AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024 Number 3357, 48 pp., 21 figures, 2 tables March 26, 2002 Frogs of the Eleutherodactylus biporcatus Group (Leptodactylidae) of Central America and Northern South America, Including Rediscovered, Resurrected, and New Taxa JAY M. SAVAGE! AND CHARLES W. MYERS? CONTENTS PCOS UDA Ee Ng 4 coe Rta Ae youre Stele teM Me Ae pettya ch 4 fee TO yong:S trat Mele Pe Dag OG Le Pee Oe. psa 2 FRSC ia a are cab oho, pte ona nang eat to yy Oln at ca adhe attec eat a eee etal wegen ens dance BeA aah et oes als 2 (BolucolCGO:SI TNSi 3i sA n eROe Arce Meh ee aCe ee OR Ea IgE AUR Rar k,L ee Set oa pn ITO 3:F 2 INFCTONC Susy yee) htt Sram 2 x ceectori ra Ses Ie YUN IRN Ny cis ss tenets pie tee Meh PUN ALAR esc 2g 3 Wa XOMOMIiG TAIStOIy Metter rece bs tas ost ne ceed nad Poh ee cette earacathh,& T eaem ese ee + IRCdISCOVEL VO SHOU VIESH BUI OSHSTE OP CS Toma foamy pent I nate eyes ca AU iat e utes eee ae 6 Identity ‘of ‘Sirabomantis biporcatus W. Peters; 1863 «x20 0:5e: teas > SAR AAD AA OO Oe HE 10 Application of the Name Lithodytes megacephalus Cope, 1875 ............0000000. e e 14 Resurrection of Aylodessfusosus W. Peterss’ T8733 24 ois che venuDeee ee w e oe 15 Systeniatic Synopsis: the biporcanis Species Group, + 30 Men y Lee Lee | WEY TOS Pe Cle Sx tbe ks eee S Sed ited i ice 2 ea at teas S Bye Pentagon lis os ee ale cot 17 FLCUNCTOAGACTYUS A PNOMIES ATA OC jr seg scien capg tat eta memtary, dodo Sete, athe beets dow Wky oR anes 17 Pleutherodactylis- biporcatus»Ovs, Peters): ipo oT te sph e CE ee lcd wp ek Db ne 21 Pel entherodacinlus SuUlasus (COPE): x1 4 ee PPS ecg, bracehowe aces eke IE EY WE Bed PS eta 24 Hleutherodactylus-mevacepnals (Ope) Ge a... Ah eee len hd dab Mites oh Ada 29 FEl eusherOAGCIVIUS OPUS AT EW BSC CICS tle Boobs. Es Bich the Be: a URE wye ae 8 ee gas 34 TE LCULNCFOAACEVIUSATH OSU NV CLELS © ortte ow tug Webege oea to hne oisL slpin eleN ew erd tet eas eeplg eane ae 38 Is the Eleutherodactylus biporcatus Group Monophyletic? ...................000008- 40 Acknowledgments?) .-2 22415 © i. Gee inet eden oleae cae ane ee. 9 ooR LS Se eee eee SIR eee oad dale 8 42 APPS: Oe CIMENS teK ATMIECC M, WOE en ae tees ty ee ol day ce tens aes MO lls ae trees: Bene 42 FETC COS Bett tty MR Re cx cite whe oh Nl 9 Ot Rc AE, 5o R Pula RA ta ie we AR En 45 ‘Research Associate, Division of Vertebrate Zoology (Herpetology), American Museum of Natural History; Pro- fessor Emeritus, Biology, University of Miami. ? Curator Emeritus, Division of Vertebrate Zoology (Herpetology), American Museum of Natural History. Copyright © American Museum of Natural History 2002 ISSN 0003-0082 AMERICAN MUSEUM NOVITATES NO. 3357 ABSTRACT A revision of the broad-headed frogs of the biporcatus species group of Eleutherodactylus s.l. has a wholly unexpected nomenclatural consequence. Eleutherodactylus biporcatus (W. Peters, 1863) is not from “‘Veragua”’ (western Panama) as originally thought, but is the proper name for the Venezuelan frog heretofore known as E. maussi (Boettger, 1893). Three names are resurrected from synonymy for Central American species currently mas- querading under the misapplied name biporcatus, and a fourth species is described as new: (1) The rediscovery of Eleutherodactylus gulosus (Cope, 1875) shows it to be a large montane frog occupying an apparently small range in the borderland of Costa Rica and Panama. (2) Eleuth- erodactylus rugosus (W. Peters, 1863) is a smaller species occurring on the Pacific versant of southwestern Costa Rica and western Panama; Lithodytes pelviculus Cope and L. florulentus Cope are synonyms of E. rugosus. (3) Eleutherodactylus megacephalus (Cope, 1875), an inter- mediate-sized frog ranging from Honduras to central Panama, is the more common species to which the name biporcatus has usually been applied. Available material from the western half of the Isthmus of Panama was too sparse to decide if another (unnamed) species is being included under the name megacephalus. (4) The name biporcatus also has been used for Eleutherodac- tylus opimus, new species, which occurs from central Panama to western Colombia. Based on the condition of the m. adductor mandibulae, the Venezuelan Eleutherodactylus biporcatus s.s. (E. maussi, auctorum) belongs to the Middle American clade of Eleutherodac- tylus (subgenus Craugaster). However, preliminary data on karyotypes, as well as morpho- logical differences, cast doubt on the closeness of E. biporcatus to the other species studied. The monophyly of the “‘biporcatus group” therefore remains to be tested. RESUMEN La revisi6n de las ranas de cabeza ancha del grupo biporcatus de Eleutherodactylus tiene una consecuencia nomenclatural completamente inesperada. Eleutherodactylus biporcatus (W. Peters, 1863) no es de “Veragua’’ (Panama occidental) como se creia previamente, sino que es el nombre propio de la rana venezolana conocida hasta ahora como E. maussi (Boettger, 1893). Se resucitan tres nombres de la sinonimia para especies centroamericanas previamente ocul- tos bajo el nombre incorrecto de biporcatus, y se describe una cuarta especie nueva: (1) El redescubrimiento de Eleutherodactylus gulosus (Cope, 1875) indica que es una rana grande, de montafia, que tiene una distribuci6n geografica pequefia en la frontera entre Costa Rica y Panama. (2) Eleutherodactylus rugosus (W. Peters, 1863) es una especie mas pequefia que ocurre en las tierras mas bajas del Pacifico del suroccidente de Costa Rica y del occidente de Panama; Lithodytes pelviculus Cope y L. florulentus Cope son sindnimos de E. rugosus. (3) Eleutherodactylus megacephalus (Cope, 1875), una rana de tamafo intermedio y distribuida desde Honduras hasta Panama central, es la especie que mas comunmente se ha llamado biporcatus. E| material disponible de la mitad occidental del Istmo de Panama no es adecuado para decidir si hay alguna otra especie (sin nombre) incluida bajo el nombre megacephalus. (4) El nombre biporcatus también se ha usado para Eleutherodactylus opimus, especie nueva, que ocurre desde Panama central hasta Colombia occidental. Basandose en la condicioén del m. adductor mandibulae, Eleutherodactylus biporcatus s.s. (E. maussi auctorum) de Venezuela pertenece al clado mesoamericano de Eleutherodactylus (subgenero Craugaster). Sin embargo, informacion preliminar sobre cariotipos, asf como di- ferencias morfolodgicas, hacer dudar del parentesco de E. biporcatus con las otras especies estudiadas. La monofilia del “‘grupo biporcatus”’ debe ser corroborada. INTRODUCTION spots in a dark reticulum. On the Atlantic side of Costa Rica, the little frogs are further In the lowlands and on the adjacent slopes characterized by a coloring of red or orange of lower Central America, some of the most on the posterior venter and undersides of the commonly seen leaf-litter anurans are small, hands and feet. These frogs have been re- broad-headed, diurnal, toadlike Eleuthero- ferred to as Eleutherodactylus biporcatus by dactylus that usually have large white ventral most recent authors following Dunn (1931). 2002 SAVAGE AND MYERS: ELEUTHERODACTYLUS BIPORCATUS GROUP 3 However, the impressions gained of these an- imens from the Berlin Museum, through the imals based on diurnal observations of the courtesy of Dr. Rainer Giinther. small juveniles are illusionary, for the adults Serendipitously, the two lines of research are large (to 70 mm SVL), seldom-seen noc- converged on the occasion of the 62nd an- turnal predators. They may typically hide in nual meeting of the American Society of Ich- burrows during the day, and come to the sur- thyologists and Herpetologists at the Ameri- face at night to ambush passing prey, includ- can Museum of Natural History, in June ing large arthropods and even small frogs 1991, when conjoint examination of the Ber- and lizards. lin type material convinced us of the need Unlike most species in this huge genus of for this collaborative analysis. In its course, over 500 species, males seemingly do not vo- we examined extensive series of biporcatus- like frogs, the type specimens of all nominal calize.* As a result, adult males, which are usually located by their sounds in other spe- Central American taxa thought to be syno- cies, are underrepresented in collections; at nyms of E. biporcatus, and the types of valid well-sampled sites, females are obviously related species. Our primary goal in the study is to determine the systematic status, distri- disproportionately represented among adults. bution, and relationships of FE. biporcatus Reproductive behavior is unknown, but eggs and its allies and to establish a proper no- are probably laid in the leaf litter where the menclature for included valid taxa. juveniles abound. Our manner of presentation of the results The present study has developed from two of study departs somewhat from traditional different lines of research, each questioning systematic papers where an a posteriori log- the allocation of all lower Central American ical framework is imposed on the not nec- specimens of these broad-headed frogs to a essarily logical sequence of investigation. single species. Savage has had a longtime Rather, as in the genre of detective fiction, commitment to elucidating the composition we have tried for the most part to present of, and relationships within the genus Eleuth- each clue leading to our conclusions in the erodactylus, especially in Central America. order of discovery. As in such tales, there For some time he had thought that more than were several real surprises and a few false one species was subsumed under the name leads as well. Eleutherodactylus_ biporcatus, but believed that a review of broad-headed frogs nomi- nally of that taxon from throughout its range METHODS was necessary to confirm or reject this hy- Description of the external morphology of pothesis. Myers was particularly interested in the included taxa follows, where appropriate, the genus in the context of its species and the characters defined in Lynch and Duell- their distribution on the Panamanian Isthmus. man (1980, 1997) and Savage (1975, 1987). His own fieldwork in western Panama had The presence or absence of paired dorsal led him also to conclude that more than one ridges (plicae) is important in the taxonomy species probably was represented by the of this group. For convenience, we use the monolithic E. biporcatus. In the process of term suprascapular plicae for the anterior beginning to resolve this problem, Myers ridges, even though they may extend poste- was able to borrow several critical type spec- riorly from the eyes to midbody in some spe- cies. One species has paravertebral plicae 3 This is an inference based on never having seen one positioned at midbody, separate from the su- calling. The males of all but one species in the group prascapular plicae. The posterior or latero- do lack vocal slits (hence vocal sacs), but absence of sacral plicae, when present, are often less these structures is not a reliable predictor of voiceless- well defined, sometimes comprising only ness. Vocal sacs have been lost in diverse species of frogs that, however, retain well-developed advertisement slightly raised, linear rows of small tubercles. calls (e.g., Myers, 1982: 4—5 [Dendrobates reticulatus]; Measurements were taken with a dial cal- Zweifel, 2000: 72—73 [Liophryne similis]). And voice in iper to 0.01 mm and rounded to the nearest female frogs does not depend on such structures (see 0.1 mm. Head-body length is indicated by Natural History in systematic section, under Eleuthero- dactylus biporcatus). the abbreviation SVL, which may be read ei- - AMERICAN MUSEUM NOVITATES NO. 3357 ther as snout-vent length or amphibian stan- time for much of what is now western Pan- dard length (distance from the tip of the ama, including today’s provinces of Bocas snout to posterior margin of vent); for these del Toro, Chiriqui, and Veraguas. Interest- frogs there is no practical difference between ingly, in the same paper he described an un- the definitions, inasmuch as the vent opens related Eleutherodactylus as Hylodes goll- in a vertical plane. Measurements, taken meri, based on two frogs purportedly from mostly by the first author over a period of Caracas, Venezuela (lectotype ZMB 3168, time, involve specimens both poorly and well paralectotype ZMB 36301; originally iden- preserved. For this reason and, especially, tified by Lichtenstein and von Martens because significant population samples of [1856] as Hylodes martinicensis from ‘‘Ca- adults are nearly lacking, we provide only vacas’’). simple quantitative values without statistical Subsequent study (Rivero, 1961; Savage, analyses. 1987) suggested that an inadvertent exchange The color descriptions emphasize colors in of data may have occurred between the life, which are based on our field notes and aforesaid type specimens of Hylodes goll- color slides. In the Systematic Synopsis these meri and some types of Phrynidium (= Ate- summarize variation in the most obvious lopus) crucigerum Lichtenstein and von Mar- and/or distinguishing features. Most bright colors, especially reds, yellows, and greens tens (1856). Although a few extant “‘types”’ are lost more or less rapidly on preservation, of Phrynidium crucigerum are from the pub- while darker ones tend to be intensified. Con- lished type locality “‘Veragoa’’, others are sequently the general pattern, uniform, spot- conspecific with a species (usually called ted blotched, striped, etc., is similar in life Atelopus cruciger) inhabiting the Cordillera and preservation. Other aspects of coloration de la Costa of Venezuela.* On the same hand, and pattern markedly affected by preserva- the types of Hylodes gollmeri are represen- tion beside those just mentioned are indicat- tatives of a rather common but distinct Pan- ed where appropriate. amanian species of Eleutherodactylus. A TAXONOMIC HISTORY 4 Tn a recent paper, Lotters et al. (1998) confirmed that The frogs of the genus Eleutherodactylus specimens labeled (in ZMB) as syntypes of the name Duméril and Bibron, 1841 were referred to crucigerum are from both Panama and Venezuela. They various genera during the 19th and early 20th however selected a nonlabeled specimen (ZMB 3380) from “‘Veragoa’’ (western Panama) as the lectotype and century (G. Myers, 1962). Original species synonymized Phrynidium crucigerum with the Central descriptions and subsequent references to the American Atelopus varius. Lotters et al. proposed a sce- taxa discussed below used the names Hylo- nario whereby ZMB 3380-—-a single specimen which has des Fitzinger, 1843, Lithodytes Fitzinger, “no indication of type status’ collected by J. Warszew- icz—-must have been cataloged with at least four other 1843, and Strabomantis W. Peters, 1863, un- specimens that may subsequently have been lost or ex- til (and even after) Stejneger (1904) conclud- changed (crucigerum supposedly was based on five ed that most species in the nominal Hylodes specimens from Veragoa). and Lithodytes should be placed in Eleuth- They asserted that Venezuelan specimens could not erodactylus. Many workers of this period ap- have been used for the original description because of a difference in skin texture. However, as is well known, plied the name Hylodes to all these frogs, but early authors commonly based descriptions on single Cope (1862: 153-154) divided that genus specimens drawn from whatever sample of “‘cotypes’’ into several groups, including Hylodes, with (syntypes) was at hand. typically granular venters, and Lithodytes, Considering the poor original description and all the with smooth venters, a distinction no longer assumptions made by L6Otters et al. as to what is or might have been one of the original types, they might have recognized. better served stability by designating a syntype that was In 1863, Wilhelm Peters erected a new cataloged as such and that would have preserved the genus and species for a broad-headed frog well-known name Atelopus cruciger for the Venezuelan from “‘Veragua”’ that he called Strabomantis frog. The matter would now seem nomenclaturally dis- posed of, except for the arguable question as to whether biporcatus. Veragua was the name of a the “‘lectotype’’ was indeed part of the syntype series as Province of New Granada and used at that postulated by Létters et al. 2002 SAVAGE AND MYERS: ELEUTHERODACTYLUS BIPORCATUS GROUP S, similar error in locality data associated with Cabello, on the Caribbean coast of Venezue- old Berlin Museum type specimens will be Fay? shown later in this paper to have serious im- Noble (1918: 329-331) used the name plications for the provenance and identity of Eleutherodactylus rugosus for examples of Strabomantis biporcatus. this group from Nicaragua and placed Lith- Peters (1873) later described a second odytes megacephalus and L. pelviculus in its broad-headed form as Hylodes rugosus from synonymy. He concluded that his “‘large se- ries of specimens shows conclusively that E. ‘“*Chiriqui’’ without reference to his earlier- megacephalus is but the adult of E. rugo- named Strabomantis biporcatus from Vera- sus’. Noble retained L. gulosus as a valid gua. The name Chiriqui was used at that species of Eleutherodactylus similar to but time for western Veragua Province and is larger than megacephalus/rugosus. essentially equivalent to today’s Panamani- Stejneger (1904) had revived usage of the an provinces of Bocas del Toro and Chiri- name Eleutherodactylus for frogs placed in qui. the Hylodes and Lithodytes of previous au- Shortly thereafter, Cope (1875) recognized thors, a course followed by most subsequent two additional broad-headed taxa, Lithodytes workers. Dunn (1931) used the name Eleuth- megacephalus and Lithodytes gulosus, col- erodactylus biporcatus for the Panamanian lected on W. M. Gabb’s expedition to the species and, according to his unpublished southern Cordillera de Talamanca of Costa notes, regarded all names based on Central Rica. A few years later Cope (1877: 89) de- American material of the group available at scribed Lithodytes pelviculus, “‘said to have that time to be synonyms, but he reluctantly been found on the west coast of Central continued to recognize E. gulosus as valid. America’. Boulenger (1882: 198—220) and Taylor (1952), in his monograph of Costa Brocchi (1882: 49-59) simultaneously Rican anurans, recognized Eleutherodactylus placed Strabomantis in the synonymy of Hy- rugosus (pp. 730-734) and E. florulentus lodes, in which they recognized all five of (pp. 765-766) as valid species. He accepted the broad-headed species mentioned above. Noble’s (1918) synonymy of Lithodytes me- Cope later (1893: 336) maintained Lithodytes gacephalus with E. rugosus, but questioned for the three species that he had named (gu- whether L. gulosus and L. pelviculus should losus, pelviculus, megacephalus) and de- be included there. scribed another (Lithodytes florulentus) from Savage (1973, 1976, 1980) and Savage Boruca in the Pacific drainage of southwest- and Villa (1986), influenced by Dunn’s un- ern Costa Rica. published synonymy, regarded all Central Giinther (1900-1901 [1885—1902]) treated American names for this group to be based on specimens of E. biporcatus. Myers and biporcatus (p. 231), rugosus (p. 233), floru- Rand (1969) and Rand and Myers (1990) fol- lentus (p. 234), megacephalus (p. 239), and lowed Dunn (1931) in using E. biporcatus gulosus (p. 240) as distinct taxa of Hylodes. for Panamanian members of this group. Giinther (op. cit.) also had intended to treat Lynch (1975), in the most recent review of Lithodytes pelviculus as a valid species of the taxonomy of these frogs, likewise con- Hylodes, as shown by its inclusion as species cluded that these various names represented no. 23 in his key on page 227 (where it is only one species in Central America, namely misspelled ’’petriculus’’ [fide erratum on p. Eleutherodactylus biporcatus. He included xx]), but subsequently changed his mind and that species in his Eleutherodactylus bipor- allocated the name to the synonymy of H. megacephalus (p. 239). >A South American broad-headed frog unrelated to Nieden (1923: 432-439, 466) followed the biporcatus group but confused with it is the nominal Giinther’s (1901) assignments for the Central Limnophys napaeus Jiménez de la Espada, 1870, which American species, and also picked up the authors have placed in the synonymy of Eleutherodac- name Hylodes maussi Boettger (1893) for a tylus biporcatus. Lynch (1975: 30—31) reviewed the ev- idence and allocated napaeus to the synonymy of South American broad-headed species appar- Eleutherodactylus cornutus in the sulcatus species ently related to this complex, from Puerto group. 6 AMERICAN MUSEUM NOVITATES NO. 3357 catus group along with E. cerastes, E. cor- within the genus. The South American clade nutus, and E. necerus of northwestern and is characterized by having a m. adductor upper Amazonian South America, and E. bu- mandibularis posterior subexternus present, foniformis of lower Central America and and the mandibular ramus of the trigeminal northwestern South America, but he placed nerve (V3) passing external (lateral) to the the Venezuelan E. maussi in the separate sul- muscle (“‘S”’ condition). The Central Amer- catus group. This grouping was reiterated in ican clade (subgenus Craugaster Cope, Lynch’s (1976) discussion of the South 1862: 153) has a m. adductor mandibularis American species groups of Eleutherodacty- externus superficialis, and the trigeminal lus. nerve passing mediad (internal) to the mus- Lynch (1975, 1976) regarded the forms in cle (“E” condition). This conclusion re- his E. biporcatus group as closely related quired a complete renovation of the two since they shared the features of very broad broad-headed species groups, because only heads (head width > 42% SVL), finger I > Eleutherodactylus biporcatus (sensu lato), E. finger II, and smooth venters. Lynch (1975) bufoniformis, E. florulentus, E. necerus, and indicated that another cluster of broad-head- E. maussi have the E condition; these were ed South American Eleutherodactylus, the assigned to the biporcatus group. Except for sulcatus group, was somehow related to the E. maussi, the remaining South American biporcatus group, but in the 1976 paper he broad-headed forms all have the S condition placed the two groups in different infrage- and currently are placed in the sulcatus neric units within the genus. group.° Subsequently, Lynch (1981) expanded the In 1980, Jonathan A. Campbell collected sulcatus group to include the nominal mono- several biporcatus-like frogs from the moun- typic genus Amblyphrynus (Cochran and tains of eastern Guatemala. Savage examined Goin, 1961) and a new form to bring its com- this material in 1980 and expressed his opin- position to five species: Eleutherodactylus ion that it was doubtless a diminutive new helonotus (Ecuador), E. ingeri (Colombia), member of the biporcatus group. Campbell E. maussi (Venezuela), E. ruizi (Colombia), (1994), after obtaining additional material, and E. sulcatus (Amazonian Ecuador and described the species as Eleutherodactylus Peru). aphanus. Savage and DeWeese (1979, 1981) and Lynch and Duellman (1997), in a major Savage (1984, 1987), while acknowledging treatment of Eleutherodactylus relationships, the utility of Lynch’s (1976) attempt to bring placed Eleutherodactylus bufoniformis and some kind of order to the classification of E. necerus in a separate bufoniformis species Eleutherodactylus, criticized Lynch’s system group, leaving the biporcatus group with as essentially phenetic and based on trivial three recognized species (aphanus, biporca- external features much given to homoplasy. tus, maussi). We accept the biporcatus and They preferred to emphasize the condition of bufoniformis species groups for purposes of jaw muscles and karyology over finger the present study, which focusses on the first length and ventral texture. group. Lynch (1986) responded to this challenge with a thorough examination of the jaw mus- REDISCOVERY OF LITHODYTES cle characters of the m. adductor mandibulae GULOSUS, COPE, 1875 of Eleutherodactylus, first discovered by In February of 1976, Myers collected a se- Starrett (1968) and touted by Savage and ries of unusual juvenile Eleutherodactylus DeWeese (1979, 1981) as systematically in- (fig. 1) in leaf litter near the site of the pre- formative. Lynch’s work has broad signifi- cance to the systematics of all Eleutherodac- tylus and their allies, but only those aspects © The sulcatus species group now includes nine spe- relating to the present paper are summarized cies according to the latest revision (Lynch, 1997). It is here. a South American assemblage. Only one species occurs (barely) in political Central America, namely E. laticor- The principal result of Lynch’s 1986 study pus Myers and Lynch (1997) from the Panamanian-Col- was the recognition of two major clades ombian frontier. 2002 SAVAGE AND MYERS: ELEUTHERODACTYLUS BIPORCATUS GROUP i this enigmatic frog. Found at elevations be- tween 1000 and 1220 m, on and near the continental divide, the adults were distinct from lower-elevation E. biporcatus in lack- ing dorsal plicae and in details of coloration. The question then naturally arose as to whether these frogs represented an undescri- bed species or were conspecific with one of the several taxa previously named from west- erm Panama and southern Costa Rica, but then regarded as junior synonyms of E. bi- porcatus. Crucial to resolving this question was examination of the type specimens of each available name. Consequently, Myers borrowed the type series of Strabomantis bi- porcatus and the holotype of Hylodes rugo- sus from the Berlin Museum. A preliminary comparison of these animals with the newly collected material left him with little doubt that the Fortuna frogs could not be associated Fig. 1. The first specimen collected during the with either of the species represented by the rediscovery of Lithodytes [Eleutherodactylus| gu- losus in 1976—a century after its original descrip- Berlin types. tion (Cope, 1875)—at the proposed Fortuna Dam But Myers was astonished to find that the site in western Panama. A juvenile female type series of S. biporcatus represented a (AMNH 95035) of 32 mm SVL, less than half species different from the frogs usually grown; note the pustulate dorsum, a juvenile con- called E. biporcatus for more than half a cen- dition (compare adult in fig. 12). [From a trans- tury, and that they were also unlike any other parency by C. W. Myers. ] known species from Central America. This puzzlement led to consultation with Savage on the occasion mentioned in the introduc- sent-day Fortuna Dam (then being surveyed tion to this paper, with the latter fully agree- for construction), at 1120 m, on the Pacific ing with Myers’ conclusions that the Fortuna versant of western Panama (fig. 2). Although frogs: (1) were not conspecific with the spe- reminiscent of the common broad-headed ju- cies usually called E. biporcatus in Central veniles identified with Eleutherodactylus bi- America; (2) were not conspecific with the porcatus, they differed from them most ob- types of S. biporcatus or H. rugosus; (3) and viously in lacking well-defined dorsal plicae that the types of S. biporcatus were not rep- and in having a different ventral coloration. resentatives of the species usually referred to In addition, the dorsum was conspicuously as E. biporcatus in Central America! covered by numerous small pustules. Direct Further discussion led us to the hypothesis comparison with juvenile E. biporcatus of that the Fortuna frogs might well be exam- the same size strongly suggested that these ples of one of the other of Cope’s nominal frogs were not conspecific with typical low- species, Lithodytes gulosus or L. megace- land E. biporcatus.’ Phalus, also from an upland area in the same Later visits to nearby areas by Myers and mountain range as Fortuna but on the Atlan- John Daly in July of 1982 and January 1983 tic slope in southern Costa Rica. The possi- led to the capture of two large females, one bility also existed that both these names were subadult female, and additional juveniles of based on a single taxon conspecific with the Fortuna examples. 7 Jungfer (1988: 16) later used the name Eleuthero- Cope’s (1875: 110—112) original descrip- dactylus biporcatus for a juvenile specimen from the tions clearly differentiated the holotypes of Fortuna regional population, but his accompanying pho- Lithodytes gulosus and L. megacephalus as tograph represents a species (E. megacephalus) from an- other locality. follows: 8 AMERICAN MUSEUM NOVITATES NO. 3357 Fig. 2. Views of the Fortuna Dam site in the upland valley of the Rio Chiriqui, in 1976, prior to road and dam construction (access was by trail or helicopter). Upper: Survey and research camp of the Instituto de Recursos Hidraulicos y Electrificacion (IHRE) above Rio Chiriqui (at right), on northern base of Cerro Fortuna, about 1000 m above sea level. Lower: View looking westward across the subsequently flooded Rio Chiriqui to Cerro Pinola (center), from 1060 m on north slope of Cerro Fortuna. [Photographs by C. W. Myers, March 2-3, 1976.] 2002 SAVAGE AND MYERS: ELEUTHERODACTYLUS BIPORCATUS GROUP ) L. gulosus L. megacephalus No dermal plicae Dermal )( folds dorsally and laterosacral plicae present Dorsum uniform dark leather brown Dorsum light ash Posterior thigh like dorsum Posterior thigh black, marbled distally with ash, black continuing as patch to vent Head width 2.2 times in head-body length (SVL); 46% of SVL Head width 2 times in SVL; 50% of SVL Adult female 103 mm SVL Adult female 70 mm SVL The types of both putative species are gulosus with the two adult females (AMNH housed at the United States National Muse- 124370, AMNH 124372), 84 and 72 mm um of Natural History and are part of the W. SVL, respectively, and a subadult female M. Gabb collection made in Costa Rica in (AMNH 124371, 46 mm SVL) from the For- 1873-1874. The holotype of both Lithodytes tuna Dam area of Panama leaves little doubt gulosus (USNM 32590) and of Lithodytes that they represent the same species. The megacephalus (USNM 32579) are from most striking similarity is the absence both **Pico Blanco”’, Limon Province, Costa Rica, of the suprascapular plicae and the laterosa- 6000 ft. However, as pointed out by Savage cral tubercle and/or ridge system. The plicae (1970), Gabb’s expedition did not ascend are characteristic of all other named forms of Pico Blanco (= Cerro Kamuk), but actually biporcatus-like frogs, and laterosacral tuber- climbed to the summit of Cerro Utyum, the cles or ridges are found in some of the other next peak to the northwest. species of the group (fig. 3). The type specimen of Lithodytes gulosus In other features, the adult females from is in excellent condition and agrees well with Panama agree closely with the holotype and Cope’s original description. Unfortunately Cope’s description of Lithodytes gulosus in the type of Lithodytes megacephalus has suf- having a uniform dark dorsum and posterior fered serious decay and today consists only thigh surface. In addition they have the chin of skin fragments and disarticulated bones. and throat pale or gray-brown, pale venters The maceration of this specimen is relatively mottled with gray-brown or with dark brown recent, as Noble (1918) and Dunn (unpub- spotting, and the seat patch is inconspicuous. lished notes) examined it while it was still The subadult female differs from the oth- whole. Noble regarded the specimen as a ers in some features of coloration, including large adult conspecific with Eleutherodacty- distinct transverse dark bars on the limbs, lus rugosus and slightly different from the some pale tan spotting on the posterior thigh type of L. gulosus, which he retained as surface, the dark seat patch mark restricted Eleutherodactylus gulosus. Dunn (1931) ap- to the area around the vent, the chin and parently considered both names as synonyms throat gray with small white spots, the un- of Eleutherodactylus biporcatus, the types of derside of the leg suffused with yellow, and which he had examined in Berlin in 1928. the venter, groin and anterior thigh surface Although the condition of the holotype of marbled blackish gray on white (life colors). Lithodytes megacephalus makes direct com- This specimen is somewhat intermediate in parisons impossible, Cope’s (1875) descrip- coloration and dorsal pustulation between the tion and illustration (reproduced herein as large adults and the series of juveniles. fig. 6) clearly demonstrate that it is not con- Unlike adults, juveniles (AMNH 95035-— specific with the adult Fortuna frogs. Unlike 95046, 124365—-124369; 18-32 mm SVL) them it had well-developed suprascapular have dorsally pustulate skin, a lighter dorsum and laterosacral ridging, the posterior thigh with obscure darker mottling, and greenish surface was black, marbled distally with ash, ventral coloring; juveniles also differ notably and the dark seat patch was continuous onto from adults in that traces of suprascapular the lower posterior surface of the thigh. plicae can be discerned on close examination Comparison of the holotype of Lithodytes in most specimens (fig. 13). A juvenile male 10 AMERICAN MUSEUM NOVITATES NO. 3357 Fig. 3. Diagrammatic representation of dorsal-ridge patterns in the Eleutherodactylus biporcatus group. A: E. gulosus, no definite dorsal plicae in adults (although weak, short anterior suprascapular plicae are discernible in most juveniles; see fig. 13). B: E. biporcatus s.s. (maussi auctorum), paired suprascapular and weak laterosacral plicae. C: E. rugosus, paired suprascapular plicae, no laterosacral plicae. D, E: E. megacephalus has chalice-shaped or hourglass-shaped suprascapular plicae, with geo- graphic variation in the laterosacral plicae or linear series of closely set tubercles; the V-shaped later- osacral pattern (D) is common from Nicaragua through Costa Rica, tending to be replaced in western Panama by nonconverging (E) parallel ridges. Two additional species, Eleutherodactylus aphanus (fig. 8) and E. opimus (figs. 17-19), resemble E. rugosus (C) in lacking paired laterosacral plicae or tuber- culate ridges, although these species are overall much less tuberculate than rugosus (fig. 20). (KU 115221, 28 mm SVL) from the Atlantic found in allied forms (see Systematic Syn- versant of Panama (juncture of Rio Claro and opsis) further supporting recognition of Rio Changena, Bocas del Toro, 910 m) also Eleutherodactylus gulosus (Cope, 1875) as a appears to be conspecific with the holotype valid species. and Fortuna Dam series; it has a pustulate dorsum and resembles the subadult female in IDENTITY OF STRABOMANTIS coloration. BIPORCATUS W. PETERS, 1863 We conclude that there can be little doubt that the Costa Rican holotype of Lithodytes The syntypic series of Strabomantis bi- gulosus and recent specimens from western porcatus consists of four specimens, ZMB Panama represent a species distinct from all 3330 (3 adults) and ZMB 3322 (a juvenile), related broad-headed Eleutherodactylus; it is all purportedly from “‘Veragua’’. The adults, the only one in which dorsal ridging is com- hereafter referred to as ZMB 3330A, B, and pletely lacking (adults) or very vague (ju- C, are about 67, 60, and 55 mm SVL and veniles). In addition, this species attains a the juvenile is 32 mm (fig. 4). much larger size than any other broad-head- These frogs resemble the several broad- ed form (table 2). Features of juvenile and headed Eleutherodactylus described from adult coloration are also different from those Central America in having cranial crests on