HERPETOFAUNA OF THE YUTAJE-COROCORO MASSIF, VENEZUELA: SECOND REPORT FROM THE ROBERT G. GOELET AMERICAN MUSEUM-TERRAMAR EXPEDITION TO THE NORTHWESTERN TEPUIS
CHARLES W. MYERS
Curator Emeritus Division of Vertebrate Zoology (Herpetology) American Museum of Natural History New York, New York
MAUREEN A. DONNELLY
Research Associate Division of Vertebrate Zoology (Herpetology) American Museum of Natural History Associate Professor, Biological Sciences Florida International University
BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY Number 261, 85 pp., 50 figures, 2 tables Issued May 9, 2001 Price: $15.00 a copy
Copyright © American Museum of Natural History 2001 ISSN 0003-0090
CONTENTS
PED SUCAC ge tse ge wes red ee Nice 00,23 3 actin Oe ea Dae Wore iued cd oo ag Myles. Bahn Phu, to See ans 3 FRESUMQCIL Bln cas hare te scope tae chan Poms ORAM Ngtdens SE AACE EEE Setos ache PIE Tey eA poke ne ahah sorts NIE acess oe 3 TNCOGUCT ONT: yous ys a FOF ie es ae ARP ne Yh Doe OY en 8 Ee vob rd Sea yet 2 + MARC TI ARS “Alt VEC LOC Sm tap cl I rR, cog Pinger eg tac ce or ok eat ery, saat ures te ens eieewar eet lek la on 6 he Naturalelandseape Af 3 nc ps sree ney be arte Aires REE Ba see eh EEG SD eee teed Bb lh 8 COME CEIT ET SINC: eae ors cas VV LIMEILID BS axle wpeiee pastime ne aie eet NE CIM Py be hacaweniesicsesosate chs ele od 12 CENTOS MILL AG Arr ies are sibe aM lat aAie a! Bae AL © te PR eek, sierra tab thats aie le C4 dis Ha seeder a 16 Hyalinobatrachium, €ccentricum, DEW SPECISS 4 a ede oe Gee ble ld ones deere pate oe baad» > 16 PDENGIrODAUT AS PEF eel Kelly ee Ae AOE Go oie lee See cbealidae Tate a led AE Ge v.5 eseaheteed ee BE 23 Colostenins HROUuIGIISANew SPECIES. i AEE Ahn. ee ee ee 23 ES ptO Gar isn CAS MUNI ere ye cc) ace abn e Baste ase tee oti gent 11,28 asco s ee Beauties LNs see Ne fe 36 Eleutherodactylus -cantitans. Myers--and’ Donnelly 6. ines et ele oe pba el 36 Eléutherodactylus yaviensis Myers ‘and. Donnelly wg ciccoce 5. 3 cee Oo ee certenenene oe PSS a9 PF SCUCODGILALC OVETLANEL EV CIL Vht 0 Osi pett emer a Tad Chr wey Sh eso neig yar wae doe pinervrimeneg tai BU tie, 43 WCRI EC BAS tae erate Eee neg ee ace ok ESP Dp rere oe EA Lee gets SA aE RUN aha dea idee bm TG & 48 AGETCOSGLTUS, "NEW CONUS). 5:5 ¢.9%3 Ae sla 8.5 NR homens REREAD 0 Pad ok gh Bete a oy lt 49 AdERCOSAHTUS VIFGANEXHS: NEW SPECIES ©. Ooo ee ee es oe eags Fae 4 Vala 2B? Prionodactylus goeleti (Myers and Donnelly), new combination ................... 60 PERATCTO AC Sh tendo 9 Lida tate Peed eae aang Gr6 Oe rete oR he oe ok Dah d le Me hat SAE Gclete gue meedd Ake 66 LFOTHARTUS: PONSHCIUS. NEW SPECIES =320.64 etry tah devi BY BS REISS IS OT Pek eal Pa eS 66 Key-to Venezuelan Tropidund Lizards: nia) seine oop sorb eel do mews pre eis ees 75 COMIDIG AGS bd <4 en once ten nate tM oe nek Sin alin 7, Banc, Bt et Sener eh Ny eS Laka ov, Roe ee EXOPHISISPCCIES? wreseh PE PE i ee ccc aecn tee Co ticln, Tle. eee add uel 2 oie. 's 75 Thamnodynastes corocoroensis Gorzula and Ayarzagiiena ..................22005- 76 DISCUSSION. 602.5 sc get locas Bre nln eh Wy ctl Come A oa lh, a ieist Mintel ath Bl gh Dake A toh SRL wet 79 ACkriowledeiments.).5 2222.08 6-2 2 3 3'5 2-2 Poe hE ees to Pepa gs hha ehh eRe oo, oh 81 FRSECROIICE SI eet ess rpeae Sele ee UTR fo Oe oe ea NT OO hp NS RE a ee SN ac pt PRN 82
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF
ABSTRACT
The Yutajé—Corocoro massif is a highly eroded sandstone table mountain, with internal drainage mainly to the central valley of the Rio Corocoro, a stream in the headwater drainage of the Rio Manapiare—some 100 km east of the middle Rio Orinoco, at the northern edge of the State of Amazonas in southern Venezuela. The rocky soil supports a mosaic of diverse scrubland and forest, with small tepui mead- ows at the higher elevations. The herpetofauna is depauperate, as is typical of the Venezuelan tepuis.
Eight species of amphibians and reptiles were collected during a 7-day period in the dry season (February). This sample includes two new frogs (Hyalinobatrachium eccentricum, n. sp., Centrolenidae; Colostethus undulatus, n. sp., Dendrobatidae) and anew genus and species of lizards (Adercosaurus vixadnexus, n. gen. & sp., Teiidae), all of which were found in humid montane mossy forest at 1700-1750 m elevation. Another new lizard (Tropidurus panstictus, n. sp., Tropiduridae) was discovered at lower elevations (180-1220 m), especially in dry scrub.
The fauna also includes a widespread lowland frog (Pseudopaludicola llanera Lynch), two tepui frogs (Eleutherodactylus cantitans Myers and Donnelly; E. ya- viensis Myers and Donnelly), a tepui lizard (Prionodactylus goeleti [Myers and Don- nelly], new combination), a snake (Liophis?) that escaped capture, and another snake (Thamnodynastes corocoroensis Gorzula and Ayarzagiiena) obtained by S. Gorzula in 1987. The two Eleutherodactylus and the Prionodactylus also occur on neighbor- ing Cerro Yavi (the type locality), although one of the frogs (EZ. yaviensis) and the lizard show evidence of differentiation. Based on the original description, the snake Thamnodynastes corocoroensis appears to be distinct from a related species on Cerro Yavi.
Two of the new species exhibit characters that are novel or not previously noted. The dendrobatid frog Colostethus undulatus, n. sp. has a glandular supracarpal pad atop the wrist, being best developed in males. This species, which also has the par- asphenoid bone curiously concealed, seems to be unusual among tepui Colostethus in lacking the recently described median lingual process. The centrolenid frog Hy- alinobatrachium eccentricum, n. sp. has a peculiar bicolored iris, with a dark median sector that conceals the pupil and which apparently dilates with the pupil. This char- acter is retained in preservative and differentiates H. eccentricum from H. crurifas- clatum Donnelly and Myers. Both species share a previously overlooked bubblelike structure in the web between the third and fourth fingers, herein termed bulla (pos- sibly parasite induced?).
RESUMEN
El macizo Yutajé—Corocoro es una meseta de piedra arenisca muy erosionada, con un dre- naje interno central dirigido hacia el valle del Rio Corocoro, un tributario del Rio Manapiare en sus cabeceras, situado a unos 100 km al este del Rio Orinoco, en el extremo norte del Estado Amazonas, Venezuela. El suelo rocoso sustenta un mosaico de arbustales y bosques diversos, con pequenos herbazales tepuyanos en las zonas mas altas. La herpetofauna del macizo esta depauperada como es tipico en un tepui en Venezuela.
Durante siete dias en la temporada seca (febrero) se recolectaron ocho especies de anfibios y reptiles. La muestra incluye dos ranas nuevas (Hyalinobatrachium eccentricum, n. sp., Cen- trolenidae, y Colostethus undulatus, n. sp., Dendrobatidae) y un género nuevo de lagartijas (Adercosaurus vixadnexus, n. gen., n. sp., Teiidae) todos colectados en el bosque himedo montano (a 1700—1750 msnm). También se encontr6 otra especie nueva de lagarto (Tropidurus panstictus, n. sp., Tropiduridae) en las zonas mas bajas (180-1220 msnm), especialmente en los arbustales secos.
BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
La herpetofauna ademas incluye una rana de tierras bajas (Pseudopaludicola llanera Lynch), dos ranas tepuyanas (Eleutherodactylus cantitans Myers y Donnelly, E. yaviensis Myers y Donnelly), una lagartija tepuyana (Prionodactylus goeleti [Myers y Donnelly], n. comb.), una culebra (Liophis?) que escap6, y otra culebra (Thamnodynastes corocoroensis Gorzula y Ay- arzagiiena) colectada por S. Gorzula en 1987. Las dos especies de Eleutherodactylus y el Prionodactylus se encuentran también en el cercano Cerro Yavi (localidad tipo), aunque una de las ranas (E. yaviensis) y la lagartija son ligeramente diferentes. La culebra Thamnodynastes corocoroensis parece ser distinta de una especie relacionada del Cerro Yavi.
Dos de las especies nuevas presentan caracteres no descritos previamente. El dendrobatido Colostethus undulatus, n. sp. tiene sobre el dorso de la mufieca una almohadilla supracarpal glandular, mas desarrollada en los machos, y tiene también el hueso paraesfenoides curiosa- mente escondido. Esta especie ademas difiere de los otros Colostethus tepuyanos porque carece del proceso lingual medial descrito recientemente. El centrolénido Hyalinobatrachium eccen- tricum, n. sp. tiene un extrafio iris bicolor, con una parte central obscura que esconde la pupila y que aparentemente se dilata con ella. Este caracter se mantiene en el material preservado y sirve para diferenciar a H. eccentrium de la especie similar H. crurifasciatum Donnelly y Myers. Las dos especies también comparten un caracter no observado previamente que es una ampolla o burbuja, aqui denominada bulla, en la membrana entre los dedos manuales tercero y cuarto (posiblemente debido a parasitismo?).
NO. 261
INTRODUCTION
This is the second of three anticipated her- petology reports from the 1995 American Museum—TERRAMAR Expedition, which es- tablished helicopter-supported camps on sev- eral sandstone table mountains (fepuis) in headwater drainages of the Rio Manapiare— some 100 km east of the middle Rio Orinoco, between 5° and 6° North Latitude, in south- western Venezuela. The first report (Myers and Donnelly, 1996) dealt with the fauna of Cerro Yavi, which has a small summit more than 2100 m above sea level. The present account deals with Cerros Yutajé and Coro- coro, which together form a much larger but generally lower massif to the west of Cerro Yavi.
The Yutajé—Corocoro massif has received little attention from biologists. The summit of neighboring Cerro Yavi was explored in 1947 by the late William H. Phelps, Jr. and his wife Kathleen D. de Phelps, in company with Charles B. Hitchcock of the American Geographic Society. Hitchcock’s report of the Phelps’ expedition showed an aerial view of Cerro Yavi, with the ‘“‘eastern cliffs of Ser- rania Yutajé”’ showing in the middle distance (Hitchcock, 1947: fig. 26 [reprinted in Myers and Donnelly, 1996: fig. 2B]). Although the existence of Cerro Yutajé was known to W. H. Phelps, Sr. and W. H. Phelps, Jr., they thought that it was too low to be of much ornithological interest (fide Kathleen D. de
Phelps, in conversation, 1995), and they made no plans to include it in their ambitious survey of the highland bird fauna of southern Venezuela (see Mayr and Phelps, 1967).
Several botanical collections were made on the Yutajé—Corocoro massif over a period of several decades, starting with work by Bassett Maguire in 1953 and ending with Otto Huber’s expeditions in the late 1980s (Huber, 1995b: 93). Huber was accompanied by Stefan Gorzula, who obtained several her- petological specimens. Gorzula and Sefiaris (1999: 255, 257) summarized this material as comprising one amphibian and two reptiles from 2150 m on Corocoro (November 8-12, 1987), and two reptiles from 1800 m on Yu- tajé (March 19-23, 1988). Included in Gor- zula’s collections was a new species of snake (Gorzula and Ayarzagtiena, *‘1995”’ [1996]).
We here report on a collection of 117 spec- imens obtained by the 1995 AMNH-—TER- RAMAR Expedition during a 7-day period in the dry season (February 24—March 2). The collection contains eight species (5 frogs and 3 lizards), of which half are new. By includ- ing one snake seen but not caught, and the additional snake taken earlier by Gorzula, the known herpetofauna of the Yutajé—Corocoro massif consists of 10 species. As is charac- teristic of tepui environments, this is a de- pauperate fauna, although other species cer- tainly wait to be found, especially in the wet season.
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 5
‘vy
EN Ny de ON -
- + i |
F SROCO
Fig. 1. Side-looking radar image showing the Yutajé—Corocoro massif. The names Cerro Yutajé and Cerro Corocoro apply to those parts of the massif lying to the east and to the west, respectively, of the upper valley of the Rio Corocoro (see footnote 2 for synonyms). Neighboring Cerro Guanay is shown to the northwest; Cerro Yavi is off the map, about 28 km northeast of the landing field at Yutajé. Three collecting sites for the Yutajé-—Corocoro massif are numbered as follow: 1. Cerro Yutajé camp at 1700 m elevation, on upper Rio Corocoro, 5°46’N, 66°08'W (see figs. 6 [bottom], 7). 2. South end of Cerro Corocoro, 1220 m elevation, 5°42’N, 66°10'W, near prominent conical peak (see figs. 4, 5). 3. Head of dugout navigation on Rio Corocoro, above Yutajé, a fishing camp at 180 m elevation, 5°37’N, 66°07'W.
The coordinates above were determined by GPS satellite receivers. However, positions of the num- bered collecting localities are placed on this radar image from helicopter and ground observation and with reference to other topographic maps (see text)—the published map coordinates for this image are accurate to only 02’. (From sheet NB-19—12, 1:250,000, elaborated by International Aero Service Cor- poration and published by Comision para el Desarrollo del Sur de Venezuela [CODESUR], Ministerio de Obras Publicas, Venezuela, 1971.)
BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
Fig. 2. Outer perimeters of the Yutajé—Corocoro massif, as viewed from the air. Top: The southern slope from small plane reconnaissance, on September 17, 1994. Bottom: Cerro Corocoro at the high northwestern side of the massif, from a helicopter on February 18, 1995.
MATERIALS AND METHODS
DEPOSITION OF SPECIMENS: Contractual agreements with PROFAUNA and INPARQUES provided that one-half of our collection, in- cluding all holotypes and half of the species
represented by single specimens, would be cataloged in the collection of the Museo de la Estaci6n Biolégica de Rancho Grande (EBRG) in Maracay, with the other half go- ing to the amphibian and reptile collections of the American Museum of Natural History
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 7
Fig. 3. The eastern escarpment of Cerro Yutajé, as viewed from a helicopter (February 18, 1995). Top: Looking northward along the eastern rim on approach to Cerro Yutajé. Bottom: Looking south- ward along the rim while leaving Cerro Yutajé. Visible landmarks in the lower photograph include (1) Cerro Camani on horizon, right of center, and (2) farther to the right but closer, the prominent small conical peak on the southern end of the Corocoro portion of the Yutajé—Corocoro massif (see figs. 1, 4,5).
8 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
(AMNH) in New York. We thank Francisco J. Bisbal, Director of the EBRG Museum, for facilitating the cataloging process and for helping in other ways.
Most specimens were collected by a her- petology team comprised of ourselves and John Daly, but, as was commonplace in the camps of this expedition, obliging colleagues of every discipline lent a hand. All material is cataloged as having been collected by the 1995 AMNH-—TERRAMAR Expedition—an ab- breviation for the official name ““The Robert G. Goelet American Museum—TERRAMAR Expedition to the Northwestern Tepuis’’.
METHODS: Insofar as practical, we find it useful to follow Ruiz-Carrenza and Lynch’s (1991la, 1991b et seq.) standardized format for definitions and descriptions of centrolen- ids, and Lynch’s long-term format and meth- odology for eleutherodactyline leptodactylids (best summarized in Lynch and Duellman, 1980, 1997). For anuran webbing formulae, we use the increasingly common notational device of Savage and Heyer (1967) as mod- ified for general use by Myers and Duellman (1982: 6); see also Myers et al. (1991: 6) regarding the desirability of using this con- vention for dendrobatids. Tadpoles were staged according to Gosner (1960). Measure- ments of less than 12 mm were taken with an ocular micrometer fitted in a Wild dis- secting microscope; larger measurements were made with digital calipers or a ruler read to the nearest 0.1 or 1 mm, depending on size. Sound spectrographs and waveforms were produced using a Kay 5500 DSP Sona- Graph.
In the field, coordinates were determined by two hand-held Global Positioning System (GPS) satellite receivers. We rounded read- ings to the nearest minute of latitude and lon- gitude.' Our receivers always were in agree- ment to the nearest minute, but GPS at that time was particularly untrustworthy for ele- vation, which therefore was determined
' Readings to seconds or to minutes with decimal no- tation would have given a false sense of precision at that time. Prior to May 2000, point-position accuracy of ci- vilian GPS devices was limited (for military reasons) “‘to about 100 m by reducing the accuracy of the broad- cast ephemeris, altering the clock epoch, and dithering the clock frequency, thus affecting both code-correlating and codeless receivers’? (Dixon, 1991: 252).
NO. 261
mainly by a Lufft pocket altimeter (20-m graduations). On two occasions, the altimeter was set at the known elevations of different airports near Caracas (about 9 a.m. each time), and subsequent readings were taken at the Yutajé base-camp landing field. An ele- vation of 180 m was accepted for the landing field at Yutajé based on these readings (Sep- tember 16, 1994: 210 m on arrival at 2 p.m., but 180 m at 7 a.m. the following morning; February 4, 1995: 180 m on arrival at 11:45 a.m.). Because of daily fluctuation in atmo- spheric pressure, a morning reading was cho- sen when possible. Elevations recorded for our two main collecting sites were about 200 m lower than shown on a topographic map (see Collecting Sites below).
THE NATURAL LANDSCAPE
The Yutajé—Corocoro massif is a highly eroded sandstone mountain on which most of the drainage flows inwardly to the central val- ley of the Rio Corocoro, which runs south- ward to exit the southern edge of the massif near Yutajé, a fishing camp and earthen land- ing field (fig. 1). The southern side of the massif (fig. 2, top) is comprised of steeply forested slopes with several prominent water- falls that have greatly reduced flows in the dry season. The western side of the massif (fig. 2, bottom) often was obscured by clouds during our brief overflights, but it is much more steeply sloped than the southern side, and there are some noncontinuous vertical cliffs marking parts of its upper edge. The eastern side is guarded by vertical cliffs, giving an appearance more in keeping with the common perception of tepui landforms (fig. 3).
The high and often jagged northern pe- riphery drops very sharply into a low rugged system of ridges and hills (fig. 1), with the northern base of the massif being higher than the low (roughly 200 m) southern base. Some of this complicated ridge system forms a narrow connection between the northeast- ern base of Cerro Yutajé and the northern base of Cerro Yavi, an otherwise isolated te- pui having a small summit entirely above 2100 m (Myers and Donnelly, 1996). Except for the broken country beyond its northern and northeastern side, the Yutajé—Corocoro massif is bordered by lowlands (180 m at
2001
Yutajé) supporting a mosaic of savanna and forest. Fires in the dry savannas near Yutajé were visible one night from our camp at 1700 m, in the upper drainage of the Rio Corocoro.
The northern edge of the Yutajé—Corocoro massif extends irregularly from east to west along the southern border of Bolivar State, with the greatest part of the massif protrud- ing southward into the northern part of Ama- zonas State. The highest elevations according to Huber (1995a: 43, 47) are 2140 m near the northeast edge of Cerro Yutajé, and 2400 m in the northwest corner of Cerro Corocoro, but the southern three-fourths of the massif lies mostly below 1700 m. Huber (1995a: 43) estimated the entire summit area as 275 km?.
There is no reason other than convention for not using a single name for the whole massif, which our party in the field simply called “‘Cerro Yutajé”’. The names Cerro Yu- tajé and Cerro Corocoro seem to be old des- ignations for the high parts of the massif on its northern perimeter. Each name also has been vaguely conjoined with Serrania, either as separate or synonymous entities (e.g., *“*Serrania Yutajé o Corocoro”’ [Hitchcock, 1947: pl. 2]). Huber (1995a: 43, 47 [map]) extended the names Yutajé and Corocoro southward to include all of the two inwardly ‘inclined summit plateaus’’ drained mainly by the Rio Corocoro (Rio Yutajé), and we follow his usage here as a practical solution.’
? CARTOGRAPHIC NOMENCLATURE: We follow older maps, and also Gorzula (in Gorzula and Ayarzagiiena, ““1995”" [1996], and Gorzula and Sefiaris, 1999), in us- ing the orthography Corocoro, which Huber (1995a, 1995b, 1995c) consistently rendered “‘Coro Coro” with- out explanation.
For the major stream exiting the south side of the Yutajé—Corocoro massif, we employ the name Rio Cor- ocoro simply because it is the name used by local people at the Yutajé base camp. In this sense, Rio Corocoro is synonymous with Rio Yutajé of some maps and authors. To confuse matters more seriously, two official govern- ment maps use Rio Maniapure for the major stream draining the Yutajé—Corocoro massif, and they assign the names Cafio Yutajé and Cano Corocoro to two other streams—tributaries that exit the massif to the east and to the west, respectively, of the ““Maniapure’’, before joining the larger stream either below or above (respec- tively) the landing strip at Yutajé (1:100,000 topographic sheets 6932 [Guaviarito|] and 6933 [Serranita de Guan- ay], published by Servicio Aut6nomo de Geografia y Cartografia Nacional in 1988 and 1990).
Hitchcock’s (1947) map showed Cano Yutajé as in the
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 9g
The summit of the massif drops generally southward (fig. 4), as well as inward, but the landscape is very irregular. A small conical peak near the southern end of Cerro Coro- coro forms an important landmark by which accuracy of map coordinates could be judged; the conical peak is identifiable in fig- ure 1 and is seen in this paper in various aerial views (see below under Collecting Sites). The sandstone base of the massif is frequently exposed (e.g., fig. 5, top). At el- evations of 1220 and nearly 1700 m, glimps- es were had along stream banks of displaced blocks (at 1220 m) or of a stratum of coarse conglomerate containing cobbles of sand- stone or quartzite (fig. 5, bottom). Conglom- erate is a rare type of rock among the sand- stones and quartzites of the Roraima For- mation. It seems likely to be the same con- glomerate that occurs at similar elevations on neighboring Cerro Yavi. (During their ascent of Yavi in 1947, the Phelps Expedition found interbedded conglomerates and sandstones between 4840 and 6000 ft [1476-1829 m],° with conglomerates becoming rare above 6000 ft [Hitchcock, 1947: 561—562].)
The rocky soil of the Yutajé—Corocoro massif supports varied plant associations, as briefly mentioned by Huber (1995a: 43; 1995c: 124-125, 135, 153). There seem to be no grasslands. The tepui meadows men- tioned by Huber (1995c: 153) are of his broad-leaved, nongramineous type dominat- ed by Kunhardtia rhodantha. We did not en- counter such meadows on the Cerro Yutajé side of the massif, but they are common on the higher parts of Cerro Corocoro. Huber
government maps above, but it used a different name (Rio Manapiare) for the main stream. All these streams and many more form the headwaters of the Rio Mana- piare (an established name), itself a major tributary of the more southern Rio Ventuari (also an established name). Instability in riverine nomenclature is mainly a phenomenon of unpopulated or sparsely populated head- waters, where boats cannot go far and where the cartog- rapher’s pen is unhindered by established tradition.
3 The metric system is used in a Spanish translation (Hitchcock, 1948: 174-175), in which 4840 and 6000 feet are inexplicably converted to 1900 and 2000 m. Similar departures from Hitchcock’s original elevations (vs. distances) occur throughout the translation (by W. H. Phelps, Jr.). The differences are too great for rounding errors, leading one to wonder whether corrections were intended.
10 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
Fig. 4. Top: Helicopter view from the north-central part of the Yutajé—Corocoro massif, looking southward over the massif to Cerro Camani on the far horizon (February 18, 1995). The closer small conical peak is on the southern end of Cerro Corocoro; it is a constant landmark. Bottom: Approaching the conical peak seen above, still looking southward over the eroded interior of the Yutajé—Corocoro massif (February 5, 1995). See figure 1 for orientation.
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 11
Fig. 5. Top: Small plateau, 1220 m elevation, just north of the conical peak (see also fig. 4). The habitat is hot, open scrubland, with frequent areas of bare rock. The numerous dark stems with terminal rosettes are the peculiar Vellozia tubiflora (Velloziaceae). This area, including an adjacent, rocky ravine (see below), is the type locality of the lizard Tropidurus panstictus. Bottom: Rocky stream bed (lower left) near the base of the conical peak shown above, habitat of Pseudopaludicola llanera and Tropidurus panstictus. Note the large block of sandstone conglomerate. (February 28, 1995)
12 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
described one Kunhardtia meadow at 1850 m elevation on Cerro Yutajé, and we landed briefly in another at 1910 m in the high northwestern side of Cerro Corocoro (fig. 6, top). Kunhardtia-dominated meadows and wet slopes are a major habitat at least on the high slopes of Cerro Corocoro, but we un- fortunately lacked time to establish a camp or to collect in them.
Scrubland (Huber, 1995c: 135) covers much of the massif, and we collected in two diverse types of scrub (fig. 5, top, and figs. 7, 8). The extensive scrublands of the Yuta- jé—Corocoro massif are broken by areas of forest, which Huber (1995c: 124—125) divid- ed into three types: (1) basimontane, semi- deciduous forest below 500 m on southern slopes; (2) dense, medium-sized, submontane to montane evergreen forest below 1500 m in the ascending interior uplands; and (3) dense, low-sized, wet, evergreen upper mon- tane forest, above approximately 1500 m. We collected only in the last type of forest—a wet mossy forest—which seems to occur es- pecially in low areas near the highland streams (figs. 6, 7, 23); it is an important habitat for some amphibians and possibly for a few reptiles as well.
COLLECTING SITES
We collected at three places on the Yutajé— Corocoro massif, as shown in figure 1, which requires comment. The side-looking radar image reproduced herein as figure 1 is pre- sumably the one used for Huber’s (1995a: 47) sketch map. The coordinates shown in our respective renditions of this map give a false impression of geographic precision, which was initially puzzling to us in the field. Our problem was that the old radar image is accurate to only 02’, as stated on a later ad- aptation of the image.* The map inaccuracy is reflected both in longitude and in parallels of latitude, as is most evident for locality 2 (figs. 4, 5 [top])—a small plateau at the northern base of the small conical peak that is readily identified on the map.
4 Sheet NB-19-—12, 1:250,000, Picto Radar, produced by the Direcci6n de Cartografia Nacional in 1982, from the 1971 Mosaico de Radar (see fig. 1), with the added qualification that geographic precision is on the order of O27:
NO. 261
If our localities were plotted on figure 1 by use of the GPS coordinates, locality 2 would lie too far west and also well north of the southwestward running escarpment that nearly isolates the little plateau from which the conical peak rises. Our locality 1 (arrow in fig. 6, bottom) also would be shown too far west and too far north. Furthermore, Hub- er and Gorzula’s 1987 camp on Cerro Cor- ocoro would lie outside of the summit area if plotted on the outline map in Huber (1995a: 47).°
The following three collecting sites were therefore placed in figure 1 by ground and helicopter observation, in conjunction with GPS coordinates referenced to available to- pographic maps (1:100,000, sheet 6932 [Guaviarito], 1988, and sheet 6933 [Serrania de Guanay], 1989, Servicio Aut6nomo de Geografia y Cartografia). These maps show localities 1 and 2 as being 180—220 m higher than our altimeter readings, but there is no guarantee of accuracy (map contours were derived from aerial photographs taken 30 years ago, with few ground checks).
LOCALITY 1: Cerro Yutajé campsite on a branch of the upper Rio Corocoro (GPS co- ordinates 5°46’N, 66°08’W), 1700 m by al- timeter (about 1920 m on sheet 6933, I: 100,000). Camp consisted of a streamside kitchen shelter, with tent sites scattered about in the adjacent forest. Some woody branches along the stream were removed so the heli- copter could safely land in the stream bed with a full load. Although the gorge of the Rio Corocoro forms a major north-to-south break in the Yutajé—Corocoro massif, the riv- er at 1700 m and above is shallow and not entrenched. The stream is a clear coffee color and flows over a bed of sandstone. A fresh vertical exposure at camp is of a fine-grained sandstone of thin, alternating pink and white strata. Several members of our party who worked their way downstream from camp commented on a thick stratum of sandstone conglomerate interbedded between sand-
> GPS coordinates and elevations for Huber and Gor- zula’s 1987-1988 camps were given by Gorzula and Sefiaris (1999: 255, 257 [localities H-O17 and H-052]).
© There was no extensive clearing of vegetation. Con- scientious attempts were made in this and all other camps to remove all garbage by helicopter. We hope that any traces of our presence are few and transient.
2001
stone—similar to the displaced blocks of conglomerate seen in another stream at lo- cality 2 (fig. 5, bottom). Although there was a sharp drop in the river below camp (notice the entrenchment in lower right of fig. 6, bot- tom), the river upstream is open and provides an easy highway through dense vegetation. A short walk upstream, to an elevation of about 1750 m, brought one to a tributary en- tering on the east bank. A new species of Colostethus was common here, although it appeared to be completely absent in the for- est downstream across from the 1700-m camp, even though the habitat was continu- ous.
The upper Rio Corocoro is a mosaic of scrub and wet mossy forest (fig. 6, bottom). Near camp, the rocky slopes on the western side of the river support a dense scrub of small trees (7yleria may be dominant; Clusia and a few palms are also present), Brocchinia, and Stegolepis (figs. 7, 8).
In great contrast to the dense, sun-exposed scrub on the west bank, a low wet forest par- allels the river along its eastern side (see fig. 6, bottom). This narrow strip of forest con- tains its own small tributary streams, canopy- covered and with sediment, in contrast to the open, scoured bed of the nearby river. There is abundant leaf litter in the forest and large bromeliads on the ground and in the trees, which bear a conspicuous moss layer remi- niscent of cloud forest. As inferred from fig- ures 6 (bottom) and 7 (top), this mossy forest is continuous with similar forest upstream along the Rio Corocoro (fig. 23).
The wetness of the mossy forest results from not only local peculiarities of drainage, but also from air currents that ascend the val- ley and bring orographic rainfall to this part of the massif. Streams were drying out dur- ing February on the southern part of the mas- sif, the grass turned brown at the Yutajé base camp, and fire was being set to the nearby savannas. On a clear night, the burning sa- vannas were visible perhaps 20 km or farther from our highland camp (which was 17 km airline north of Yutajé).
Although it clearly was dry season below, daily rainfall was recorded at the 1700-m camp by both groups who worked there from February 11 through March 1. Light or heavy, the montane rain came every day dur-
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 13
ing early or midafternoon. On March 1 (our last night in camp), rainfall upstream from camp was particularly heavy, causing the riv- er to rise alarmingly by dusk. Camp activities were disrupted and foot travel along the river was difficult. The next morning, on very sog- gy ground, Paul Sweet picked up the only known specimen of a new genus and species of lizard.
Loca.iry 2: South end of Cerro Corocoro (GPS coordinates 5°42'N, 66°10'W), 1220 m (about 1400 m on sheet 6933), on small pla- teau with an incised streambed, 2.9 km (air- line) due north of the conical peak that is a prominent landmark on the southern side of the massif. The area was examined by small- plane reconnaissance in September 1994, but the locality was too low for the single camp to be established on the Yutajé—Corocoro massif. We visited this site by helicopter from the 1700-m camp; locality 2 was worked for several hours on February 28 and again on February 29, 1995. The only her- petological specimens collected or seen were the frog Pseudopaludicola llanera and a new species of the lizard genus Tropidurus.
The little plateau supports a dry scrub that is very different and much more open than the scrubland at locality 1. The openness must be largely due to thin soil and the fre- quent exposure of sandstone at ground level (fig. 5, top). The dominant vegetation in- cludes a low stratum of Vellocia tubiflora, Stegolepis, and Brocchinia, and a_ sparse growth of small trees, especially Tyleria.
There is a very narrow strip of gallery for- est on the plateau along a small incised stream, in which little water was flowing in late February. The streambed is full of sand- stone blocks and boulders, as well as large, displaced blocks of conglomerate (fig. 5, bot- tom) such as seen in situ below the 1700-m camp. As mentioned earlier, conglomerate is an unusual rock in the Roraima Formation.
LOCALITY 3: Rio Corocoro above Yutajé, 180 m (GPS coordinates for Yutajé, 5°37'N, 66°07'W). This locality (El Salto on sheet 6932) is about 15 minutes by motorized dug- out upstream from Yutajé and cannot be missed, since it is the head of navigation at the southern base of the Yutajé—Corocoro massif. It is an area of forest with some open rocky slopes. Considering the drainage area
14 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
Fig. 6. Top: A broad-leaved (nongramineous) tepui meadow on Cerro Corocoro, 1910 m above sea level, in the high northwestern quadrant of the Yutajé—Corocoro massif. The dominant plants are Kun- hardtia (Rapateaceae), with green straplike leaves, and Brocchinia (Bromeliaceae), the broader, yellow- ish-leaved tank bromeliads, all growing on a sphagnum substrate (February 14, 1995). Bottom: Heli- copter view near the northern edge of Cerro Yutajé, looking northward over upper headwaters of the Rio Corocoro, showing scrubland with interspersed forest (February 18, 1995). The arrow indicates a campsite at 1700 m above sea level. The forest edge across the stream from camp is the type locality of the lizard Adercosaurus vixadnexus. East of camp, paralleling the open main stream, a strip of dense, mossy gallery forest marks the course of a small woodland brook, the type locality of the frog Hyali- nobatrachium eccentricum. The type locality of the frog Colostethus undulatus is situated farther up- stream in this gallery forest, at about 1750 m elevation (see fig. 23).
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 15
Fig. 7. Views of the upper Rio Corocoro valley, from the slope above the 1700-m camp (see arrow in fig. 6, bottom). Top: Looking northeastward (upstream) toward the Colostethus locality on far side of river. Bottom: Looking southeastward to helicopter about to sit down in the streambed at camp.
The hillside above the 1700-m camp is covered in Tyleria-Stegolepis-Brocchinia scrub growing on rocky soil. Peering uphill (looking for the cameraman who had just called to him), ornithologist George Barrowclough provides a sense of scale for the density and height of the scrub vegetation. See figure 23 for interior views of the mossy gallery forest across the river.
16 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
NO. 261
Fig. 8. short distance below the 1750-m Colostethus locality. The open, scoured streambed, flanked here by rocky scrubland on the west bank, is in stark contrast to the small woodland streams in dense, mossy forest on the eastern side of the valley (for overviews, see figs. 6 [bottom], 7; for interior of mossy forest, see fig. 23).
involved, the Corocoro seems in the dry sea- son to be a small river, although it can be- come torrential. There are low falls over hard sandstone or quartzite that has a colorful wa- ter-polished surface. The rocky channel is well scoured, although there are small pock- ets of the characteristically fine, white Ro- raima sand. The place is used as a swimming hole by visitors to Yutajé. Only Tropidurus was collected at this locality (on March 2, £995).
SPECIES ACCOUNTS
FROGS FAMILY CENTROLENIDAE
Hyalinobatrachium eccentricum, new species Figures 9-12, 13B, 13C, 14
HOLOTYPE: EBRG 3049 (field no. CWM 19713), an adult male from waterfall in small forest stream on Cerro Yutajé, 1700 m (5°46’N, 66°08’W), Amazonas, Venezuela; collected February 26, 1995, AMNH—TER- RAMAR Expedition. See locality 1 on map
(fig. 1)
Looking southward down the upper Rio Corocoro toward direction of the 1700-m camp, a
PARATOPOTYPE: AMNH 159164 (CWM 19712), an adult male with same collecting data as holotype.
ETYMOLOGY: The species name is a Latin adjective meaning “‘not concentric’’, in al- lusion to the unconventional elliptical to domelike shape of the median part of the iris.
DEFINITION AND DIAGNOSIS: A Hyalinoba- trachium belonging to the fleischmanni spe- cies group and characterized by the follow- ing combination of characters (in the format of Ruiz-Carranza and Lynch, 1991b, with unusual characters in italics): (1) Womerine teeth absent; (2) bones white in life; (3) pa- rietal peritoneum clear; liver, stomach, and intestines, and a variable part of pericardium, silvery white, but gallbladder exposed; (4) in life, head and dorsum with pale green or pale yellow spots set in a light green reticulum dotted with dark green—changeable to pure green with green dots; hind legs with narrow, faint green crossbars; in preservative, spots and limb bands lost, dorsum becoming cream with purplish black melanophores; iris bi- colored, with a brown circumpupillary zone concealing pupil (pupillary ring absent) and
2001
a golden yellow peripheral zone with gray flecks; (5) webbing absent between inner fin- gers; webbing formula for outer fingers HI 2—2 IV—a bulla in outer-finger webbing; (6) webbing formula for foot I 2—2 II 1%—2 Ill 14-Q-2%) IV (2-2%)-1 V; (7) snout trun- cate in dorsal and lateral views; (8) dorsal skin smooth in preservative; (9) very weak ulnar and tarsal folds arguably present; (10) humeral spine absent; (11) tympanum cov- ered by skin (visible through skin); (12) size medium, males about 21—23 mm SVL, fe- males not known; (13) nuptial excrescence present as small patch of gland cells on dor- somedial base of thumb, continuing proxi- mally along thenar tubercle.
In general morphology and coloration— including presence of a bulla in the outer- finger webbing (fig. 12) and narrowly cross- banded legs in life (fig. 9)—Hyalinobatrach- ium eccentricum closely resembles A. crur- ifasciatum Myers and Donnelly, but eccentricum is immediately separated by its bicolored iris. The median dark zone sur- rounding and concealing the pupil (which lacks the usual bright pupillary ring) is di- agnostic even in preservative (fig. 13).
MEASUREMENTS OF HOLOTYPE AND PARA- TOPOTYPE (in mm): Both undissected speci- mens are adult males as revealed by presence of vocal slits and distensible, subgular vocal sacs. The following measurements are for the holotype first, paratopotype second: SVL 21.6, 22.9; tibia length 10.5, 11.4; foot length from proximal edge of inner metatarsal tu- bercle to tip of longest toe 9.6, 9.7; head width 8.6, 8.9; greatest body width 9.3, 9.4; head length on the diagonal from angle of jaw to tip of snout 6.4, 6.4; upper eyelid width 1.4, 1.3; interorbital distance 2.9, 3.0; eye to posterior edge of nostril 1.8, 1.9; eye length 2.9, 2.9; tympanum not well defined, about 1.0, 1.0; width of fourth finger disc 1.4, 1.4; width of third toe disc 1.0, 1.2; width of fourth toe disc 1.3, 1.4.
DESCRIPTION
The two specimens in the type series are moderate-sized adult males, falling within the range of 21—23 mm SVL. See above for measurements.
MORPHOLOGY: Head slightly narrower than
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 17
body; head width 39-40% of SVL. Snout short, truncate in dorsal view and in profile; end of snout nearly vertical in profile or (in holotype) with a slight posterodorsad incli- nation; canthus rostralis indistinct; loreal re- gion sloping outward to lip; lips not flared; nostrils terminal, protuberant, directed an- terolaterally; internarial region slightly con- cave. Eyes large, directed anterolaterally, length greater than eye—nostril distance; eye length/eye—nostril distance = 1.5—1.6. Upper eyelid width 43—48% of interorbital distance. Supratympanic fold absent or indistinct; tym- panum covered by skin but visible through skin (at least in preserved specimens). Vo- merine teeth absent; choanae small or mod- erate (~ 0.5 mm), rounded; tongue round, entire, barely free posteriorly. Vocal slits ex- tending from lateral base of tongue toward angle of jaw; distensible subgular vocal pouch.
Humeral spine absent; a very weak low, glandular ulnar fold,’ faintly pigmented white, extending from elbow along ventro- lateral edge of arm to disc on finger IV. Rel- ative length of appressed fingers I] > IV > I ~ Il. Webbing absent between inner fin- gers; webbing formula for outer fingers HI 2-2 IV; a bulla in outer-finger webbing (see fig. 12 and Remarks). Finger discs broadly rounded to truncate. Thenar tubercle ellipti- cal, low; palmar tubercle elliptical or some- what rounded, low and barely distinct; sev- eral indistinct supernumerary palmar tuber- cles; subarticular tubercles small, low, entire. Nuptial excrescences not distinctively pig- mented (not white), comprised of a line of gland cells (barely discernible in holotype) along edge of thenar tubercle on medial base of thumb and distally expanded as a larger and more evident patch of cells on dorso- medial surface of third joint.* Prepollical Spine absent (or not externally detectable by probing).
’ The glandular nature of the ulnar fold is evident un- der magnification, but the structure is weak and not dis- tinctively colored (white pigment present but sparse) in the preserved material and could easily be overlooked. A similar metatarsal fold is weaker still.
8 These nuptial excrescences are medially much nar- rower and dorsally less extensive than in the type-1 pad illustrated by Flores (1985: fig. 3A—C), but the distri- bution of gland cells follows the same pattern.
18 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 9. Hyalinobatrachium eccentricum, new species. Two views of the paratopotype in life (AMNH 159164<d). Note the weak but distinct crossbanding on the hindlimb. The bicolored iris has a golden yellow periphery and a dark brown median section in which the expanded pupil is barely visible as a broad ellipse. Absence of the usual bright pupillary ring causes the pupil to be virtually concealed because of lack of contrast with the adjacent brown part of the iris. The dark median section of the iris appears to be somewhat changeable in size and shape (see under Re- marks).
Hind limbs slender; tibia length 49-50% of SVL. Faintest trace of a slightly whitish, low metatarsal fold from heel to disc of fifth toe. Inner metatarsal tubercle elliptical, low; outer metatarsal tubercle smaller, low, barely distinguishable; subarticular tubercles low, round, entire. Feet moderately webbed; web- bing formula I 2—2 IT 14-2 I 1%-2 or 2% [holotype] IV 2% [holotype] or 2—1 V. Toe discs broadly rounded, slightly smaller than those on fingers.
NO. 261
Fig. 10. Hyalinobatrachium eccentricum, new species. Ventral view of the holotype before pres- ervation (EBRG 3049<).
Skin relatively smooth overall except are- olate on belly and ventral thigh surfaces. Vent directed posteriorly at upper level of thighs; cloacal opening concealed by short anal flap; anal ornamentation consisting of folds and tubercles with a slight whitish frosting overall.
COLOR AND PATTERN: Changeable in life, pure green with dark green dots when seen at night (not photographically documented). Otherwise, head and dorsum with a pattern of clear, pale yellow or pale green spots (about size of tympanum) set in a light green reticulum that is dotted with dark green chro- matophores; limbs grayish or yellowish, forelimbs with green dots irregularly ar- ranged, hind legs with green-dotted, faint green crossbars on thigh and shank (fig. 9).
Parietal peritoneum clear. Lineae mascu- lineae present. Pericardium variable: in ho- lotype, heart mostly exposed, with pericar- dium appearing (from ventral view in life) to
2001
Fig. 11. Hyalinobatrachium eccentricum, new species. Note the loss of crossbars on the limbs after preservation (compare with fig. 9). Left to right: AMNH 159164 3, EBRG 3049 2, approx. A ll.
have white pigmentation principally above the heart; in paratopotype, pericardium most- ly white except ventromedially, where heart was partly exposed. Liver, stomach, and in- testines closely bound in silvery white peri- toneum; gallbladder exposed (fig. 10). Bones white.
Eye coloration peculiar (figs. 9, 13B, C), with iris comprised of two strikingly differ- ent zones: (1) Peripheral zone of iris golden yellow, sparsely flecked with gray; and (2) medial, circumpupillary zone uniformly brown (little contrast with horizontal pupil, which lacks a bright pupillary ring) and markedly changeable in shape. The brown, medial section of the iris forms a broad el- lipse around the expanded pupil in life (fig. 9, bottom). However, after euthanasia of the two frogs in chlorotone solution, the hori- zontal pupil contracted and the surrounding brown area became noticeably flattened be- low the pupil and domelike above it (fig. 13C). During formalin fixation, however, the brown circumpupillary zone regained an el- liptical shape that has been retained in pre- servative (fig. 13B).
In preservative (fig. 11), dorsal surfaces are cream, eyelids paler; dorsal surfaces (in- cluding eyelids, lower arm, thigh, shank, tar- sus, foot, and base of toes IV—V) are dotted with purplish black chromatophores of vary- ing size. Crossbars on hind legs no longer evident. Ventral surfaces uniform cream, in- ternal organs not discernible.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 19
Fig. 12. Hyalinobatrachium eccentricum, new species. A, B: Right hand and left foot of male paratopotype (AMNH 159164). The raised or bubblelike ventral expansion of the web between fingers HI and IV is an apparently novel structure, the bulla, that is shared with H. crurifasciatum (see fig. 15). C: Dorsal outline of the above hand, showing the distinct dorsal depression of the bul- la. D: Right hand of the holotype (EBRG 3049). The bullae are not differentiated by pigmentation and are shaded for emphasis only. A possibility that the bullae might be induced by parasitism cannot be discounted and needs investigation (see Remarks in text). Scale line = 1 mm.
NATURAL HISTORY AND VOCALIZATION
This species—the only centrolenid seen or heard during our six nights on Cerro Yuta- jé—was found perched near a little waterfall above a small stream in closed, mossy forest. It was a small group of fewer than 10 males calling in a concentrated area; the frogs were calling by night up to 3 m aboveground in dense vegetation. They were the devil to find even with three headlights triangulating on the calls. The only two frogs located were
20 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 13. Diagnostic iris patterns retained in preservative, roughly X<10—-12. A: HAyalinoba- trachium crurifasciatum Myers and Donnelly, right eye of holotype (AMNH 131329¢)—show- ing horizontal pupil through the transparent nic- titating membrane. B: Hyalinobatrachium eccen- tricum, new species, left eye of holotype (EBRG 3049 3 )—showing dark central section of iris (the horizontal pupil is not evident in this photograph owing to lack of contrast); the nictitating mem- brane is deflected downward. C: Enlarged rendi- tion of field sketch of eye of Hyalinobatrachium eccentricum, after specimen had been relaxed in chloretone solution. In addition to the rough field drawing, the central part of the iris was noted as ““domelike, with flattened lower side, and brown (little contrast with horizontal pupil)’. The me- dian dark zone regained an elliptical shape after preservation (as in part B above).
NO. 261
calling upside down from the undersides of leaves.
The vocalization (fig. 14) is a single-note ‘“‘peep’’ infrequently given. A recording was obtained of a solitary note, captured early in nearly 2 min of tape. This note, 0.22 sec in duration, is irregularly but distinctly pulsed (see waveforms). There is initial frequency modulation, with the note abruptly rising from about 3480 Hz to a flat frequency of 4160 Hz (fig. 14, top).
REMARKS
The two type specimens were not dissect- ed. Although the gallbladder is exposed, the liver is bound in white peritoneum (fig. 10) as is characteristic of Hyalinobatrachium (Ruiz-Carrenza and Lynch, 199la: 23). Shape of the presumably bulbous liver, how- ever, was not explicitly determined. Field- notes stated “2 lobes visible through white covering”’.
Hyalinobatrachium eccentricum is named for its peculiar eye, which was not noted un- til the two specimens were examined with a jeweler’s loupe, after they had been killed in chloretone solution prior to preservation. The eccentricity of a sharply bicolored iris, with a dark, dome-shaped circumpupillary zone, seemed so bizzare that a field sketch was made (fig. 13C). There was no realization then that the median zone of the iris could change shape, but photographs of a living frog (fig. 9) suggest that the dark, median dome becomes an ellipse concurrent with pu- pillary expansion; a more shallow elliptical shape was regained during fixation (fig. 13B).
The fact that frog pupils contract to ex- traordinarily diverse shapes is well known (e.g., Boulenger, 1879; Duke-Elder, 1958: 339, figs. 403-404: 339; Walls, 1942: fig. 87) and is widely documented in countless published photographs of species in different genera and families. Nevertheless, we are un- aware of any report of a nonsymmetrical change in the appearance of the central part of the iris itself. Iris constriction in the few anurans studied seems to be mainly an au- tonomous response to light that is mediated by the sphincter pupillae muscle (e.g., Rubin
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 21
8 50 msec 6 MN T4 = ies = ee ee ee ee eee, ee ee 2 0 i i ‘ 8) 0.5 1.0 1.5
Time in Seconds
Fig. 14. The single-note advertisement call of Hyalinobatrachium eccentricum, new species, from a frog calling on underside of Philodendron leaf 1.3 m above small forest stream. Top: Sound spectro- gram graphed with narrow-band (59 Hz) filter, with frequency cursors set at 3480 and 4160 Hz, and matching waveform. Bottom: Wide-band (300 Hz) spectrogram of the same note and expanded wave- form. The time between waveform cursors is 0.219 sec. Specimen (AMNH 159164) recorded February 26, 1995, at air temperature of 17.0°C (AMNH herpetology reel 264:1).
22 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 15. AHyalinobatrachium crurifasciatum Myers and Donnelly. New drawings of the left hand and left foot of the male holotype empha- sizing the bulla in the webbing between fingers III and IV (compare with fig. 12) and correcting the shape of the distal subarticular tubercle on toe V. Scale line = 1 mm.
et al., 1986), but littke comparative informa- tion is available.
Another notable aspect of the eye of Hy- alinobatrachium eccentricum is the absence of a bright rim close around the pupil. Al- though absent in a few groups, a bright pu- pillary ring (continuous or broken) is so widespread among frogs (e.g., see fig. 16 of this paper) that it is usually ignored in de- scriptions of iris color. Walls (1942: 549) ob- served that
The common frogs, and many fishes and birds have the pupil outlined by a thin gold or silver line, the rest of the iris being so dark that the pupil would be beautifully concealed in it were it not for this metallic frame. To the adaptive colorationists, putting this ring around the pupil must seem about as mean a trick as hanging a bell on a cat.
A bright pupillary ring is usually present in centrolenids (e.g., see photographs in Lynch and Duellman, 1973; Ruiz-Carranza et al., 1996). There is little contrast between the pupil and the dark median zone of the iris of Hyalinobatrachium eccentricum, and the ab- sence of the bright ring therefore causes the horizontal pupil to virtually disappear in pho- tographs of both living and preserved spec- imens.
Another unusual characteristic of Hyali- nobatrachium eccentricum is the structure
NO. 261
here termed a bulla? in the webbing between fingers III and IV. It is a shallow bubble on the ventral side of the web (fig. 12A, D), corresponding to a narrower, definite depres- sion or pocket on the dorsal side (fig. 12C). We overlooked this structure in H. crurifas- ciatum, taking it only to be an artifact of preservation and not showing it in a diagram- matic drawing of the hand (Myers and Don- nelly, 1997: fig. 8). We rectify that omission here with a more carefully drawn figure of the hand and foot of A. crurifasciatum (fig. 15). The bullae in figures 12 and 15 are em- phasized by shading—there is no difference in pigmentation (or lack thereof, since the fingers and webs are essentially unpigment- ed; see figs. 9, 10), and the structure is easily discounted as an unimportant irregularity in the web or else it is overlooked entirely.
In the preceding discussion, we treat the bulla as a potentially informative taxonomic character, but there is another possibility that needs exploring. Our colleague Julian Fai- vovich has called our attention to parasitism in the Neotropical hylid genus Scinax, in which parasitic mites infect the web between toes IV and V. Examination of a few speci- mens of Bolivian Scinax nasicus (AMNH 144434, 144562—144563) revealed a vari- able-sized swelling, without a sharply de- fined perimeter, on the dorsal side of the out- er toe webbing; these dorsal swellings ap- peared to be caused by acarid mites that clus- ter in the shallow corresponding concavity on the ventral side of the web (a kind of ‘“‘mite pocket” usually associated with liz- ards); a sample of the mites from one spec- imen of Scinax nasicus (AMNH 144434) were identified as Microtrombicula sp. (Trombiculidae) by Dr. Ricardo Guerrero. Although the dorsal swellings in Scinax clearly are not the same as the well-defined bullae described above, one has to wonder whether a different parasite might account for (or be associated with) the dorsal pock- eting/ventral swelling in the centrolenid fin- ger web. The possibility cannot be discount- ed at this time, even though there is no direct evidence.
° Latin: bulla, -ae (bubble, knob, ornament), dim. bullula.
2001
Hyalinobatrachium eccentricum and H. crurifasciatum are structurally very similar. Unusual shared characters include the bullae (figs. 12, 15) and the narrow green crossbars on the legs (e.g., fig. 9, top) that disappear in preservative. The advertisement call of each species is a well-spaced, frequency- modulated, irregularly pulsed “‘peep’’, the single note starting at roughly 3.5 kHz and rising to = 4 kH, with a duration of 0.2 sec or more. However, only a single note of ec- centricum was recorded (fig. 14), whereas Myers and Donnelly (1997: 14-15) de- scribed call variation in note duration and frequency modulation in what appeared to be a single calling individual of crurifascia- tum; they suggested that such variability might be an antipredator mechanism. Except for their very different irises, we likely would have identified the specimens of H. eccentricum as being conspecific with H. crurifasciatum.
The bicolored iris of Hyalinobatrachium eccentricum appears to be shared with a spe- cies in Surinam, based on three transparen- cies of AMNH 87629 (CWM 11033), col- lected at Brown’s Berg in 1972.!° These slides show a Hyalinobatrachium that is col- ored similarly to H. eccentricum (pale green with yellowish spots); the iris has a periph- eral golden yellow zone (like eccentricum), but the large median black section appears to be circular (vs. elliptical). In addition to the circularity of the median black section, this frog differs from H. eccentricum in hav- ing a more truncate snout, in having the hind legs spotted like the body (vs. cross- barred), and in having the gallbladder con- cealed (liver, gallbladder, and digestive track closely bound in white peritoneum, heart and lungs exposed). Fieldnotes suggest that the median black part of the iris was simply thought to be a greatly expanded pupil; from the transparencies, however, it seems to be too large and too circular for that (with cor- rect lighting, the pupil should be discernible under the microscope).
10 Specimens were not compared directly, since the Surinam material (AMNH 87629, 87630) is on loan to the Nationaal Natuurhistorisch Museum for study by Dr. Marinus Hoogmoed.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 23
FAMILY DENDROBATIDAE
Colostethus undulatus, new species Figures 16—22
HOLoTyPE: EBRG 3021 (field no. CWM 19700), an adult female from forest stream on Cerro Yutajé, 1750 m (5°46'N, 66°08'’W), Amazonas, Venezuela; obtained February 24-25, 1995, by the AMNH-—TERRAMAR Ex- pedition. The type locality is a short walk upstream from from the Expedition’s 1700- m camp; see locality 1 on map (fig. 1).
PARATOPOTYPES: AMNH 159118—159142 (159139-159140 skinned carcasses, 159141—-159142 cleared & stained), 159143 (7 tadpoles), EBRG 3022-3047 (3024 cleared & stained), 3048 (8 tadpoles), from same locality as holotype, collected in the period February 24-28, 1995, by the AMNH-TERRAMAR Expedition.
ETYMOLOGY: The species name is a Latin adjective meaning undulating or wavy, in reference to the characteristic dorsal mark- ing.
DEFINITION AND DIAGNosIS: A medium- sized Colostethus (about 20—25 mm SVL) with black sides and a usually conspicuous wavy-edged dorsal marking; feet webless ex- cept for rudimentary basal web between toes III and IV. Lacking pale lateral or discrete dorsolateral lines, lacking cloacal tubercles, and lacking a median lingual process. Un- usual external characters include a swollen supracarpal pad in males and an unusually small subarticular tubercle on base of fourth toe; an unusual osteological character is the concealment of all but the cultriform process of the parasphenoid. (See Remarks).
The undulating dorsal pattern also is un- usual in Colostethus, although similar but less constant markings occur in the variation of a few species, including C. mandelorum (Schmidt), a moderately webbed species from the eastern coastal range of northeast- ern Venezuela. The absence of a pale ingui- nal line and its nearly webless feet immedi- ately differentiates C. undulatus from that species, as well as from C. guanayensis on an adjacent tepui.
MEASUREMENTS OF HOLOTYPE (in mm): The holotype is an adult female as deter- mined by examination of its reproductive tract (and as indicated by size, absence of
24 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
Fig. 16. Colostethus undulatus, new species, in life. Top: A male paratopotype (AMNH 159188). Middle: The female holotype (EBRG 3021). Bottom: Same specimens in ventral view, male on left, female holotype on right.
2001
TABLE 1 Size (in mm) and Proportions of Colostethus undulatus, New Species, from the Type Locality?
Character N Mean + 1 SE Snout-vent length (SVL) 14¢ 20.69 + 0.24 27 2 22.20 + 0.21 6 juy. 13.32 + 0.65 Tibia length>/YSVL 14¢ 0.465 + 0.004 27 & 0.439 + 0.004 6 juv. 0.433 + 0.015 Head width*/SVL 143 0.338 + 0.002 27 9 0.329 + 0.002 6 juv. 0.348 + 0.003 Center naris to edge 14¢ 0.707 + 0.014 eye/eye length Pee 0.703 + 0.010 6 juv. 0.629 + 0.026 Hand length*/SVL 14¢ 0.273 + 0.003 271 2 0.257 + 0.002 6 juv. 0.237 + 0.008 Hand length/head width l4¢ 0.809 + 0.008 eae 0.782 + 0.008 6 juv. 0.681 + 0.028 Width 3rd-finger disc/ 14¢ 1.367 + 0.023 finger width below disc¢ 27 2 1.443 + 0.020 6 juv. 1.249 + 0.053
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF
SD CV (%) Range
0.90 4.36 19.6-21.9 1.09 4.90 20.1-25.3 1.60 12.01 12.0-16.2 0.015 3.15 0.44-0.49 0.020 4.45 0.41-0,48 0.036 8.23 0.38-0.47 0.009 2.74 0.32-0.35 0.013 3.86 0.30-0.35 0.008 2.29 0.34-0,36 0.051 TS 0.61-0.77 0.050 7.05 0.62-0.81 0.063 9.97 0.55-0.70 0.013 4.78 0.24-0.30 0.012 4.61 0.24-0.29 0.020 8.56 0.21-0.26 0.031 3.81 0.77-0.86 0.040 5.14 0.68-0.87 0.070 10,23 0.58-0.75 0.087 6.35 1.17-1.50 0.102 7.04 1.20-1.60 0.129 10.33 1.00-1.33
4Five additional adult female paratopotypes were skinned in the field and are excluded from this table. bTibia length is the shank measured from the heel to the convex surface of the knee (with limb segments flexed at right angles),
roughly approximating length of the tibiofibula. ‘Greatest head width as measured between jaw articulations.
dHand length measured from proximal edge of large medial palmar tubercle to tip of longest (3rd) finger.
eDigit width measured near distal end of penultimate phalanx.
vocal slits, and pale throat). Length from snout to vent 22.2; tibia length between heel and outer surface of flexed knee 10.1; great- est width of body 8.3; head width between angles of jaws, and between outer edges up- per eyelids, 6.8, 6.1, respectively; approxi- mate width of interorbital area 2.7; head length from tip of snout to angle of jaw 6.0 (sagittal) or 7.1 (diagonal); snout length from tip to edge of eye (sagittal) 2.6; center of naris to anterior edge of eye 1.9; distance between centers of nares 2.9; eye length from anterior to posterior edge also 2.9; horizontal diameter of tympanum about 1.2 (concealed dorsally and posteriorly); corner of mouth to lower edge of tympanic ring 0.7; hand length from proximal edge of large medial palmar tubercle to tip of longest (third) finger 5.6;
width of disc of third finger (and width of penultimate phalanx below disc) 0.6 (0.4); width of 3rd finger at base 0.6; width of discs (and penultimate phalanges below discs) of third and fourth toes 0.9 (0.5) and 0.7 (0.5), respectively.
DESCRIPTION
There are 52 specimens in the type series, including 32 adult females, 14 adult males, and six unsexed juveniles. Five of the adult females were skinned in the field for subse- quent myological and osteological examina- tion; measurements and proportions of the other 47 specimens are summarized in table 1. Maturity of males was judged by presence of open vocal slits and usually by gray vocal
26 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
pouches (unpigmented in two small adults). All females were opened by lateral incision on the right side, and all were seen to have active ovaries and large or enlarging ovi- ducts; dissection of skinned carcasses re- vealed about 3—4 large (~ 2 mm diameter), pigmented ova in each ovary.
MorRPHOLOGY: Dorsum weakly granular in life, becoming markedly tuberculate on hind legs (fig. 16). Dorsal skin becoming nearly smooth in preservative; ventral skin very finely granular to smooth in preservative. Vent opens at upper level of thighs, without a pronounced dorsal flap above vent; no anal sheath; posterior thighs coarsely granular or tuberculate proximally, but no pronounced or symmetrically arranged tubercles in cloacal region.
Head little wider than long; greatest head width (between angles of jaws) 30-36% of SVL, averaging largest in juveniles (table 1). Snout sloping, rounded in profile, broadly rounded in dorsal and ventral view. Nares sit- uated near tip of snout and directed slightly posterolaterally; nares visible from front, barely or not visible from above or below; posterior rim of naris raised slightly (but lacking any tuberclelike prominence); pos- terodorsal portion of raised rim often bor- dered behind by a crescent-shaped, finely in- cised groove. Canthus rostralis rounded; lo- real region weakly concave, sloping slightly outward to lip. Interorbital region much wid- er than upper eyelid. Snout a little longer than eye length; center naris—edge eye/eye = 0.55—0.81, shortest in juveniles (table 1). No discrete postrictal tubercles. Tympanum rath- er inconspicuous, concealed dorsally and posteriorly by a diffuse supratympanic bulge.'' Tympanum width revealed by dis- section to be about one-half of eye length; tympanum positioned posteroventrally be- hind eye, with its lower edge close (0.5—0.8 mm in adults) to angle of jaws.
Hand (fig. 17) length in adults 24-30% of SVL, 68-87% of greatest head width. Rela- tive lengths of appressed fingers HI > IV ~ I ~ II, with slight individual variation; the first, second, and fourth fingers nearly equal when appressed, but with any one being
'! This bulge, which reflects the underlying muscle, is too diffuse to be called a supratympanic fold.
NO. 261
Colostethus undulatus, new species.
Right hand and left foot of female holotype (EBRG 3021). Scale line = 1 mm.
Fig. 17.
slightly longer than others in one specimen ori another stesso Sav or TV¥> ee = II). Discs of all fingers moderately expanded; third finger disc in adults 1.2—1.6 times wider than distal end of adjacent phalanx.
The third finger appears relatively slender in most females, slightly stouter in most males and perhaps a few females. In some males all the fingers appear stouter than in most females (fig. 18), although differences are subtle and not readily quantified.'?
Base of palm with large median metacar- pal tubercle, generally rounded; smaller el- liptical inner metacarpal tubercle on base of first finger; one or two low, rounded subar- ticular tubercles (one each on fingers 1, 2, two each on fingers 3, 4). No finger fringes (lateral keeling, especially on third finger, barely detectable under magnification). Faint trace of an outer metacarpal fold detectable
' A modified third finger characterizes some species of Colostethus s.1. and Epipedobates and may be ple- siomorphic in these genera (summary discussion in My- ers and Donnelly, 1997: 24—25). Although very well ex- pressed in some species, sexually dimorphic finger wid- ening is easily overlooked when barely developed (e.g., figs. 17, 18; also Myers et al., 1998, fig. 4), and the taxonomic distribution of this trait is therefore uncertain.
2001
Fig. 18.
on some specimens but not others, some- times bearing a minute tubercle on edge of palm between palmar tubercle and proximal subarticular tubercle of fourth finger. No ul- nar tubercles or fold.
A fleshy, glandular supracarpal pad atop wrist (fig. 18), little developed in females but swollen in males, which also have a more pronounced overhang or crease around an- terior margin of pad (supracarpal pad also associated with increased black pigmentation in males, see Color and Pattern; see also Re- marks).
Hind legs of moderate length, with heel of adpressed limb usually reaching eye (not reaching eye in some females, a little beyond eye in some males); tibia 41—49% of adult SVL, relatively shorter in juveniles than adults on average (table 1). Relative lengths of appressed toes IV > I] > V>II>I1; first toe short, reaching or failing to reach
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 27
Sexual dimorphism in hand and wrist of adult Colostethus undulatus, new species, as seen in dorsal view. Left: Right hand of female (EBRG 3042, 22 mm SVL), 9.7. Right: Right hand of
male (AMNH 159134, 21.5 mm SVL), showing stouter digits, swollen supracarpal pad atop wrist, and increased black pigmentation on forearm and hand, 10.0.
base of subarticular tubercle of second toe. Discs weakly to moderately expanded on first and fifth toes, with discs on toes H-IV distinctly expanded and larger than finger discs. Toes nearly webless, rudimentary basal webbing sometimes arguably discernible be- tween toes II and III, but distinct basal web- bing only between HI and IV (webbing for- mula III 3—4 or 4% IV); this web variably pigmented, sometimes clear distally. Toes not fringed, lateral keeling weak or very faint even on facing sides of toes III and IV.
A small to moderate-sized elliptical inner metatarsal tubercle and a round outer meta- tarsal tubercle are subequal in size (either may be larger than the other); rarely a weak indication of a third tubercle between the in- ner and outer ones (e.g., AMHN 159128, EBRG 3044). One to three nonprotuberant subarticular tubercles (one each on toes I, I, two each on II and V, three on IV); subar-
28 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
ticular tubercle near base of fourth toe very small, offset laterad. The faint keel along outer edge of the fifth toe is continuous with a weak outer metatarsal fold that terminates in a small but usually distinct tubercle distad from the outer metatarsal tubercle; no other supernumerary tubercles. Faint keel along free edge of first toe is continuous with a fold on distal third to distal half of tarsus; this usually weak tarsal fold becomes pronounced only at its proximal end, where it abruptly curves laterad and dorsad to terminate either as a prominent keel or a definite tubercle (fig. 17).
Teeth present on maxillary arch. Tongue longer than wide; free posteriorly, with rounded margin; no median lingual process. Vocal slits large, each extending from ante- rior edge of tongue nearly to angle of jaw.
COLOR AND PATTERN: In life (fig. 16), the dorsa were orangish brown with a usually prominent, wavy-edged, darker brown or grayish brown figure that remains conspicu- ous in preservative (fig. 19). Although no two markings are identical, there is constan- cy in general aspect and a tendency toward bilateral symmetry. The undulating figure ex- tends posteriorly from between the eyes to past midbody, with one constriction before the arms and one after. This marking is ob- scure in only a few specimens (most notably AMNH 159122, see fig. 19); one specimen (EBRG 3034) has the anterior part of the marking broken off as an interocular triangle (fig. 19).
A black face mask, continuous around snout, widens behind the eye and becomes a broad lateral stripe, narrowly confluent above thighs with black seat patch. The contrast be- tween black sides and brown dorsum is em- phasized by the suggestion of a pale dorso- lateral line (thin and ill-defined, not to be called a stripe), which was tan or pale gray- ish in life. There is some inconspicuous white flecking on the flank and into the groin (but no suggestion whatever of a pale oblique lateral stripe nor any clear development of a ventrolateral stripe). A pale bronze labial stripe is posteriorly continuous to the upper arm, often emphasized by black pigmenta- tion along the mouth edge.
Arms pale orange in life, without discrete crossbands but with tendency for a black
NO. 261
stripe along anterior (medial) side of upper arm. A diagonal black wrist marking of var- iable size, mostly small and confined to top of wrist in females, but in males larger and extending proximally and medially along lower arm, sometimes covering most of me- dial side of arm nearly to the black stripe on forearm. Upper sides of thighs and shanks brown—or shanks brown and _ thighs or- ange—with darker brown crossbands; pos- terior thigh surface either brown with a suf- fusion of orange or overall bright orange.
Throats pale to dark gray in most males (see comment on preserved specimens be- low), with an ill-defined paler chin spot hav- ing a greenish tinge. Throats yellow or gray- ish yellow in females, sometimes with an area of white (although a well-defined light gray blotch on chin and median throat was noted in a preserved specimen, EBRG 3038); venter yellow, brightened by underlying (sil- very?) peritoneum. Undersides of arms or- ange, undersides of hind legs brighter or- ange. Undersides of hands and feet blackish brown or black, this color extending proxi- mally along ventrolateral sides of lower arm and tarsus to elbow and heel.
Iris orangish bronze, with some dark suf- fusion or venation, and a clear bronze pupil- lary ring.
Juveniles look like the adults except that the smallest specimens (about 12 mm SVL) had the throat and venter pale gray in life (vs. yellow venters in adults) and the orange coloring on the undersides of limbs was paler and less bright.
In preservative (fig. 19), the dorsal ground color has faded to a duller brown or grayish brown. The axilla (not noted in life) is whit- ish, presumably colored like the venter in life. The undersides of the heads in males remain pale gray to blackish gray in preser- vative, usually with a vaguely paler chin spot; two small adult males have unpig- mented throats. Ventral surfaces otherwise mostly yellowish white, with a tinge of pale orange lingering on undersides of thighs. Un- derside of shank whitish in preservative, with a variable dusting of melanophores.
MyYOLOGY AND OSTEOLOGY
Myo.ocy: Basic information is provided by five adult female paratopotypes that were
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 29
Fig. 19. Colostethus undulatus, new species, in dorsal and ventral view, X1.1. Top row, females: Left to right: EBRG 3021 (holotype), AMNH 159122, EBRG 3025, AMNH 159133, EBRG 3034. Bottom row, males: EBRG 3026, AMNH 159131, 159132, EBRG 3031, 3046. Virtually the full range of variation in dorsal color pattern is shown; the vestigial and broken patterns in two of these frogs (2nd and 5th from left in top row) are unique in the type series of 52 specimens.
30 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
skinned in the field. Selected deep muscles were examined by dissection on three of these specimens prior to clearing and stain- ing. The muscle fasciae bear only scattered melanophores and the flesh appears white (vs. gray or black in some dendrobatids).
The massive superficial slip of the m. de- pressor mandibulae originates from the dor- sal fascia and overlaps the posterodorsal edge of the tympanum; a much greater part of the dorsal and posterior part of the tym- panum is concealed externally owing to the bulk of the muscle, which prevents close as- sociation with the skin. A concealed, deeper slip of the m. depressor mandibulae has a combined origin from the otic ramus of the squamosal and the posterior margin of the tympanic ring.
There is no m. adductor mandibulae ex- ternus superficialis. The mandibular branch of the trigeminal nerve (visible without dis- section on both left and right sides of all five carcasses) laterally crosses the deeper m. ad- ductor mandibulae posterior subexternus, which passes under the anterior part of the tympanic ring, apparently originating partly from the tympanic ring as well as from the zygomatic ramus of the squamosal.
On the thigh, the tendon of the deep m. semitendinosus distally passes through the loose, ligamentous attachments of the pos- terodorsal edge of the m. gracilis complex, to insert on the tibio-fibula adjacent (distad) to the insertion of the single gracilis tendon.
OsTEOLOGY: Three specimens skinned in the field were cleared and stained—one stained for bones only (AMNH 159141), and two double-stained for bones and cartilage (AMNH 159142, EBRG 3024). The speci- mens are variously damaged and the follow- ing brief description is a composite.
Skull approximately as wide as_ long. Frontoparietals well ossified, in medial con- tact except for anterior notch. Sphenethmoid large, in close contact with frontoparietals. Nasal bones of moderate size, widely sepa- rated from one another, and barely in contact with sphenethmoid at their posteromedian corners. Elongated maxillary process of na- sal extending posteroventrally past, and slightly overlapping, facial lobe of maxilla.
‘*Palatine’’ (neopalatine) bones absent. Ventrolateral process of sphenethmoid prom-
NO. 261
inent. Vomers toothless and widely separat- ed. Septomaxillae small. Sphenethmoid fused midventrally, ventrolaterally in contact with very large optic foramen in one double- stained specimen, narrowly separated by a rim of cartilage in the other. Only anterior ramus (cultriform process) of parasphenoid visible, with rest of parasphenoid seemingly fused with otoccipital by overlying bone that is anteriorly delimited by a distinct transverse suturelike edge, in front of posterior edge of optic foramen. Extending forward from this is the cultriform process, a stout bone, with sides nearly parallel except for slight diver- gence rostrad; cultriform process overlain by bone anteriorly, at its abrupt terminus in pos- terior concavity of sphenethmoid.
Alary processes of premaxillae projected dorsally (nearly vertical), with a lateral in- clination. Moderate-size maxillary and pre- maxillary teeth markedly recurved at tips, the posterior maxillary teeth small but retain- ing the recurvature; teeth possibly nonpedi- cellate,'* although a few breaks can be found near maxilla. Squamosal triradiate, but ante- rior zygomatic ramus very short; otic ramus somewhat longer than ventral ramus. Ante- rior ramus of pterygoid long and slender, posterior ramus short and robust, medial ra- mus very short. Retroarticular process of mandible long and prominent.
Hyoid plate (damaged) cartilaginous, wid- er than long; elongate posteromedial pro- cesses ossified except for distal cartilaginous tips.
Eight presacral vertebrae, procelous, with nonimbricate neural arches; no vertebral fu- sions. Transverse processes of vertebrae II— VIII subequal, anteriorly directed on II, tend- ing to be posteriorly directed on IV—VI. Sa- cral processes not dilated distally. Sacral— coccygeal articulation bicondylar.
Pectoral girdle firmisternal, with coracoids in medial contact; epicoracoid cartilage me- dially reduced and ossified. Coracoids much wider than long, greatly expanded along their medial contact. Clavicles long and very slen- der (threadlike), laterally fused with scapu-
'3 Myers et al. (1991: 11) noted that ‘‘The loss or sig- nificant obfuscation of the usual amphibian pedicellate condition warrants attention as a possible additional syn- apomorphy for the Dendrobatidae.”’
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF a TABLE 2
Measurements (in mm) of Colostethus undulatus Tadpoles?
Stage N Total Head—body Body Body Tail fin Oral disc
length length width depth depth width
25 13 19.0-32.6 HiPlae 4.9-9.2 3.8-7.8 4.3-9.1 2.0-3.2 26.92 + 3.82 10.79 + 1.41 7.22 41.21 5.634113 64341.36 2.73 +0.30
37 1 37.2 14.8 9.6 7.2 6.8 2.8
39 l 41.5 14.9 10.7 10.2 9.3 3.7
4 Values include means + | SD when NV > I. lae, medially not quite in contact with one TADPOLES
another. Omosternum an elongate, mostly bony structure, cartilaginous across its wide base, anteriorly narrowing, and distally car- tilaginous with a clublike terminus. Small sternum posteriorly bifurcated, partly ossi- fied.
Phalangeal formula of hands 2-2-3-3, of feet 2-2-3-4-3. Each terminal phalanx prom- inently T-shaped.
Nurse frogs carrying dorsal tadpoles were not seen. One lot of 15 free-living tadpoles (AMNH 159143, MBUCV 3048) was netted in water 7-45 cm deep at the edge of the forest-stream habitat of the type series, at places where there was little current (fig. 23, bottom). The lot includes 13 larvae in Gos- ner stage 25 and one each in stages 37 and 39 (measurements in table 2).
Fig. 20. tadpole below (AMNH 159143).
Colostethus undulatus, new species. Stage-37 tadpole above, mouthparts of a stage-25
32. BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
A a Da alicia dae
25 msec
0 0.5 1.0 1.5 Time in Seconds
Fig. 21. Vocalization of Colostethus undulatus, new species; 1.9 sec of a 90 sec recording of variably spaced note couplets. Top: Sound spectrogram graphed with narrow-band (59 Hz) filter, with frequency cursor set at 3600 Hz, and matching waveforms. Bottom: Wide-band (300 Hz) spectrogram of the same notes above and expanded waveform of the first note. The time between waveform cursors is 0.069 sec. Specimen (not caught) recorded 10:30 a.m., February 28, 1995, at air temperature of 19.8°C (AMNH herpetology reel 264:3).
The following description is based on the HABITUS AND PROPORTIONS: Viewed from single tadpole in stage 37 (fig. 20), followed above, the head and body form a somewhat by concluding comments on ontogenetic elongated ellipse rounded at the ends, with change. midbody width 66% of head—body length.
2001
1) 0.5
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 33
50 msec
1.0 1.5
Time in Seconds
Fig. 22. Vocalization of Colostethus undulatus, new species; a pair of pulsed notes from the same train of couplets as the unpulsed notes in figure 21. Wide-band (300 Hz) spectrogram, with frequency cursor at 3600 Hz. Expanded waveforms show both pusatile notes together (bottom) and the first (mid-
dle) and second (top) notes separately.
Head—body depressed (midbody depth = 74% of midbody width), somewhat flattened dorsally and ventrally. Eyes and nostrils dor- sal, directed dorsolaterally. Spiracle high on body (nearly midway between venter and dorsum), sinistral, a posterodorsally project- ing tube about 1.2 mm long; its opening 8 mm behind snout, at 54% of head—body length. Vent tube dextral to caudal fin.
Lateral line system with a supraorbital and an infraorbital branch originating near upper labium and meeting behind eye. A dorsolat- eral branch rising behind eye and apparently terminating near end of body. A _ lateral branch curving above spiracle and dropping posteroventrally.
Caudal musculature not robust, tapering gradually, terminating anterior to tail tip. Tail 60% of total length and relatively high, with a greatest depth of 23% of total length. Tail high for most of its length, rounded at end;
dorsal and caudal fins of approximately equal height.
PIGMENTATION IN PRESERVATIVE: Head- body dark brown above, turning pale below. Caudal musculature pale tan, fins transpar- ent; conspicuous dark brown blotching over caudal musculature and on dorsal fin, only small clusters of melanophores on ventral fin (but ventral fin conspicuously blotched in some other specimens, or both fins nearly de- void of pigment). Grayish white iridophores conspicuous ventrally and on lower sides of body, sparsely distributed elsewhere on body and tail.
MovutTH PARTS: Mouth ventral. Oral disc emarginate. Labial teeth in 2/3 rows: Ante- rior rows about equal in length, extending to near marginal papillae; second anterior row interrupted widely (in all specimens) above upper jaw sheath (row Al continuous except for a seemingly anomalous narrow median
34 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
break in the stage-37 larva but not in any other). First and second posterior rows sub- equal, slightly longer than third row. Upper jaw sheath a moderately wide arch, shallowly keratinized, with slender lateral processes; small blunt serrations. Lower jaw sheath V- shaped, shallowly keratinized; small blunt serrations like upper jaw. Marginal papillae bluntly pointed, single from lateral edges of anterior labium to posterior labium, where they become arranged in a weak double row.
ONTOGENETIC CHANGE: The smallest stage- 25 tadpole (7.7 mm head—body length, 19.0 mm total length) seems to have a single row of marginal papillae. All the other larvae have a weakly doubled row of papillae on the posterior labium, although a single row is re- tained laterally and anteriorly.
NATURAL HISTORY AND VOCALIZATION
Colostethus undulatus is a diurnal, terres- trial frog that was discovered in mossy gal- lery forest (fig. 23) at an elevation of about 1750 m, not far upstream from our camp at 1700 m (fig. 6, lower). The forest at the type locality is continuous with the mossy forest at base camp, where Colostethus did not oc- cur. The tendency of dendrobatids to occur in localized, apparently allopatric demes seems to be commonplace across all genera. Frequent if not constant diurnal calling by males makes it easy to locate demes such as this one. Excluding juveniles, there are 32 females and 14 males in the sample of 46 adults; this difference from a 1:1 sex ratio is statistically significant (x? = 7.0435, P = 0.00795), but whether it might reflect a dif- ference in the actual sex ratio or in activity patterns is unknown.
The advertisement call of Colostethus un- dulatus is a pair (couplet) of notes repeated at varying intervals, usually at a rate of <1 couplet/sec. A single note seems to be only rarely interspersed in a train of couplets; one 62-sec segment of tape contains 27 couplets and one single note. Narrow-band analysis of calls from two frogs shows the notes to be narrowly tuned at a nearly flat frequency of either 3600 Hz (fig. 21, top) or 3720 Hz. Waveforms in a series of calls from a single frog vary from nonmodulated (fig. 21) to dis-
NO. 261
tinctly pulsatile, with two pulses per note (fig. 22).
Note duration is 0.04—0.11 sec, with suf- ficient separation to hear each note of a cou- plet. These are heard as peep-peep, coequal in loudness or either note louder than the oth- er. The first note is longer than the second (occasionally equal), as shown by the follow- ing data for 27 couplets from a single frog, with a one-sided f-test for note 1 > note 2:
lst note 2nd note ‘K 0.073 sec 0.060 sec min. 0.05 0.04 max. 0.11 0.08 SD 0.015 0.011 F = 1.84768, P = 0.06197 t = 3.68791, P = 0.00029
The internote interval for 27 couplets from one frog is 0.08—0.30 sec (* = 0.157, SD = 0.055). Time between couplets varies from 0.76 to 4.19 sec in a sequence of 19 couplets (x = 1.699, SD = 0.754) from the same frog, but the recording of another frog includes several couplets that are more closely spaced (0.44—0.49 sec). The intercouplet interval therefore is too variable to characterize the call as a uniform “‘train”’ of couplets.
REMARKS
The affinities of Colostethus undulatus are not known. It probably differs from most other named and unnamed tepui species in lacking a median lingual process (Grant et al., 1997: 5; Myers and Donnelly, 1997: 24).'4 Somewhat similar dorsal markings ap- pear as pattern variants in the repertoire of a few other species, including such divergent species as C. lacrimosus Myers (1991: fig. 6) and C. mandelorum (Schmidt, 1932; Rivero,
‘4 We predict that most if not all of the web-footed highland species reviewed by La Marca (“1966”’ [1998] will be found to have a median lingual process. Among those species, for example, the lingual process is present in C. parkerae Meinhardt and Parmelee, Colostethus shrevei Rivero, and Colostethus tepuyensis La Marca. Photographs of the median processes of C. shrevei and C. tepuyensis (as Colostethus sp.) are given in Grant et al. (1997: fig. 3A, B), and the median process in C. shrevei is further illustrated in Myers and Donnelly (1997: fig. 17A, B), along with that of of another Ve- nezuelan species (C. tamacuarensis, fig. 17C, D).
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 2)
nll
Fig. 23. Interior views of mossy forest on east side of the upper Rio Corocoro, elevation about 1750 m (see fig. 6, bottom). Reflections from the surface of a small, sluggish stream can be seen in the lower photograph (right of center). This forest is the type locality of the frog Colostethus undulatus. (February 27, 1995)
36 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
“1982” [1984]: fig. 1; La Marca, 1993: fig. 2), but they are not constant enough to be diagnostic. Complex dorsal markings, when they occur, generally are quite variable in Colostethus and rarely provide reliably di- agnostic features. The remarkably consistent, sinuous dorsal figure of C. undulatus is therefore highly unusual in this regard.
Another unusual characteristic among den- drobatids is the concealment in C. undulatus of the parasphenoid. An overlay of bone ap- parently fuses the parasphenoid with the braincase, leaving the anteriorly projecting cultriform process as the only visible part of the parasphenoid.
Still another character of possible note is the subarticular tubercle on the base of the fourth toe—it seems unusually small in this species. Tubercle reduction on digits having two or more subarticular tubercles usually seems to involve the distal tubercle (as on the 4th finger of this species) rather than a basal one. This tubercle is offset laterad from the midline of the digit, a peculiarity that also pertains to some other species of the genus (e.g., see Myers and Donnelly, 1997: fig. 12), but these are casual observations and we can- not summarize its taxonomic distribution.
A character worthy of special attention is the fleshy, glandular swelling atop the wrist—the supracarpal pad (fig. 18), which seems either to be a novel structure in the Dendrobatidae or else one that has been overlooked in other species. Superficially, this swelling does not appear (at X50) to be much different from adjacent skin (except in being blacker), but a difference is obvious when the pad is bisected from the wrist to the forearm (e.g., on left wrist of EBRG 3041). In longitudinal section, the pad is seen to be several times thicker than the adjacent skin and to be comprised of large glands in the dermis. The glands, which are easily seen even at low (X6) magnification, are sur- rounded by melanophores, thickly so above and to the sides of the glands and thinly so below them. Below this is a conspicuous, bright layer of presumed iridophores. Finally, there is another heavy deposition of mela- nophores in the loose fascia on the innermost side of the thickened skin, which is loose above the wrist.
The supracarpal pad is superficially best
NO. 261
developed in males; it is weak or vestigial in females. The function of the supracarpal pad is not obvious, although it conceivably might play a role in the unusual dendrobatid trait of cephalic amplexus.'!? Only two types of integumentary glands—mucous and granu- lar—are known to occur in most amphibians, including the Dendrobatidae (Neuwirth et al., 1979). The glands comprising the pad will therefore deserve special study to determine if they are modified granular glands or some- thing else.
FAMILY LEPTODACTYLIDAE
Eleutherodactylus cantitans Myers and Donnelly Figures 24, 25
Eleutherodactylus cantitans Myers and Donnelly, 1996: 11-14, figs. 7, 8. Holotype EBRG 3005 (field no. CWM 19684), an adult male from summit of Cerro Yavi, 2150 m, Amazonas, Venezuela; collected February 23, 1995, AMNH-TERRAMAR Expedition.
MATERIAL: AMNH_ = 159144-159159, 159160 (4 froglets), EBRG 3055-3071, 3072 (4 froglets), from Cerro Yutajé, 1700-1750 m, Amazonas, Venezuela; collected February 26-28, 1995, AMNH-—TERRAMAR Expedition.
This series agrees well with our original description of the species, which was based on specimens from Cerro Yavi (Myers and Donnelly, 1996). However, as indicated be- low, the much larger sample from Cerro Yu- tajé gives better perspective on variation in size and proportions, and in color pattern, as well as in a few aspects of morphology. We also call attention to the ‘“‘axillary tubercle’’, a diagnostically useful character of possible phylogenetic value.
SUPPLEMENTAL DATA FROM THE YUTAJE SAMPLE
Females are larger than males (x female SVL/x male SVL = 1.47). Adults are 22—45
‘The dorsal sides of the fingers are pressed up against the female’s head in dendrobatid cephalic am- plexus, which would bring the supracarpal pads close to (or touching) the side of her chin. Cephalic amplexus, first demonstrated in Dendrobates (now Epipedobates) tricolor by Myers et al. (1978: fig. 6), is postulated as a dendrobatid synapomorphy that has been lost in Den- drobates + Phyllobates.
2001
mm SVL (20 males 22.5-30.6 mm, * = 28.29 + 0.45 mm; 6 females 36.9—45.2 mm, x = 41.47 + 1.15 mm). All males have vocal slits and at least traces of a nuptial pad on each thumb, and all adult females have large or enlarging ova and oviducts. Seven smaller juvenile or subadult females (23.1—31.6 mm SVL, * = 26.63 mm) have small ovarian ova and nonconvoluted oviducts. Eight unsexed froglets are 12.8-17.1 mm SVL (« = 14.99 mm).
Head wider than body, wider than long, or slightly longer than wide; head width 39— 43% of SVL in adults, with no evident sex- ual dimorphism (« = 43% in each sex), but relative head width slightly smaller in the ju- venile samples (39-40%, * = 39.3% SVL in 8 unsexed juveniles <18 mm; 40—41%, x = 40.3% SVL in 7 juvenile females). Eye-nos- tril distance 71-110% of eye length (20 males 76-90%, x = 83.4 + 0.011%; 6 fe- males 88-110%, x = 96.0 + 0.038%; juve- nile females 83-97%; small juveniles <18 mm SVL 71-85%). Supratympanic fold pre- sent, weakly developed, or absent. Posterior edge of tongue notched or entire.
Middorsal raphe present or absent. A large axillary tubercle (whitish in preservative), coded as absent only in four juveniles, is pre- sent in both males and females. The axillary tubercle varies from indistinct to very dis- tinct (fig. 24), sometimes being as large as the tympanum (tympanum/tubercle width ~ 1.0—2.8).'°
Tibia 48—60% of SVL (20 males 51-60%, x = 56.0 + 0.006%; 6 females 48-56%, x = 53.6 + 0.012%; juvenile females 56-59%, small unsexed juveniles <18 mm SVL 5l- 57%).
COLOR PATTERN: Fieldnotes and color transparencies show that the dorsal ground color of living Eleutherodactylus cantitans was various shades of brown, sometimes with a greenish suffusion (fig. 25). Although the dorsal pattern is polymorphic, most in- dividuals have indication of a dark interor- bital bar and a vaguely W- or X-shaped scap- ular marking (fig. 25, top right), followed
'©6 The axillary tubercle was not noticed in our original description of Eleutherodactylus cantitans. It is identi- fiable in all three of the AMNH paratopotypes from Cer- ro Yavi, but is distinct only in AMNH 143359@¢.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 37
Eleutherodactylus cantitans Myers and Donnelly, from Cerro Yutajé (EBRG 30712). Right side, approximately X<6.8.
posteriorly by other such dark markings as a few chrevrons or blotches and lumbar bar. Some are nearly uniformly dark, with only faint traces of darker markings, including two specimens with a distinct pale (whitish in preservative) vertebral line. Several spec- imens have scattered white or yellowish dots. One individual has a pale middorsum with slightly darker, wavy median and paraverte- bral stripes (fig. 25, bottom right).
Although not evident in figure 25, most individuals in life had a few yellowish oblique lines on the flanks, and yellowish to pale brown interspaces between the dark thigh bands. The rear surface of the thigh was gray with indefinite pale speckling and often with a rose suffusion. The ventral sur- faces were gray in adults, usually with a weak suffusion of rose under the hind limbs and in the groin and sometimes on the belly; juveniles <18 mm SVL had the ventral sur- faces grayish black with areas of bronze. Ex- cept for a reddish brown horizontal stripe, the iris was pale gray overall, or gray below
38 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fie: 25; row: Adult males, left to right, EBRG 3055 (31 mm SVL), EBRG 3056 (30 mm SVL). Bottom row: Adult females, left to right, AMNH 159144 (45 mm SVL), AMNH 159145 (37 mm SVL). (Not to same scale.)
the stripe and pale bluish gray or greenish gray above it.
In preservative, the dorsal ground color is light to dark brown. Ventral surfaces are light to medium brownish gray, sometimes vague- ly mottled. The seat patch varies from light to dark. A few females and several males have traces of a pale median line on the throat. Lip bars radiating from the eye to the lip vary from distinct to vague as in life. For a few years after preservation, the rose wash could seen in the groin, on the flanks and ventral surfaces of the hind limbs, and in the axilla.
NATURAL History: All specimens were captured at night, mainly in wet, mossy for- est. Males were calling frequently, both day and night. (The name cantitans means “‘sing- ing often’’, in allusion to the day-and-night
NO. 261
Eleutherodactylus cantitans Myers and Donnelly, from Cerro Yutajé, 1700-m camp. Top
calling behavior.) Eleutherodactylus_ canti- tans was much more common at 1700—1750 m on Cerro Yutajé than in the higher, cooler forest at 2150 m on Cerro Yavi, whereas the situation was reversed in the case of E. ya- viensis, which was sympatric with cantitans in both places.
REMARKS: At the type locality, all speci- mens of Eleutherodactylus cantitans were noticed to develop a temporary blue-green metallic sheen during formalin fixation. Al- though transient, this color was useful in the initial sorting of the several sympatric spe- cies of Eleutherodactylus on Cerro Yavi. We neglected to look for this unusual coloring when fixing the series of specimens on Cerro Yutajé.
Fieldnotes describe Eleutherodactylus cantitans as having the rear of the thigh
2001
black at the type locality, but gray on Cerro Yutajé. Otherwise, the series from Cerro Yu- tajé agrees well with color and pattern vari- ation shown by the type specimens from Cer- ro Yavi. Each sample includes a specimen of a morph having three irregularly wavy dark lines on a lighter middorsum (fig. 25, bottom right; Myers and Donnelly, 1996: fig. 7, up- per). The unusual axillary tubercle also oc- curs in both populations. Although we failed to get comparable sound recordings, we have no doubt that the Yavi and Yutajé popula- tions represent a single species.
Eleutherodactylus yaviensis Myers and Donnelly Figures 26-28
Eleutherodactylus yaviensis Myers and Donnelly, 1996: 17-22, figs. 13, 14. Holotype EBRG 3017 (field no. CWM 19677), an adult female from summit of Cerro Yavi, 2150 m, Amazon- as, Venezuela; collected February 22, 1995, AMNH-TERRAMAR Expedition.
?Eleutherodactylinae series b, Gorzula and Sefi- aris, 1999: 55 (part: MHNLS 11374 [not seen], from Cerro Corocoro; collected November 10, 1987, S. Gorzula).
MATERIAL: AMNH = 159161-159163, EBRG 3050-3051, from Cerro Yutajé, 1700 m, Amazonas, Venezuela; collected February 14 and February 26—28, 1995, AMNH-—TER- RAMAR Expedition.
In general morphology, these five speci- mens are very similar to the type series of Eleutherodactylus yaviensis from neighbor- ing Cerro Yavi. But E. yaviensis was de- scribed as lacking an ear, whereas a variably flimsy tympanum was found hidden under the skin in four of the five specimens from Cerro Yutajé. The apparent absence of a tym- panum in the fifth specimen gives us an ex- cuse for not describing another species of Eleutherodactylus. Assigning these speci- mens to E. yaviensis, however, does require altering two character states in the definition of that species, as follow (from Myers and Donnelly, 1996: 18): (char. 2) Tympanum ab- sent or, at least on Cerro Yutajé, often rep- resented by a vestigial structure concealed under the skin. (char. 3) Canthus rostralis slightly concave to nearly straight.
The following description of the Yutajé frogs mirrors our original description of
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 39
Eleutherodactylus yaviensis. Except for the tympana, differences are very minor. See Re- marks for discussion of the ear region.
DESCRIPTION OF THE YUTAJE SAMPLE
One male 20.8 mm SVL; four females 25.5—27.8 mm SVL (x = 26.63 mm). Head slightly longer than wide or as wide as long; head width 41% of SVL in the male, 40— 42% of SVL in females. Snout rounded (nearly pointed in AMNH 159163) in dorsal view, rounded in profile; eye-nostril distance 72% of eye length in the male, 79-83% in females; nostrils slightly protuberant, direct- ed dorsolaterally; canthus rostralis slightly concave or (in two specimens) straight, edge rounded; loreal region concave, sloping out- ward to lip. Upper eyelid with small low tu- bercles; upper eyelid width 93% of interor- bital distance in the male, 93—100% in fe- males. Tympanum feebly developed, con- cealed beneath skin in four individuals (fig. 28, top), but apparently absent in AMNH 159162. Low glandular ridge from posterior edge of eye toward arm. Several small, in- conspicuous postrictal tubercles. Choanae small, round, not concealed by palatal shelf of maxillary arch; vomerine odontophores triangular, posterior and median to choanae, bearing 2—4 teeth; odontophores as wide or slightly wider than choanae. Tongue slightly longer than wide with slight posterior notch, posterior half free. Vocal slits absent in male.
Dorsal skin granular with some inconspic- uous, small scattered warts, becoming smoother in preservative; a weak middorsal raphe in the male, absent in the four females; throat smooth, venter areolate; triangular area of glandular skin present on posteroventral thigh surfaces; discoidal fold anterior to groin; ulnar tubercles absent.
Relative lengths of appressed fingers III > IV > II > I; first finger not reaching edge of disc of finger I]; male with whitish nuptial pad on posterolateral side of thumb; fingers lacking definite fringes or keels. Finger discs broader than long, rounded apically, with subdigital pads wider than long; disc on thumb slightly expanded. Palmar tubercle large and cordiform in the male and in one female (19789), divided in the rest; when di- vided, elongate inner portion much larger
40 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
than small outer part; thenar tubercle oval, large; prominent subarticular tubercles round, conical; supernumerary palmar tuber- cles round, low or slightly protuberant, slightly smaller than subarticular tubercles.
Hind legs relatively long; heels overlap when held at right angles to the sagittal plane; tibia 53% of SVL in the male, 51-— 54% of SVL in females. Relative lengths of appressed toes IV > V > HI > Il > I; tip of toe V extending to distal edge of ultimate subarticular tubercle of IV, tip of toe HI to the distal edge of the penultimate subarticu- lar tubercle of IV. Toe discs broad, as broad or slightly broader than those on fingers, with wide subdigital pads. Toes lacking lateral fringes or keels, with basal webbing (I 3-3 II 2-3% Hl 3-4% IV 4%-3%V). Inner meta- tarsal tubercle large, elongate, about 1.8—2.6 times longer than small round outer metatar- sal tubercle; small supernumerary plantar tu- bercles round, low to somewhat protuberant; subarticular tubercles large and round on toes I and II, protuberant on II, IV, and V; in- conspicuous low calcar tubercles; tubercula- tion on outer side of tarsus inconspicuous or absent (more conspicuous in AMNH para- topotypes, but no tarsal fold or defined tarsal tubercle in either sample).
COLOR PATTERN: Three of the five speci- mens were photographed alive (fig. 26), and written descriptions were made after they were anesthetized (in chloretone solution). The dorsal ground color ranged from orang- ish brown (AMNH 159162) to yellowish brown. One photograph shows a faint, dull greenish tinge in one specimen (fig. 26, top), probably suggestive of the slight metachro- matic changes of which these (and many oth- er) frogs are capable. One specimen had a yellow upper lip (fig. 26, bottom). The throat was greenish gray in the male, and pale bronzy green (EBRG 3050) or bronzy white in the two females. The male had the venter and undersides of the limbs light brown, with bronze flecking on the chest; the two females had paler venters with extensive chest fleck- ing of the same color as the throat (pale bronzy green or bronzy white). The rear of the thigh was light brown in the male, or- ange-red in the two females. Unlike the male, the females also had orange-red in the groin and an orange suffusion underneath the hind
Fig. 26. Eleutherodactylus yaviensis Myers and Donnelly, from Cerro Yutajé, 1700-m camp (not to same scale). Top: Adult male, 21 mm SVL (AMNH 159161. Middle: Adult female, 26 mm SVL (EBRG 3050). Bottom: Adult female, 28 mm SVL (AMNH 159162).
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 41
Fig. 27. The sample of Eleutherodactylus yaviensis from Cerro Yutajé, X1.1. Left to right, top to bottom: AMNH 1591614, 1591622, EBRG 3050@, 30512, AMNH 1591638.
legs. The iris was pale gray in the male, or overall bronzy tan in one female, or bronzy green above and pale gray below in the other female (EBRG 3050), but all had inconspic- uous black venation and a reddish brown horizontal stripe.
In preservative (fig. 27), the dorsal ground color of all five frogs varies from brown in the male to light tan in the females. An ill- defined dark stripe below the canthus ros- tralis also is visible in photographs from life (fig. 26), as are vague dark brown dorsolat- eral lines in one specimen (AMNH 159162). Two females (EBRG 3051, AMNH 159163) each have a W-shaped dark mark on the scapula; one (AMNH 159163) of these also has a distinct interorbital bar, a small V- shaped mark anterior to the interorbital bar, and vague chevrons posterior to the scapular W-shaped mark. Crossbars on the limbs vary from vague to distinct. All the preserved specimens have a variably developed (vague to distinct) dark line extending from the rear edge of the eye toward the arm insertion. Granules on the anterior flank tend to be un-
pigmented and to form the pale centers of a variably distinct dark reticulum. The upper lip has a dusky appearance owing to the den- sity of melanophores, but one specimen has a pale, poorly defined stripe that was yellow in life (fig. 26, bottom); a few others have vague dark markings suggestive of incipient lip bars (fig. 26, middle). Most subarticular, palmar, and plantar tubercles are pigmented, although some are not.
NATURAL HISTORY
Three frogs (AMNH 159161-159162, EBRG 3050) were found at night during February 26—28 on low vegetation in wet, mossy forest near the 1700-m camp. Two others (AMNH 159163, EBRG 3051) were found in bromeliad tanks (Brocchinia) by en- tomologists, who were working out of this camp on February 14. Eleutherodactylus ya- viensis was less common at 1700 m on Cerro Yutajé than in the cooler forest at 2150 m on Cerro Yavi, whereas the reverse situation pertained to E. cantitans.
42 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
REMARKS
Four of the five frogs in the sample from Yutajé have a feebly developed ear that is concealed beneath the skin. This has caused us to worry that we may be confounded by having two sibling species in our small Yu- tajé sample, with the one earless specimen perhaps representing E. yaviensis and the others being undescribed. In describing E. yaviensis, we dissected only a few specimens to confirm absence of the tympanum, but in view of the Yutajé situation, we subsequently have examined all the AMNH paratopoty- pes!’ from Cerro Yavi in order to determine if there is variation in that sample. Compar- isons of the ear region between the Yavi and Yutajé samples follow.
CERRO YAVi: There are 13 paratopotypes of Eleutherodactylus yaviensis at the Amer- ican Museum; in all, the ear region on the left side was exposed. None have a tympa- num or any obvious vestige that we can iden- tify, but there is interesting variation in the disposition of the m. depressor mandibulae. Some specimens have it as shown in figure 28 (bottom), in which the anterior edge of this muscle superficially appears to originate from the epimysium of the adductor mandib- ulae muscle. But the fibers in the anterior part of the depressor mandibulae actually originate from a thin white tendinous edge that overlaps and lies close to the adductor mandibulae, and which may be anchored by one or more connective-tissue attachments to the adductor mandibulae. Other specimens (e.g., AMNH 143372, both sides dissected) show a distinct gap between the lower pos- terior part of the m. adductor mandibulae and the anterior tendinous edge of the m. depres- sor mandibulae. The anterodorsal part of the depressor mandibulae seems to originate partly from the zygomatic ramus of the squa- mosal bone. The tendinous anterior edge in- serts on the mandible.
CERRO YUTAJE: Figure 28 (top) shows the best development of the tympanum among the four specimens that have it. A portion of the m. depressor mandibulae originates from the posteroventral part of the tympanic ring; the posterodorsal part of the tympanum is
'7 The holotype and other paratopotypes had already been deposited in the EBRG collection in Venezuela.
NO. 261
Fig, 28: viensis. Top: Concealed tympanum (arrow) in specimen from Cerro Yutajé (EBRG 3051). Bot- tom: Absence of the ear in a paratopotype from Cerro Yavi (AMNH 143374). Abbreviations: A = m. adductor mandibulae; D = m. depressor man- dibulae.
Ear region of Eleutherodactylus ya-
tipped under the edge of the depressor man- dibulae. The tympana are flimsier in three other specimens, only two (AMNH 159161, EBRG 3050) of which have a slip of de- pressor mandibulae originating apparently from the tympanic ring.
A very faint vestige of the tympanum could be seen with careful lighting in AMNH 159163; it evidently is too weak a structure for anchoring part of the depressor mandib- ulae, which originates instead from the ad- ductor mandibulae. A tympanum was not de- tected in AMNH 159162, although there is an equivalent-sized cavity; as in the preced- ing specimen, part of the depressor mandib- ulae originates directly from the epimysium
2001
of the adductor mandibulae rather than from a tympanic ring. A tendinous edge along part of the anterior depressor mandibulae (as mentioned above) was noticed in a few spec- imens and might be present in all.
OTHER CONSIDERATIONS: Museum speci- mens of pattern-polymorphic, sympatric spe- cies of Eleutherodactylus can be exceedingly hard to differentiate, although differences in body and iris coloration in life often help to distinguish them in the field (e.g., see com- mentary on EF. avius and E. memorans in Myers and Donnelly, 1997: 58).
Excluding the ear character for the mo- ment, the diagnostic characters shared by ya- viensis and the Yutajé specimens include or- ange or red flash marks in groin, anterior and posterior thigh, and concealed part of shank (the flash color is suffused, without sharply defined edges). The flash color was absent in life in the one male from Yutajé, as well as in six small frogs from Cerro Yavi. The field catalogue mentions a bronzy aspect to the throat and bronze chest flecking for speci- mens in both populations. Iris color was var- iable in both samples, except for the constant feature of a reddish brown horizontal stripe through the pupil.
Color features that tend to disappear in preservative are at least consistent with our hypothesis that a single species is represent- ed. Color pattern itself is so variable, and the Yutajé sample so small, that any significant differences cannot be determined. The con- spicuous W-shaped scapular marking in two specimens from Yutajé (fig. 27) is not men- tioned in our description of variation in E. yaviensis, although the associated dark chev- rons were reported as occurring. However, one of the paratopotypes (AMNH 143371) dorsally has a large oblong marking with thin anterior extensions that precisely match the outer legs of the ““‘W”’ mark in the two Yutajé specimens. A dark reticulum on the anterior flank is conspicuous in a few specimens from each sample.
In summary, we cannot satisfactorily sep- arate the Yavi and Yutajé samples in any consistent way, nor can we convince our- selves that the one earless Yutajé specimen represents a taxon different from the ones with tympana. We imagine that there must
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 43
be a taxonomic purgatory!® for those who carelessly and thoughtlessly establish unnec- essary nominal species of Eleutherodactylus, and we therefore treat the specimens from the Yutajé—Corocoro massif as representing a partially differentiated population of Eleutherodactylus yaviensis. Insofar as we are aware, however, intrapopulational varia- tion in presence or absence of tympana has not been reported for any other species in the genus.
The one amphibian taken by Gorzula on Cerro Corocoro in 1987, at 2150 m, was in- cluded with several other unidentified mon- tane frogs under “‘Eleutherodactylinae series b” by Gorzula and Sefaris (1999: 55, 255). Among aspects of its color in life were “‘dor- sum a light-brown filigree over a golden cream base color . . . upper lip with a pale yellow line [and] brown subarticular tuber- cles.’ One of our specimens (fig. 26, middle) has a light brown reticulum over a lighter base, another one (fig. 26, bottom) has a yel- low lip stripe, and all tend to have pigmented subarticular tubercles, leading us to suggest that Gorzula’s frog is a specimen of Eleuth- erodactylus yaviensis as here defined.
Pseudopaludicola llanera Lynch Figures 29, 30
Pseudopaludicola llanera Lynch, 1989: 583. Ho- lotype ICNMHN 13576, from Puerto Gaitan, Depto. Meta, Colombia, collected June 8, 1984 by Olga V. Castafio and Maria C. Ardila.
MATERIAL: AMNH 1591652, 1591666, 159167 subad.d, EBRG 30522, 30536, 3054 subad.d, from rocky stream north of conical peak on south end of Cerro Coroco- ro, 1220 m (5°42'N, 66°10’W). Of the six specimens, two are adult females (enlarged oviducts and large pigmented ova) 16.2 and 16.5 mm SVL, two are adult males (vocal slits open) 13.4 and 13.8 mm SVL, and two are juvenile or subadult males (testes well
'8 Thomas Barbour (1928: 25) conveyed a similar sen- timent when he wrote that “It seems wicked and sinful in this day and generation to describe new species of Eleutherodactylus.”’ G. Kingsley Noble seemed not to enjoy receiving preserved specimens for identification, as suggested by his private generic synonym ‘“‘Eleuth- erodactylice”’ (Myers, 2000: 40). It is a large genus, with over 500 named species.
44 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
NO. 261
Fig. 29. Pseudopaludicola llanera Lynch, X1.6. Pattern polymorphism in sample from 1220 m elevation on Cerro Corocoro. Left to right: AMNH 1591652, EBRG 30522, AMNH 159166¢, EBRG
30536.
developed but vocal slits not opened) 12.1 and 12.6 mm SVL.
DESCRIPTION OF SAMPLE
The following account, which excludes characters common to the pusilla species group sensu Lynch, closely mirrors Lynch’s (1989) description of P. llanera, which he established on the basis of 23 specimens in the type series (some 50 additional speci- mens were referred to the species in his list of specimens examined). Disparities with Lynch’s description are shown in italics (see Remarks).
Snout acuminate or subacuminate in dor- sal view (fig. 29); interorbital space flat and narrow (narrowest part 71-92% of median width of upper eyelid, as follows: 71-75% in two subadult males, 77—83% in two adult males, and 86-92% in two adult females). Skin of dorsum bearing various-sized, low (‘‘flattened’’) warts with smoothly rounded surfaces, with degree of tuberculation pos- sibly reflecting sexual dimorphism as well as differences attendant with preservation. The adult female in life (fig. 30, top) was clearly tuberculate over the body, and also had a few small warts on the upper eyelid, although the top of the snout appears smooth in the pho- tograph. In preservative, however, the skin of the two females is somewhat smoother over the body than in life, with the eyelid warts inconspicuous and the top of the snout being quite smooth in the photographed specimen and very weakly tuberculate in the other one. In contrast, all four males in preservative
have the eyelids and the top of the snout markedly tuberculate; more subjectively, the males also seem more tuberculate on the body when compared with the preserved fe- males.
All specimens with two distinct antebra- chial tubercles, the distal one largest. Adult males with large vocal slits and unpigmented nuptial asperities on base of thumb.
An indistinct to distinct low rounded tu- bercle on upper edge of heel (but not pro- jecting and pointed as in most P. boliviana). Row of minute, inconspicuous tubercles along outer edge of tarsus (not evident in AMNH 159165). Outer metatarsal tubercle conical, small but nearly as long as the pro- tuberant inner metatarsal tubercle. Toe tips round, not expanded, lacking clearcut ter- minal groves, although an occasional toe shows a faint trace of a terminal groove un- der high magnification. Toes bearing distinct lateral fringes, but no basal webbing (dis- counting basal fusion of lateral fringes be- tween toes). Tibia 46-51% of snout—vent length, with indication of the expected on- togenetic change and sexual dimorphism (tibia/SVL = 0.46, 0.46 in adult males, 0.50, 0.51 in subadult males, and 0.49, 0.50 in adult females).
Color-pattern polymorphism present. In life, dark brown or grayish brown above, with or without a pale grayish vertebral line in each sex. In preservative (fig. 29), a wavy- edged brown dorsal band is conspicuous in the male and female with pale vertebral lines, and is faintly indicated in the other adult
2001
male; the other females and the subadult males are overall dark without a definite pat- tern except for crossbars present on the limbs of all specimens. Throat and venter white and undersides of limbs light gray in life; throat, chest, and anterior venter sparsely to heavily flecked with dark pigment, with sparse to moderate flecking on undersides of thighs. Rear of thighs dark brown below crossbars, but this pigment not forming *‘a dark line along lower edge of posterior sur- faces of thighs’’; several specimens with the dark pigment forming an ill-defined brown stripe along anterior side of thigh. The iris was grayish with brown suffusion. The base of the tongue is very weakly “‘pigmented”’ (see Lobo, 1995, character 15) with a few to several melanophores, except in EBRG 3052, which lacks melanophores.
NATURAL HISTORY
These miniature frogs were taken by day from sparse aquatic vegetation in shallow water along the edge of a drying, boulder- strewn stream (fig. 5, bottom). Some tiny eggs (not preserved) were found in a plastic bag containing only AMNH 159165, but it was not clear whether the eggs had been picked up unnoticed with the frog or laid in the bag overnight. The stream is near the base of the conical peak seen in several pho- tographs (e.g., fig. 5, top). This part of the Yutaje—Corocoro massif is well drained and seasonally arid.
REMARKS AND COMPARISONS
Although the specimens were collected on a relatively low and arid part of the Yutajé— Corocoro massif, the occurrence of Pseudo- paludicola at an elevation of 1220 m was nonetheless unexpected. The few differences between Lynch’s description of Pseudopalu- dicola llanera and the present specimens are assumed not to have taxonomic significance, but to merely extend the range of variation in P. llanera. Concerning Lynch’s original concept of this species as being monomor- phic in color pattern, he and colleagues (Ruiz-Carranza et al., 1996: fig. 83) subse- quently published a photograph of a living female specimen that has a pale vertebral line like the one in figure 30 (top), and Gorzula
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 45
and Senaris (1999: 81) mentioned the exis- tence of three basic color morphs in Vene- zuela.
The Cerro Corocoro sample described above was compared with a series of Pseu- dopaludicola (AMNH 100614—100633) col- lected by Myers and Daly on February 26— 27, 1978, in the lowlands (100 m) at Santa Barbara, junction of the Rio Ventuari with the Rio Orinoco—some 200 km to the south- southwest. Fieldnotes for this series read that the frogs were
Active by day (disappearing at night) on muddy bank of a drying cafio near its entrance into the Orinoco. Very much like cricket frogs (Acris) in their behavior and habitat and appearance: i.e., small diurnal frogs, active on muddy banks, showing pronounced pattern polymor- phism, and hopping about like crickets when pursued.
The 20 specimens in this series seem to represent a composite of two sibling species, which were not differentiated in the field and which may differ most noticeably in size. Two of the specimens are assigned to Pseu- dopaludicola, species inquirenda, and the others represent P. boliviana Parker, sensu Lynch, thus providing an interesting instance of microsympatry in the genus despite the problems of identification.
Pseudopaludicola boliviana: As currently recognized, this species ranges from Bolivia, Paraguay, and Brazil north to Guyana, south- ern Venezuela, and eastern Colombia (Lynch, 1989; Lobo, 1994)—a distribution suggest- ing the likelihood of boliviana being a com- posite species in need of critical review, par- ticularly when bioacoustical and additional data become available (Lynch [1989: 581] mentioned the problem of distinguishing pre- served specimens “‘from one another, in spite of the vast geographic distances between lo- calities’’). The present Venezuelan sample (AMNH 100615, 100617—100633) reveals variation in the diagnostically useful antebra- chial tubercles and in the conical heel tuber- cle (see below).
Pseudopaludicola boliviana is a _ small frog. There are two presumed adult males (vocal slits open) 11.7 and 12.1 mm SVL, and two slightly larger males of 12.2 and 12.9 mm SVL that lack vocal slits. Four ju-
46 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 30. Pseudopaludicola (not to same scale). Top: Pseu- dopaludicola llanera Lynch. Adult female
Southern Venezuelan species of
NO. 261
venile females (with threadlike oviducts) are 9.4-11.1 mm SVL. The remaining 10 spec- imens are sexually inactive females 12.0— 13.7 mm SVL (* = 12.84 mm); these have convoluted albeit small oviducts and lack en- larged ova. The interorbital space is relative- ly wide—0.91-1.44% (* = 1.10%, N = 18) of the width of an upper eyelid.'? The tibia is 50-56% of SVL in the four males and 51— 55% in 10 adult or subadult females (54— 57% in the four juvenile females). In life, frogs are mostly gray above with vague dark- er gray markings; a male and a female each had a conspicuously contrasting vertebral line of pale tan, and another female shows a faint hairline in preservative. Four females (AMNH 100630—100633, which include 3 of the 4 juveniles) had the distal two-thirds of the head and body uniformly reddish brown, with this color being sharply demarcated at its posterior margin; in preservative, these specimens also have faint indication of a pale vertebral hairline and also faint indication of dark smudging on the anterior dorsum. Ven- tral surfaces were white, with some variable but mostly inconspicuous dark speckling.
'°Tn preserved specimens of these tiny frogs, there often is no clear demarcation to the medial side of the upper eyelid. Therefore, the eyelid and the interorbital distance are subject to significant measuring error, which is why statistics are not separately given for males and females in the small sample at hand. The character Eye- lid/IOD nonetheless is of some use in species compari- sons (see Lynch, 1989: table 1).
<
(AMNH 159165, 16.5 mm SVL) from 1220 m elevation on Cerro Corocoro. Middle: Pseudo- paludicola boliviana Parker. Adult or subadult fe- male (AMNH 100615, 13.7 mm SVL) from Santa Barbara, 100 m. Bottom: Pseudopaludicola, spe- cies inquirenda. Juvenile female (AMNH 100614, 14.0 mm SVL) also from Santa Barbara. These photographs help to visualize certain diagnostic characters. Note the narrow tips of the toes in P. llanera vs. the weakly expanded toe tips in the other two. P. boliviana usually has a conical tu- bercle on each heel (see left side; right side shows a bit of debris), and the interorbital space is rel- atively wide. The possibly unnamed species in the lower photograph has expanded toe tips but lacks a heel tubercle, and it seemingly is a larger spe- cies than P. llanera or P. boliviana (see text).
2001
Among the suite of diagnostic characters that help identify Pseudopaludicola bolivi- ana are a conical heel tubercle and a rela- tively wide interorbital space, both of which are visualized by reference to figure 30 (mid- dle). But the interorbital space is unfortu- nately hard to quantify, and the heel tubercle is evidently subject to some variation. The heel tubercle was coded as present on both heels in 14 of the 18 specimens, as present on one heel only in three specimens, and as absent in one juvenile female (AMNH 100623). The tubercle seems rather fragile, and its occasional absence or blunted ap- pearance may be due to damage or abrasion. For example, the frog shown in figure 30 (middle) is seen to have a noticeably pointed tubercle on the left heel, but this tubercle seems smaller and is blunted in the preserved specimen; the tubercle was coded as absent on the right heel of this specimen, although an indication of it is arguably present at high magnification (the photograph shows what seems to be a bit of debris on the right heel).
About half the specimens have a single an- tebrachial tubercle, as is supposedly charac- teristic of P. boliviana, but many also have a slightly elevated posterior tubercle; there may also be indication of a few smaller ac- cessory tubercles in line with the larger ones (e.g., AMNH 10062436). In any case, the an- tebrachial tubercles are less elevated than in the highland frogs assigned above to P. Ila- nera. Pseudopaludicola boliviana also is said to be characterized by basal foot webbing, but it is not evident in this sample; basal fu- sion of toe fringing makes determination of basal webbing especially subjective in spe- cies such as P. boliviana and P. llanera (see Lynch, 1989: figs. 2B, 2D); toe fringing in the present sample seems less developed than in P. llanera above. All specimens have the toe discs slightly expanded in contrast to P. llanera.
Pseudopaludicola, species inquirenda: There are two specimens (AMNH 1006148, 100616¢) that were not distinguished from P. boliviana in the field. They were both gray with blackish markings and a tan or pale brown vertebral line (fig. 30, bottom). The interorbital space is about 85% (¢) and 92% (2) of upper eyelid width, and the tibia is 47% (3) and 53% (2) of SVL. There are
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 47
two weak antebrachial tubercles, but no con- ical heel tubercles.
The female, at 14 mm SVL, has threadlike oviducts and is therefore judged to be a ju- venile, although it is slightly larger than the largest females in the sympatric sample of P. boliviana. The male, at 15.9 mm, is an adult with well-developed vocal slits and black testes; it is about 19-27% larger than the four male boliviana. This specimen also is 13-16% larger than the two adult male J/la- nera from Cerro Corocoro, and direct com- parison makes it seem unlikely that a single species is involved (the juvenile female is uninformative, being smaller than the adult female /lanera).
Despite the size difference, we would be inclined to assume interpopulational varia- tion and assign these specimens to Pseudo- paludicola llanera. However, in marked con- trast to llanera, the toe tips are slightly but definitely expanded to about the same degree as the sympatric specimens of boliviana, and, similarly, these two specimens and sympatric boliviana have the distal antebrachial tuber- cle somewhat less pronounced than in Ila- nera. Thus, we suspect that these two spec- imens represent a third species of Pseudo- paludicola in southern Venezuela.
Specimens from the same locality (Santa Barbara) were referred to Pseudopaludicoa llanera by Gorzula and Sefiaris (1999: 81), leaving open the possibility that the postu- lated three southern Venezuelan species all occur in sympatry. But it seems conceivable that at least some of their specimens may represent the presumably undescribed sibling species as well as P. boliviana. Specimens collected by Gorzula at Santa Barbara might be from a different microhabitat than the mixed sample of P. boliviana and species in- quirenda reported here, but his material was obtained in the same week as those speci- mens, and on the same expedition.”°
20 The 1978 CODESUR-—American Museum Expedi- tion to Cerro Yapacana, in which Gorzula participated, used Santa Barbara as a starting place for boat travel. Dr. Otto Huber served as expedition leader and coordi- nated the unpublished expedition report (see Myers, 2000: 157-158).
48 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
LIZARDS FAMILY TEIIDAE
The diverse genera of “‘microteiids’’ are sometimes placed in a family (Gymnophthal- midae) separate from their larger relatives (‘“‘macrotetids’’), but the reasons for so doing are either based on undocumented characters, incomplete data, or on authoritarianism. In critically examining a claim by Presch (1983) that macroteiids and microteiids are not related, Harris (1985: 564) presented very persuasive evidence in support of their ‘“‘exclusive common ancestry’. Harris fur- ther concluded that available evidence does not confirm that the microteiids themselves comprise a single natural (monophyletic) group. This level of knowledge has not changed since 1985.
The matter of authoritarianism comes in with the impressive work of Estes et al. (1988), which was influenced (see p. 142 of their Character Analysis) by the flawed work of Presch (see Harris, 1985). Even though Estes et al. confirmed teiid monophyly, they nonetheless, according to Myers and Don- nelly (1996: 21-22),
continued the separate family arrangement for no obvious reason except that one of them (Es- tes) also had earlier recognized the Gymno- phthalmidae—based on synapomorphies that ‘““were for the most part variable characters”’ (Estes et al.: 217). The newer lists of putative synapomorphies (Estes et al.: 215-217) for the macroteiid and microteiid ‘families’? are based on an unknown (unstated) number of genera and species and their extrapolated use requires perhaps more faith than should be necessary.
The following additional remarks are in response to a reviewer’s suggestion that we should perhaps follow the authority of Estes et al. in the matter of teiid classification, in- asmuch as they provided supporting syna- pomorphies.
However, except for one unequivocal character (anterior nasal scales separated by one or two frontonasals) dating from Bou- lenger, one cannot usually tell how many mi- croteiid genera or species have been checked for the synapomorphies listed by Estes et al. (1988: 217)—there is no summary of evi- dence. Furthermore, among the 12 putative synapomorphies given by Estes et al. for mi-
NO. 261
croteiids, there are other problems involving either accuracy or pragmatism, as in the fol- lowing examples.
The presence of frontal tabs projecting posteriorly over the parietal bone was given by Estes et al. (1988: 144, 217) as a defining synapomorphy of microteiids (with indepen- dent acquisition only by some chamaeleon- tids)—overlooking the observation by Harris (1985: 564) that frontal tabs occur also in at least one macroteiid genus and that the con- dition may simply be easier to see in small teiids. Another purported synapomorphy giv- en by Estes et al. is “descending processes of frontals in contact below olfactory tracts’’, which condition, however, is displayed only in “‘some’”’ gymnophthlamids (Estes et al., 1988: 144; see also Harris, 1994: 229).
The common derived squamate condition of “‘second epibranchial lost’? was listed as a microteiid synapomorphy by Estes et al. (1988: 164, 217) on the authority of Camp (1923), MacLean (1974), and Presch (1980). However, Camp (1923: 339) only said that the ‘‘second epibranchial is apparently absent and the second ceratobranchial is still pre- sent’? in a specimen of the microteiid Bachia, whereas neither MacLean nor Presch explic- itly stated that the second epibranchial was lost in microteiids. MacLean (1974: 191-— 192), in fact, wrote that ““Many Gymno- phthalminae have disjunct epi-elements, one portion attached to the cerato-element and the other to the braincase.’’ Presch (1980) dealt only with the loss in some microtetids of the second ceratobranchial (note reversal of labels in his fig. 5).
Two character states (“‘frontals fuse in em- bryonic or early postembryonic ontogeny’’, and “‘parietal foramen lost’’) in the synapo- morphy list in Estes et al. (1988: 217) are, according to the text (pp. 143, 148), indistin- guishable from the usual conditions in ma- croteiids (the first character also was given [p. 215] as a synapomorphy for macrotetids based on interpretation of fossil evidence).
The apparently derived character of “‘cal- cified spines embedded in hemipenial flounc- es’’ (Estes et al., 1988: 142) describes what seems to be the usual condition in micro- teiids. However, whether this is synapo- morphic for all microteiids or only a large subset is very much open to question. In the
2001
case of in-group variation, Estes et al. (1988: 142) tended to interpret “‘the derived state as plesiomorphic for that taxon.”’ In this case, that assessment would have agreed with Presch’s (1983: 192) conclusion that absence of ‘“‘definite folds and teeth [hemipenial spi- nules]’’ is a secondary loss in microteiids, based on his earlier descriptions of unorna- mented hemipenes in Bachia and Anadia— an observation perhaps conceivable for the first genus but in error for the last one.?! Cal- careous spinules, however, are absent in the hemipenes of Alopoglossus and Ptychoglos- sus, Which “‘possibly comprise a sister group to all other microteiids”’ (Harris, 1994: 229). The new genus Adercosaurus, described be- low, also lacks series of comblike calcareous spinules, and it may be that the absence of such spinules represents the plesiomorphic condition for microteiids.
As suggested above, it remains to be de- termined whether any of the purported mi- croteiid synapomorphies (other than separat- ed nasal scales) can be used to define any- thing other than various subsets of small te- lids. The division between larger and smaller—macroteiids and microteiids—is a convenience for taxonomic discourse without the burden of formal nomenclature. MacLean (1974) separated teiids into the new subfam- ilies Teiinae and Gymnophthalminae; his ac- tion recognized the monophyly of the Teiidae and was not an unreasonable hypothesis for its time. Estes (1983: 74-75, 97) and Presch (1983) independently elevated MacLean’s subfamilies to family rank, with a major dif- ference—Estes considered them as _ sister
71 Presch (1978) examined and described everted mi- croteiid hemipenes without realizing that some had been only partially everted in the field. Thus, an organ of Anadia metallica (= A. ocellata) was described and il- lustrated as a smooth bulbous structure, with a dimpled tip (almost always an indication of incomplete eversion). The hemipenis of this species is in fact bilobate with comblike rows of calcareous spinules (observation based on everted hemipenes of AMNH 114306 and 129779 from western Panama).
Also described as smooth were the illustrated organs of two species of Bachia (a genus characterized by loss of the external ear and trends toward loss or reduction of various head scales and limb elements). However, these hemipenes also may be incompletely inflated, and they need reexamination in order to confirm degree of eversion and absence of spinules.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 49
groups, whereas Presch considered them to be unrelated to one another. Presch’s reason- ing was discredited by Harris (1985), who pointed out errors of fact and provided con- vincing proof of monophyly for the two groups. The argument for monophyly was subsequently supported by Estes et al. (1988), who also discounted (on p. 217) the gymnophthalmid synapomorphies earlier given by Estes (1983) as being “‘for the most part variable characters’’. Nonetheless, Estes et al. continued recognizing macro- and mi- croteiids as separate families on the basis of other purported synapomorphies, which, upon examination, start to disintegrate into those objectionable “‘variable characters’’.
Macroteiids and microteiids may indeed differ in some fundamental way yet to be elucidated, and it is at least conceivable that the single character of separated nasals might be corroborated in a way that will demon- strate the monophyly of all microteiids. But in the face of inadequately presented data, insufficient character analysis, and uncorrob- orated hypotheses, we join Harris (1985) and Cole et al. (1990: 2) in preferring nomencla- tural caution when change is neither man- datory nor purposeful.
Venezuela has a rich teiid fauna, with 45 species disposed in 18 genera as of 1997, which compares with 42 species in 21 genera in the much greater area of Brazilian Ama- zonia (La Marca, 1997; Avila-Pires, 1995). In each region, microteiids comprise about three-quarters of the teiid species diversity.
One of two species of microteiids from the Yutajé—Corocoro massif does not fit in any teiid genus known from the Guayana region, nor can we place it in any of the extralimital genera. Accordingly, the following generic name is proposed as a hypothesis of distinc- tiveness of this newly discovered lizard (figs. SUIS):
Adercosaurus, new genus
TYPE SPECIES: Adercosaurus vixadnexus, n. sp.
ETYMOLOGY: From the Greek aderkes (something unexpected or unseen) + sau- ros—an unexpected lizard. Gender mascu- line.
CONTENT: Monotypic.
50 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
Bien 31, mm SVL).
DEFINITION AND DIAGNOSIS: Small lizards, maximum size probably less than 80 mm SVL, with tail probably less than twice body length. Tongue anteriorly and posteriorly bearing oblique, anteriorly converging pli- cae, with intervening midsection of imbricate scalelike papillae; three pairs of nonswollen chrevron-shaped infralingual plicae. Ptery- goid teeth present. Phalangeal formulae 2-3- 4-5-3 for hands, 2-3-4-5-4 for feet. Hemi- penis lacking spines or spinules, symmetri- cally bifurcate, with thickened, lobate apical discs; asulcate side of hemipenis with encir- cling nude ridges. Head scales smooth. Nasal scales separated by rostral and undivided frontonasal. Loreal separated from labials by frenocular. Supraoculars separated from eye-
Adercosaurus vixadnexus, new species. Views of the holotype in life (EBRG 3126¢, 55
lids by a complete superciliary series; ante- rior superciliary large, not expanded dorsal- ly. Translucent palpebral disc composed of about six vertical panes. Frontoparietals me- dially in point contact. Interparietal longer than parietals, their common sutures forming a jagged line across rear of head. Tympanum slightly recessed, lightly pigmented. Single postmental scale followed by three pairs of genials in contact with labials. Anterior gular crease, incomplete guttural fold, and collar fold all conspicuous. Paramedian gulars en- larged, in short double row. Dorsal and lat- eral scales elongate with parallel sides, hex- agonal, in transverse rows only; dorsals sharply keeled, strongly mucronate, laterals becoming less so. Lateral fold absent, a gra-
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF Sil
Fig. 32. Adercosaurus vixadnexus, new species. Dorsal, ventral, and radiographic views of the pre- served holotype (EBRG 3126<¢), 1.6. Approximately the distal third of the tail is regenerated, as shown by the cartilaginous rod in the radiograph, although there is no clear external evidence of this.
dation between lateral and ventral scales; me- dian ventrals smooth, rectangular, gently rounded posteriorly, subimbricate, forming both transverse and longitudinal rows. Pre- anal scales in two rows. Femoral pores and preanal pores on same line. Limbs pentadac- tyl, all digits clawed; subdigital lamellae di- vided; base of pollex with enlarged thenar scales having produced median keels. Caudal scales similar to body scales, in uninterrupted annuli, no paramedian series of supracaudals along a vertebral suture.
In having oblique plicae on the anterior as well as the posterior part of the tongue, Ad- ercosaurus seems to stand apart from all other microteiids (including Arthrosaura) except Alopoglossus, Ecpleopus, Ptychoglossus, and
Riolama, but it stands with Ecpleopus and Riolama in having the lingual plicae inter- rupted by a midsection of scalelike papillae. Among these genera, Adercosaurus most closely resembles Arthrosaura and some Ptychoglossus in habitus and body scalation, but those two genera and Alopoglossus are distinctive in having the posterior margins of the parietals and interparietal forming a more- or-less straight line across the rear of the head (interparietal posteriorly projecting, forming a jagged or irregular line in Adercosaurus and Riolama, a gently rounded line in Ecpleopus). Ecpleopus differs from all in having hexago- nal ventrals that do not form longitudinal rows, the ventrals being being arranged in transverse rows only.
52 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
The endemic tepui genus Riolama differs from Adercosaurus in lacking a claw on the first finger,’ in having the subdigital lamellae mostly undivided, and in having smooth lat- eral scales that are smaller than the keeled dorsals.
The primarily Andean genus Anadia also reaches the tepuis, but it differs most con- spicuously from other tepui microteiids in having a tight covering of nonmucronate (and usually smooth) dorsal scales that are juxtaposed or only subimbricate, as well in having a very long tail and a relatively at- tenuated, flat-topped snout that give it a dis- tinctive aspect. Anadia evidently differs from Adercosaurus in hemipenial morphology (comblike rows of spinules in Anadia) and in the dorsal surface of the tongue (scalelike papillae anteriorly), but summary data for Anadia have not been published. Anadia, however, has been characterized as having (in 9 species) 6—10 swollen infralingual pli- cae (Harris, 1985), which differs from the three thin (nonswollen) infralingual plicae in Adercosaurus.
See below for additional comparisons (un- der Diagnosis and Remarks in species ac- count).
DISTRIBUTION: Known only from 1700 m on the Yutajé—Corocoro massif.
Adercosaurus vixadnexus, new species Figures 31—36, 37 (top), 38, 39
HOLOTYPE: EBRG 3126 (field no. CWM 19809), an adult male from wet gallery forest on Cerro Yutajé, 1700 m (5°46'N, 66°08’W), Amazonas, Venezuela; collected February 25, 1995, AMNH-TERRAMAR Expedition. See locality 1 on map (fig. 1).
ETYMOLOGY: The species name is com- pounded from the Latin adverb vix (barely, hardly) + the passive past participle adnexus (tied together or joined, connected), calling
. Both Boulenger (1900) and Uzzell (1973: 54) stated that the holotype of Prionodactylus leucostictus—type species of Riolama Uzzell—has all digits clawed. But E. N. Arnold kindly examined the specimen at our re- quest and assured us that the first finger is clawless, as in a specimen of Riolama from Cerro Duida (Myers and Donnelly, 1996: 22n). D. M. Harris (personal commun.) and R. W. McDiarmid (personal commun.), who have studied the holotype and other specimens, consider a clawless first finger to be diagnostic of Riolama.
NO. 261
attention to the possibly diagnostic point contact of the frontoparietal plates (fig. 34).
DIAGNOSIS: The generic definition and di- agnosis are thought to differentiate Aderco- saurus vixadnexus from all other teiids. The point contact of the frontoparietals, if a con- sistent character, will alone separate A. vix- adnexus from other teiids with few excep- tions (see Remarks).
On the Guayanan tepuis, A. vixadnexus is most likely to be confused with the upland species of Arthrosaura, which have brown dorsa, similar body scalation, and sometimes dark ventral markings; also the third pair of genials are in lateral contact with the infral- abials in both genera. However, as indicated above, Arthrosaura is immediately distin- guished from other tepui microteiids in hav- ing the posterior margins of the parietals and interparietal forming a nearly straight line across the rear of the head, and Adercosaurus further differs significantly from Arthrosaura in tongue morphology and in having a linear series of curved, nude tissue ridges across the asulcate side of the hemipenis (rather than oblique folds with comblike rows of miner- alized spinules).
The general appearance of the black-out- lined, white-centered ventral markings of the holotype of Adercosaurus vixadnexus, includ- ing the mottled throat (figs. 31, 32, 35), prob- ably is distinctive among teiids generally and is quite different from the several tepui lizards with dark ventral surfaces. Arthrosaura tyleri tends to have dark-centered ventral scales and lacks the throat mottling (Donnelly et al., 1992: fig. 4). Male Prionodactylus goeleti may have extensive black on the throat and venter (Myers and Donnelly, 1996: fig. 18), and the male holotype of P. nigroventris was “shiny jet black’’ over all ventral surfaces (Gorzula and Sefaris, 1999: 137), but these Prionodactylus differ noticeably from Ader- cosaurus in having small lateral scales and a vivid pale labial line extending posteriorly to the shoulder. The tepui endemic Riolama leu- costicta differs in being overall black with small, vividly defined yellow spots and yel- low mandibular stripes (Gorzula and Sefaris, 1999: color photo 91), and Riolama (nclud- ing unnamed species) also is immediately dis- tinguished from Adercosaurus by lacking a claw on the first finger.
2001
DESCRIPTION OF HOLOTYPE
The undissected male holotype is the only known specimen. It is adult judged by the well-developed hemipenes. Snout—vent length 55 mm; tail length 79 mm (regener- ated); head length (obliquely from tip of snout to edge of ear) 10.6 mm; greatest head width 7.7 mm; greatest head depth 5.5 mm; snout—axilla length 20.5 mm; length of neck (posterior edge of ear to forearm) 8 mm; trunk length (axilla—groin) 28 mm; length of forelimb and of hindlimb (from axilla or groin to tip of longest digit) 11 and 15 mm, respectively.
Head length 19% of SVL, 1.4 times longer than wide, and 1.4 times wider than high; head wider than neck. Neck long, 75% of head length, 29% of trunk length. Snout—ax- illa length 73% of trunk length, 37% of SVL. Body wider than deep. Tail externally ap- pears to be original, but a radiograph (fig. 32) shows that distal third is regenerated (i.e., caudal vertebrae replaced by cartilaginous rod); regenerated tail 1.5 times SVL; tail nearly cylindrical, with barely perceptible lateral compression and slight ventral flatten- ing. Limbs pentadactyl, all digits clawed (fourth digit lost from left hand). Forelimbs 21% of SVL, 39% of trunk length; hind legs 28% of SVL, 54% of trunk; adpressed limbs not overlapping, widely separated by five or six lateral scales.
Tongue (fig. 33) lanceolate, anterior half gray. Most of anterior half behind fork with oblique, anteriorly converging plicae, which extend around onto ventrolateral sides of tongue as distinct folds. Posteriorly, these elongate plicae start to become subdivided into shorter sections before they are replaced by a zone of scalelike papillae, which occupy the midsection of the tongue in an area over- lapping the pigmented anterior and unpig- mented posterior halves. Behind the scalelike papillae, the mostly unpigmented posterior half of the tongue again bears distinct, ante- riorly converging plicae in the shape of chev- rons; the proximal half dozen or so of the posterior plicae are emphasized by gray pig- mentation along their free edges. Raised mid- ventral side of tongue with a median groove, which continues anteriorly through three pairs of thin (nonswollen), oblique, anteriorly
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 33
Fig. 33. new species (holotype).
Tongue of Adercosaurus vixadnexus,
converging (chevronlike) infralingual plicae immediately behind fork; first pair of infral- ingual plicae slightly larger than second pair, third pair barely indicated.”°
Anterior maxillary and dentary teeth near- ly conical, slightly recurved; posterior max- illary and dentary teeth primarily bicuspid, with a rare tricuspid tooth; small pterygoid teeth present.** Phalangeal formula of hands 2-3-4-5-3, of feet 2-3-4-5-4.
°3'The appearance of the infralingual plicae is similar to the condition illustrated for Ptychoglossus by Harris (1985: 562, fig. 2b), except that the apices of the ante- riorly converging plicae are pointed in Adercosaurus, not slightly rounded.
*4'Tiny teeth on the pterygoid could be seen under magnification after the lower jaw was widely reflected (for the purpose of photographing the tongue). However, the pterygoid teeth are not discernible by magnification of a radiograph of this specimen, which lends support to Harris’ (1994: 230) suggestion that pterygoid teeth in some small teiids have been overlooked by investigators who relied on radiographs.
54 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 34. Adercosaurus vixadnexus, new spe- cies. Dorsal, ventral, and lateral views of head of holotype (EBRG 3126). Scale line = 5 mm.
Head scales (fig. 34) smooth, with numer- ous minute scale organs (pits) especially con- centrated on anterior head plates (dorsal, lat- eral, and ventral), becoming posteriorly sit- uated mainly on margins of head plates and nape scales. Similar minute pits also sparsely distributed on edges of dorsal scales.
Snout rounded. Rostral plate much wider than deep, laterally in contact with nasal and first supralabial, dorsally in contact with frontonasal. Anterior and posterior edges of frontonasal nearly straight. Paired prefrontals medially in broad contact. Frontal slightly hexagonal, with weak anterior apex and pro- nounced posterior one, longer than wide. Frontoparietals two, rhomboidal, barely in median contact. Supraoculars four, in contact with superciliary series; posterior two supra- oculars on right aberrantly subdivided (fig. 34). Interparietal longer than wide, heptago- nal, with anterior apex and short, straight posterior edge; lateral edges of interparietal not quite parallel, slightly diverging posteri- orly. Parietals longer than wide, but shorter than interparietal, which extends farther cau-
NO. 261
dad—posterior margins of parietals and in- terparietals forming a jagged line across rear of head. Occipitals (postparietals) three, fol- lowed by a medially interrupted transverse row of four smaller postoccipitals, which in size are subequal with dorsal neck scales.
Nostril situated below center of nasal, which is undivided albeit with a very faint, thin groove below nostril. Loreal anteriorly inclined, higher than wide, dorsally in con- tact with frontonasal and first supraocular at their narrow suture, ventrally well separated from supralabials by frenocular. Preocular scales two, very small, between first super- ciliary and first subocular. Superciliaries five, with an azygous scale on both sides lying above the suture between small second and third superciliaries, and below suture be- tween second and third supraoculars; first su- perciliary long and wide. Suboculars four, moderately large, behind frenocular, none greatly elongate. Postoculars two, upper smallest, between superciliary and subocular series. Supralabials seven on left side, six on right, ultimate plate smallest on each side.
Upper eyelid with eight or nine ciliaries, none exceptionally enlarged. Ocular recess with one row of small, flat scales separating all but first few and last few ciliaries from superciliaries. Lower eyelid with nine cili- aries of subequal size. Lower eyelid scaled, with a large but not well-defined opaque me- dian window (palpebral disc) with approxi- mately six vertical divisions. Palpebral disc mostly in contact with lower eyelid ciliaries, with only a few minute scales intervening.
Temporal region between postoculars and ear with approximately 21 irregularly shaped, juxtaposed smooth scales. Ear open- ing vertically oval, edged around with small, smooth scales, those on anterior edge swol- len; a row of small flat scales around inner rim of opening. Auditory meatus shallow, tympanum lightly pigmented.
Posterior margin of mental nearly straight. Postmental large, slightly pentagonal, wider than long, laterally in contact only with elon- gated first infralabial on each side. Genials in three pairs, all in contact with infralabials; first two pairs in broad median contact; third pair of genials widely separated by interven- ing pregular and other scales. A pair of side- by-side, moderate-size postgenials (or pre-
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF so)
Fig. 35. Adercosaurus vixadnexus, new species (holotype). Details of gular—pectoral (left) and sub-
pelvic (right) regions.
gulars) in posterior contact with each mem- ber of the third pair of genials; lateral mem- ber of each pair of postgenials separated from infralabials by an elongated sublabial. The third pair of genials and the two pairs of postgenials widely separated by wedge of three pairs of pregulars, first pair in contact, second and third pairs separated by irregular smaller scales.
A narrow but distinct gular crease crossing anterior edge of throat from ear to ear, evi- dent in both lateral and ventral view; gular crease bearing a line of tiny scales contigu- ous laterally but becoming widely separated ventrally. Seven transverse rows of gulars between gular crease and collar scales; first several gular rows somewhat irregular (third row incomplete); two middle scales in last three rows forming double row of enlarged, paramedian gulars, each succeeding pair larger than the preceding (fig. 35). Gulars im-
bricate. Well-defined guttural fold, with tiny scales, not crossing throat, ending ventrolat- erally on each side at suture between second and third double rows of enlarged gulars. Collar row with seven scales, the median scale about as wide as long and much larger than the longer-than-wide lateral collar scales. Collar fold very evident, with 2-3 rows of tiny scales broken by median gap. Conspicuousness of anterior gular crease, in- complete guttural fold, and collar fold not due to distention of neck with preservative— all evident in life (as determined by enlarge- ment of transparency used for fig. 31, bot- tom).
Sides of neck with covering of medium- sized, oval or rounded, smooth juxtaposed scales, arranged in about eight oblique rows, which do not (or only weakly) form annuli with nape and gular scales. Last two rows of neck scales anteriorly delimited by 1—2 rows
56 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
NO. 261
Fig. 36. Adercosaurus vixadnexus, new species (holotype). Dorsal midbody scales (anterior is left), x18. An odd, nonmucronate rectangular scale is visible in the lower right quadrant.
of tiny scales in guttural fold and posteriorly delimited by tiny scales anterior to arm. Area of roughly 6—8 rows of tiny scales in axilla continuing as semicircle up over arm and then ventrad, with a few rows then confluent with tiny scales in collar fold.
Dorsal and lateral scales (fig. 36) elongat- ed, parallel-sided, hexagonal,?? keeled and strongly mucronate—keels extending poste- riorly as elongated, acute points. Scale keels sharpest dorsally, becoming sharpest poste- riorly on body, losing definition on nape and sides of body, where narrow median keels are replaced by wider, smooth-top elevated thickenings. Dorsals and laterals otherwise homogeneous, in uninterrupted transverse rows; no lateral fold or line of small scales. Dorsals in 36 transverse rows between inter-
°° Although misleading, “hexagonal” is a venerable and still useful descriptor for the body scales of some microteiids. Hexagonal scales are parallel-sided, elon- gated pentagonal scales that are arranged in transverse but not in longitudinal rows. The free pointed tips of laterally adjacent scales fall into sutures between scales in the next transverse row, causing each single scale to appear six-sided.
parietal and posterior edge of thigh held at right angle to body.
Caudal scales (fig. 37, top) similar to body scales, in uninterrupted annuli (no interven- ing small scales); supracaudals not forming paramedian series along a vertebral suture; supracaudals strongly keeled to end of tail, keels becoming broader and blunter on lat- eral scales and disappearing ventrally.
Ventral scales forming both transverse and weak longitudinal rows. Ventrals in 21 trans- verse rows between collar and pair of ante- rior preanals, with some irregularity in size of scales in first and last rows. Ventrals in about eight longitudinal rows of smooth scales at midbody—a subjective count inas- much as the mucronate lateral scales grade imperceptibly into ventrals, both in shape and in a gradual shift to longitudinal rows; lateralmost one or two scales bluntly pointed; scales in median four or five ventral rows nearly rectangular, but with their free edges slightly rounded and barely overlapping the bases of scales following (G.e., weakly im- bricate).
Marginal preanal scales three, outer two
Fig. 37. Caudal scutellation in dorsal view (anterior is left). Top: Adercosaurus vixadnexus, new species (holotype), showing scales arranged in transverse rows only; there is no paramedian series along a vertebral suture. Bottom: Ptych- oglossus plicatus (AMNH_ 114307¢, Panama), showing supracaudals arranged in both transverse and longitudinal rows; arrows indicate vertebral suture separating paramedian series.
more than twice the size of median one; pair of elongate preanal scales in posterior contact with the small, median marginal scale (fig. 35). Femoral pores 2/2 on posteroventral sur- face of thigh, preanal pores 2/2 in same line as femoral pores; each femoral and preanal pore surrounded by 3-5 irregularly shaped small scales.
Scales on dorsal surface of arm, hand, and ventral side of forearm large (wider than dor- sal body scales), imbricate to juxtaposed, smooth, rhomboidal or irregularly shaped; scales on underside of upper arm small, smooth, slightly swollen, juxtaposed. Scales on anteroventral surface of thigh and most of lower leg large, mostly smooth (a few keels on posterior scales atop thigh); posterodorsal and posterior surface of thigh pebbled with smooth, swollen granules. Supradigital scales glossy, squarish or longer than wide, distal scale proximal to ungual scale especially
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF a6
Fig. 38. cies. Palmar view of right hand (holotype).
Adercosaurus vixadnexus, new spe-
elongated. Palms and soles covered with small, rounded juxtaposed scales that are not thickened or rounded. Two enlarged thenar scales on inner margin of palm below pollex (fig. 38), each with produced inner edge; a similar but smaller pair of enlarged keeled scales below first toe. Subdigital lamellae di- vided (each half tends to be inclined up to- ward the middle, forming median subdigital keel between the paired scales, but this may be an artifact of preservation or desiccation).
Subdigital lamellae as follow (Roman nu- merals indicate digits, Arabic numbers indi- cate pairs of lamellae on left/right feet): fore- foot, I 4/4 If 6/6 III 8/8 IV—/9 V 6/6; hind foot, I 4/3 If 6/7 If 10/10 IV 14/14 V 8/8. (A few of these counts are approximations owing to a combination of desiccation and variation in scales at bases of digits; the di- vided lower ungual sheath scale is omitted from the counts.)
COLORATION: In life (fig. 31), the dorsum was brown, with indication of a paler brown dorsolateral line; the sides were black with whitish flecks; the ventral surfaces were strikingly patterned in black and white.
58 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Close inspection of the preserved speci- men (fig. 32) shows the brown dorsal scales to be finely mottled with darker brown. The paler tan dorsolateral line extends posteriorly from the eye, across the upper temporal re- gion, and well above the ear to the anterior body. The dorsolateral line reappears at the end of the body, extending above the hind legs and onto the base of the tail for a dis- tance of 7—8 caudal annuli. The sides of the body are mostly black, with irregular small whitish and tan spots. The lips are dark ex- cept for a few small white areas, which are situated mainly along the edges of the mouth at sutures between the labial scales. The un- derside of the head is mottled black and white, turning darker on the throat. The basal and lateral edges of the ventral and most sub- caudal scales are black, offsetting grayish white centers.
HEMIPENIS: The genitalia of the holotype were field everted, and the left organ was subsequently removed and inflated with red petroleum jelly (fig. 39). This hemipenis is 5 mm long, with a greatest width of nearly 4 mm. The organ is bilobed, each of the short symmetrical divisions terminating in a flat- tened, trilobed apical disc of thick tissue. The two discs tilt slightly mediad toward one an- other, with the medial lobe of each disc being deflected distad, resulting in paired projec- tions between the outer flat parts of the discs.
The sulcus spermaticus is a broad, indefi- nite channel, from which two narrow, ill-de- fined groves form below the crotch (without visible bifurcation); these weak, distal sulcate branches diverge with centrolineal orienta- tion, each branch extending into the gap be- tween the medial and sulcate-side lobes of the apical disc. The thickened tissue com- prising the apical discs is proximally contin- uous with the broad medial band of tissue that bears the sulcus spermaticus (fig. 39).
There is a conspicuous series of 11 or 12 curved, nude tissue ridges across the asulcate surface, encircling the side of the organ and terminating laterally on the sulcate face. The distal ridge at the base of a hemipenial lobe is small and ill-defined; the continuity of the basal seven ridges is broken on the side of the organ by a nude area extending up from the base. No mineralized spines, spinules, or hooklike papillae are evident at high mag-
NO. 261
nification, nor were such structures evident after the organ had been immersed for 24 hours in a solution of Alizarin Red.
REMARKS
The unique specimen of Adercosaurus vix- adnexus was found near camp by our col- league Paul Sweet, at about 8 a.m. on a sun- ny morning, after a night of local flooding (which sometimes displaces small inhabi- tants of the ground litter). It was on soggy ground at the edge of wet mossy gallery for- est. We assumed in the field that the speci- men was an Arthrosaura, but, in preparing this paper, we found that it fits neither there nor in such extralimital genera as the physi- ognomically similar Ptychoglossus (see Har- ris, 1994). Adercosaurus thus joins Riolama (Uzzell, 1973) as a second genus of micro- teiids that seems to be endemic to the Ve- nezuelan tepuis. Their relationships remain to be determined.
Our colleague Dennis Harris (personal commun.) called our attention to the possibly diagnostic point contact of the frontoparietal plates. Harris noted that the form of the fron- toparietals is very conservative in “‘group II teiids’’, except for obvious (bilaterally assy- metrical) anomalies or when there are radi- cally different head scales (as in Neusticurus, Echinosaura, Teuchocercus). Otherwise, the frontoparietals, unless absent (as in Gymno- phthalmus), characteristically have a broad median suture. Interestingly, one exception is Arthrosaura reticulata, 11 which the medial suture varies “‘from short to relatively wide”’ (Avila-Pires, 1995: 337, fig. 111).
A resemblance in tongue morphology be- tween Adercosaurus and Riolama is of inter- est, although variation in lingual surface morphology is not well accounted for. Ac- cording to Uzzell (1973: 52), the holotype of Riolama leucosticta has a tongue ornamen- tation of chevron-shaped plicae on the ante- rior and posterior thirds, separated by scale- like papillae on the middle third—similar to Adercosaurus as described herein. However, at least in some Riolama spp. (e.g., AMNH 141343, Cerro Duida, and AMNH 132118- 132119, Cerro de la Neblina) the tongue seems almost entirely plicate, with, at most, a few short, scalelike subdivisions of the pli-
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 59
Fig. 39. sulcate (left) and asulcate view, X12.
cae occurring on the lateral edges of the tongue at its midsection. A Brazilian micro- te1id, Ecpleopus gaudichaudi, has a variable tongue morphology according to Uzzell (1969: 5), who stated that “‘the middle and sometimes the posterior part’? are covered with scalelike papillae, but that the usual pli- cae at the anterior tip were lacking in one specimen. Another specimen of E. gaudi- chaudi (AMNH 131869) not then available to Uzzell clearly has anterior and posterior lingual plicae interrupted by a midsection of scalelike papillae.
Macroteiids and nearly all microteiids are characterized by scalelike papillae over most of the tongue. Often it is not convenient to examine the proximal part of the tongue without risk of damage to the specimen, but a combination of anterior scalelike papillae and posterior plicae is the common condition according to Ruibal (1952: 511), who stated that ‘“‘Actually most teiids that have imbri- cate scale-like papillae on the tongue may have the posterior bifurcation covered with oblique plicae.”’
Heretofore, only Alopoglossus, Ecpleopus, Ptychoglossus, and Riolama seem to have been reported to have plicae also on the an- terior part of the tongue (Uzzell, 1973: 54),
Hemipenis of Adercosaurus vixadnexus, new species. Everted left organ of holotype, in
with the fully plicate tongues of Alopoglos- sus and Ptychoglossus being “‘practically a unique feature”’ (Harris, 1994: 229). But the taxonomic distribution of anterior lingual pli- cae, as well as extent of intraspecific varia- tion, is not known with certainty. This was brought home by casual examination of spec- imens of the long-tailed microteiid genus Pantodactylus. Specimens of Pantodactylus schreibersii (AMNH 32776, 144571) have, as expected, mostly scalelike papillae and posterior plicae, but a specimen of P. guad- rilineatus (AMNH 131877, Sao Paulo) has anterior plicae that give way to scalelike pa- pillae posteriorly (proximal part of tongue not examined).
Most of the morphological features com- monly used to diagnose genera of Teiidae have been in use for over a century (see Bou- lenger, 1885: 330 et seq.). However, in a se- ries of important papers in the 1960s and 1970s, Thomas Uzzell removed and stained the uneverted hemipenes of many microteiids (see especially Uzzell, 1973: 40—46). Uzzell showed that, in addition to various small spines, the hemipenes of many if not most microteiid genera have facing series of oblique or chevron-shaped folds or flounces that bear comblike rows of mineralized spi-
60 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
nules, which are easily stained with Alizarin Red. The complex folds formed by the lobes at the tip of the uneverted organ were not amenable to study by dissection, and the spi- nule-bearing flounces on the body of organ were characterized as being in pockets. With more field-inflated hemipenes becoming available, the spinule-bearing flounces are seen to be superficial on the surface of the everted organ (e.g., Donnelly et al., 1992: fig. 3; Myers and Donnelly, 1996: fig. 20), and the everted apices take on recognizable shapes of systematic value (see especially Harris, 1994).
‘“‘Nude”’ hemipenes lacking mineralized spines or spinules, although uncommon among microteiids, are characteristic of Ec- pleopus (Uzzell, 1969: fig. 8, uneverted) and Ptychoglossus, with the latter having ‘‘flounces of the form found in macrotetids and the closely related genus Alopoglossus, a primitive condition among microtetids”’ (Harris, 1994). Adercosaurus resembles Ptychoglossus and Alopoglossus in general hemipenial morphology. The similarity of the asulcate nude ridges or flounces can be seen by comparing photographs of the left evert- ed hemipenes of Adercosaurus vixadnexus (fig. 39) and Ptychoglossus plicatus (Harris, 1994: fig. 17 [AMNH 114307]). Not visible in photographs are extremely minute bumps or microknobs, which are barely discernible on these two organs under the dissecting mi- croscope; their structure is not resolvable at a magnification of X50. Possibly these are the “‘densely packed minute hook-like pa- pillae”’ said by Harris (1994: 229, 231) to ornament the nude flounces of Ptychoglos- sus and Alopoglossus. If so, they are too easily overlooked to provide taxonomic content at this time. They probably repre- sent the largest features in the epithelial landscape, which is best studied by scanning electron microscopy.
Prionodactylus goeleti (Myers and Donnelly), new combination Figures 40, 41, 42 (top), 43 (part)
Euspondylus goeleti Myers and Donnelly, 1996: 23-31, figs. 16-20. Holotype EBRG 3112 (field no. CWM 19639) from summit of Cerro Yavi, 2150 m, Amazonas, Venezuela; collected Feb-
NO. 261
ruary 19-20, 1995, AMNH-—TERRAMAR Expe- dition.
Prionodactylus phelpsorum (Lancini): Gorzula and Sefaris, 1999: 139-141.
MATERIAL: AMNH 147038 (Guv.), 147039 (egg shells), EBRG 3127 (egg shells), Cerro Yutajé, 1700 m, February 24—28, 1995.
An additional juvenile (““MHNLS 11387”) was reported by Gorzula and Sefi- aris (1999: 139) as being collected first on Cerro Corocoro on November 10, 1977, and secondly in the Serrania de Yutajé on March 22, 1988—presumably there actually are two specimens, although on page 141 they re- peated the same number for ‘“‘a juvenile male” from each locality.
REMARKS
The only lizard of this species obtained by us on Cerro Yutajé is an unsexed juvenile taken on rocks at the edge of a stream, at the 1700 m camp (fig. 40, left). A nesting site was found in a small mat (0.09 m7) of veg- etation atop bare sandstone, in Tyleria—Ste- golepis—Brocchinia scrub. Hatched eggshells of different-aged clutches were found on the rock underneath the small mat of vegetation and also in cavities within the mat. Two mea- surable shells were 10.6 and 11.0 mm long by 6.0 and 6.8 mm wide. Communal nest sites on Cerro Yavi were under rocks, mostly on layers of fibrous plant material or occa- sionally on fine sand (Myers and Donnelly, 1996: 30).
Our specimen from Yutajé is a juvenile of 25 mm SVL. In life, the pale middorsal stripe was beige and the dorsolateral and lateral lines were pale brown, with the lateral line being anteriorly confluent with a white labial line. The underside of the head was whitish and the venter pale green. Iris orange. Tongue sharply bicolor, the anterior third black.
The specimen compares favorably with three juveniles in the type series of Priono- dactylus goeleti, collected several days ear- lier on neighboring Cerro Yavi, although a few differences can be found, especially the following: (1) the Yutajé specimen has rela- tively smaller scales in the midgular region (fig. 41). Although subtle, the difference could be seen with a jeweler’s loupe when
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 61
Fig. 40. Possible interpopulational difference in width of pale middorsal stripe in juvenile Priono- dactylus goeleti (Myers and Donnelly), 1.75. Left: Cerro Yutajé, a juvenile (AMNH 147038) 25 mm SVL. Right: Cerro Yavi, a juvenile paratopotype (AMNH 141331) 26 mm SVL. The middorsal stripe is narrower in juveniles than in adults, and its posterior widening becomes more noticable in adults (compare with fig. 43); see Taxonomic Notes.
Fig. 41. Possible interpopulational difference in size of midgular scales in Prionodactylus goeleti (Myers and Donnelly). Left: Cerro Yutajé, a juvenile (AMNH 147038) 25 mm SVL. Right: Cerro Yavi, a juvenile paratopotype (AMNH 141331) 26 mm SVL.
62
NO. 261
BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Fig. 42. Top: Prionodactylus goeleti (Myers and Donnelly), the adult male paratype (AMNH 141329, 50 mm SVL) from Cerro Yavi, Venezuela. Middle: The generic type species, Prionodactylus manicatus (O’Shaughnessy), an adult female (AMNH 113016, 44 mm SVL) from Cordillera Azul, eastern Peru. Bottom: Prionodactylus vertebralis (O’Shaughnessy), an adult male (AMNH 119368, 64 mm SVL) from Cerro Mali, on the Panamanian—Colombian border.
2001
the freshly preserved specimens were com- pared in the field. (2) The Yutajé specimen has a narrower middorsal stripe than do any of the three juveniles from Cerro Yavi. The last specimens are pictured in Myers and Donnelly (1996: fig. 19). One of the Yavi juveniles is shown here (fig. 40, right) in di- rect comparison with the Yutajé specimen, in which the median stripe is 1 scale wide be- tween the arms, 1% scales at midbody, and 3 scales wide above the inguinal region. Gor- zula and Sefaris (1999: 141) described the stripe of another juvenile from the Serrania de Yutajé as being similarly narrow (1 scale wide at nape, 2 scales at midbody, 3 scales above inguinal region). (See below for dis- cussion of stripe width on Cerro Yavi.)
The samples are of course too small for reliable extrapolation. Nonetheless, the nar- row median stripes in the aforementioned two juveniles from Cerro Yutajé suggest that the adults may also have narrower stripes than the adult lizards on Cerro Yavi. If so, Prionodactylus goeleti on the Yutajé—Coro- coro massif is slightly differentiated from the population at the type locality.
TAXONOMIC NOTES
GENERIC STATUS: We follow Gorzula and Sefiaris (1999: 140) in their reassignment of Euspondylus phelpsorum (including E. goe- leti) to Prionodactylus, although the generic situation remains unsatisfactory. Uzzell (1973) considered the type species of Eus- pondylus to be “‘essentially devoid of unusu- al external morphological’’ features, and he revived the younger name Prionodactylus from the synonymy of Euspondylus for spe- cies having ‘“‘a double row of widened gular scales and keeled hexagonal scales’’. But these characters, even in combination, do not uniquely define Prionodactylus, and Euspon- dylus remains loosely defined. Paired median gulars seem to be a tendency in several mi- croteiid lineages, and the character apparent- ly is not strongly fixed in some Prionodac- tylus. Avila-Pires (1995: 453, 456, 470) not- ed that the gulars are occasionally irregular and not arranged in pairs in P. argulus and P. oshaughnessyi.
The hexagonal dorsal scales in Prionodac- tylus (and some other genera) appear distinc-
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 63
tive when sharply keeled and strongly an- gular, but it is only a degree of difference from Euspondylus maculatus (type species), which has relatively wider (less elongate) rectangular scales. As in Priondactylus, how- ever, the scales are disposed in transverse rows only, so that acquisition of points on the posterior scale margins would give the appearance of hexagonal scales. The pres- ence or absence of keeling is not an absolute difference, inasmuch as some specimens of Peruvian Euspondylus maculatus have weak- ly keeled dorsals posteriorly on the trunk (keels absent in AMNH 1704 but present in AMNH 56268; posterior keeling was report- ed for a syntype by W. Peters, ‘1862’ [1863]: 207).
Whether all the species of Prionodactylus form a monophyletic group that excludes Euspondylus maculatus is a problem for fu- ture revisionary work. Meanwhile, the ge- neric reallocation of Venezuelan tepui spe- cies by Gorzula and Sefiaris has the practical advantage of associating them in the same group with the similar-appearing Andean Prionodactylus vertebralis (fig. 42, bottom).
SPECIES RECOGNITION: Prionodactylus goe- leti was described by Myers and Donnelly (1996: 23-31) on the basis of six specimens (3 adults, 3 juveniles) collected on Cerro Yavi in 1995. The species was recognized as being probably related to Prionodactylus Phelpsorum (Lancini, 1968), which was de- scribed from an adult female taken in 1967 on Cerro Sarisarifama,”° some 240 km south-
6 The type locality of Prionodactylus phelpsorum must be corrected as pointed out by Gorzula and Sefiaris (1999: 140), who called attention to a locality emenda- tion by Steyermark and Brewer-Carias (1976: 180). Phelps (1977: 16 and map 1) provided further details of the geographic confusion that occurred on his pioneering 1967 expedition, when the lizard was collected. Owing to topographic complexity and misleading information from Maquiritare Indians, Phelps established a “‘Cam- pamento Cerro Jaua’’ on what actually was the western side of the mesa of Cerro Sarisarifiama.
Phelps (1977: 16—17) changed his specimen data from Cerro Jaua to Cumbre Occidental, Meseta de Sarisari- flama, with coordinates estimated at 4°45'N, 64°26’W. The helicopter-supported camp was occupied by Phelps’ party for 10 days (March 20—29, 1967), with collecting carried out between 1900 and 2050 m above sea level. The holotype of Prionodactylus phelpsorum was taken at 1917 m on March 29 (Lancini, 1968: 2).
The locality was shown by Phelps (map 1) as ““Cam-
64 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
east of Cerro Yavi. The main diagnostic dif- ference was that goeleti ‘‘is distinguished by a posteriorly expanded middorsal stripe.”
The three adult and three juvenile P. goe- leti were separately described, with clear in- dication of ontogenetic change in relative width and posterior widening of the middor- sal stripe: ““Three juveniles are colored sim- ilar to the adults although the pale brown middorsal stripe is relatively narrower and its posterior widening is less pronounced in the juveniles than in adults’? (Myers and Don- nelly, 1996: 27). The point was made with facing-page photographs comparing all spec- imens in the type series of goeleti, and equiv- alent stripe widths (total scales covered) were given for adults and juveniles (p. 27), as summarized below:
Adults Juveniles Stripe width (scales) (scales) Between arms 2-3 2” Midbody 3%-4 3-4 Inguinal region 5 +
We reiterate the above variation because we evidently should have dwelled on its ob- vious taxonomic implications, namely that ontogenetic variation should not be confused with noncorrelated individual variation and that the dorsal stripe in phelpsorum (juve- niles unknown) is best compared with the adult condition in goeleti; juvenile goeleti have narrower stripes (although they too show the diagnostic posterior widening). In allocating the name Euspondylus goeleti to the synonymy of Prionodactylus phelpso- rum, Gorzula and Sefiaris apparently con- fused ontogenetic variation with the kind of extensive intraspecific variation that is not correlated with age, sex, or geography. This caused them to differently interpret Lancini’s photograph of the phelpsorum holotype—an adult female containing two eggs.
Lancini (1968) does not mention whether or not the middorsal stripe is expanded posteriorly. However, [Lancini’s] Figure 7 shows a[n]
<
pamento Phelps—Steyermark, 1967, 2000 m’’, and a pho- tograph (fig. 2) of the habitat at 2000 m also was given. A camp on Cerro Jaua was later established, as shown on the same map.
NO. 261
oblique view of the dorsum of the holotype where it can be seen that the middorsal stripe is narrow anteriorly [and posteriorly] and wid- ens to cover the dorsal surface of the tail [em- phasis added]. In addition, the photographs of Myers’ and Donnelly’s own specimens show variation in the width of the middorsal stripe. For example, it is narrow in AMNH 14133 [sic (presumably 141331, a juvenile)]. Of the spec- imens reported here, the middorsal stripe of MHNLS 11387 Guvenile male from Serrania de Yutajé) is only 1 scale wide at the nape, 2 scales wide at midbody and 3 scales wide above the inguinal region. (Gorzula and Sef- aris, 1999: 140-141)
Stripe width on the tail is not relevant to the present discussion, and the Yutajé juve- nile described above obviously has the stripe posteriorly widened on the body, as does our juvenile specimen from Cerro Yutajé (fig. 40, left). Lancini (1968: 3) said that the middor- sal stripe of phelpsorum was up to two scales wide (hasta dos escamas de anchura), which seems consistent with his photograph show- ing a median stripe that is apparently narrow for the length of the trunk. An enlarged copy of Lancini’s photograph of Prionodactylus Phelpsorum is reproduced in figure 43, in side-by-side comparison with the adult type specimens of Prionodactylus goeleti. The Phelpsorum holotype is obviously more ro- bust than the specimens of P. goeleti, but that may be due to a combination of gravidness and distention with preserving fluid. In any case, despite the oblique view, it seems clear that Lancini’s description of the median stripe was accurate, and that this stripe is not posteriorly expanded in P. phelpsorum.
Myers and Donnelly (1996: 30-31) took what they could from Lancini’s overly brief description of the phelpsorum holotype (which is lost according to verbal commu- nication from Alfredo Paolillo) and pointed out similarities with goeleti. The only differ- ences were the obvious one in adult middor- sal-stripe width, a slight difference in dorsal- scale count (40 in one phelpsorum, 33-38 in six goeleti), a possibly subjective difference in degree of keeling, and an apparent differ- ence in ventral coloration.
Only the last character above was used as a secondary diagnostic character for differ- entiating goeleti from phelpsorum, and it de-
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 65
Fig. 43. first specimen on left) and Prionodactylus goeleti (next three specimens, from left to right: EBRG 3112¢ [holotype], AMNH 1413302, 1413293). Holotype of P. phelpsorum enlarged X2.6 from photograph in Lancini (1968: fig. 7); specimens of P. goeleti from Myers and Donnelly (1996: fig. 18).
pended on an interpretation of Lancini’s de- scription. As also noted by Gorzula and Sefi- aris (1999: 141), Lancini “‘did not describe the ventral color of the body, but simply stat- ed that it was similar to that of the tail which was ‘gris plomizo salpicado de negro’ (slate gray with black speckling).”’ In contrast, to- potypic adult female goeleti were either white or bluish white under the head, and greenish white on other ventral surfaces, ex- cept for a tinge of orange under the tails. Gorzula and Sefiaris (1999: 141) quoted Gor- zula’s fieldnotes describing the ventral col- oration of another female from Cerro Yavi as ‘pale yellowy brown’’, and then obfuscated the matter by discussing ventral color in ju- veniles (pale green as described for the ju- venile paratopotypes or “‘dirty pale grayish green with faint gray dusting” in a juvenile male from Cerro Corocoro). Gorzula and Sefiaris concluded their discussion of ventral color by saying that “‘Such variation is not convincing to diagnose a new species.”’ Nonetheless, an adult female ventral color of slate gray in life is suggestive to us of taxonomic difference when compared with a geographically remote population in which female venters are pale greenish white or ‘pale yellowy brown’’. But we took Lancini at face value that the throat and venter of the
Adult dorsal color patterns of type specimens of Prionodactylus phelpsorum (holotype &,
female specimen of phelpsorum were slate gray in life, and assumed that if the ventral ground color in life had been pale he would have said so. If, however, he had misleading- ly extrapolated ventral color from the pre- served specimen, it would be a different mat- ter, since female goeleti became light gray in preservative “‘apparently owing to expansion of melanophores that were not evident in life’? (Myers and Donnelly, 1996: 27). Ob- viously, posterior widening of the middorsal stripe—most evident in adults, albeit present in juveniles—is probably the most reliable character for separating Prionodactylus goe- leti from P. phelpsorum (fig. 43). The ab- sence of a middorsal stripe seems to distin- guish a third tepui species, as discussed be- low.
Gorzula and Sefaris (1999: 129-139) de- scribed Prionodactylus nigroventris from 1650 m elevation on Cerro Guanay, a neigh- boring tepui immediately to the northwest of the Yutajé—Corocoro massif (fig. 1). The type series of P. nigroventris comprised one adult male, a subadult male, and a subadult female. The new species was compared with P. goe- leti (as “‘phelpsorum’’)—its nearest geo- graphic relative—only in the Diagnosis, where P. nigroventris was said to differ by having a “‘dark black venter in adult males
66 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
(orange in P. phelpsorum) and 6 to 9 rows of temporals between the middle postocular and the scales bordering the ear opening.”’
The comparison of the black venter of adult male nigroventris with the orange ven- ter of adult male phelpsorum seems based on two specimens—the nigroventris holotype and our earlier description of a male goeleti, which developed extensive black ventral col- oration after preservation (Myers and Don- nelly, 1996: fig. 18, lower right). Concerning the temporal scales, Gorzula and Sefiaris did not give data for phelpsorum/goeleti, but lat- eral head drawings of two nigroventris do not show the temporal scales as being clearly organized in rows. However, the lateral head views shown by Gorzula and Sefiaris (1999: figs. 17b, 18b) for P. nigroventris do bear some resemblance to a drawing of P. goeleti, which was described as having the ““Tem- poral region between postoculars and ear opening with about 15—25 irregularly shaped juxtaposed scales, of which the upper ones are large and the lower ones small to medium in relative size”” (Myers and Donnelly, 1996: 26 and fig. 17). The smallest scales are wedged obliquely from near the postoculars toward the corner of the mouth. But there are other characters not discussed.
Examination of the description of P. ni- groventris reveals an especially striking dif- ference between it and the other tepui species (i.e., P. goeleti and P. phelpsorum [as well as the extralimital P. vertebralis, fig. 42, bot- tom]). Prionodactylus nigroventris evidently lacks the pale middorsal stripe that is so con- spicuous in the other species! We admittedly extrapolate, since Gorzula and Sefiaris nei- ther published a photograph nor explicitly stated that nigroventris lacks a pale middor- sal stripe, but its absence has to be assumed from their description: In life, ““‘Dorsum of head dark brown. Dorsal scales brown, slightly lighter than the head. Dorsolateral scales greenish yellow, forming two dorso- lateral lines’? (Gorzula and Sefiaris, 1999: 137):
Therefore, as described by Gorzula and Sefiaris, the dorsal color pattern of their Prionodactylus nigroventris seems to be much like that of the generic type species, P. manicatus (fig. 42, middle) from western South America. The absence of a median
NO. 261
stripe should immediately separate P. ni- groventris from the other two tepui species.
FAMILY TROPIDURIDAE
Only one species of tropidurid lizard was obtained by the 1995 AMNH-—TERRAMAR Ex- pedition. It is close to a species that we pre- viously named in the genus Plica (Donnelly and Myers, 1991), which subsequently was placed in the synonymy of TJropidurus by Frost (1992).
Tropidurus panstictus, new species Figures 44A, 45—48
HOLOTYPE: EBRG 3130 (field no. CWM 19795), an adult male from south end of Cer- ro Corocoro, 1220 m (5°42’N, 66°10’W), Amazonas, Venezuela; collected March 1, 1995, AMNH-—TERRAMAR Expedition. See locality 2 on map (fig. 1).
PARATOPOTYPES: AMNH 147040—147045, EBRG 3128-3129, 3131-3132, from same locality as holotype, collected February 28— March 1, 1995, AMNH-—-TERRAMAR Expedi- tion.
PARATYPES: AMNH 147046, EBRG 3133- 3134, from above Yutajé, Rio Corocoro, 180 m (5°37'N, 66°07'W); collected March 2, 1995, AMNH-—TERRAMAR Expedition. See locality 3 on map (fig. 1).
ETYMOLOGY: The species name panstictus is an adjective derived from the Greek pan- (all, all over) + stiktos (dappled, spotted, punctured), in allusion to the profuse speck- ling and spotting over the head, body, and limbs of adults.
DIAGnosis: A large Tropidurus having several tufts of elongate, spinous scales on the neck; complete gular and antegular folds; anteriorly imbricate head scales; small, slightly imbricate, acutely pointed, thornlike body scales; and a middorsal crest from rear of head onto the tail.
Tropidurus panstictus most nearly resem- bles T. lumarius, from which panstictus dif- fers in larger size and a different color pat- tern, which is much lighter and which in- cludes profuse pale speckling in adults (see later, under Comparison with Tropidurus lu- marius). Tropidurus panstictus differs from other spiny-necked tropidurines in the same way that 7. lumarius does (Donnelly and
2001
Myers, 1991: 31-32). We provide a key to Venezuelan tropidurines following the Re- marks section.
SIZE AND PROPORTIONS
Head large, head length (measured obliquely from tip of snout to angle of jaw) 26-30% of SVL; head longer than wide and wider than high, wider than neck; neck nar- rower than body. Snout rounded to broadly rounded in dorsal view, not projecting over mouth. Body wider than deep. Limbs long and slender, pentadactyl; relative finger and toe lengths 4 > 3 >2>5 > 1. Tail 1.6—2.1 times SVL; tail becoming slightly vertically compressed and then rounded toward end. Four adult males with maximum snout—vent length of 119 mm (96-119 mm, x* = 108.3 mm, SD = 9.6). One adult female 93 mm SVL, five juvenile males 62—94 mm SVL, two juvenile females 66-71 mm SVL, and two unsexed juveniles 40-54 mm SVL.
SQUAMATION, POCKETS, AND FOLDS
HEAD (fig. 44A): Interparietal large, roughly triangular, wider than long or longer than wide; posteriorly bordered by 1—3 rows of small smooth or rounded scales separating the interparietal from small pointed scales on rear of head; laterally in contact mainly with irregularly shaped smooth scales. Scales of interocular and frontonasal region irregularly shaped, usually smooth (with sensory pits), occasionally a few scales keeled; weakly im- bricate anteriorly; 2—3 rows between circu- morbital series (three in holotype). Circu- morbitals mainly in 2 rows, inner row with largest scales, anteriormost and especially posteriormost scale in inner (and sometimes outer) row often resembling adjacent supra- oculars in being longer than wide. Three or (usually) four large supraoculars (4/3 in ho- lotype). Superciliaries longer than high, in three stacked rows; scales in middle row largest; superciliaries forming a laterally or dorsolaterally projecting crest extending be- yond orbit.
Nasals separated from rostral and first su- pralabial by 1-2 scales. Rostral separated from internasals by 1—2 rows of postrostral scales. Internasals about 15—23 in 3—4 vague rows (holotype, 18 scales in 4 rows). Super-
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 67
ciliary series terminates in a single large can- thal; canthal separated from nasal by 1-2 scales. Loreal region with 6—13 small to moderate-sized smooth scales with pits. Lor- ilabials in 3—6 rows. Enlarged median su- boculars 4—6; suboculars keeled (with sen- sory organs on keels). Temporal region with moderate-sized keeled or pointed scales, which imbricate anteriorly. Rostral much wider than high, little higher than adjacent supralabial. Supralabials 4—6, usually 4/4 (as in holotype).
Mental scale projecting slightly posteriorly beyond anterior margins of adjacent infrala- bials. Postmental series not well defined. In- fralabials 5—7, usually 5/5 (as in holotype, which has second infralabial on right par- tially divided). Gular scales smooth, increas- ing in size anteriorly and laterally; lateral gu- lars laterally imbricate; anterior gulars ante- riorly imbricate; medial posterior gulars im- bricate posteriorly, then grade into conical and pointed granules, with pointed granular scales to either side.
TYMPANUM: Ear canal moderately deep, opening higher than wide; a tuft of 4—6 large pointed scales at anterior margin; usually two tufts of large pointed scales near posterov- entral margin, sometimes merged into an obliquely elongated larger tuft. Internal rim of ear canal anteriorly lined with small smooth scales; dorsal, ventral, and posterior rim lined by smaller, smooth or weakly keeled, elongate scales.
NeEcK: Nape with anteriorly imbricate small pointed scales (large granules). Side of neck with pointed granules, anterior ones dorsally imbricate, becoming posteriorly im- bricate caudad. Two tufts of elongated spiny scales more or less in line with upper edge of ear opening; spines in anteriormost tuft largest; posteriormost tuft situated at termi- nus of oblique neck fold.
POCKETS AND FOLDs: Terminology follow- ing Frost (1992: 25-27): Antegular fold com- plete (faint in smallest juvenile), although sometimes interrupted medially by longitu- dinally raised ridge of granules extending posteriad from throat; antegular fold overlap- ping or lying just anterior to the gular fold. Antegular fold continuous with oblique neck fold, which terminates in posteriormost tuft of scales on side of neck. Lateral mite pocket
68 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Evnag wr
ESE
ptt Nee, pe ENC
hoa ae cD eal
a
UD Vee RES i" Pye ete Le
eer aa ee =e
is. ea 3
i
NO. 261
bes, Pin tte is feth ie eiicereroitee stom ees
a 1 ZS eset wae SS
an a, “4
ea
52 oe
cane Ser
Fig. 44. Dorsal, lateral, and ventral views of heads of adult male holotypes. A: Tropidurus panstic- tus, new species (EBRG 31306, 111 mm SVL). B: Tropidurus lumarius Donnelly and Myers (AMNH
136176, 98 mm SVL). Scale line = 10 mm for both.
present, under junction between antegular and oblique neck folds. Gular fold complete, its lateral mite pocket continuing under the antehumeral fold. Scales lining folds not (or but little) reduced in size, except for tiny smooth granules in mite pockets. Rictal fold present. Dorsolateral fold present, anteriorly continuous with antehumeral fold; granular scales on dorsolateral fold elongate and pointed, somewhat larger than other body scales. Weak ventrolateral fold present, its
scales little larger than other body scales. No mite pockets in axilla or groin.
Bopy: Middorsal crest of laterally com- pressed, long pointed scales directed poster- odorsally, starting about 3—7 scales behind interparietal and continuing onto tail; crest scales on body about 68—89 (* = 76.2), usu- ally decreasing in height posteriorly and then increasing somewhat on base of tail. Mid- dorsal crest present on all specimens, being best developed in adult males and weakly de-
2001
veloped in small juveniles, in which enlarged crest scales are prominent but are not ele- vated.
Body to either side of middorsal crest with slightly imbricate, acutely pointed scales di- rected posterodorsally. Paravertebral scales (from posterior edge interparietal to level of posterior margin of thigh) 153-175 ( = 169.7). Scales around midbody 143-164 (« = 157.0).
Ventral scales larger than dorsals and smooth; ventrals somewhat irregularly shaped, those on anterior venter usually pos- teriorly pointed and imbricate, often becom- ing quadrangular and subimbricate on pos- terior belly, in wavy transverse rows only; 30-36 scales across chest (* = 32.2); 95-116 transverse rows from gular fold to vent (x = 108.5). Umbilical scar still present in small- est juvenile (40 mm SVL).
Limss: Dorsal surface of upper arm with moderate-sized, imbricate, acutely pointed scales, much larger than granular body scales. Dorsal surface of forearm with larger, imbricate scales, weakly to sharply keeled, slightly mucronate; scales on lateral side of forearm with mucrones elevated and acutely pointed. Elbow scales smooth. Ventral side of arm with moderate-sized, imbricate, smooth scales, sometimes faintly mucronate. Top of hand with smooth (some slightly mu- cronate) imbricate scales, some scales be- coming keeled distally on fingers; palm with either smooth or weakly keeled scales, tend- ing to be distally elevated and mucronate. Subdigital lamellae single, tending to be bi- carinate on fingers I and H, becoming tricar- inate on fingers IIJ—V, distally changing to carinate. Subdigital lamellae of fourth finger 23-30 (* = 26.1, SD = 2.04).
Dorsal and posterior thigh with small pointed scales not much larger than granular body scales; raised scale points (and direc- tion of imbrication) projected posteriorly on top of thigh, dorsally on posterior thigh. An- teroventral thigh with larger smooth scales (some slightly mucronate) imbricating to- ward knee; ventral thigh with smooth scales posteriorly imbricate. Lower leg scales keeled and mucronate on upper surfaces, smooth on ventral side, all imbricate toward foot. Knee scales smooth, smaller than ad- jacent dorsal or ventral scales. Dorsal scales
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 69
of foot proximally small, essentially smooth, distally becoming larger, keeled, mucronate. Sole with smooth, slightly mucronate scales. Subdigital lamellae with carination similar to that of fingers. Subdigital lamellae of fourth finger 31-39 (« = 34.8, SD = 2.26).
TAIL: Enlarged middorsal crest scales con- tinuous from body crest, declining in size posteriorly; distinct on anterior third to half of tail. Dorsal and lateral caudal scales pos- terodorsally imbricate or subimbricate, acute- ly pointed on base of tail, becoming weakly mucronate distally. Subcaudal scales poste- riorly imbricate, smooth, becoming weakly keeled toward tip of tail.
COLORATION
In life (or shortly after death, fig. 45),”’ head of adults with black, light yellow, yel- lowish brown, or greenish markings, includ- ing a small to large pale spot on the parietal in all. A variably configured and broken black nuchal collar and a V-shaped scapular band (posterior to antehumeral folds). Dorsal surfaces of body greenish gray or light brown, with several ill-defined blackish blotches (posterior to scapular band) visible on middorsum of the lighter specimens. Dor- sum and sides overall densely and rather uni- formly speckled with small spots and dots of pale tan, pale yellow, or pale green.
Limbs colored like dorsa, with black or grayish brown bands on the paler ground col- or; bands on forelimbs tending to be darker and better defined than those on hind legs. Dark tail bands on the lighter ground color.
Underside of head with black markings on variable ground color of pale green, or bluish gray, or yellow on chin and orange on throat. Ventral surface of body and limbs bluish white, with an ill-defined but conspicuous suffusion of orange along ventrolateral sides of venter and under tail. Iris brown, tongue and mouth lining unpigmented.
A juvenile paratopotype (fig. 47) had a grayish head with brown markings, promi-
27 Color in “‘life’’ is based on transparencies and field- note descriptions made by daylight, from dead or stunned upland paratopotypes that had been recently col- lected with .22-caliber dustshot. A few lowland paraty- pes reached camp after dark and were only photo- graphed (fig. 46).
70 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
cat 4 P| tteés mele
ona
Fig. 45. Tropidurus panstictus, new species. Coloration of adult males before preservation. Top: Holotype (EBRG 3130) in dorsolateral view. Bottom: Dorsal view, left to right, EBRG 3130 (holotype), AMNH 147041, 147042 (paratopotypes).
2001 MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 71
Fig. 46. Tropidurus panstictus, new species. Dorsal and ventral views of lowland paratypes, EBRG 31336, AMNH 1470462 (lower).
72 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
NO. 261
Fig. 47.
nent black nuchal and scapular markings, and a light brown body with darker brown, round spots; the limbs and tail were light brown with darker brown bands.
In preservative, dorsal ground color gray- ish brown to blackish, with numerous light spots and flecks scattered over body, except light flecking inconspicuous in smaller ju- veniles. Collar and cape often emphasized by light internal spots and edging of light mark- ings. Three to five irregular dark middorsal blotches on body posterior to scapular band, sometimes (especially in juveniles) with ir- regular smaller blotches on sides; dorsal markings often vague. Light interorbital bar (anterior to supraocular scales) present in all, distinct in most (V-shaped in one); a less dis- tinct light bar across snout discernible in most. A light triangular to round, small to large parietal spot, usually distinct. Light spots or one or two light lines from eye to upper lip; a dark-edged light stripe or a few light lines on side of head from eye to ear; upper lip black with light spots (still showing yellow or light blue color in a few) distinct or vague; lower lip spotted black or gray.
Larger specimens with large black spot on base of throat, this spot smaller or absent in others. Underside of head rather densely marked with gray lines or spots tending to radiate forward and outward from black spot (when present) or from base of throat. A
Tropidurus panstictus, new species. A juvenile paratopotype (AMNH 147045, 40 mm SVL). Compare with color pattern of adults (figs. 45, 46).
midventral dark-edged, light gular line in some specimens, extending from middle of throat to base.
Forelimb, especially lower arm, with dis- tinct black bands and narrower light inter- spaces. Hind legs usually less distinctly banded. Venter and undersides of limbs and tail cream, often with light gray suffusion. Tail bands distinct in smaller individuals.
HEMIPENIS
The right hemipenis of a paratopotype (AMNH 147041), which had been partially everted in the field, was detached from the specimen and softened in a solution of KOH. After removing as much of the major retrac- tor muscle as possible, the organ was inflated with red-dyed petroleum jelly. Each lobe had to be manually everted separately, and each lobe ruptured at its delicate apex. The com- plete eversion otherwise was successful, as shown in figure 48.
The hemipenis is tiny considering the size of the adult male (119 mm SVL). When straightened from its curved position, the or- gan measures about 11 mm from the base to the tip of one lobe. It is divided for about half of its length. The sulcus spermaticus bi- furcates proximal to the crotch, and the sul- cus branches diverge in centrolineal orienta- tion and extend apparently to the tips of the
2001
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 73
Fig. 48.
Hemipenis of Tropidurus panstictus, new species Right organ of AMNH 147041 (parato-
potype, 119 mm SVL), in sulcate and asulcate view, X<7.9.
lobes. The basal stalk of the hemipenis lacks ornamentation; there are thin lateral alae on the sides of the stalk, extending from about halfway up the stalk and terminating at the base of each lobe.
The lobes are nearly fully calyculate, from shortly above the base of each lobe, with the calyces decreasing in size distally from the large basal calyces. The calyculate areas are confined to the lobes and are widely sepa- rated across the nude crotch.
COMPARISON WITH TROPIDURUS LUMARIUS
Tropidurus panstictus is most similar to T. lumarius from Cerro Guaiquinima, about 300 km east of the Yutajé—Corocoro massif, on the other side of the Sierra de Maigualida. Tropidurus panstictus 1s a larger species: Four adult male panstictus average 108.3 mm SVL (96-119 mm), compared with a mean of 88.5 mm (76-100 mm) for four adult male lumarius. One adult female pan- stictus measured 93 mm SVL, compared with a range of 76-79 mm (* = 77.1) for five adult female /umarius. The difference in SVL appears to be correlated with a relative- ly stockier head and body in T. panstictus.
Apart from size, it is hard to find consis- tent differences in the morphology of these two lizards. Differences seen when compar- ing one individual with another mostly dis- appear when intrapopulational variation is considered. For example, when comparing heads of the holotypes (fig. 44), there are vis- ible differences in the disposition of spinous scales on the anterior ear margin, and in alignment of the two spiny tufts on the neck, but these traits are variable in each species. The elongate spines on the granular neck and body scales seem to us to be longer in lu- marius than in panstictus, but any difference would be tediously difficult to quantify. One difference, although not an absolute one, in- volves the nature of the tiny scales in contact with the rear margin of the large interparietal. In 7. Jumarius, the interparietal usually is in contact with small pointed scales anterior to the middorsal crest, but sometimes a row of small smooth scales intervenes. In JT. pan- stictus, there usually are 2—3 rows of small smooth scales separating the interparietal from the small pointed scales, but sometimes there is only one such row, as in lumarius. Such a difference, not detectable in the draw-
74 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
ings (which cannot show elevated points or mucrones on granular scales at the magnifi- cation drawn), would seem both biologically and taxonomically insignificant.
Differences in living coloration are pro- nounced. Tropidurus lumarius is essentially a black lizard, with small yellowish spots and dots, which are frequently coalesced to form ill-defined transverse bars or rows of spots on the body and tail or narrow crossbars on the limbs. Adult males have golden yellow patches on the posterior belly, preanal area, and undersides of thighs, whereas females have a pink wash in the same area.
Tropidurus panstictus differs from the above in having a lighter ground color and (especially in adult males) a profuse dappling of small light spots and dots that do not co- alesce to form bars or rings (figs. 45, 46). Rather than golden yellow ventral patches as in lumarius, adult male panstictus are suf- fused with orange ventrolaterally and under the tail. Some specimens of panstictus have dark dorsal spots and other dark markings not observed in /umarius.
Both species share variable but similar black nuchal and scapular markings, but these tend to be obscured in Tropidurus lu- marius and more prominent in 7. panstictus. Direct comparison of the hemipenes suggests that the organ of T. panstictus (fig. 48) is virtually identical to that of 7. lumarius (Donnelly and Myers, 1991: fig. 18), al- though the latter organ is less well prepared. The earlier published drawing for 7. /umar- ius shows only representative calyces on the lobes, which prove to be incompletely evert- ed; the calyculate areas resemble those in T. panstictus (fig. 48).
REMARKS
We described Tropidurus lumarius from 14 specimens that we collected on the sum- mit of Cerro Guaiquinima in 1990 (Donnelly and Myers, 1991) on an expedition spon- sored by the Fundacion para el Desarrollo de las Ciencias Fisicas, Matematicas y Naturales (FUDECD); this expedition included different groups of scientists working on various pro- jects during different times, with no coordi- nation between groups. We were not in- formed that a German group had been sched-
NO. 261
uled to work on amphibians and reptiles be- fore us; later, we were given reason to believe that their interests were not system- atic. We were unaware that a professional herpetologist was involved with the German effort, and we certainly made a mistake in not pressing for more particulars on the Ger- man party and in not trying to make con- tact—we probably missed a chance at fruitful collaboration. But we thought no more about it until we saw in late March, 1991, their collection report that had been published a month earlier (Magdefrau et al. 1991). We were surprised by their report—as they doubtless also were when our own report ap- peared several months later on June 28 (Don- nelly and Myers, 1991). A short paper by Magdefrau (1991) appeared a few days after that, with the description of Plica nigra, a junior synonym of 7. lumarius (synonymy by Frost, 1992: 5). This hapless duplication of effort, and a comparison of collection re- sults, is discussed in a postscript to our 1991 paper. The only reason for bringing it up again emanates from insinuations made re- cently by Gorzula and Sefnaris (1999: 153). Among other things, these authors com- plained that Gorzula had collected the first specimens of Tropidurus lumarius/nigra six years before the 1990 FUDECI expedition, and that unnamed persons in FUDECI were aware that he had been planning to describe the new lizard. Gorzula and Sefaris (1999: 153) asserted that “Certainly by late 1990, staff of both the AMNH and the ZSM were aware of the series of [Gorzula’s] specimens collected in 1984 and 1985.”’ This statement is fabricated! We did not learn of Gorzula’s earlier collection of this lizard on Cerro Guaiquinima until much later, long after the 1991 publications (Magdefrau et al., 1991; Magdefrau, 1991; Donnelly and Myers, 1991), and we have no reason to believe that our German colleagues knew of it either. In 1995, Gorzula first complained about the matter in a letter to Myers; in that letter, Gor- zula did not mention FUDECTI, but twice laid the blame on unidentified persons in TER- RAMAR, where “I spent a whole morning giv- ing them maps of Guaiquinima and telling them where and what I had collected”’ (letter, Gorzula to Myers, August 24, 1995). In re- sponse, Gorzula was advised that (1) it had
2001
been not a TERRAMAR but a FUDECI-spon- sored expedition, and that we had had no as- sociation with TERRAMAR in 1990; (2) his in- formation had not been passed on to us (by either institution); and (3) that his prior man- uscript name would have been honored had we known about it.”8 There is much water under such bridges.
Tropidurus panstictus is not endemic to Cerro Yutajé, but occurs also on neighboring Cerro Guanay (Myers and Donnelly, in pro- gress).
KEY TO VENEZUELAN TROPIDURID LIZARDS
With the present description of Tropidurus panstictus, eight species of tropidurid liz- ards—all members of the tribe Tropidurini— are known from eastern Venezuela. One spe- cies is in the monotypic genus Uranoscodon, which was considered by Frost (1992) to be the sister taxon of all other genera of Tropi- durini. The remaining seven species are cur- rently assigned to Tropidurus following Frost (1992), who substantially extended the concept of this genus by placing Plica and Uracentron in its synonymy.
Three of the eight tropidurines in Vene- zuela are endemic, but the others are wide- ranging lizards that are illustrated by color photographs and line drawings in Avila-Pires (1995). The following key is based on a few easily discerned characters. Asterisks indi- cate species that have been found on tepuis; a double asterisk denotes species that appear to be endemic to particular tepuis.
1. Tail ‘normal’? (compressed or not)—slender and gradually tapering, longer than snout—
VETI Demet = COV ee dasetch tet sel ick eee one 2,
— Tail depressed and much shorter than SVL, wide for most of its length and armored with whorls of spiny scales; Amazonian fowland: forest <5 fk wee oss Tropidurus [Uracentron] azureus (Linnaeus)
2. No projecting tufts of elongate, spinous scales ONPNEC KP ee ee ee ae at temeel ate. 8 3
°8 Two-page letter, Myers to Gorzula, February 2, 1996; receipt acknowledged by Gorzula on July 27, 1996; also unanswered letter, Myers to Gorzula, March 22, 2000, with copies attached of pertinent 1991 letters to FUDECI and 1991 and 1993 letters between Myers and Andreas Schliiter, Stuttgart. (Correspondence in AMNH Dept. Herpetology Archives, Myers collection, Gorzula folder.)
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF JS
— Spiny tufts present on neck ........... 5
3. No middorsal crest; widespread in open coun- try and dry forest Ae ae eae Tropidurus hispidus (Spix)*
— Middorsal crest of enlarged scales present; lowland forest in Amazonas and Bolivar
ey
4. Body and tail rounded
Pe Se Tropidurus umbra (Linnaeus)
— Body and tail laterally compressed ......
Uranoscodon superciliosus (Linnaeus)
5. Prominent middorsal crest of enlarged scales;
gular fold medially complete, with mite
pocketinefold Vaterally.. :: 4 840 wsudeeen 5535, 6
— Middorsal crest absent; gular fold incomplete
medially, lacking a mite pocket in fold; Au-
yantepui .... Tropidurus bogerti Roze**
6. Mainly grayish to black in life, without a bold
dark pattern (dark bands lacking, incon-
spicuous or very fragmented on body); a
tuft or line of markedly enlarged scales
projecting from anterior edge of ear open-
ATS WA onde Pree Pot Ree Pons oe cakes 7
— Mainly green in life, with conspicuous dark
brown or blackish bands (often broken) on
body and limbs; enlarged scales not pro-
jecting from anterior edge of ear opening; geographically widespread ..........
es oe Tropidurus plica (Linnaeus)*
7. Mainly black in life and death, sparsely
marked with small yellowish spots; SVL to
about 79 mm in females, 100 mm in males;
Cerro Guaiquinima ....... Tropidurus
lumarius Donnelly and Myers**
— Lighter gray or brownish with more conspic-
uous dark markings, and (in adults) with
dense and uniform distribution of small
spots or speckling of pale tan, yellow, or
green; larger, SVL to at least 93 mm in fe-
males, 119 mm in males; Yutajé—Corocoro
massif, Cerro Guanay, and perhaps adja-
Céntsareas: 45.52 33, Tropidurus panstictus,
new species*
ey
SNAKES
FAMILY COLUBRIDAE Liophis species?
MATERIAL: No specimen collected. A small snake of Liophis size was seen by John Daly one day at the type locality of Colos- tethus undulatus (q.v.), at 1750 m elevation. The snake slipped into a forest stream and there were no further sightings despite inten- Sive search.
76 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
Thamnodynastes corocoroensis Gorzula and Ayarzagiiena
Thamnodynastes corocoroensis Gorzula and Ay- arzagtiena, “1995” [1996]: 6-8, fig. 1 (draw- ings of dorsal and lateral head scalation). Ho- lotype MHNLS 11376 from Tepuy Corocoro, 2150 m, Estado Bolivar [Estado Amazonas], Venezuela (05°46'N, 66°11'W), collected on November 10, 1987.
This recently described snake, known only from the holotype, is from 2150 m elevation in the high northwestern quadrant of Cerro Corocoro. The type locality, originally given as being in Bolivar state, was corrected to Amazonas state by Gorzula and Sefaris (1999: 180, 255 [locality H-017]).
COMPARISONS
The unique specimen of Thamnodynastes corocoroensis comes from only some 30 km WNW of neighboring Cerro Yavi, type lo- cality of the almost simultaneously described Thamnodynastes yavi Myers and Donnelly. Based solely on comparison of the original descriptions, the holotype of 7. corocoroen- sis differs from the type specimens of T. yavi in aspects of coloration in life and in relative tail length.
Thamnodynastes corocoroensis was de- scribed as having in life a whitish vertebral line (presumably confined mainly to the ver- tebral scale row), as well as two dorsolateral light lines along scale rows 5 and 6.”? In con- trast, 7. yavi in life (Myers and Donnelly, 1996: fig. 25) lacked a whitish vertebral line and had suggestions either of dark dorsolat- eral lines or rows of vague dark spots, which extended posteriorly from two parallel dark stripes on the nape and which tended to fuse again as a pair of solid lines on the rear of
°° The description of the two pale “‘paravertebral’’ (i.e., high lateral or dorsolateral) lines, as given, would normally mean that the line on each side occupies all or adjacent parts of scale rows 5 and 6 along the length of the body. However, since there is scale-row reduction, the authors more likely described line position from only one place on the body. There must be a posterior drop in position of the lateral stripe unless scale-row reduc- tion involves paravertebral or other dorsal scales above row 6. Method of reduction was not specified for Tham- nodynastes corocoroensis, but only lateral rows (usually 3 + 4) have been implicated for 7. chimanta, T. duida, and T. yavi.
NO. 261
Fig. 49. Thamnodynastes yavi, a paratopotype (AMNH 1413426) from Cerro Yavi; for same specimen in life, see color photograph in Myers and Donnelly (1996: fig. 25, upper). Note that this is not a full dorsal view; the coiled body is in- clined to its left, giving a nearly lateral view that includes only part of the dorsum. The light gray areas, where stratum corneum has been lost, show hints of pale vertebral and dorsolateral lines said to be evident in life in T. corocoroensis from Cer- ro Yutajé (Gorzula and Ayarzagiiena, “£1995” [1996]: 8).
the body and tail. Thamnodynastes yavi in life also had slightly darker (grayish brown) sides that were edged above with a narrow blackish line extending mainly along row 5 anteriorly and row 4 posteriorly.
There would seem to be little resemblance in the dorsal colorations of T. corocoroensis and 7. yavi in life. After preservation, how- ever, repeated handling for examination eventually causes loss of the stratum corne- um in most snakes, and specimens of Tham- nodynastes then become a lighter gray. The available specimen (AMNH 1411326) of Thamnodynastes yavi has faded somewhat after four years in preservative, turning ligh- ter brown over areas of still-existing stratum corneum, and gray where the corneal layer has been lost (fig. 49). This specimen was reexamined to see if it might have indications of the pale vertebral and dorsolateral lines
2001
that were ascribed to living T. corocoroensis. Although there is a reduction in density of melanophores, the vertebral row in the spec- imen of yavi is only very slightly paler than the paravertebral rows; it does not give the impression of being a pale line or narrow stripe. Also owing to a reduction in mela- nophores, there is a paleness above the vague blackish lateral line, mainly along rows 5 (anteriorly) and 4 (posteriorly). Under certain conditions of lighting, when the specimen is immersed in alcohol, this pale edging seems to include row 6 on the anterior part of the body, as well as the upper part of row 5; one can then imagine a pale stripe on rows 5 and 6. These linear areas of reduced melanophore density seem likely to be either precursors or remnants of the pale lines described for T. corocoroensis. Thamnodynastes yavi, appar- ently unlike T. corocoroensis, has vague dor- solateral black markings rising from black nape stripes and posteriorly fusing to form a pair of paravertebral black lines; these black markings are more evident after loss of the stratum corneum (fig. 49).
Thamnodynastes yavi seems to be notice- ably different from 7. corocoroensis in hav- ing anteriorly distinct dark ventrolateral stripes. Ventral or ventrolateral striping is not mentioned in the description of 7. coroco- roensis, but the ventral surfaces of the two taxa seem otherwise to be similarly colored.
The female holotype of Thamnodynastes corocoroensis was stated to be 258 mm in total length (“‘longitud total’’), 74 mm in tail length (tip of tail possibly incomplete)—the tail accordingly being at least 28.7% of total length. Relative tail length therefore may be greater in T. corocoroensis than in T. yavi, for which lengths of 22.9 and 23.3% were recorded for two females and 24.1% for one male. If these two named taxa represented only one species, the range in variation of nearly 6% (22.9—28.7%) for females (and a minimum overlap of 4.6% between sexes) is conceivable, but unlikely to be represented in a sample of only four specimens. For com- parison, relative tail length for 10 Thamno- dynastes chimanta was 18.5—21.0% in three females and 18.6—25.6% in seven males, with means of 19.93% @ and 21.60% ¢ (as summarized by Myers and Donnelly [1996: 47] from available data).
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 77
REMARKS
Considering the geographic proximity of the type localities on Cerro Corocoro and Cerro Yavi, we had rather expected these snakes to be conspecific, although the pre- ceding comparisons suggest that they are dis- tinct. Nonetheless, direct comparison of specimens would be useful. Except for one paratopotype (AMNH 141342) of T. yavi, the type specimens of both taxa are now in Ve- nezuelan institutions.*°
The possible conspecificity of two addi- tional species named from neighboring tepuis (Duida/Marahuaka)—Thamnodynastes mar- ahuaquensis Gorzula and Ayarzagtiena, and Thamnodynastes duida Myers and Donnel- ly—is perhaps more problematic inasmuch as each is known only from a single speci- men. The holotype of 7. duida is a male 455 mm in total length (tail = 22.2%), whereas the holotype of 7. marahuaquensis is a fe- male 396 mm in total length (tail = 23.0%). The specimen of T. duida has only 17 scale rows on the anterior half of the body (17-— 17-15), whereas T. marahuaquensis has 19 rows up to the level of ventrals 87/88 (i.e., the standard scale-row formula = 19—19— 153'). Each has four ventral stripes (a com-
30 Along with the holotype and a paratopotype, there is an additional specimen of Thamnodynastes yavi in the Museo de la Estacién Biol6gica de Rancho Grande at Maracay. All specimens were taken on the 1995 Amer- ican Museum—TERRAMAR Expedition, but the additional specimen was caught when we were on another tepui, and it was accidentally transported to Maracay in a con- tainer of alcohol-preserved bats. We were unaware of its existence until February 1995, after delivering half our collection and a copy of our in-press manuscript to the PROFAUNA Office in Caracas.
3! The original description of 7. marahuaquensis gave a formula of “‘19—17—15” to indicate reduction. How- ever, the shorthand scale reduction formula most com- monly used shows counts (1) about a head length behind the head, (2) at midbody, and (3) a head length or less anterior to the vent (e.g., J. Peters, 1964: 312). Thus, the standard formula is 19-19-15 for 7. marahuaquensis, which implies that reductions to 17 and to 15 rows take place posterior to midbody. “‘Midbody” can be estimat- ed if dorsal-scale reductions are linked to ventral counts. For example, in 7. marahuaquensis reduction from 19 to 17 rows was said to occur at the level of ventrals 87/ 88—well past midbody at roughly 64% of the distance from the first to the last (137th) ventral plate. These differences in scale-row annotation appear to have con- fused Kornacker (1999: 140), whose unsuccessful key to Venezuelan Thamnodynastes shows only one species
78 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
mon character in the genus) and generally seems to resemble the smaller 7. chimanta more than the geographically closer T. cor- ocoroensis and T. yavi.
Any revisionary work involving these taxa unfortunately will have to contend with the fallacious “‘Junio 1995” publication date of Gorzula and Ayarzagtiena’s paper, which ap- peared as number 6 of the Publicaciones de la Asociacién de Amigos de Donana (Sevil- la). In acknowledging receipt of this paper on July 1, 1996, one of us expressed regret at having overlooked a year-old publication (letter, Myers to Ayarzagiiena). Soon after- wards, however, herpetologist Ignacio de la Riva, a member of the journal’s editorial committee, visited the American Museum in August 1996, and said that he had not pre- viously seen that issue and had been unaware of its existence. Subsequent query revealed that the paper also had been received in Bo- gota in the summer of 1996, but that inves- tigators were still unaware of it at such her- petological centers as the Paris Museum, the U.S. National Museum of Natural History, and the Museum of Comparative Zoology.
Attempts to find holding libraries, from which dates of receipt could be obtained, were unsuccessful. According to John Cadle (in litt.), the reference librarian at the Mu- seum of Comparative Zoology reported to him in August 1996 that “‘Neither the library databases nor the directories of associations list them [i.e., La Asociaci6n de Amigos de Dofiana]. I tried searching the Spanish h- brary catalogues directly, but didn’t have any luck.”’ More recently (December 1999), a computer search was performed by reference librarians at the American Museum, with the following results:
We looked for the publication in the National Library of Spain, other European libraries, uni- versity libraries, and OCLC [Online Computer Library Center, Inc.]. We used the [name of the journal] and the author and title of the article,
<
(T. yavi) as having 19 midbody scale rows, whereas that count also pertains to 7. corocoroensis, T. marahua- quensis, and probably 7. strigilis. Northern Venezuelan specimens assigned to the last species (Roze, 1966: 229-230) should be compared with Thamnodynastes gambotensis, described from the Caribbean lowlands of northern Colombia (Pérez-Santos and Moreno, 1989).
NO. 261
on the off-chance that the article itself was in- dividually catalogued. This gave us no results. We accessed the ISSN site online and deter- mined that the ISSN number [1132-8398] is a registered number, but after re-running the same search with the ISSN no. (which should be a sure way of tracking it down, as numbers are less messy than text), we still came up with no hits. Unfortunately, we were not able to ac- cess the University of Seville’s online catalog, as it is only available to the University of Se- ville community. We were hoping that since the publication is put out in Seville, they would have a copy.”
Evidently not! During a trip to the Univer- sidad de Sevilla in February 2000, Ms. Tracy Hogan inquired about the publication, and her colleagues there initiated a search, only to find that, as of March 3, 2000, the general library had no records of the Publicaciones de la Aso- ciaci6n de Amigos de Donana, and that it had not been found during a search by the librarian in the School of Biology. Attempts were being made by Universidad de Sevilla representa- tives to find out more about the Seville-based Asociacion de Amigos de Dofiana and to ob- tain their publications for the library.
Therefore, it would appear that issues of the Publicaciones de la Asociaci6én de Ami- gos de Dofana are not routinely sent to h- braries, and that the separately issued num- bers are distributed mainly by the authors.
Darrel Frost, who has tried to document time of publication for various journals that frequently bear erroneous dates, wrote to the President of the Asociaci6n in March 1997, asking if there were records for precise dates of publication. He received a polite response that, however, did not address the question. Dr. de la Riva (in litt.) said that issue number 6 was sent out around June 1996, but that nobody could provide a particular date of publication.
In summary, no evidence could be found that issue number 6 of the Publicaciones de la Asociacién de Amigos de Dofiana was published any earlier than some unknown time in June 1996—a year later than the pub- lished date (“‘Junio 1995’’). In lacking an un- biased, objectively determined date, the day of publication for taxonomic purposes must
32 Some issues have been printed in Venezuela, but number 6 was printed in Spain.
2001
be assigned according to the last-day-of-the- month rule (i.e., June 30, 1996) of the Jnter- national Code of Zoological Nomenclature (International Commission, 1999: art. 21.3.1 and 21.4).
DISCUSSION
Our sampling of the the amphibians and reptiles of the Yutajé—Corocoro massif was not expected to provide a complete faunal list, but the results do reinforce several emerging generalities about the herpetofauna of Pantepui: (1) Tepuis have relatively de- pauperate herpetofaunas; (2) neighboring te- puis are likely to have significantly different faunas; (3) tepui endemics outnumber wide- spread highland species; (4) some endemic species have widespread counterparts, the remnants perhaps of a once widespread tepui fauna; and (5) lowland species find their way onto tepuis in an irregular, unpredictable manner. Following is a brief elaboration of these points.
1. Species diversity clearly is limited on the tepuis, perhaps a general reflection of bi- otic impoverishment on poor soils. Beyond that, experience suggests that sizes of tepui faunas are (as would be expected) positively correlated with area and negatively correlat- ed with elevation (Myers and Donnelly, 1997: 64). The recorded number of species in a given tepui herpetofauna also reflects collecting effort. In terms of collecting effort and actual area sampled, the size of our pre- sent collection from the Yutajé—Corocoro massif is very roughly comparable with sev- eral other collections that we have made:
Cerro Guaiquinima: 15 spp. from 1030-1150 m (12 days)
Pico Tamacuari: 12 spp. from 1160-1460 m (6 days)
Yutajé—Corocoro: 9 spp. from 1220 m and 1700— 1750 m (7 days)
Cerro Yavi: 7 spp. from 2150 m (5 days)
Increased collecting effort over a greater area naturally yields more species, with known faunas of 15—25 species on the best-sampled large tepuis (Myers and Donnelly, 1997: 64).
2. Data showing profound differences be- tween the herpetofaunas of neighboring te- puis are only just accumulating. The best documented example is that of Auyantepui
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF Hie)
and Chimanta—these large tepuis are contin- uous at the 1000 m contour and their high summits (>2000 m) stand less than 50 km apart. However, their summits share less than 50% of the families represented (5 of 11), only 38% of the genera (9 of 24), and only about 15% of the known species (about 4 of 26 spp. total), as tabulated by Myers (1997)?
The northwestern tepuis (Yavi, Yutajé— Corocoro, Guanay) provide additional data on differences among neighboring tepuis. The small summit of Cerro Yavi is less than 20 km airline distance from equivalent ele- vation on the Yutajé—Corocoro massif. Fur- thermore, a system of ridges connects the northern bases of these mountains, so past (but probably not current) faunal interchange seemed likely and evidently did occur. Of the combined fauna of 14 species, three (21%) are known to be shared. The lack of more extensive overlap between neighboring Yavi and Yutajé—Corocoro can be explained in part by ecological differences. Cerro Yavi, for example, lacks summit streams, which at the Yutajé camp yielded two frogs and a sighting of a riparian snake, and the summit of Yavi is perhaps too high for the lizard Tro- pidurus. Nonetheless, the snake genus Tham- nodynastes appears to be represented by dif- ferent species on Yavi and Yutajé—Corocoro, and our samples of two of the three shared species seem to show evidence of geographic differentiation! Eleutherodactylus yaviensis seems to be an earless frog on Cerro Yavi, whereas a flimsy concealed tympanum is present in part of the small Yutajé sample. Furthermore, the meager data suggest the ex- istence of population differences in stripe width and size of median gular scales in the lizard Prionodactylus goeleti.
Study of our collection from Cerro Guan- ay is in progress. The new lizard Tropidurus panstictus appears to be the same species as one on Cerro Guanay, but otherwise there is
33 Chimanta fauna based mainly on the works of Gor- zula (1992) and Roze (1958b). Auyantepui fauna based mainly on results of the 1937-1938 Phelps Venezuelan Expedition of the American Museum of Natural History (Roze, 1958a) and the 1994 Robert G. Goelet American Museum—TERRAMAR Expedition to Auyantepui (Myers and Donnelly, in prep.). See Myers (1997: 1—2) for ac- knowledgment of other collectors on Auyantepui.
80 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 261
eps elie Geo 5 poets ;
Fig. 50. A ridge corridor between the northern base of Cerro Yavi and the northeastern base of Cerro Yutajé. Viewed from a helicopter, looking northeastward. The ridge crests were very roughly estimated from the air as ranging from perhaps 600 m to more than 1000 m above sea level (highest elevations of Cerro Yavi and the Yutajé—Corocoro massif >2100 m). Although the ridges between these neighboring tepuis are heavily forested, there is no continuity with higher elevation forest—at least on Cerro Yutajé, where rocky cliffs and scrubland intervene. (Photograph by C. W. Myers, May 8, 2000)
virtually no overlap between Guanay and the other northwestern tepuis. Cerro Guanay lies close to Cerro Corocoro, but is separated by a lowland valley (fig. 1).
3. Excluding various lowland species that invade tepuis (see below), endemic species of amphibians and reptiles are much more likely to be found than widespread highland species. Of 14 species in the combined fau- nas of Cerro Yavi and Yutajé—Corocoro, at least 11 (79%) appear to be endemic to those tepuis, including the three shared species. In confusing ontogenetic variation with noncor- related individual variation, Gorzula and Sefiaris (1999) expressed the opinion that one of the three shared endemic species (Prionodactylus goeleti) is identical with one (P. phelpsorum) named earlier from the tepui Sarisarifiama (see page 63). They bolstered this idea with the following scenario:
Prionodactylus phelpsorum {including P. goe- leti] is known from the cerros Corocoro, Yutajé and Yavi in the State of Amazonas, and from Cerro Jaua (the type locality) in the State of Bolivar, at elevations ranging from 1800 to 2100 m asl. The distance between Cerro Jaua and the group of cerros Corocoro, Yutajé and Yavi is about 180 km in a straight line and about 300 km following the connecting water- sheds. At no point does the watershed drop be- low an elevation of 950 m asl. Much of the
intervening terrain reached [reaches] elevations of 2000 m or more. (Gorzula and Sefiaris, 1999: 140)
Although conceivable, such a broad distri- bution would be highly unusual for a high- land reptile in Pantepui. It is contraindicated by variation as presently understood, includ- ing indication of differentiation between pop- ulations of P. goeleti on neigboring tepuis. Samples are regrettably small.
4. Tepui endemics are likely to have coun- terparts** on other tepuis, as shown for ex- ample by the dwarf members of the snake genus Thamnodynastes, now represented in Pantepui by five named species, including one each on Cerros Corocoro and Yavi. If there is a relatively old and widespread high- land fauna of amphibians and reptiles in Pan- tepui, it is represented by such groups as the the dwarf Thamnodynastes complex, the Or- eophrynella complex, the Phenacosaurus neblininus complex, and the endemic Riola- ma, assuming for the moment that all of these comprise monophyletic highland groups rather than separate invasions from
34 This is a purposely vague term that we use for con- generic species that we judge to be phenotypically and ecologically similar. Phylogenetic closeness is probable, but cladistic analyses are lacking and degree of relat- edness is unknown.
2001
the lowlands. Current known distributions seem very spotty, as would be expected by random extinction events. However, the slow ongoing investigation of tepui herpetofaunas will eventually fill in some of the gaps.
We have speculated on aspects of the problem elsewhere (Myers and Donnelly, 1996: 53-54; 1997: 63-67), but additional faunal sampling is needed much more than additional hypothesizing. The highland her- petofaunas of Pantepui remain very poorly known, and we suspect that many more en- demics remain to be discovered. Our current report of a new species of microteiid lizard on Cerro Yutajé was unpredictable but scarcely surprising—the only thing truly re- markable is that a new generic name (Ader- cosaurus) was needed to accommodate it, since there are few endemic genera in high- land Pantepui (Riolama is the only other en- demic genus of reptiles). This finding under- scores the point that the tepuis are poorly ex- plored in a herpetological sense.
5. Lowland species doubtless invade all te- puis, although they may not necessarily reach the highest summits (e.g., none is known on Cerro Yavi), or, if they do, they may not es- tablish permanent populations there. Avail- able habitat for vertical dispersal can be an important factor, as on the huge but relatively low Cerro Guaiquinima, which has a strong lowland component in the herpetofauna (Donnelly and Myers, 1991; Magdefrau et al., 1991). On occasion, species at first re- garded as “‘tepui species”’ later come to be seen as widespread animals with extensive elevational ranges (e.g., the lizard Neusticu- rus racenist; see Myers and Donnelly, 1997: 61, 65). Lizards of the Anolis chrysolepis complex are perhaps the most frequent in- vaders of tepui habitats.
Although lowland invaders are expected, they can seldom be successfully predicted. For example, there was no way of anticipat- ing the presence of Pseudopaludicola llanera at 1220 m on Cerro Corocoro—the first re- ported tepui occurrence of this lowland ge- nus of miniature frogs. However, discovery of the new lizard Tropidurus panstictus on Cerro Corocoro led to the expectation that T. hispidus or the macroteiid Ameiva ameiva might also occur, since both widespread low- land species occurred in sympatry with Tro-
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 81
pidurus lumarius on Cerro Guaiquinima. But neither was found, although another macro- teiid (Kentropyx sp.) occurred prominently with Tropidurus panstictus on Cerro Guan- ay! Appearance of lowland species in tepui faunas seems to be regulated mostly by op- portunity and chance, and we see no general pattern emerging.
ACKNOWLEDGMENTS
The 1995 expedition was made possible by the generosity and commitment of Mr. Robert G. Goelet, former president and Chairman Emeritus of the Board of Trustees of the American Museum of Natural History.
As acknowledged in our earlier report, the expedition would not have been successfully concluded or even started without the essen- tial help and splendid collaboration from many Venezuelan friends and colleagues. For help with permits and contracts, we thank Dr. Armando Michelangeli Ayala, President of Fundacion TERRAMAR; General Francisco Loreto, then President of the Instituto Na- cional de Parques (INPARQUES); and several officials in the Servicio Aut6énomo para la Proteccion, Restauraci6n, Fomento y Racion- al Aprovechamiento de Fauna Silvestre y Acuatica del Pais (PROFAUNA)—especially former Director-General Dr. José Luis Mén- dez Arocha, Directora-General Lic. Mirna Quero de Pefia, and Ms. Magaly Ojeda C., Directora de Fauna Silvestre.
Mrs. Kathleen D. de Phelps, of the Colec- cidn Ornitologica Phelps, expressed great in- terest in the work and expedited it in every possible way.
Expedition planning was facilitated in September 1994, when Dr. Luis Arturo Ay- ala, of TERRAMAR, piloted his Piper Seneca in an initial reconnaissance that gave Mich- elangeli, Guerrero, and Myers a close look at the northwestern tepuis. We gratefully ac- knowledge the Venezuelan Air Force (FAV) for subsequently providing a military trans- port plane for moving personnel and supplies to base camp on February 4, 1995, and for evacuating the expedition on March 4. The expedition base was Yutajé, a backcountry fishing camp run by José Félix Raggi.
Owing to limits placed by PROFAUNA on the number of persons in a single camp, the
82 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
expedition had two or three camps simulta- neously open in the northwestern tepuis dur- ing February 4—March 2. It was an extremely efficient operation owing to logistical plan- ning by expedition leader Armando Miche- langeli, and support from helicopter pilot Ri- cardo Trevisi and mechanic Juan Vasconce- los. Their experience allowed the scientific teams to establish five wilderness camps on three tepuis in a month’s time.
We were air-lifted on February 24 from Cerro Yavi to the 1700-m Cerro Yutajé camp, evacuating it on March 2. The Yutajé camp had been first opened on February 11 by TERRAMAR members José Clavijo and José Luis Garcia (Museo del Instituto de Zoologia Agricola, Univ. Central de Venezuela, Ma- racay), Miguel Lentino and the late Gilberto Pérez-Chinchilla (Coleccion Ornitologica Phelps), Juan Carlos Navarro (Instituto de Zoologia Tropical, Univ. Central de Vene- zuela, Caracas), with expert support from TERRAMAR staff members Williams Sarmien- do and Aleizer Zerpa.
Our campmates in the 1700-m camp on Cerro Yutajé included George FE Barrow- clough and Paul R. Sweet (AMNH Dept. Or- nithology), Francisco Delascio Chitty and Aniusk Kazandjian (Herbario Nacional de Venezuela, INPARQUES), John W. Daly (Na- tional Institutes of Health and AMNH Dept. Herpetology), Ricardo Guerrero (Instituto de Zoologia Tropical, Univ. Central de Vene- zuela, and TERRAMAR), and Petia Alcécer and Adriana Sanchez H. (TERRAMAR).
For comments on various parts of the manuscript or for other help, we are grateful to George FE Barrowclough, John E. Cadle, Charles J. Cole, Julian Faivovich, Linda S. Ford, Darrel R. Frost, Taran Grant, Ricardo Guerrero, Dennis M. Harris, David A. Kizir- ian, Ignacio de la Riva, Paul R. Sweet, and Richard G. Zweifel. Drawings were prepared by Patricia J. Wynne. Photographs of pre- served specimens are the work of Peter Goldberg (photographs of living specimens and landscapes were taken by the first au- thor). Thomas Trombone prepared a radio- graph of the lizard Adercosaurus.
For their help in trying to determine the actual date of publication of a problematic paper, we thank Michelle Anastasia and Meghan Manahan, Reference Librarians in
NO. 261
the American Museum, and Mary Sears, Reference Librarian in the Museum of Com- parative Zoology at Harvard University. Their investigations were extended on a visit to Seville by Tracy Hogan (Montclair State University), who received valuable assis- tance from Drs. Isabel L. Calderon and Ig- nacio Martinez of the Universidad de Sevilla.
REFERENCES
Avila-Pires, Teresa C. S.
1995. Lizards of Brazilian Amazonia (Repti- lia: Squamata). Zool. Verh. 299: 706 pp.
Barbour, Thomas
1928. New Central American frogs. Proc. New England Zool. Club 10: 25-31 + 4 color pls.
Boulenger, George Albert
1879. Quelques observations relatives a la forme de la pupille et a la coloration de Tiris chez certains batraciens. Bull. Soc. Zool. France 4: 129-131. Catalogue of the lizards in the British Museum (Natural History), 2nd ed. Vol. 2, xili, 497 pp. + 24 pls. Reptiles and Batrachians. Jn E. R. Lan- kester, Report on a collection made by Messrs. E V. McConnell and J. J. Quelch at Mount Roraima in British Guiana. Reptiles. Trans. Linn. Soc. London, Zool., ser. 2, 8(2): 53-56 + pl. 5: Camp, Charles L.
1923. Classification of the lizards. Bull. Am. Mus. Nat. Hist. 48(11): 289-481. [Re- printed in facsimile in 1971 by Society for the Study of Amphibians and Rep- tiles, with a new Preface by Camp and a new introduction by G. Underwood. ]
Cole, Charles J., Herbert C. Dessauer, Carol R. Townsend, and Margaret G. Arnold
1990. Unisexual lizards of the genus Gym- nophthalmus (Reptilia: Teiidae) in the Neotropics: genetics, origin, and sys- tematics. Am. Mus. Novitates 2994: 29 Pp.
Dixon, T. H.
1991. An introduction to the global position- ing system and some geological appli- cations. Rev. Geophys. 29(2): 249-276.
Donnelly, Maureen A., and Charles W. Myers
1991. Herpetological results of the 1990 Ve- nezuelan expedition to the summit of Cerro Guaiquinima, with new tepui
1885.
1900.
2001
reptiles. Am. Mus. Novitates 3017: 54
pp. Donnelly, Maureen A., Roy W. McDiarmid, and Charles W. Myers
1992. A new lizard of the genus Arthrosaura (Teiidae) from southern Venezuela. Proc. Biol. Soc. Washington, 105(4): 821-833.
Duke-Elder, Sir Stewart
1958. The eye in evolution. System of oph- thalmology, vol. 1, xvi + 843 pp. St. Louis: C. V. Mosby.
Estes, Richard
1983. Sauria terrestria, Amphisbaenia. Handb.
Palaoherpetol. 10A: xxii + 249 pp. Estes, Richard, Kevin de Queiroz, and Jacques Gauthier Phylogenetic relationships within Squa- mata. In R. Estes and G. Pregill (eds.), Phylogenetic relationships of the lizard families: 119-281. Stanford, CA: Stan- ford Univ. Press. Flores, Glenn
1985. A new Centrolenella (Anura) from Ec- uador, with comments on nuptial pads and prepollical spines in Centrolenella. J. Herpetol. 19(3): 313-320.
Frost, Darrel R.
1992. Phylogenetic analysis and taxonomy of the Tropidurus group of lizards (Igu- ania: Tropiduridae). Am. Mus. Novita- tes 3033: 68 pp.
Gosner, Kenneth L.
1960. A simplified table for staging anuran embryos and larvae with notes on iden- tification. Herpetologica 16(3): 183- 190.
Gorzula, Stefan
1992. La herpetofauna del macizo del Chi- manta. Jn O. Huber (ed.), El macizo del Chimanta: 267-280 + 304-310 (= photographs 152-171). Caracas: Oscar Todtmann Editores.
Gorzula, Stefan, and José Ayarzagiiena
“1995” [1996]. Dos nuevas especies del gé- nero Thamnodynastes (Serpentes; Co- lubridae) de los tepuyes de la Guayana Venezolana. Publ. Asoc. Amigos de Dofana 6: 17 pp. [assigned publ. date = June 30, 1996; see text].
Gorzula, Stefan, and J. Celsa Sefiaris
1999. Contribution to the herpetofauna of the Venezuelan Guayana I. A data base. Sci. Guaianae 8: xviii + 268 pp. + 129 color photos [in 32 end pls.], maps 2— 5 in pocket [1998” on cover, but ex- plicitly January 20, 1999 on reverse of inside title page].
1988.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 83
Grant, Taran, Elaine C. Humphrey, and Charles W. Myers The median lingual process of frogs: a bizarre character of Old World ranoids discovered in South American dendro- batids. Am. Mus. Novitates 3212: 40 pp. Harris, Dennis M. 1985. Infralingual plicae: support for Boulen- ger’s Teiidae (Sauria). Copeia 1985 (3): 560-565. 1994. Review of the teiid lizard genus Ptych- oglossus. Herpetol. Monogr. 8: 226— 275. Hitchcock, Charles B.
1997.
1947. The Orinoco-Ventuari region, Venezue- la. Geogr. Rev. 37(4): 525-566 + fold- out map.
1948. La regi6dn Orinoco-Ventuari, Venezue-
la. Bol. Soc. Venezolana Cienc. Nat. 11(72): 131-179 + foldout map (trans- lation of 1947 paper).
Huber, Otto
1995a. Geographical and physical features. Jn J. A. Steyermark, P. E. Berry, and B. K. Holst (general eds.), Flora of the Ve- nezuelan Guayana, vol. 1: 1-61. St. Louis: Missouri Botanical Garden.
1995b. History of botanical exploration. In J. A. Steyermark, P. E. Berry, and B. K. Holst (general eds.), Flora of the Ve- nezuelan Guayana, vol. 1: 63-95. St. Louis: Missouri Botanical Garden.
1995c. Vegetation. In J. A. Steyermark, P. E. Berry, and B. K. Holst (general eds.), Flora of the Venezuelan Guayana, vol. 1: 97-160. St. Louis: Missouri Botani- cal Garden.
International Commission on Zoological Nomen- clature
International code of zoological no- menclature, 4th ed. London: Interna- tional Trust Zoological Nomenclature, Xxxix + 306 pp.
Kornacker, Paul M.
1999. Checklist and key to the snakes of Ven- ezuela. Lista sistematica y clave para las serpientes de Venezuela. Rheinbach, Germany: Pako-Verlag, 270 pp. + 90 color illus. in 16 pls.
La Marca, Enrique
1993. Phylogenetic relationships and taxono- my of Colostethus mandelorum (Anu- ra: Dendrobatidae), with notes on col- oration, natural history, and description of the tadpole. Bull. Maryland Herpe- tol. Soc. 29(1): 4-19.
“1996” [1998]. Ranas del género Colostethus (Amphibia: Anura; Dendrobatidae) de
1999.
84 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY
la Guayana Venezolana con la descrip- cidn de siete especies nuevas. Publ. Asoc. Amigos de Dofiana 9: 64 pp. Lista actualizada de los reptiles de Ven- ezuela. In E. La Marca (ed.), Vertebra- dos actuales y fosiles de Venezuela, vol. 1: 123-142. Mérida, Venezuela: Museo de Ciencia y Tecnologia de Mé- rida. Lancini V., R. Abdem 1968. El género Euspondylus (Sauria: Tei- idae) en Venezuela. Publ. Ocas. Mus. Cienc. Nat. (Caracas), Zool. 12: 8 pp. Lobo, Fernando
1997.
1994. Primera cita de Pseudopaludicola bo- liviana Parker, 1927, para el Brasil y nuevos registros el Paraguay. Cuad. Herpetol. 8(2): 230-231.
1995. Analisis filogenético del género Pseu-
dopaludicola (Anura: Leptodactylidae). Cuad. Herpetol. 9(1): 21-43. Lynch, John D.
1989. A review of the leptodactylid frogs of the genus Pseudopaludicola in northern South America. Copeia 1989 (3): 577— 588.
Lynch, John D., and William E. Duellman
1973. A review of the centrolenid frogs of Ecuador, with descriptions of new spe- cies. Occas. Pap. Mus. Nat. Hist., Univ. Kansas 16: 66 pp.
1980. The Eleutherodactylus of the Amazo- nian slopes of the Ecuadorian Andes (Anura: Leptodactylidae). Univ. Kan- sas Mus. Nat. Hist. Misc. Publ. 69: [iv] + 86 pp.
1997. Frogs of the genus Eleutherodactylus in western Ecuador. Univ. Kansas Nat. Hist. Mus. Spec. Publ. 23: iv + 236 pp + 8 pls.
Magdefrau, Helmut
1991. Plica nigra, ein neuer Leguan von Guaiquinima Tepui (Venezuela) (Sau- ria, Iguanidae). Spixiana 14: 229-234.
Magdefrau, Helmut, Karin Magdefrau, and An- dreas Schliiter
Herpetologische Daten vom Guaiquin- ima-Tepui, Venezuela. Herpetofauna (Germany) 13(70): 13-26.
MacLean, William P.
1974. Feeding and locomotor mechanisms of teiid lizards: functional morphology and evolution. Pap. Avulsos Zool. (Sao Paulo) 27(15): 179-213.
Mayr, Ernst, and William H. Phelps, Jr.
1967. The origin of the bird fauna of the
South Venezuelan highlands. Bull. Am.
1991.
NO. 261
Mus. Nat. Hist. 136(5): 269-327 + pls. 14-21, foldout map. Myers, Charles W.
1991. Distribution of the dendrobatid frog Colostethus chocoensis and description of a related species occurring macro- sympatrically. Am. Mus. Novitates 3010: 15 pp., 8 figs.
Preliminary remarks on the summit herpetofauna of Auyantepui, eastern Venezuela. Acta Terramaris 10: 1-8. A history of herpetology at the Amer- ican Museum of Natural History. Bull. Am. Mus. Nat. Hist. 252: 232 pp. Myers, Charles W., and Maureen A. Donnelly
1996. A new herpetofauna from Cerro Yavi, Venezuela: first results of the Robert G. Goelet American Museum—TERRAMAR expedition to the northwestern tepuis. Am. Mus. Novitates 3172: 56 pp.
A tepui herpetofauna on a granitic mountain (Tamacuari) in the borderland between Venezuela and Brazil: report from the Phipps Tapirapeco Expedition. Am. Mus. Novitates 3213: 71 pp. Myers, Charles W., and William E. Duellman 1982. A new species of Hyla from Cerro Col- orado, and other tree frog records and geographical notes from western Pan- ama. Am. Mus. Novitates 2752: 32 pp. Myers, Charles W., John W. Daly, and Borys Mal- kin A dangerously toxic new frog (Phyllo- bates) used by Embera Indians of west- ern Colombia, with discussion of blow- gun fabrication and dart poisoning. Bull. Am. Mus. Nat. Hist. 161(2): 307— 366 + color pls. 1-2. Myers, Charles W., Alfredo Paolillo O., and John W. Daly Discovery of a defensively malodorous and nocturnal frog in the family Den- drobatidae: phylogenetic significance of a new genus and species from the Venezuelan Andes. Am. Mus. Novita- tes 3002: 33 pp. Myers, Charles W., Lily O. Rodriguez, and Javier Icochea Epipedobates simulans, a new cryptic species of poison frog from southeast- ern Peru, with notes on E. macero and E. petersi (Dendrobatidae). Am. Mus. Novitates 3238: 20 pp. Neuwirth, Maria, John W. Daly, Charles W. My- ers, and Lois W. Tice Morphology of the granular secretory glands in skin of poison-dart frogs
1997.
2000.
1997.
1978.
1991.
1998.
1979.
2001
(Dendrobatidae). Tissue Cell 11(4): 755-771. Pérez-Santos, Carlos, and Ana G. Moreno
1989. Una nueva especie de Thamnodynastes (Serpentes: Colubridae) en el norte de Colombia. Boll. Mus. Reg. Sci. Nat. Torino 7(1): 1-9.
Peters, James A.
1964. Dictionary of herpetology. A brief and meaningful definition of words and terms used in herpetology. New York: Hafner, ix, 392 pp. + 30 figs.
Peters, Wilhelm.
“1862” [1863]. Uber Cercosaura und die mit dieser Gattung verwandten Eidechsen aus Stidamerica. Abh. Phys. KI. K. Akad. Wiss. Berlin 1862(4): 165-225 + 3 pls. [Available in 1995 SSAR fac- simile reprint, The herpetological con- tributions of Wilhelm C. H. Peters (1815-1883): 175-197.]
Phelps, William H., Jr.
1977. Aves colectadas en las mesetas de Sar- isarifama y Jaua durante tres expedi- ones al macizo, Estado Bolivar. Des- cripciones de dos nuevas subespecies. Bol. Soc. Venezolana Cienc. Nat. 33, 134: 15-42 + 2 maps.
Presch, William
1978. Descriptions of the hemipenial mor- phology in eight species of microteiid lizards (family Teiidae, subfamily Gymnophthalminae). Herpetologica 34(1): 108-112. Evolutionary history of the South American microteiid lizards (Teiidae: Gymnophthalminae). Copeia 1980 (1): 36-56. The lizard family Teiidae: is it a mono- phyletic group? Zool. J. Linn. Soc. 77: 189-197. Rivero, Juan A.
“1982” [1984]. Sobre el Colostethus mande- lorum (Schmidt) y el Colostethis inflex- us Rivero (Amphibia, Dendrobatidae). Mem. Soc. Cienc. Nat. La Salle 42(118): 9-16.
Roze, Janis A.
1958a. Los reptiles del Auyantepui, Venezuela, basandose en las colecciones de las ex- pediciones de Phelps-Tate, del Ameri- can Museum of Natural History, 1937— 1938, y de la Universidad Central de Venezuela, 1956. Acta Biol. Venezuel- ica 2(22): 243-270.
1958b. Los reptiles del Chimanta Tepui (Esta- do Bolivar, Venezuela) colectados por la expedicién botanica del Chicago
1980.
1983.
MYERS AND DONNELLY: YUTAJE-COROCORO MASSIF 85
Natural History Museum. Acta Biol. Venezuelica 2(25): 299-314. La taxonomia y zoogeografia de los of- idios en Venezuela. Caracas: Univ. Central Venezuela, 362 pp. Rubin, L., P. Eller, and J. Nolte 1986. Fine structure of the photosensitive iris of the toad, Bufo marinus. J. Morphol. 188: 225-238. Ruibal, Rodolfo 1952. Revisionary studies of some South American Teiidae. Bull. Mus. Comp. Zool. 106(11): 477-529. Ruiz-Carranza, Pedro M., and John D. Lynch 199la. Ranas Centrolenidae de Colombia I. Propuesta de una nueva clasificacién genérica. Lozania (Acta Zool. Colom- biana) 57: 30 pp. 1991b. [Same title] I]. Nuevas especies de Centrolene de la Cordillera Oriental y Sierra Nevada de Santa Maria. Lozania (Acta Zool. Colombiana) 58: 26 pp. Ruiz-Carranza, Pedro M., Maria Cristina Ardila- Robayo, and John D. Lynch 1996. Lista actualizada de la fauna de Am- phibia de Colombia. Rev. Acad. Col- ombiana Cienc. 20(77): 365-415. Savage, Jay M., and W. Ronald Heyer 1967. Variation and distribution in the tree- frog genus Phyllomedusa in Costa Rica, Central America. Beitr. Neotrop. Fauna 5(2): 111-131. Schmidt, Karl P. 1932. Reptiles and amphibians of the Mandel Venezuelan Expedition. Field Mus. Nat. Hist., Zool. Ser. 18(7): 159-163. Steyermark, Julian A., and Charles Brewer- Carias 1976. La vegetacion en la cima del macizo de Jaua. Bol. Soc. Venezolana Cienc. Nat. 32,132/133: 179-405. Uzzell, Thomas
1966.
1969. The status of the genera Ecpleopus, Ar- throseps and Aspidolaemus (Sauria, Teiidae). Postilla (Peabody Mus. Nat. Hist. Yale Univ.) 135: 23 pp.
1973. A revision of lizards of the genus
Prionodactylus, with a new genus for P. leucostictus and notes on the genus Euspondylus (Sauria, Teiidae). Postilla (Peabody Mus. Nat. Hist. Yale Univ.) 159: 67 pp. Walls, Gordon Lynn
1942. The vertebrate eye and its adaptive ra- diation. Cranbrook Inst. Sci. Bull. 19: xiv + 785 pp. [Reprinted 1963, with same pagination, by Hafner Publ. Co., New York and London. ]