Astyanax varzeae, Vinícius Abilhoa & Luiz Fernando Duboc, 2007

Vinícius Abilhoa & Luiz Fernando Duboc, 2007, A new species of the freshwater fish genus Astyanax (Ostariophysi: Characidae) from the rio Iguacu basin, southeastern Brazil., Zootaxa 1587, pp. 43-52 : 45-50

Astyanax varzeae , new species

(Figs. 2-4; Table 1)

Astyanax sp. F . Haluch & Abilhoa, 2005: 384 (fig.1), 387 (comparative material).

Holotype. MHNCI 11688 , 83.1 mm SL, Brazil, Paraná, Tijucas do Sul, rio Sao Joao , rio da Varzea headwaters, an affluent of rio Negro, upper rio Iguacu basin , approx. 25°55’S 49°10’W, A. Schwarz, March 1999.

Paratypes. Collected with the holotype: MCP 40535 , 14, 43.9-72.0 mm SL ; MHNCI 9146 , 1 c&s, 53.1 mm SL ; MHNCI 9147 , 15, 43.0-81.8 mm SL (including 2 c&s) .

Diagnosis. The combination of shallow body depth (28.7-32.9% vs. more than 35% of SL) and the lower number of branched anal-fin rays (15-20 vs. 20-45) include Astyanax varzeae in the Astyanax scabripinnis species complex of Bertaco & Lucena (2006). Astyanax varzeae differs from A. ojiara and A. paris by its shallower body depth (28.7-32.9% vs. 34.0-40.0% of SL in A. ojiara and A. paris ). Shallow body depth (28.7- 32.9%, mean=30.1) also differs Astyanax varzeae from A. ita (34.3-39.8%, mean=37.1), A. rivularis (35.0- 46.0%), A. chico (37.3-43.0%, mean=40.8), A. tumbayaensis (39.2-45.3%, mean=42.4), A. hermosus (33.9- 38.7%, mean=36.3 in females, 36.1 in males), A. endy (36.6-42.2%, mean=39.7 in females, 38.4 in males), A. pampa (36.9-42.2%, mean=38.7), and A. laticeps (35.3-40.6%, mean=37.9). Astyanax intermedius can be distinguished from A. varzeae by its smaller dorsal-fin length (10.7-16.4%, mean=13.4 vs. 19.9-25.9%, mean=23.2). Astyanax varzeae differs from A. giton by its larger head length relative to SL (27.1-29.7%, mean=28.6 vs. 21.4-26.3%, mean=23.7), and from A. taeniatus by its shorter snout length compared to head length (16.4-23.3%, mean=20.1 vs. 25.3-32.6%, mean=28.6). Astyanax varzeae is distinguished from A. jenynsii by its larger orbital diameter relative to head length (30.9-44.5%, mean=37.9 vs. 24.7-27.6%, mean=26.3), and from A. obscurus by the greater interorbital width (29.8-37.7%, mean=34.5 vs. 27.7-28.6%, mean=28.2). The presence of bony hooks on all fins but pectoral-fin rays of males separates A. leonidas and A. troya from A. varzeae , and the presence of bony hooks on pelvic and anal-fin rays of males separates A. totae , A. cremnobates and A. brachypterygium from A. varzeae . The occurrence of two humeral spots (second one faint) distinguishes A. varzeae from A. paranae , A. microschemos , and A. pelecus (single humeral spot).

Description. Morphometric data of holotype and paratypes are presented in Table 1. Body moderately elongated, compressed, greatest depth at dorsal fin origin, 3.1 to 3.5 times in SL.

Dorsal profile of head straight. Dorsal body profile between the supraoccipital spine through dorsal-fin origin gently convex, and almost straight from the end of dorsal-fin base to caudal peduncle. Base of dorsal fin straight. Ventral profile of head gently rounded. Body profile from the margin of lower lip to anal-fin origin convex. Anal-fin base straight. Caudal peduncle elongated with dorsal and ventral profiles nearly straight.

Snout profile rounded, shorter than orbital diameter. Postero-ventral edge of third infraorbital distant from preopercle, leaving wide naked border between these bones. Mouth terminal, angled posteroventrally. Maxilla reaching a vertical line through the middle of the orbit, with 1 to 3 (usually 2) tricuspid teeth.

Premaxilla with two tooth rows. Outer row with 3 (2), 4* (21) tricuspid teeth, with central cuspid developed. Inner row with 5 teeth gently expanded distally, slightly compressed at distal tips. Symphysial tooth narrower and deeper, with 4 or 5 cusps. Second, third and fourth teeth with 5 cusps. Fifth tooth smaller, with 4 cusps. In all teeth central cusp slightly larger than remaining ones. Dentary with 7 teeth. First four anterior teeth larger, usually with 5 cusps, followed by 3 very smaller tricuspid teeth (Fig. 4).

Dorsal-fin rays ii, 9. Posterior margin of dorsal-fin straight. Dorsal-fin origin approximately at middle of SL. Pelvic-fin rays i, 7. Pelvic-fin origin located slightly anterior to vertical through dorsal-fin origin. Pectoral-fin rays 11 (3), 12 (14), 13* (13). Anal-fin with iii-iv, 15 (1), 16 (5), 17* (16), 18 (7), 20 (1) rays. Adiposefin located approximately at vertical through insertion of base of last anal-fin rays. Caudal-fin forked, lobes similar in size, slightly rounded, and with 19 principal rays.

Lateral line complete, with 37 (1), 38 (6), 39* (16), 40 (6), 42 (1) perforated scales. Six series of scales between dorsal-fin origin and lateral line. Four or five series of scales between lateral line and pelvic-fin insertion. Circumpeduncular scales 14* (4), 15 (7), 16 (19).

In three cleared and stained specimens procurrent caudal-fin rays 8 or 10 dorsal and ventral. First gill arch with 19 gill rakers (7 epibranchial, 9 ceratobranchial, 1 on cartilage between epibranchial and ceratobranchial, and 2 hypobranchial), vertebrae 36, 37 or 38 (18 precaudal and 18, 19 or 20 caudal). Supraneurals 5.

Color in alcohol. Preserved specimens with dorsal and lateral parts of the body medium brown to pale yellowish. Dark chromatophores concentrated on middorsal surface of head and body. Humeral spot narrow and vertically elongated, two scales wide above lateral line. Body with black, pigmented, midlateral stripe extending from humeral region to the median caudal-fin rays. The anterior portion of the stripe expanded, sometimes forming a faint second humeral spot, more clearly visible in smaller specimens (Fig. 3). Fins with scattered dark chromatophores.

Sexual dimorphism. None of the specimens examined have bony hooks on the fins or any other sexually dimorphic feature.

Etymology. The species name, varzeae, is in reference to the sample locality, rio da Várzea drainage, a tributary of the upper rio Iguaçu basin.

Distribution. Astyanax varzeae is known only from the rio da Várzea headwaters, an affluent of rio Negro, a tributary of the upper portions of the rio Iguaçu in Paraná State, Brazil.

Ecological notes. Astyanax varzeae was collected in a headwater stream with clear water, light to moderate current water and bottom with stones, sand and moderate amount of vegetal debris.

Discussion

Despite the fact that the Iguaçu River is a large tributary of the Parana River, many fishes that occur in rio Paraná basin are not found in rio Iguaçu basin. In fact, the fish fauna of this hydrographic basin has a unique evolutionary history, and the endemism of its ichthyofauna is probably related with the geographic isolation imposed by the Iguaçu falls (Severi & Cordeiro 1994; Garavello et al. 1997).

Furthermore, recent collecting efforts in the upper rio Iguaçu basin showed that the biogeographic history of the ichthyofauna in that watershed is likely to be more complex than described herein. The fish fauna of the upper rio Iguaçu basin, for example, consists mostly of species that does not occur in the lower portions of the river, and also components endemic to the coastal drainage.

The characterization of the upper reaches of the rio Iguaçu basin as an area of endemism was already proposed by Abilhoa (2004) and Ingenito et al. (2004), who subdivided the upper rio Iguaçu basin in two portions: rio Negro basin and rio Iguaçu headwaters. These authors listed several species with their geographic distribution restricted in each region ( Pareiorhaphis parmula , Astyanax totae , Ancistrus sp., Rineloricaria sp., Heptapterus stwearti , Trichomycterus naipi , Cnesterodon carnegiei , and Jenynsia eigenmanni ). The identification of a new species of Astyanax endemic to one portion of the upper Iguaçu river basin supports that hypothesis.

Regarding still undescribed Astyanax species occurring in the rio Iguaçu basin recorded by Sampaio (1988) and Severi & Cordeiro (1994), A. varzeae can be distinguished by the lower body depth compared to SL (28.7-32.9%, mean=30.1) from Astyanax sp. B (35.2-42.5%, mean=38.5), Astyanax sp. C (35.8-44.9%, mean=39.7), Astyanax sp. D (34.2-39.7%, mean=36.3), and Astyanax sp. E (32.0-46.9%, mean=37.7).

Although the species described here was first identified by Haluch & Abilhoa (2005) as Astyanax sp. F , another endemic morphotype recorded by several authors to the medium and lower portions of the rio Iguaçu basin (Severi & Cordeiro, 1994; Garavello et al., 1997; Baumgartner et al, 2006), the shorter snout length relative to head length differs A. varzeae from Astyanax sp. F (16.4-23.3%, mean=20.1 vs. 24.6-32.6%, mean=28.2). This morphotype was created by Sampaio (1988) to describe what seems to be a heterogeneous Astyanax , which exhibit body depth of 28.5 to 39.6% in relation to SL, 2 to 5 maxillary teeth, 1 to 4 teeth in outer row premaxilla, 17 to 22 branched anal-fin rays, and 7 to 9 pelvic-fin rays. Given that Sampaio (1988) did not analyze specimens from rio da Várzea headwaters, and the analysis carried out here clearly distinguish Astyanax varzeae from Astyanax sp. F , we recognize them as distinct species.

The frequent presence of two humeral spots differs the new species from A. janeiroensis Eigenmann and Astyanax aff. scabripinnis (from coastal drainage of southeastern Brazil - MHNCI 7824, MHNCI 7844, MHNCI 7858) (humeral spot well-defined, black and round, with a vertically elongated line extended anteroventrally), and from A microschemos and A. pelecus . Furthermore, these last two species inhabit drainages very distant from the rio Iguaçu basin. Astyanax varzeae also differs from the holotype of A. janeiroensis (MCZ 21057) by the lower body depth (28.7-32.9, mean=30.1 versus 36.9% of SL).

Despite de absence of secondary sexual characteristics (bony hooks) in the specimens examined, the larger orbital diameter also separates A varzeae from A. brachypterygium , A. cremnobates and A. totae (30.9- 44.5%, mean=37.9 vs. 24.8-34.8%, mean=29.0 in A. brachypterygium , 28.1-35.4%, mean=32.1 in A. cremnobates , and 25.4-42.6%, mean=35.7 in A. totae ; Fig. 5). Although the orbital diameter in A. varzeae partially overlaps the range of A. brachypterygium , A. cremnobates and A. totae , significant differences using one-way analysis of variance test were found among the species (F=67.91, P=0.001; Fig. 6).

Astyanax varzeae differs from Eigenmann’s Astyanax scabripinnis subspecies, all recognized as valid species by Bertaco & Malabarba (2001), Lucena & Lucena (2002) and Lima et al. (2003). Astyanax varzeae differs from the holotype of A. scabripinnis (Jenyns) (BMNH 1917.7.14:15) by the lower number of branched anal-fin rays (15-20 versus 21) and from the syntypes of D. longirostris (Steindachner) (NMW 57633, 2) by the shorter snout length compared to head length (16.4-23.3%, mean=20.1 versus 29.3%). Although it was not possible to analyze the type material of A. paranae , we examined one lot from the upper rio Tibagi basin (type-region of A. paranae ): MHNCI 8138 (rio Tibagi, locality of Palmeira), identified by Bertaco & Lucena (2006) as morphologically similar to the A. paranae holotype. One-way analysis of variance test indicated significant differences (F=11.18, P=0.001; Fig. 7) in snout length compared to head length between A varzeae (16.4-23.3%, mean=20.1) and A. paranae (16.0-20.4%, mean=18.1). The occurrence of two humeral spots (second one faint) also distinguishes A. varzeae from A. paranae (single humeral spot). Characters distinguishing A. jenynsii , A. laticeps , A. intermedius , and A. rivularis from A. varzeae were discussed under diagnosis.