Recharacterization of Rhinophis Dorsimaculatus Deraniyagala, 1941 (Serpentes: Uropeltidae), Including Description of New Material

Home , Rhinophis, Uropeltidae

Zootaxa 4158 (2): 203–212 ISSN 1175-5326 (print edition) http://www.mapress.com/j/zt/ Article ZOOTAXA Copyright © 2016 Magnolia Press ISSN 1175-5334 (online edition) http://doi.org/10.11646/zootaxa.4158.2.3 http://zoobank.org/urn:lsid:zoobank.org:pub:4E5AB214-9CD0-4A60-8C16-E847433A9409 Recharacterization of Rhinophis dorsimaculatus Deraniyagala, 1941 (Serpentes: Uropeltidae), including description of new material

DAVID J. GOWER1,3 & L. J. MENDIS WICKRAMASINGHE2 1Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK 2Herpetological Foundation of Sri Lanka, 31/5, Alwis Town, Hendala, Wattala, Sri Lanka 3Corresponding author. E-mail: [email protected]

Abstract

The Sri Lankan uropeltid (shieldtail) snake Rhinophis dorsimaculatus Deraniyagala, 1941 was described originally from two specimens that were subsequently lost. The small amount of previously published data and lack of published colour photographs made this one of South Asia’s most poorly known snake species, and this resulted in at least one instance of taxonomic misidentification. An additional 10 specimens from a historical collection from the vicinity of the type locality recently came to light. This material is reviewed and documented and the species recharacterized. An additional locality for the species is reported. The newly reported material helps to corroborate the taxonomic validity and distinctiveness of Rhinophis dorsimaculatus. The species is readily distinguished from congeners by having 227 or more ventral scales; a large, dorsally carinate rostral shield; posterior margins of paired anals that are largely separated by the posteriormost ventral scale; and a distinctive colour pattern with bilaterally asymmetrical dark blotches within a broad, pale middorsal stripe and regularly punctate flanks.

Key words: Rhinophis porrectus, Rhinophis punctatus, shieldtail, Sri Lanka, systematics, taxonomy

Introduction

The uropeltid snake Rhinophis dorsimaculatus was described on the basis of two specimens from a single locality in Sri Lanka, deposited in the Department of National Museums, Colombo, Sri Lanka (NMSL: Deraniyagala 1941). Those two specimens could not be located by us or several other people during multiple visits to the NMSL collection since 2001 and are presumed lost or destroyed along with other older NMSL specimens (R. Pethiyagoda, pers. comm. 2006). Gans (1966) reported that the two R. dorsimaculatus specimens were lost during evacuation, citing “Director of National Museums, in litteris”. Deraniyagala (1955) presented a colour painting of the species, confirming his previous description of a distinctive orange and black dorsum. However, the loss of the types, lack of any additional material, and lack of colour photographs of living or preserved specimens has subsequently caused at least one instance of taxonomic confusion in which a previously undescribed species (R. zigzag Gower & Maduwage, 2011) was identified as R. dorsimaculatus (Somaweera 2006; Wickramasinghe et al. 2009). Uropeltid taxonomy in general is in need of reassessment and revision (e.g., Gower et al. 2008). Material from the late Carl Gans’ collection that has been deposited relatively recently in the California Academy of Sciences, San Francisco (CAS) includes 10 specimens of R. dorsimaculatus. In addition, photographs of live animals document a new locality for the species and further information on variation. The purpose of this paper is to recharacterize the species and to clarify its taxonomic status.

Materials and methods

Relevant type and comparative material was examined in NMSL and the Natural History Museum, London, UK (BMNH prefix). Total length was measured to the nearest 1 mm using a ruler or tape measure. Circumference was

Accepted by Z. Nagy: 22 Jul. 2016; published: 29 Aug. 2016 203 measured to the nearest 1 mm using a piece of thread and a ruler. All other measures were taken under stereo dissecting microscopes with vernier calipers to 0.1 mm. Ventral scale counts include all midventral scales between the mental and anal scales, following Gower & Ablett (2006). Dorsal scale-row reduction formulae are based on Dowling (1951; see Appendix). Selected specimens were sexed by looking for oviducts and/or ova for females and epididymis and/or testes for males, through an approximately midventral incision into the coelom. Uropeltid taxonomy follows McDiarmid et al. (1999).

Rhinophis dorsimaculatus Deraniyagala, 1941

Rhinophis dorsimaculatus Deraniyagala, 1941: Deraniyagala (1941: 800–802, fig. 1, plate 1); Smith (1943: 88, 526); Taylor (1950: 533); Deraniyagala (1955: 12, 15, plate 36); Gans (1966: 13–14); de Silva (1980: 113, 131, 183, 190, 191, 195); Mahendra (1984: 63, 65); de Sliva (1990: 45, 69); Cadle et al. (1990); de Silva (1996: 70); de Silva (1998: 111); McDiarmid et al. (1999: 136); Abeysiriwardana et al. (2000: 195); Bambaradeniya & Samarasekara (2001: 36); de Silva (2001: 63); Bossuyt et al. (2004: fig. 2C); Somaweera (2006: 227, 234, 235) (in part); Wickramasinghe et al. (2009: 1, 6, 11) (in part); Gower & Maduwage (2011: 55, 59, 60, 63, 65); Pyron et al. (2013: 977, fig. 1); Wallach et al. (2014: 638).

Holotype. The “type” designated by Deraniyagala (1941). NMSL 86, collected 1938 or 1941 from “the coastal village of Marichchukate [= Marichchikattuwa; Marichchukkaddi], North Western Province, Ceylon [= Sri Lanka]”, now lost (Gans, 1966: 14). Approximate coordinates from maps: 8º 35’ N, 79º 57’ E, < 40 m elevation. “Paratype”. NMSL unnumbered, collected 1938 or 1941 from the type locality. In his description of the species, Deraniyagala (1941) mentioned a second specimen. However, he did not explicitly designate it as a type or report a specimen number. There is no evidence that any of the data or figures reported for R. dorsimaculatus pertain to this second specimen, which we thus do not consider a paratype. The second specimen, which is also lost (Gans 1966: 14), was referred to as a paratype by Gans (1966: 13). Referred material. Ten specimens, all from the type locality, newly referred here. CAS 225802, 225803 and 225804 (collected 27 November, 1974), 225842 and 225843 (29 April, 1976), 226077 and 226078 (9 February, 1977), 226662 (29 May, 1974), and 244583 and 244584 (4 December, 1974). These specimens were in the personal collection of, and under tight proprietary control by, Carl Gans until they were donated to the CAS shortly before his death (C.J. Bell pers. comm. 2016). The CAS material is referable to Rhinophis dorsimaculatus on account of it being topotypic, and matching the holotype in its long and dorsally carinate rostral, large posteriormost ventral projecting between the posterior margins of the anal scales, high number of ventrals (227+), and distinctive colour pattern with a highly regularly punctate flanks and belly and approximately midvertebral, somewhat irregular black blotches lying in a broad, pale, middorsal stripe. Revised diagnosis. In having 227 or more ventral scales, R. dorsimaculatus differs from all congeners except R. punctatus Müller, 1832 and R. porrectus Wall, 1921. Rhinophis dorsimaculatus differs from R. porrectus in having fewer than 260 (227–250 among the only known material) versus more than 260 ventral scales (283 in holotype of R. porrectus). Rhinophis dorsimaculatus differs from R. punctactus in its colour pattern. Although both species have a broad pale dorsal stripe, in the former this bears dark dorsal blotches and in the latter a continuous narrow dark vertebral line. Additionally, the colour of R. punctatus resembles that of R. porrectus in that the pale dorsal stripe is bordered by a broad dark, dorsolateral stripe (broader than the dark lines on the lower flanks and belly), whereas in R. dorsimaculatus the pale stripe is bordered by a narrow dark line that is undifferentiated from the similar lines on the lower flanks and belly. Rhinophis dorsimaculatus also differs clearly from the non-Rhinophis Sri Lankan uropeltids that, although currently classified in other genera (Pseudotyphlops, Uropeltis, e.g., McDiarmid et al., 1999), likely form part of the Sri Lankan uropeltid radiation together with Sri Lankan Rhinophis (Cadle et al. 1990; Bossuyt et al. 2004; Pyron et al. 2013). Thus, R. dorsimaculatus differs (beyond in its distinctive colour pattern) from Pseudotyohlops philippinus Müller, 1832, Uropeltis phillipsi (Nicholls, 1929) and U. melanogaster (Gray, 1858) in having far more than 170 ventrals and a strongly dorsally carinate rostral. The same differences apply to U. ruhunae Deraniyagala, 1954, known from a single specimen that otherwise bears features found only among some Indian uropeltids. Description of referred specimen CAS 225842. Generally good condition; split ventrolaterally and ventrally at vent. Head small; snout pointed. Rostral large, protruding far anterior (1.8 mm) to mouth, pointed, trihedral anteriorly, much longer than wide, dorsally strongly carinate and raised/arched (in lateral view); rostral widest at

204 · Zootaxa 4158 (2) © 2016 Magnolia Press GOWER & WICKRAMASINGHE level of anterior superior corner of first supralabials. Rostral many times longer (in dorsal view) than short rostral- frontal gap. Frontal subtriangular with convex anterior margin, distinctly longer than wide, little break in angle between anterolateral (ocular) and posterolateral (parietal) margins. Frontal much shorter than, as wide as, rostral. Paired nasals (and most of prefrontals) separated from each other by rostral. External naris small, subcircular, slightly countersunk within small depression, located in anteroventral corner of nasal. Nasal contacts supralabials 1 and 2. Prefrontals briefly (for less than 25% of their length) in contact with each other along midline, briefly separating frontal from rostral. Prefrontals taller than long, longer than frontal. Supralabials four, first smallest, making the least contribution to margin of mouth; second larger but only slightly longer at mouth; fourth much the largest. Ocular contacts supralabials 3 and 4. Eye small but distinct, diameter less than one third length of ocular, located near anteroventral corner of ocular, bulging slightly from ocular surface, pupil subcircular. Paired parietals longer than frontal, broadly rounded posteriorly (a little > 90°). Opposite parietals in brief midline contact, left overlapping right. Parietals a little longer than wide, wider than frontal and rostral. Each parietal contacts four scales other than head shields and supralabials. Three infralabials, first and third subequal in length, notably shorter than second. First infralabials not in midline contact behind mental. First ventral longer than wide, second to fourth approximately as long as wide, fifth and subsequent ventrals wider than long. First ventral not in contact with mental. Seven maxillary and seven or eight mandibular teeth on each side. Teeth simple, pointed, distinctly retrorse, straight, evenly spaced. Anteriormost maxillary tooth aligned approximately with posterior end of lower margin of second supralabial, posteriormost maxillary tooth close to posterior edge of lower margin of third supralabial; mandibular row similar in length and alignment, with anteriormost member a little further forward than maxillary row. Inside of mouth, including tongue, unpigmented.

FIGURE 1. Rhinophis dorsimaculatus. Topotype CAS 225842, whole specimen in two views. Scale bar in mm.

Body subcylindrical to slightly dorsoventrally compressed. Head and body scales macroscopically smooth, lacking keels. Body scales generally evenly sized on dorsum and along body. Midline ventral scales between mental and anal of even size though anterior- and posteriormost ones gradually narrow. Posteriormost ventral much larger, V-shaped, separating posterior margins of paired anals. Right anal possibly slightly overlaps left, but overlap almost entirely absent. Ventrals 230, at midbody same width as exposed part of adjacent first dorsal scale row. Posteriormost three or four ventrals slightly wider, 1.05 times as wide as adjacent first dorsal row; final ventral much wider, V-shaped, 1.5 times as wide as first dorsal row. Dorsal scale rows 19 anteriorly, reducing to 17 along most of body, until slightly anterior of vent. Scale row formula:

4+5 (11) 3+4 (224) 19 ------17 ------15 4+5 (11) 3+4 (225)

Dorsal scale rows approximately 13 at base of tail. Paired anal scales slightly larger than posteriormost dorsal scales and subcaudals, slightly smaller than last ventral; only briefly in midline contact, substantially overlapped by last ventral. Distal margin of each anal overlaps three other scales in addition to anteriormost subcaudals. Subcaudals 7 (left), 8 (right), either all paired/divided with posteriormost on right much wider than others, or under an alternative interpretation posteriormost subcaudals all single/undivided. Tail scales macroscopically smooth though many with two or three inconspicuous, low keels on posterior portion of posteriormost subcaudals on lower flanks. Caudal 'shield' bluntly conical, forming tip of tail, longer than wide in dorsal view, only slightly shorter than shielded part of head, more visible from above than below, base (only a littler narrower than base of tail) surrounded by last pair of subcaudals and 10 other scales. In posterior view shield regular broad oval, widest point approximately mid height. Shield surface matt, covered with tiny tubercles in approximately radial distribution. Narrow halo around base of shield glossy and without tubercles. Left hemipenis everted, perhaps not fully. Moderately long (ca. 4.8 mm), slender, subcylindrical, slightly tapering distally. Asulcate surface ornamented with short, slender, curved, proximally-directed, evenly spaced spines extending to base of organ. Sulcate surface mostly lacking spines or other ornamentation, more encroachment of spines towards sulcus margins distally than proximally. Sulcus spermaticus shallow, inconspicuous, walls smooth.

FIGURE 3. Map indicating position of type locality (Marichchikattuwa = Marichchukate) and newly reported locality (Sannar) for Rhinophis dorsimaculatus in northwestern Sri Lanka.

FIGURE 4. Rhinophis dorsimaculatus. Two live, uncollected specimens from Sannar near Vidattalativu. Photographs by L.J.M.W.

-( ( * $ @ 1 . = ?.. $ ( 6% * $ 33 1 33 ( (* $ %9 ( ( %9 @ ( * # % 8%9 6% 7% %9 8%9 ! % 8% % %6 % 8%6 %7 % %6 6%! %9 7% 9% !%6 % !%8 % 6%9 7%! 6%6 %8 7% 6%76%7 6%! 7% !% %8 6%7 !%6 %! 6% 7%! 6%8 % %9 % 7%8 %! 6%! 7%9 8%8 %! 9%9 !% %6 6% 7%8 %9 % % 89% 7% 6% 6% 7%! 8%! % %7 %! !%6 7%! % %8 !% %8 6%6 % 7% % % %6 8% !8%6 7%8 % % 6%6 7%9 %9 6% % 8% 7%8 %! % 6%6 % 8% 7%8 %! %8 6%7 % 7% % 6% 7% %9 9%7 % 8% 7%6 % 01/$ 0- ( (* ( 03 ( (* ( 033 4$5 ($ 67"% -: * : (* ( ( "( * < * ( - ;" *( = 8"= = > 68*$ 9 8" ! !1 ' 8 8* ( 67 6676= ?*$ 8 8 ( * 3 ! ! 6678& @ 8 8 -%$ 68 667! 6% " -% * !8 6 6 8 ! ! 1 ' * 66!6 9 9 %7 8 81 9 6" 8% 66!8 6 ; ! 9 9!" 8 9 9 9 66799 8 8 %6 %7 7% 9 9" 687 6679 8 ! ! ! 6 6 %7 9" 8 8 6% 68 666 %8 %8 68 9 98" 6!!8 ! ! 8%6 6 6 %! 6!6 8 8 %7 6!!! % 9 9 669 9 8% ! ! 6 8 8 %6 68 89 8 8 % 9 9 68 %6 % ; !% 8% ! ! ( ' * 6 6 6 9%7 6!6 ! % 6 8 8 -(1 ' !% 6! 9 9 %6 % %8 ( 6% %8 ( ! ! 68 66 % 9%8 6 9%9 8 8 9 9 %! 9%8 ! ; 6%8 6% 9%7 86 !%6 !% 6 %7 6 8 8% !% 6%! ! 8 6 !%! % 9%7 67 ! ! 8 8 !% % %6 8 8 6 % % !% 8% % !% !%6 %9 9 9 8% 8% !%6 %9 8%6 % % % !%7 !% % % %8 !%! %6 % %! 8%! %7 %6 8%9 % %7 % 8% % "% %& ' ( ' %) (* ' +,-%.

208 · Zootaxa 4158 (2) © 2016 Magnolia Press GOWER & WICKRAMASINGHE In alcohol, background body colour pale tan with much darker (brown to blackish) markings. Broad, regular, pale middorsal stripe/band across five scale rows extending from shortly behind head to tail shield. Pale middorsal band marked with asymmetrical dark blotches, mostly discontinuous, more continuous anteriorly, darker posteriorly, becoming blackish on tail. All other dorsal scale rows of body and all ventrals each with single dark dot together forming series of regular, narrow, punctate lines along length of body. Dark spots on body scales confined to scale bases, distal margins paler and translucent. Tail with more unmarked scales laterally and ventrolaterally than on body. Anals coloured as posteriormost ventrals and dorsal scales of lower flanks. Tail shield matt, mostly pale orange with broad, irregular, subterminal pale brown ring. Head brownish grey with small pale flecks, pale borders to many shields. Rostral somewhat yellowish, paler anteriorly and ventrally. Posteromedial parts of parietals pale. Edges of mouth (‘lips’) paler than dorsal and ventral surface of head. Variation. See Table 1 for meristic and morphometric details. CAS 225843 and 226077 (the latter with much of inside of mouth blackish, this atypical condition possibly artefactual) in generally good condition; other CAS specimens dehydrated or otherwise damaged. CAS 225804 incomplete, in several parts, and poorly preserved; CAS 244583 eviscerated, largely skinned and missing anterior of head; CAS 244584 partly eviscerated and lacking most of vertebral column, end of tail including shield missing. CAS 225842 and the other new specimens closely match the lost holotype. The holotype was seemingly less attenuate than the best-preserved CAS specimens (Table 1). The number and disposition of head shields (including left over right overlap of parietals) is similar in all specimens as far as can be ascertained. Mental notably more prominent in CAS 225843 and separated from anteriormost chin shields (as also in CAS 226078) by midline contact between first pair of infralabials. Ventrals 230–250; subcaudals 6 to 8 on each side. In some specimens (e.g., CAS 225843) it is difficult to determine whether some of the posteriormost scales on the ventral surface of the tail are ‘true’ subcaudals or not. Anals always paired, making little or no midline contact except slight right over left overlap in CAS 226662. Dorsal scale rows always 19 anteriorly, 17 by level of 13th ventral and then without reductions until close to the vent (see Appendix). Tail shield slightly variable in shape and size, for example it is more squarely-ended and strongly overhangs the posteriormost subcaudal in CAS 225843 and, to a lesser extent, in 226077. The shield tubercles are all small but vary from being low and worn (e.g., CAS 225842, 226662) to more prominent, pointed and spine-like (e.g., CAS 225843). Colour pattern generally constant and matching holotype (as figured by Deraniyagala 1941). Posteromedial edges of parietals generally pale (as in holotype and CAS 225842), pigmented in CAS 225843. Tail shield always somewhat orange, but brown marks more or less diffuse and variable in shape (e.g., U- rather than O-shaped in CAS 226662). Pale dorsal band on body always present but variable. For example, it is three (rather than five) scales wide on the anterior half of CAS 226662 and anterior two thirds to three quarters of CAS 225843, 226077, and 244584. The narrow, punctate, darker lines on the other dorsal scale rows and ventrals are a constant feature. Last ventral unpigmented in CAS 225843, 226077, 226662, and 244584. Suggested ‘common’ names: Blotched Rhinophis, Marichchukate Rhinophis (English). The names ‘orange shield tail’ (English) and ‘thambapanni walga ebaya’ (Sinhalese) were used by Wickramasinghe (2012: 112). Distribution, natural history, and conservation: The lost types and only known existing museum specimens of Rhinophis dorsimaculatus are all from a single locality, Marichchukate, on the northwestern coast of Sri Lanka (Fig. 3). This is lowland and in the dry zone (sensu Legg & Jewell 1995), in contrast to most Sri Lankan uropeltids that occur in the hills of the wet zone. We here report a second locality for R. dorsimaculatus. On 21 and 23 January 2014, eight live R. dorsimaculatus were seen at Sannar near Vidattalativu, Northern Province (8º 59’ 30.69” N, 80º 06’ 26.34” E, < 20 m elevation) while digging in loose soil and/or on the surface at night following monsoon rainfall. This locality is also northwestern coast dry zone lowland, approximately 40 km north of the type locality (Fig. 3). The habitat was typical dry-zone evergreen forest mixed with shady home garden vegetation. None of the animals was collected but two were photographed (Fig. 4). Data were not recorded for the two photographed animals but their colour pattern, head shields, and vertebral scale counts (235 and 238) closely match those of the lost holotype and the CAS specimens of R. dorsimaculatus reported here. One of the photographed specimens (Fig. 4, left panel) has a more orange dorsal band (as reported for the type material by Deraniyagala 1941, 1955) and the other (Fig. 4, right panel) a paler tan band. Both specimens differ slightly from the CAS specimens in that the dark dorsal blotches are more continuous and less asymmetric, and the pale band has more scales with dark marks. The type and the newly reported locality were, until recently, inaccessible during three decades of civil unrest in Northern Province. After the war (itself cited as a cause of habitat degradation by Abeysiriwardana et al. 2000),

RHINOPHIS DORSIMACULATUS Zootaxa 4158 (2) © 2016 Magnolia Press · 209 the region has experienced rapid deforestation for human settlement. Much of the forest cover in the region of Marichchukate has been cleared. In addition to expanding human settlements, forest has been cleared for agriculture, which might also pose a conservation threat to R. dorsimaculatus (see also Abeysiriwardana et al. 2000). The second author and colleagues have also seen roadkill specimens of R. dorsimaculatus in the region.

Discussion

Although the holotype of Rhinophis dorsimaculatus is lost, there is no need for a neotype to be designated. Deraniyagala’s (1941) description and illustrations are clear, there are reasonable locality data, topotypic material matches Deraniyagala’s description, and the species is valid. In Deraniyagala’s (1955) colour plate of R. dorsimaculatus (reproduced in Somaweera 2006: 235 plate B), the dorsal scale rows are too irregular, tail shield too flat, and shield spines too large. In reality, the dorsal scales of R. dorsimaculatus are regular. Hemipenis morphology of uropeltids has been little studied to date. The two species of Rhinophis for which hemipenis descriptions are available (R. dorsimaculatus, R. lineatus) differ in the relative size and disposition of spines. As is apparent from their figure, Gower & Maduwage (2011: fig. 5) were in error in describing the hemipenis of R. lineatus as “deeply forked”; instead it is simple, as in R. dorsimaculatus. Species of Rhinophis generally resemble other uropeltids in having a pair of large anal shields that are in broad, overlapping contact and that are not substantially overlapped by the last ventral (e.g., Wickramasinghe et al. 2009; Gower & Maduwage 2011). In this respect, R. dorsimaculatus is atypical—it has a pair of less enlarged anals that are not, or only barely, in contact and which are substantially overlapped by the last ventral (the tip of which overlaps the first subcaudals). We suspect that the anteriormost subcaudals were misidentified as the anals by Deraniyagala (1941: fig. 1C), such that the type likely had seven rather than six subcaudals, the first of which are paired. The condition of the anals in R. dorsimaculatus (relatively small and being overlapped substantially by the posteriormost ventral) is similar to that in the BMNH material of R. porrectus and R. punctatus. Rhinophis dorsimaculatus was described as less attenuate than R. punctatus and R. porrectus (Deraniyagala 1941) but our observations suggest the three species are similar in this respect. Total length divided by midbody width is 46–56 in the four measured specimnes of R. dorsimaculatus housed at CAS (43 reported for the type material: Deraniyagala 1941), 67 in the holotype of R. porrectus (BMNH 1920.8.25.1), and 48–65 in four R. punctatus specimens (BMNH 71.11.13.1-2; 74.4.29.220-221). Reported differences among R. dorsimaculatus, R. porrectus and R. punctatus in the ratio of the lengths frontal:rostral and frontal:parietal (Deraniyagala 1941) do not hold up based on the material examined here for the three species. A photograph of R. punctatus in life (Somaweera 2006: 247 plate A) shows a whitish/pinkish pale dorsal stripe and thin, dark continuous, punctate vertebral line. However, from this photograph we count approximately 270 vertebral scales, and thus we identify it instead as more probably R. porrectus. An “undescribed species” of Rhinophis (Das & de Silva 2005: 67) that has the colour pattern of a R. porrectus or R. punctatus is here identified tentatively as R. punctatus on the basis of the reported ventral count of 251. Smith (1943: 526) suggested that R. dorsimaculatus was “closely related to R. punctatus” (under Smith’s taxonomy, R. punctatus is a senior subjective synonym of R. porrectus). We agree with Smith’s interpretation, based on the shared derived conditions of a high number of ventrals, greatly attenuate body, dorsally strongly carinate rostral (also in some other Rhinophis), small frontal, large posteriormost ventral projecting between posterior margins of anals, and pale, broad, dorsal midline band. Biochemical data did not resolve the relationships of R. dorsimaculatus (Cadle et al. 1990) and the only DNA sequence-based phylogenies published to date (Bossuyt et al. 2004; Pyron et al. 2013) did not sample R. punctatus or R. porrectus.

Acknowledgements

We thank Jens Vindum for his patient loaning of CAS specimens to London. For practical help with various aspects of this study, we are grateful to Mohomed M. Bahir, Lalith Kariyawasam, Kalana Maduwage, Kelum Manamendra-Arachchi, Sudath Nanayakkara, Rohan Pethiyagoda, Dinarzarde Raheem, and Ruchira Somaweera. DJG thanks the Zoology Department of the Natural History Museum, London (NHM) for funding a visit to Sri

210 · Zootaxa 4158 (2) © 2016 Magnolia Press GOWER & WICKRAMASINGHE Lanka that benefited this project. Harry Taylor (NHM) produced Figures 1 and 2, and Prasanna Samarawickrama produced Figure 3. LJMW thanks Dilmah Conservation for funding project activities, and especially Asanka Abayakoon, Department of Wildlife Conservation, Sri Lanka for permission granted, and Jaliya Dehideniya and Ruwan Jayalath Dandeniya for their hospitality at Vidattalativu, Commando Training School. Christopher Bell and Jennifer Olori provided constructive criticisms of an earlier draft.

References

Abeysiriwardana, K.E., Daniel, B.A., de Silva, A., de Silva, P., Ferdinando, J.L., Ganegama, K., Hewamathes, K.K., Herath, N.D., Jayawickrama, A., Jayasinghe, S.S.S., Jayasoorya, L.G., Lasantha, S.A., Premasiri, P.G.D.R., Srinath, D., Rathnayake, L.K., Samarawickrama, V.A.M.P.K., Somaweera, R.K. & Wijewardhana, S. (2000) Rhinophis dorsimaculatus Deraniyagala, 1941. In: de Silva, A., Molur, S. & Walker, S. (Eds.), Conservation Assessment and Management Plan. CAMP Report for Amphibians and Selected Taxa of Reptiles of Sri Lanka. Conservation Breeding Specialist Group, Sri Lanka, pp. 195. Bambaradeniya, C.N.B. & Samarasekara, V.N. (2001) An Overview of the Threatened Herpetofauna of South Asia. IUCN - The World Conservation Union, Sri Lanka Country Office and Asia Regional Biodiversity Programme, Homagama, Sri Lanka, 118 pp. Bossuyt, F., Beenaerts, N., Meegaskumbura, M., Gower, D.J., Pethiyagoda, R., Roelants, K., Mannaert, A., Wilkinson, M., Bahir, M.M., Manamendra-Arachchi, K., Oommen, O.V., Ng, P.K.L., Schneider, C.J. & Milinkovitch, M.C. (2004) Local endemism within the Western Ghats-Sri Lanka biodiversity hotspot. Science, 306, 479–481. http://dx.doi.org/10.1126/science.1100167 Cadle, J.E., Dessauer, H.C., Gans, C. & Gartside, D.F. (1990) Phylogenetic relationship and molecular evolution in uropeltid snakes (Serpentes: Uropeltidae): allozymes and albumin immunology. Biological Journal of the Linnean Society, 40, 293– 320. http://dx.doi.org/10.1111/j.1095-8312.1990.tb00541.x Das, I. & de Silva, A. (2005) Photographic guide to snakes and other reptiles of Sri Lanka. New Holland Publishers, London, 144 pp. de Silva, P.H.D.H. (1980) Snake fauna of Sri Lanka. National Museums of Sri Lanka, Colombo, Sri Lanka, xi + 472 pp. de Silva, A. (1990) Colour guide to the snakes of Sri Lanka. R & A Publishing Ltd., Avon, 130 pp. de Silva, A. (1996) The herpetofauna of Sri Lanka: A brief review. Graphic Land, Kandy, Sri Lanka, 99 pp. de Silva, A. (1998) Snakes of Sri Lanka: A checklist and an annotated bibliography. Department of Wildlife Conservation/ GEF/UNDP/FAO, Colombo, Sri Lanka, 112 pp. de Silva, A. (2001) The herpetofauna of Sri Lanka. An historical overview, current status with checklists. Amphibian and Reptile Research Organization of Sri Lanka, Sri Lanka, 91 pp. Deraniyagala, P.E.P. (1941) A new fossorial snake (Rhinophis dorsimaculatus) from Ceylon. Journal of the Bombay Natural History Society, 42, 800–802. Deraniyagala, P.E.P. (1954) Two new snakes from Ceylon. Proceedings of the 10th Congress of the Ceylon Association for the Advancement of Science, 1, 27–28. Deraniyagala, P.E.P. (1955) A colored atlas of some vertebrates from Ceylon. Vol. 3. Serpentoid Reptilia. Government Press, Ceylon, xvii + 121mpp., 49 pls. Dowling, H.G. (1951) A proposed method of expressing scale reductions in snakes. Copeia, 1951, 131–134. http://dx.doi.org/10.2307/1437542 Gans, C. (1966) Liste der rezenten Amphibien und Reptilien. Uropeltidae. Das Tiereich, 84, 1–29. Gower, D.J., Captain, A. & Thakur, S.S. (2008) On the taxonomic status of Uropeltis bicatenata (Günther) (Reptilia: Serpentes: Uropeltidae). Hamadryad, 33, 64–82. Gower, D.J. & Ablett, J.D. (2006) Counting ventral scales in Asian anilioid snakes. Herpetological Journal, 16, 259–263. Gower, D.J. & Maduwage, K. (2011) Two new species of Rhinophis Hemprich (Serpentes: Uropeltidae) from Sri Lanka. Zootaxa, 2881, 51–68. Gray, J.E. (1858) On a new genus and several new species of Uropeltidæ, in the collection of the British Museum. Proceedings of the Zoological Society of London, 1858, 260–265. http://dx.doi.org/10.1111/j.1469-7998.1858.tb06374.x Legg, C. & Jewell, N. (1995) A 1:50,000-scale forest map of Sri Lanka: the basis for a National Forest Geographic Information System. The Sri Lanka Forester (Special Issue: Remote Sensing), 3–24. Mahendra, B.C. (1984) Handbook of the snakes of India, Ceylon, Burma, Bangladesh and Pakistan. Annals of Zoology, 22, 1– 412. McDiarmid, R.W., Campbell, J.A. & Touré, T. (1999) Snake Species of the World. A Taxonomic and Geographic Reference. Volume 1. The Herpetologists’ League, Washington, 511 pp. Müller, J.P. (1832) Beiträge zur Anatomie und Naturgeschichte der Amphibien. Zeitschrift für Physiologie, 4, 190–275. Nicholls, L. (1929) A new species of Earth Snake of the genus Silybura. Spolia Zeylanica, 15, 153–155.

RHINOPHIS DORSIMACULATUS Zootaxa 4158 (2) © 2016 Magnolia Press · 211 Pyron, R.A., Kandambi, H.K.D., Hendry, C.R., Pushpamal, V., Burbrink, F.T. & Somaweera, R. (2013) Genus-level phylogeny of snakes reveals the origins of species richness in Sri Lanka. Molecular Phylogenetics and Evolution, 66, 969–978. http://dx.doi.org/10.1016/j.ympev.2012.12.004 Smith, M.A. (1943) The Fauna of British India, Ceylon and Burma, Including the whole of the Indo-Chinese sub-region. Reptilia and Amphibia. Vol. 3. Serpentes. Taylor and Francis, London, 583 pp. Somaweera, R. (2006) Sri Lankawe Sarpayin [in Sinhalese; “Snakes of Sri Lanka”]. WHT Publications, Colombo, Sri Lanka, x + 297 pp. Taylor, E.H. (1950) The snakes of Ceylon. University of Kansas Science Bulletin, 33, 519–603. Wall, F. (1921) Ophidia Taprobanica or the snakes of Ceylon. Government Printer, Colombo, Sri Lanka, 581 pp. Wallach, V., Williams, K.L. & Boundy, J. (2014) Snakes of the World: A catalogue of living and extinct species. CRC Press, Boca Raton, 1237 pp. http://dx.doi.org/10.1201/b16901 Wickramasinghe, L.J.M. (2012) The taxonomy and conservation status of the reptile fauna in Sri Lanka. In: Weerakoon, D.K. & Wijesundara, S. (Eds.), The national Red List 2012 of Sri Lanka; Conservation status of the fauna and flora. Ministry of Environment, Colombo, Sri Lanka, pp. 99–113 Wickramasinghe, L.J.M., Vidanapathirana, D.R., Wickramasinghe, N. & Ranwella, P.N. (2009) A new species of Rhinophis Hemprich, 1820 (Reptilia: Serpentes: Uropeltidae) from Rakwana massif, Sri Lanka. Zootaxa, 2044, 1–22.

APPENDIX

Scale row formulae (based on Dowling, 1951) for five of the 10 CAS specimens of Rhinophis dorsimaculatus. CAS 225804 is too incomplete to be included, and data were not recorded for other specimens. Possible scale row reduction in final 20-30 ventrals was not assessed for CAS 225802 or 225803. Upper and lower values correspond to right and left side, respectively.

4+5 (11) 3+4 (229) CAS 225843: 19 ------17 ------15 4+5 (13) 3+4 (231)

4+5 (13) 3+4 (231) CAS 226662: 19 ------17 ------15 4+5 (12) 3+4 (241)

4+5 (11) 3+4 (224) CAS 225842: 19 ------17 ------15 4+5 (11) 3+4 (225)

3+4 (12) CAS 225802: 19 ------17 3+4 (11)

5+6 (9) CAS 225803: 19 ------17 5+6 (9)