Apteronotus ferrarisi, Santana & Vari, 2013

Santana, Carlos David De & Vari, Richard P., 2013, Brown ghost electric fishes of the Apteronotus leptorhynchus species-group (Ostariophysi, Gymnotiformes); monophyly, major clades, and revision, Zoological Journal of the Linnean Society 168 (3), pp. 564-596 : 577-586

APTERONOTUS FERRARISI SP. NOV.

( FIGS 5 , 9 , 10 ; TABLE 1)

Diagnosis: Apteronotus ferrarisi is distinguished from A. anu by the condition of the fifth infraorbital (absent versus present as a tubular laterosensory canal segment), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of scales above the lateral line at the midbody (10–13 versus seven to nine), the number of scales along the lateral line (73–80 versus 59–63), the number of caudal-fin rays (15–17 versus 10–13), the extent of the middorsal groove (extending two to six scales beyond the vertical through the posterior terminus of the anal fin versus seven to eight) and the tail length (6.9–9.9% of LEA versus 17.1–37.5%). Apteronotus ferrarisi differs from A. baniwa in the number of scales above the lateral line at the midbody (10–13 versus seven to eight), the number of scales along the lateral line (73–80 versus 61–70), the mouth length (49.4–56.4% of HL versus 36.6–47.2%), the tail length (6.9-9.9% of LEA versus 13.3–24.2%), and the tail depth (22.2– 28.7% of CL versus 8.2–15.3%). Apteronotus ferrarisi is differentiated from A. galvisi by the orbital diameter (4.0–5.9% of HL versus 6.4–11.3%) and the tail length (6.9-9.9% of LEA versus 11.8–15.9%). Apteronotus ferrarisi is distinguished from A. leptorhynchus by the mouth length (49.4–56.3% of HL versus 35.1–48.8) and the tail length (6.9-9.9% of LEA versus 17.1–18.6%). Apteronotus ferrarisi differs from A. macrostomus in the tail length (6.9-9.9% of LEA versus 13.9–20.6%), the number of caudal-fin rays (15–17 versus 18–21), and the number of total vertebrae (61–71 versus 75). Apteronotus ferrarisi is differentiated from A. pemon by the mouth length (49.4– 56.3% of HL versus 30–44.5%), the tail length (6.9-9.9% of LEA versus 13.0–17.5%), the tail depth (22.2–28.7% of CL versus 9.9–15.4%), and the number of caudal-fin rays (15–17 versus 18–20). Apteronotus ferrarisi is distinguished from A. rostratus by the condition of the fifth infraorbital (absent versus present as a tubular laterosensory canal segment), the condition of the lateral ethmoid (ossified versus cartilaginous), the tail length (6.9-9.9% of LEA versus 10.6–12.4%), and the number of scales along the lateral line (73–80 versus 63–67). Apteronotus ferrarisi differs from A. spurrellii in the condition of the fifth infraorbital (absent versus present as a tubular laterosensory canal segment), the condition of the lateral ethmoid (ossified versus cartilaginous), the mouth length (49.4–56.3% of HL versus 42.9–48.4%), the tail length (6.9-9.9% of LEA versus 12.4–28.7%), the number of scales above the lateral line at the midbody (10–13 versus eight to nine), the number of anal-fin rays (150–167 versus 171–173), and the number of caudal-fin rays (15–17 versus 14).

Description: Head and body shape and pigmentation illustrated in Figures 9 and 10 . Morphometrics for holotype and paratypes presented in Table 1. Body laterally compressed, greatest body depth located at, or slightly posterior to, abdominal cavity. Dorsal profile of body nearly straight. First perforated scale of lateral line located above pectoral-fin origin. Lateral line extending posteriorly to base of caudal fin. Scales along lateral line 73(1), 74(1), 77*(2), or 80(1). Scales above lateral line to mid-dorsal line at midbody 10(2), 11*(2), or 13(1).

Head laterally compressed, widest at opercular region and deepest at nape. Dorsal profile of head nearly straight but with snout longer in mature males ( Fig. 9 ) than mature females ( Fig. 10 ). Eye small, located laterally on head and completely covered by thin membrane. Anterior naris located at end of small tube and close to tip of snout. Posterior naris ellipsoidal, without tube and positioned closer to tip of snout than to anterior margin of eye. Mouth terminal with rictus extending posterior of vertical through anterior border of eye. Branchial opening located anterior to vertical through pectoral-fin insertion. Anus and urogenital papillae located posterior to vertical through eye and not varying ontogenetically in position.

Pectoral fin elongate, with 13(1), 15*(1), 16(2), or 17(1) rays. Anal-fin origin located anterior to vertical through posterior margin of opercle. Unbranched anal-fin rays 21(1), 23(1), 24(1), 25*(1), or 33(1); total anal-fin rays 150(1), 153*(1), 156(1), or 167(1). Tail compressed and short; ending in small, elongate, caudal fin. Caudal-fin rays 15*(3) or 17(1).

Origin of dorsal sagittal electroreceptive filament located on posterior half of body. Filament inserted into narrow mid-dorsal groove and extending 2(3), 4*(1), or 6(1) scales posterior of vertical through posterior terminus of anal fin. Precaudal vertebrae 17(2) or 18*(3). Total vertebrae 61(1), 68*(1), 70(2), or 71(1).

Secondary sexual dimorphism: Sexually mature males (N = 2, 231– 274 mm TL; Fig. 9 ) and females (N = 3, 94– 111 mm TL, Fig. 10 ) of A. ferrarisi differ in diverse components of head morphology. The length of the snout, mouth length, and the distance from the posterior naris to the eye are proportionally longer in mature males than in mature females ( Table 1).

Coloration in alcohol: Body dark and head dark brown. Prominent pale, broad, longitudinal band extending from chin along dorsal midline of head and body to beginning of dorsal filament. Pectoral and anal-fin rays dark brown and inter-radial membranes translucent. Pale band encircling base of caudal fin. Caudal fin pale at base and dark brown posteriorly.

Distribution: Apteronotus ferrarisi is known from the type locality in the Río Yaracuy and also occurs in the adjoining Río Aroa, both of which are coastal Caribbean versant drainages of northern Venezuela ( Fig. 5 ). These rivers are a centre of endemism within the Tocuyo-Yaracuy subprovince as defined by Rodríguez-Olarte et al. (2009: 87), with 24 endemic fish species previously reported in these basins.

Etymology: The species name, ferrarisi , is in honour of Carl J. Ferraris Jr, in recognition of his many contributions to our knowledge of tropical freshwater fishes worldwide and his invaluable assistance to the authors, particularly the second author, over the years.

Remarks: The report of Apteronotus sp. from the Tocuyo-Yaracuy region by Rodríguez-Olarte et al. (2009: 73) is presumably based on A. ferrarisi , which is the only member of the genus known to occur in those basins.

Material examined: Holotype: Venezuela. Yaracuy. MCNG 55635 , male, 231 mm; Río Yaracuy basin (Caribbean coast), Quebrada Guaquira on Hacienda Guaquira, south-east of El Peñon Reservoir , 10°17′40″N, 68°39′35″W, collected by D. Rodríguez, A. Amaro, H. Aguldelho, J. Coronel & D. Taphorn, 23.x.2005 . Paratypes: Venezuela. Yaracuy. MCNG 54584 , 1 (male), 160 mm; collected at holotype locality, 21.x.2005. MCNG 54597 , 2 (females), 176– 196 mm; collected at holotype locality. MCNG 52077 , 1 (male), 274 mm; Río Carabobo, Río Aroa drainage (Caribbean coast), 10°30′08″N, 68°47′02″W, D. Taphorn et al., 15.i.2004 .

APTERONOTUS GALVISI DE SANTANA , MALDONADO- OCAMPO & CRAMPTON

( FIGS 5 , 11 ; TABLE 2)

Apteronotus galvisi de Santana, Maldonado-Ocampo & Crampton, 2007: 118 , fig. 1 [original description; type-locality Río Cusiana, piedmont of Cordillera Oriental, 305–424 m asl, upper Río Meta; secondary sexual dimorphism]. – Maldonado-Ocampo et al., 2008: 213 [ Colombia]. – Urbano-Bonilla et al., 2009: 162 [ Colombia, Departamento de Casanare, Ríos Pauto, Cravo Sur, Cusiana, Upía]. – Machado-Allison et al. 2010: 242 [Río Meta]. – de Santana & Cox-Fernandes, 2012: 284 [Río Orinoco basin; occurrence in streams; presence of secondary sexual dimorphism].

Diagnosis: Apteronotus galvisi is differentiated from A. anu by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of scales above the lateral line at the midbody (11–12 versus seven to nine), and the tail length (11.8–15.9% of LEA versus 17.1–37.5%). Apteronotus galvisi is distinguished from A. baniwa by the number of scales above the lateral line at the midbody (11–12 versus seven to eight). Apteronotus galvisi differs from A. ferrarisi in the orbital diameter (6.4–11.3% of HL versus 4.0–5.9%) and the tail length (11.8–15.9% of LEA versus 6.9-9.9%). Apteronotus galvisi is differentiated from A. leptorhynchus by the number of caudal-fin rays (10–16 versus 17–21), the length of the anal-fin base (69.4–81.1% of LEA versus 81.7–83.7%), and the tail length (11.8–15.9% of LEA versus 17.1–18.6%). Apteronotus galvisi is distinguished from A. macrostomus by the number of caudal-fin rays (10–16 versus 18–21). Apteronotus galvisi differs from A. pemon in the ocular diameter (6.4–11.3% of HL versus 4.0– 6.3%), the number of scales above the lateral line at the midbody (11–12 versus eight to nine), and the number of caudal-fin rays (10–16 versus 18–20). Apteronotus galvisi is differentiated from A. rostratus by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), and the number of scales above the lateral line at the midbody (11–12 versus nine to ten). Apteronotus galvisi is distinguished from A. spurrellii by the ocular diameter (6.4–11.3% of HL versus 3.7-5.8%), the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the number of scales above the lateral line at the midbody (11–12 versus eight to nine), the number of anal-fin rays (145–165 versus 171–173), and the condition of the lateral ethmoid (ossified versus cartilaginous).

Description: Head, body, and fin shape and pigmentation illustrated in Figure 11 . Morphometrics for holotype and paratypes presented in Table 2. Body laterally compressed, greatest body depth located at, or slightly posterior to, abdominal cavity. Dorsal

N, number of specimens; H, holotype. Range includes holotype. Tail depth data unavailable for A. galvisi

profile of body nearly straight. First perforated scale of lateral line located above pectoral-fin origin. Lateral line extending posteriorly to base of caudal fin. Scales above lateral line to mid-dorsal line at midbody 11(11) or 12*(4).

Head laterally compressed, widest at opercular region and deepest at nape. Dorsal profile of snout and overall head nearly straight with snout longer in mature males than mature females. Eye small, located laterally on head, and completely covered by thin membrane. Anterior naris located at end of small tube and close to tip of snout. Posterior naris ellipsoidal, without tube and positioned closer to tip of snout than to anterior margin of eye. Mouth terminal with rictus extending posterior of vertical through anterior border of eye. Branchial opening located anterior to vertical through pectoral-fin insertion. Anus and urogenital papillae located posterior to vertical through eye without ontogenetic variation in position.

Pectoral fin elongate, with 13*(3), 14(3), 15(7), 16(2), or 17(1) rays. Anal-fin origin located anterior to vertical through posterior margin of opercle. Total anal-fin rays 145(1), 146(2), 147(1), 148(2), 152*(3), 153(1), 155(1), 156(1), 161(1), or 165(2). Tail compressed and short; ending in small, elongate caudal fin. Caudal-fin rays 10 (2), 11(1), 12*(6), 14(2), or 16(1).

Origin of dorsal sagittal electroreceptive filament situated on posterior half of body. Filament inserted into narrow mid-dorsal groove and almost extending to, or slightly beyond, vertical through posterior terminus of anal-fin base.

Secondary sexual dimorphism: Sexually mature males (N = 3, 160– 174 mm TL) and females (N = 9, 124– 189 mm TL) of A. galvisi differ in overall head morphology with the dorsal profile of head straight in mature males versus slightly curved ventrally in immature and mature females ( Fig. 11 ; de Santana et al., 2007: fig. 1).

Coloration in alcohol: Body and head grey to light brown. Prominent, broad, cream-coloured or yellow band extending from chin along dorsal midline of head and body to beginning of dorsal filament. Pectoral and anal-fin rays brown with inter-radial membranes hyaline. Pale band encircling base of caudal fin. Caudal fin light coloured basally and brown distally.

Distribution: Apteronotus galvisi is known from the Ríos Cusiana, Cravo, Pauto, and Upia in the piedmont of the Cordillera Oriental, 305–424 m asl, in the upper Río Meta system, Colombia ( Fig. 5 ).

Electric organ discharges (EODs): As is the case in all other recorded species of the Apteronotidae , A. galvisi generates wave-type EOD with frequencies in the species ranging from 700–957 Hz, without overlapping frequencies between mature males and females ( de Santana et al., 2007).

Remarks: The Río Orinoco basin is home to four of the nine species in the A. leptorhynchus species-group. This high percentage correlates with this river system being by far the largest basin within the range of the species-group. Two of the species, A. baniwa and A. pemon , are limited to southern portions of the Orinoco basin with known areas of occurrence distant from both each other and from A. galvisi and A. macrostomus , the two other members of the species-group in the system. Apteronotus galvisi and A. macrostomus , in contrast, both occur within a relatively limited region in the western portion of the Orinoco basin.

Notwithstanding their geographical proximity, A. galvisi and A. macrostomus have a distinct separation in the ranges of caudal-fin rays (10-16 versus 18-21, respectively); a difference all the more noteworthy because the samples of these two species are the largest available in the A. leptorhynchus speciesgroup. The results of the PCA showed the separation between A. galvisi and A. macrostomus along the second and third principal components ( Fig. 12 ). PC2 and PC3 reflected 4.6 and 1.4% of the variation, respectively. The best predicators of the differences between the two species were: the internarial dis- tance (0.54), the ocular diameter (0.76), the interorbital distance (0.58), the branchial opening (0.44), and the pectoral fin length (0.35). The jack-knife crossvalidation procedure classified 100% of the specimens between these species, further supporting the recognition of these as valid species. Finally, although mature males of both species are relatively rare, it is noteworthy that mature males of A. galvisi (three specimens, 160–174 mm TL) can be distinguished from the only known mature male of A. macrostomus (the holotype, 265 mm TL) by the head length (18.6– 20.3% of LEA versus 21.4%), the head width (25.3– 36.5% of HL versus 39.3%), the head depth (37.9– 39.4% of HL versus 50.1%), the snout length (40.0– 43.8% of HL versus 44.7%), the distance from the posterior naris to the eye (17.8–20.3% of HL versus 23.6%), the width of the mouth (40.6–50.1% of HL versus 56.2%), the intraocular width (14.1–18.6% of HL versus 19.9%), and the postocular distance (49.2– 58% of HL versus 62.1%). Many of these differences are more pronounced than those between mature males of different species in other groups within the Apteronotidae ; however, additional material is necessary to test whether these differences, or some subset of them, serve to unambiguously delimit the two species.

Material examined: Colombia. Casanare. IAvHP 8133 , holotype, male, 184.4 mm TL; Río Meta drainage, Río Cusiana, at El Venado bridge, 305 m asl, 05°00′50′′N, 72°41′30″W. Paratypes: IAvHP 8130 , two immatures, 97.5–99.2 mm TL; same data as holotype. IAvHP 8129 , three males, 125–174.2 mm TL, three females 132.8–163.3 mm TL; Yopal, Quebrada La Aguatoca at bridge on road to Morro, 357 m asl, 05°26′21″N, 72°27′09″W. IAvHP 8131 , two females, 124.8– 162.6 mm TL; Río Pauto at bridge on road to Yopal- Pore, 306 m asl, 05°33′44′′N, 72°08′44″W. IAvHP 8132 , one immature, 98.3 mm TL, one female, 158.4 mm TL, Río Upia at bridge on road to Sisga, 327 m asl, 04°49′09″N, 73°04′57″W. IAvHP 8167 , three females, 170.6–189.8 mm TL; Yopal, Río Cravo Sur at Colgante bridge, 424 m asl, 05°25′37″N, 72°27′11″W .

APTERONOTUS LEPTORHYNCHUS (ELLIS)

( FIGS 5 , 13 ; TABLE 2)

Sternarchus leptorhynchus Ellis in Eigenmann, 1912: 439 [original description; type-locality Amatuk, Guyana]. – Ellis, 1913: 147, pl. 23, fig. 4 [redescription based on Ellis, in Eigenmann, 1912; first illustration of species].

Apteronotus leptorhynchus, Fowler, 1951: 426 [comb. nov.; in part, citations from Guyana; not citations from upper Amazon]. – Lasso, Machado-Allison & Pérez Hernández, 1989: 122 [ Essequibo River]. – Mago-Leccia, 1994: 29 [in listing of species; not illustrated specimen from Río Orinoco basin]. – Albert, 2003: 498 [in part; in listing of species; Guyana; not citations from other regions of South America]. – Albert, 2009: 46 [in part; Guyana; not citations from other regions of Guiana Shield]. – de Santana & Cox-Fernandes, 2012: 284 [ Essequibo River basin; occurrence in streams].

Diagnosis: Apteronotus leptorhynchus is differentiated from A. anu by the condition of the fifth infraor- bital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of scales above the lateral line at the midbody (12–13 versus seven to nine), the number of pectoral-fin rays (17–18 versus 14–16), the number of anal-fin rays (151–156 versus 159–176), the number of caudal-fin rays (17–21 versus 10–13), and the number of scales along the lateral line (78–82 versus 59–63). Apteronotus leptorhynchus differs from A. baniwa in the number of scales above the lateral line at the midbody (12–13 versus seven to eight), the number of scales along the lateral line (78–82 versus 61–70), the number of pectoral-fin rays (17–18 versus 14–16), the number of precaudal vertebrae (18 versus 16–17), and the number of total vertebrae (78–81 versus 68–77). Apteronotus leptorhynchus is distinguished from A. ferrarisi by the mouth length (35.1–48.8% of HL versus 49.4–56.3%) and the tail length (17.1–18.6% of LEA versus 6.9-9.9%). Apteronotus leptorhynchus is differentiated from A. galvisi by the number of caudal-fin rays (17–21 versus 10–16), the length of the anal-fin base (81.7–83.7% of LEA versus 69.4– 81.1%), and the tail length (17.1–18.6% of LEA versus 11.8–15.9%). Apteronotus leptorhynchus differs from A. macrostomus in the number of pectoral-fin rays [17–18 (17 pectoral-fin rays present in one of four specimens examined) versus 14–17 (17 pectoral-fin rays present in two of 13 specimens examined)]; in addition, the holotypes of A. leptorhynchus and A. macrostomus , mature males of nearly identical sizes, differ from each other in the head depth (43.6% of HL versus 32.5%), the snout length (41.9% of HL versus 44.7%), the mouth length (48.8% of HL versus 56.2%), and the postorbital distance (57.9% of HL versus 48.7%). Apteronotus leptorhynchus is distinguished from A. pemon by the number of scales above the lateral line at the midbody (12–13 versus eight to nine) and the number of pectoral-fin rays (17–18 versus 16). Apteronotus leptorhynchus is differentiated from A. rostratus by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the number of scales above the lateral line at the midbody (12–13 versus nine to ten), the number of caudal-fin rays (17–21 versus 14–16), the number of scales along the lateral line (78–82 versus 63–67), the tail length (17.1–18.6% of LEA versus 10.6–12.4%), and the condition of the lateral ethmoid (ossified versus cartilaginous). Apteronotus leptorhynchus differs from A. spurrellii in the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of pectoral-fin rays (17–18 versus 12 -13), the number of anal-fin rays (151–156 versus 171–173), the number of scales along the lateral line (78–82 versus 73–76), and the extent of the mid-dorsal groove (extending eight scales beyond the vertical through the posterior terminus of the anal fin versus four scales).

Description: Head, body, and fin shape and pigmentation illustrated in Figure 13 . Morphometrics for holotype and paratypes presented in Table 2. Body laterally compressed, greatest body depth located at, or slightly posterior to, abdominal cavity. Dorsal profile of body nearly straight. First perforated scale of lateral line located above pectoral-fin origin. Lateral line extending posteriorly to base of caudal fin. Scales along lateral line 78(1), 80(1), or 82*(1). Scales above lateral line to mid-dorsal line at midbody 12*(3) or 13(1).

Head laterally compressed, widest at opercular region and deepest at nape. Dorsal profile of snout and overall head nearly straight. Eye small, located laterally on head, and completely covered by thin membrane. Anterior naris located at end of small tube and close to tip of snout. Posterior naris ellipsoidal, without tube and positioned closer to tip of snout than to anterior margin of eye. Mouth terminal with rictus situated posterior of vertical through anterior margin of eye. Branchial opening located slightly anterior to vertical through pectoral-fin insertion. Anus and urogenital papillae located posterior to vertical through eye and without ontogenetic variation in position.

Pectoral fin elongate, with 17(2) or 18*(2) rays. Anal-fin origin located at, or slightly anterior to, vertical through posterior margin of opercle. Unbranched anal-fin rays 19*(1) or 22(2); total anal-fin rays 151(1), 152*(1), or 156(1). Tail compressed and short; ending in small, elongate caudal fin. Caudal-fin rays 17(1), 19(1), or 21*(1).

Origin of dorsal sagittal electroreceptive filament located on posterior half of body. Filament inserted into narrow mid-dorsal groove extending 8*(2) scales beyond vertical through posterior terminus of anal fin. Precaudal vertebrae 18*(4). Total vertebrae 78(2) or 81*(1).

Coloration in alcohol: Body and head light to dark brown. Prominent, pale band extending from chin along dorsal midline of head and body to beginning of dorsal filament. Pectoral- and anal-fin rays dark brown with inter-radial membranes translucent. Pale band encircling base of caudal fin. Caudal fin pale at base and dark brown posteriorly.

Distribution: Apteronotus leptorhynchus is known only from the Essequibo River drainage in Guyana ( Fig. 5 ).

Remarks: Apteronotus leptorhynchus has been reported from broad regions of northern South America (cis- and trans-Andean regions of Colombia – Mojica-C, 1999: 563; Lago Maracaibo – Mago-Leccia, 1970: 77; Taphorn et al., 1997: 80; Río Caura, Rodríguez-Olarte et al., 2003: 198; Lasso et al., 2003: 238; Lasso et al., 2010: 70; Río Orinoco basin and Río Paria – Cala, 1977: 10; Lasso et al., 2004a: 181, 2004b: 142; Pacific and Caribbean drainages of Colombia – Maldonado-Ocampo et al., 2008: 213; Peru – Fowler, 1945: 182, 1951: 426; and Venezuela – Lasso et al., 2004a: 181). Available information, however, indicates that A. leptorhynchus is endemic to the parts of the Essequibo River basin in Guyana. The above records are either misidentifications of specimens of other species-groups or may be based on species described as new herein. Lasso et al. (2010: 70) reported A. leptorhynchus from the Río San Juan, a river draining into the Golfo de Paria north of the mouth of the Río Orinoco. Owing to lack of access to the voucher specimen we were unable to evaluate this record.

Although the primary diagnostic feature discriminating A. leptorhynchus from A. macrostomus , the number of pectoral-fin rays, shows a limited degree of overlap in the available samples of the species, the holotypes of the two species, both mature males of approximately the same body size, show trenchant differences in various details (see Diagnosis for each species). In addition, various features of mature males not readily expressed via standard meristics and morphometrics differ notably between mature males of the two species. Most obvious of these is the form of the snout and dorsal surface of the head, which is straight in the case of A. leptorhynchus versus notably convex in A. macrostomus (see Figs 13 , 15 ). Furthermore, the results of the PCA showed partial notable, albeit not absolute, separation between A. leptorhynchus and A. macrostomus along the first and second principal components ( Fig. 14 ). PC1 and PC2 reflected 91.6 and 5.8% of the variation, respectively. The best predicators of the differences between the two species were: the distance from the posterior snout to the eye (0.36) and the head depth (-0.94). A jack-knife cross-validation procedure correctly classified 86.6% of the specimens by species, further supporting the hypothesis of the differences between the two species. Those morphological differences, supplemented by the distinct gap in the known distributions of the two species, support the recognition of the two nominal forms as valid.

Material examined: Guyana. BMNH 1974.5 .22.136, 1, 124; Potaro River . CAS 62333, 1 , 100 ; Essequibo River , Potaro River at Amatuk, 1908. FMNH 53294 , 1 , 260 ; holotype, Amatuk, 1908. FMNH 53295 , 1 , 95 , Warraputa, 1908. INHS 49524 , 1 , 94 ; Potaro River , Essequibo River basin, Amatuk Cataract, Macaroni- Potaro 05°18′13″N, 59°18′40″W .

APTERONOTUS MACROSTOMUS (FOWLER)

( FIGS 15 , 16 ; TABLE 2)

Sternarchus macrostomus Fowler, 1943: 263 , fig. 63 [original description; type-locality: Villavicencio, Río Meta basin, Colombia].

Apteronotus macrostomus, Mago-Leccia, 1994: 30 , fig. 35 [comb. nov.; in listing of species]. – Albert, 2003: 499 [in listing of species; Río Meta basin, Colombia]. – Lasso et al., 2004a: 181 [Orinoco basin]. – Lasso et al., 2004b: 142 [western Orinoco basin]. – de Santana & Cox-Fernandes, 2012: 284 [Río Orinoco basin; occurrence in rivers].

‘Apteronotus’ macrostomus, Albert & Campos-da- Paz, 1998: 431 [phylogenetic position].

Diagnosis: Apteronotus macrostomus is distinguished from A. anu by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of anal-fin rays (140–152 versus 159–176), the number of caudalfin rays (18–21 versus 10–13), and the number of scales above the lateral line at the midbody (11–14 versus seven to nine). Apteronotus macrostomus is differentiated from A. baniwa by the number of scales above the lateral line at the midbody (11–14 versus seven to eight), the number of caudal-fin rays (18–21 versus 14–17), and the number of precaudal vertebrae (18 versus 16–17). Apteronotus macrostomus differs from A. ferrarisi in the tail length (13.9–20.6% of LEA versus 6.9-9.9%), the number of caudal-fin rays (18–21 versus 15–17), and the number of total vertebrae (75 versus 61–71). Apteronotus macrostomus is distin- guished from A. galvisi by the number of caudal-fin rays (18–21 versus 10–16). Apteronotus macrostomus is differentiated from A. leptorhynchus by the number of pectoral-fin rays [14–17 (17 pectoral-fin rays present in two of 13 specimens examined) versus 17–18 (17 pectoral-fin rays present in one of four specimens examined)]; in addition, the holotypes of A. macrostomus and A. leptorhynchus , both mature males of nearly identical sizes, differ from each other in the head depth (32.5% of HL versus 43.6%), the snout length (44.7% of HL versus 41.9%), the mouth length (56.2% of HL versus 48.8%), and the postorbital distance (48.7% of HL versus 57.9%). Apteronotus macrostomus is distinguished from A. pemon by the number of scales above the lateral line at the midbody (11–14 versus eight to nine). Apteronotus macrostomus is distinguished from A. rostratus by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of scales above the lateral line at the midbody (11–14 versus nine to ten), the number of anal-fin rays (140–152 versus 153–162), the number of caudal-fin rays (18–21 versus 14–16), and the tail length (13.9– 20.6% of LEA versus 10.6–12.4%). Apteronotus macrostomus is differentiated from A. spurrellii by the condition of the fifth infraorbital (present as a tubular laterosensory canal segment versus absent), the condition of the lateral ethmoid (ossified versus cartilaginous), the number of pectoral-fin rays (15–17 versus 12–13), the number of scales above the lateral line at the midbody (11–14 versus eight to nine), the number of anal-fin rays (140–152 versus 171–173), and the number of caudal-fin rays (18–21 versus 14).

Description: Head, body, and fin shape and pigmentation illustrated in Figure 15 . Morphometrics for holotype and paratypes presented in Table 2. Body laterally compressed with greatest body depth located at, or slightly posterior to, abdominal cavity. Dorsal profile of body nearly straight. First perforated scale of lateral line located above pectoral-fin origin. Lateral line extending posteriorly to base of caudal fin. Scales along lateral line 79(3), 80(4), or 81*(5). Scales above lateral line to mid-dorsal line at midbody 11(5), 12*(3), 13(1), or 14(3).

Head laterally compressed, widest at opercular region and deepest at nape. Dorsal profile of head nearly straight. Eye small, located laterally on head, and completely covered by thin membrane. Anterior naris located at end of small tube and close to tip of snout. Posterior naris ellipsoidal, without tube and positioned closer to tip of snout than to anterior margin of eye. Mouth terminal with rictus extending posterior of vertical through anterior border of eye. Branchial opening located slightly anterior to vertical through pectoral-fin insertion. Anus and urogenital papillae located posterior to vertical through eye without ontogenetic variation in position.

Pectoral fin elongate, with 14*(3), 15(2), 16(2), or 17(5) rays. Anal-fin origin located at, or slightly anterior to, vertical through posterior margin of opercle. Unbranched anal-fin rays 19*(1), 22(4), 23(2), 25(3), or 27(2); total anal-fin rays 140(1), 141(1), 143(2), 146*(3), 148(1), or 152(1). Tail compressed and short; ending in small, elongate caudal fin. Caudal-fin rays 18(3), 19(4), or 21*(3).

Origin of dorsal sagittal electroreceptive filament located on posterior half of body. Filament inserted into narrow mid-dorsal groove ranging from slightly anterior to slightly posterior of vertical through posterior terminus of anal fin. Precaudal vertebrae17(3) or 18*(2). Total vertebrae 71(1), 73(2), 74(2), or 75(1).

Coloration in alcohol: Body and head lightly coloured to dark brown. Pale, broad, longitudinal band extending from chin along dorsal midline of head and body to beginning of dorsal filament. Pectoral- and anal-fin rays dark brown with inter-radial membranes translucent. Pale band encircling base of caudal fin. Caudal fin pale basally and dark brown distally.

Distribution: Apteronotus macrostomus is known from the Río Meta at Villavicencio, Colombia ( Fig. 16 ).

Remarks: See under Remarks for A. galvisi concerning differences between that species and A. macrostomus and the diagnostic features discriminating the single available mature male (holotype) of A. macrostomus from mature males of A. galvisi . Features distinguishing A. galvisi from A. leptorhynchus are discussed in the Remarks for the latter species.

Material examined: Colombia. Meta. ANSP 70528, 1, holotype, 265; Río Meta at Villavicencio (4°08′S, 73°40′W). CAS [ SU 23728 ], 1, 205; Guaicaramo. FMNH 92638 , 8 , 83–145 ; Caño Negro on road to Puerto Porfia , east of Villavicencio. FMNH 92715 , 1 , 85 , Campo Alegre , Campo Aguasclaras east of Villavicencio. USNM 100808 , 1 , 172 , Villavicencio , Río Meta basin. UF 33218, 1, 157; Río Meta, Caño Quenane (4°02′S, 73°10′W). UF 36604, 5, 71–230, Río Guatiguia, near Villavicencio .

Figures

Figure 5. Map of north portion of South America showing geographical distribution of Apteronotus anu (1 = holotype locality); Apteronotus baniwa (2 = holotype locality); Apteronotus ferrarisi (3 = holotype locality); Apteronotus galvisi (4 = holotype locality), and Apteronotus leptorhynchus (5 = holotype locality). Some symbols represent more than one locality and/or lot of specimens.

Figure 9. Apteronotus ferrarisi, 231 mm total length, male, holotype, MCNG 55635; Venezuela, Yaracuy, Río Yaracuy basin, Quebrada Guaquira on Hacienda Guaquira, south-east of El Peñon Reservoir, 10°17′40″N, 68°39′35″W.

Figure 10. Apteronotus ferrarisi, 196 mm total length, female, paratype, MCNG 54597; Venezuela, Yaracuy, Río Yaracuy basin, Quebrada Guaquira on Hacienda Guaquira, south-east of El Peñon Reservoir, 10°17′40″N, 68°39′35″W.

Figure 11. Apteronotus galvisi, top, 184.4 mm total length, male, holotype, IAvHP 8133; bottom, 162.6 mm total length, female (missing caudal fin), paratype, IAvHP 8131. Colombia. Casanare, Río Meta drainage, Río Cusiana, at El Venado bridge, 305 m asl, 05°00′50″N, 72°41′30″W.

Figure 12. Principal components analysis on covariance matrix of log-transformed measurements of Apteronotus galvisi (squares) and Apteronotus macrostomus (circles). Scatter plot of scores on second (PC2) and third (PC3) principal components.

Figure 13. Apteronotus leptorhynchus, 260 mm total length, male, holotype, FMNH 53294; Guyana, Amatuk.

Figure 14. Principal components analysis on covariance matrix of log-transformed measurements of Apteronotus leptorhynchus (circles) and Apteronotus macrostomus (squares). Scatter plot of scores on first (PC1) and second (PC2) principal components.

Figure 15. Apteronotus macrostomus, 265 mm total length, male, holotype, ANSP 70528; Colombia, Meta, Río Meta at Villavicencio.

Figure 16. Map of north portion of South America showing geographical distribution of Apteronotus macrostomus (1 = holotype locality); Apteronotus pemon (2 = holotype locality); Apteronotus rostratus (3 = holotype locality); and Apteronotus spurrellii (4 = type locality). Some symbols represent more than one locality and/or lot of specimens.