide, lateral (ocular) margins slightly converging posteriorly, posterolateral margins straight to very slightly concave; lateral (ocular) margin shortest, posterolateral edges longest. Frontal longer, wider than rostral. A pair of nasals, separated from each other by posterior half of rostral. External naris small, subcircular, slightly countersunk within small depression, located in anteroventral corner of nasal. Nasal in contact with first and second supralabials. Prefrontals (for most of their length) in contact with each other along midline (left overlapping right), separating frontal from rostral. Prefrontals wider than long, shorter than frontal. Supralabials four, first smallest, making the least contribution to margin of mouth; fourth much the largest. Ocular in contact with third and fourth supralabials. Eye distinct, diameter approximately one third length of ocular, located near anteroventral corner of ocular, bulging slightly from ocular surface, pupil circular. Paired parietals longer than wide, shorter, very slightly wider than frontal, posteriorly broadly rounded, angle between postermedial and posterolateral edges approximately 90°. Opposite parietals in brief midline contact, 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 and single postmental (= 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. Six or seven maxillary and approximately seven mandibular teeth on each side. Teeth simple, pointed, distinctly retrorse, straight, evenly spaced. Body cylindrical. 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. Ventrals 168, posteriormost ventral notably smaller, penultimate ventral paired. Dorsal scale rows 19 anteriorly, reducing to 17 by level with 30th ventral and maintained along most of body; scale row reduction formula: 3 + 4 (30) 19 --------------- 17 3 + 4 (30) Dorsal scale rows approximately 14 at base of tail. Head and body scales macroscopically smooth, lacking keels. Inconspicuous keels on scales on posteriormost portion of body and on tail, increasingly prominent posteriorly, more obvious ventrally (including on anals) and ventrolaterally. Paired anal scales (right overlying left) considerably larger than posteriormost ventrals and subcaudals. Distal margin of each anal overlaps three other scales in addition to anteriormost subcaudals. Seven right and seven left subcaudals. Tail 'shield' mildly conical, forming tip of tail, small, longer than wide in dorsal view, shorter than the frontal in dorsal view, visible from below and especially above, base (much narrower than base of tail) surrounded by last pair of subcaudals and 6 other scales. In posterior view shield oval to slightly egg-shaped, wider ventrally than dorsally ( Fig. 3 ). Shield surface sparsely spinose, most spines small, inconspicuous but four (arranged in two pairs, one above the other) much longer and substantial, pointing straight backwards; ventral pair of larger spines notably longer than dorsal pair ( Fig. 3 ). Colour in life. Dorsum and lateral background uniform black, with sparse, very small yellow flecks ( Fig. 2 ). Ventral background dark brown for most of length, gradually paler anteriorly, darker posteriorly (similar to dark colour of dorsum). Anterior and underside of snout paler than rest of head. Venter with conspicuous yellow blotching, blotches notably larger than on dorsum and lateral surfaces of body; ventral blotching absent on tail, head and anteriormost and posteriormost of body. Colour in alcohol. Colour pattern remains with a little fading, black to dark brown, yellow to off white and brown to a paler brown ( Fig. 3 ). Paratypes and variation. Paratype DWC 2016.05.0 3 is slightly longer ( 218.2 mm SVL) than the holotype and the other paratype (DWC 2016.05.04) slightly shorter ( 205.2 mm SVL), both are female. The two paratypes are very similar to the holotype with respect to the description presented above, including identical scale row reductions (19 to 17 rows by level of 30th ventral). DWC 2016.05.0 3 differs from the holotype in having: parietals more notably wider than frontal ( Fig. 4 A); seven rather than six scales plus last pair of subcaudals surround base of tail shield; posteriormost ventral paired; supernumerary scale between second pair of subcaudals ( Fig. 4 A). DWC 2016.05.0 4 differs from the holotype in having six rather than seven subcaudals on right side, and in having three rather than four major spines on the tail shield, two posteroventrally and one posterodorsally ( Fig. 5 ). Both paratypes appear to have seven maxillary teeth on each side; mandibular counts are more difficult but are estimated at six or seven on each ramus. Paratypes closely resemble holotype in colour pattern. Etymology. The species epithet roshanpererai is named for the late Roshan Perera, who was an Instructor of the Reptiles group of the Young Zoologist’s Association of Sri Lanka , Department of National Zoological Gardens, in recognition of his dedicated services to wildlife conservation in Sri Lanka . The species name roshanpererai is a noun in the genitive case. Suggested vernacular names. Roshan Pererage thudulla, Roshan Pereravin nilakael pambu, Roshan Perera’s sheildtail (or Roshan Perera’s Rhinophis ) in Sinhala, Tamil, and English, respectively. Distribution, habitat and threats. The first author first encountered the new species as a single roadkill specimen at the type locality in 1999. In five or six subsequent visits to the type locality approximately 30 individuals of Rhinophis roshanpererai sp. nov. have been observed, including a second roadkill specimen. The type series of R. roshanpererai sp. nov. was found within a 1 m radius, dug during the day from soil ca. 150 mm deep among banana plants in a home garden. Other specimens have been seen at or close to (within a couple of kilometers) of this site in a wide range of habitats, including shaded patches of grassland, tea plantations, and disturbed riverine forest, always dug from soil or leaf litter during the day. A few specimens have been seen moving on the surface, only at night. Several other individuals of the new species were dug from soil in disturbed riverine forest in 1999 from Uda Diyaluma, approximately 10 km away ( 6° 44’ 08.55” N , 81° 01’ 57.10” E , elevation 750 m ), and from Haldummulla, approximately 6 km from the type locality ( 6° 45’ 39.95” N , 80° 54’ 05.73” E , elevation 938 m ). Despite a substantial amount of fieldwork (including digging through soil and leaf litter) at similar altitudes, we have not observed this species outside this region, including at, for example, Haputale, less than 2 km North of the type locality but approximately 400 m higher in elevation. Rhinophis roshanpererai sp. nov. has not been found in sympatry with other uropeltid species. The nearest observation of other species that we know of (L.J.M.W., pers. obs.) is for R. drummondhayi at 960 m elevation at Kubalwela, approximately 16 km to the northeast by north (bearing of 30°) of the type locality of R. roshanpererai sp. nov. . We suspect that the vertical and horizontal distributional range of the new species is small, and that substantial human disturbance in the form of intensive agriculture and urbanization represent the likely greatest conservation threats.