Institutional Abbreviations

Torpedo Houttuyn, 1764

Type Species: Raja torpedo Linnaeus, 1758 (by tautonomy).

Apomorphic Characters: Derived skeletal features of the genus (and family) include paired, slender rostral extensions formed by anterodorsal outgrowths of the lamina orbitonasalis (Holmgren, 1940; Miyake et al., 1992), projecting anteriorly in a relatively straight fashion; conspicuous shape and extent of antorbital cartilages (very slender, curved, and with comblike projections extending anteriorly to support anterior disc area); first hypobranchial (basihyal?) continuous with first ceratobranchial cartilage; multiple and subdivided hypobranchial elements radiating anteriorly or anterolaterally from anterior margin of basibranchial copula.

Species Diversity: Of the 46–53 nominal species (the availability of 7 of these is problematical; cf. Eschmeyer [1998] and Carvalho [1999]), only 17–20 are here considered valid. At least five additional new species have been recognized and are presently being described.

Torpedo adenensis, new species Figures 1–7, table 1

Holotype: ZMUAS 5121 (109204), 407 mm TL adult male, Gulf of Aden, 15°15′03"N, 52°00′04" E, 125–140 m, R/V Dm. Stefanov (cruise 5, bottom trawl station no.125), 29 December 1989, collectors: S.I. Usachev and Y.G. Domaschenko (YugNIRO Kerch) (fig. 1).

Paratypes: Four specimens: ZMUAS 5121 (109203), 397 mm TL adult male (same data as holotype) (fig. 2); ZMUAS 5418, 390 mm TL adult female, Gulf of Aden, 15°07′09"N, 51°58′02"E, 230 m depth, R/V Dm. Stefanov (cruise 5, bottom trawl station no. 126), 29 December 1989, collectors: S.I. Usachev and Y.G. Domaschenko (YugNIRO Kerch); AMNH 231441, 295 mm TL (subadult?) female (same data as holotype) (fig. 3a); ISH 4–1993, 277 mm TL subadult male, Gulf of Aden, 15°17′00"N, 51°34′05"E, 26–33 m, R/V Dm. Stefanov (cruise 5, bottom trawl station no. 120), 28 December 1989, collectors: S.I. Usachev and Y.G. Domaschenko (YugNIRO Kerch) (fig. 3b).

Diagnosis: A species of Torpedo from the Gulf of Aden distinguished from all congeners by its unique dorsal coloration, a uniform reddish-, rusty-, or orange-brown without any distinctive spots, blotches, or reticulations. A unique combination of characters further distinguishes Torpedo adenensis: spiracles with few, knoblike, and somewhat inconspicuous spiracular papillae (at least in specimens sexually mature or almost mature); spiracles and eyes relatively close together (without large gap in between); base of first dorsal fin extending just posterior to level of pelvic fin axil (not entirely positioned over bases of pelvic fins); distance between second dorsal and caudal fin greater than distance between first and second dorsal fins; claspers with fleshy integumental flap in clasper glans region.

Description: Measurements and counts are summarized in table 1. External morphology—Disc broadly rounded in outline with relatively straight anterior margin, but some specimens (e.g., ISH 4–1993) with a slight median protuberance anteriorly. Disc wider than long (disc width ranging from 58.6 to 67% of TL and disc length ranging from 52.3 to 55.9% of TL), and widest anteriorly at about one-third of its length. Disc only faintly overlaps origin of pelvic fins, not leaving a prominent free lobe posteriorly (e.g., ZMUS 5121 [109203]), but in most specimens disc does not appear to overlap pelvic fins due to slight postmortem shrinking. Preorbital snout length about one-seventh of disc length and just greater than prenasal length (ranging from 6.9 to 8.4% of TL compared to 5.5 to 6.9% of TL, respectively), and smaller than preoral snout length in all specimens. Snout anterior to orbits about equal to length of eyes and spiracles combined. Eyes moderately small, horizontal eye diameter ranging from 2.8 to 3.4% of TL, much smaller than spiracles. Eyes and spiracles relatively close together, space in between them not as great as horizontal eye diameter. Spiracles relatively large and rounded, with slightly elevated rims, especially on posterior contour. Very small, inconspicuous knoblike spiracular papillae are present around posterior margin of spiracle, sometimes only clearly seen under magnification; eight papillae in subadult male paratype (277 mm TL, ISH 4–1993), but only three in holotype (407 mm TL adult male). Papillae relatively closely spaced on inner margin of spiracle in ISH 4–1993 (fig. 4), but more sparsely grouped on spiracular outer margin. Pseudobranchial folds present inside spiracles, lining anterior spiracular wall, ranging from 8 in smaller male paratype to 12 in larger male paratype. Spiracular opening itself constricted by anterior spiracular wall in some specimens, and relatively small and crescent-shaped. Distance between eyes greater than interspiracular distance in all specimens. Electric organs without a clearly visible outline dorsally, but easily observed in ventral view. Electric organs originate very close to anterior disc contour, well anterior to eyes and nostrils, and terminate posteriorly just caudal to last gill openings. Electric organs with concave inner margins posteriorly and convex outer contours, and widest anteriorly in between mouth and first gill slit. Electric organ length about twice its width at widest point. Anterior to gill openings, electric organs inflect somewhat abruptly toward midline.

First gill slit situated at about anterior one-third of disc length; last gill slit just anterior to two-thirds of disc length. Gill slits crescent-shaped, situated lateral to level of outer corners of mouth and nostrils (more or less level with pelvic fin origin). Distance between first gill slits just slightly greater than distance between last gill slits. Third gill slit larger than others, and fifth gill slit smallest. The nasal curtain short and wide, its length just a little smaller than internarial width. Nasal curtain extends posteriorly in a more or less straight line to mouth opening. Posterior margin of nasal curtain with a slight median lobe present (median lobe may be obscured because nasal curtain distorts easily in preservative). Nostrils with prominent folds mostly surrounding outer margins (fig. 5); outer margins of nostrils situated just lateral to level of mouth corners. Posterior contour of nostrils confluent with conspicuous median lobes that contact nasal curtain at its corners and separate nostrils from mouth. Mouth arched, with a relatively large gape. Width of mouth is slightly greater than or equal to distance between nostrils. Teeth in 33/32–47/39 rows in type series (from 277 mm TL to 407 mm TL). Teeth amber in color, set in quincunx, somewhat flattened labial-lingually. Tips of individual cusps overlap bases of teeth positioned in next inner tooth row. Teeth morphologically similar among both upper and lower jaws, with well-developed single cusps, these more prominent toward central rows. Cusps slightly oriented toward lateral mouth corners on all teeth, but more prominently on teeth with longer cusps. Teeth of outermost rows with relatively wider bases and smaller cusps, but cusps also slender. Cusps of inner and middle rows tall, pointed, and somewhat sharp. Tooth bases slightly arched at center.

Pelvic fins originate just anterior to posterior disc margins. Pelvic fins long but not very wide, greatest width of pelvic fins just less than one-half of disc width, and slightly wider in male specimens. Anterior margins of pelvics relatively straight, not projecting laterally to a large degree. Posterior margins of pelvics more convex. Pelvic fins much longer than wide; pelvic fin length ranging from 20.6 to 24.4% of TL. Pelvics widest posteriorly, at about two-thirds of their length. Cloaca situated at about midlength of pelvic fins. Claspers project significantly from distal tip of pelvic fins in all male specimens, mature or subadult, reaching posteriorly to at least level of midbase of second dorsal fin. Claspers relatively slender and not tapered greatly from origin to distal tip. Clasper groove extends on dorsolateral aspect of clasper from pelvic fin to about three-fourths of its length, where it curves toward the midline and continues to clasper tip. A fleshy integumental flap, bearing a distinctive notch, present close to the hypopyle (fig. 6). Both ventrolateral and dorsolateral slits (presumably the ventral pseudosiphon and dorsal pseudosiphon, respectively) present, continuing posteriorly to close to level of integumental flap.

Tail relatively short and stout, ranging from 18.6 to 21.9% of TL as measured from second dorsal fin origin, but tail length from cloaca is about 40% of TL. Tail tapers considerably from posterior tips of pelvic fins toward caudal fin. Tail oval in cross-section close to pelvic and dorsal fins, being much wider than tall almost along its entire extent, but at caudal fin origin tail more circular in cross-section. Lateral tail folds ridgelike and not very prominent in preserved specimens, extending from underneath level of anterior second dorsal fin base to end posteriorly on ventral aspect of caudal peduncle. Lateral tail-folds ranging from 9.8 to 15.4% of TL, and may vary slightly in length among both sides of the same specimen. First dorsal fin origin anterior to level of posterior axil of pelvic fins. First dorsal fin base entirely anterior to pelvic fin posterior apex, but posterior free lobe of first dorsal just posterior or at same level of posterior apex of pelvic fins (fig. 7). Posterior insertion of first dorsal fin just posterior to pelvic fin axil in all specimens, but more so in holotype and other males. First dorsal fin originates at about same level of greatest pelvic fin width. First dorsal fin moderately slanted at its anterior margin, with a posterior free lobe about equal to one-third of its base length (less so in the female paratype). First dorsal fin greater than second in both height and length of base in all specimens (first dorsal fin height ranges from 7.6 to 10.6% of TL; second dorsal fin height ranges from 5.5 to 7.2% of TL). Dorsal fins more or less similar in shape, but second dorsal fin more slanted and lower, with a more acute apex and relatively greater free lobe posteriorly in the male specimens. Female paratype (AMNH 231441, 295 mm TL) with relatively smaller dorsal fins in relation to male specimens, even smaller proportionally than smaller male paratype (ISH 4–1993, 277 mm TL). Dorsal fins of female also more broadly rounded at apex and less slanted. Distance between dorsal fins clearly smaller than distance between second dorsal and caudal fin (in ISH 4–1993, distance between dorsals less than half of distance between second dorsal and caudal). Caudal fin moderately tall, ranging from 12.5 to 17.2% of TL, and taller proportionally in larger specimens. Upper lobe of caudal fin longer than lower lobe in all male specimens, but lower lobe just longer than upper in the female paratype (AMNH 231441). Upper lobe of caudal fin slightly more sloping than lower lobe (especially in the holotype), and relatively straight from origin to apex. Posterior margin of the caudal fin more or less straight, especially in males, but it may have a small protrusion at mid-height. Both upper and lower apexes of caudal fin slightly acute and not broadly rounded (fig. 7). Coloration—In preservative, dorsal coloration uniform reddish-, rusty-, or orange-brown without distinctive blotches, spots, vermiculations, or reticulations. Tip of claspers dorsally slightly darker than rest of dorsal surface. Very small faint whitish spots, roughly the size of sensory pores, may be present on dorsal disc surface, unevenly scattered over outer and central disc, electric organs, and to some extent over tail region (these are more apparent on the female paratype and are perhaps preservational, as they are very faint, irregular, and inconspicuous). Posterior margin of dorsal and caudal fins with a very narrow edging of light cream color. Some specimens (e.g., ISH 4–1993) have a lighter brown dorsal color, but this may be preservational. Coloration when freshly captured not significantly different (perhaps more intense). Holotype dorsally with a small unpigmented area at center of disc (an abrasion). Ventral coloration uniform creamy-white, but disc and pelvic fins clearly have darker margins similar to dorsal color. This edging stronger on the ventral surface at lateral and posterior disc and posterior margins of pelvic fins. Ventral surface of tail irregularly blotched with rusty-brown in most specimens (slightly less in the holotype). Irregular blotches also present on ventral snout region of female paratype (AMNH 231441), and at cloaca of one male paratype (ZMUAS 5121). Pores of ampullary system highlighted with a reddish-brown pigmentation on ventral disc surface, particularly those outlining electric organs. Holotype with faint, indistinct blotches of darker color on ventral surface of mid-disc and right pelvic fin region. Selected skeletal features—Superficial calcification present on most skeletal elements, but more pronounced on neurocranium (except nasal capsules), synarcual cartilage, and pectoral and pelvic girdles. Neurocranium slender and somewhat hourglass-shaped, about one-third of disc length. Rostral structures not visible in radiographs. Antorbital cartilages slender and anterolaterally directed, forming a partial anterior border to electric organs. Jaws slender and highly arched; labial cartilages absent. Hyomandibulae short, devoid of rays, with very stout proximal portions that articulate with neurocranium, and taper greatly to contact outer jaw corners. Synarcual relatively long and wide, extending posteriorly to level of suprascapulae; its total length (including posterior segment with distinct centra) about equal to neurocranial length. Anteriorly, synarcual notched on either side of foramen magnum. Anterolateral margin of lateral stays sloping and concave; posterior margin more straight. About 12 spinal nerve foramina exiting from synarcual anterior to first separate vertebral centrum. Basibranchial copula large, trapezoidal, and widest just posterior to its midlength. Hypobranchials 2 and 3 radiating from anterior margin of copula obliquely, extending anteriorly underneath neurocranium. Ceratobranchial and epibranchial elements slender, but last ceratobranchial relatively large, articulating with basibranchial copula posterolaterally, and contacting scapulocoracoid posteriorly. Scapulocoracoid slender medially but stouter at corners, where it articulates with pectoral pterygia; ascending scapular processes not fused with suprascapula. Pro- and mesocondyles protrude slightly (metacondyle not visible in dorsoventral view). Propterygium stout, divided into six segments (decreasing in size distally from procondyle), and also contacting mesopterygium posteriorly. Mesopterygium small and slender, extending laterally obliquely and articulating with anteriorly situated procondyle at its midlength. Metapterygium very slender and inconspicuous, articulating with very few weakly calcified radials. Puboischiadic bar slender, with concave anterior margin and laterally projecting prepelvic processes at anterior corners. Posterior margin of girdle convex, with median indentation at level of vertebral column. At least three obturator foramina present at each anterolateral aspect of pelvic girdle, mesial to small triangular process that projects laterally from each corner. Each basipterygium about as long as half of pelvic girdle width, and with sinuous external margin for articulation with pelvic radials. First pelvic radial about twice as stout as remaining radials, and articulating distally with small additional radial element. Dorsal caudal radials more elongate than ventral caudal radials; radial elements with slight proximal thickening.

Etymology: The specific epithet adenensis (an adjective) is in reference to the only known locality, the Gulf of Aden. Gender masculine.

Geographic Distribution: Torpedo adenensis has been collected from three distinct but adjacent localities from the eastern Gulf of Aden, close to the Yemen coastline, at depths ranging from 26 to 140 m (fig. 8).

Torpedo panthera Figures 9, 10a, table 2

The remaining syntype of T. panthera (lectotype) is in relatively good condition (ZMB 4560; fig. 9), as it is not significantly desiccated (measurements in table 2). However, no trace of its original coloration remains, which was described by von Olfers (1831: 15) as darker brown with small whitish spots (“dunkler braun mit wenigen weisslichen Flecken”; darker presumably in reference to T. marmorata). ZMB 4560 was identified as the sole remaining syntype by Paepke and Schmidt (1988: 172; but not by Klausewitz, 1960), and it presumably was the only specimen available only a decade later to Müller and Henle (1841: 193). The lectotype is roughly equal in total length to one of the paratypes of T. adenensis (ISH 4–1993), allowing for a detailed comparison. It contains within its viscera an undigested teleost, ingested head-first, which is about one-third of its total length.

We have examined specimens of T. panthera from the northern Red Sea (Egypt; fig. 10a), and reliable literature records confirm its presence in the southern Red Sea (Rüppel, 1835) and Gulf of Oman (Randall, 1995). Annandale (1909) recorded a specimen of Torpedo (misidentified as T. marmorata) from Puri (off eastern India, in the Bay of Bengal) that is very similar in coloration to T. panthera (see his photograph on plate 5, fig. 3), thus greatly increasing its range if confirmed. Bianchi (1985) cited this species from Pakistan, but perhaps not based on hard evidence, even though it probably does occur there. Torpedo panthera apparently does not extend farther south than the Red Sea along the east African coast (contrary to Sommer et al., 1996), and it has not been recorded from the Gulf of Aden even though it probably occurs there as well. Further reliable Red Sea references include Klunzinger (1871), Bamber (1915), Gohar and Mazhar (1964), Dor (1984), Baranes and Golani (1993), and Field and Field (1998), but not Fowler (1956, both species mentioned may be T. sinuspersici). Fowler's (1934) mention of this species from South Africa is also a misidentification. The accounts of Blegvad (1944) and Kuronuma and Abe (1986) of T. panthera from the Persian Gulf are based on misidentifications (see below). Carpenter et al. (1997) also record T. panthera from the Persian Gulf, but whether this is compiled or based on specimens is not clear.

Specimens of T. panthera from throughout its confirmed range present small clusters of more or less isolated, and sometimes blurry, whitish spots over the disc, pelvic fins, and tail, contrasting with the more uniform coloration of T. adenensis. There is some variation in the amount, size, and arrangement of spots in different specimens of T. panthera, but whether this variation is significant systematically is unknown (Rüppell's stylized depiction of T. panthera from the Red Sea is similar to Annadale's specimen from the Bay of Bengal). The very small (sensory pore-sized), faint lighter markings irregularly present in the female paratype of T. adenensis (which may be preservational) are very distinct from the much larger and conspicuous spots of T. panthera, which are constantly present and much larger than sensory pore diameter (may be as large as eye diameter) (fig. 10a). Torpedo adenensis can be further distinguished from T. panthera by having the distance between the second dorsal and caudal fin markedly greater than interdorsal distance. In T. panthera, the interdorsal distance is roughly equal to the distance between the second dorsal and caudal fin (according to all reliable accounts and specimens examined, including the lectotype). Compared to T. panthera, T. adenensis has spiracles relatively closer together, less pronounced spiracular papillae (comparing individuals of roughly equal dimensions; specimens of T. panthera also commonly have a more pronounced postero-central papilla), and a straighter posterior caudal fin margin (in many specimens of T. panthera the posterior margin is more curved). Torpedo adenensis also has more tail vertebrae (and hence total vertebrae) than does T. panthera (tables 1, 2). The lectotype of T. panthera has 61 tail vertebrae (and 86 total vertebrae), compared to the AMNH and ISH paratypes of T. adenensis (which are roughly the same size as the lectotype of T. panthera), which have, respectively, 73 (98 total) and 71 (97 total) vertebrae (however, more specimens of T. panthera need to be compared). Both species are similar in external clasper structure, presenting a unique median integumental flap in the clasper glans region, interpreted as a derived feature supporting their monophyly as a species pair (fig. 6). Judging from the lectotype, T. panthera reaches sexual maturity at least before 281 mm TL as its claspers are fully calcified, and therefore probably at a smaller size than T. adenensis (paratype ISH 4–1993 is a subadult male at 277 mm TL; this species matures sexually between 280 and 395 mm TL). This may constitute further evidence distinguishing both species.

Torpedo sinuspersici Figure 10b

Torpedo sinuspersici was originally used by von Olfers (1831) as a variation of T. marmorata (as was T. panthera), a variation that also included ”Raja maculata” and ”R. bicolor” of Shaw (1804) (both names are referred to Narcine; Carvalho, 1999), as von Olfers (1831) erroneously regarded these nominal species as synonymous. The brief original description of T. sinuspersici only mentions the presence of two dorsal fins and a white- and dark brown-spotted dorsal coloration pattern as diagnostic characters (von Olfers, 1831: 17, citing and translating Kämpfer's original Latin diagnosis). No type specimen exists, and designating a neotype is premature at present.

According to published accounts and material examined, this species appears to be the most widespread of the four previously described western Indian Ocean forms, with a reported range extending to South Africa (Wallace, 1967; van der Elst, 1993; Compagno, 1995), Somalia (Sommer et al., 1996), the Red Sea (e.g. Compagno and Randall, 1985), the Arabian Sea (Bianchi, 1985; Qureshi, 1972), the Gulf of Oman (Debelius, 1993, 1999; Randall, 1995), the Persian Gulf (Carpenter et al., 1997), Sri Lanka, and the Andaman Sea (Kuiter and Debelius, 1997). Specimens or more reliable records of T. sinuspersici are still required from the northern Arabian and Andaman seas (we have not examined material from the Red Sea either, even though it is commonly reported from there). Fourmanoir (1963) may have misidentified T. sinuspersici as T. fuscomaculata from Madagascar, and Smith and Smith's (1963) T. fuscomaculata from the Seychelles may also include T. sinuspersici or another similar form. Jones and Kumaran (1980) illustrated a specimen from the Laccadive Islands, misidentified as T. marmorata, that is probably this species as well. Other misidentifications of T. sinuspersici as T. marmorata in the literature of the western Indian Ocean include Günther (1870), von Bonde and Swart (1923), Smith (1961), Qureshi (1972) and perhaps Prashad (1920). We have examined material of T. sinuspersici from Sri Lanka and the Persian Gulf (Iran).

Dorsal coloration appears to be the most secure means of identifying T. sinuspersici (Compagno et al., 1989; Compagno, 1995), even though it may be somewhat variable, and also easily distinguishes it from T. adenensis with which it may co-occur in the Gulf of Aden (although no records or material of T. sinuspersici in the Gulf of Aden were found). In the Persian Gulf (type locality), its coloration is composed of strong cream or whitish and rather thick vermiculations over disc, pelvic fins, and tail, with many cream-colored and irregular spots, no larger than eye diameter, on anterior and lateral disc regions (fig. 10b). These vermiculations may be formed from fused irregular spots. A specimen from South Africa, of undetermined size, has cream-colored brain-shaped figures or rosettes over disc and tail, decreasing in size toward disc margins, along with numerous small spots laterally on disc and the anterior snout region (van der Elst, 1993). Compagno et al. (1989) and Compagno (1995) illustrated roughly the same arrangement (based on a specimen from Natal), but with more closely packed brain-shaped figures and rosettes, approaching the pattern of specimens from the Persian Gulf and Sri Lanka (see also illustration in Wallace, 1967: fig. 27). Field and Field (1998) depicted a specimen identified as T. sinuspersici from the northern Red Sea that varies considerably from the typical pattern, resembling T. marmorata from the Mediterranean more than typical T. sinuspersici (if confirmed, this would indicate that the former species probably dispersed through the Suez Canal). Specimens from the full distribution of T. sinuspersici need to be compared in detail to determine if it can be subdivided into more than one species.

Torpedo fuscomaculata Figures 11–13, table 2

Peters (1855) described the original coloration of T. fuscomaculata as having a yellowish- to reddish-brown background with closely packed brownish-black spots and brownish-black dorsal fins with a white periphery. The type specimens are slightly distinct in coloration (cf. figs. 11 and 12), as the smaller paralectotype is more markedly spotted and less faded, but both specimens retain details of their original color. The lectotype has faded dark brown blotches over dorsal disc, pelvic fins, dorsal and caudal fins, and tail, and is probably late subadult or adult, as the clasper is somewhat rigid with calcified terminal cartilages. Accurate measurements could not be made on the ventral surface of the lectotype, and the disc and pelvic fins are distorted (our measurement of TL is more or less equivalent to that of Peters at 223 mm, but other measurements cannot be directly compared). The paralectotype is about one-half the length of the lectotype, and it was not measured (table 2). Its coloration is composed of numerous spots that become progressively smaller toward disc margins (much smaller than eye diameter), with larger spots (larger than spiracles) over posterior mid-disc area. The distance between the second dorsal and caudal fin is greater than the distance between both dorsal fins in the lectotype (this is also reported for 13 specimens from South Africa by Wallace, 1967), but these dimensions are about equivalent in the paralectotype. This is in contrast to the assertion by Fraser-Brunner (1949; repeated by Bigelow and Schroeder, 1953; Sommer et al., 1996), who used the greater distance between the first and second dorsal fins to separate this species from T. sinuspersici. The first dorsal fin base appears to be entirely above the pelvic fin base. The lateral tail fold of the lectotype is ridgelike, extending from about midlength of the second dorsal fin base to the lower caudal peduncle. The clasper does not have the integumental flap in the clasper glans region as in T. adenensis and T. panthera. Specimens of T. fuscomaculata are easily separated from T. adenensis on the basis of coloration, but size at maturity may provide another distinction if the lectotype of T. fuscomaculata (223 mm TL) is either adult or close to adult, and if sexual maturation closely follows clasper rigidity.

The differences in coloration in both original syntypes may be because of their distinct sizes, but a complete ontogenetic series of T. fuscomaculata from Mozambique still needs to be studied. Alternatively, both specimens may represent distinct color morphs of this species. Variation in coloration has been previously reported by many authors, such as Playfair and Günther (1866) from Zanzibar, and Wallace (1967) and Compagno et al. (1989; also Compagno,1995) from South Africa. The variation in coloration of T. fuscomaculata is confined not only to number, size, and disposition of the darker spots, but also to whether they are present at all, because some examined specimens from South Africa, roughly the same size of the paralectotype, have a uniform dark grayish- or blackish-brown dorsal coloration without any spots. Other specimens are reported to have gray flecks along with darker spots, or even a gray background color (e.g., Smith, 1961; Compagno, 1995). As in T. sinuspersici, this variation still awaits a more thorough documentation. A lectotype is chosen for T. fuscomaculata so as to avoid taxonomic confusion in case its type series is shown to be heterogeneous.

The holotype of Torpedo smithii Günther, 1870 (fig. 13) agrees with the lectotype of T. fuscomaculata in coloration (even though faded), being dark brown with darker brownish-black irregular spots, about the same size as the spiracles, over the disc. The specimen has a shrunken disc and pelvic fins and dried dorsal fins. As in the lectotpye of T. fuscomaculata, the distance between the second dorsal and caudal fin is greater than the distance between both dorsal fins (also as in T. adenensis). The eyes are deeply embedded and set apart from the spiracles by a small gap, and the central posterior spiracular papilla is larger than the remaining inconspicuous papillae, similar to the holotype of T. panthera. There are some 50–54 upper tooth rows, and teeth are very sharp with pointed cusps and relatively wide bases. The lateral tail folds remain as slight ridges, similar in extension to T. fuscomaculata. The second dorsal fin is smaller and more slanted than the first dorsal fin. There is no reason to consider T. smithii a separate species distinct from T. fuscomaculata, as previous authors have suggested (e.g., Fraser-Brunner, 1949; Compagno, 1995).

Torpedo fuscomaculata appears to be more restricted geographically than T. sinuspersici, being distributed in the western Indian Ocean from South Africa (Wallace, 1967; Compagno et al., 1989; Compagno, 1995) to Zanzibar (Playfair and Günther, 1866) and perhaps as far north as the Kenyan coast (Sommer et al. [1996] reported it as possibly occurring off Somalia, but more definite records do not exist). It has also been recorded from Madagascar (as Torpedo sp.; Fourmanoir, 1963), Seychelles (perhaps only in part; Smith and Smith, 1963), and Mauritius (Fraser-Brunner, 1949). We have examined material from Mauritius (Agalega Islands), South Africa, and Mozambique, but specimens from Mauritius may in fact be distinct, as they have a lighter background color with smaller and more regularly displayed spots.

Torpedo suessi Figures 14, 15

Torpedo suessi was described by Steindachner (1898a) in an abstract preceding his more detailed account of it (Steindachner, 1898b). Two of his three original specimens remain in Vienna and are depicted in figures 14 and 15. Torpedo adenensis is distinct from T. suessi in coloration, described originally as having a bright brown background with eight large, deep brown spots with lighter outlines (ocelli), with slender reticulations over the disc and disc edges, and with smaller dark spots on the anterior disc region, pelvic fins (in the female syntype, an anterior pelvic spot merges with a larger ocellus from the disc), and dorsal fin bases (Steindachner, 1989b: 786). Even though slightly faded, the striking original coloration of T. suessi is still present in the type specimens (figs. 14, 15), easily separating it from congeners from the western Indian Ocean, the Red Sea, and adjacent gulfs. Steindachner's original plate (taf. II) appears to be of the lectotype (there are no signs of claspers in the illustration), but it may also be a composite (the ocelli are not in strict accord with any of the two remaining types) or the missing original syntype (a 130 mm TL male specimen). The lectotype of T. suessi (originally reported to be 300 mm TL) can be directly compared to the AMNH and ISH paratypes of T. adenensis, as they are similar in TL, confirming that their distinction in coloration is not dependent on size. The paralectotype of T. suessi has elongated claspers devoid of the integumental flap present in both T. adenensis and T. panthera.

Torpedo suessi has been considered a junior synonym of T. sinuspersici by almost all previous authors, following Fraser-Brunner (1949; e.g., Bigelow and Schroeder, 1953; Dor, 1984), but those authors did not examine the type specimens or provide other justification. Garman (1913) considered T. suessi to be a valid species, but noted that it may synonymous with T. panthera. It is inconceivable that the ornate color pattern of T. suessi, originally described in specimens of 130, 240, and 300 mm TL, will transform into the typical pattern of small, irregular whitish spots present in T. panthera, or into the vermiculate or marbled whitish pattern over a dark brown background of T. sinuspersici. We have examined specimens of T. panthera smaller than 200 mm TL (fig. 10a) and confirm that the color pattern of juveniles is essentially the same as in adults, and never composed of dark brown spots or ocelli. We have also examined specimens of T. sinuspersici smaller than 240 and 300 mm TL, and clearly the same typical color pattern of larger specimens is present. We therefore refute that T. suessi is a synonym of any one of these two species, and consider it a valid species that occurs at least in the southern Red Sea, known from two extant type specimens.

Photographs of a species of Torpedo taken by Jack Randall in the Sudanese Red Sea (fig. 16) may constitute a further record of T. suessi. The specimens photographed resemble T. suessi in dorsal background coloration (composed of a slender reticulate pattern) and in having large brown spots over the dorsal disc region (another photo of this species by J. Randall is featured in Michael [1993: 79] as “Ocellated torpedo ray or Torpedo sp. 1”). The differences in color pattern between the type specimens of T. suessi and the specimens from Sudan may be ontogenetic or intraspecific. The transformation from proportionally larger and prominent ocelli into more uniformly colored and fewer spots is possible, as a similar transformation has been recorded previously for another electric ray species (Narcine entemedor; Carvalho, 1999). Specimens of this distinctive Torpedo from off Port Sudan are needed for comparisons to clarify if they represent an undescribed species or simply a range extension for T. suessi.

Another possibility (not supported here) is that T. suessi can be identified with the paralectotype of T. fuscomaculata (ZMB 4573; fig. 12). This specimen presents larger brown spots posteriorly over the disc that may be encircled by lighter color (ocelli?), with numerous smaller spots over disc and pelvic fins. However, this identification is very remote, as the paralectotype of T. fuscomaculata is from the coast of Mozambique and the similarities in coloration are not convincing. Torpedo suessi would still be a valid species even if it were conspecific with the paralectotype of T. fuscomaculata.

Comments on Other Nominal Species of the Subgenus Torpedo

The original descriptions of Narcacion polleni Bleeker, 1866 and Torpedo zugmayeri Engelhardt, 1912, and the examination of the holotype of the former species, clearly indicate (on the basis of dorsal coloration) that they cannot be used for our new species. These nominal species have been relatively neglected in the literature, and their taxonomic status has never been critically examined.

The holotype of T. polleni (fig. 17) does not have a uniform dorsal coloration as in T. adenensis. Described originally from Réunion (not figured), T. polleni was distinguished by Bleeker from both T. panthera and T. fuscomaculata (T. sinuspersici was overlooked) on the basis of its teeth, shape of disc and fins, and dorsal color. The dorsal coloration was originally described as having a blackish-violet background, with off-white, white, and irregular pearllike spots that coalesce into a strong and varied marbled pattern. The holotype has faded considerably in color, and the background is now brown or orange-brown, but traces of creamy spots that appear to fuse into rosettes (giving it a vermiculate or marbled appearance) are still evident over the disc. The dorsal, caudal, and pelvic fins, along with the edges of the disc, the eye, nasoral regions, and the dorsal snout area, are poorly preserved in the holotype. Garman (1913) mistakenly cites this species as being described from Madagascar.

There are only three possible interpretations concerning T. polleni in our view: it is a junior synonym of T. panthera, a junior synonym of T. sinuspersici, or it is a valid species (it is not a junior synonym of T. fuscomaculata). The first of these is the most improbable outcome, as T. panthera does not occur off islands in the southwestern Indian Ocean or along the eastern African coast, as far as we can discern, and the dorsal vermiculate pattern of T. polleni is more similar to the coloration of T. sinuspersici than to T. panthera. Bigelow and Schroeder (1953) synonymized this species with T. panthera, but they did not examine its type specimen. We tentatively regard T. polleni as a synonym of T. sinuspersici on the basis of the similarity in coloration, fully realizing that it may eventually prove to be a valid species (T. sinuspersici has not been reported from Réunion, Mauritius or the surrounding islands).

Torpedo zugmayeri was described by Engelhardt (1912) from a single 330 mm TL female specimen (not figured) from off Gwadar (southwestern Pakistan; fig. 8). This specimen apparently no longer exists (the same fate of other of Engelhardt's types originally in Munich). Engelhardt (1912) described its coloration as bright brown dorsally with blackish mottles or a blackish-marbled pattern, and ventrally yellowish-white with brown-stained disc edges. This darkly mottled or spotted color pattern clearly distinguishes this nominal species from T. adenensis as well as from T. panthera. Bigelow and Schroeder (1953) placed T. zugmayeri in the synonymy of T. panthera, but without elaborating, and overlooking the clear distinctions in coloration present in the original description. It could be either a junior synonym of T. fuscomaculata, because of the described similarities in dorsal color (therefore extending its range considerably), a junior synonym of T. sinuspersici, or a valid species. Torpedo fuscomaculata has not been recorded from the northern Arabian Sea, and its occurrence there is not probable. There is no mention of any light-colored dorsal markings in the original description of T. zugmayeri, which is in strong contrast to T. sinuspersici (however, see below). In our view this species is either a synonym of T. sinuspersici (leaving open the issue regarding the lack of mention of white spots or vermiculations in the original description) or valid, but material from off western Pakistan is needed. Engelhardt (1912) compared his new species only to T. marmorata.

The accounts of Blegvad (1944) and Kuronuma and Abe (1986) from the Persian Gulf described a species (misidentified as T. panthera) that is brownish with darker irregular spots. Both accounts were based on freshly captured specimens and do not mention white, cream-colored, or lighter spots, mottles, or vermiculations indicative of T. sinuspersici, originally described from the Persian Gulf. Belgvad (1944: 44) even included T. zugmayeri as a synonym of T. panthera (the first citation that we could find of this species after its original description), noting however the differences in coloration between his specimens and Rüppell's (1835) account of T. panthera from the Red Sea. Blegvad's depiction (1944: pl. III, fig. 1; an adult male 350 mm TL) shows a specimen that slightly resembles T. sinuspersici, with some lightly colored blotches, but this is inconclusive. If T. zugmayeri is indeed a valid species, then perhaps these two accounts are the first to record it subsequently, extending its range to the Persian Gulf.

Other valid species of the subgenus Torpedo are primarily Mediterranean and west African in distribution and are easily distinguished from T. adenensis in coloration, as all present characteristic spots, ocelli, and blotches dorsally. These species are T. torpedo (Linnaeus, 1758), T. marmorata Risso, 1810, and T. bauchotae Cadenat, Capapé, and Desoutter, 1978. Torpedo andersoni Bullis, 1962, a species known only from the holotype and paratype from the Bahamas, has reddish spots and blotches over the dorsal disc, tail, and dorsal and caudal fins.

Indo-Pacific Nominal Species of the Subgenus Tetronarce

There are few records of species of Tetronarce in the Indian Ocean. One species has been reported from South Africa (as T. nobiliana Bonaparte, 1835), but it does not extend farther east than Algoa Bay (Compagno et al., 1989; Compagno, 1995). This species is uniformly dark brown, black, or gray dorsally, and may be distinct from true T. nobiliana and from another species, possibly new and also of the subgenus Tetronarce, that has been recorded from the Mozambique channel (Compagno, 1995). Duhamel (1989) confirmed the occurrence of a Tetronarce species from the southern Indian Ocean islands of St. Paul and Amsterdam, tentatively identified as T. macneilli. We have also examined a uniform dark brown specimen of Tetronarce (ZMUAS 3735) from off southwestern India (previously identified as T. fairchildi). None of these occurrences can be confused with T. adenensis.

All species of Torpedo from the Pacific Ocean, including those from the eastern Pacific, also belong to the subgenus Tetronarce. These species are, in the western Pacific (junior synonyms in brackets), T. fairchildi Hutton, 1872 [T. fusca Parker, 1884], T. tokionis (Tanaka, 1908), and T. macneilli (Whitley, 1932; a junior synonym of T. fairchildi?); and in the eastern Pacific, T. californica Ayres, 1855, and T. tremens de Buen, 1959 [?T. peruana Chirichigno, 1963,?T. microdiscus Parin and Kotylar, 1985,?T. semipelagica Parin and Kotylar, 1985]. The species described by Parin and Kotylar (1985) from the eastern Pacific submarine ridge of Sala y Gomez appear to differ only in disc and pelvic fin proportions, characters that are commonly preservational, and they may be synonymous. In fact, distinguishing both species from those occurring off the western coast of South America is not straightforward. The distinctions between T. fairchildi and T. macneilli from New Zealand and Australia, respectively, and even between T. tremens and T. peruana from Chile and Peru, respectively, are also in need of evaluation (as are the limits of most species of Tetronarce). However, all Indo-Pacific species of Tetronarce are drab, dark brown, gray, or blackish-purple dorsally, lacking papillae around the spiracles, and are therefore easily distinguished from T. adenensis.