SMITHIANA Publications in Aquatic Biodiversity Bulletin No. 7, May 2007 Published by the South African Institute for Aquatic Biodiversity SAIAB Margaret Mary Smith (1916-1987), James Leonard Brierley Smith (1897-1968) with their dog Marlin The publication series (Monographs, Bulletins) of SAIAB (formerly the JLB Smith Institute of Ichthyology), in its new format, honours James Leonard Brierley Smith and Margaret Mary Smith with the name Smithiana, in recognition of their many years of devoted service to African aquatic biology. Their life's work, a team effort, established modern ichthyology in southern Africa and laid the groundwork for the expansion of aquatic biology throughout the region. All contents of this publication © 2007, The South African Institute for Aquatic Biodiversity, Grahamstown Front cover photograph: Left pectoral fin of a preserved coelacanth specimen by Wouter Holleman. © SAIAB, 2006 SMITHIANA BULLETIN NO. 7 CONTENTS John E. McCosker. Luthulenchelys heemstraorum, a new genus and species of snake eel (Anguilliformes: Ophichthidae) from KwaZulu-Natal, with comments on Ophichthus rutidoderma (Bleeker, 1853) and its synonyms. . 3 John E. Randall. Iniistius grijfithsi, a new razorfish (Perciformes: Labridae) from Mauritius . 9 Leonard J. V. Compagno and Phillip C. Heemstra. Electrolux addisoni, a new genus and species of electric ray from the east coast of South Africa (Rajiformes: Torpedinoidei: Narkidae), with a review of torpedinoid taxonomy. . 15 Wouter Holleman. Fishes of the Genus Helcogramma (Blennioidei: Tripterygiidae) in the Western Indian Ocean, including Sri Lanka, with descriptions of four new species. . 51 Smithiana Bulletins and Monographs are publications of the South African Institute for Aquatic Biodiversity, for original scientific articles in the fields of taxonomy, systematics, ethology, ecology, biogeography and conservation of the fishes of Africa and its surrounding waters. Priority will be given to papers by staff and associates of the Institute. Manuscripts from outside the Institute will be considered if they are pertinent to the work of the Institute or use the Institute's collections. For instructions to authors, please see the publications page at http://www.saiab.ru.ac.za Smithiana Bulletin 7: 3-82 Digitized by the Internet Archive in 2017 with funding from JRS Biodiversity Foundation https://archive.org/details/smithiana7200sout Luthulenchelys heemstraorum, a new genus and species of snake eel (Anguilliformes: Ophichthidae) from KwaZulu-Natal, with comments on Ophichthus rutidoderma (Bleeker, 1853) and its synonyms John E. McCosker California Academy of Sciences, San Francisco, California 94118, USA e-mail: Jmccosker@calacademy.org (Received 17 September 2006; accepted 2 January 2007) Abstract. Luthulenchelys heemstraorum genus and species novum, subfamily Ophichthinae, tribe Ophichthini, is described from a 472 mm eel trawled off Durban, KwaZulu-Natal, South Africa, in 450-460 m. Luthulenchelys differs from all known ophichthids in having the following suite of characters: an extremely elongate body, long tail, dorsal-fin origin in anterior trunk region, an elongate pectoral fin, posterior eye/ jaw location, blunt snout, posterior nostril within upper lip, slender dentition, a single vomerine tooth, uniquely developed lateral-line ossicles, and five gill arches, with a very reduced fifth ceratobranchial. Several changes in taxonomy are proposed: Ophisurus lumbricoides Bleeker 1853, Ophisurus rutidodermatoides Bleeker 1853, and Ophichthus derbyensis Whitley 1941 are junior synonyms of Ophichthus rutidoderma (Bleeker 1853); Sphagebranchus lumbricoides Bleeker 1864 is a species of Yirrkala. Key Words: Ophichthidae, Luthulenchelys heemstraorum gen. & sp. nov., KwaZulu-Natal, South Africa, Ophichthus rutidoderma, Ophichthus rutidodermatoides, Ophichthus derbyensis, Yirrkala lumbricoides INTRODUCTION The snake eel fauna of the Western Indian Ocean is rich, diverse and poorly known. J.L.B. Smith (1962) included 55 species in his monograph of the Western Indian Ocean and Red Sea ophicthids. McCosker and Castle's (1986) account of the southern African species (from northern Namibia to Mozambique) treated 29 species and made several changes in taxonomy. In a forthcoming volume on coastal fishes of the Western Indian Ocean and Red Sea (Phillip Heemstra and Jack Randall, editors), I will report at least 70 species and 30 genera of ophichthids, including numerous changes in taxonomic status. In preparation for that work I have examined many of the recently collected specimens from South Africa and beyond that are deposited in the collections of the South African Institute for Aquatic Biology, the Bernice P. Bishop Museum, the United States National Museum of Natural History, the California Academy of Sciences, the British Museum of Natural History, and other institutions, and have discovered at least six species of ophichthids that appear to be undescribed, as well as a number of eels whose taxonomic history is as twisted as the state of their holotypes. Undoubtedly additional ophicthid species from this area remain to be discovered. The majority of the known undescribed species await description or explanation in a variety of generic revisions that are underway. One new species, however, is so distinct that it deserves a new genus as well, and in preparation for the publication of the Western Indian Ocean volume, I herein describe it and take pleasure in naming it after Phil and Elaine Heemstra in recognition of their prodigious efforts in this and previous regional works. MATERIALS AND METHODS The single specimen is deposited at the South African Institute for Aquatic Biodiversity (SAIAB), Grahamstown. Specimen measurements are straight-line, made either with a 300 mm ruler with 0.5 mm gradations (for total length, trunk length, and tail length) and recorded to the nearest 0.5 mm, or with dial calipers (all other measurements) and recorded to the nearest 0.1 mm. Body length comprises head and trunk lengths. Head length is measured from the snout tip to the postero-dorsal margin of the gill opening; trunk length is taken from the gill opening to mid-anus; maximum body depth does not include the median fins. Head-pore terminology follows McCosker et al. (1989: 257), in which supraorbital pores include the ethmoidal pore + pores in supraorbital canal, i.e. 1+3, and the infraorbital pores include pores along the upper jaw + those in the vertical canal behind the eye ("postorbital pores"), i.e. 4 + 2, as the last pore included along the upper jaw is frequently part of the postorbital series. Gill arch and lateral-line ossicle examination was accomplished after removal and clearing and counterstaining with alcian blue and alizarin red dyes (Dingerkus & Uhler 1977). Vertebral counts (which include the hypural) were taken from radiographs. Radiographic techniques are described in (Bohlke 1989). The mean vertebral formula (MVF) is expressed as the average of predorsal, preanal, and total vertebrae (Bohlke, 1982). Institutional abbre¬ viations follow (Leviton et al., 1985). Smithiana Bulletin 7: 3-7 4 John E. McCosker Luthulenchelys een. nov. Figs. 1-4 Type species: Luthulenchelys heemstraorum sp. nov. Diagnosis. An elongate ophichthid, subfamily Ophichthinae, tribe Ophichthini (sensu McCosker 1977), with tail much longer than head and trunk; median fins low; dorsal fin arising well behind pectoral fin; pectoral fin base arising above and occupying more than half of gill opening; gill openings lateral, elongate, nearly vertical and crescentic; eye moderately developed, its centre above posterior quarter of upper jaw, its posterior margin slightly in advance of rictus; jaws moderately developed, but not elongate; snout conical, tapering evenly, its tip rounded; underside of snout with a median sulcus exposing the teeth in advance of anterior nostril bases; anterior nostrils within short tubes; posterior nostrils a hole within upper lip, covered by a flap; teeth conical, erect, numerous and small, uniserial on mandible and biserial on maxillary, a single vomerine tooth; 2 preopercular pores; gill arches similar to those of Ophichthus (cf. Nelson 1966; McCosker 1977), however, fifth ceratobranchial is a thin, nearly ossified rod for anterior 80%, the remainder cartilaginous; upper pharyngeal tooth plates fused. Other characters those of the single species. Etymology. Named Luthulenchelys in honour of Chief Albert John Mvumbi Luthuli of KwaZulu-Natal, Africa's first winner of the Nobel Peace Prize and former President of the African National Congress; and enchelys, an ancient Greek word for eel, feminine. Remarks. In general appearance, the type species of Luthulenchelys appears similar to some of the very elongate species of Ophichthus, such as O. rutidoderma and O. microcephalus. Luthulenchelys heemstraorum differs from them in the condition of its upper lip and associated posterior nostril, its large posterior eye (ending nearly above rictus, rather than in advance of it), its reduced vomerine dentition, its fifth ceratobranchial reduced to a minor cartilaginous rod (rather than an ossified structure), and in the condition of its lateral-line ossicles. Adaptations displayed by the new species such as the nearly uniform dark coloration, small and numerous teeth, fairly large eye, tumid snout (associated with burrowing in soft mud, rather than the sharp-pointed snouts and tails of shallow-water ophichthids that are adaptations for burrowing into sand), moderately developed cephalic pores, and poorly developed median fins are shared by other deep-dwelling ophichthids (cf. McCosker et al. 1989; McCosker 1999). Those similarities and adaptations are so apparent that upon first glance the new species appears very similar to several deep-dwelling Atlantic myrophine ophichthids (McCosker, 1989), such as Asarcenchelys longimanus McCosker 1985, Mixomyrophis pusillipinna McCosker 1985 and Pseudomyrophis nimius Bohlke 1960, whose ancestry is very different to that of the new species. Smithiana Bulletin 7: 3-7 Fig. 1. Holotype of Luthulenchelys heemstraorum sp. nov., SAIAB 75732, 472 mm TL. New snake eel from South Africa 5 Luthulenchelys heemstraorum sp. nov. Holotype: SAIAB 75732, 472 mm TL, sex undetermined (gonads not apparent), from KwaZulu-Natal, South Africa, off Durban (29°57'14"S, 31°16'23"E), 450-460 m, collected by Grant van der Westhuizen aboard F/V Ocean Spray using an otter trawl, field no. ORI 180-3/5, on 16 November 2004. Diagnosis. An elongate species of ophichthine with: tail 62%, head 7.2%, and body depth at gill opening 1.3% of total length; dorsal-fin origin nearly 2 head lengths behind pectoral-fin tips; pectoral fin elongate; posterior nostril a hole within upper lip, covered by a flap that extends below edge of lip; no barbels on upper lip and a no slit at posterior nostril; pores small but conspicuous, SO 1 + 4, IO 4 + 2, POM 5 + 2; teeth small and conical, biserial on upper jaw, uniserial on lower jaw, a single vomerine tooth; coloration uniform grey-black; vertebral formula 20/61/165. pores 1+4, infraorbital pores 4 + 2, lower jaw pores 5, preopercular pores 2, supratemporal pores 3. Lateral¬ line pores apparent only in anterior trunk region; 9 before gill opening in a high-arching sequence, the remainder difficult to discern due to condition of specimen. Lateral-line canal unique in appearance (Fig. 3) among the Ophichthini, exhibiting horizontal tubes indented at each end; pores in tail region appear as prominent sharp spikes as seen from above (these are lateral processes of the vertebrae as evidenced by radiography, perhaps exaggerated by the desiccated condition of the specimen), reaching nearly to tail tip. Fig. 2. Head of holotype of Luthulenchelys heemstraorum sp. nov., SAIAB 75732, 472 mm TL. Counts and measurements of the holotype (in mm). Total length 472; head 34.0; trunk 144; tail 294; predorsal distance 64; pectoral-fin length 8.25; pectoral base 2.4; body depth ~6 at gill openings; body width ~6 at gill openings; body depth ~5 at anus; body width -5 at anus; snout 7.5; tip of snout to rictus 11.2; snout overhang 3.5; eye diameter 3.0; interorbital width 3.5; gill opening height 3.1; isthmus width 2.6. Description. Body very elongate, trunk subcircular, tail laterally compressed; body depth at gill openings 79 in TL. Branchial basket moderately expanded; 17 pairs of overlapping branchiostegal rays visible by radiograph. Head 4.2 in trunk. Head and trunk 2.7 and head 14 in TL. Snout rounded, moderately acute when viewed from above; a short groove bisecting underside of snout nearly to tip of upper jaw. A pair of small bumps on snout flanking dorsal midline in advance of anterior nostril base (their appearance possibly exaggerated by condition of specimen). Lower jaw included, its tip reaching well beyond base of anterior- nostril tubes. Upper jaw not elongated, rictus immediately behind a vertical line at posterior margin of eye. Eye fairly large, 3.7 in upper jaw and 11.3 in head. Anterior nostrils tubular, extending ventro- laterally from snout at -45° from horizontal, reaching below upper lip and beyond tip of chin. Posterior nostrils an elongate opening within upper lip, not visible externally, covered by a flap (expanded edge of lip) that extends below edge of mouth and lacks a vertical slit. No barbels along upper lip. Dorsal-fin origin well behind pectoral fin about a head length into trunk length. Median fins low but obvious, ending a little more than eye diameter before bluntly pointed and laterally compressed tail tip. Pectoral fins elongate. Head pores small but apparent (Fig. 2). Single median interorbital and temporal pores. Supraorbital . A/ Fig. 3. Left lateral-line pores 21-23 of holotype of Luthulenchelys heemstraorum sp. nov., Scale indicates 1 mm. Teeth (Fig. 4) small, conical, slightly recurved; biserial in upper jaw and uniserial on mandible. An intermaxillary rosette of 4, followed by a gap, then a diamond of 4, followed by a single vomerine teeth. Maxillary tooth rows linear, separated by a narrow gap; inner row of -16-17 larger teeth, flanked medially by -22-24 smaller teeth in outer row. Lower-jaw teeth uniserial, -28-30 descending in size to become very small posteriorly. Fig. 4. Dentition of holotype of Luthulenchelys heemstraorum sp. nov., SAIAB 75732, 472 mm TL. Gill arches removed, cleared and counterstained. Gill arches developed, similar to those of Ophichthus (cf. McCosker 1977: 32) except for fifth ceratobranchial condition; first basibranchial ossified, all others cartilaginous; hypobranchials 1-2 ossified, 3-4 cartilaginous; epibranchials 1-4 ossified; infra- pharyngobranchials 2-3 ossified; fifth ceratobranchial very reduced, a thin nearly ossified rod for anterior 80%, the remainder cartilaginous; upper pharyngeal Smithiana Bulletin 7: 3-7 6 John E. McCosker tooth plates fused, containing 2-4 rows of conical recurved teeth; lower pharyngeal tooth plates contain 2-3 rows of 6-14 conical recurved teeth. Colour in ethanol: uniform grey-black; throat, snout and chin slightly darker; median and pectoral fins pale; anterior nostrils, inside of mouth, anal region and peritoneum pale. A photograph of the dead specimen taken soon after its capture indicates that in life it was uniform dark brown. Etymology. Named in honour of Phil and Elaine Heemstra in recognition of their efforts to understand, illustrate, and explain the fishes of the Indian Ocean to scientists and the general public. Distribution. Known only from the type specimen, collected off Durban in 450-460 m depth. Remarks and Comparisons. The characteristics of the new species are those of the monotypic genus. The new species is easily separable from any known ophichthid on the basis of its eye size and location, its extremely reduced vomerine dentition, its body elongation, and other characters unique to the genus. Comparisons with species of Ophichthus that are similar in appearance are described in the following section. COMMENTS ON OPHICHTHUS RUTIDODERMA During my initial attempts to identify the specimen herein described as Luthulenchelys heemstraorum, I examined the extremely elongate species of Ophichthus, O. microcephalus and O. rutidoderma. Ophichthys microcephalus Day 1878 was described from Malabar, India, and is currently known only from Day's description and an extant 625 mm TL syntype in the Australian Museum (B.7843). ( Ophichthus microcephalus may ultimately deserve relocation to another genus, however, the extant syntype is somewhat desiccated and its head is damaged, which would disallow an adequate generic decision.) I compared that syntype to the holotype of Luthulenchelys heemstraorum and found them to differ in numerous characters: the teeth of O. microcephalus are more numerous, more closely-set, and stouter than those of L. heemstraorum-, the snout is shorter and sharper, the body less elongate, the eye is comparatively smaller and located more anteriorly, and O. microcephalus has more vertebrae than L. heemstraorum 12/69/209 vs. 20/61/165); and its lateral-line ossicles are more typical of the Ophichthus condition (McCosker 1977) than are those of L. heemstraorum. I also compared the shallow-water Indonesian species Ophichthus rutidoderma (Bleeker 1853) (treated by Kaup 1856: 18, as Pisoodonophis rutidoderma, later emended to Ophichthys rhytidoderma by Gunther 1870: 63) to Luthulenchelys heemstraorum. It has a similar dorsal-fin origin (above the 16th vertebra) and biserial Smithiana Bulletin 7: 3-7 dentition, but the snout is shorter and sharper, its eye is smaller and located more anteriorly than that of L. heemstraorum, its lateral-line ossicles are more typical of the Ophichthus condition (McCosker 1977), and it has more total vertebrae (195-199 vs. 165). While examining the type (BMNH 1867.11.28:226, 94.5 cm TL) of Ophisurus rutidoderma Bleeker (1853: 30) at the British Museum, I attempted to decipher its identity and that of related Bleeker types. I ultimately concluded that Ophisurus rutidodermatoides Bleeker (1853: 31) [treated by Kaup 1856: 18, as Pisoodonophis rutidermatoides (sic.), later emended to Ophichthys rhytidodermatoides by Gunther 1870: 62] and Ophisurus lumbricoides Bleeker (1853: 32) are junior synonyms of O. rutidoderma. The types of O. rutidodermatoides (BMNH 1867.11.28:292, 61 cm TL) and O. lumbricoides (BMNH 1867.11.28:300, 41.5 cm TL) are in poor condition, however, that of O. rutidoderma is in fair condition. Eschmeyer (1998: 1491) cited O. rutidoderma and O. rutidodermatoides as valid species whose unique holotypes were "whereabouts unknown." My examination of those Bleeker specimens in the British Museum demonstrated that the length and locality data were appropriate, and I therefore identify those specimens, with confidence, as the holotypes. Eschmeyer (1998: 951) listed O. lumbricoides as a valid species of Yirrkala, and also suggested that the whereabouts of the unique holotype was unknown. It is in fact the British Museum specimen, as evidenced by the length of the specimen and additional records in the collection; Eschmeyer's assumption of its validity was based upon McCosker and Castle's (1986: 185) mistaken referral of Ophisurus lumbricoides Bleeker 1853, rather than Sphagebranchus lumbricoides Bleeker 1864, to Yirrkala in their treatment. I have also examined the holotype (AM. 1. 840, 258 mm TL) of Ophichthus derbyensis Whitley 1941 described from a specimen from Derby, Western Australia, and found it to be a synonym of Ophichthus rutidoderma. It has the same proportions, meristics, and dental condition as that of Bleeker's type specimens. The vertebral formulae of the four holotypes are as follows: O. derbyensis 14/61/195; O. lumbricoides 16/64/199; O. rutidoderma 16/61/19 7; and O. rutidodermatoides 7/63/19 7. In summary, the twisted tale of these elongate ophichthids is as follows: Ophisurus lumbricoides Bleeker 1853, Ophisurus rutidodermatoides Bleeker 1853, and Ophichthus derbyensis Whitley 1941 are herein considered to be junior synonyms of Ophichthus rutidoderma (Bleeker 1853), and Sphagebranchus lumbricoides Bleeker 1864 is a valid species of Yirrkala. ACKNOWLEDGEMENTS My list of acknowledgments could easily exceed the length of the elongate holotype of the new species. I sincerely thank the following for their assistance: Grant van der Westhuizen and Sean Fennessy of the New snake eel from South Africa 7 South African Oceanographic Research Institute (ORI) for collecting and photographing the type specimen; Beth Herd Guy for preparing the illustrations; Mysi Hoang (CAS) for clearing and staining the gill arches and skin segment; William N. Eschmeyer (CAS) for advice concerning nomenclature; the research and curatorial staff members of the Australian Museum (AMS), the British Museum of Natural History (BMNH), the Bernice P. Bishop Museum (BPBM), the California Academy of Sciences (CAS), the South African Institute for Aquatic Biodiversity (SAIAB), the United States National Museum of Natural History (USNM), and the Universiteit van Amsterdam Zoologisch Museum (ZMA) for their assistance with specimens; Phil Heemstra and the late Margaret Mary Smith for their advice and assistance during my visits to SAIAB; Tomio Iwamoto (CAS) for reading a draft of this manuscript; and Don Linker and the Jewish Community Endowment Fund for their financial support of the artist. LITERATURE CITED Bleeker, P. 1853. Bijdrage tot de kennis der Muraenoi'den en Symbranchoiden van den Indischen Archipel. V erhandelingen van het Bataviaasch Genootschap van Kunsten en Wetenschappen 25: 1-76. — 1864. Atlas ichthyologicjue des Indes Orientales Neerlandaises, publie sous les auspices du gouvernement colonial. Tome 4, Murenes, Synbranches, Leptocephales. Leiden. 132 pp, plates 145-193. Bohlke, E. B. 1982. Vertebral formulae of type specimens of eels (Pisces: Anguilliformes). Proceedings of the Academy of Natural Sciences of Philadelphia 134: 31-49. — 1989. Methods and Terminology. In: Fishes of the Western North Atlantic, Part 9, Volume 1, ed. E.B. Bohlke, pp. 1-7. Memoirs of the Sears Foundation for Marine Research, New Haven. Bohlke, J.E. 1960. A new ophichthid eel of the genus Pseudomyrophis from the Gulf of Mexico. Notulae Naturae No. 329: 1-8. Day, F. 1878. The fishes of India; being a natural history of the fishes known to inhabit the seas and fresh waters of India, Burma, and Ceylon. Part 4: 553-779. Dingerkus, G. & L.D. Uhler. 1977. Enzyme clearing of alcian blue stained whole small vertebrates for demonstration of cartilage. Stain Technology 52: 229-232. Eschmeyer, W. N., ed. 1998. Catalog of Fishes. California Academy of Sciences, San Francisco. 3 vols., 2905 pp. Gunther, A. 1870. Catalogue of the Physostomi, containing the families Gymnotidae ... Leptocardii, in the British Museum. Catalogue of the fishes in the British Museum. Vol. 8. British Museum, London. 549 pp. Kaup, J. 1856. Catalogue of apodal fish in the collection of the British Museum. London. 160 pp. Leviton, A.E., R.H. Gibbs, JR., E. Heal & C.E. Dawson. 1985. Standards in herpetology and ichthyology: part I. Standard symbolic codes for institutional resources collections in herpetology and ichthyology. Copeia 1985: 802-832. McCosker, J.E. 1977. The osteology, classification, and relationships of the eel family Ophichthidae. Proceedings of the California Academy of Sciences ser. 4, 41(1): 1-123. — 1985. Two new genera and two new species of deepwater western Atlantic worm eels (Pisces: Ophichthidae). Proceedings of the California Academy of Sciences 44(2): 9-15. — 1999. Pisces Anguilliformes: Deepwater snake eels (Ophichthidae) from the New Caledonia region. Southwest Pacific Ocean. In: A. Crosnier (ed.), Resultats des Campagnes MUSORSTOM Vol. 20, ed. A. Crosnier. Memoires du Museum national d'Histoire naturelle 180: 571-588. McCosker, J.E., E.B. Bohlke & J.E. Bohlke. 1989. Family Ophichthidae. In: Fishes of the Western North Atlantic, Part 9, Volume 1, ed. E.B. Bohlke, pp. 254-412. Memoirs of the Sears Foundation for Marine Research, New Haven. McCosker R, J.E. & P.H.J. Castle. 1986. Family Ophichthidae. In: Smith, M.M. and P.C. Heemstra (eds.), Tlie Sea Fishes of Southern Africa, pp. 176-186. Macmillan South Africa, Johannesburg. Nelson, G.J. 1966. Gill arches of teleostean fishes of the order Anguilliformes. Pacific Science 20(4): 391-408. Smith, J.L.B. 1962. Sand-dwelling eels of the western Indian Ocean and the Red Sea. Rhodes University Ichthyological Bulletin 24: 447-466. Whitley, G.P. 1941. Ichthyological notes and illustrations. Australian Zoologist 10: 1-50. Smithiana Bulletin 7: 3-7 . Iniistius griff ithsi, a new razorfish (Perciformes: Labridae) from Mauritius John E. Randall Bishop Museum, 1525 Bernice St., Honolulu, HI 96817-2704, USA. e-mail: jackr@hawaii.rr.com (Received 24 October 2006; accepted 2 January 2007) Abstract. The razorfish Iniistius griffithsi is described from three male specimens caught by handline off the south coast of Mauritius in 120 m. It is distinct in lacking scales on the cheek except for a curved row of 6 small scales from behind to below the eye; 19-20 gill rakers; body depth of 25-2.6 in standard length; first two dorsal spines slender, flexible, and not long (the first 2.45-2.55 in head length); space between first and second dorsal spines more than twice that between second and third spines; and in colour. The most conspicuous colour markings are a broad oblique pale blue band containing many small black spots from high on the nape to above the opercle, and black spots of about pupil size, each edged in pale blue, on the side of the body above the tip of the pectoral fin. Key words: Taxonomy, Labridae, Iniistius griffithsi sp. nov., Mauritius INTRODUCTION Labrid fishes of the six genera of the subfamily Xyrichtyinae (Hemipteronotinae of Norman 1957) contain species that are able to escape predation by quickly diving into sand: Ammolabrus, Cymolutes, Iniistius, Novaculichthys, Novaculoides and Xyrichtys. Three of the genera are monotypic: Ammolabrus dicrus Randall & Carlson is a zooplankton feeder that forms aggregations over open sand substrata; Novaculichthys taeniourus (Lacepede) is usually seen over sand and rubble areas near reefs; and Novaculoides macrolepidotus (Bloch) typically hides in seagrass or algal beds, which it resembles in its colouration. The species of the other three genera may be found over broad open stretches of sand far from the shelter of reefs or dense plant growth needed for shelter by other benthic shore fishes. The three species of Cymolutes have been given the common name knifefishes, and those of Xyrichtys and Iniistius are called razorfishes, in both cases alluding to their compressed bodies and the sharp leading edge of their forehead and snout, specialisations for quick entry into sand. Some razorfishes have been classified in the genus Hemipteronotus Lacepede. As recommended by Randall & Bauchot (1993), this generic name was suppressed by Opinion 1799 of the International Commission on Zoological Nomenclature to preserve the names Naucrates and Xyrichtys. Iniistius has been considered as a synonym of Xyrichtys by some authors. Randall & Earle (2002) followed Tri-thuc Nguyen (1974) who differentiated the two genera osteologically. Externally, the two genera can be distinguished by three characters of the dorsal fin: the origin of the fin is over or less than half an orbit diameter behind the eye in Iniistius, but an eye diameter or more behind the eye in Xyrichtys. The first two dorsal spines of Iniistius are flexible, whereas only the first is flexible in Xyrichtys. The space between the second and third dorsal spines in Iniistius is much broader than the space between the third and fourth spines; by contrast these two spaces are about equal in Xyrichtys. Randall & Earle provided a diagnosis of Iniistius. Species of Xyrichtys occur in the Atlantic, eastern Pacific and the Indo-Pacific region, whereas those of Iniistius are confined to the Indo-Pacific, except I. pavo (Valenciennes), which has extended its range to the eastern Pacific. As might be expected from their habits, razorfishes are not well represented in museum collections. One is described here in the genus Iniistius from three specimens from Mauritius. MATERIALS AND METHODS Type specimens are deposited in the Australian Museum, Sydney (AMS); Bernice P. Bishop Museum, Honolulu (BPBM); and U.S. National Museum of Natural History, Washington, D.C. (USNM). The length of specimens is given as standard length (SL), measured from the median anterior end of the upper lip to the base of the caudal fin (posterior end of the hypural plate); body depth is the greatest depth from the base of the dorsal spines to ventral edge of the abdomen (correcting for any malformation of preservation); body width is measured just posterior to the gill opening; head length is taken from the upper lip to the posterior end of the opercular flap; orbit diameter is the greatest fleshy diameter, and interorbital width the least bony width; snout length is measured from the median anterior point of the upper lip to the nearest fleshy edge of the orbit; upper-jaw length from the same anterior point to the posterior end of the maxilla; caudal-peduncle depth is the least Smithiana Bulletin 7: 9-13 10 John E. Randall depth, and caudal-peduncle length the horizontal distance between verticals at the rear base of the anal fin and the caudal-fin base; lengths of spines and rays are measured to their extreme bases; caudal-fin and pectoral-fin lengths are the length of the longest ray; pelvic-fin length is measured from the base of the pelvic spine to the tip of the longest soft ray. Morphometric data are presented in Table 1 as percentages of the standard length. Proportional measurements in the text are rounded to the nearest 0.05. Lateral-line scale counts include the last pored scale that overlaps the end of the hypural plate; scales in transverse series are counted from the origin of the anal fin obliquely upward to the base of the first dorsal fin; the count of gill rakers is made on the first gill arch and includes all rudiments. Meristic and morphometric data in parentheses refer to paratypes. Fig. 1. Holotype of Iniistius griffithsi, BPBM 40486, male, 140 mm SL, Mauritius, 120 m (photograph by D. Lebreton). Iniistius griffithsi, sp. nov. (Figs. 1-3; Table 1) Holotype: BPBM 40486, male, 140 mm SL, Mauritius, south coast, 0.5 km south of mouth of Riviere des Anguilles, about 1 km offshore from point at 20°31.2'S, 57°33'E, 120 m, handline, Jeremy Griffiths and Tonio Isidore, 1 May 2006. Paratypes: AMS 1.43854-001, male, 138 mm and USNM 388078, male, 143 mm, same data as holotype, except March, 2006. Diagnosis. Dorsal rays IX,12; anal rays 111,12; pectoral rays 12; lateral line interrupted, the pored scales 19-20 + 5 or 6; cheek naked except for a curved oblique row of 6 small scales from behind to below eye; 1 or 2 small scales dorsoanteriorly on opercle; gill rakers 19-20; body depth 2. 5-2. 6 in SL; dorsal profile of snout nearly vertical; first two dorsal spines slender and flexible, the first slightly longer, 2.45-2.55 in head length; space between second and third dorsal spines more than twice that between first and second spines; colour when fresh, pale grey, scale edges dorsally on body grey-brown; head grey-brown, operculum suffused with orange, with vertical blue bands; a broad oblique pale blue band containing numerous small black spots Smithiana Bulletin 7: 9-13 from interorbital space to above opercle; one to several black spots, edged in pale blue, on side of body above tip of pectoral fin; a dusky streak below posterior end of maxilla and lower lip; orangish patch on side of caudal peduncle, each scale with a pale blue spot; median fins bluish grey; pectoral fins orange. Description. Dorsal rays IX,12, all soft rays branched (first 4 or 5 rays of paratypes unbranched), the last to base; anal rays 111,12, all branched (first ray of one paratype unbranched), the last to base; pectoral rays 12, the uppermost very short and spine-like, the second unbranched; pelvic rays 1,5, all soft rays branched; principal caudal rays 12, the uppermost unbranched; upper procurrent caudal rays 6; lower procurrent caudal rays 5; lateral-line scales 20 + 5 (19-20 + 5-6), the last on base of caudal fin; scales above lateral line to origin of dorsal fin 4; scales above lateral line to middle of spinous portion of dorsal fin 2 (uppermost small); scales below lateral line to origin of anal fin 9; circumpeduncular scales 14; gill rakers 20 (19-20); branchiostegal rays 5; vertebrae 25. Body deep, the depth 2.55 (2.5-2. 6) in SL; body very compressed, the width 2.8 (2.75) in body depth; head length 2.9 (2.9-2.95) in SL; snout length (as measured from lower edge of orbit to front of upper lip) 1.9 New razorfish from Mauritius 11 (1.85-1.9) in head length; dorsal profile of snout nearly vertical to level of lower edge of eye, then convex to above eye (indentation of profile above upper lip of holotype the result of a wound, probably from the hook); front of snout narrowing to sharp ridge that extends to above eye; front of chin also a ridge though not as sharp as that of snout; eye set high on head, but nearly an orbit diameter vertically below dorsal profile of interorbital space; orbit diameter 6.0 (5. 9-6. 2) in head length; interorbital width 5.45 (5.4-5. 5) in head length; caudal-peduncle depth 2.6 (2.5-2.75) in head length; caudal-peduncle length 4.05 (3. 9-4.0) in head length. Mouth moderately large, the maxilla extending to or nearly to a vertical at anterior edge of orbit, the upper-jaw length 2.8 (2.75-2.85) in head length; mouth slightly oblique, forming an angle of about 20° to horizontal axis of body; a pair of large, recurved, outdaring canine teeth at front of jaws that overlap lips when mouth closed, the lower pair medial to upper; side of jaws with a close-set series of strong conical teeth, 14 on upper jaw and 13 in lower jaw of holotype; two to three irregular rows of small, slightly nodular teeth medial to row of conical teeth in upper jaw, and one to two rows in lower jaw. Tongue broadly rounded, set far back in mouth. Lips thin, the lower with a well-developed flap along side of mandible. Gill rakers short, the longest on first arch about one-half length of longest gill filaments. Posterior edge of preopercle free nearly an eye diameter below ventral edge of orbit, the ventral edge nearly or just reaching a vertical at posterior edge of maxilla; a narrow fleshy flap surrounding orbit from below anterior edge of pupil to above posterior edge of pupil. Nostrils very small, before lower fourth of eye, the anterior two-thirds orbit diameter before eye, pore¬ like with a short rim and dorsoposterior flap; posterior nostril a short, near-horizontal slit dorsoposterior to anterior nostril, the internarial space nearly one-half pupil diameter. Cephalic sensory pores tiny; sub¬ orbital pores 6, with 2 isolated pores one above the other, nearly an orbit diameter below front edge of eye; a series of 8 preopercular pores, continuing anteriorly as 3 mandibular pores. Scales cycloid and very thin; lateral-line scales with a single horizontal tubule, ending posteriorly in a pore; scales on chest about half height of largest scales on side of body; head naked except for one or two small scales dorsally on opercle and a curved series of six scales, progressively smaller ventrally from behind lower part of eye to below posterior edge of pupil (paratypes with two or three small partially embedded scales ventral to lowermost scale in oblique row); no scales on dorsal and anal fins; two or three rows of scales basally on caudal fin; no pelvic axillary scale; a single scale extending posteriorly from between base of pelvic fins, its length about equal to half length of pelvic spine. Origin of dorsal fin above posterior edge of orbit, the predorsal length 3.6 in SL; first two dorsal spines slender and flexible, the remaining spines sharp- pointed and stiff; space between second and third dorsal spines more than twice space between first two spines; membrane deeply notched between second and third spines; first dorsal spine longest, slightly longer than second, 2.55 (2.45) in head length; third dorsal spine 4.85 (4.7-4.75) in head length: remaining dorsal spines progressively longer, the ninth 3.95 (3.95-4.05) in head length; fifth and sixth dorsal rays longest, 2.5 (2.55-2.6) in head length; origin of anal fin below base of first dorsal soft ray, the preanal length 1.85 (1.85-1.9) in SL; third anal spine longest, 4.35 (4.15-4.2) in head length; fifth anal soft ray longest, 2.6 (2.6-2.7) in head length; caudal fin slightly rounded, 4.85 (4.65) in SL; third and fourth pectoral rays longest, reaching to above origin of anal fin, 4.15 (3.85-3.9) in SL (distal end of about upper fourth of pectoral fin of right side of holotype missing); pelvic fins nearly or just reaching anus, the first soft ray longest, 5.15 _ (4.8-5. 0) in SL; pelvic spine 2.25 (2.2-2.25) in length of longest pelvic soft ray. Table 1. Proportional measurements of type specimens of Iniistius griffithsi as percentages of standard length (SL). Holotype Paratypes BPBM 40486 AMS 1. 43854 USNM 388078 Standard length (mm) 140 138 143 Body depth 39.1 39.8 38.5 Body width 13.9 14.4 14.0 Head length 34.3 34.1 34.6 Snout length 18.3 18.2 18.2 Orbit diameter 5.7 5.8 5.6 Interorbital width 6.3 6.3 6.3 Upper-jaw length 12.2 12.4 12.2 Caudal-peduncle depth 13.3 13.6 12.6 Caudal-peduncle length 8.5 8.6 8.9 Predorsal length 27 8 28 0 27 8 Preanal length 53.8 52.5 54.2 Prepelvic length 31.3 31.0 31.9 Rase of dorsal fin 74 4 76 1 74 5 First dorsal spine 13.5 13.8 14.0 Second dorsal spine 12.4 11.9 12.7 Third dorsal spine 7.1 7 2 7.4 Ninth dorsal spine 8.7 8.6 8.5 Longest dorsal ray 13.6 13.3 13.2 Base of anal fin 40.8 41.7 41.3 First anal spine 4.1 4.1 4.2 Second anal spine 5.8 6.6 6.3 Third anal spine 7.9 8.1 8.3 1 ongest anal ray 13.3 12.7 13 2 Caudal-fin length 20.7 21.4 21.6 Pectoral-fin length 24.2 25.8 25.9 Pelvic-spine length 8 7 9 2 9 0 Pelvic-fin length 19 5 20.8 20.0 Smithiana Bulletin 7: 9-13 12 John E. Randall Colour ofholotype in alcohol: head, chest and most of abdomen purplish grey; body above lateral line dull greyish orange; broad middle zone of body yellowish white with a narrow midlateral dull orange stripe along middle of scale row; three black spots the size of pupil forming a triangle on side of body above distal end of pectoral fin, the uppermost spot on seventh lateral-line scale; scales on posterior fourth of body and base of caudal fin purplish grey, the scale edges paler; a broad oblique whitish band containing numerous small black spots of variable size from above eye, passing above operculum to level of lower edge of eye; lips and chin pale with a grey streak below ventral flap of lower lip to end of maxilla; dorsal and caudal fins translucent grey; anal and paired fins pale yellowish. Colour ofholotype when fresh as shown in Fig. 1. One of the paratypes has only a single pupil-size black spot below the seventh lateral-line scale and no others below. The other has black spots on nine scales, including sixth and seventh lateral-line scales of left side and sixth scale of right side, with one spotted scale above and six or seven on the two scale rows below. In addition, the second paratype has an isolated scale on one side below outer part of pectoral fin with two black spots. All of these spots were edged in pale blue in the newly caught fish. It is regretted that no female specimens were obtained. In view of the sexual dichromatism known for the species of Iniistius, the female can be expected to exhibit some difference in colour pattern. Juveniles will probably have still another pattern, most likely one with dark bars. Etymology. This species is named for the collector, Jeremy Griffiths. Remarks. The three specimens of Iniistius griffithsi were caught by handline from 120 m, hence deeper than normal scuba-diving depths. Fishes of this genus are highly esteemed as food fishes in spite of their relatively small size. They are often consumed by the fishermen themselves, therefore not often found in markets. Two specimens of Iniistius pavo (Valenciennes) were caught from the same location and depth as the three of I. griffithsi. Iniistius pavo and the western Pacific I. dea (Temminck & Schlegel) are easily distinguished from all other species of the genus by the much longer and more forward position of the first two dorsal spines and their complete separation from the remaining spinous portion of the fin. As noted by Randall et al. (2002), the remaining species of the genus Iniistius are divisible into two lineages, one with a broad band of scales that extend ventrally from below the eye to at least the level of the corner of the mouth, and other with just a few scales below the eye. Iniistius grijfithsi falls in the latter group for which only I. aneitensis (Gunther) and I. cyanifrons (Valenciennes) are known. It is easily distinguished from these two species by its striking black-spotted pattern. ACKNOWLEDGEMENTS I am most grateful to Owen Griffiths in Mauritius for noting that this species is an unusual fish, arranging for the collection of an additional specimen, and for its photograph by Dominique Lebreton. When informed of the plan to describe the fish in his honour, he asked that it be named for his son Jeremy, one of the two fishermen who caught the type specimens. Griffiths' first two specimens were given to the Australian Museum; Mark A. McGrouther kindly passed them on loan to me. Thanks are also due Loreen R. O'Hara for taking x-rays. LITERATURE CITED Nguyen, T. 1974. Osteological studies on the labrid fishes (Family Labridae) of Japan-Morphology, taxonomy and phylogeny. Unpublished PhD thesis. Ocean Research Institute, University of Tokyo. Norman, J. R. 1957. A draft synopsis of the orders, families and genera of Recent fishes and fish-like vertebrates. Trustees of the British Museum (Natural History), London. Randall, J. E. & M. L. Bauchot. 1993. Case 2843. Naucrates Rafinesque, 1810 and Xyrichtys Cuvier, 1814 (Osteichthyes, Perciformes): Proposed conservation. Bulletin of Zoological Nomenclature 50(4): 277-281. Randall, J. E. & J. L. Earle. 2002. Review of Hawaiian razorfishes of the genus Iniistius (Perciformes: Labridae). Pacific Science 56(4): 389-402. Randall, J. E., D. R. Robertson & J. L. Earle. 2002. Iniistius auropunctatus, a new razorfish (Perciformes: Labridae) from the Marquesas Islands. Cybium 26(2): 93-98. ADDENDUM While this paper was in press, the author determined that Gerald R. Allen observed about 10 individuals of Iniistius griffithsi varying from about 8-18 cm in total length in the outer section of Loky Bay, northeastern Madagascar (12°43.532'S, 49°42.386'E) on a clean sand bottom at a depth of 8 m in March 2006. His underwater photographs of two adult individuals are reproduced here as Figs. 2 and 3. Allen observed a similar fish at Christmas Island in the Indian Ocean in July, 2006. Smithiana Bulletin 7: 9-13 New razorfish from Mauritius 13 Fig. 2. Iniistius griffithsi , presumed female, Loky Bay, northern Madagascar (photograph by G.R. Allen). Fig. 3. Iniistius griffithsi , presumed male, same locality as Fig. 2 (photograph by G.R. Allen). Smithiana Bulletin 7: 9-13 Electrolux addisoni, a new genus and species of electric ray from the east coast of South Africa (Rajiformes: Torpedinoidei: Narkidae), with a review of torpedinoid taxonomy Leonard J.V. Compagno1 and Phillip C. Heemstra2 1 Shark Research Centre, Iziko-South African Museum, Cape Town, South Africa, e-mail lcompagno@iziko.org. za 2 South African Institute for Aquatic Biodiversity, Grahamstown, South Africa, e-mail p.heemstra@ru.ac.za (Submitted 7 December 2006; accepted 23 February 2007) Abstract. A new genus and species of sleeper ray, Electrolux addisoni (Family Narkidae), with two dorsal fins is described from two adult males (total lengths 50 and 52 cm) caught on a shallow reef off the east coast of South Africa. Electrolux is distinguished from other genera of Narkidae by its prominent spiracular papillae, the morphology of its nostrils, nasal curtain, mouth, jaws, chondrocranium, basibranchial skeleton, pectoral and pelvic girdles, and unique and complex colour pattern. It has higher vertebral, pectoral radial, tooth and intestinal valve counts than other narkids and reaches a greater size than all species with the possibly exception of Typhlonarke aysoni. Taxonomic definitions are provided for the electric rays, for the family Narkidae, and for Electrolux, as well as keys to families of electric rays and to the genera of Narkidae. The systematics of the narkid genus Heteronarce is reviewed and the genus validated. Members of the Narkidae may include the smallest, or at least the shortest, living chondrichthyans ( Temera hardwickii and an undescribed species of Narke). Electrolux addisoni is a reef-dweller that eats polychaete worms and small crustaceans, and has been photographed and videotaped by divers while actively feeding in the daytime. The conspicuous dorsal colour pattern may be aposematic, as the ray was seen to make a possible threat display when closely approached. Electrolux addisoni is recorded from four localities along an approximately 310 km. strip of coastline from Coffee Bay, Eastern Cape Province, to just north of Durban, kwaZulu-Natal inside the 50 m isobath. This conspicuous, active ray is known only from a few diver records from reefs reported over approximately two decades, and its conservation status needs to be critically assessed. Key words: Electrolux addisoni, new genus and species, Narkidae, sleeper ray, Torpedinoidei, description, taxonomy, distribution, biology, conservation status. The coastal fish fauna of the east coast of South Africa is a mixture of tropical and subtropical Indo-West Pacific fishes (mainly coral-reef species), many endemic warm temperate species, and several worldwide species. This South African fish diversity is not well known, and in the past 20 years, our cursory fish survey work has produced numerous range extensions, many new records and several new species. In addition to our own collecting efforts and those of fisheries biologists on South African research vessels (particularly RV Africans, see Compagno, 1999b) and fishing companies and observers on fishing vessels, our knowledge of the South African marine fish diversity is significantly enhanced by the specimens and photographs provided by various amateur ichthyologists (anglers, aquarists, divers and underwater photographers). This paper describes a remarkable new genus and species of electric ray (Family Narkidae) from the east coast of South Africa. The species was first made known to us by diver-photographer Peter Chrystal (through Rudy van der Elst, pers. comm.), who photographed this spectacular ray on a patch of sand on Aliwal Shoal, kwaZulu-Natal, South Africa in 1984 (Figure 1). The ray was subsequently rediscovered on Protea Banks off Shelly Beach, kwaZulu-Natal in 1997 by underwater photographers Stephania and Peter Lamberti (pers. comm.) who sent us a video-clip of the ray (Figure 2). It has also been photographed underwater at the Tee Barge north of Durban by Dennis King. A live specimen was seen underwater by P. Fleemstra at Coffee Bay, Eastern Cape in 2001, but the ray fled before he could net it. From the 1984 photographs we immediately recognized an undescribed species of electric ray (Torpedinoidei or Torpediniformes) with a unique dorsal colour pattern that was far more elaborate and ornate than that of any electric ray known at the time, although Last & Stevens (1994) and de Carvalho (1999) subsequently described species of Narcine off Australia and in the Western Hemisphere with ornate but simpler colour patterns. The ray was tentatively considered as either a member of the family Narkidae or Narcinidae by Compagno (in Smith & Heemstra, 1995) but after video footage became available, Compagno (1999b) thought it to be an undescribed narkid most probably of the genus Heteronarce because of its general morphology and twin dorsal fins. Smithiana Bulletin 7: 15-49 16 Leonard J.V. Compagno and Phillip C. Heemstra However, the ray frustrated us by eluding capture for nearly two decades despite efforts by colleagues to collect it. In September 2003 a specimen (SAIAB 78777) was collected by Mark Addison on a reef off Manaba Beach near Margate in southern KwaZulu- Natal. A second specimen was collected by Mark's father, Brent Addison at the same locality in October 2003. Both specimens (Figures 3-4) were presented by Mark Addison to the South African Institute of Aquatic Biodiversity and the second specimen was transferred to the fish collection of Iziko - South African Museum (SAM 36908). Study (including radiography) of the two specimens and dissection of the paratype (SAM 36908) confirmed our hypothesis that they represented a new species of the Family Narkidae (sleeper rays), which in South African waters also includes the common onefin electric ray, Narke capensis and the rare Natal electric ray, Heteronarce garmani (Compagno in Smith & Heemstra, 1986, Compagno et al., 1989). However, this new electric ray proved sufficiently different from known narkid genera Heteronarke, Narke, Temera and Typhlonarke to warrant a new genus and species (Compagno, 2005). We also discuss the systematics of Heteronarce and the Family Narkidae, and the taxonomy of electric rays, and comment on the size of narkid rays and the biology, habitat and conservation status of the new ray. Figure 1. Electrolux addisoni photographed by Peter Chrystal at Aliwal Shoal, KwaZulu-Natal, South Africa in 1984. These two photographs were sent to the junior author by Rudy van der Elst and show the intricate concentric black stripe of the dorsal colour pattern on the disc of a live ray. Figure 2. Two frames from a videotape of Electrolux addisoni taken by Stephania and Peter Lamberti in 1997 on Protea Banks off Shelly Beach, KwaZulu-Natal, South Africa. Left, ray vigorously feeding on substrate and moving towards the photographer, using its pelvic fins as legs. Right, ray stationary and giving a possible threat display, with disc curled and tail raised. Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 17 METHODS External Morphology and Morphometrics: Terminology and abbreviations for torpedinoid external morphology and morphometrics are based on Bigelow and Schroeder (1953) for batoids, Compagno & Roberts (1982, 1984) for stingrays, and Compagno (1984, 1988, 2001) and Compagno et al. (2005) for sharks. Measurements and their abbreviations for electric rays are based on the system for sharks in Compagno (1984, 2001) with some modifications and additions. For the bases of precaudal fins the term origin is used for the anterior end of the fin base and insertion for the posterior end of the base. The morphometric abbreviations and definitions for torpedinoid measurements are presented below (Table 1). Meristics. Terminology and abbreviations for torpedinoid vertebral counts is derived from that for sharks in Compagno (1988) and for stingrays in Compagno & Roberts (1982, 1984) and are defined as follows: SYN, total synarcual vertebrae; sum of SYC, synarcual segments anterior to synarcual centra and determined by count of neural canals through lateral walls of synarcual, and SYC, synarcual centra in posterior end of synarcual. MP, monospondylous precaudal centra, between posterior end of synarcual and monospondylous- diplospondylous transition, sum of MPN, monospondylous precaudal centra without ribs, and MPR, monospondylous precaudal centra with ribs. DP, diplospondylous precaudal centra, between MP- DP transition and upper origin of caudal fin; sum of DPN, diplospondylous precaudal centra without ribs, and DPR, diplospondylous precaudal centra with ribs. DC, diplospondylous caudal centra from upper origin of caudal fin to end of vertebral column. PC, precaudal vertebrae, including SYN, MP, and DP vertebrae. PCC, precaudal centra, including SYC, MP, and DP centra. TF, total free centra from posterior end of synarcual to end of caudal centra; sum of MP, DP, and DC centra. TC, total free centra plus synarcual centra; sum of SYC, MP, DP, and DC. TS, total segments, all centra plus SYS. SYN%, total synarcual centra as percentage of total free centra, 100*(SYN/TC). MP%, monospondylous precaudal centra as percentage of total free centra, 100*(MP/TC). DP%, diplospondylous precaudal centra as percentage of total free centra, 100*(DP/TC). DC%, diplospondylous caudal centra as percentage of total free centra, 100*(DC/TC). Tooth and intestinal valve terminology and count methodology is after Compagno (1988). Terminology and information on torpedinoid anatomy including chondrocranial and hyobranchial morphology follows Henle (1834), Gill (1862), Haswell (1895), Garman (1913), Holmgren (1941), and Compagno (1973, 1977, 1988, 1999a). Clasper terminology for torpedinoids follows Leigh-Sharpe (1922, 1924, 1926) and Compagno (1988). ORDER RAJIFORMES, SUBORDER TORPEDINOIDEI - ELECTRIC RAYS Definition (derived from Compagno (1973, 1977, 1999a). Rays (Rajiformes) with head broadly depressed and included with body and hypertrophied pectoral fins to form a thick, flattened, oval, circular or subquarate fleshy pectoral disc with broadly rounded apical margins. Trunk thick, broad, and depressed, not rising abruptly dorsal to pectoral bases. A pair of large, kidney-shaped electric organs in disc between propterygium, branchial region and pectoral girdle, generally visible through skin of ventral surface. Precaudal tail usually stout and muscular, more or less elongated, but diminutive in Hypnidae; tail with two, one or no dorsal fins, ventro-lateral folds (often present) but no median dorsal or ventral skin folds; tail without caudal electric organs or a sting. Preoral snout short to moderately long, 6-21% of total Total length, broadly rounded or truncated, front edge continuous with that of the pectoral disc and not angular, not formed into a tooth-studded rostral saw. Eyes small to moderate-sized in most taxa (eyes vestigial in some Narke dipterygia and absent externally in Typhlonarke) dorsal on the head, well medial to lateral margins of disc and anterior to spiracles; eyes with a velum on the cornea but no semilunar groove below them. Spiracles either contiguous with rear margins of eyes or separated from them by a space less than their width; spiracles with internal pseudobranchs. Nostrils moderately large, close together, just anterior to mouth; circum-narial flaps and grooves well developed on incurrent apertures of nostrils; anterior nasal flaps fused medially to form a broad nasal curtain, which is free posteriorly and usually reaches mouth; broad nasoral grooves present between excurrent apertures of nostrils and mouth. Mouth always subterminal on head although varying in position (close behind front edge of snout in Electrolux), small to moderate-sized, straight or arcuate. No "ill sieves or rakers on internal gill slits. Teeth small and not fused into crushing plates; dental bands limited to medial half of jaws, not extending to mouth corners; teeth in 8-68 / 7-75 rows, total rows 15- 141; tooth crowns carinate, monocuspidate or tricuspidate. Skin completely naked on all surfaces of disc, tail, fins, and claspers. Pectoral fins expanded, fused medially with head and trunk, usually not obviously distinct externally; pectoral fin bases very long, extending from nasal capsules to pelvic fin origins; pectoral axils much closer to vent than to first gill apertures. Propterygia of pectoral fin skeleton greatly elongated, longer than metapterygia, segmented anteriorly and forming a propterygial axis; propterygia reaching level of nasal capsules or ending behind them, radials extending in front of nostrils but not reaching snout tip. Mesopterygia large and expanded anterolaterally to opposite posterior quarter to half of basal segment of propterygial axis, and Smithiana Bulletin 7: 15-49 18 Leonard J.V. Compagno and Phillip C. Heemstra carrying its radials. Scapulocoracoids with a pair of lateral bars connecting coracoid bar with scapular processes on each side, separated by a large fenestra from lateral face of scapulocoracoid, which forms unique hollow tube, more or less elongated posterolaterally and terminating in a small, rounded- oval articular surface with a distinct procondyle for the propterygium, small anterodorsal and anteroventral foramina above and below a broad anterior bridge between the procondyle and a small mesocondyle for the mesopterygium, small postdorsal and postventral foramina, and a moderate-sized metacondyle for the metapterygium. Suprascapulae fused together above synarcual or behind it and above neural spines of free monospondylous vertebrae; suprascapulae without the complex attachment to vertebral column as in other batoids; distal ends of suprascapulae straight and not forked, articulating with scapular processes. Pelvic fins not divided into distinct anterior and posterior lobes (except in Typhlonarke, where the anterior lobes are fused with and protrude from the pectoral disc and the posterior lobes are fused to the posterior disc). Origins of pelvic fins anterior to pectoral free rear tips; pelvic fins with straight, convex, or concave posterior margins. Pelvic girdle with strong lateral prepelvic processes, short ischial and iliac processes, but without a medial prepubic process on puboischiadic bar; puboischiadic bar posteriorly arched or transverse, not anteriorly arched. Claspers short, stout, and protruding a short distance past pelvic free rear tips or not at all in adult males; clasper glans very simple, with small pseudosiphon and pseudopera but without clasper spines. Dorsal fins usually two (one in Narke and Typhlonarke, absent in Temera ), moderately large (small in Hypnidae) and rounded-angular (not falcate). First dorsal fin when present with base over or just behind pelvic bases and over anterior half of precaudal tail. Caudal fin usually large to moderate-sized (small in Hypnidae); caudal fin vertebral axis horizontal to weakly diagonal and elevated, diphycercal or weakly heterocercal; dorsal and ventral caudal fin margins broadly rounded, terminal lobe mostly rounded or occasionally pointed, a low ventral caudal lobe present or absent; caudal fin without a differentiated postventral margin and without a discrete terminal lobe and subterminal notch. Vertebral column with cervicothoracic synarcual but no thoracolumbar synarcual; synarcual variable in length, ending before or behind suprascapulae; synarcual not formed anteriorly into a collar-like sheath around spinal cord but with a peg¬ like ventral projection fitting in foramen magnum and between occipital condyles. Cranium with rostrum variably developed, absent (Hypnidae), more or less reduced (Torpedinidae, Narkidae) or moderately large, wide, and trough-shaped (Narcinidae). Precerebral cavity more or less expanded into rostrum but not roofed dorsally, truncated anteriorly in Narkidae. Rostral nerves not enclosed in rostrum. Nasal capsules expanded laterally, ventrolaterally, ventrally or anteroventrally; internasal septum broad and depressed (Narcinidae) to more or less narrow and compressed (Hypnidae, Narkidae), broadly separating the nasal capsules or not. Antorbital condyles on anterior, lateral or posterior surfaces of nasal capsules; antorbital cartilages distally expanded, directed anteriorly or anterolaterally, fan-shaped or antler-shaped and branched; antorbital cartilages not articulating with propterygia. Preorbital processes rudimentary or absent, no supraorbital crests, no postorbital processes and no low suborbital shelves. Anterior fontanelle not delimited anteriorly by a transverse ridge. Cranial roof with frontoparietal fenestra absent or very small to huge; separated from anterior fontanelle by an epiphysial bar or not. Basal plate flat or arched and without basal angle; basal plate with a single internal carotid foramen or two narrowly separated carotid foramina at its midline. Occipital condyles small, ventral, and not covering occiput. Dorsal and ventral labial cartilages present in Narcinidae and Narkidae; absent and presumably lost in Hypnos and Torpedo. Hyobranchial skeleton without basihyoid element; hypobranchials discrete, parasagittal, not fused to form a midventral plate and not fused to well-developed basibranchial copula. Branchial rays expanded distally as broad circular plates. Mode of reproduction ovoviviparous as far as is known, with prenatal young nourished primarily by their yolk sacs. ELECTROLUX COMPAGNO & HEEMSTRA, genus novum. Type species: Electrolux addisoni Compagno & Heemstra sp nov. described below. Generic Diagnosis: Narkid electric rays with subcircular disc, length 51% TL; snout short, broadly rounded and nearly straight (Figs 1- 4). Eyes well- developed, close to front edge of disc, but mostly hidden by loose skin. Spiracles (Fig. 5) contiguous with eyeballs, rim a low, rounded, ridge with 5 or 6 long, slender stiff papillae and 2 or 3 short, soft papillae, including one minute papilla on eye; spiracle diameter ~ 1.1-1. 3 times eye diameter. Incurrent apertures of nostrils nearly circular, flaps of incurrent apertures (circumnarial flaps) broad, large, flattened, trumpet¬ like and elongated, their length 2/ 3-3/ 4 length of anterior nasal flaps; posterior margin of nasal curtain deeply incised, with prominent lateral lobes and a small medial lobe; ventral surface of curtain with a shallow medial groove; length of anterior nasal flap about 1.1 in outer internarial width; lateral margins of curtain nearly parallel. Mouth and nostrils projecting ventrally from disc as a prominent nasoral turret (Figs 6A & 7) near front edge of disc. Lips between labial folds and dental bands thin and with small transverse pleats of skin; lower lip narrow and thin, no mental Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 19 Figure 3. Holotype of Electrolux addisoni, 515 mm TL adult male, in A, dorsal and B, ventral views. Note differences in dorsal disc colour pattern compared with live animals (Figs. 1 & 2). Figure 4. Paratype of Electrolux addisoni , 502 mm TL, adult male, in A, dorsal and B, ventral views. Smithiann Bulletin 7: 15-49 20 Leonard J.V. Compagno and Phillip C. Heemstra Figure 5. Photos of left eye and spiracle (anterior to left) of: A. Electrolux addisoni, holotype, with papillose depressed spiracular rim. Pin penetrates skin at base of ocular papilla, forceps are clamped on fleshy lower eyelid. B. Heteronarce garmani, SAM 34813, 289 mm TL adult male, with smooth elevated spiracular rim. Photographs by authors. Scale bars = 5 mm SYG Figure 6. Mouth and nostril drawings of A. Electrolux addisoni, holotype. B. Heteronarce garmani, RV Benguela G13531 88N 30-08, 125 mm TL female. C. Narke capensis, SAM uncataloged, 265 mm TL adult male. D. Temera hardwickii, SU-35728, 104 mm TL adolescent male. E. Typhlonarke aysoni from SIO 61-149-6A and Garrick (1951). Drawings by Elaine Heemstra based on specimens and sketches or photographs by LJVC . Abbreviations: ANF, anterior nasal flap; CNF, circumnarial fold; CNG, circumnarial groove; INA, incurrent aperture; LLF, lower labial fold; LLG, lower labial groove; MG, medial groove on nasal curtain; NCU, nasal curtain; SYG, symphysial groove. Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 21 groove, but ventral dental band bisecting it medially; lower labial folds and grooves short, ending far lateral to midline of mouth (Figure 7); labial cartilages very small, their bases well lateral to symphyses. Total tooth row counts 32-34 or 15-16 / 17-18. Superscapulae C-shaped (Fig. 13A), situated anterior to coracoid bar. Total pectoral radials 47-53; metapterygial axis not greatly reduced in length, with equal numbers of radials to propterygium (17-19 each); no neopterygial radials on scapulocoracoid. Pelvic fins not divided into discrete anterior and posterior lobes. Puboischiadic bar with stout angular iliac processes (Fig. 14A). Total basipterygial pelvic radials 17-19, Claspers short and extremely flat in adult males (Figure 15), not extending past pelvic free rear tips. Two dorsal fins, subequal in area, second dorsal slightly smaller than first, first dorsal origin over pelvic free rear tips and well behind pelvic insertions. Lateral tail folds broad. Total synarcual segments 14-16, monospondylous precaudal centra 30, diplospondylous precaudal centra, 61- 62, precaudal centra 96-97, precaudal total segments 105- 108, total free centra 118-120, total centra 123-125, total segments 132-136. Neurocranium with a narrow, large, erect, flattened medial rostral cartilage, and two lateral rostral cartilages below it forming a large ventrally-directed yoke-like structure (Fig. 9); antorbital cartilages with narrow anterior branched ends (Figs 11 and 12); ethmoid region of chondrocranium laterally compressed, strongly bent and expanded ventrally; no frontoparietal fenestra; otic capsules small and not expanded laterally. Jaws antero-ventro-medially expanded, palatine processes of palatoquadrates nearly straight. Second hyobranchial cartilages large. Intestinal valve count 17 (paratype). Mature males 50-52 cm. TL. Elaborate colour pattern of pale spots on dark brown background present on most of body; dorsal surface of disc of living rays with concentric black lines and pale spots. Derivation of generic name: The name alludes to the well-developed electrogenic properties of this ray (collectors and photographers have experienced the shocking personality of this bold, active and brightly patterned electric ray first-hand), the discovery of which sheds light (Latin, lux) on the rich and poorly- known fish diversity of the Western Indian Ocean. And the vigorous sucking action displayed on the videotape of the feeding ray that was taken by Stephania and Peter Lamberti may rival a well-known electrical device used to suck the detritus from carpets, furniture, and other dust-gathering surfaces in modern homes. The gender of the name Electrolux is feminine. Comparison with other genera: Electrolux differs from all other narkids including Heteronarce in having large spiracular papillae (Fig. 5); very broad, elongated and flattened circumnarial flaps (Figure 6; flaps narrower and usually shorter in other narkids); prominent anterior nasoral turret; narrow thin lower lips without a mental groove (lower lips large and with a mental groove in other genera); small labial cartilages and labial folds ending well lateral to midline of mouth (Fig. 7) labial folds enlarged and meeting at midline of mouth in other genera); broad lateral tail folds (tail folds narrow in other narkids and obsolete in Typhlonarke); pectoral radials more numerous (47-53 versus 40-43 radials in other narkids); metapterygia not reduced and radials equal to propterygial radials (metapterygia shorter and their radials fewer than propterygial radials in other genera); claspers not extending past pelvic fin free rear tips (extending past pelvic fin tips in most narkids except Typhlonarke aysoni and some Narke capensis ); higher vertebral counts for most vertebral count groups (Table 6); chondrocranium with enlarged medial rostral cartilage and enlarged yoke-shaped lateral rostral cartilages (Fig. 9); rostral cartilages small and slender in other genera (as in Fig. 10); compressed, ventrally bent and expanded ethmoid region (laterally expanded in other taxa); antorbital cartilages with very narrow branched anterior section (broadly branched in most narkids); palatine processes of palatoquadrates straight, not curved medially; higher intestinal valve counts (17 in Electrolux vs. 8-10 in other genera); greater size (see discussion below); and unique coloration. Electrolux and Heteronarce are the only narkid genera with two dorsal fins. Two dorsals are primitive for torpedinoids and for batoids, but Narke and Typhlonarke have one dorsal fin in the position of the first dorsal fin of Electrolux and Heteronarce and presumably have lost their second dorsal fins, while Temera has no dorsal fins. Electrolux and Heteronarce also agree in their nearly parallel-edged anterior nasal flaps (nasal flaps more divergent in other narkids, particularly Typhlonarke and Narke), long and thinner circumnarial flaps (short and thick in other narkids), larger rounded-angular basibranchial copula (copula reduced and tack-shaped in Narke and Temera), and the lack of a frontoparietal fenestra (Narke, Temera, and Typhlonarke with prominent fenestra), nostrils with circular rather than elongate-oval incurrent apertures; a V - shaped posterior margin on its nasal curtain (nearly straight in Heteronarce); short, stout iliac processes on the pelvic girdle (Fig. 14A, iliac processes slender, long, curved and attenuated at least in H. bentuvai, H. garmani and H. mollis ); and more tooth rows (Table 4; total 32-34 vs 20-24 in Heteronarce). Electrolux additionally differs from Typhlonarke in having well-developed eyes (rudimentary in the latter genus and absent externally), a medial groove on its nasal curtain (absent in Typhlonarke), a trilobate V- shaped posterior edge on its nasal curtain (transverse and undivided in Typhlonarke), more tooth rows and smaller teeth (32-34, vs. 15-24 total in Typhlonarke), a stouter precaudal tail, undivided pelvic fins without leg-like anterior lobes, and longer more slender jaws ( jaws short and very stout in Typhlonarke). Electrolux differs from Narke in having shorter iliac processes on its pelvic girdle. Smithiana Bulletin 7: 15-49 22 Leonard J.V. Compagno and Phillip C. Heemstra Species: A single known species, Electrolux addisoni Compagno & Heemstra, described below. Electrolux addisoni sp. nov. Ornate sleeper-ray Figures 1-4, 5A, 6A, 7A, 8, 9A, 11A, 12A, 13A, 14A, 15, 16A, 17A, 18. Heteronarce ? sp. nov. Compagno, 1999b: 116. Undescribed genus and species. Compagno, 2005: 529. Holotype: SAIAB 78777, adult male, 515 mm TL, 305 mm disc width, Indian Ocean on reef off Manaba Beach near Margate, southern KwaZulu-Natal, 30°51.4' S, 30°23.1' E; depth 6-12 m; collected by Mark Addison, September 2003. Paratype: SAM 36908, adult male, 502 mm TL, 291 mm DW, locality the same as in holotype but collected by Brent Addison, October 2003. Description. Measurements in millimetres and proportions as percentages of total length (TL) are presented in Table 2 for the types of Electrolux addisoni. Disc subcircular, thick and fleshy; greatest width slightly more (1.1-1. 2 times) than its length. Edge of disc a continuous curve, anterior edge nearly straight but broadly curving to sides of disc, with snout, pectoral anterior margins, pectoral apices, and the posterior and inner pectoral margins not distinct but continuously curving rearwards, mesially, and recurving anteriorly to form broad but discrete free rear tips before merging with the sides of tail base. Snout short, preorbital length 6.4-8. 2 % TL, preoral length 6. 2-8. 4 % TL. Prespiracular head length 2.4-2. 7 times interspiracular width, preoral length about 1.8- 2.6 times as great as outer internarial width. A pair of small endolymphatic foramina on nuchal region of head about eye-length behind spiracles and about on anterior-posterior line tangent to their inner margins. Electric C- or narrowly kidney-shaped and about three times as long as wide; organs not obvious through skin of dorsal surface but prominent on white medial area of ventral surface; electric organ cells rounded-oval to hexagonal, count of cells for right organ approximately 289 on paratype. Vent anterior to pelvic inner margins and free from them, vent long and with broad lateral folds and conspicuous abdominal pores on the folds of the posterior third of vent. Tail stout, depressed, horizontally oval in cross section, with cutaneous lateral folds (about 7 mm wide) along lower part of tail from above rear end of pelvic fins and below midbase of first dorsal fin to base of caudal fin; tail moderately flattened below folds but broadly convex above; tail from vent to caudal tip about 1.3 times as long as snout-vent length; caudal peduncle nearly flat below folds, narrowly convex above them. Interdorsal space 75-106 % of first dorsal fin base; dorsal-caudal space 70 to 71 % of second dorsal fin base. Eyes well-developed, protruding above surface of disc, posterior to a transverse vertical plane at mouth; upper eyelids thickened, lower eyelids enlarged and apparently mobile, capable of covering palpebral apertures. Eyeball included in anterior edge of spiracle; dorsal rim of spiracle with 8 slender papillae (Fig. 5), their configurations symmetrical between right and left spiracles and arranged as follows: measurements of left spiracle papillae of holotype, followed by lengths of paratype's papillae in parentheses, first (anterior-most) papilla short, soft, recumbent on top of fleshy orbit, 2.5 mm (2.8 mm); second and third papillae (proceeding clockwise around spiracle rim) cuneate, flattened and stiff, lengths 4.4 (4.5) and 4.7 (3.8) mm respectively; fourth papilla stiff, flattened, resembling a forefinger and short thumb, lengths 6.1 (6.1) and 1.6 (2.0) mm respectively; fifth papilla short, simple, sausage¬ shaped, 2.6 (1.7) mm; sixth, seventh and eighth papillae long, slender, stiff, and finger-like, lengths 4.8 (4.1), 4.2 (3.5) and 8.0 (4.8) mm respectively. The spiracle papillae are bilaterally symmetrical for the left and right spiracles, with the three long, slender papillae on the lateral margins, short fleshy papilla on top of eyes, finger and thumb papilla on posterior rims and two cuneate papillae on medial edges of spiracles. Nostrils undivided, incurrent apertures with trough- or trumpet-shaped circum-narial folds (Fig. 6A), resembling a funnel cut in half vertically and separate from mouth; circumnarial folds somewhat expanded ventrally. Nasal curtain fleshy, elongated, with numerous small pores and a median longitudinal groove or sulcus on its ventral surface; distal (posterior) edge of curtain emarginate and trilobate, with a short, bifid fleshy barbel-like anterior nasal flap at each lateral corner with broadly rounded posterior ends, and a short barbel-like lobe on midline of inner surface ; with mouth closed, the nasal curtain extends over front of lower jaw. Each anterior nasal flap has a strong mesonarial flap above (dorsal to) its posterior tip. Excurrent apertures moderately large, hidden ventrally by nasal curtain and circumnarial folds but open broadly above them to posterior surface of snout in front of mouth, internarial space between inner ends of excurrent apertures hidden, internarial space about 3. 7-4.0 in mouth width when jaws are retracted and mouth closed. Gill openings small, width of fifth 0.4-0. 8 times width of first gill opening and 1.9-2. 8 in spiracular length; distance between inner ends of first pair of gill openings about 2. 9-3.0 times interspiracular width and about 1.5-1. 9 times distance between fifth pair. Mouth protrusile but scarcely distensible; mouth and jaw tips apparently forming a short tube when protracted and opened; lips thin, transversely-pleated, projecting (Fig. 7A) and surrounded by a shallow groove; lips pleated between lateral edges of dental bands and small labial folds; no prominent groove between labial folds, which include labial cartilages at Smithinna Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 23 Figure 7. Mouth with nasal curtain lifted of A. Electrolux addisoni, paratype, mouth opened. B. Heteronarce garmani , SAM 34813, 256 mm TL adult male, mouth closed, upper dental band partly obscured by debrisScale bars = 5 mm Abbreviations: EXA, excurrent aperture; LDB, lower dental band; LLF, lower labial fold; LLG, lower labial groove; LLP, lower labial pleats; NOG, nasoral groove; SYG, symphysial groove; UDB, upper dental band; ULG, upper labial groove; ULF, upper labial fold; ULP, upper labial pleats. corners of mouth; mouth width 3. 7-4.0 times in interspiracular distance and 1.5-1. 6 times spiracle length. Tooth bands occupy about half of mouth width, firmly connected to jaw cartilages by connective tissue, band widths greater than inner internarial distance. Teeth in quincunx arrangement, forming a tessellated pavement. Tooth formula 7-8~l~7 / 8~l~8-9, tooth row counts 15-16 / 17-18 rows or 32-34 total rows; ca. 6-8 series functional, as indicated by wear on crowns. Teeth very small, similar in both jaws and similar from symphysis to mouth corners, teeth in midline of mouth about 1 mm wide across crown foot and lateral teeth slightly smaller; labial surface of crown obliquely flattened, extending lingually as a single small stout cusp (Fig. 8) worn-off on most functional teeth); prominent basal ledge and groove on crown; roots small and lingually projecting, divided ventrally by a transverse groove; no teeth exposed on ventral surfaces of jaws when mouth is closed, but dental bands visible along anterior edge of jaws. Pectoral girdle crescentic in dorso-ventral view (Fig 13A). Anterior margin of coracoid cartilage concave medially but convex laterally. Propterygium a narrow segmented axis, articulates with procondyle of lateral face of scapulocoracoid, curves anteriorly, and is divided into 6 axial segments, the distal-most segment poorly calcified; mesopterygium elongated and unsegmented, articulating with the mesocondyle on lateral scapulocoracoid face and runs parallel to proximal segment of propterygium; metapterygium forming an axis of 4 segments, the proximal segment articulates with metacondyle on posterior corner of lateral face of scapulocoracoid and the distal segments extend rearwards into the pectoral base and free rear tip. 17 or 20 pectoral radials on propterygium, 12 or 13 on mesopterygium and 17 or 20 on metapterygium, total radials 46 or 53; no neopterygial radials articulating on lateral face of scapulo-coracoid; distal 2-4 segments of pectoral radials are bifid. Pelvic fins separate, deltoid-rounded anterior corner rounded, posterior (lateral) edge slightly convex, fairly thick and fleshy, pelvic fin origins well anterior to pectoral fin insertions; anterior edge stiff and thick, supported by the densely calcified anterior radial, apices broadly rounded, posterior (lateral) edge nearly straight or slightly convex when fins are fully spread; free rear tips well-developed, narrow, rounded-subangular, inner margin of pelvics distinct, not attached to sides of tail at free rear tips; no frenum between pelvic inner margins; pelvic fin length ~ 40- 43 % disc width. In live animals the pelvics are often broadly spread and appear rounded rather than angular. Pelvic girdle (puboischiadic bar) broad (Fig. 14A), its median antero-posterior width equals one sixth of girdle's transverse length; prepelvic process at each lateral end of girdle thin and weakly calcified extending anteriorly for the lengths of 7 centra. Iliac Smithiana Bulletin 7: 15-49 24 Leonard J.V. Compagno and Phillip C. Heemstra A B Figure 8. Lower replacement tooth (cusp unworn) from symphysis, removed from lingual end of dental band in the paratype of Electrolux addisoni, A. Lateral view. B. Labial view. C. Basal view. D. Lingual view. Abbreviations: BG, basal groove; BL, basal ledge; CF, central foramen; CFT, crown foot; CR, crown; PC, primary cusp; RT, root; TG, transverse groove. processes well-developed on pelvic girdle, expanded dorsally from lateral node, these short, stout, nearly straight, and angular; large basal condyle present on postero-lateral surface of each lateral node for articulation of long pelvic basipterygium, a short but prominent, blunt ischial process present on lateral node medio-posteriorly to basal condyle and mesial to basipterygium. 17 or 19 radials articulating with pelvic basipterygium, plus anterior enlarged radial (apparently a fused double radial) articulating with lateral node of pelvic girdle; the enlarged anterior radial projects postero-laterally and supports the leading edge of the pelvic fin. Claspers broad, flat and short (Fig. 15), nearly reaching free rear tips of pelvics ; clasper depth at base about 2.7 in base width and with a depressed elliptical cross-section; in dorso-ventral view claspers with parallel sides and a broad, bluntly rounded tip; in lateral view, dorsal and ventral surfaces nearly parallel along their lengths but with a short tapering posterior tip. Dorsal clasper groove open, with apopyle dorsolateral on base of clasper and hypopyle dorsal on glans; clasper glans with a long, low flap or cover rhipidion on the mesial edge of groove, a long slit-like pseudosiphon (slot of Leigh-Sharpe, 1922) mesial to cover rhipidion and near rear tip of clasper, and a long pseudopera (slit of Leigh-Sharpe, 1922) on postero¬ lateral tip of clasper. Dorsal and caudal fins compressed, flexible and close together; first dorsal fin origin above axil between inner edge of pelvic fin and clasper, or about opposite pelvic insertions; apex of first dorsal fin reaches past level of second dorsal fin origin; apex of second dorsal fin reaches well past caudal fin origin; dorsal fins similar in shape, but first dorsal fin slightly larger than second, second dorsal fin height 81- 86 % of first dorsal height, and second dorsal base 86-97 % of first dorsal base; anterior margin of dorsal fins strongly sloping and broadly convex, apices broadly rounded; posterior and inner margins weakly differentiated and convex; first dorsal base about 41-47 % of anterior margin; second dorsal fin base about 41% of anterior margin. First dorsal fin with 6 radials, second dorsal fin with 7; dorsal-fin radials divided into 3 or 4 segments. Caudal fin fairly high, short, and oval, its lower rear edge forming a continuous broad convex curve to tip of fin, upper margin slightly less convex, tip broadly rounded; vertebral column axis slightly raised; hypaxial lobe narrower than epaxial lobe; fin height 70-78 % of dorsal caudal margin, which is 1.5- 1 .6 times in distance from first dorsal origin to upper caudal origin. Caudal fin with 25 dorsal radials (radials 21 & 22 fused at the base) and 31 ventral radials, plus a V-shaped terminal radial. Antorbital cartilages short, directed anteriorly, with expanded posterior condyle for articulating with sockets on nasal capsules, with narrow shaft and moderately expanded, weakly branched anterior end, separated from anterior extensions of proterygia by ~ 4 cm; rear surface of chondrocranium with two prominent occipital condyles fitting two lateral sockets on anterior surface of synarcual. Hyomandibula large (Figs 11A & 12A), articulating directly to lateral edge of lower jaw (Meckel's cartilage). On radiographs (Fig. 12A; also in Heteronnrce, Fig. 12B), the synarcual seemed to be divided into anterior and posterior segments above the fourth gill arch, but detailed investigation of radiographs and dissection of the paratype revealed that the 'joint' between the two sections was the superimposed anterior edge of the basibranchial copula; a lateral strut on each side that seems to extend from the synarcual to the scapulocoracoid is the fifth ceratobranchial, which extends from the basibranchial to the scapulocoracoid. Vertebral counts are listed in Table 6. Spiral intestinal valve with 17 turns in the paratype (Table 7). Live colour. The spectacular and elaborate dorsal colour pattern of live Electrolux addisoni is shown in Figures 1 and 2. The dorsal surface and most of the ventral surface of the disc and pelvic fins, the claspers, the tail and median fins are dark brown covered with dense small pale spots which are white in the preserved types (Figs 3 and 4) but pale yellow in life. The pale spots are larger on middle of dorsal disc, where there are several scattered, short, pale streaks; also light streaks irregularly present in dark ventrolateral surfaces of disc and anteroventral surfaces of pelvic fins near apices. In life the ray is covered with mucus, and there are several curved. Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 25 Figure 9. Chondrocranium of Electrolux addisoni, paratype, in A. dorsal, B. ventral, C. lateral views. Abbreviations: ACS, articular socket for antorbital cartilage; AF, anterior fontanelle; AVF, foramen for orbitonasal vein draining nasal sinus; BP, basal plate; CR, cranial roof; FI I , fenestra for optic nerve; FIV, foramen for trochlear nerve; FIX, foramen for glossopharyngial nerve; FX, foramen for vagus nerve; FC, carotid foramina; FEN, foramen for endolymphatic duct; FM, foramen magnum; FOC, fenestra for superficial ophthalmic nerve; FPF, frontoparietal fenestra; FPN, fenestra for perilymphatic space; HF, hyomandibular facet (double); LR, lateral rostral cartilage; MR, medial rostral cartilage; NA, nasal aperture; NC, nasal capsule; NF, nasal fenestra; O, orbit; OC, occipital condyle; ONF, orbitonarial foramina; OPC, socket for optic pedicel; ORF, orbital fissure; OT, otic capsule; PTP, pterotic process. (some concentric) black stripes on the dorsal disc which disappear when the mucus is rinsed off; the light spots and streaks are often obscured by sediment when the ray is at rest. Ventral surface of disc and pelvic bases abruptly white in center, forming a pear- shaped symmetrical blotch from the nasoral turret to the vent and pelvic insertions and including the gill slits and most of the electric organs (Figs 3B, 4B). Etymology. The species is named for Mark Addison, Managing Director of Blue Wilderness dive charters of Widenham, KwaZulu-Natal. Mr Addison collected the holotype and instigated the capture of the paratype. He has an extensive knowledge of the marine fish fauna of South Africa and has provided much valued assistance in our fish survey research. Size: The holotype (515 mm TL adult male) weighed Figure 10. Chondrocranium of Narke capensis, Africana A12093 095 040 1035, 231 mm adult female, in A. dorsal, B. ventral, C. lateral views. Abbreviations as in Fig. 9. 1.8 kg., and the paratype (502 mm TL adult male) weighed 1.9 kg. Electrolux addisoni is apparently one of the largest members of the Narkidae, although females have yet to be examined, and the size range for maturation is unknown for males. Compagno & Last (1999a) noted that the Narkidae includes species that are adult at 9-46 cm TL and possibly reached a greater length, but Electrolux extends this to 52 cm, or possibly greater. Whitley (1940) stated that Typhlonarke aysoni reaches a far greater size than what is known for E. addisoni, with a maximum DW of about. 91 cm. and estimated TL over 122 cm. However, adult male specimens of T. aysoni examined by Whitley (1940), Garrick (1951) and by ourselves were only 21-38 cm. TL, and Garrick (1951: 5) repeated Whitley's comments but indicated that "most specimens taken are under 400 mm in total length". We wonder if Whitley's maximum DW figure (not his direct observation but based on the comment that "Graham records a maximum width of 36 inches") might be based on mistaken identity of the much larger Torpedo fairchildi for T. aysoni, although it is extremely difficult to mistake the two. The family Narkidae includes the smallest living batoids (Compagno et ai., 1999) and perhaps the smallest or at least the shortest, of the living chondrichthyans (Fig. 16). Males of Temera hardwickii examined by us are fully mature at 82-109 Smithiana Bulletin 7: 15-49 26 Leonard J.V. Compagno and Phillip C. Heemstra Figure 11. Narkid hyobranchial skeleton in ventral view. A. Electrolux addisoni , composite of SAIAB 78777 and SAM 36908, derived from radiograph of holotype and dissection of paratype. B, Narke japonlca , drawing modified and simplified from Garman (1913: pi. 67, fig. 3) in ventral view. Abbreviations: AOC, antorbital cartilage; BBC, basibranchial copula; CB1-5, cerato- branchials 1-5; CH, ceratohyal; HB1-2, hyobranchials 1 and 2; HMD, hyomandibula; LC, labial cartilages: LR, lateral rostral cartilages; MC, Meckel’s cartilage (lower jaw); MNC, mental or symphysial cartilage; MR, medial rostral cartilage; NC, nasal capsules; PQ, palatoquadrate (upper jaw); PSH, pseudohyoids; ROA, rostral appendix; SCP, scapular process of scapulo- coracoid (shoulder girdle); SPC, spiracular cartilage; SUS, supra- scapulae of scapulocoracoid; SYN, synarcual. mm TL and females at 105-148 mm TL; the species has been reported as reaching 458 mm TL. Adult males of a dwarf species of Narke from the Taiwan Straits are 99-109 mm TL and possibly reach 149 mm (see study material). In comparison to these tiny narkids, Electrolux addisoni is a giant, with the holotype 6.3 times longer and 139 times heavier than the smallest adult male T. hardivickii examined by us (82 mm TL and 13 grams). Available data for the smallest living sharks suggest that they may be almost twice as long as the smallest narkids. Compagno (1988: 24) discussed the minimum size of sharks and noted that although the diminutive dalatiid shark Squaliolus laticaudus is popularly considered the smallest living shark at a minimum adult length of 150 mm (Figure 16), it has several rivals of similar size. The proscylliid catshark Eridacnis radcliffei matures at 166-242 mm TL, with a 186 mm adult male weighing 14 gm, and a 242 mm pregnant female 37 gm. The etmopterid lantern shark Etmopterus carter i is adult at 186-212 mm, E. perryi is adult at 160-200 mm, E. polii is adult at 229-241 cm and probably smaller (a recently examined adult male is 195 mm long), and £. virens is adult at 181-257 mm. The dalatiid kitefin shark Euprotomicrus bispinatus is mature at 200-266 mm while the two Squaliolus spp. (S. aliae and S. laticaudus ) are mature at 15 or 20-25 cm Biological Notes. Electrolux addisoni belongs to the sleeper ray family Narkidae, based on the genus Narke and Greek narke, numbness, torpor, that alludes to the sluggish nature of these sedentary rays and the numbing effects of their electric organs. Electrolux addisoni however, is far from torpid while feeding on the substrate (Fig. 2A), and vigorously thrusts its mouth into loose sand or gravel while walking actively on its spread pelvic fins. It may lie motionless on the substrate, but when approached can arch its back, curl its disc, and raise its tail to perform a possible threat- display directed at the photographer (Fig. 2B). The stomach contents of the paratype included the semi- digested and fragmentary remains of approximately eight polychaete worms (including a tube-worm) and at least one small shrimp-like crustacean. Stomach contents weighed 5.6 grams. Electrolux as an infauna or meiofauna feeder agrees with the South African narkids Narke capensis, which mostly eats polychaetes (Compagno et al. 1989), and Heteronarce garmani (one specimen examined, Benguela G13531 88N 30-08, 127 mm immature female), which had a stomach filled with mud-balls. The conspicuous dorsal colour pattern of Electrolux addisoni combined with the ray's boldness and activeness near divers, and its possible threat display (Figure 2B) may be aposematic and indicates that the ray is well-armed with electric organs and should be avoided. On the shallow, well-lit reefs where E. addisoni has been found, its main potential predators may be large carcharhinoid sharks (requiem sharks, Carcharhinidae, and hammerheads, Sphyrnidae) and lamnoid sharks (ragged-tooth sharks and white sharks). Ebert (1990) found that two species of electric rays successfully defended themselves from much larger sixgill sharks ( Hexanchus ), which are apex predators with a broad prey spectrum. The small blind narcinid Benthobatis yangi from Taiwan was observed to repel a Hexanchus nakamurai. And two individuals of the larger torpedinid Torpedo cf. nobiliana from South Africa that were examined showed bite patterns of Hexanchus griseus, indicating that the sharks grabbed the rays but were repelled (probably with a shock) before they could inflict a strong and lethal bite. These Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 27 Figure 12. Narkid syncranium (neurocranium and splanchnocranium) and associated structures, monochrome inverted photographs from radiographs. A. Electrolux addisonl, holotype. B. Heteronarce garmani, BMNH 1921.3.1.3, 169 mm adult male holotype. C. Heteronarce mollis, CAS 58351, 218 mm TL female. D. Narke dipterygia, SU-41717, female ca. 135 mm TL. E. Temera hardwickii, CAS 58369, 108 mm TL female. F. Typhlonarke aysoni, SIO 61-149-6A, 92 mm TL immature female. Abbreviations: AOC, antorbital cartilage; CB5, 5th ceratobranchial (attaching basibranchial copula to scapulocoracoid; CVJ, occipito-cervical joint of chondrocranium and synarcual; HMD, hyomandibula; LC, labial cartilages; MC, Meckel’s cartilage; MNC, mental (symphysial) cartilage; NC, nasal capsules; PPR, palatine process of palatoquadrate; PQ, palatoquadrate; SCO, scapulocoracoid (shoulder girdle); SYN, cervicothoracic synarcual. incidents probably occurred in deep water where the sixgill sharks may have been using non-visual senses to locate their potential prey. Neither of these deep¬ water electric rays has a prominent colour pattern, and aposematic coloration and threat displays might be of little use to them in a visually limited environment. However, a quick defensive shock apparently can minimize damage by aborting a predator's attack. For the inshore Electrolux addisonl, aposematic coloration and a threat display might prevent a shark attack if the visual warnings are reinforced by a shock. There are other conspicuously marked inshore narcinids and torpedinids as well as the boldly marked narkid Heteronarce bentuviai, but their behaviour is for the most part poorly known. Distribution and Habitat: As presently known, Electrolux addisoni is endemic to the east coast of South Africa in warm-temperate or subtropical waters along approximately 310 km of coastline with a very narrow continental shelf (10-36 km wide to the 200 m isobath) but the few sightings were all inside the 50 m isobath. It could be more wide-ranging elsewhere in warm waters of southern Africa and even off East Africa. Known localities (Fig. 17, map) are mostly from dive sites off south-central kwaZulu-Natal (depth and Smitliiana Bulletin 7: 15-49 28 Leonard J.V. Compagno and Phillip C. Heemstra Figure 13. Narkid pectoral girdles (scapulocoracoids), radiographs showing left side in ventral view. A, Electrolux addisonl, holotype. B. Heteronarce mollis, CAS 58351, 218 mm TL female. C, Narke japonica, SU 61723, ca. 215 mm TL, female. D, Temera hardwickii, SU 35736, 105 mm TL, female. Abbreviations: ARS, articular surface; MSC, mesocondyle or mesopterygial condyle; MTC, metacondyle or metapterygial condyle; PRC, procondyle or propterygial condyle; SCP, scapular process; SLB, lateral bar; SUS, suprascapulae; TSS, tubular section; VC, vertebral column. Figure 14. Narkid pelvic girdles (puboischiadic bars). A, Electrolux addisoni, holotype (note pin posterior to pelvic girdle). B, Heteronarce mollis, CAS 58351, 218 mm TL female. C, Narke dipterygia, SU-41717, female ~ 135 mm TL. D, Temera hardwickii, SU 35736, 119 mm TL, female. Abbreviations: BPT, basipterygium; ILP, iliac process; ISP, ischial process; LPP, lateral prepelvic process; MPR, ribs on monospondylous precaudal vertebrae; PR1, first pelvic fin radial; PUI, puboischiadic bar; VC, vertebral column. Sinilhinna Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 29 Figure 15. Claspers of Electrolux addisoni, paratype: A, claspers and vent, ventral view; B, right clasper in dorsal view; C, clasper glans in dorsal view. Abbre-viations: AP, apopyle; CG, clasper groove; CGL, clasper glans; CRH, cover rhipidion; HP, hypopyle; PO, abdominal pores; PSP, pseudopera; PSS, pseudo¬ siphon; VN, vent. Scale bars = 10 mm. locality data in part from Koornhof, 1995). These include from southwest to northeast (Fig. 17, numbered 1-5): 1, Coffee Bay, Eastern Cape (ca. 31°58'S, 29°9'E; depth ~ 10 m); 2, Manaba Beach, the type locality near Margate (30°51.4'S, 30°23.TE, depth 6-12 m); 3, Protea Banks, about 8 km off Shelly Beach near Margate (ca. 30°49.8'S, 30°28.8'E, depths ~ 28-35 m); 4, Aliwal Shoal, 4.8 km. off Park Rynie (ca. 30° 19.2'S, 30°48'E, depths ~ 14-30 m.); 5, Tee (or T-) Barge north of Durban and about 3 km off Virginia Beach (an artificial reef habitat at ca. 29°47'S, 31°05'E, depths ~ 20-27 m.). Wallace's (1967) survey of electric rays from the east coast of southern Africa did not report this species. To our knowledge, this ray has not been seen south of Coffee Bay or along the south coast to False Bay. Electrolux addisoni was not taken by the South African Marine and Coastal Management's (MCM) research vessel Algoa during Cruise 014 off Mozambique in 1994 with 28 stations on soft bottom at 37-500 m depth. The Algoa collected torpedinoids in small numbers including Heteronarce garmani, Narcine rierai, and a second species of Narcine similar to the Malagasy N. insolita (Compagno in Smith & Heemstra, 1995, Compagno, 1999b, de Carvalho et al., 2002). The MCM research vessel Africana did not collect either E. addisoni or H. garmani in thousands of inshore and offshore bottom trawl stations at 17-200 m. during two decades of fisheries survey (ongoing) on the east coast of South Africa from Cape Agulhas to Port Alfred. However, Narke capensis were commonly caught by the Africana in this area at 27- 90 m depths (average depth 55 m) and apparently also occur in shallow inshore waters. M. Marks ( pers . comm.) caught N. capensis by hand while diving at 4.5 m in False Bay in the Western Cape Province. Electrolux addisoni occurs on the continental shelf on reefs with sandy or gravelly areas from close inshore to less than 50 m depth and including patches of appropriate habitat on inshore and offshore rocky banks and reefs. We wonder if this conspicuous ray is largely restricted to soft bottom patches on reefs off subtropical South Africa because of its having not been seen or collected elsewhere in southern Africa, including dive sites south of Coffee Bay and off Mozambique. Conservation Status: The conservation status of this ray is uncertain but worrisome, because it is only known from a few records to date on a heavily utilized narrow strip of habitat with extensive and intensive recreational diving and sport and commercial fishing, along with runaway coastal housing development. Its known habitat and geographic distribution suggest that Electrolux addisoni could be at risk from human activities including harassment and disturbance by divers, as well as fisheries, pollution, and habitat degradation. There are no known fisheries that target this species or include it as bycatch, although it might become of some interest in the aquarium fish trade, and it would make a spectacular aquarium exhibit provided one could collect live specimens and keep them successfully in captivity. In terms of its known range and area 'footprint' (perhaps not more than a few square kilometers), rarity, and exposure to human activities, this species might rank high on the IUCN Red List criteria for threatened species, possibly Critically Endangered (IUCN, 2006), although there are problems with ranking it with IUCN criteria, because data on trends in abundance are non-existent. The senior author suggests that species of electric rays with limited ranges in the tropical-subtropical southwestern Indian Ocean, particularly species of insular Torpedo and quite possibly narkids and Smithiana Bulletin 7: 15-49 30 Leonard J.V. Compagno and Phillip C. Heemstra C - Squaiiolus laticaudus Figure 16. Size comparisons of the possibly largest narkid. A, Electrolux addisoni (holotype at 515 mm TL) with B, the smallest adult male examined of the diminutive narkid Temera hardwickii (ZRC no number, 82 mm TL) and C, the adult male holotype of the dalatiid shark Squaiiolus laticaudus (USNM 70259, ca. 150 mm long, after Smith 1912). Figure 17. Digital map of South Africa showing localities for Electrolux addisoni (filled circles) and 200 m and 1000 m isobaths. Localities are numbered from southwest to northeast (1 from Eastern Cape, 2-5 from KwaZulu-Natal): 1, Coffee Bay; 2, Manaba Beach (type locality); 3, Protea Banks; 4, Aliwal Shoal; 5, Tee Barge, north of Durban. Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 31 Figure 18. Dorsal (A, C, E, G, I) and ventral (B, D, F, H, J) views of A-B. Electrolux addisoni , holotype. C-D. Heteronarce garmani, 145 mm TL adolescent male, RV Benguela G13531 88N 30-08. E-F. Narke capensis, 231 mm TL adult female, RV Africana A12093 095 040 1035. G-H. Temera hardwickii , SU 35728, 104 mm TL adolescent male. I-J. Typhlonarke aysoni, 354 mm TL adult male. Drawings mostly by the senior author and Elaine Heemstra, l-J modified from a drawing in Garrick (1951). Smithiana Bulletin 7: 15-49 32 Leonard J.V. Compagno and Phillip C. Heemstra Figure 19. Lateral views of A, Electrolux addisoni. B, Heteronarce garmam. C, Narke capensls. D, Temera hardwlckii. E. Typhlonarke aysoni. Same specimens and artists as in Fig. 18. narcinids, are a major (if limited) concern for conservation, as are offshore scyliorhinid catsharks of the genus Holohalaelurus in the area (Human, 2006). Electrolux addisoni would be an appropriate subject for a dive survey project by fish-watchers, professional ichthyologists and conservationists on the numerous dive sites of the northeast coast of the Eastern Cape Province, kwaZulu-Natal and Mozambique to attempt to better understand its distribution and estimate its abundance for the purposes of assessing its conservation status. TORPEDINOID CLASSIFICATION AND NOMENCLATURE The classification and nomenclature of torpedinoids used here follows Gill (1862, 1895) and Compagno (1973, 1999a, 2005). For discussion in the text below, detailed data on nomenclature, authors, dates, synonymies and modern equivalents are given for torpedinoid family groups (Table 3), torpedinoid genera (Table 4), and species of Narkidae (Table 5). Citations of families, genera and species below are abbreviated by omitting the authors and dates except in formal synonymies. The electric rays (suborder Torpedinoidei or order Torpediniformes) have long been recognized as very distinct from other batoids, with a variety of classifications including a separate order Torpediniformes (eg., Compagno, 1973, 1977) or suborder Torpedinoidea or Torpedinoidei (Bigelow & Schroeder, 1953, Compagno, 2005). Discussion of the higher classification and interrelationships of torpedinoids to other batoids is not elaborated here but has been considered by many authors. Suffice it to note that in modern studies torpedinoids are always regarded as monophyletic and distinct from other batoid groups and have numerous shared derived characters not found in other batoids. The classification of electric rays at the familial level had not varied much until the last half of the 20th Century and torpedinoids have usually been placed in a single family without subdivision. Bonaparte (1838) proposed a subfamily, Torpedinini, and Muller & Henle (1841) proposed a single family, Torpedines, for the four valid genera known at the time ( Astrape = Narke, Narcine, Torpedo and Temera). Torpedinidae or its numerous synonyms (Table 3), were utilized by 19th Century and most 20th Century authors for all electric rays. Gill (1862) proposed the division of his family Torpedinoidae (or Narcaciontoidae) for all electric rays into three subfamilies and four subgroups (equivalent to tribes) as follows: Subfamily Hypninae. Disc pyriform, formed by the union of the true disc with the ventrals, [pelvic fins] which are united beneath the tail; tail very short; head emarginated in front; spiracles far behind eyes; teeth with three points; dorsals [fins] two. Hypnos. Subfamily Narcaciontinae. Disc and tail nearly equal; head emarginated in front; spiracles far behind eyes; teeth transverse, with one point. Narcacion (= Torpedo), with two subgenera Narcacion and Tetronarce Subfamily Narcininae. Disc and tail nearly equally long; head entire or convex in front; spiracles close behind eyes; teeth rhombic or hexagonal. [Tribe] Discopygae. Dorsals two, ventrals [pelvic fins] united beneath the tail; teeth rhombic, acute behind; disc orbicular. Discopyge. [Tribe] Narcinae. Dorsals two, ventrals separated; teeth rhombic, with a median point. Narcine, with three subgenera Narcine, Cyclonarce, and Gonionarce. [Tribe] Astrapae. Dorsal [fin] single. Teeth rhombic, each with a median point; disc sub-circular. Astrape (= Narke). [Tribe] Temerae. Dorsal [fin] obsolete; teeth hexagonal and flat; disc sub-circular. Temera. Gill (1893) mentioned the division of his family Torpedinidae into three subfamilies, Torpedininae, Narcininae and Hypninae with tribes and generic allocations not specified. Fowler (1934, 1941) proposed the division of the family Torpedinidae into three subfamilies based on presence and number of dorsal fins, as with Gill's (1862) tribal classification within his Narcininae, but with the union of two-dorsal fin taxa in subfamily Torpedininae for Narcine, Heteronarce, Benthobatis, Hypnos and Torpedo; subfamily Narkinae for the one-finned Narke (including Bengalichthys) and Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 33 Typhlonarke, and subfamily Temerinae for the finless Temera. Bigelow & Schroeder (1953) followed Fowler's classification but raised the rank of his subfamilies to family (Temeridae, Narkidae, and Torpedinidae) and added Discopyge and Diplobatis to the Torpedinidae. Their classification was followed by a few subsequent authors, including Lindberg (1971) and Rass & Lindberg (1971). Bigelow & Schroeder (1953: 86) cautioned that Fowler's system based on the number of dorsal fins was artificial, and that it ran "counter to a dichotomous grouping based on the firmness of articulation of the upper and lower jaws and the presence or absence of labial cartilages, characters which are probably of greater importance phylogenetically than the number of dorsal fins." Bigelow and Schroeder retained Fowler's arrangement as a matter of convenience, as being appropriate to a general work, and because several of the torpedinoid genera had not been examined for jaw morphology. They expected that a torpedinoid classification based on jaw morphology would eventually replace the dorsal fin scheme. Fowler (1970) retained a single family Torpedinidae but included a modified arrangement of five subfamilies with a somewhat different composition than his earlier work and apparently incorporating elements of Gill's (1862) arrangement: Temerinae for the no dorsal fin Temera ; Narkinae for the one-dorsal Narke, Typhlonarke and Bengalichthys (= Narke); Discopyginae for the two- dorsal Discopyge; Hypninae for the two-dorsal Hypnos; and Torpedinidae for the one-dorsal Crassinarke (= Narke) and two-dorsal genera Benthobatis, Diplobatis, Heteronarce, Narcine, and Torpedo. Compagno (1973, 1977, fig. 12) proposed a revised classification of the electric rays based on anatomical comparisons of members of all the torpedinoid genera. The torpedinoids were divided into two well-defined superfamilies, Torpedinoidea and Narcinoidea, that verified and expanded the alternate dichotomous arrangement suggested by Bigelow & Schroeder (1953) as follows: Superfamily Narcinoidea: Mouth straight, with stout jaws; strong labial cartilages; rod-shaped hyomandibulae; well-developed ceratohyals; branched, antler-like antorbital cartilages; short and broad crania; well-developed frontoparietal fenestrae; rostrum present; occipital condyles not exserted; and disc rounded anteriorly. Families Narcinidae and Narkidae. Family Narkidae: shallow groove around mouth; narrow, rod-shaped rostrum; nasal capsules anteroventrally directed, contiguous, with a narrow internasal plate; precerebral fossa very small and terminated anteriorly by abrupt constriction of the rostrum; jaws short, stout and weakly protrusile; anterior hypobranchial elements and ceratohyals enlarged; posterior hypobranchials narrow and separated from each other by a wide space; and basibranchial copula small. Genera: Heteronarce, Narke, Temera, and Typhlonarke. Family Narcinidae. Deep groove entirely surrounding mouth and lips; rostrum broad, trough or shovel shaped; nasal capsules directed ventrolaterally, separated by a wide, flat internasal plate; precerebral fossa very large; jaws long, stout, and strongly protrusile; anterior hypobranchial elements and ceratohyals small; posterior hypobranchials very broad and nearly meeting mid-ventrally; and basibranchial copula moderately large. Genera: Benthobatis, Diplobatis, Discopyge, and Narcine. Superfamily Torpedinoidea: Mouth arcuate, very extensile, with extremely slender jaws; no labial cartilages; flattened triangular hyomandibulae; no ceratohyals; pinnate antorbital cartilages; crania elongated and narrow; with poorly developed frontoparietal fenestrae; rostrum absent or reduced; occipital condyles exserted; disc truncate and emarginate anteriorly. Families: Hypnidae and Torpedinidae. Family Hypnidae. Disc pear-shaped, tail rudimentary with two tiny dorsal fins and a small caudal fin; teeth tricuspid; "rostral appendices" articulating with cranium (Haswell, 1885); ethmoid region strongly bent ventrally, with nasal capsules expanded anteroventrally, contiguous, internasal plate narrow and compressed, not separating nasal capsules; preorbital processes absent; otic capsules large, with their outlines expanded abruptly from orbital walls; and one pair of hypobranchials articulating with basibranchial copula. Genus Hypnos. Family Torpedinidae. Disc subcircular, tail well- developed, with two moderate-sized dorsal fins and large caudal fin; monocuspid teeth; "rostral appendices" (Holmgren, 1941) fused to cranium; ethmoid region not bent ventrally, nasal capsules expanded laterally and separate, internasal plate wide and flat; preorbital processes present; otic capsules small, their lateral surface sloping gradually into orbital walls; two pairs of hypobranchials articulating with basibranchial copula. Genus Torpedo. This four-family torpedinoid classification was utilized by several authors including McEachran & Compagno (1982), Carroll (1988), Compagno (1990, 1999a,b, 2005), Last & Stevens (1994), McEachran et al. (1996), Compagno & Last (in Carpenter & Niem, 1999a, b, c), Compagno et al. (in Carpenter & Niem, 1999), and Carvalho et al. (in Carpenter & Niem, 1999). 'Downranked' variants of this classification include Nelson (1976, 1984) with a single family Torpedinidae and two subfamilies: Torpedininae and Narcininae with two tribes each. Torpedininae has the tribes Smithiana Bulletin 7: 15-49 34 Leonard J.V. Compagno and Phillip C. Heemstra Torpedinini and Hypnini while Narcininae has the tribes Narcinini and Narkini. Nelson (1994) modifies these to two families (Torpedinidae and Narcinidae) with two subfamilies each. Zhu & Meng (1979) used the families Torpedinidae and Narkidae for the two main groups, while Eschmeyer (1990, 1998) used two families Torpedinidae and Narcinidae. Shirai (1996) used one family Torpedinidae and three subfamilies, Torpedininae for Hypnos and Torpedo, Narkinae for the narkids Heteronarce, Narke, Temera, and Typhlonarke, and Narcininae for Benthobatis, Diplobatis, Discopyge, and Narcine. Our comparison of narkid genera with one another and with other torpedinoids reinforces the distinctiveness of the four families and their grouping into two higher groups (superfamilies) which corresponds to morphological similarities that reflect very different trophic specializations in the two superfamilies. Torpedinoidea feed on large prey that are stunned by the electric organs and swallowed whole through their distensible (snake-like) mouths and flexible jaws. Narcinoidea are bottom feeders that use their more or less protrusible jaws to feed on small prey on or in the substrate. Both groups can use their electric organs defensively against predators, but their utility in feeding is uncertain among the Narcinoidea. A detailed review of torpedinoid systematics and interrelationships of the families and a revision of the Narkidae (particularly Narke including describing new species) are needed, but these tasks are beyond the scope of this paper and more appropriate elsewhere. The present work concentrates on the systematics of Electrolux but has revealed numerous additional characters distinguishing the genera of Narkidae and provides materials for a future cladistic analysis of the genera that would be preliminary and inappropriate here without a revision of the family. KEY TO TORPEDINOID FAMILIES Key modified from Compagno et al. (in Carpenter & Niem, 1999) la. Mouth broadly arcuate, protrusile and greatly distensible; no labial folds and cartilages at corners of mouth . 2 lb. Mouth nearly straight, more or less protrusile but not much distensible; strong labial folds and cartilages at corners of mouth . 3 2a. Disc longer than wide, heart or pear-shaped; teeth tricuspid; tail much reduced; caudal fin about as high as dorsal fins . Hypnidae (Australia) 2b. Disc transversely elliptical, not pear-shaped; teeth monocuspid; tail not greatly reduced, caudal fin much higher than dorsal fins . Torpedinidae (Wide-ranging in all temperate and tropical seas) 3a. Snout firm, with broad, stiff, shovel-shaped rostral cartilage, readily felt by palpitation of snout; deep groove around mouth; teeth extending onto outer surfaces of jaws in most species . Narcinidae Wide-ranging in most temperate and tropical seas except for Eastern Atlantic. 3b. Snout soft, with a slender, rod-shaped rostral cartilage; shallow groove around mouth; teeth not extending onto outer surfaces of upper and lower jaws . Narkidae. (Eastern South Atlantic and temperate and tropical Indo-West Pacific from South Africa and Red Sea to Indonesia, Japan and New Zealand) FAMILY NARKIDAE FOWLER, 1934. SLEEPER RAYS Group Astrapae Gill, 1862: 387 (Family Torpedinoidae or Narcaciontoidae, subfamily Narcininae Gill, 1862). Type genus: Astrape Muller & Henle, 1837 (= Narke Kaup, 1826). Proposed as a group that is equivalent to tribe in rank for purposes of nomenclature. Subfamily Narkinae Fowler, 1934: 240 (Family Torpedinidae). Type genus: Narke Kaup, 1826. Replaces Astrapae Gill, 1862. Family Narkidae Bigelow & Schroeder, 1953: 87; and Narkidae Compagno, 1973: 41. Group Temerae Gill, 1862: 387 (Family Torpedinoidae or Narcaciontoidae, subfamily Narcininae). Type genus: Temera Gray, 1831. Subfamily Temerinae Fowler, 1934: 240 (Family Torpedinidae), Family Temeridae Bigelow & Schroeder, 1953: 87. Compagno (1973: 41) synonymized Temeridae with Narkidae although Temerae Gill, 1862 has priority. Family Diagnosis (derived and expanded from Compagno, 1973, 1977, 1999a and Compagno & Last in Carpenter & Niem, 1999a): Electric rays with short preorbital snouts, 5-13% TL; snout broadly rounded anteriorly or nearly truncate ( Electrolux and some Heteronarce species). Spiracles contiguous with posterior edges of eyeballs, not situated behind them; margins of spiracles usually smooth and flat or with a low ridge, elevated and occasionally with a few low papillae in Narke capensis, or with several long and prominent papillae that screen the spiracles ( Electrolux ). Nasal curtain elongated, narrow, posteriorly expanded and thickened, with prominent ampullal pores on its ventral surface. Mouth transverse, small and narrow, not highly distensible; shallow circumoral groove surrounding mouth and lips; labial folds and grooves strong. Tooth row counts low, 8-17 / 7-21 or 15-38 total rows; teeth concealed when mouth is closed; teeth small, rounded-oval, unworn crowns with keels or a single low blunt, broad cusp. Tail fairly large and stout but variably short to moderately elongated. Disc circular, ovate, rounded- Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 35 angular or pear-shaped, often about as broad as long. Pectoral girdle crescentic; tubular section of lateral faces of scapulocoracoids greatly elongated, longer than medial fenestrated section. Suprascapulae V- or C- shaped, with fused midline above free vertebrae behind synarcual. Superscapulae articulating with scapulae entirely in front of coracoid bar or crossing above bar. Metapterygial axis subequal to propterygial axis or much shorter; propterygial radial count equal to metapterygial radial count or much more numerous. Puboischiadic bar with prominent short to greatly elongated iliac processes. Two dorsal fins (. Electrolux , Heteronarce ) or one dorsal fin ( Narke , Typhlonarke) , or dorsal fin absent ( Temera ). Caudal fin larger than dorsal fins or fin (when present) and subequal to or somewhat smaller in size than pelvic fins ( Typhlonarke with pelvic fins fused to disc); caudal fin without prominent ventral lobe. Cranium short and fairly broad; rostrum incomplete, medial floor reduced to narrow medial rostral cartilage and paired lateral rostral cartilages below it, lateral walls of rostrum truncated around precerebral cavity and fused to ethmoid region of cranium; medial rostral cartilage rod-shaped with short bifurcated rostral node, less than one-third nasobasal length. Rostral appendices small, separate from rostrum, just lateral to rostral node, and articulating with antorbital cartilages. Precerebral fossa very small and terminated anteriorly by an abrupt constriction of the lateral walls of the rostrum. Ethmoid region anteriorly directed or strongly bent ventrally ( Electrolux ); nasal capsules expanded anteroventrally or laterally; nasal cartilages (ala nasalis) greatly expanded posterolaterally from the nasal capsules to support the expanded and thickened nasal curtain; internasal plate narrow, compressed, narrowly separating the nasal capsules. Antorbital cartilages more or less branched and antler¬ like (narrower distally in Electrolux and in Heteronarce garmani than in other narkids examined); bases of cartilages articulating on posterolateral surfaces of nasal capsules; shafts of cartilages directed more or less anterolaterally; a lateroposteriorly directed spur or process present or absent on shafts of antorbital cartilages. Preorbital processes apparently absent. Cranial roof perforated by frontoparietal fenestra (Narke, Temera, Typhlonarke) or not ( Electrolux , Heteronarce); when present long, U-shaped and contiguous with anterior fontanelle or separated from it by an epiphysial bridge. Otic capsules large and broad, length about 33-40% of nasobasal length, width across capsules about 59-68% of nasobasal length, capsules rather inflated and expanded laterally, distally rounded-angular; lateral outlines of otic capsules sloping gradually into orbital walls. Occipital condyles relatively short and low, not strongly exserted from occiput. Jaws stout and transverse, weakly protrusile. Palatoquadrates thick, straight and subtriangular, with strong overlapping processes on their articulation with Meckel's cartilages; orbital cartilages obsolete to strong on palatoquadrates. Meckel's cartilages very stout, flat and broad, strongly expanded symphysially and distally, with a weak to strong distal process. A large flat oval mental cartilage just posterior to lower symphysis in most genera (possibly absent in Typhlonarke). Well-developed upper and lower labial cartilages present and close to symphyses, dorsal labial cartilage simple or forked. Hyomandibulae heavy, elongated and sigmoid¬ shaped, with expanded bases and apices; ceratohyals well-developed and large, about size of anterior hypobranchials. Prespiracular cartilages present but postspiracular cartilages absent. Two pairs of hypo¬ branchials, the posterior pair articulating with basibranchial copula; anterior hypobranchials large, as broad or broader than posteriors but not as long; posterior hypobranchials broad to very narrow and separated from each other by a narrow ( Electrolux ) to wide space (Narke, Temera, with Heteronarce intermediate). Basibranchial copula small and tack¬ shaped (Narke and Temera), larger and more rounded- angular in Electrolux and apparently Heteronarce. Dorsal surface usually brownish or reddish-brown, white or brownish below; dorsal surface either plain or with a few large dark spots or blotches, paired white spots, and white side bands on the tail and posterior pelvic bases, usually without a complex colour pattern or ocelli on pectoral fins (Electrolux exceptional with its elaborate colour pattern). Dwarf to moderate-sized batoids, adult males 82-515 mm TL and possibly longer (see size discussion for Electrolux, above). Genera of Narkidae and status of Heteronarce: Heteronarce, Narke, Temera, and Typhlonarke were included in the Narkidae (Compagno, 1973), to which we add Electrolux as the second genus of two-dorsal narkid. We initially considered Electrolux addisoni as a possible species of Heteronarce but it soon became apparent that it was very different from any species of Heteronarce and that the four valid species of Heteronarce formed a coherent genus that is separable from Electrolux and other narkids. We present meristic data of Heteronarce and other narkids (Tables 6, 7 & 8) and morphometric data (Table 9) for Heteronarce and and can easily distinguish Electrolux and Heteronarce from each other and from other narkids. Figures 18 and 19 summarize external differences within narkid genera. Figures 6 and 7 show differences in mouth and nostril structures, Figures 5, 9, 10, 11 and 12 differences in head anatomy. Figure 13 pectoral girdle structure and Figure 14 pelvic girdle structure. There has been considerable confusion in the literature on the status, familial position, and species of Heteronarce. Some authors have doubted its distinction from Narcine and have synonymized the two genera or have mistaken species of Heteronarce for Narcine or vice versa. Part of the problem is that Smithiana Bulletin 7: 15-49 36 Leonard J.V. Compagno and Phillip C. Heemstra Heteronarce was originally defined on a few nasoral characters that although partially valid were not seen as definitive by some authors, particularly after a few species of Narcine were discovered with elongated, relatively narrow, Heteronarce- like nasal curtains. The first valid species of Heteronarce was described by Lloyd (1907) as Narcine mollis from the Gulf of Aden at 238 m. This was distinguished from the Indian species N. timlei by its enlarged anterior nasal valves (circumnarial folds) and more elongated nasal curtains about as long as wide (about three times wider than long in N. timlei). Lloyd (1909) and Annandale (1909) gave additional information on N. mollis including illustrations of the whole ray, teeth, and oral anatomy. Garman (1913) included N. mollis in Narcine without comments. The genus Heteronarce was proposed by Regan (1921) for a second new species of two-dorsal electric ray, H. garmani, collected about 15-22 miles off the Umvoti River, kwaZulu-Natal, South Africa, in 120- 130 fms depth, and for Narcine mollis Lloyd, 1907. Heteronarce was distinguished from Narcine by "the minute nostrils, the length of the anterior nasal valves, which are confluent to form a curtain that is not much broader than long and is studded with pores, and the lateral position of the posterior nasal valves." (Regan, 1921). Regan noted that H. garmani was very similar to H. mollis but had much smaller eyes and spiracles, a longer snout, a smaller mouth and nasal valves, and different coloration (brown above and white below in H. garmani, dark brown above and gray-brown below in H. mollis). Regan did not assign a type species for Heteronarce, but this was subsequently designated as H. garmani by Fowler (1941). Von Bonde & Swart (1923) described a third species, Heteronarce regani from a 190 mm specimen from kwaZulu-Natal, South Africa (two stations mentioned, depth 211-329 m), but didn't compare or distinguish it from H. garmani which they also recognized. Their specimen (pi. 22, fig. 2) had a truncated caudal fin tip that is unusual compared to the rounded caudal tips of other Heteronarce species and might be abnormal. The genus Heteronarce was not characterized by these authors. Fowler (1925a) described a fourth species of Heteronarce as Narcine natalensis, from a 260 mm specimen trawled from off kwaZulu-Natal at 40 fathoms, but didn't refer to Regan's or von Bonde & Swart's accounts and didn't compare it with either H. garmani or H. regani. Fowler (1925b) subsequently synonymized his N. natalensis with H. garmani, and later (Fowler, 1941) included both N. natalensis and H. regani in synonymy of H. garmani. Fowler's (1941) synonymy was recognized by Wallace (1967) and is followed here. Fowler (1941) included H. mollis and H. garmani in Heteronarce and separated this genus from Narcine by its more elongated nasal curtain. Bigelow & Schroeder (1953) included Heteronarce and Narcine in their Family Torpedinidae as separate genera, but noted that Heteronarce (including H. garmani and H. mollis) was so close to Narcine that its generic validity was doubtful, and differed only in its more elongated nasal curtain. However Compagno (1973) noted that Heteronarce "had two dorsals, and had usually been placed in the vicinity of Narcine, but examination of its exterior and skeleton revealed its affinity with Narke. McKay (1966) described Narcine westraliensis from Western Australia which has an unusually elongated nasal curtain for a narcinid that resembles that of Heteronarce. McKay used this similarity to synonymize Heteronarce with Narcine. However, the chondrocranium and oral structure of N. westraliensis as described by McKay is like that of other Narcine species and of narcinids and unlike that of Heteronarce and other narkids. Talwar (1981) reviewed Heteronarce and named a fifth species from the southwest coast of India, H. prabhui. Talwar recognized three valid species, including H. garmani from southern Africa (with synonyms H. regani and Narcine natalensis), H. mollis from the Arabian Sea, and H. prabhui from India. Talwar defined Heteronarce as having the disc rounded anteriorly, a shallow groove around the mouth, jaws short and weakly protractile, eyes well developed and almost contiguous with spiracles, nasal curtain only slightly broader than long, and two dorsal fins. Baranes & Randall (1989), described Narcine bentuviai from 80-200 m. in the Gulf of Aqaba, Red Sea, which proved to be a sixth and unusual species of Heteronarce. Heteronarce bentuviai has a unique bold and black-blotched color pattern and a second dorsal fin noticeably smaller than the first dorsal. H. mollis, H. garmani, and H. prabhui in contrast are uniformly colored and have second dorsal fins about as large as the first. According to Baranes & Randall's detailed account, the external morphology and anatomy of H. bentuviai agrees with that of other narkids and particularly with other species of Heteronarce but not with Narcine or other narcinids. These authors tentatively placed their species in Narcine because they considered Heteronarce a junior synonym following McKay (1966), but noted that an alternate generic arrangement might remove it from Narcine. De Carvalho (1999) excluded H. bentuviai from Narcine in his revision of the genus. Lloris & Rucabado (1991) described a seventh species, Heteronarce rierai from off Mozambique, but this proved to be a narrow-bodied Narcine with a narrow nasal curtain as in N. westraliensis but like that species has mouth and anatomical characteristics typical of narcinids (Compagno, in Smith & Heemstra, 1995, de Carvalho, 1999). There are several other narrow-bodied species of Narcine in Australian waters (MacKay, 1966, Last & Stevens, 1994, de Carvalho, 1999). The superficial external similarity of Heteronarce to Narcine is contradicted by its anatomical and external Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 37 differences (particularly in the oral and chondrocranial morphology) which are similar to those of other narkids rather than Narcine or other narcinids. Based on our examination of Heteronarce specimens, as well as literature data, we confirm that Heteronarce is a valid and well-defined genus including the four species H. bentuviai, H. garmani, H. mollis, and H. prabhui, with the latter species morphologically very similar to H. mollis. Heteronarce is separable from narcinids by characters in the key to families and definition of the Narkidae above, and separable from other narkids in the key to narkid genera and in comparison with Electrolux. KEY TO NARKID GENERA la. No dorsal fins . Temera (Indo-West Pacific from Andaman Sea near southern Thai-Burma border through Straits of Malacca and Malay Peninsula to Singapore, Thailand and Viet-Nam, doubtful from Philippines) lb. One or two dorsal fins . 2 2a. One dorsal fin . 3 2b. Two dorsal fins . 4 3a. Eyes not visible externally; anterior lobes of pelvic fins form isolated leg-like structures protruding from ventral surface of pectoral disc, posterior lobe of pelvic fins fused to pectoral disc . Typhlonarke (New Zealand) 3b. Eyes usually visible externally; no separate pelvic fin lobes . Narke (Southeastern Atlantic and Indo-West Pacific, South Africa, northern Arabian Sea, India, Malaysia, Singapore, Indonesia, Thailand, Viet- Nam, Philippines (doubtful), China, Taiwan, Japan, Korea) 4a. Spiracles with long, slender, stiff papillae (Fig. 5 A); nostrils and mouth projecting ventrally as prominent nasoral turret near front of disc (Figs 3B; lower lips thin, without a chin (mental) groove and with labial cartilages not meeting at midline (Fig. 7 A & 12A); tooth rows numerous, 32-34 total; dorsal side of disc with numerous pale spots on dark brown background, live ray with elaborate concentric pattern of black lines and pale spots; underside of disc white medially, the broad distal margin dark brown, with numerous small pale spots . Electrolux, (Western Indian Ocean, South Africa) 4b. Spiracles without papillae (Fig. 5B); nostrils and mouth not projecting as a prominent nasoral turret at front of disc, more posterior and slightly projecting from ventral surface of disc; lower lips thick, with prominent mental groove and labial cartilages meeting at symphysis (Figs 7B & 12B); tooth rows fewer, 20-24 total; uniform pale to dark brown or grayish above, without markings or with a few large black blotches on disc, first dorsal fin and caudal fin, white, grayish or gray- brown below . Heteronarce (Western Indian Ocean, South Africa to Gulf of Aqaba, Arabian Sea and west coast of India) MATERIAL EXAMINED Institutional abbreviations: Institutional abbreviations for specimens of Narkidae mostly follow Compagno (1988): BMNH - Natural History Museum, London, UK, formerly British Museum (Natural History). CAS - California Academy of Sciences, San Francisco, California, USA. GVF - George Vanderbilt Foundation fish collection, Stanford University, Stanford, California, USA, housed at the California Academy of Sciences. ISH - Institut fur Seefischerei, Hamburg, Germany. KUMF - Kasetsart University. Faculty of Fisheries, Museum, Bangkok, Thailand. LJVC - XXXX and LJVC - YYMMDD (Year, Month, Day), L.J.V. Compagno accession number and field- accession number. PCH - Phillip C. Heemstra field numbers. NMS - National Museum of Singapore (formerly Raffles Museum) zoology collection, housed in the Department of Zoology, National University of Singapore. SAIAB - South African Institute of Aquatic Biodiversity, formerly RUSI, for the J.L.B. Smith Institute of Ichthyology, Grahamstown, South Africa. SAM - Iziko - South African Museum, Natural History Division, department of Marine Biology, Cape Town, South Africa. SIO - Scripps Institution of Oceanography, La Jolla, California, USA. SU - Stanford University fish collection, Stanford, California, USA, housed at the California Academy of Sciences. USBCF - United States Bureau of Commercial Fisheries, Department of the Interior field number. Now US National Marine Fisheries Service, Department of Commerce. ZRC - Zoological Reference Collection, Department of Zoology, National University of Singapore. COMPARATIVE MATERIAL - FAMILY NARKIDAE 'Crassinarke dormitor' (? = Narke japonica), SIO 4-257-6B, 278 mm TL adult male. Yellow Sea. Heteronarce garmani: BMNH 1921.3.1.3, holotype, adult male, 169 mm TL, 77 mm DW, 1921, Umvoti River, kwaZulu- Natal, South Africa, 220-238 m. SAM 34813, two adult males, 256-289 mm TL and 124-135 mm DW, RV Algoa, C00813 014 011-3115, 19940612, Western Indian Ocean, Mozambique, 23° 28.0' S, 35°43.00' E, 185 m. SAM uncataloged, two females, 125-132 mm TL and 65-72 mm DW, adolescent male, 145 mm TL and 64 mm DW, RV Benguela G13531 88N 30-08, 19880822, kwaZulu-Natal, South Africa, 29°44.0' S, 31°23.00' E, 154 m. Heteronarce mollis: CAS 58352, 206 mm TL, 92 mm DW adult male, RV Anton Bruun, AB 9-444, 19641216, N. Indian Ocean, Somalia, 9° 36.00'N, 51° l.OO'E, 78-82 m. CAS 58351, 6 females, 217, 226, 255, 218, 211 and 212 mm TL, 115, 120, 122, 111, 108 and 100 mm DW, 2 adult males, 220 and 199 mm TL, 118 and 100 mm DW; immature male, 165 mm TL and 90 mm DW, RV Anton Bruun, AB 9-464, 19641218, N. Smithiana Bulletin 7: 15-49 38 Leonard J.V. Compagno and Phillip C. Heemstra Indian Ocean, Somalia, 11° 37.00'N, 51° 27.00'E, depth? ISH 254/75, adult male, 199 mm TL, 100 mm DW, Dr. F. Nansen, FAO, 19750303, North coast, N. Indian Ocean off Somalia, 11°41.00' N, 51°36.00'E, 82 m. Nnrke capensis: (all from South Africa): Africana A04752 048 030-1039, immature males, 87, 94, 98 mm TL, 52, 53, 57 mm DW, adult males, 206, 250 mm TL, 142, 165 mm DW, immature female, 89 mm TL, 49 mm DW, adult (?) females, 205, 215, 223, 233, 255 mm TL, 133, 133, 137(?), 153, 164 mm DW, adult females, 193, 208, 220, 220, 224, 232, 234, 245 mm TL, 120, 139, 136, 145, 146, 156, 146, 173 mm DW, 19860920, southeastern Cape coast, , 34°16.0' S, 22°1.0'E, depth 42 m. Africana A04770 048 043-1093, adult males, 176, 237, 245 mm TL, 119, 153, 154 mm DW, immature female (?), 155 mm TL and 101 mm DW, females, 145, 165, 171, 192, 218 mm TL, 96, 108, 105, 128, 140 mm DW, 19860923, southeastern Cape coast, , 33°48.0'S, 26° 7.0' E, depth 56 m. Africana A06215 056 023-1042, adult males, 243 and 253 mm TL, 154 and 153 mm DW, adult females, 196 and 213 mm TL, 115 and 120 mm DW, 19870917, southeastern Cape coast, , 34°7.0'S, 22° 15.0'E, depth 40 m. Africana A07115 063 011-2123, adult male 243 mm TL and 172 mm DW, 19880513, southeast Cape coast, 34°30.0'S, 21°14.0'E, depth 59 m. Africana A07116 063 012 2132, male (adult?) 253 mm TL, 155 mm DW, 19880513, East coast cruise. Cape coast, 34°31.0'S, 21°17.0'E, depth 58 m. Africana A07128 063 020-1039, immature female, 118 mm TL, 75 mm DW, adult females, 225, 229, 246, 254, 263 mm TL, 136, 143, 156, 162,169 mm DW, adolescent male, 257 mm TL, 168 mm DW, adult males, 243, 253 mm TL, 173 mm DW (253 mm TL), 19880515, southeastern Cape coast, 34°16.0'S, 22°1.0'E, depth 35 m. Africana A07152 063 038-1093, adult males, 175, 178, 215, 255 mm TL, 117, 123, 134, 170 mm DW, adult females, 155, 175 mm TL., 95, 112 mm DW, 19880519, southeastern Cape coast, 33°45.0'S, 26°5.0' E, depth 32 m. Africana A13339 102 014-2332, immature male 73 mm TL, 50 mm DW, 1992040, southeastern Cape coast, 33°53.0'S, 26°46.0'E, depth 90 m. Africana A16338 122 043-1042, immature male, 105 mm TL, 68 mm DW, adult female, 176 mm TL, 109 mm DW, 19940618, eastern Cape coast, 34°9.0'S, 22°13.0'E, depth 39 m. SAM 34347, adult male, 174 mm TL, 118 mm DW, Africana A18168 135 077-1090, 19960429, southeastern Cape coast, , 33°45.0'S, 26°1.0'E, depth 27 m. depth. LJVC 961014, SAM uncataloged, adult male, 265 mm TL, 180 mm DW, M. Marks site 1, 19961013, Buffels Bay, False Bay, Western Cape, , 34°19.06'S, 18° 27.75'E, depth 4.5 m. RUSI 11932, adult male, 195 mm TL, 115 mm DW , TBD- 3, off Swartkops, Eastern Cape, . SAM 22796, female, 188 mm TL, 120 mm DW, 19590919, Algoa Bay, Eastern Cape, . SAM 22799, female 170 mm TL, 101 mm DW, 19590916, Sandy Point, Mazeppa Bay, Eastern Cape, 32°27.00'S, 28°39.00'E. SAM 30992, adult male, 239 mm TL, 156 mm DW, 19781103, Muizenberg, False Bay, Western Cape. Narke dipterygia: BMNH 1909.7.12.13, female, 121 mm TL, 57 mm DW, syntype of Bengalichthys impennis Annandale, 1909, Balasore Bay, Orissa coast, India. CAS uncataloged, female ca 120 mm TL, 50 mm DW, Indo-West Pacific. CAS 66840, female, 146 mm TL, 78 mm DW, collected by J. Mee, 19890222, Sudah, Oman, 6 m depth. LJVC 0508, five females, 170, 150, 150, 162 and 155 mm TL, immature male, 165 mm TL, F. Steiner, 19751006, Taiwan straits?, Taiwan. LJVC 0514, 157 mm TL adult (?) female, F. Steiner, 19751006(?), Taiwan straits, Taiwan? LJVC 0515, adult (?) females, 146 and 151 mm TL, F. Steiner, 19751006, Taiwan straits, Taiwan(?). J. Randall uncataloged, female, 131 mm TL, India. NMS 3119, adult males, 150 and 131 mm TL, 84 and 70 mm DW, A. K. Tham, 1964, Fisheries Biology Unit, SPR 425, Singapore. SU 32406, adult male, ca. 135 mm TL, 65 mm DW, India. SU 41717, adult male ca 150 mm TL and female ca 135 mm, India. USBCF F.H. Berry no number, two adult males, 170 and 180 mm TL, five females, 178, 153, 166, 130, 153 mm TL, Porto Novo, Madras, India. USBCF F.H. Berry, SOSC-381, female ca 160 mm TL, Porto Novo, Madras, India, 15-22 m. depth, Narke cf. dipterygia, KUMF 2464, female, 120 mm TL, 75 mm DW, 19740105, Samut Songkram, Thailand,. KUMF 0807, female, 123 mm TL, 69 mm DW, 1968, collected by P. Wongrat, Prachuat Khiri Khan, Gulf of Thailand, KUMF 0834, females 109, 114 and 112 mm TL, 64, 66, 66 mm DW, 1972021-10, collected by P. Wongrat, offshore near Sataheys, Gulf of Thailand. Narke japonica: CAS Acc. 1972:1:5, female, 233 mm TL, 19711200, East. China Sea, between Japan and Korea. GVF Naga 60-61 (GVF-2077), adult male ca 225 mm, 19600227, South China Sea, 15°40.00'N, 109°25.50'E. PCH Jan. 1988, female 168 mm TL, 93 mm DW, 1988, from Tachi, South China Sea, Taiwan. SU 3363, adult male, ca. 365 mm TL, Japan. SU 61723, female ca 215 mm TL, Japan. SU 7267, adolescent male, ca 195 mm TL, Japan. Narke spp. CAS Acc 1972-XII: 18, immature male, 67 mm TL, adolescent males, 94 and 109 mm TL, adult males, 99 and 103 mm TL, females, 58 and 93 mm TL, southwest of Kao Ksiung, Taiwan. Anton Bruun, AB 4B-263, immature male, 100 mm TL, adult males, 143 and 147 mm TL, females, 117, 121, 127, 127, 132, and 139 mm TL, Arabian Sea, 22°54'N, 68°06'E. GVF-2430, female ca 150 mm TL, Gulf of Thailand, Thailand. GVF-2449, adult males, 153 and 154 mm TL, female, ca. 125 mm TL, Gulf of Thailand, Thailand. GVF- 2663, female ca 150 mm TL, Gulf of Thailand, Thailand. CAS Anton Bruun AB 4B-223A, females, 130, 150 and 165 mm, Arabian Sea, 22°54'N, 68°06'E, depth 16 m. Temera hardwickii: BMNH 1953.8.10.9-10, female, 139 mm TL, 73 mm DW, and adult male, 108 mm TL, 55 mm DW, from Penang, Malaysia, syntypes of Temera hardivickii Gray, 1831. CAS 58369, adolescent males, 74 and 84 mm TL, 44 and 49 mm DW, female, 108 mm TL, 59 mm DW, Anton Bruun Sta. 0248, 9°54'N, 97°42'E, Andaman Sea, SSW of Kawthaung, Isthmus of Kraa, Burma, 0-200 m. KUMF 0014, female, 136 mm TL, 86 mm DW, 19650823, Phuket (Puket), Thailand, depth 91-105 m. KUMF 2916, females, 78 and 142 mm TL, 36 and 85 mm DW, adult or late adolescent male, 88 mm TL, 44 mm DW Songkhla Research Vessel, S. Mongkolprasit, 198412, Thailand. NMS 2090, adult female, 148 mm TL, 76 mm DW, M. Tweedie, identified by A.W.C.T. Herre, 1940, fish market, Singapore. NMS 2110, adult (?) female, 140 mm TL, 81 mm DW, R.L. Chermin, 1954, det. E.R. Alfred, 680615, Changi, Singapore. SU 35728, adolescent male, 104 mm TL, 64 mm DW, A.W.C.T. Herre, 1936-1937, Telok Kurau, Perak, Malaysia, SU 35736, immature female, 74 mmTL, 38 mm DW, adolescent female, 105 mm TL, 62 mm DW, adult female, 119 mm TL, 62 mm DW (cranium dissected), A.W.C.T. Herre, Singapore. ZRC no number, adult male, 82 mm TL, 48 mm DW, no locality data or other Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 39 information, presumably Singapore, ZRC no number, female (possibly adult), 121 mm TL, 72 mm DW, Ahmad Draman, 19630811, Ponggol, Singapore. ZRC 10588, adult male, 109 mm TL, 66 mm DW adult male, Ahmad Draman, 19640801, Ponggol, Singapore. ZRC 38918, adult female, 122 mm TL, 78 mm DW, K. Lim, P.K.L. Ng, et al., June 1995, Pulau Bintan, north coast, Tanjung Tondang, Sumatra, Indonesia. Typhlonarke aysoni: LJVC 0424, adult male, 205 mm TL, 110 mm DW, Kaikoura coast. South Island, New Zealand, 110 m. SIO 61-149-6A, immature female, 92 mm TL, 48 mm DW, 19610129, NW of Mernoo Bank, South Island, New Zealand, 43°39.0'S, 175°15.0'E, 0-119 m. ACKNOWLEDGEMENTS We thank Mark Addison and his father Brent Addison for their efforts in collecting and donating the two specimens of Electrolux addisoni to the South African Institute for Aquatic Biodiversity. We are grateful for the images of the live fish supplied by Stefania and Peter Lamberti, Dennis King, Avril and Len Fish, and Peter Chrystal through Rudy van der Elst. We thank Elaine Heemstra for her assistance with the drawings. Thanks also to Mark Marks, Mike Bougaardt and Liz Hoenson of Iziko-South African Museum. The first author thanks the curators and collection managers at the following institutions for allowing him to examine narkid specimens in their collections: Natural History Museum, London, UK; California Academy of Sciences, San Francisco, California, USA; Institut fur Seefischerei, Hamburg, Germany; Kasetsart University. Faculty of Fisheries, Museum, Bangkok, Thailand; National Museum of Singapore (formerly Raffles Museum) zoology collection and Zoological Reference Collection in the Department of Zoology, National University of Singapore; Scripps Institution of Oceanography, La Jolla, California, USA. Smithiana Bulletin 7: 15-49 40 Leonard J.V. Compagno and Phillip C. Heemstra Table 1. Morphometric abbreviations and definitions for torpedinoids Code Description Code Description ANF anterior nasal flap length, from anterior edge of incurrent apertures to rear end of nasal curtain IG5 distance between medial ends of fifth gill slits CDM dorsal caudal margin, upper caudal fin origin to rear tip of fin INO least inter-orbital width between eyeballs CH greatest vertical height of caudal fin INS least width between spiracles CLB width across base of clasper INW inter-narial width, least distance between excurrent narial apertures CLI clasper medial length from anterior end of vent IOW inter-narial outer width, nasal curtain base width at incurrent apertures CLO clasper length, from pelvic fin base to clasper tip MOL mouth length, mid upper lip to line joining corners COH height of exposed cornea MOW mouth width, distance between mouth corners COL anterior - posterior length of exposed cornea NOW nostril width from lateral edge of incurrent aperture to medial edge of excurrent aperture CPH peduncle height at upper caudal fin origin PCS pelvic fin insertion to ventral origin of caudal fin CPW caudal peduncle width at upper caudal origin PDI least distance between verticals at pelvic fin insertion and first dorsal fin insertion CVM ventral caudal fin margin, from ventral fin origin to rear tip PDO pelvic fin origins to first dorsal fin origin DCS dorsal caudal space, D2 insertion to upper caudal fin origin PDI pre-DI , snout tip to first dorsal fin origin DL disc length, snout tip to line at rear tips of disc PD2 pre-D2, snout tip to second dorsal fin origin DT greatest thickness at middle of disc PGL prebranchial length, snout tip to level of first gill slits DW greatest transverse width across disc PGW snout tip to level of greatest disc width D1 A D1 anterior margin from origin to fin apex PIW body width at pectoral fin insertions DIB first dorsal fin base, from origin to insertion POB pre-orbital, snout tip to line at front edge of eyes D1H D1 height, vertical distance from fin base to apex POR pre-oral length, snout tip to front edge of mouth Dll D1 inner margin, from insertion to rear edge PP2 pre-pelvic length, snout tip to level of pelvic fin origins D1L D1 length, from origin to rear edge of fin PRC pre-caudal length, from snout tip to dorsal caudal fin origin DIP D1 posterior margin from apex to free rear tip PRN pre-narial length, midline snout tip to level of nostrils D2A D2 anterior margin from origin to fin apex PSP pre-spiracular length, snout tip to level of spiracles D2B length D2 base, from origin to insertion P1I pectoral disc insertion to free rear tip D2H D2 height, vertical distance from base to apex P2A pelvic fin anterior margin from fin origin to apex D2I D2 inner margin from insertion to rear edge of fin P2B pelvic fin base from pelvic fin origin to insertion D2L D2 length, from origin to rear edge of fin P2H pelvic fin height, perpendicular distance from base to apex of fin D2P D2 posterior margin, from apex to free rear tip P2I pelvic fin inner margin, pelvic fin insertion to rear tip ESL distance from rear edge of eyeball to spiracle P2L pelvic fin length, from pelvic fin origin to rear tip EH height of protruding eyeball P2P pelvic fin postero-lateral margin, fin apex to rear tip EL anterior-posterior length of protruding eyeball P2S pelvic fin span, distance between pelvic fin apices GS1 width from medial to lateral ends of first gill slit SPL anterior-posterior length (diameter) of spiracle GS2 width of second gill slit SPW lateral-medial width (diameter) of spiracle GS3 width of third gill slit SVL snout tip to anterior end of vent GS4 width of fourth gill slit TBH tail base vertical thickness at pelvic fin insertions GS5 width of fifth gill slit TBW tail base width at pelvic fin insertions HL head length, tip of snout to level of fifth gill slits TL total length, from snout tip to rear end of caudal fin IDS interdorsal distance, D1 insertion to D2 origin VNL anterior-posterior length of vent (cloaca) IG1 distance between medial ends of first gill slits Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 41 Table 2. Measurements in mm and proportions as %TL for type specimens of Electrolux addisoni. Abbreviations for measurements are defined in Table 1. Holotype Paratype Holotype Paratype TL 515 mm %TL 502 mm %TL mm %TL mm %TL PRC 425 83 410 82 GS3 10 1.9 9 1.8 DW 305 59 291 58 GS4 10 1.9 9 1.8 DL 264 51 255 51 GS5 5 1.0 7 1.4 DT 46 8.9 53 11 IG1 50 9.7 51 10 PRN 37 7.2 41 8.2 IG5 34 6.6 27 5.4 POR 32 6.2 42 8.4 VNL 24 4.7 22 4.4 POB 42 8.2 32 6.4 PIW 99 19 80 16 PSP 46 8.9 40 8.0 TBH 32 6.2 36 7.2 PGL 79 15 71 14 TBW 62 12 55 11 HL 132 26 127 25 CPH 15 2.9 15 3.0 PGW 137 27 129 26 CPW 13 2.5 16 3.2 PP2 194 38 186 37 P1I 25 4.9 27 5.4 SVL 225 44 221 44 P2L 123 24 126 25 PD1 300 58 282 56 P2A 81 16 82 16 PD2 369 72 357 71 P2B 82 16 81 16 IDS 33 6.4 27 5.4 P2H 63 12 68 14 DCS 21 4.1 22 4.4 P2I 48 9.3 51 10 PCS 144 28 142 28 P2P 116 23 116 23 PDO 132 26 62 12 P2S 202 39 197 39 PDI 64 12 58 12 CLO 22 4.3 24 4.8 COL 9 1.7 9 1.8 CLI 102 20 95 19 COH 5 1.0 5 1.0 CLB 9 1.7 10 2.0 EL 11 2.1 12 2.4 D1A 75 15 76 15 EH 10 . 1.9 8 1.6 DIB 31 6.0 36 7.2 INO 24 4.7 24 4.8 D1H 56 11 57 11 NOW 12 2.3 12 2.4 Dll 16 3.1 18 3.6 INW 6 1.2 5 1.0 D1L 52 10 53 11 IOW 18 3.5 16 3.2 DIP 44 8.5 39 7.8 ANF 17 3.3 15 3.0 D2A 73 14 75 15 SPL 14 2.7 13 2.6 D2B 30 5.8 31 6.2 SPW 13 2.5 12 2.4 D2H 48 9.3 46 9.2 INS 17 3.3 17 3.4 D2I 21 4 23 4.6 ESL 0 0.0 0 0.0 D2L 46 8.9 43 8.6 MOL 3 0.6 5 1.0 D2P 35 6.8 37 7.4 MOW 22 4.3 20 4.0 CDM 81 16 80 16 GS1 12 2.3 9 1.8 CVM 58 11 66 13 GS2 10 1.9 9 1.8 CH 63 12 56 11 Smithianci Bulletin 7: 15-49 42 Leonard J.V. Compagno and Phillip C. Heemstra Table 3. Torpedinoid family-group taxa assigned to the two superfamilies of Compagno (1973), and sorted into currently recognized families. Superfamily Narcinoidea (Narcinidae + Narkidae) Compagno 1973 Taxon Original Family Current Family Tribe (group) Astrapae Gill, 1862 Narcaciontoidae Narkidae Tribe (group) Discopygae Gill, 1862 Narcaciontoidae Narcinidae Subfamily Discopyginae Gill, 1895 Narcobatidae Narcinidae Tribe (group) Narcinae Gill, 1862 Narcaciontoidae Narcinidae Family Narcinidae Compagno, 1973 Narcinidae Subfamily Narcininae Gill, 1862 Narcaciontoidae Narcinidae Subfamily Narcininae Gill, 1893 Torpedinidae Narcinidae Family Narkidae Bigelow & Schroeder, 1953 Narkidae Family Narkidae Compagno, 1973 Narkidae Subfamily Narkinae Fowler, 1934 Torpedinidae Narkidae Tribe (group) Temerae Gill, 1862 Narcaciontoidae Narkidae Family Temeridae Bigelow & Schroeder, 1953 Narkidae Subfamily Temerinae Fowler, 1934 Torpedinidae Narkidae Superfamily Torpedinoidea (Hypnidae + Torpedinidae) Compagno 1973 Subfamily Hypninae Gill, 1862 Narcaciontoidae Hypnidae Subfamily Hypninae Gill, 1893 Torpedinidae Hypnidae Family Hypnidae Compagno, 1973 Hypnidae Family Narcaciontoidae Gill, 1862 Torpedinidae Subfamily Narcaciontinae Gill, 1862 Narcaciontoidae Torpedinidae Family Narcobatidae Gill, 1895 Torpedinidae Subfamily Narcobatinae Gill, 1895 Narcobatidae Torpedinidae Subfamily Torpedinae Fowler, 1934 Torpedinidae Torpedinidae Family Torpedines Muller & Henle, 1841 Torpedinidae Family Torpedinidae Owen, 1866 Torpedinidae Family Torpedinidae Bigelow & Schroeder, 1953 Torpedinidae Family Torpedinidae Compagno, 1973 Torpedinidae Subfamily Torpedinini Bonaparte, 1838 Rajidae Torpedinidae Subfamily Torpedinini Gill, 1893 Torpedinidae Torpedinidae Family Torpedinoidae Gill, 1862 Torpedinidae Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 43 Table 4. Torpedinoid genera and synonyms. Genus group names that are apparently incorrect subsequent spellings of previously described names are designated by an asterisk (*). Citations for incorrect subsequent spellings are not included in Literature Cited list. Superfamily Narcinoidea Superfamily Torpedinoidea Genus Valid genus Genus Valid genus Astrape Muller & Henle, 1837 Narke Eunarce Fowler, 1910 Torpedo Bengalichthys Annandale, 1 909 Narke Fimbrio torpedo Fritsch, 1886 Torpedo Benthobatis Alcock, 1 898 Benthobatis Gymnotorpedo Fritsch, 1886 Torpedo Crassinarke Takagi, 1951 Narke Hypnarce Waite, 1 902 Hypnos Cyclonarce Gill, 1 862 Narcine Hypnarea* Sharp, 1903 Hypnos Diplobatis Bigelow & Schroeder, 1948 Diplobatis Hypos* Cappetta, 1988 Hypnos Discopyge Heckel in Tschudi, 1846 Discopyge Hypnos Dumeril, 1 852 Hypnos Gonionarce Gill, 1862 Narcine Narcacion Gill (Klein) 1862 Torpedo Heteronarce Regan, 1 92 1 Heteronarce Narcobatis* Blainville, 1825 Torpedo Narcina* Jordan & Seale, 1 905 Narcine Narcobatus Blainville, 1816 Torpedo Narcine Henle, 1834 Narcine Notastrape Whitley, 1932 Torpedo Narcinops Whitley, 1940 Narcine Tetranarce* Gill, 1895 Torpedo Narke Kaup, 1 826 Narke Tetronarce Gill, 1 862 Torpedo Syrraxis Jourdan in Bonaparte, 1835 Narcine Tetronarcine* Tanaka, 1908 Torpedo TemeraG ray, 1831 Temera Torpedo Houttuyn, 1764 Torpedo Temerara Tirant, 1929 Temera Torpedo Dumeril, 1 806 Torpedo Typhlonarke Waite, 1909 Typhlonarke Table 5. Family Narkidae: species and synonyms, including name changes and one species (Heteronarce rierai ) originally allocated to the Narcinidae Nominal species Valid species Astrape aysoni Hamilton, 1902 Typhlonarke aysoni Narcine bentuviai Baranes & Randall, 1989 Heteronarce bentuviai Raja capensis Gmelin, 1788 Narke capensis Raja dipterygia Bloch & Schneider, 1801 Narke dipterygia Crassinarke dormitor Takagi, 1951 ? = Narke japonica Heteronarce garmani Regan, 1 92 1 Heteronarce garmani Temera hardwickii Gray, 1 83 1 Temera hardwickii Temerara hardwickii Tirant, 1929 Temera hardwickii Bengalichthys impennis Annandale, 1909 Narke dipterygia Torpedo japonica Temminck & Schlegel, 1850 Narke japonica Narcine mollis Lloyd, 1907 Heteronarce mollis Narcine natalensis Fowler, 1925 Heteronarce garmani Heteronarce prabhuiTalwar, 1981 Heteronarce prabhui Heteronarce regani von Bonde & Swart, 1 924 Heteronarce garmani Heteronarce reirai Lloris & Rucabado, 1 99 1 Narcine reirai Typhlonarke tarakea Phillipps, 1929 ? Typhlonarke tarakea Smithiana Bulletin 7: 15-49 44 Leonard J.V. Compagno and Phillip C. Heemstra Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 45 Table7. Narkid intestine spiral valve counts (number of turns); total lengths in mm. Species Specimens Count Electrolux addisoni: SAM-36908, adult male, 502 mm 17 Heteronarce garmani: SAM-34813, adult male, 256 mm 9 Heteronarce garmani. SAM-34813, adult male, 289 mm 8 Narke capensis. A 12093 +040-1035, adult female, 231 mm 10 Narke capensis: A14742 1 1 1 010-2074, adult male, 215 mm 10 Typhlonarke aysoni: LJVC-0424, adult male, 208 mm 10 Table 8. Tooth row counts of narkids; total lengths in mm; upper jaw rows / lower jaw rows. Species Specimens and counts Electrolux addisoni SAIAB - holotype adult male, 515 mm; 15/17 Electrolux addisoni SAM - paratype, adult male, 502 mm; 16/18 Heteronarce bentuvai HUJ - holotype, adult male, 191 mm; 11/11 Heteronarce garmani SAM-34813, adult male, 256 mm; 11/11 Heteronarce garmani SAM-34813, adult male, 289 mm; 11/10 Heteronarce mollis Lloyd (1907): 10- 12/ 10- 12 Heteronarce prabhuai Talwar (1981): Types: 10-12/11-12 Narke capensis A 12093 adult female, 231 mm; 16/14 Narke dipterygia SU? adult female, 164 mm; 17/21 Typhlonarke aysoni Garrick (1951): 10-12/10-12 Typhlonarke aysoni LJVC-0424, adult male, 205 mm; 10/11 Typhlonarke aysoni SIO-61-149-6A, immature female, 92 mm; 8/7 Typhlonarke tarakea Garrick (1951): 11/11 Table 9. Measurements of Heteronarce spp. in %TL for H. garmani and holotypes of H. bcntuvai (after Baranes & Randall, 1989) and H. prabhui (after Talwar, 1981) garmani garmani garmani garmani garmani garmani bentuvai prabhui SAM- SAM- Benguela Benguela Benguela BMNH- HUJ ZSI 34813 34813 30-98 30-98 ' 30-98 1921.3.1.3 13612 F7614/2 Holotype Holotype Holotype Adult Adult Adolescent Juvenile Immature Adolescent Adult Adult male male male female female male male male TL 256 289 145 132 127 164 191 220 PRC 82 80 83 80 84 85 79 79 DW 48 43 48 52 51 42 48 51 DL 47 45 50 49 50 49 48 49 DT 13 12 13 12 12 9.8 PRN 10 10 9.9 8.0 11 12 3.7 POR 12 12 12 10 13 10 7.3 11 POB 9.4 11 13 10 13 13 8.4 11 PSP 12 13 15 13 15 19 13 PG1 17 17 18 18 19 30 16 HL 29 28 31 31 31 24 26 PGW 31 32 36 37 35 43 PP2 44 41 42 43 43 49 SVL 47 47 49 51 50 45 49 46 PD1 61 61 63 63 64 58 57 PD2 71 70 72 70 72 74 71 67 IDS 4.4 4.4 3.1 2.4 1.6 5.8 4.7 4.8 DCS 3.4 2.0 4.2 2.7 2.4 5.5 3.7 PCS 23 24 23 15 15 27 PDO 21 19 20 18 19 23 PD1 11 9.6 9.7 5.5 6.3 11 COL 1.7 1.6 2.0 1.8 1.9 1.2 COH 1.0 0.6 1.0 1.0 0.2 0.6 EL 3.2 2.7 3.1 4.2 3.6 1.8 2.6 EH 3.0 2.1 2.4 2.0 1.9 1.2 INO 5.1 5.5 4.7 6.3 6.3 7.9 8.9 NOW 1.8 1.3 1.4 2.3 2.0 2.1 INW 1.2 0.9 1.6 1.2 1.6 0.6 IOW 3.3 3.6 3.8 3.1 3.4 5.2 ANF 3.4 2.8 2.9 2.2 2.4 1.8 3.7 SPL 1.3 1.7 1.4 1.0 1.3 1.2 2.1 SPW 1.7 1.7 1.5 1.4 0.9 1.2 INS 6.5 6.9 6.4 7.0 7.9 6.7 6.8 6.1 ESL 0.4 0.3 0.3 0.7 0.0 0.3 1.0 MOL 0.0 0.3 0.0 0.0 0.0 1.2 MOW 3.0 2.9 3.4 3.5 1.8 4.3 5.2 4.8 GS1 1.5 1.0 1.6 1.2 1.3 1.2 Smithiana Bulletin 7: 15-49 46 Leonard J.V. Compagno and Phillip C. Heemstra Table 9 continued GS1 1.5 1.0 1.6 1.2 1.3 1.2 GS2 1.1 1.1 1.3 1.4 1.3 1.8 GS3 1.3 1.1 1.5 1.4 1.2 1.8 GS4 1.3 1.2 1.7 1.3 1.1 1.8 GS5 1.4 1.1 1.6 1.3 0.9 1.2 IG1 10 9.3 12 11 12 12 13 15 IG5 6.4 6.2 6.4 6.4 6.3 7.3 8.9 8.2 VNL 5.6 3.3 3.4 4.5 3.9 4.9 PIW 20 17 20 24 22 TBH 5.3 5.8 6.2 6.7 5.5 6.7 TBW 11 9.9 12 13 11 11 CPH 3.4 3.0 3.0 2.5 2.4 3.0 CPW 4.0 3.8 4.3 4.1 4.7 3.0 Pll 43 41 43 47 0.0 41 P2L 5.2 2.7 5.9 6.1 0.0 0.0 P2A 22 21 21 24 24 20 P2B 13 14 12 11 11 11 11 P2H 16 17 17 20 21 15 P2I 9.7 9.2 9.6 10 15 9.1 P2P 6.2 4.7 6.9 6.1 3.9 4.0 1.6 P2S 16 15 16 17 17 14 19 CLO 34 30 35 33 33 28 CLI 5.2 4.2 3.0 5.5 11 CLB 20 21 15 15 D1L 2.5 2.4 2.1 2.4 D1A 7.6 8.6 8.5 7.8 10 7.9 DIB 9.3 12 9.1 10 11 9.5 9.4 D1H 5.3 6.3 5.2 5.1 5.8 ■ 5.5 7.3 5.7 Dll 6.6 5.7 4.0 6.1 5.5 5.5 6.8 DIP 2.3 2.6 2.7 2.4 1.6 3.4 D2L 4.9 3.9 4.2 5.2 3.9 4.9 D2A 7.7 9.4 8.3 9.2 8.7 8.5 D2B 11 13 10 12 11 11 6.8 D2H 6.1 6.9 5.5 6.3 5.5 4.9 3.1 5.7 D2I 7.1 5.6 4.9 6.2 6.3 4.3 1.6 D2P 1.8 2.4 2.0 2.4 1.6 5.5 CDM 5.0 4.2 4.6 4.9 4.7 4.6 CVM 17 18 17 17 16 15 24 CH 12 14 13 13 13 13 18 Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 47 LITERATURE CITED Annandale, N. 1909. Report on the fishes taken by the Bengal fisheries steamer "Golden Crown". Part I. - Batoidei. Memoirs of the Indian Museum, 2(1): 1-60. Baranes, A., & J.E. Ramdall. 1989. Narcine bentuviai, a new torpedinoid ray from the northern Red Sea. Israel Journal of Zoology, 36: 85-101, pi. 1, fig. 1-9. Bigelow, H.B., & W.C. Schroeder. 1953. Chapter one. Sawfishes, Guitarfishes, Skates and Rays. In Fishes of the Western North Atlantic. Memoirs of the Sears Foundation for Marine Research (1) 2: 1-514. Bonaparte, C.L. 1838. Selachorum tabula analytica. Nuovi Annali delle Scienze Naturali, Bologna, series 1, 2: 195- 214. Carroll, R.L. 1988. Vertebrate paleontology and evolution. W.H. Freeman, New York, pp. i-iv, 1-698. Compagno, L.J.V. 1973. Interrelationships of living elasmobranchs. In P. FI. Greenwood, R. S. Miles and C. Patterson (editors). Interrelationships of fishes. Zoological Journal of the Linnaean Society, London, Supplement 1, 53: 15-61. Compagno, L.J.V. 1977. Phyletic relationships of living sharks and rays. American Zoologist, 17(2): 303-322 Compagno, L.J.V. 1984. FAO Species Catalogue. Vol. 4, Sharks of the World. An annotated and illustrated catalogue of shark species known to date. FAO Fisheries Synopsis No. 125, 4(1): i-viii, 1-250, 4(2): i-x, 251-655. Compagno, L.J.V. 1986. Family Narkidae, in M.M.Smith & P.C. Heemstra, eds. Smith's Sea Fishes, 1st edition, first impression. Macmillan South Africa Publishers, Johannesburg, South Africa, 113-114. Compagno, L.J.V. 1988. Sharks of the Order Carcharhiniformes. Princeton University Press, Princeton, New Jersey, pp. i- xxii, 1-572. Reprinted by Blackburn Press, 2003, pp. i-xxii, 1-572, with new forward. Compagno, L.J.V. 1990. Evolution and diversity of sharks. In S. H. Gruber, ed., Discovering sharks. A volume honoring the work of Stewart Springer. Underwater Naturalist, Bulletin of the American Littoral Society, 19-20(4/1): 15- 22, figs Compagno, L.J.V. 1995. Additions to and revisions of Family Narkidae and Narcinidae, in M.M.Smith & P.C. Heemstra, eds. Smith's Sea Fishes, 1st edition, third impression (revised). Southern Book Publishers, Johannesburg, South Africa, vi-vii. Compagno, L.J.V. 1999a. Chapter 1. Systematics and body form. Chapter 3. Endoskeleton. Appendix. Checklist of living elasmobranchs. In W. C. Hamlett, ed.. Sharks, skates and rays. The biology of elasmobranch fishes. Johns Hopkins Press, pp. 1-42, 69-92, 471 to 498, figs. 1.1-1.15, 3.1-3.8. Compagno, L.J.V. 1999b. An overview of chondrichthyan systematics and biodiversity in southern Africa. Transactions of the Royal Society of South Africa, 1999 54(1): 75-120, fig. 1-9, tab. 1-2. Compagno, L.J.V. 2001, Sharks of the World. Volume 2. Bullhead, mackerel and carpet sharks (Heterodontiformes, Lamniformes and Orectolobiformes). An annotated and illustrated catalogue of the shark species known to date. FAO Species Catalogue for Fisheries Purposes (1): i-v, 1-269, figs. 1-163, 160 maps. Compagno, L. J. V. 2005. Chapter 16. Checklist of living Chondrichthyes. In W. C. Hamlett, ed., Reproductive biology and phylogeny of Chondrichthyes. Enfield, New Hampshire, USA, Science Publishers, Inc.: 503-548. Compagno, L.J.V., M. Dando & S. Fowler. 2005. A Field Guide to the Sharks of the World. First printing, first impression. Harper-Collins, London. Pp i-vii, 9-368, monochrome figs. 1-60, color pis. 1-64. Compagno, L.J.V., D.A. Ebert & M.J. Smale. 1989. Guide to the sharks and rays of southern Africa. Struik Publishers, Cape Town, 160 pp. Compagno, L.J.V., & P.R. Last. 1999a. Order Torpediniformes. Narkidae. Sleeper rays. In Carpenter, K.E. and V.H. Niem (eds), 1999. FAO species identification guide for fishery purposes. The living marine resources of the Western Central Pacific. Rome, FAO, 3: 1443-1446, figs. Compagno, L.J.V., & P.R. Last. 1999b. Order Torpediniformes. Hypnidae. Coffin rays. In Carpenter, K.E. and V.H. Niem (eds), 1999. FAO species identification guide for fishery purposes. The living marine resources of the Western Central Pacific. Rome, FAO, 3: 1447-1448. Compagno, L.J.V., & P.R. Last. 1999c. Order Torpediniformes. Torpedinidae. Torpedos. In Carpenter, K.E. and V.H. Niem (eds), 1999. FAO species identification guide for fishery purposes. The living marine resources of the Western Central Pacific. Rome, FAO, 3: 1449-1451, Compagno, L.J.V., P.R. Last, B. Seret, & M. De Carvalho. 1999. Batoid Fishes. In Carpenter, K.E. and V.H. Niem (eds), 1999. FAO species ide?itification guide for fishery purposes. The living marine resources of the Western Central Pacific. Rome, FAO, 3: 1398-1409, figs. Compagno, L.J.V., & T.R. Roberts. 1982. Freshwater stingrays (Dasyatidae) of Southeast Asia and New Guinea, with description of a new species of Himantura and reports of unidentified species. Environmental Biology of Fishes, 7(4): 321-339. Compagno, L.J.V., & T.R. Roberts. 1984. Marine and freshwater stingrays (Dasyatidae) of West Africa, with description of a new species. Proceedings of the California Academy of Sciences, series 4, 43: 283-300, fig. 1-10. De Carvalho, M.R. 1999. A systematic revision of the electric ray genus Narcine Henle, 1834 (Chondrichthyes, Torpediniformes: Narcinidae), and the higher level-phylogenic relationships of the orders of elasmobranch fishes (Chondrichthyes). Unpublished Ph.D. Thesis, City University of New York, pp. i-xxxii, 1-704, fig. 1-116. De Carvalho, M.R., L.J.V. Compagno, & P.R. Last. 1999. Order Torpediniformes. Narcinidae. Numbfishes. In Carpenter, K.E. and V.H. Niem (eds), 1999. FAO species identification guide for fisherx/ purposes. The living marine resources of the Western Central Pacific. Rome, FAO, 3: 1433-1442, figs. De Carvalho, M.R., B. Seret, & L.J.V. Compagno. 2002. A new species of electric ray of the genus Narcine Henle, 1834 from the southwestern Indian Ocean (Chon¬ drichthyes: Torpediniformes: Narcinidae). South African Journal of Marine Science, 24: 135-149, figs. 1-7, tab. 1. Ebert, D.A. 1990. The taxonomy, biogeography and biology of cow and frilled sharks (Chondrichthyes: Hexanchiformes). Unpub. Ph.D. thesis, Rhodes University, Grahamstown, 308 pp. Eschmeyer, William N. 1990. Catalog of the genera of Recent fishes. California Academy of Sciences, San Francisco, 697 pp. Eschmeyer, William N. 1998. Catalog of fishes. Vol. 1, Introductory materials. Species of fishes (A-L): 1-958. Vol. Smithiana Bulletin 7: 15-49 48 Leonard J.V. Compagno and Phillip C. Heemstra 2. Species of fishes (M-Z): 959-1820. Vol. 3. Genera of Fishes. Species and genera in a classification. Literature Cited. Appendices: 1821-2905. Also CD-ROM with Access database. Fowler, H. W. 1925a. Descriptions of three new marine fishes from the Natal coast. Annals of the Natal Museum, 5(2): 195-200. fig. 1-3. Fowler, H. W. 1925b. Fishes from Natal, Zululand and Portuguese East Africa. Proceedings of the Academy of Natural Sciences, Philadelphia, 1925, 77: 187-268, fig. 1-4. Fowler, H. W. 1934. Descriptions of new fishes obtained 1907 to 1910, chiefly in the Philippine Islands and adjacent seas. Proceedings of the Academy of Natural Sciences, Philadelphia, 85: 233-367 Fowler, H.W. 1941. The fishes of the groups Elasmobranchii, Holocephali, Isospondyli, and Ostariophysi obtained by United States Bureau of Fisheries Steamer Albatross in 1907 to 1910, chiefly in the Philippine Islands and adjacent seas. Bulletin of the United States National Museum (100) 13: 1-879. Fowler, H. W. 1970. A catalog of world fishes (XII). Quarterly Journal of the Taiwan Museum 23(1-2): 39-126. Garman, Samuel. 1913. The Plagiostoma. Memoirs of the Museum of Comparative Zoology, Harvard, 36: 1-515. Garrick, J. A. F. 1951. The blind electric rays of the genus Typhlonarke (Torpedinidae). Zoological Publications, Victoria University College, (15): 1-6. Gill, T. 1862. Analytical synopsis of the Order of Squali; and revision of the nomenclature of the genera. Squalorum generum novorum descriptiones diagnosticae. Annals of the Lyceum of Natural History of New York, 7: 367-413. Gill, T. 1893. Families and subfamilies of fishes. Memoirs of the National Academy of Sciences, United States of America, 6(6): 125-138. Gill, T. 1895. Notes on the synonymy of the Torpedinidae or Narcobatidae. Proceedings of the United States National Museum, 18(1050): 161-165. Haswell, W.A. 1885. Studies on the elasmobranch skeleton. Proceedings of the Linnaean Society of New South Wales, 9: 71-119. Henle, F. G. J. 1834. Ueber Narcine, eine neue Gattung electrischer Rochen nebst einer synopsis der electrischer Rochen. Koniglichen anatomischen Museum, Berlin, 1, pp. 1-44, pis. 1-4. FIolmgren, N. 1941. Studies on the head in fishes. Embryological, morphological, and phylogenetical researches. Part II. Comparative anatomy of the adult selachian skull, with remarks on the dorsal fins in sharks. Acta Zoologica, 22: 1-100. figs. 1-74. Human, B. A. 2006. A taxonomic revision of the catshark genus Holohalaelurus Fowler 1934 (Chondrichthyes: Carcharhiniformes: Scyliorhinidae), with descriptions of two new species. Zootaxa (1315): 1-56, figs. 1-17, figs. 1-5, append. 1. IuCN. 2006. International Union for the Conservation of Nature, Species Survival Commission. Red List web site at http:/ / www.iucnredlist.org Koorhnhof, A. 1995. The dive sites of South Africa. Comprehensive coverage of diving and snorkeling. New Holland Publishers, London, Cape Town, Sydney, Singapore, 1-183. Last, P.R., & J.D. Stevens. 1994. Sharks and rays of Australia. CSIRO, Australia, pp. i-ix, 1-513, 1400 black and white ill., color pis. 1-84. Leigh-Sharpe, W. H. 1922. The comparative morphology of the secondary sexual characters of Holocephali and elasmobranch fishes. Memoir IV. Journal of Morphology 36: 199-220. Leigh-Sharpe, W. H. 1924. The comparative morphology of the secondary sexual characters of elasmobranch fishes. Memoir VI. Journal of Morphology 39: 553-566. Leigh-Sharpe, W. H. 1926. The comparative morphology of the secondary sexual characters of elasmobranch fishes. Memoir VIII. Journal of Morphology 42: 307-320. Lindberg, G. U. 1971 (1974). Fishes of the world. A key to families and a checklist. Pp. 1-545, fig. 1-986. Translated by Israel Program for Scientific Translations, 1974. Halsted Press, New York. Lloris, D. & J.A. Rucabado. 1991. Heteronarce rierai, a new narkid ray from Mozambique, Western Indian Ocean. Japanese Journal of Ichthyology, 37(4): 327-332, fig. 1-3. Lloyd, R. E. 1907. Contributions to the fauna of the Arabian Sea, with descriptions of new fishes and Crustacea. Records of the Indian Museum, 1(1): 1-12. Lloyd, R. E. 1909. A description of the deep sea fish caught by the R.I.M.S. Ship Investigator since the year 1900, with supposed evidence of mutation in Malthopsis. Memoirs of the Indian Museum, 2(3): 139-180. McEachran, J.D., & L.J.V. Compagno. 1982. Class Chondrichthyes. In S. Parker, ed. Synopsis and classification of living organisms. McGraw-Hill, 2: 831-858, figs, 1113-1114. McEachran, J.D., K.A. Dunn, & T. Miyake. 1996. Interrelationships of the batoid fishes (Chondrichthyes: Batoidei). In M.L.J. Stiassny, L.R. Parenti, & G.D. Johnson, Eds, Interrelationships of fishes. Academic Press, San Diego, London, 63-84, fig. 1-13. McKay, R.J. 1966. Studies on western Australian sharks and rays of the families Scyliorhinidae, Urolophidae, and Torpedinidae. Journal of the Royal Society of Western Australia. 49(3): 65-82. MOller, J. & F. G. J. Henle. 1841. Systematische Beschreibung der Plagiostomen, (part 3) pp. 103-200. Veit, Berlin. Nelson, J. S. 1976. Fishes of the world. Pp. i-ix, 1-416. Wiley- Interscience, New York. Nelson, J. S. 1984. Fishes of the world, second edition. Pp. i- xv, 1-523. Wiley-Inter science. New York. Nelson, J. S. 1994. Fishes of the world, third edition. Pp. i- xvi, 1-600. John Wiley & Sons, Inc., New York. Rass, T. S., & G. U. LINDBERG. 1971. Modern concepts of the classification of living fishes. Journal of Ichthyology 11: 302-319 (translated by the American Fisheries Society). Regan, C. T. 1921. New fishes from deep water off the coast of Natal. Annals and Magazine of Natural History, series 9, 7(41): 412-420. Shirai, S. 1996. Phylogenetic interrelelationships of neoselachians (Chondrichthyes, Euselachii). In M.L.J. Stiassny, L.R. Parenti, & G. D. Johnson, Eds, Interrelationships of fishes. Academic Press, San Diego, London, 9-34, fig. 1-4. Smith, H. M. 1912. The squaloid sharks of the Philippine Archipelago, with descriptions of new genera and species. Proceedings of the United States National Museum, 41(1877): 677-685, figs. 1-4, pis. 50-54. Talwar, P. K. 1981. The electric rays of the genus Heteronarce Regan (Rajiformes; Torpedinidae), with the description of a new species. Bulletin of the Zoological Survey of India, 3(3): 147-151, fig. 1. Smithiana Bulletin 7: 15-49 A new electric ray from South Africa and review of sleeper rays 49 Von Bonde, C., & D. B. Swart. 1923. The platosomia (skates and rays) collected by the S. S. "Pickle". Union of South Africa, Fisheries Marine Biological Survey, Report (3), 1922, special Report (5), pp. 1-22 pis. 20-23. Wallace, J.H. 1967. The batoid fishes of the east coast of southern Africa. Part III: Skates and electric rays. South African Association for Marine Biological Research, Oceanographic Research Institute, Investigational Report (17): 1-62, fig. 1-29. Whitley, G. P. 1940. The fishes of Australia. Part I. The sharks, rays, devilfish, and other primitive fishes of Australia and New Zealand. Australian Zoology Handbook, pp. 1-280. Royal Zoological Society of New South Wales, Sydney. Zhu Yuanding & Meng Quingwen. 1979. A study of the lateral-line canals system and that of Lorenzini ampullae and tubules of Chondrichthyes fishes of China. Monographs of Fishes of China, Science and Technology Press, Shanghai, i-iv, 1-132, figs. 1-71. Smithiana Bulletin 7: 15-49 Fishes of the genus Helcogramma (Blennioidei: Tripterygiidae) in the Western Indian Ocean, including Sri Lanka, with descriptions of four new species. Wouter Holleman South African Institute of Aquatic Biodiversity, Private Bag 1015, 6140 Grahamstown, South Africa, e-mail: w.holleman@ru.ac.za Received 11 October 2006; accepted 23 February 2007. Abstract. Fifteen species of the tripterygiid fish genus Helcogramma are recognised from the Western Indian Ocean (including Sri Lanka and the southeastern coast of India). Helcogramma shinglensis Lai Mohan is recognised as a valid species and four species are described as new: Helcogramma alkamr, which is similar to H. chic a Rosenblatt, is known from the Comoro Islands, the Seychelles and Mauritius; Helcogramma serendip, apparently confined to Sri Lanka and possibly a sibling species to H. alkamr; Helcogramma ememes from the Seychelles and the East African coast, which is similar to H. hudsoni Schultz, and Helcogramma rharhabe, one of the H. obtusirostre complex of species, and which appears confined to the east coast of Africa from northern Mozambique to East London. Diagnoses of the other Western Indian Ocean species and a key are provided. Keywords: taxonomy, Tripterygiidae, Helcogramma, new species. Western Indian Ocean INTRODUCTION The genus Helcogramma McCulloch & Waite 1918 has been revised only once, by Hansen, in 1986. Subsequent studies on tripterygiid fishes have been regional (e.g. Fricke 1994 & 1997), have described new species (e.g. Fricke & Randall 1992; Randall & Clark 1993), or dealt specifically with the Helcogramma fuscopinna complex (Williams & McCormick 1990; Williams & Howe 2003). Hansen recognised 12 species, describing four as new. She also synonymised many nominal species, several as Helcogramma obtusirostre (originally described by Klunzinger from the Red Sea), a species which she considered to occur widely in the Indo-West Pacific Ocean. Williams and McCormick (1990) recognised that H. fuscopinna is a complex of at least eight species. Fricke (1994, 1997) described a further three new species for the genus. This study recognises 15 species from the Western Indian Ocean of which four are described as new, bringing the total number of species in Helcogramma to some 35. Notwithstanding Fricke's record of variations of 4, 5 or even 6 in counts of fin spines or rays in a single species, this review shows little morphometric variation in Helcogramma species. At most, fin elements will vary by 3, and then the number of counts off the norm (mode) are generally very low. Lateral¬ line scale counts do show greater variation in some species. This study also shows that several species have fin-element counts that do not separate them from each other. It was this factor, I believe, that encouraged Hansen to synonymise a number of species as Helcogramma obtusirostre. However, individuals from different areas can be separated on the basis of colour pattern (see e.g. Williams & Howe 2003; Holleman 2006). In the Western Indian Ocean there are three species that can be ascribed to the Helcogramma obtusirostre complex. There are additional species in the Western Pacific Ocean-as well as one in the Southern Atlantic-that can also be ascribed to this complex. METHODS All measurements were made with pin dial calipers under a binocular microscope and are given in millimeters to a single decimal place. Ratios are calculated as the number of times a given measured length is contained in either the standard length or the head length. The measurement of head profiles, as measured in Holleman (1982), are fairly 'crude7, but serve to indicate differences in skull structure. All counts were made under a microscope, including vertebrae, which were counted from radiographs. Pectoral-fin ray and lateral-line counts were made on the left side. Specimens on rare occasion have lost or gained a single ray in one of the pectoral fins. If the count was not as expected, the right side was counted, and if "normal", recorded; if not, the "non-normal" count was recorded. The counts are given, from the dorsalmost fin ray, as number of undivided rays, number of divided rays, number of undivided rays, e.g. 2, 7, 7. The numbers of divided and undivided rays were found to be remarkably consistent for most species. Body scales of tripterygiid fishes are deciduous and often missing. If the last lateral-line scale on the left side was not followed by an ordinary body scale, the right side's scales were counted. If neither side was complete, the count was not recorded. Most species of Helcogramma have a naked area laterally along the base of the first and often the second dorsal fins. In species in which the body scales do not extend to the base of these fins, the scales decrease in size and become very thin as they approach the fin bases. They are best seen Smithiana Bulletin 7: 51-81 52 Wouter Holleman Table 1. Selected characters of WIO species of Helcogramma\ usual/modal counts are given in parenthesis. Species Second & third dorsal fins Anal fin Pectoral fins Lateral line Total lateral scales Mandibular pores Vertebrae Nape Ht. D1 cf D2 alkamr WIO islands n = 174 XIII-XIV + 10-11 19- 20 16: 1, 8, 7 19-25 (mode 21 ) few with 1 9 or 23-24 36-39 (37-38) 5+1+5 (%) 6 + 1+6 (Vh) 10 + 25-26 naked y2 0" c f = ? obtusirostre Red Sea and Oman n = 20 XII-XIII + 9-10 rarely XII spines or 9 rays 18-19 (19) 16: 2, 7, 7 20-23 (21-22) 36-38 (37-38) 4 + 1+4 10 + 25-26 naked ~y2 rharhabe East London to Bazaruto n = 236 XII-XIV + 10-11 (XIII + 10) 18-20 rarely 18 or 20 16: 1,8, 7 sometimes 15 20-31 (mode 24) 37-38 (38) 5+1+5 10 + 24-25 naked ~ V2 rosea Sri Lanka to Phuket n = 42 XI-XIVI + 10-12 rarely 10 18-20 (20) rarely 18 16: 2, 7, 7 23-29 usually 25-27 36-38 usually 37 3-4 + 1 + 3-4 1 1 + 24-26 (25) scaled >