Colostethus breweri, Barrio-Amorós, César L., 2006

Barrio-Amorós, César L., 2006, A new dendrobatid frog (Anura: Dendrobatidae: Colostethus) from Aprada tepui, southern Venezuela, Zootaxa 1110, pp. 59-68 : 60-66

Colostethus breweri sp. nov.

( Fig. 1 )

Holotype — MHNLS 17044. An adult female with circunvoluted oviducts from the entry of Cueva del Fantasma, northwestern slope of Aprada tepui, 05º 27’N, 62º 27’W, 660 m above sea level, Estado Bolívar, Venezuela, collected by Charles Brewer­Carías, Alberto Tovar and Fernando Tamayo, on February 9, 2004.

Paratypes—MHNLS 17045, 17046, two adult males, and MHNLS 17047, an unsexed juvenile with the same data as the holotype.

Diagnosis —A relatively small species (male SVL up to 21 mm, only known female 23.8 mm SVL), Finger I approximately equal to Finger II; Finger III of males not swollen; venter in preservative dirty white in males, immaculate white in the only known female; in life venter of males dirty white, female yellow; toes moderately webbed; dorsolateral stripe absent; oblique lateral stripe short, white; ventrolateral stripe absent; median lingual process (MLP) present; cloacal tubercles absent; anal sheath absent; black armband absent.

Comparison with other species — Colostethus breweri shares with the following species from Venezuelan Guayana the presence of a MLP: C. parkerae , C. shrevei , C. tamacuarensis , C. tepuyensis , C. triunfo , C. wothuja ( Grant et al. 1998; Myers and Donnelly 1997; Barrio­Amorós et al. 2004), C. praderioi and C. roraima (T. Grant, pers. comm.). Species from the Venezuelan Guayana lacking the MLP are C. aff. brunneus (probably a similar species to C. brunneus : see Morales 2000), C. aff. marchesianus (see Morales 1994 and Caldwell et al. 2002), and C. undulatus ( Myers and Donnelly 2001) .

Colostethus breweri can be also distinguished from those species having scanty toe webbing ( C. beebei , C. brunneus , C. praderioi and C. roraima ) because of its moderately developed toe webbing. It is distinguishable from other species from the Guiana Shield that also present MLP and moderate toe webbing by the following character (those of C. breweri in parentheses). Colostethus ayarzaguenai has a dorsally­rounded snout (nearly truncate), Finger I shorter than II (equal); and an uniform pattern without dorsal spots (consistent pattern). Colostethus degranvillei has Finger I shorter than II (equal), oblique lateral stripe absent (present), a post tympanic white bar (absent) and ventral surfaces brown with white spots (whitish or yellow). Colostethus guanayensis has Finger I shorter than II (equal), dark ventral coloration (dirty white or yellow). Colostethus murisipanensis has a Finger I shorter than II (equal), no oblique lateral stripe (present), dark ventral colouration (dirty white or yellow). Colostethus parimae has non consistent pattern (consistent), a dorsally rounded snout (truncate), Finger I shorter than II (equal). Colostethus parkerae has a rounded snout in dorsal view (truncate), oblique lateral stripe absent (present). Colostethus sanmartini has tympanum large, 57% of eye diameter (indistinct), pale dorsolateral stripes (absent). Colostethus shrevei is larger, up to 36 mm (up to 23.8), Finger I shorter than II (equal). Colostethus tamacuarensis has anal tubercles (absent), Finger III slightly swollen in males (not swollen). Colostethus tepuyensis has a rounded snout in dorsal view (truncate), Finger I shorter than II (equal), oblique lateral stripe absent (present). Colostethus triunfo has smooth dorsal skin (granular in life), metacarpal tubercle triangular (rounded), tarsal fold short, 1/3 of the tarsus (1/2). Colostethus wothuja has a distinct tympanum (indistinct), and no particular pattern (consistent).

Description of holotype —Dorsal and ventral skin smooth in preservative; shagreen in life with small but protuberant tubercles on the posterior part of the dorsum. Dorsal skin forming usually a well­defined rounded, posteriorly projecting flap well above vent, which opens at upper level of thighs; no anal tubercles; anal sheath absent.

Head slightly longer than wide, head width (between angles of jaws) 33.6% of SVL. Snout protruding in profile, nearly truncate in dorsal and ventral views. Nares situated near the tip of the snout and directed slightly posterolaterally; nares visible from the front, barely or not visible dorsally, but well visible from below. Canthus rostralis rounded, rather indistinct; loreal region barely concave. Interorbital region slightly wider than upper eyelid. Snout slightly longer than eye diameter. Tympanum barely visible; positioned close behind eye and low, nearly touching angle of jaws.

Hand length moderate, 27.7 % of SVL. Relative lengths of appressed fingers III>IV>I=II; Finger I equal in length to Finger II. Discs of all fingers moderately expanded; Finger III disc 1.5 times width of distal end of adjacent phalanx. Base of palm with large, nearly round, median palmar tubercle; oval inner thenar tubercle on base of Finger I; one subarticular tubercle on Fingers I and II, and two on Fingers III and IV, distal tubercles small, indistinct; all tubercles low, round. No fringes on fingers.

Hind limbs of moderate length; shank 49.5 % of SVL. Relative lengths of appressed toes IV>III>V>II>I; first toe reaching when appressed half subarticular tubercle of second toe. Toe discs moderately expanded; toe IV disc 1.4 times width of distal end of adjacent phalanx. Feet moderately webbed, the web distally continuous with a narrow fringe on all toes, including external edges of toes I and V. Webbing formula I 1–2 II 1 ½–3 III 2 ½–3½ IV 3 ½–1½ V. One to three non protuberant subarticular tubercles on toes as follows: one on toes I and II, two on toes III and V, and three on toe IV (distal tubercle ill defined). Sole with two metatarsal tubercles, similarly sized, a round outer metatarsal tubercle, and an elliptical inner metatarsal tubercle. A narrow tarsal fold or keel, straight, extending almost half the length of tarsus, distally continuous with the narrow fringe on free (preaxial) edge of toe I; fringe distinctly raised proximally, not tubercle­like.

Maxillary teeth present. Tongue longer than wide, elliptical, posterior half free; MLP very small, longer than wide, positioned on the anterior third of the tongue.

Colour in life —Dorsal ground colour pale brown with diffuse markings on the back, consisting of a dark brown interorbital bar, straight, usually with an apex at the posterior margin, a moderately large chevron between shoulders, three spots at midbody, forming an inverted V; and a single small and median posterior spot near the end of the body. Canthal stripe dark brown; supratympanic stripe black, and very distinct. In the holotype the upper lip bar is immaculate white, containing the indistinct tympanum. Upper lips also white in the other specimens. Supratympanic stripe is continuous above the arm, widening at mid body, forming a dark brown to black flank, where the white oblique lateral stripe is present. Flanks darker than dorsum but without definite borders. Upper flank dark brown. Throat, chest, and belly immaculate yellow; inferior surfaces of thighs dark orange. Iris bronze.

Upper arms and forearms pale brown, the latter with dark crossbars. Thighs with illdefined dark brown crossbars.

After one year and a half in preservative, dorsal ground colour became dark brown; flanks dark brown; ventral surfaces white. The rest of the body remains without changes.

Measurements of holotype (in mm)—SVL: 23.8; SL: 11.8; ThL: 12; HeL: 8.3; HW: 8; UEW: 2; IOD: 2.2; ED: 3; TD: 1; ETS: 3.2; FD: 1; 4TD: 1; 1FiL: 3.2; 2FiL: 3.3.

Variation —All specimens have a similar dorsal colour pattern, which is more or less conspicuous. It is most distinct in MHNLS 17045 (a male), and the juvenile (MHNLS 17047) and is less defined (but still detectable) in the female holotype and male MHNLS 17046. The oblique lateral stripe occurs in all specimens, although varying in shape (always short), and never well defined. It is best defined in MHNLS 17045, or fragmented into three small white spots, as in the holotype. Males differ from the single female in at least three external characters (these may be due to sexual dimorphism rather than individual variation): First, the belly of adult males is immaculate white, whereas that of the adult female is immaculate yellow; second, the presence of short vocal slits, extending from near tongue insertion to nearly the end of tongue; and third, the presence of a larger lingual process, longer than wide; this last character was also observed in Colostethus wothuja ( Barrio­Amorós et al. 2004) .

In preservative, holotype and MHNLS 17046 dorsal ground colour became dark brown with diffuse markings; MHNLS 17045 and juvenile (MHNLS 17047) pale gray with conspicuous dark gray markings; flanks dark brown to dark gray. Black arm band (sensu Grant and Castro 1998) absent, but in all specimens except the juvenile, a black longitudinal stripe on the anterior surface of the arms and forearms is present.

Males (MHNLS 17045 and 17046) with dirty white throat (white with profusion of melanophores) and larger white spots especially at edges of throat; anterior part of chest also with melanophores in MHNLS 17045 but not in MHNLS 17046; posterior part of chest, belly and ventral surfaces of the thighs immaculate white. Testes white.

Distribution —The species is known only from one locality at Aprada tepui, a flattopped sandstone mesa (tepui) of Guayana highlands in Venezuela, at 660 masl ( Fig. 2 ). This locality lies in medium to tall, evergreen, basimontane and lower montane forests ( Huber & Alarcón 1988). To found the type series inside a “cave” must not be significant, as this “Cueva de El Fantasma” ( Fig. 2 ) is a huge, collapsed steep gorge and is not strictly a cave. The maximum altitude of Aprada is at 2500 m, with a tepui surface of 4.37 km 2 ( Huber 1995), although its slopes embraces at 1000 m of altitude a wider area including the Chimantá massif and even the whole Gran Sabana area. The species could be an endemic, or more widely expanded through uplands of the Gran Sabana.

Aprada tepui was first explored by an expedition leaded by C. Brewer­Carías on February 25, 1978, accompanied by Roy McDiarmid, as herpetologist, Luis José Joly as enthomologist, Julian Steyermark as botanist and G.C.K. Dunsterville and his wife Nora as orchidologists ( Dunsterville 1979). This paper becomes the first written reference about this huge "cave" at N 05º 27.480’W 62º 27.588’ on the northwestern slope of the Aprada tepui.

Natural history —This species is a fast­moving frog that lives along creeks and in quiet pools along small streams on the slopes of Aprada­tepui. An undescribed species of Eleutherodactylus and the lizard Neusticurus rudis (Gymnophtalmidae) were found syntopically. Tadpoles and call of the new species are unknown.

Etymology —I dedicate this frog to the indefatigable explorer of the Venezuelan Guayana, Charles Brewer­Carías, in recognition for his endless support of biological research in the Guiana highlands.

Discussion —Currently, knowledge of Guayanan Colostethus is increasing, and several characters have been observed recurrently for most of the species. The most apparent is the presence of a MLP (Grant et al. 1997); although it is not exclusive of Guayanan Colostethus , it seems most of the species share it. Furthermore, all species known to have a MLP ( Colostethus breweri C. parkerae , C. praderioi , C. roraima , C. shrevei , C. tamacuarensis , C. tepuyensis , C. triunfo , C. wothuja and C. sp. from Sarisariñama (Barrio­Amorós and Brewer­Carías, in press), also share other characters, such as a short oblique lateral stripe, absence of dorsolateral and ventrolateral stripes, moderate to scant webbing, and in at least two species, C. triunfo and C. wothuja , absence of palatine bones. I assume that most of the species of the Venezuelan Guayana (and probably the rest of the Guiana Shield) that do not belong to the C. trilineatus group (sensu Morales 2002), will have a MLP (which can be larger in males) and absence of palatine bones, which suggest they might form a monophyletic group.

The remaining species are not known to have the MLP, but it is probable to occur in C. ayarzaguenai , C. guanayensis , C. murisipanensis , C. parimae , and C. sanmartini (unpublished data).

The geography of the MLP is of interest. Grant et al. (1997) recognized eight species (only four named at that time) of Colostethus having a MLP. I mentioned above Venezuelan species currently known to have a MLP, which are 11 (plus two more without proper name). Of all species with the MLP, only three ( C. atopoglossus , C. lacrimosus and C. aff. chocoensis from Ecuador) are known from the Chocoan bioregion (Pacific versant of Panama, Colombia and Ecuador), and the rest (13) occur at the Guiana Shield. The distance between the closest Guayanan and Chocoan species, respectively C. shrevei from Venezuelan Amazonas state and C. lacrimosus from Valle del Cauca in Colombia, is of about 1200 km, separated by Los Llanos, the Cordillera Oriental and Central de Colombia. No relationship can be inferred between the Chocoan and Guayanan groups of Colostethus having MLP. Grant et al. (1997) considered the MLP to be a dendrobatid plesiomorphy. I consider here the presence of the MLP in both biogeographically separated groups, as a convergence.

Due to its restricted distribution to a single tepui of 4.37 km 2, I consider the species to be potentially endangered and sensitive to natural or artificial catastrophes, as was predicted by Barrio­Amorós (2001) for other tepui endemics. I should locate the species in the category VU D2 of the IUCN, following Young et al (2004).

Figures

FIGURE 1. Female holotype of Colostethus breweri in life. Photo by Charles Brewer­Carías.

FIGURE 2. View of the type locality (Cueva de El Fantasma, Aprada tepui) from inside. Note the size of the two helicopters at the entrance. This is the first geographic report and photograph of such immense “ cave ”. Photo by Charles Brewer­Carías.