Tepuihyla edelcae (Ayarzagüena, Señaris, and Gorzula)

Myers, C. W. & Donnelly, M. A., 2008, The Summit Herpetofauna Of Auyantepui, Venezuela: Report From The Robert G. Goelet American Museum-Terramar Expedition, Bulletin of the American Museum of Natural History 2008 (308), pp. 1-147 : 60-70

Tepuihyla edelcae (Ayarzagüena, Señaris, and Gorzula) Figures 37–40

Ololygon sp. : Gorzula, 1992: 269, color photos 152, 154 (observations on distribution and natural history).

Osteocephalus sp. : Duellman and Hoogmoed, 1992: 15 (undescribed species from Auyantepui).

Osteocephalus edelcae Ayarzagüena, Señaris, and Gorzula, 1992 a: 122 , fig. 4 (hand and foot). Holotype MHNLS 10626 (8) from Auyantepuy , 10.8 km al este del Salto Angel, Estado Bolívar, Venezuela (5 ° 589N, 62 ° 299W), 1970 m, collected May 25, 1984 by S. Gorzula.

Tepuihyla edelcae: Ayarzagüena, Señaris, and Gorzula, 1992 b: 215 .

MATERIAL: Camp 1, 1700 m: AMNH A-164865–164878, 164879–164880 (larvae), EBRG 2596–2609, 2628 (larvae). Camp 2, 1750 m: AMNH A-164881–164899, 164900– 164909 (larvae), EBRG 2610–2619, 2629– 2631 (larvae). Camp 3, 1850 m: AMNH A- 164910–164917, 164918–164923 (larvae), EBRG 2620–2627, 2632–2634 (larvae). Camp 4, 1600 m: AMNH A-164924–164926, EBRG 2637–2639. Camp 5, 2100 m: AMNH A-164927–164931, 164932–164935 (larvae), EBRG 2640–2644, 2635–2636 (larvae). All from the 1994 AMNH –TERRAMAR Expedition to Auyantepui.

ADDITIONAL MATERIAL: Auyantepui Plateau, 2200 m (7218 ft): AMNH A-46033– 46035, 146339 (larvae), from the 1937–1938 AMNH –Phelps Venezuelan Expedition.

Females are larger than males (x¯ female SVL / x¯ male SVL 5 1.24) and there is little overlap in size between sexually mature males and females. The following measurements were obtained by combining all adult males and all adult females from the 1994 Expedition:

Adult 8 (51), x¯ 5 36.53, SD 2.084, 32.0– 41.5 mm SVL

Adult ♀ (23), x¯ 5 45.43, SD 2.858, 39.3–50.5 mm SVL

There is no obvious variation in size across sites (data not shown); analysis of variance for males from camps 1–5 shows no statistically significant difference between the mean SVL at the 95 % confidence level (F 5 2.30, P 5 0.0731).

All adult males have a well-developed nuptial pad on the base of the thumb; the pad is a pale brown color owing to minute centers of cornification (visible with a dissecting microscope). The vocal slits of males are relatively small apertures situated near the corner of the mouth; it is difficult to find these openings on some specimens, and a few individuals of adult size with nuptial pads seem to lack them. There is a degree of sexual dimorphism in skin texture, with males well covered with small spicules and with females tending to have smoother or slightly pustulate skin, as seen in figure 38 (males above, female below); some adult females, however, do have sparse spicules (especially anteriorly) additional to smoother pustules. Adult males have enormous white testes that fill most of the body cavity; females in the sample have convoluted oviducts and numerous small pigmented ova.

COLORATION: Based on a large sample of 95 specimens, Tepuihyla edelcae was diagnosed as having a uniformly dark dorsal color (Ayarzagüena et al., 1992a: 122), with some limited variation being mentioned in the description (p. 126). However, in life this is a colorfully variable frog, at least on Auyantepui (figs. 37–38), with a basic dorsal color of brown, orange-brown, red-brown, or gray ; a few specimens had the dorsolateral areas slightly lighter than the darker middorsum and one individual ( AMNH A- 164916) has distinct pale dorsolateral stripes that are retained in preservative. Some frogs had orange or yellowish spots or blotches over the head and dorsum and also on the throat. Limbs colored like the dorsum, rarely with indication of dark crossbanding; rear of thigh orange-brown to gray, sometimes with small areas of silvery gray frosting. Upper lip white or variably suffused with dark pigment. Throat and venter dirty white, turning gray under the limbs. Bones white. The iris varies from light bronze to dark brown or gray, with variable black venation and a tendency for a dark horizontal stripe through the pupil.

In preservative the frogs are darker, but any light markings are well retained in our material. Two specimens were noted as having either faint or distinct dark crossbanding on the limbs ( AMNH A-164869, 164916, respectively); the rear of the thigh is light to dark orange. Some specimens retain a distinct pale labial line in preservative; Ayarzagüena et al. (1992a: 126) seemed to suggest that this is a juvenile character retained in some adults. We see no ontogenetic linkage in the Auyantepui material,

inasmuch as the upper lip is dark in some small specimens such as AMNH A-164872 (a metamorph 16 mm SVL), EBRG 2639 (20 mm), and EBRG 2641 (25 mm).

TADPOLES

The tadpole of Tepuihyla edelcae has not been described, although there was reference to its habitat and behavior in the original description (Ayarzagüena et al., 1992a). Ayarzagüena et al. (1992b: 216) cited a formula of 2/3 for the genus, without specifying which species were examined. Tepuihyla edelcae differs from that generic diagnosis in having 2/4 larval tooth rows as the most common condition in a variable formula including also 2/2 and 2/3.

We netted tadpoles from shallow sandstone pools at four of our five camps. At camp 1, 11 tadpoles (stages 25–32) were taken from a surface pool near a clump of Brocchinea. Three collections were made at Camp 2, with one large sample containing only stages 24–25 (N 5 64), and with two small samples containing a greater range of stages (27–42, N 5 9, and 25–39, N 5 12). At camp 3, two collections included only early stage-25 tadpoles (N 5 36, 55), whereas a third included stages 25–40 (N 5 75). Two small collections of 13 larvae each from at Camp 5 consisted of stage-25 tadpoles and a single stage-36 tadpole.

We did not find tadpoles of other species in pools with Tepuihyla , although some other pools contained tadpoles of two species ( Hypsiboas jimenezi and Anomaloglossus tepuyensis , or H. jimenezi and Leptodactylus rugosus ).

The following description is based on one stage- 33 larva ( AMNH A-164902) and one in stage 37 ( AMNH A-164907 [fig. 39]), followed by notes on ontogenetic change. See table 4 for additional measurements.

HABITUS AND PROPORTIONS: Head and body elliptical and rounded at both ends in dorsal view; maximum body width 53.8 % – 54.6 % of head-body length. Body higher than wide (maximum body depth about 117.6 % – 118.2 % of maximum body width); body flattened dorsally and somewhat wedgeshaped in profile, increasing in depth from anterior to posterior. Eyes dorsal, directed laterally, not visible from below. Nares dorsal, directed anterolaterally, approximately equidistant between tip of snout and eye. Spiracle sinistral, low on body, a short tube (1.4– 2.2 mm long) directed posteriorly with medial wall attached to body; spiracle opening at 58.2 % –59.5 % of head-body length. Vent tube attached to ventral fin, opening dextral to fin.

Lateral line system somewhat indistinct. Supraorbital and infraorbital branches forming a loop around eye; superior and middle trunk branches originate behind eye; superior branch continues posteriorly in dorsal fin above musculature; middle trunk branch continues posteriorly in the middle of tail musculature; inferior body branch originates on body near spiracle and curves ventrally to the ventral body.

Tail 64.7 % –66.5 % of total length; maximum tail height approximately equal to body depth; dorsal fin not extending onto body, higher than ventral fin; musculature height greater than fin height; margins of dorsal fin slightly scalloped; tail tip rounded.

TABLE 4 Measurements (in mm) of Tepuihyla edelcae Tadpoles (Values are ranges followed below by means ± 1 SD when N. 2)

PIGMENTATION: In preservative, headbody brown dorsally and laterally; ventral skin clear, with underlying tissue having an iridescent blue sheen; tail musculature tan, flecked with melanophores, becoming vaguely blotched posteriorly. Dorsal and ventral fins brown (heavy melanophore deposition), paler at tail tip and adjacent to body on lower fin.

MOUTH PARTS: Mouth ventral. Oral disc not emarginate. Labial teeth in 2/3 (AMNH A-164907) or 2/4 rows (AMNH A-164902); A1 and A2 rows subequal in length, A2 row with distinct median gap; P1 slightly shorter than P2; P3 either shorter than P1 or equal to P2; P1 with or without a short gap in median curvature; P4 row (in AMNH A-164902) small and barely keratinized. Upper jaw sheath a broad arch bearing pointed serrations and with slender lateral processes, its surface with a minutely pebbled texture that is highly reflective under certain light conditions. Lower jaw sheath V-shaped with pointed serrations. Marginal papillae with a broad anteromedial gap on upper labium; one row of marginal papillae on upper labium and along lateral margins of disc; two rows of slightly blunter papillae on lower labium; several pointed to blunt submarginal papillae laterally.

ONTOGENETIC CHANGES: In three available tadpoles of Tepuihyla edelcae in stage 24/ 25, maximal tail height exceeds body depth, and body depth and width are approximately equal. The body depth exceeds tail height and body width in all other stages. Changes in body size are summarized in table 4.

The three larvae in stage 24/25 have paired patches of raised, pigmented skin posterior and lateral to the oral disc. These raised patches may be adhesive glands; they are reduced to pigmented spots in early stage- 25 larvae, with the spots disappearing by late stage 25. 16

The stage 24/ 25 larvae have 2/2 tooth rows that are barely keratinized; jaw sheaths are keratinized in this stage and pointed serra-

16 According to Altig and McDiarmid (1999: 34), adhesive glands (when present) are transient structures usually visible between stages 18–25 and sometimes persisting as pigmented spots for several stages later. They serve ‘‘to provide stabilization prior to the further development of the oral disc and tail which afford more coordinated adhesion and locomotion for active tadpoles.’’

tions are present; marginal papillae are also present in these tadpoles but submarginal papillae are lacking. Early stage-25 tadpoles have 2/3 labial tooth rows; the P3 row is the smallest and lightly keratinized; these tadpoles also lack submarginal papillae. The posteriormost tooth row (P4) is the last to appear and it is often not well developed. The submarginal papillae also appear relatively late in development.

Variation in the tooth-row formula is as follows: The three available tadpoles in stage 24/25 have 2/2 tooth rows; some tadpoles in stages 25 (N 5 6), 26 (1), 27 (2), 33 (1), and 37 (1) have 2/3 tooth rows; all others (stages 25–40) have 2/4 tooth rows.

Some unusual configurations of mouthparts are not associated with ontogeny and may be environmentally caused (see Natural History). In one stage-33 tadpole (AMNH A- 164902), the P4 row is short, is situated to the right of the midline, and is positioned in the marginal row of papillae. In some cases, the tooth-row ridges are comprised of short and long sections (i.e., A 1 in three sections—two short lateral and one long medial section; A1, A2, or P 2 in two sections—one short and one long). In some tadpoles, the tooth-row ridges are unusually short (P1 and P2 missing on the left, A1 and A2 short on right, P3 short on right). In many tadpoles, the P4 row is small, difficult to discern, and mixed with the marginal papillae. In one late stage- 25 larva, there are large pointed papillae between tooth rows), and in one there is a single papilla in the median gap on the upper labium.

DISTRIBUTION, NATURAL HISTORY AND VOCALIZATION

Tepuihyla edelcae is the most common frog on Auyantepui and occurred at all five of our camps (although only tadpoles were found at Camp 5). It was the only amphibian taken from the summit by the 1937–1938 AMNH– Phelps Venezuelan Expedition (it was not found at the lower-elevation camps). The species was described from an abundance of specimens collected mostly by S. Gorzula on Auyantepui (the type locality) and various parts of the Chimantá massif (Ayarzagüena et al., 1992a). Señaris and MacCulloch (2005: 21) stated that Tepuihyla edelcae is ‘‘Endemic to summits of Auyán-tepui and Macizo de Chimantá … 1630–2600 m’’, but Gorzula and Señaris (1999: 49) had additionally reported a specimen from ‘‘Altiplanicie S Terekyurén-tepui’’ in the Los Testigos range to the east of Auyantepui. It occurs on Cerro La Luna (Gorzula and Señaris, 1999: 49), an erosional remnant standing north of Auyantepui (map 1), and might also be expected on the poorly known Uaipantepui, a satellite mountain south of Auyantepui ( Huber, 1995a: fig. 1–25).

Tepuihyla edelcae is best known through the work of Stefan Gorzula, who studied it on the Chimanta massif and on Auyantepui. Gorzula noted its affinity for tube bromeliads, of which the species Brocchinia hechtioides and B. reducta occur on Auyantepui (Steyermark, 1967; Givnish et al., 1997). On Chimantá, Gorzula recounted that this frog

is a very easy species to locate, since the

majority utilize Brocchinia hechtioides as hiding

places during the day. It is enough to walk

carefully through the turberas where this

bromeliad abounds, in order to observe them

in the upper part of the tube of the Brocchinia .

They are frightened by a vibration or a slight

shadow, hiding themselves in the water reserve

in the base of the tube. Even if the Brocchinia

provides an appropriate habitat for the refuge

of adult specimens of this frog, the small reserve

of water in the base of the tube has character-

istics that are not favorable for spawning. This

frog spawns in temporary puddles in the

turberas or in shallow pools between rocks

adjacent to colonies of Brocchinia . (translated

from Gorzula [1992: 270], under Ololygon sp. )

Aside from the possibly unfavorable characteristics of the bromeliad water mentioned by Gorzula, the relatively large number of small ova contained by gravid female Tepuihyla , and the enormous size of the testes, suffice to show that it is not a bromeliad breeder—the small tube bromeliads could hold only a small fraction of the clutch. On Auyantepui, Gorzula found 12 T. edelcae in a transect of some 100 m × 10 m, representing one frog for every 30 Brocchinia ; both Brocchinia hechtioides and (in our experience) B. reducta are utilized, or sometimes the tank bromeliad B. acuminata , usually with only a single frog in a bromeliad (Ayarzagüena et al., 1992a: 126–127). The claim (loc. cit.) that this species uses its head as a plug to prevent evaporative water loss needs confirmation, inasmuch as the skin is not co-ossified with the skull as in the case of other hylid genera exhibiting the behavior (Trueb, 1970).

We collected Tepuihyla at night at Camp 1, on low vegetation (, 0.5 m aboveground). The majority of frogs were on, or peering from within, tube bromeliads growing in patches on flat sandstone outcrops (fig. 3, bottom). One froglet peering out of a Brocchinia tube was a recent metamorph.

We had the impression at Camp 1 that Tepuihyla edelcae was indeed associated primarily with tube bromeliads, but our experience with them at Camp 3 suggested much greater behavioral plasticity. Although we avoided destroying bromeliads because the frogs are easily spotted in the beam of a headlight at night, Tepuihyla seemed less common in the tube bromeliads at Camp 3 than at Camp 1. Instead we found them in a greater diversity of situations, such as in the middle of the stream bed at night, on various kinds of low vegetation at night (usually, 1 m aboveground), and by day under rocks on open sandstone flats, sometimes in company with other concealed frogs ( Leptodactylus rugosus ) or lizards ( Tropidurus bogerti ).

Two groups of unusually large stage-25 tadpoles collected from drying pools near Camp 5 have a high incidence of damaged fins, suggesting that high tadpole density in a decreasing volume of water may have had led to tail damage from abrasion and/or cannibalistic events.

Some abnormalities of the larval tooth rows (see above) may have an environmental cause. Bresler (1954) experimentally found that tooth-row abnormalities in Bufo and Rana developed more frequently at high temperature. Temperature in small sandstone pools at high elevation fluctuates widely, as shown by Gorzula (1992: 270), who recorded a 15-hour temperature range of 12.2–27.0 ° C in a pool containing larvae of Tepuihyla edelcae (reported as Ololygon sp. ), at an elevation of 2100 m on the neighboring Chimantá. Gorzula noted that the tadpoles of this species are inactive on the bottom during the colder hours of night.

At Camp 1, Tepuihyla edelcae was calling back and forth from bromeliads around a small pond at night. Two individuals record- ed seemed to be responding to one another in a duet (fig. 40). One male (AMNH A- 164873) gave single-note calls, his neighbor double-note calls. The note sounds like a soft chickenlike ‘‘cluck’’. Each well-spaced call note is 0.02 sec in duration and nearly nonpulsatile, with but slight indication of two poorly separated pulses. Narrowband analysis shows a fundamental frequency of about 1400 Hz, with poorly resolved harmonics at roughly 1400 and 2300 Hz.

REMARKS

The gender of Tepuihyla was stated to be masculine by its authors (Ayarzagüena et al., 1992b: 215), presumably because the Pemón Indian word tepui (or tepuy) is treated as masculine in Spanish. But gender must be based on the last component of a compound generic name ending in a Latin word ( ICZN, 1985: art. 30a; 1999: art. 30.1.1). Although Hyla was derived from the name of the mythological boy Hylas (Laurenti, 1768: 32–33), the generic name itself is universally and correctly treated as feminine, as recently reaffirmed by Myers and Stothers (2006).

Required correction of the gender of Tepuihyla from masculine to feminine has no effect on the endings of the six specific names transferred to that genus by Ayarzagüena et al. (1992a: 215). Most are singular nouns in the genitive case (i.e., four are dedicatory names or ‘‘patronymics’’, including two, aecii and edelcae , derived from acronyms). A fifth name, rimarum (‘‘of or belonging to cracks or crevices’’), is a noun in the genitive plural.

As originally explained by Ayarzagüena et al. (1992b: 131), the sixth specific name seemed as if it should be a noun in apposition: ‘‘ luteolabris, derivado del latín luteo (amarillo) y labris (labio)’’. However, labris is not the Latin substantive equivalent of Spanish labio (5 lip, a noun). The name luteolabris is an adjective, derived from the stem of luteus + connecting o + labris (‘‘lipped’’). The last component is an adjective derived from the Latin noun labrum and suffix - is, thus labris (masculine and feminine) and labre (neuter)—the declension like that of collis (‘‘necked’’), from collum ( Brown, 1956: 45). Compared with alterna- tive adjectives and participles such as labralis (from labrum), labialis, and labiatus (the last two from labium), the shorter labris is a better choice for use in compound names, and it is employed mainly (if not entirely) as a suffix in taxonomic compounds.

Words formed with adjectival suffixes like – is (with a short vowel) must not be confused with the dative and ablative plural cases (normally with long i) of the original nouns, because these cannot qualify as specific names. Although the ablative plural noun labris (‘‘with lips’’) might have been preempt- ed for use as an adjective, the adjective properly derived from the stem of labrum is formed with - is —the adjective-forming suffix meaning ‘‘with or having the nature of’’ ( Brown, 1956: 549).17

The Venezuelan frog Hyla rodriguezi Rivero (1968b) was transferred to the genus Osteocephalus by Duellman and Hoogmoed (1992: 16). Ayarzaqüena et al. (1992a) described five new species of Osteocephalus and established a rodriguezi species group. In the following issue of the same journal, Ayarzaqüena et al. (1992b) elevated the group to generic status under the name Tepuihyla . Unfortunately, no summary was given of species possessing the purported diagnostic characters, including various cranial features, 2/3 larval tooth rows, and sperm with two tail filaments.

Ayarzagüena et al. (1992b) said that, according to Fouquette and Delahoussaye

17 The inclusion of etymological information for new names in zoology is customary nowadays and is recommended by the Code (International Commission, 1999: art. 25B, app. B5). But it has escaped the attention of many authors, reviewers, and editors (nor is it mentioned in the Code) that an important reason for so doing is to enable future workers to determine easily if a new generic combination will require changing the ending of a specific name for mandatory agreement in gender. The difference between adjectival specific names (adjectives and participles in the nominative singular), which must agree in gender with the generic name, and unchangeable nouns in the genitive case or in the nominative singular (standing in apposition to the generic name), is not always selfevident and sometimes impossible to determine by inspection. It would be helpful if authors were to explicitly give the part of speech of new names; if the author cannot do so and cannot find the name or its components in standard Latin or Greek dictionaries, a little research or a substitution might be in order prior to publication. It needs remembering, however, that Latin is a dead language. Many words, including adjectives formed by regular transformations of Latin nouns, may not be shown in Latin dictionaries unless they are established in the classical literature.

(1977), Ololygon [now Scinax ] and Sphaenorhynchus were the only other hylid genera with double-tailed sperm (although Fouquette and Delahoussaye also mentioned a species of Hemiphractus ) and therefore paid particular attention to comparing Tepuihyla with Scinax . Duellman and Yoshpa (1996) commented that most morphological features mentioned by Ayarzagüena et al. ‘‘are homoplastic in their appearance in diverse genera of hylids’’. Duellman and Yoshpa said that if ‘‘all species of Tepuihyla have a double-filamented sperm, this may be the only synapomorphy [if independently derived] for the genus’’, or else a synapomorphy uniting Tepuihyla , Scinax , and Sphaenorhynchus . ( Tepuihyla most recently has been allied with Osteocephalus and Osteopilus [ Faivovich et al., 2005: fig. 12]).

However, there are two problems involving the sperm character in the above context. Although Fouquette and Delahoussaye (1977) did in effect use the character as a derived feature to define Ololygon [now Scinax ], they also presented valid reasons for considering two or more sperm tails (as they understood the character) as the primitive condition among frogs—a plesiomorphic not a synapomorphic character. But this is now unimportant, because it seems that the two-tailed sperm character was a misinterpretation of morphology that is not well studied by standard light microscopy. After examination of ultrastructural data from transmission electron microscopy of the spermatozoan tail of Scinax ranki (as Hyla ranki ), Taboga and Dolder (1993: 85) summarized:

The flagellum contained a typical vertebrate type axoneme and an electron dense, F-PTA positive axial rod, in parallel association and interconnected by an undulating membrane. The parallel arrangement had previously and erroneously been interpreted as a separate structure, or double filaments, when examined by light microscopy. These results suggest that this structure is not a dependable taxonomic character.

Kuramoto (1998: 113) independently noted the presence of an axonemal fiber and axial rod in hylid sperm tails and, in reference to Fouquette and Delahoussaye’s (1977) taxon- omy, suggested that ‘‘If a single tail fiber involves two flagellar components … it seems not essential to separate ‘two-flagellated’ from ‘single-flagellated’ species, until further studies reveal that each of the two ‘flagellas’ contains an axoneme’’. Sperm morphology appears to have potential for taxonomic insight, but comparative data from ultrastructural studies are needed for it to be useful. The recognition of Scinax , however, is not considered to be dependent on the dubious sperm character ( Faivovich, 2002; Faivovich et al., 2005).

Although the monophyly of Tepuihyla remains to be rigorously tested ( Faivovich et al., 2005: 110), Ayarzagüena et al. (1992a, 1992b) have usefully called attention to this interesting group of high-elevation frogs. Tepuihyla , like Stefania below, is a Guayana endemic comprised primarily of tepui species, but with representatives in nontepui habitats.

FAMILY HEMIPHRACTIDAE (FAMILY CRYPTOBATRACHIDAE AUCTORUM )

One genus, Stefania , occurs on the summit. It is one of five genera of Neotropical ‘‘marsupial frogs’’ that formerly were grouped in the hylid subfamily Hemiphractinae . Consistent with growing evidence from other authors, the hemiphractines were formally removed from Hylidae by Faivovich et al. (2005: 49–50), who questioned the group’s monophyly while tentatively placing it as a subfamily of paraphyletic Leptodactylidae . Subsequently, Frost et al. (2006) separated the group into three families— Hemiphractidae (Hemiphractus) , Cryptobatrachidae ( Cryptobatrachus , Stefania ), and Amphignathodontidae ( Flectonotus , Gastrotheca ). However, Wiens et al. (2005, 2007) argue for recognition of a monophyletic family Hemiphractidae . Differences between these alternative classifications remain to be resolved.

Stefania schuberti Señaris, Ayarzagüena, and