Two new species of the shieldtail snake genus Rhinophis Hemprich, 1820 (Serpentes: Uropeltidae), from the Rakwana and the Knuckles Massifs of Sri Lanka Author Gower, David J. Department of Life Sciences, The Natural History Museum, London SW 7 5 BD, UK Author Sampaio, Filipa L. Department of Life Sciences, The Natural History Museum, London SW 7 5 BD, UK & Department of Genetics, Evolution and Environment, University College London, London WC 1 E 6 BT, UK Author Vidanapathirana, Dulan Ranga Herpetological Foundation of Sri Lanka, 31 / 5, Alwis Town, Hendala, Wattala, Sri Lanka Author Mendis Wickramasinghe, L. J. Herpetological Foundation of Sri Lanka, 31 / 5, Alwis Town, Hendala, Wattala, Sri Lanka text Zootaxa 2024 2024-05-29 5458 3 332 360 http://dx.doi.org/10.11646/zootaxa.5458.3.2 journal article 298361 10.11646/zootaxa.5458.3.2 18d6f856-ced1-49b9-b460-5b7bdf073060 1175-5326 11547469 54609B08-07DE-4009-B15E-FEBE2C577AAE Rhinophis dinarzardae sp. nov. Figs. 7–12 ; Tables 4–6 ; Appendix 2 urn:lsid:zoobank.org:act: 7D4E0EF2-DED5-4440-B17B-0355D22B2830 Chresonymy Rhinophis philippinus (Cuvier, 1829) : Cadle et al. (1990 , in part); Gower (2003) ; Bossuyt et al. (2004) ; Gower & Ablett (2006) ; Douglas & Gower (2010) ; Pyron et al. (2016 , in part); Cyriac & Kodandaramaiah (2017) ; Jins et al. (2018) Rhinophis cf. philippinus : Sampaio et al. (2023 : in part) Holotype . USNM 548092 ( Fig. 7 ), adult male , Rattota, near Illukkumbura ( Fig. 8 ), Central Province , Sri Lanka . Rattota is a small town in Matale District ( 7° 31’ 14” N , 80° 40’ 47” E , ca. 400 m elevation according to GoogleEarth). Collected by Lalith Jayawickrama , 29 October 1979 . Paratypes (n = 7). USNM 548089 (female), 548090 (female), 548091 (female) and 548093 (male) ( Figs. 9 and 10 ), collection data as for holotype . CAS 225847 (female), 225857 (male) and 225856 (male), collected from Palatenne (Opalgalla Group), below Pride’s Gap, Central Province , Sri Lanka (ca. 7° 32’ 30” N , 80° 40’ 20” E , 500 m based on GoogleEarth data for “Pallethanna”), 1 May, 1976. Referred material (n = 64). Palatenne (Opalgalla Group) below Pride’s Gap, 620 m : CAS 225750, 225752, 225771, 225864, 225867, 225872, 225875, 225877, 225880, 225881, 225890, 225894, 225914, 225915, 225916, 225927, 225938, 225943, 225944, 226480, 226484, 226489, 226494, 226499, 226501, 226502, 226508, 226527, 226536, 226597, 244519, 244541, 244578 [ 33 specimens ]; Palatenne (Opalgalla Group) below Pride’s Gap, 515 m : CAS 225932 [1]; Palatenne (Opalgalla Group) below Pride’s Gap, below 420 m : CAS 226598 [1]; Palatenne: CAS 225940, LSUMZ 39255, 39256 [3]; Palatenne Lines, 618 m : CAS 226227 [1]; Palatenne, above 680 m : CAS 226229, 226232 [2]; Palatenne, below 600 m : CAS 226236, 226237, 226238, 226258, 226259 [5]; Bambaragala, near Midlands on Rattota-Illukkumbura road, 760 m : CAS 226267 [1]; Palatenne (Opalgalla Group), Hunuketa, Rattota, 480 m : CAS 226457, 226459, 226465 [3]; Gammaduwa Road, Hunuketa, Rattota, 480 m : CAS 226466 [1]; Gammaduwa Road, Gammaduwa high, 750 m : CAS 226539, 226541, 226543 [3]; Gammaduwa, below Church, 710 m : CAS 225861, 225862, 225863 [3]; Gammaduwa: Kalugaltenne, 5 miles below church on crest, 690 m : CAS 226596 [1]; Kandehana, near Rattota, Matale District, Central Province ( 7° 33.73’ N , 80° 40.97’ E ; 673 m ): NMSL MW1739, NMSL MW1741 [2]; Dombawela, near Palapatwela, Matale District, Central Province ( 7 33.08 N , 80 37.14 E , 418 m : NMSL MW1754 [2]; Kalugaltenne, Knuckles: NMSL WHT5155, WHT5157, WHT5158 [3]. Diagnosis. A Rhinophis restricted to the Knuckles Massif of Sri Lanka with 17 dorsal scale rows at midbody, more than 170 and fewer than 205 ventral scales ( 191–204 in females; 178–190 in males), and a colour pattern with uniform brown-grey dorsum and punctate venter (dark scale bases and pale peripheries), and a large, protuberant, domed tail shield with very small, homogenous spines. Rhinophis dinarzardae sp. nov. differs from all six Indian species of Rhinophis by having a ventral count of> 170 and <205 (versus fewer than 150 in R. travancoricus ; more than 210 in R. goweri ) and in having 17 dorsal scale rows at or just behind midbody (versus 15 in R. fergusonianus , R. sanguineus , R. melanoleucus , and R. karinthandani ). Among Sri Lankan congeners, Rhinophis dinarzardae sp. nov. differs from R. saffragamus in lacking a flat tail shield and 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. melanogaster , R. phillipsi , R. roshanpererai and R. gunasekarai by having homogenously sized spines on a regularly domed, wide tail shield (versus a less-regularly prominent tail shield narrower than the end of the tail and with heterogeneously sized spines). The new species differs from R. dorsimaculatus , R. porrectus , R. punctatus and R. zigzag by having fewer than 205 ventral scales (versus more than 209), and by having a rostral that is dorsally transversely rounded (versus with a longitudinal ridge) and less (versus more) pointed. Rhinophis dinarzardae sp. nov. differs from R. blythii and R. erangarviraji in having more than 170 (versus fewer than 170) ventral scales, and in having a more protuberant, conical tail shield (versus less-conical tail shield). The new species differs from R. martin sp. nov. in having more ventral scales ( 178–190 in males and 191–204 in females versus ca. 169–178 and 175–186, respectively) and in having a venter that is paler (versus darker) than the dorsum. FIGURE 7. Holotype of Rhinophis dinarzardae sp. nov. , USNM 548092. Upper row: head end in dorsal, right lateral, ventral and left lateral views; Central two rows whole specimen in ventral (upper) and dorsal (lower) views); Lower row: tail end in left lateral, ventral, right lateral, and dorsal views. Scale bar increments in millimetres. Photographs by H. Taylor. FIGURE 8. Map showing places relevant to the distribution of Rhinophis dinarzardae sp. nov. and R. philippinus . Inset map A shows localities for R. philippinus (white circles), type locality of R. dinarzardae sp. nov. (triangle), questionable locality for the new species (Handapangala: black polygon), and locality of R. cf. dinarzardae sp. nov. (Pelmadulla: black circle). Inset map B shows type (triangle) and other localities (white circles) for R. dinarzardae sp. nov. and two localities for R. cf. dinarzardae (Matalapitya, Illukkumbura: black circles). Abbreviations for upland regions: CH—Central Highlands, KN—Knuckles, V–W— Vicarton-Wiltshire Range. Black squares indicate selected towns. Rhinophis dinarzardae sp. nov. differs from R. mendisi by having more ventrals ( 178–190 in males and 191– 204 in females versus 159–169 and 170–177, respectively). The new species has fewer ventral scales (<205) than R. oxyrhynchus (214 and 222 in the two types ). Rhinophis dinarzardae sp. nov. differs from R. lineatus and R. drummondhayi in having a conical tail shield that at its base is almost as wide as the end of the tail (versus narrower and less conical) and in lacking (versus having) large pale blotches on the lateral surface of the tail. Rhinophis dinarzardae sp. nov. differs from R. homolepis in males having fewer ventrals (178–190 versus 192–196 for the specimens reported by Gower 2020 : table 3), in having a wider base to the tail shield, and in lacking pale transverse markings on the dorsolateral aspect of the body. The new species differs from R. tricoloratus by having a lessstrongly tapering head and tail shield in dorsal and ventral views, and in lacking pale transverse markings on the dorsolateral aspect of the body or tail. Rhinophis dinarzardae sp. nov. is superficially similar (including in overall colour pattern and form of the head and tail shield) to R. philippinus but the two species can be distinguished on the basis of the new species having more ventral shields in males (178–190, n = 33 versus 154–168, n = 11) and in females (191–204, n = 33 versus 166–179, n = 19). On average, R. dinarzardae sp. nov. males have more subcaudal scales than do R. philippinus males, typically 6–7 per side versus typically 5–6 per side, respectively ( Fig. 11 ). Rhinophis dinarzardae sp. nov. is also distinct from all named species of Rhinophis in DNA sequences ( Table 4 ; Sampaio et al. 2023 ). FIGURE 9. Four paratypes of Rhinophis dinarzardae sp. nov. Whole specimens in approximately dorsal (upper image of each pair) and ventral (lower) views. A. USNM 548089, B. 548090, C. 548091, D. 548093. Scale bars 20 mm. Photographs by H. Taylor. Description of holotype . See Table 5 for morphometric and meristic data and Fig. 7 for photographs. Fair condition; somewhat dehydrated, with artefactual longitudinal grooves and ridges; posterior end of body sharply downturned just anterior to tail; some body scales damaged and with soil attached. Head small, snout pointed. Rostral pointed, longer than wide, without dorsal crest; widest at level of anterior upper corner of first supralabials. Rostral many (> 4) times longer (in dorsal view) than rostral-frontal gap. Frontal subtriangular with three convex edges, anterior edge least convex, longer than wide, posterolateral edges longest. Frontal much shorter, wider than rostral. Paired nasals separated from each other by posterior of rostral. External naris small, subcircular, slightly countersunk within small depression, located in 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. Prefrontals wider than long, shorter than frontal. Supralabials four, first smallest, making the least contribution to margin of mouth; fourth much the largest. Ocular contacts supralabials 3 and 4. Eye distinct but small, diameter approximately 0.2 times length of ocular, located close to anteroventral edge of ocular, marginally anterior of halfway along, bulging slightly from ocular surface, pupil circular. Parietals about as long as wide, each similar to frontal in width and length, posteriorly broadly rounded, angle between postermedial and posterolateral edges approximately 90°. Opposite parietals in brief midline contact (longer than contact between opposite prefrontals), left overlapping right. Each parietal contacts four scales other than head shields. No mental groove; mental much wider than long, much smaller than infralabials, contacting first infralabials but not first ventral; three pairs of infralabials, second largest, third shortest and smallest on right, third and first equally small on left. First and second ventrals longer than wide, third approximately as long as wide, fourth and subsequent ventrals wider than long. 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 188. Dorsal scale rows 19 anteriorly, reducing to 17 by level with 37 th ventral and maintained along most of body except for slight fluctuation on right, level with approximately the 60 th ventral; no such fluctuation occurs on the left side in this position but here the fourth-row scale level with the 57 th ventral and the third-row scale level with the 58 th ventral are enlarged and the fourth-row scale level with the 58 th ventral is small. Scale row reduction formula: 4 + 5 (37), +4 (59), -4 (61) 19 ------------------------------------ 17 4 + 5 (32) Dorsal scale rows approximately 13 at base of tail. Head and body scales macroscopically smooth, almost all lacking keels. Very faint, low, short keels and tubercles towards posterior edges of lowermost couple of dorsal scale rows at far posterior of body, on last few ventrals, anals, small scales overlapped by anals, and on subcaudals and dorsal scales of tail. Paired anal scales (right overlying left) considerably larger than posteriormost ventrals and subcaudals. Distal margin of each anal overlaps two other scales in addition to anteriormost subcaudals. Six subcaudals on each side, second, third, fifth and sixth undivided. Tail ‘shield’ long, projecting, blunt-ended dome with much longer dorsal than ventral margin in lateral view. Shield approximately as wide as long and approximately same length as head in dorsal view, base surrounded by 12 scales (including last subcaudal). In posterior view, shield base subcircular. Shield surface bears small, low, laterally compressed spines arranged in narrowly spaced subparallel rows that converge towards shield apex; spines less prominent toward apex. TABLE 4. Uncorrected p-distances (%) for intra- and interspecific comparisons for Rhinophis dinarzardae sp. nov. and closest sampled relatives, for two nuclear ( prlr , cmos ) and four mitochondrial genes ( 12s , 16s , cytb , nd4 ) as analysed by Sampaio et al. (2023) . Values are means, with ranges in parentheses (where there is variation), and number of pairwise comparisons in square brackets. "–" indicates insufficient data for pairwise comparison. The single R. cf. dinarzardae specimen is “ Rhinophis cf. philippinus ” sample KNU017 reported in Sampaio et al. (2023) ; the ten R. dinarzardae sp. nov. specimens are all the other “ Rhinophis cf. philippinus ” samples reported in Sampaio et al. (2023) .

12s	16s	cytb	nd4	prlr	cmos
Within R.	0.2 (0–0.3) [3]	0 (0–0.1) [3]	–	0 [3]	–	0 [1]
dorsimaculatus
Within R. philippinus	0.5 [1]	0 [1]	0 [1]	0 [1]	–	–
Within R. dinarzardae	0 [45]	0 [45]	0 (0–0.1) [36]	0.3 (0–0.6) [36]	0 [36]	0.1 (0–0.3) [21]
sp. nov.
R. dorsimaculatus –R.	7.5 (7.2–7.7) [6]	1.6 (1.5–1.6) [6]	4.3 [2]	15.8 (15.4–16.4) [6]	1.1 [1]	1.1 [2]
philippinus
R. dorsimaculatus –R.	6.5 (6.1–6.8) [30]	1.5 (1.4–1.5) [30]	4.5 (4.4–4.5) [9]	16.8 (16–17.9) [27]	1.1 [9]	1.1 [14]
dinarzardae sp. nov
R. dorsimaculatus – R.	7.2 (7–7.4) [3]	1.4 (1.4–1.5) [3]	4.7 [1]	16.4 (15.9–16.7) [3]	1.1 [1]	1.4 [2]
cf. dinarzardae
R. philippinus –R.	1.4 (1.1–1.6) [20]	0.3 (0.3) [20]	1.1 [18]	4.1 (4–4.2) [18]	0 [9]	0 [7]
dinarzardae sp. nov.
R. philippinus – R. cf.	2.5 (2.5) [2]	0.3 (0.3) [2]	1.3 (1.2–1.4) [2]	4.9 [2]	0 [1]	0.2 [1]
dinarzardae
R. dinarzardae sp.	1.7 (1.6–1.9) [10]	0 (0) [10]	0.9 (0.8–1) [9]	3.6 (3.4–3.9) [9]	0 [9]	0.3 (0.2–0.5) [7]
nov.– R. cf. dinarzardae

Colour in alcohol. Darker above than below; upper surfaces dark, blackish brown; lower surfaces paler brown. Paler parts of scales on head and anterior of body pale tan (slightly orange-tan on head); paler parts increasingly off-white posteriorly. Brown colour on dorsal surface of head paler than dorsal surface of body. Head scales brown with pale tan free edges, larger pale patch on right ocular behind eye. Lower halves of supralabials pale tan; anterior half of rostral orange brown. Scales on underside of head (except infralabials and mental) with more extensive brown patches than on underside of anterior end of body; infralabials and mental almost entirely pale tan with few small, isolated darker brown flecks; inside of mouth pale. Dorsally, body scales mostly dark with pale laterodistal edges, pale edges decreasing in extent posteriorly. Ventrally, body highly punctate; anteriorly scales with small brown patch at base, increasing in size posteriorly but off-white edges remain even at far posterior end of body. Anal shields brown at base but mostly off-white, with pale transverse bar extending across left and right anal. Tail shield dark brown at base, otherwise paler brown, with pale yellowish patch on tip, extending along narrow irregular middorsal streak. FIGURE 10. Four paratypes of Rhinophis dinarzardae sp. nov. Head and tail ends in four views. Scale bars 10 mm. A. USNM 548089, B. 548090, C. 548091, D. 548093. Photographs by H. Taylor. Variation among paratypes . Condition generally good; some specimens preserved when close to ecdysis. See Table 5 for some meristic and morphometric data, and Figs. 9 and 10 for photographs of the USNM paratypes . In scalation patterns, all paratypes resemble the holotype except for minor differences. For example, the frontal can be more hexagonal rather than subtriangular (e.g., CAS 225847, 225856 and 225857) and/or slightly wider than long (e.g., CAS paratypes and USNM 548090). Several paratypes have only a single point of reduction in scale rows from 19 to 17, by level with the 42 nd ventral (Appendix 2), though in these cases there is a point on each side with an enlarged third or fourth row scale adjacent to a small fourth of fifth row scale. Teeth were counted in 548089 and 548093, both of which have five teeth in each maxilla; the former has seven teeth in each mandible, the latter has seven left and eight right mandibular teeth. Colour of paratypes is generally similar to holotype . Specimens close to shedding are more gray than brown, and less obviously punctate anteroventrally. CAS 225847 has a paler underside of head, with a few entirely unpigmented anteriormost ventrals; also with posteriormost three ventrals and anteriormost pair of subcaudals unpigmented; off white rather than orange rostral tip. CAS 225856 with yellowish rather than off white distal ends of anals. Inner parts of infralabials with more brown flecking in some specimens (e.g., CAS 225856 and 225857). Apex of shield paler than base but with variation in extent of pale region and whether pale brown or off white (yellowish). Referred specimens. Some meristic and morphometric data for 58 specimens are presented in Table 6 , and a bivariate plot of ventral and subcaudal scales of the referred and type specimens is presented in Fig. 11 . The sample displays sexual dimorphism in tail length, and in numbers of subcaudal and ventral scales. As in many other uropeltids, females of the new species have more ventrals, fewer subcaudals, and relatively shorter tails than conspecific males. TABLE 5. Meristic and morphometric (in mm) data for holotype (*) and all seven paratypes of Rhinophis dinarzardae sp. nov. Subcaudal scales data are reported as counts for: left, right (with number of fused pairs in parentheses). Head length = distance between snout tip and posterior end of last supralabial. Head width = distance between posterior edges of last supralabials. “–“ indicates shortest distance between two points.

USNM 548092*	USNM 548089	USNM 548090	USNM 548091	USNM 548093	CAS 225847	CAS 225856	CAS 225857
Sex	M	F	F	F	M	F	M	M
Total length	204	265	231	221	201	224	205	170
Tail length	6.9	5.7	5.5	5.0	7.6	6.2	7.4	7.0
Midbody width	4.9	6.1	6.4	6.1	5.6	6.8	6.3	5.3
Midbody circumference	17	19	22	18	16	21	20	16
Head length	5.3	5.5	5.6	5.3	5.2	4.9	4.9	4.5
Head width	3.5	3.8	4.0	3.7	3.5	3.1	3.1	3.1
Eye diameter	0.3	0.3	0.3	0.4	0.3	0.2	0.2	0.2
Eye–snout tip	2.9	3.0	3.2	3.1	2.9	2.7	2.6	2.5
Eye–eye	2.6	2.6	2.9	2.8	2.6	2.2	2.1	2.1
Rostral length	2.3	2.0	2.2	2.1	2.0	2.0	2.0
Frontal length	1.9	1.6	1.6	1.8	1.7	1.6	1.7	1.5
Frontal width	1.8	1.6	1.5	1.7	1.7	1.5	1.5	1.6
Ventral scales	188	200	198	197	188	199	188	181
Subcaudal scales	6,6 (4)	4,4 (3)	4,4 (3)	3,4 (0)	6,7 (2)	4,4 (1)	6,5 (4)	6,6 (0)
Scales around shield base	11	13	15	14	12	12	14	13
Shield base height	5.7	5.5	5.4	4.5	4.8	5.2	5.2	4.2
Shield base width	4.8	5.2	5.1	4.5	4.4	5.0	4.5	3.9
Shield length (middorsal)	5.6	5.0	6.0	6.0	4.7	5.2	5.6	4.6
Shield length (midventral)	2.4	2.1	2.5	2.0	1.7	2.4	2.4	2.2

TABLE 6. Sex, total (TL) and tail (tL) length (in mm) and scalation data for 58 referred specimens of Rhinophis dinarzardae sp. nov. and 15 specimens of R. cf. dinarzardae sp. nov. (indicated by *). Subcaudal scales given in left, right order (number of fused or undivided subcaudals in parentheses). "–" indicates data not recorded.

Specimen	Sex	TL	tL	Ventral scales	Subcaudal scales
CAS 225750	F	111	3.4	195	4,4 (1)
CAS 225752	M	107	5.0	185	6,6 (3)
CAS 225771	M	–	–	189	6,6 (6)
CAS 225861	F	208	5.3	189	4,3 (1)
CAS 225862	M	191	7.9	183	7,7 (1)
CAS 225863	F	–	–	200	4,4 (2)
CAS 225864	M	200	7.4	182	5,6 (1)
CAS 225867	M	187	7.4	189	6,6 (1)
CAS 225872	M	206	7.5	186	6,6 (0)
CAS 225875	F	227	5.5	192	4,3 (1)
CAS 225877	F	220	4.8	185	4,4 (3)
CAS 225880	F	230	5.1	193	4,4 (1)
CAS 225881	F	223	4.9	199	4,4 (2)
CAS 225890	F	138	3.4	193	3,3 (1)
CAS 225894	F	137	4.2	199	4,4 (1)
CAS 225914	M	–	–	187	6,5 (2)
CAS 225915	M	195	7.6	183	6,6 (3)
CAS 225916	M	–	–	190	6,6 (0)
CAS 225927	M	151	6.2	186	6,6 (5)
CAS 225932	F	200	5.6	185	3,4 (1)
CAS 225938	F	249	4.9	193	4,4 (3)
CAS 225940	F	–	–	195	3,4 (1)
CAS 225943	M	208	8.5	187	6,6 (5)
CAS 225944	M	90	3.6	178	6,6 (1)
CAS 226227	F	188	5.1	198	4,4 (2)
CAS 226229	M	216	8.7	179	6,5 (3)
CAS 226232	M	128	5.9	184	6,6 (4)
CAS 226236	M	–	–	185	6,6 (1)
CAS 226237	M	–	–	187	6,6 (1)
CAS 226238	M	–	–	187	6,6 (2)
CAS 226258	F	–	–	195	4,3 (0)
CAS 226259	F	–	–	193	3,4 (0)
CAS 226267	F	228	5.8	191	4,4 (1)
CAS 226457	F	224	5.0	199	3,3 (2)
CAS 226459	M	189	6.6	188	6,6 (5)
CAS 226465	F	203	5.3	187	3,3 (2)
CAS 226466	F	173	4.9	195	3,3 (1)
CAS 226480	M	150	5.8	187	6,6 (1)
CAS 226484	M	174	6.5	181	5,5 (1)
CAS 226489	M	180	6.9	186	6,6 (0)
CAS 226494	F	158	3.9	195	4,3 (0)
CAS 226499	F	260	6.6	203	4,4 (1)

.....Continued on the next page TABLE 6. (Continued)

Specimen	Sex	TL	tL	Ventral scales	Subcaudal scales
CAS 226501	F	218	5.4	197	4,4 (2)
CAS 226502	M	233	9.3	190	6,6 (1)
CAS 226508	F	224	5.6	199	3,3 (1)
CAS 226527	M	201	6.9	183	7,7 (2)
CAS 226536	M	144	5.0	188	6,7 (2)
CAS 226539	F	104	3.3	199	3,3 (2)
CAS 226541	M	140	5.9	190	6,7 (1)
CAS 226543	F	–	–	204	4,4 (3)
CAS 226596	F	156	3.8	191	4,4 (3)
CAS 226597	F	–	–	199	4,4 (3)
CAS 226598	M	185	7.4	184	6,6 (3)
CAS 244519	F	–	–	201	4,4 (3)
CAS 244541	M	205	7.3	189	7,6 (3)
CAS 244578	F	–	–	200	4,4 (4)
LSUMZ 39255	M	181	6.8	181	5,6 (1)
LSUMZ 39256	M	223	7.6	188	7,6 (2)
*AMNH 104444	F	156	3.5	189	4,4 (3)
*AMNH 104445	F	98	2.9	182	3,3 (2)
*BMNH 1964.1705	M?	115	3.9	177	6,6 (1)
*CAS 225794	M	164	7.5	172	6,7 (2)
*CAS 225933	F	184	4.1	188	3,4 (1)
*CAS 225945	F	212	5.2	193	3,4 (1)
*CAS 226619	F	89	2.4	186	3,3 (3)
*CAS 226622	F	224	4.7	199	4,4 (0)
*CAS 244497	F	–	–	198	3,4 (0)
*CAS 244514	M	165	5.7	181	6,5 (2)
*CAS 257351	M	–	–	182	6,7 (3)
*CAS 244592	M	192	6.6	192	6,6 (2)
*CAS 244797	F	208	5.5	204	4,3 (0)
*CAS 225947	F	165	3.9	198	4,4 (2)
*NMSL-NH 2024.01.03	F	182	4.3	195	3,3 (2)

Colour in life. Specimens generally brownish, with off-white edges to scales ( Fig. 12 ). Off-white areas on venter seen in preserved specimens slightly flushed with pale pink. Etymology. Named in honour of our colleague and friend Dr. Dinarzarde C. Raheem, in recognition of her high-quality contributions to the knowledge of the terrestrial (especially snail) biodiversity of Sri Lanka . Dinarzarde organised and partook in the fieldwork in 2001 that led to the collection of some of the referred specimens of the new species. For nomenclatural purposes, the species epithet is considered a noun in the genitive case. Suggested vernacular name. Dinarzardege thudulla (Sinhala); Dinarzardgin nilakael pambu (Tamil); Dinarzarde’s’s rhinophis or Dinarzarde’s shieldtail (English). Distribution, natural history, and conservation. As far as is known, Rhinophis dinarzardae sp. nov. is restricted to central Sri Lanka in the Knuckles Massif and nearby Makulussa Hill (Dombawela) to the West ( Fig. 8 ), at elevations of ca. 400– 760 m . Its most similar congener and close relative, R. philippinus appears to be restricted to the Central Highlands of Sri Lanka , a short distance to the South. One possible specimen-based challenge to the interpretation that R. dinarzardae sp. nov. and R. philippinus are allopatric is a R. philippinus specimen (CAS 244597, collected in 1978) with the recorded locality “Illukkumbura”, which otherwise has yielded only R. cf. dinarzardae sp. nov. (see Remarks section below). While of potential interest, a single exceptional locality record among tens of other records made at a time when only one of these species was considered distinct might be considered questionable. We suggest that unless and until modern records of R. philippinus from the Knuckles are found (or R. dinarzardae sp. nov. specimens are found in the Central Highlands), the two species can be considered to have non-overlapping geographical distributions. TABLE 7. Sex, total (TL) and tail (tL) length (in mm) and scalation data for 36 specimens of Rhinophis philippinus . Subcaudal scales given in left, right order (number of ‘fused’ or undivided subcaudals in parentheses). "–“ indicates data not recorded. Data for 33 other BMNH specimens (including the holotype of R. planiceps ) plus the holotype of R. philippinus (MNHN 6994) are presented in Huntley et al. (2021) .

Specimen	Sex	TL	tL	Ventral scales	Subcaudal scales
AMNH 24674	M	158	7.2	154	6,6 (1)
MNHN 1895.0070	F	188	5.9	166	4,4 (1)
ZMB 3831	M	154	6.6	168	6,6 (4)
ZMB 29612	F	149	4.5	176	4,4 (2)
ZMB 80291	F	237	5.8	175	3,3 (1)
ZMB 80292	F	181	4.5	175	3,3 (1)
ZMB 80293	F	153	4.0	175	3,3 (1)
ZMB 7267	F	198	5.5	172	4,4 (1)
CAS 226620	F	179	4.4	171	4,4 (0)
CAS 226621	F	212	4.4	170	3,3 (1)
CAS 225810	F	–	–	169	3,3 (0)
CAS 225811	M	–	–	161	6,6 (5)
CAS 225812	F	–	–	168	3,3 (0)
CAS 225813	M	–	–	157	6,6 (0)
CAS 225814	F	–	–	179	4,4 (0)
CAS 225917	M	–	–	158	6,6 (5)
CAS 225918	M	–	–	157	5,4 (0)
CAS 225920	F	–	–	173	3,3 (0)
CAS 225921	F	–	–	173	4,4 (0)
CAS 226082	M	172	7.2	160	5,5 (?)
CAS 226083	F	130	3.9	171	3,2 (?)
CAS 226084	F	113	3.2	164	3,3 (1)
CAS 226603	F	–	–	169	4,4 (1)
CAS 226605	M	–	–	158	6,5 (0)
CAS 244597	M	119	5.0	157	6,6 (2)
CAS 257358	M	–	–	159	5,5 (0)
CAS 257360	F	–	–	171	4,4 (0)
USNM 548083	F	258	5.5	171	3,3 (0)
USNM 548094	M	172	6.0	162	5,6 (1)
USNM 548095	M	174	7.0	159	6,6 (4)
BMNH 58.10.19.53	F	277	6.3	173	3,4 (0)
BMNH 68.3.17.20	M	120	5.5	158	6,7 (1)
BMNH 61.2.21.2	F	238	5.4	176	4,4 (0)
BMNH 65.5.11.189	M	182	6.8	159	5,5 (0)
NMSL KU0046	F	241	6.2	172	3,3 (1)
NMSL KU0085	F	211	4.7	173	3,3 (1)

Huntley et al. (2021) presented and analysed external and skeletal morphological data for 35 specimens of R. philippinus . They noted that one of these (BMNH 1964.1705) was something of an outlier because its subcaudal count (6,6) suggested it was a male but its ventral count (177) was much higher than the maximum count (166) for the other R. philippinus males in their sample. Examination of the internal organs did not clarify the sex of that specimen. Cranial and tail-shield osteological data for that specimen were not highly disparate, although it does lie on the edge of scatter plot distributions of centroid size versus shape ( Huntley et al. 2021 : figs. S2, S3). Based on ventral and subcaudal counts, we here identify BMNH 1964.1705 as R. cf. dinarzardae . The 31 Rhinophis specimens studied by Huntley et al. (2021) that have BMNH catalogue numbers starting “1964.” are all from unknown localities, so the presence of one R. cf. dinarzardae (possibly R. dinarzardae sp. nov. ) in this sample cannot be taken as compelling evidence for sympatry of the new species and R. philippinus . Although precise data for R. philippinus localities are sparse, it seems that this species might extend to greater elevations than R. dinarzardae sp. nov. , with Hantana Estate ( 835 m ) and Agrapatana ( 1,540 m ) exceeding the highest known elevation of 760 m for the latter species. The new species might also extend to lower elevations than R. philippinus , but more data are required to establish elevational patterns with greater confidence. There are few ecological data associated with the available specimens of R. dinarzardae sp. nov. but NMSL MW1739 and MW1741 were excavated from the top 20 cm of soil in low intensity agriculture, where they were syntopic with the caecilian amphibian Ichthyophis glutinosus (see Gower et al. 2005 ). The CAS catalogue reports that CAS 225932 was found in a roadside ditch. Many specimens of Rhinophis dinarzardae sp. nov. have been collected over several decades from multiple localities, including habitats under low-intensity agriculture, and thus the new species possibly qualifies as Least Concern under IUCN Red List criteria. The main caveat to this is that the ecology of this species is very poorly known, and most of the verified field records (from museum collections) are decades old. Remarks . Rhinophis dinarzardae sp. nov. was previously and repeatedly mistakenly identified as R. philippinus . Comparative material examined for the latter species is listed in Appendix 3. The two species are sisters among those Rhinophis sampled thus far for DNA phylogenetics ( Sampaio et al. 2023 ), and are similar in overall appearance (including colour and form of the tail shield), but they differ most obviously in that the new species is more attenuate with more ventrals. It might be noted that number of ventral (and subcaudal) scales is much more useful taxonomically for Rhinophis when accompanied by information on the sex of specimens ( Fig. 11 ). On the basis of locality information and overall appearance, the CAS collection includes many tens of additional specimens likely referrable to R. philippinus or R. dinarzardae sp. nov. , but for which ventral scales have not yet been counted. It is unclear if some historical references to R. philippinus might instead pertain to R. dinarzardae sp. nov. For example, Gans & Baic (1977) illustrate scale microornamentation in “ R. philippinus ” but without specifying locality or voucher information, and the source of the data for the “ R. philippinus ” in Pyron et al. ’s (2013) molecular phylogenetic study is also unclear. Although the type locality of R. philippinus is not known, mtDNA sequence data for the holotype are very similar to those for the only other specimen for which data are available, from Ampitiya, Kandy ( Sampaio et al. 2023 ), giving us some confidence that specimens currently referred to this species are a single taxonomic unit. The holotype of R. planiceps (type locality “ Ceylon ”) is much more similar to the holotype of R. philippinus than to R. dinarzardae sp. nov. and we accept that the former taxon is a junior synonym of R. philippinus . Cadle et al. ’s (1990) analyses of allozyme and albumin immunological variation found evidence for a distinction between their samples of “ R. philippinus ” from the Knuckles (Palatenne) and the Vicarton-Wiltshire Range (Matalapitya). Sampaio et al. (2023) did not find notable differences in DNA-sequence data between a single specimen of R. dinarzardae sp. nov. (labelled as R. cf. philippinus ) from the eastern edge of Makulussa Hill (Dombawela) and the western Knuckles (Palatenne, Kalugaltenna, Kandehana), and the sequences from these four localities were inferred to be a single, distinct species in all of the species-delimitation analyses reported by Sampaio et al. (2023) . However, the R. dinarzardae sp. nov. from these four localities do differ more substantially in DNA-sequence data from a single specimen (also labelled as R. cf. philippinus by Sampaio et al. 2023 ) from further East in the Knuckles (Illukkumbura) (see also Table 4 ), which was also inferred to be distinct from both R. dinarzardae sp. nov. and R. philippinus in all of Sampaio et al’s (2023) species-delimitation analyses. Given the molecular differences, we refrain from referring any Matalapitya or Illukkumbura specimens to R. dinarzardae sp. nov. and instead consider them R. cf. dinarzardae . As seen in Fig. 11 , the Matalapitya and Illukkumbura specimens are more similar to R. dinarzardae sp. nov. than to R. philippinus in scalation, though they have fewer and more ventrals, respectively, than the average R. dinarzardae sp. nov. We suggest that a more thorough examination of the morphological and molecular diversity of the Matalapitya and Illukkumbura populations is warranted. We note that these two localities are somewhat separated from the known localities of R. dinarzardae sp. nov. by lower elevation regions and/or by distance, and in the case of at least Illukkumbura represent areas of drier habitats. Two additional specimens examined for this study resemble the new species in scalation but are associated with localities far beyond the Knuckles, and so are identified for now as R. cf. dinarzardae . CAS 226622 (“Kandy”) and CAS 244497 (“Handapangala”) are both singleton specimens with imprecise locality information that is here considered questionable, especially Handapangala which is at a much lower elevation (ca. 150 m ) and in the Dry Zone ( Fig. 8 ). AMNH 104444 and 104445 (both females) from Pelmadulla are superficially slightly more similar to R. dinarzardae sp. nov. than to R. philippinus although they have relatively low ventral counts ( Tables 5–7 ; assuming, based on number of subcaudals, that both are female specimens). Assuming these two specimens are from the Pelmadulla that is in Ratnapura District, Sabaragamuwa Province , they are from a place far from localities of unambiguously identified R. dinarzardae sp. nov. (ca. 80 km , the other side of the Central Highlands: Fig. 8 ). The centre of Pelmadulla lies below 150 m elevation but the surrounding landscape rises to 500 m within a 2 km radius, and even higher beyond that, so without more precise collection information it is not possible to ascribe elevational data to the likely locality. Pending further investigation, we consider the status of R. philippinus- and R. dinarzardae sp. nov. -like uropeltids from Pelmadulla unresolved.