Hemitrygon yemenensis sp. nov. (Figs. 1–5, Table 1) Holotype. NMW 60783, adult male 223 mm DW, Gischin, Yemen, collected 1902. Paratype. NMW 99841, female 206 mm DW, collected with holotype. Diagnosis. A small species of Hemitrygon (attaining at least 22 cm DW) with the following combination of characters: disc width subequal to length; snout elongate, tip narrowly pointed, angle 107–110°, length 2.1–2.3 times interorbital width; preoral length 2.6–2.7 times mouth width; internasal distance 1.8–1.9 in prenasal length; body and tail mostly naked, denticles confined largely to head; single continuous median row (broken in paratype) of 8–18 small to large thorns on disc (two scapular thorns on each side of adult male), 4–13 similar thorns on tail before caudal sting (more thorns on male than female); tail moderately elongate, slender, whip-like beyond sting, width 0.5–1.2 times its depth, postcloacal tail length 1.9–2.2 times precloacal length; ventral cutaneous fold long, very slender, its length 1.9–2.0 in DW, height 0.3–0.4 in tail depth at its midlength; dorsal skin fold elongate, low, much shorter than ventral fold; distance from cloaca to sting 1.5–1.6 in precloacal length; pectoral-fin radials ~115; total vertebral centra 125–126, monospondylous vertebrae (including all synarcual) 38–39. Description. Disc quadrangular, bluntly angular anteriorly and slightly produced; length subequal to width, width 0.99 times length in adult male holotype (0.98 in female paratype); axis of greatest width of disc slightly forward of scapular region, its distance from snout tip 1.83 (1.89) times in distance from tip of snout to pectoral-fin insertion; body moderately robust, thickness 7.1 (7.4) times in disc width, raised slightly above cranium (marginally more so in scapular region); anterior margin of disc concave anteriorly, straight medially, strongly and evenly convex just before pectoral-fin apex; apex broadly rounded; posterior margin weakly convex; free rear tip broadly rounded. Pelvic fins weakly triangular, anterior and posterior margins almost straight, apices narrowly rounded, free rear tip broadly rounded; relatively large, length 24.3% (25.3%) DW; 1.49 (1.66) times width across fin bases. Tail moderately elongate, slender, with a long, very low ventral skin fold and a shorter and lower dorsal fold (modified somewhat through partial desiccation); postcloacal tail 2.22 (1.88) times precloacal length; width at axil of pelvic fin 1.31 (1.63) times its depth; tapering gradually and evenly to sting base (missing in holotype but position evident as a scar; sting broken at base in paratype); narrowly oval in cross-section near origin of ventral skin fold, width 1.03 (1.25) times height at fold origin; tapering evenly in dorsoventral view below sting scar; very slender, whip-like beyond sting scar, becoming progressively more compressed toward tail tip; subquadrangular in cross-section above mid ventral fold, depth 1.36 (1.16) times width; at end of fold suboval, weakly compressed, depth 1.37 (1.07) times width; filamentous towards tail apex. Dorsal skin fold reduced to low, elongate ridge, length 77 (36) times its height, 1.37 (2.03) in snout length, 2.68 (3.99) in length of ventral fold; its height 1.91 (1.04) in height of mid ventral fold; origin near likely position of undamaged sting apex. Ventral skin fold long, very narrow, length 2.03 (1.87) in disc width, 3.82 (3.05) in post cloacal tail; commencing almost below sting origin, origin 1.1% (1.0%) before sting origin; depth at quarter length 0.43 (0.27), at mid length 0.39 (0.30), at three quarter 0.32 (0.29) in adjacent tail height; distance from cloaca to sting origin 1.54 (1.62) in precloacal length; length of tail beyond ventral fold 0.65 (1.00) in fold length, 2.48 (3.04) in tail length. Lateral line on ventral surface distinct. Snout elongate, subtriangular; apex lobe-like, narrowly and bluntly pointed; angle 107 o (110 o); preoral snout length 2.58 (2.66) times mouth width, 2.34 (2.48) times internarial distance, 1.16 (1.23) times distance between first gill slits; direct preorbital snout length 2.05 (2.31) times interorbital length; snout to maximum disc width 1.97 (2.10) in DW; interorbital space broad, almost flat; eyes small, almost lateral, barely protruding in both types, ventral margin partly covered by thin skin fold; orbit not elevated above disc, diameter 1.00 (0.91) in spiracle length, eye diameter 1.58 (1.35) in spiracle length; inter-eye distance 3.81 (3.31) times eye diameter. Spiracles subrectangular to suboval, enlarged, opening dorsolaterally. Nostril elongate, suboval, directed posterolaterally; anterior margin fleshy; anterior nasal fold internal, broad, membranous; weak oronasal groove present; internasal distance 1.78 (1.93) in prenasal length, 2.51 (2.84) times nostril length. Nasal curtain skirt-like, relatively broad, short, width 1.74 (1.70) times length; not bilobed; surface flat, smooth, without longitudinal medial groove, not covered with minute pores; apex partly recessible within lateral margin of oronasal groove; lateral margin concave (possibly distorted through preservation; paratype almost straight), smooth edged; posterior margin not strongly fringed, weakly concave, not following contour of lower jaw, well removed from symphysis of lower jaw when mouth closed. Jaws asymmetric with teeth visible when mouth closed. Upper jaw strongly arched (teeth highly visible), symphysial part of jaw projecting ventrally; lower jaw strongly convex with truncate apex, broad band of symphysial teeth visible when mouth closed; lateral grooves deep (most evident on right side of holotype), almost straight, extending from nostril to well below lower jaw, longer than nasal curtain length. Lower jaw not projecting forward when mouth open, mouth not protrusible; skin on chin very fleshy, corrugate (more so in less dehydrated paratype); jaws of types not prised apart to reveal oral papillae. Teeth of adult male holotype rather large, variable in shape; those in mid lateral part of upper jaw almost plate-like, their crowns more than twice size of those either side; those at symphysis of upper jaw much smaller, upright, crowns with well-developed slender cusps; cusps much shorter in those teeth posterolaterally; crowns on symphysial teeth in lower jaw somewhat globular, slightly large than those at symphysis of upper jaw; those toward angle of lower jaw concealed; teeth not close-set in either jaw, in straight to semi-oblique rows, not obviously arranged quincuncially; rows in upper jaw ~31 (counted from photograph). Teeth of female paratype smaller than adult male, more similar in size and shape, more closely arranged, quincuncial. Gill openings S-shaped; length of first gill slit 1.36 (1.32) times length of fifth gill slit, 3.64 (3.38) times in mouth width; distance between first gill slits 2.02 (2.01) times internasal distance, 0.43 (0.41) times ventral head length; distance between fifth gill slits 1.31 (1.24) times internasal distance, 0.28 (0.25) times ventral head length. Squamation. Disc and tail with well-developed thorns and an otherwise largely naked disc; holotype with small, widely spaced stellate, dermal denticles on raised part of head; denticles sparse elsewhere, scattered few on post sting tail. Mid-dorsal thorn series in holotype in a single continuous, closely spaced row from nuchal region to sting base; 18 thorns on disc (8 on head, 10 on posterior disc), length 2.0– 6.3 mm, width 1.1 to 2.8 mm; 13 thorns on tail before sting scar, length 3.1–7.9 mm, width 2.1–2.8 mm; thorns very well developed, with rectangular bases and long, semi-erect, pungent lanceolate crowns; two similar thorns on each side of scapulocoracoid; thorns on tail much larger and more widely spaced than those on disc. Squamation of paratype less well developed; fewer dermal denticles on head (~ 12 in orbito-spiracular region); median thorn row less well developed, discontinuous, 8 thorns on disc (4 on head, 4 on posterior disc), length 2.8–5.5 mm, width 1.5–2.5 mm; 4 thorns on tail before sting base, length 6.2–7.2 mm, width 2.5– 2.5 mm. Both type specimens lacking an intact sting; distance from sting base to pectoral-fin insertion 49.4% (49.6%) DW; distance from cloaca to sting base 0.54 (0.53) in disc length. Clasper of adult moderately depressed, robust basally, convex distally and tapering to a blunt or narrowly rounded point; post cloacal length 31.0% DW. Total pectoral radials ~115 left side only (~115 right side only); propterygials ~51 (~47), mesopterygials 16–17 (17–18) and metapterygials ~48 left side only (~51 right side only). Total pelvic radials in right side of paratype 1 + ~26. Total vertebral segments (including first synarcual centra) 125 (126); all synarcual and monospondylous centra 38 (39); total diplospondylous centra 87 (87). Colouration. No information on live colour. In preservative (holotype): Uniformly brownish on dorsal surface, lightest around orbit, orbital membrane darker brown; thorns typical paler than surrounding skin, obvious. Ventral surface yellowish brown (blotching likely to be due to preservation and partial desiccation. Paratype: Dorsal surface pale brownish, lighter than holotype (possible artefact of preservation); scapular region, orbit and interorbit palest. Ventral surface almost uniformly white, paler than dorsal surface. Size.— Male holotype mature at 223 mm DW; stage of development of female paratype 206 mm DW unknown. Distribution. Type material reported as collected from Gischin, assumed to be Qishn in eastern Yemen, on the Arabian Sea coast (Fig. 6). Etymology. Epithet derived from the country of collection. Vernacular name: Heins’ stingray, after Marie and Wilhelm Hein who collected the type material shortly before Wilhelm’s death, and who also collected the holotype of the rare shark Carcharhinus leiodon. Remarks. The dasyatid genus Hemitrygon Müller & Henle 1838 presently comprises several small to medium-sized stingray species distributed in shallow marine, estuarine and freshwater environments. The new species H. yemenensis is currently known only from its collection locality (off eastern Yemen in the northwestern Indian Ocean), whereas almost all species of the genus Hemitrygon are largely restricted to the western Pacific (H. bennetti and H. parvonigra also occur in the eastern Indian Ocean) (Last et al., 2016). The extent of this disjunct distribution is unusual for a small-bodied coastal batoid, so additional scrutiny of the provenance of these specimens was required. However, we believe the collection locality of Gischin (Qishn) in Yemen is most likely correct for a number of reasons. Firstly, Wilhelm and Marie Hein undertook a well-documented expedition from Vienna to South Arabia from December 1901 – April 1902, most of which was spent in Gischin, although they stopped in Aden and Mukalla (Fig. 6). In Gischin they researched Mehri, a Modern South Arabian language that is unique to eastern Yemen and southern Oman. Wilhelm Hein is not known to have visited the western Pacific region (i.e. the core distribution of Hemitrygon) during his lifetime, and died in 1903, the year after the expedition to Gischin (Müller 1909, Klein-Franke 2006). Secondly, listings of NMW holdings of reptiles and teleost fish collected by the Heins on the same expedition appear to be consistent with the fauna of southern Arabia (ABMM, unpubl. data). Finally, the only other elasmobranch that the Heins collected in Gischin, the smoothtooth blacktip shark C. leiodon, has been recorded in eastern Yemen and the adjacent waters of southern Oman, despite having a highly limited known distribution (Henderson & Reeve, 2011; Moore et al., 2013), providing further evidence that the collection locality of H. yemenensis is likely correct. Hemitrygon yemenensis sp. nov. most closely resembles H. akajei (Müller & Henle, 1841) from the western North Pacific, H. bennetti (Müller & Henle, 1841) from the Indo-West Pacific and the Bay of Bengal, and H. fluviorum (Ogilby, 1908) from Australian seas, than any other 10 members of the genus treated recently in a guide to rays of the world (Last et al., 2016). These species all have well-developed thorns and tubercles in a median row on the disc and tail, and a short thorn row on each shoulder of adult males. However the disc shape of H. yemenensis is characteristically marked by a longer and more narrowly pointed snout than these species (and any other member of the genus), and the disc’s length is shorter than its width for all species of Hemitrygon other than H. yemenensis. Compared to the most similar species, H. bennetti, it has a shorter tail (cloaca origin to tail tip 163–188% DW in H. yemenensis vs. 209–258% DW in H. bennetti, n=6), longer disc (length 101–102% DW vs. 91–97% DW), head (length 51–52% DW vs. 44–49% DW) and snout (direct length 25–27% DW vs. 22–24% DW), slightly narrower interorbit (width 11.8–12.3% DW vs. 12.7–14.0% DW), larger eye and orbit (eye diameter 4.6–4.7% DW vs. 3.1–3.7% DW); longer proportions around the head of H. yemenensis are reflected by longer prenasal and preoral lengths. Although the elasmobranch fauna of Yemen has been poorly documented, at least fourteen other species of the family Dasyatidae either occur or are likely to occur there (Last et al., 2016). Of these, H. yemenensis is most similar in morphology to Brevitrygon walga (Müller & Henle, 1841), Maculabatis species (e.g. M. ambigua Last, Bogorodsky & Alpermann, 2016), and Pateobatis jenkinsii (Annandale, 1909) (Last et al., 2016) but unlike all of these species has dorsal and ventral folds on the tail. Hemitrygon yemenensis can also be distinguished from these species by a combination of small size of adult males, squamation (notably an absence of a denticle band, and the arrangement of thorns on disc, tail, and shoulder in adult males), and tail morphology (i.e. length, thickness). Although no fresh material was available for this study, colouration of live or freshly caught H. yemenensis may also be an important distinguishing field character and should be documented at the earliest opportunity. It is noteworthy that nearly 120 years have elapsed since the type specimens of H. yemenensis were collected, and the species has not been recorded nor additional material collected since. It remains to be seen if this apparent rarity is simply a reflection of a paucity of sampling effort near the type locality or a change in its conservation status.