<<

Herpetologica, 50(3), 1994, 282-295 (? 1994 by The Herpetologists'League, Inc.

FOUR SPECIES OF BRAZILIAN POISON RELATED TO PICTUS (DENDROBATIDAE): AND NATURAL HISTORY OBSERVATIONS

C'ELIo F. B. HADDAD' AND MARCIO MARTINS2 'Departamento de Zoologia, Instituto de Biociencias, Universidade Estadual Paulista, Caixa Postal 199, 13506-900 Rio Claro SP, Brasil 2Laborat6rio de Zoologia, DB/ICB, Universidade do Amazonas, 69077-000 Manaus AM, Brasil

ABSTRACT: We presentmorphological and naturalhistory data on Epipedobatespictus and threeadditional taxa currently regarded as synonymsof the former:Epipedobates braccatus, E. flavopictus,and E. hahneli.Topotypes of the fourspecies were examined.Morphological data on tadpolesand adults,in additionto distributionpatterns and vocalizations,indicate that the four taxarepresent distinct species. Two groupsare evidentwithin them: (1) the slender-bodied,Am- azonianE. hahneliwith small flash marks on the thighsand a low-finnedtadpole, and (2) the robust bodied, almost extra-AmazonianE. braccatus,E. flavopictus, and E. pictus with large flash marks on the thighsand tadpoleswith finsof moderateheight. Key words: Taxonomy;Amphibia; Dendrobatidae; Epipedobates; Vocalizations; Distribution; Tadpoles;Natural history

RECENTLY, Myers (1987) proposed the more (Rondonia), Brazil and used Fitzin- Epipedobates to accommodate cer- ger's label names, braccatus tain species of Dendrobates, including and Dendrobates eucnemis, respectively, those in the pictus group. Seven species for these populations. Cope (1887) de- comprising this group were reviewed by scribed additional specimens from Cha- Silverstone (1976): E. bolivianus, E. ingeri, pada dos Guimaraes, Mato Grosso, Brazil, E. parvulus, E. petersi, E. pictus, E. pul- and also named them Dendrobates brac- chripectus, and E. smaragdinus. Vigle and catus. Boulenger (1883) described Den- Miyata (1980) described E. erythromos and drobates hahneli from Yurimaguas, Peru, suggested that this species is closely allied and A. Lutz (1925) described Hylaplesia to E. ingeri. Myers and Burrowes (1987) flavopicta from Belo Horizonte, Minas tentatively included a new species, E. an- Gerais, Brazil. dinus, in the pictus group, suggesting a B. Lutz (1952) briefly reviewed the spe- close relationship between this species and cies above and considered the later four as E. erythromos. Although the pictus group subspecies of Dendrobates pictus. Follow- might not be a natural group (Silverstone, ing B. Lutz, Silverstone (1976) conserva- 1976), authors are conservatively adopting tively regarded all the above species as the scheme proposed by Silverstone (1976) conspecific. Following these two authors, while awaiting revisions of the species in- the species that occur throughout northern cluded therein. South America have been referred to un- Five taxa conservatively considered as der the name Dendrobates pictus in recent E. pictus in the literature are reviewed decades (e.g., Duellman, 1978; Heatwole here (see Frost, 1985). Hylaplesia picta et al., 1965; Lescure, 1976; Myers and Bur- was described by Bibron (in Tschudi, 1838) rowes, 1987; Schliiter, 1980; Stebbins and based on material from Santa Cruz de la Hendrickson, 1959; Vigle and Miyata, Sierra, Bolivia. Later Steindachner (1864) 1980). Haddad et al. (1988) and Myers briefly described specimens from two pop- (1987) cited Epipedobatesfiavopictus as a ulations, one from Mato Grosso (probably valid species, although they did not com- Sao Vicente) and the other from Rio Ma- ment on this decision, and Caldwell and

282 September 1994] HERPETOLOGICA 283

Myers (1990) suggested that "the current Dendrobates braccatus Cope, 1887:53-54. name Epipedobates pictus is still being ap- Holotype: ANSP (Academy of Natural plied to two or more distinct species." The Sciences, Philadelphia) 13414, Chupada changes proposed here were anticipated (= Chapada dos Guimaraes),Mato Gros- in a popular account by Martins and Sa- so, Brazil. zima (1989). Dendrobates pictus braccatus, B. Lutz, In this paper, we analyze museum spec- 1952:599-601, pl. II-IV, Figs. 2, 5, and 6. imens and present field data and obser- pictus (part.), Silverstone, vations on habitat and vocalizations of E. 1976:38-42, Fig. 9E (only the syntypes braccatus, E. hahneli, E. flavopictus, and of Dendrobates braccatus Steindachner, E. pictus. We propose E. eucnemis as a 1864, and the holotype of Dendrobates junior synonym of E. pictus. Except for E. braccatus Cope, 1887). eucnemis, we analyzed specimens from the Epipedobates braccatus, Martins and Sa- type localities to assure proper identifica- zima, 1989:34. and Haddad, 1990). tion (Martins Description. -A small Epipedobates (18.3-21.8 mm SVL, sexes grouped) with MATERIALS AND METHODS slightly granularskin (Fig. 1). Small finger We measured snout-vent length (SVL), discs (Fig. 1), first finger longer than sec- head width (HW), head length (HEL), ond. Measurements and mean ratios are thigh length (THL), and tibia length (TIL) given in Tables 1 and 2. In life, E. brac- of adults with a caliper. We took additional catus has a dark brown dorsum with pale measurementsof adults with an ocular mi- brown marbling and yellowish spots (some crometer: foot length (FL), hand length specimens lack dorsal spots); black flanks; (HAL), eye diameter (ED), eye to nostril two yellowish white to yellow dorsolateral distance (EN), internarial distance (ID), stripesextending from tip of snout to groin; and tympanum diameter (TD). In order whitish stripes on upper lips extending to avoid damaging the material examined, from below nares to axillae; dorsolateral sex was not determined, except for indi- stripes become orange to reddish orange viduals carrying tadpoles. We also took flash marks that extend along the upper measurements of tadpoles with an ocular and lower surfaces of the thighs; orange to micrometer.Gosner's (1960) table was used reddish orange flash marksare also present to identify tadpole stages. Tooth row for- on the proximal half of the lower surface mula notation of tadpoles follows Altig of tibiae and behind axillae; upper surfaces (1970). Specimens analyzed are listed in of limbs brown with darkerirregular trans- Appendix I. verse stripes; belly brownish white with We recorded vocalizationsof E. hahneli scattered small black spots concentrated in and E. pictus with a Sony TCM 2 tape the ventrolateral region; chest and throat recorder and a Sony F-VlOT microphone brown (specimens from Chapada dos Gui- and those of E. flavopictus with a Uher maraes, MT, Brazil). Preserved specimens 4000 Monitor tape recorder and a Uher from Barra do Bugres, MT, have whitish M538 microphone. The recorded tapes spots on the dark brown dorsum and/or were analyzed in a Voice Identification two dorsal stripes formed by the juxtapo- Series 700 sound spectrographwith a wide sition of spots. In preservative, all colors, band filter (300 Hz). except the darker, fade to brownish white. SPECIES ACCOUNTS Diagnosis.-Epipedobates braccatus is distinguished from E. pictus and E. hah- Epipedobates braccatus neli by having a dark brown dorsum (in- (Steindachner, 1864) stead of deep black), by the lack of blueish Dendrobates braccatus Steindachner, marbling on the belly, and by a more ro- 1864:258-259. Two syntypes: NHMW bust body. From E. flavopictus by having (NaturhistorischesMuseum, Wien) 3818: the upper surfaces of limbs brown with 1-2, Sao Vicente, Mato Grosso, Brazil. darker transverse stripes (instead of black 284 HERPETOLOGICA [Vol. 50, No. 3

A

E~~~~~

FIG. 1.-Dorsal and ventral views of Epipedobatesbraccatus (ZUEC 9143) from the type locality, Chapada dos Guimar&es,Mato Grosso,Brazil. with bright yellow spots), brownish white white to yellow dorsolateral stripes (in- belly with scattered black spots (instead of stead of white), and larger, orange to red- yellowish white with black marbling), by dish orange flash marks on the groin (in- relatively larger eyes and smaller tympa- stead of yellow). From E pictus also by na, and by smaller size. From E. hahneli having dark stripes on the upper surfaces also by more robust members, yellowish of limbs and by smaller fingers and toes.

TABLE 1.-Measurements of four Brazilian species of Epipedobates: mean and standard deviation (x ? SD) and range (in parentheses).

E. braccatus E. flayoictus E. hneli E. pictus (n 9) (n =38) (n =64) (n 6) SVL 20.19 + 1.32 26.66 ? 2.22 19.02 ? 1.50 22.20 + 0.83 (18.3-21.8) (21.4-30.5) (15.3-23.0) (21.3-23.1) HW 6.26 ? 0.46 8.03 ? 0.72 5.76 ? 0.41 6.38 ? 0.17 (5.7-6.9) (6.6-9.6) (4.8-6.7) (6.2-6.6) HEL 6.60 ? 0.47 8.70 ? 0.63 6.35 ? 0.41 7.03 ? 0.14 (6.0-7.2) (7.5-10.0) (5.3-7,2) (6.9-7.2) THL 9.59 ? 0.68 12.45 ? 0.85 8.97 ? 0.67 10.43 ? 0.43 (8.9-10.9) (10.3-13.9) (7.1-10.4) (10.0-11.0) TIL 10.12 ? 0.82 12.99 ? 0.92 9.33 ? 0.64 10.75 ? 0.33 (9.3-11.6) (10.8-14.6) (7.5-10.7) (10.2-11.0) FL 9.07 ? 0.67 11.90 ? 0.89 8.78 ? 0.64 10.62 ? 0.28 (8.3-10.1) (10.1-14.0) (7.0-10.4) (10.3-11.0) HAL 5.59 ? 0.61 7.08 ? 0.44 5.32 ? 0.34 6.38 ? 0.10 (4.8-6.5) (6.2-8.0) (4.5-6.2) (6.3-6.5) ED 2.50 ? 0.12 2.73 ? 0.29 2.35 ? 0.19 2.42 ? 0.08 (2.3-2.7) (2.3-3.4) (2.1-3.1) (2.3-2.5) TD 1.16 ? 0.15 1,44 ? 0.26 1.15 ? 0.10 1.12 0.04 (0.9-1.4) (0.9-1.9) (0.9-1.4) (1.1-1.2) EN 2.23 ? 0.14 2.71 ? 0.24 1.85 ? 0.15 2.25 ? 0.08 (2.1-2.5) (2.2-3.2) (1.6-2.3) (2.1-2.3) ID 2.52 ? 0.22 3.08 ? 0.28 2.45 ? 0.17 2.62 ? 0.12 (2.3-2.9) (2.4-3.6) (2.1-3.0) (2.4-2.7) September 1994] HERPETOLOGICA 285

TABLE 2.-Mean ratios and standard deviation (x ? SD) and range (in parentheses) of four Brazilian species of Epipedobates.

E. braccatus E. flavopictus E. hahneli E. pictus (n =9) (n =38) (n =64) (n =6) HW/SVL 0.31 ? 0.01 0.30 ? 0.02 0.30 + 0.04 0.29 ? 0.01 (0.30-0.32) (0.25-0.32) (0.27-0.33) (0.28-0.30) HEL/SVL 0.33 ? 0.01 0.33 ? 0.01 0.33 ? 0.01 0.32 ? 0.01 (0.30-0.34) (0.30-0.35) (0.30-0.36) (0.30-0.33) THL/SVL 0.47 ? 0.01 0.47 ? 0.02 0.47 ? 0.06 0.47 ? 0.01 (0.46-0.50) (0.43-0.50) (0.43-0.52) (0.46-0.48) TIL/SVL 0.50 ? 0.02 0.49 ? 0.02 0.49 ? 0.02 0.49 ? 0.01 (0.47-0.53) (0.44-0.53) (0.45-0.55) (0.47-0.51) FL/SVL 0.45 ? 0.01 0.45 ? 0.03 0.46 ? 0.02 0.48 ? 0.01 (0.43-0.46) (0.39-0.51) (0.45-0.52) (0.46-0.49) HAL/SVL 0.28 ? 0.02 0.26 ? 0.01 0.28 ? 0.02 0.29 ? 0.01 (0.25-0.31) (0.24-0.29) (0.24-0.37) (0.28-0.30) TD/ED 0.46 ? 0.05 0.54 ? 0.06 0.49 ? 0.05 0.47 ? 0.02 (0.38-0.52) (0.39-0.63) (0.38-0.61) (0.44-0.50) EN/ED 0.90 ? 0.06 0.99 ? 0.09 0.79 ? 0.07 0.93 ? 0.05 (0.81-1.00) (0.82-1.20) (0.55-0.92) (0.84-1.00) ID/HW 0.40 ? 0.03 0.38 ? 0.04 0.43 ? 0.03 0.41 ? 0.02 (0.35-0.44) (0.29-0.47) (0.39-0.58) (0.38-0.42)

Distribution.-Epipedobates braccatus mm); tail length 60-65% of total length; is known from the type locality (Chapada greatest depth of the medium-finned tail dos Guimaraes, MT, Brazil), from two 16-19% of total length. In preservative, nearby localities (Barrado Bugres and Ca'- head and body dark brown above and be- ceres, MT, Brazil), from Aquidauana, MS, low; tail brown with light brown fins Brazil, and from Santa Rita do Araguaia, sparsely speckled with brown. Oral disc GO, Brazil(Fig. 2). directed ventrally;tooth row formula 2(2)/ Natural history. -Individuals of E. 3, second anterior row broadly broken braccatus were found amidst leaf litter in above the jaw sheath; jaw sheath kerati- gallery forests in the "cerrado" (savanna- nized, finely serrated;posterior jaw sheath like vegetation). On 8 December 1988, at V-shaped; oral disc laterally emarginate; Estagao Ecologica Serra das Araras, MT, anterior edges of anterior labia without one male was found carrying seven tad- papillae; lateral edges of anterior labia and poles on its dorsum, near a rivulet in the edges of posterior labia with a single row "cerrado" (U. Caramaschi, personal com- of pointed papillae (Fig. 3A). munication). Comments.-The specimens here re- Vocalization.-Unknown. ferred to as E. braccatus fit well the orig- Tadpoles.-The following description inal description provided by Steindachner is based on a series of seven tadpoles in (1864). However, two specimens (ZUEC stage 25 that were found while being car- 9015 and 9017) from Chapada dos Gui- ried by a male at Estagio Ecologica Serra maraes, MT, have intermediate color pat- das Araras, MT, on 8 December 1988. terns between E. braccatus and E. pictus. Head and body depressed (depth nearly It is possible that both of these species oc- two-thirds width), convex above and cur and may hybridize at Chapada dos slightly flattened below (Fig. 3A); eyes and Guimaraes. Additional data from this lo- nostrilsdirected dorsolaterallyfrom a dor- cality would help to solve this problem. solateral position (Fig. 3A). Head-body Two specimens from the Rio das Mortes, length 3.6-4.2 mm (x = 3.9 ? 0.2 mm); MT (MZUSP 4271 and 6486), are inter- total length 10.1-11. 3 mm (x = 10.7 ? 0.4 mediate between E. braccatus and E. fla- 286 HERPETOLOGICA [Vol. 50, No. 3

700 620 540 460 A braccatus

X flavopictus A&hahneli 00 .- o pictus

A A AmoZ0

80-

160-~ ~ ~ ~~~~~~~~~

0

FIG. 2.-Map of central and northern South America showing localities of the four species of Epipedobates.

vopictus (additional localities where inter- Phyllobates pictus (part.), Silverstone, mediate forms were found are in B. Lutz, 1976:38-41, Figs. 3D and 9D (only the 1952). holotype of Hylaplesia flavopicta A. Lutz, 1925). Epipedobates flavopictus Epipedobates flavopictus, Myers, 1987: (A. Lutz, 1925) 303. Hylaplesia flavopicta A. Lutz, 1925:139. Epipedobates flavopictus, Haddad et al., Syntypes: AL 853 and 854, USNM (Na- 1988:12, Fig. 2c. tional Museum of Natural History, Epipedobates flavopictus, Martins and Sa- Washington)96986, Belo Horizonte, Mi- zima, 1989:34, Figs. on p. 36. nas Gerais, Brazil. Dendrobates braccatus, Miranda-Ribeiro, Description.-A medium sized Epipe- 1926:179-180, text Fig. 96, pl. 20, Figs. dobates (21.4-30.5 mm SVL, sexes 3a and b. grouped) with slightly granular skin (Fig. Dendrobates pictus flavopictus, B. Lutz, 4). Finger and toe discs small (Fig. 4), first 1952:597-601, pl. I-IV, Figs. 1, 2, 5, finger slightly longer than second. Mea- and 6. surements and mean ratios are given in Dendrobates flavopictus, Cochran, 1955: Tables 1 and 2. In life, E. flavopictus has 8-9, pl. 1, Figs. G and H. a deep black dorsum with bright yellow September 1994] HERPETOLOGICA 287 A , t

'C ' ,f,otli'l'ff 'l'lt''l'ttl X

{ ,

7 00/

7 ~ ~ ~ ~ ~ ~~~~7 C~~~~~~~~~~~~~~~~~~~~~C

FIG. 3.-Lateral view and oral disc of tadpoles (stage 25) found on the dorsumof male Epipedobates:(A) E. brraccatusfrom EstaciioEcologica Serradas Araras,MT, Brazil;(B) E. flavopictus from Santanado Riacho, MG, Brazil; (C) E. hahneli from Presidente Figueiredo, AM, Brazil; (D) E. pictus from Santa Cruz de la Sierra, Santa Cruz, Bolivia. Bars under tadpoles = 5 mm and beside oral discs = 0.5 mm. round spots often arranged in two longi- yellow round and elongate spots;belly and tudinal rows; deep black flankswith sparse throat yellowish white with black, scat- bright yellow round spots; two bright yel- tered marbling (specimens from the Serra low dorsolateralstripes extending from tip do Cipo, Santana do Riacho, MG, and the of snout to groin; dorsolateral stripes be- Serrados Caraj'as,Paraopebas, PA, Brazil). come orange to red flash marksthat extend In preservative,all except the darkercolors along the anterior, superior, and posterior fade to white. surfaces of the thighs; orange to red flash Diagnosis. Epipedobates flavopictus marks also on proximal half of the lower is distinguished from the other three spe- surfacesof tibiae and behind axillae;upper cies reviewed here by the presence of yel- surfaces of limbs deep black with bright low spots on the flanks and upper surfaces 288 HERPETOLOGICA [Vol. 50, No. 3

t t ~~~~~~E~ ~ E

FIG. 4. Dorsal and ventral views of Epipedobatesflavopictus (ZUEC 3014) from Santana do Riacho, MG, Brazil.

of limbs, by relatively larger tympana, and calling in rock crevices at creek margins by larger size. It is distinguished from E. in rupestrianfields. On 8 November 1989, braccatus also by having black dorsum and at Santana do Riacho, a male found near limbs (instead of dark brown) and by rel- a creek was carrying 18 tadpoles on its atively smaller eyes; from E. hahneli also dorsum. At Paraopebas,PA, males and fe- by larger flash marks on the groin, by hav- males were found in rock crevices, within ing yellow dorsolateral stripes (instead of the vegetation on a forest waterfall, and whitish), by the lack of blueish marbling amidst leaf litter (apparently foraging) at on the belly, by the presence of yellow the margins of this creek (no calling ac- spots on the dorsum, and by relatively tivity was observed); the creek is located smaller eyes; from E. pictus also by having in a ravine within "canga" fields (a ru- black dorsum and limbs (instead of dark pestrian field vegetation on iron-rich soil: brown), by the presence of spots on the Morellato and Rosa, 1991). dorsum, and by the lack of blueish mar- Vocalization.-The advertisement call bling on the belly. of E. flavopictus consists of frequency- Distribution.-Epipedobates flavopic- modulated notes separated by about 480- tus is known from southeastern (MG, GO, 630 ms in a narrow frequency range be- and TO), northern (PA), and northeastern tween 3.2-4.2 kHz (Fig. 5A). The duration (MA) Brazil (Fig. 2). of individual notes is about 110 ms. Each Natural history.-At Santana do Ria- note is composed of six pulses. cho, Sao Roque de Minas, and Alpinopolis, Tadpoles.-The following description MG, Epipedobates flavopictus was heard is based on a series of 10 tadpoles in stage

FIG. 5.-Sonagrams of vocalizations of Epipedobates: (A) E. flavopictus from Santana do Riacho, MG, Brazil, air temperature 24 C, recording of Adao J. Cardoso (AJC) 13/07; (B) E. hahneli from Presidente Figueiredo, AM, Brazil, air temperature 28 C, recording of Marcio Martins (MRM) 08/07; (C) E. pictus from Santa Cruz de la Sierra, Santa Cruz, Bolivia, air temperature 22.5 C, recording MRM 08/06. September 1994] HERPETOLOGICA A 289

6- N

z

O-1

cr

C

24 - - - - - _ _ - Rol

TIME,01s 290 HERPETOLOGICA [Vol. 50, No. 3

25 that were found while being carried by Dendrobates pictus, (part.), Silverstone, a male at Santana do Riacho, MG, on 8 1976:38-42, Figs. 3, 9 (only the syntypes November 1989. of Dendrobates hahneli Boulenger, Head and body depressed (depth nearly 1883). two-thirds width), convex above and Epipedobates hahneli, Martins and Sazi- slightly flattened below; eyes and nostrils ma, 1989:34, top Figs. on p. 38 (wrong directed dorsolaterallyfrom a dorsolateral caption as E. pictus). position (Fig. 3B). Head-body length 3.9- Description. -A small Epipedobates 4.3 mm (x = 4.2 ? 0.1 mm); total length (15.3-23.0 mm SVL, sexes grouped) with 10.4-11.6 mm (x = 11.0 ? 0.3 mm); tail slightly granularskin (Fig. 6). Small finger length 60-64% of total length; greatest and toe discs first medium-finned tail 17-19% (Fig. 6), finger slightly depth of the longer than second. Measurements and total In head and of length. preservative, mean ratios are given in Tables 1 and 2. brown above and throatand body laterally; In E. hahneli has a black dorsum with tail cream life, belly cream; sparsely speckled pale grayish or greenish two with brown. Oral disc directed ventrally; marbling; white dorsolateral stripes extending from tooth-row formula 2(2)/3(1), second an- tip of snout to broken above groin; white stripe on the terior row broadly jaw upper from below sheath keratinized ser- lip extending the eye to sheath; jaw finely the axillae; dorsolateral become jaw sheath oral stripes a rated; posterior V-shaped; small yellow flash mark on anterior of groin; yellow disc laterally emarginate; edge flash marks also on the without lateral proximal half of upper labia papillae; edges lower surfaces of and of tibiae and behind axillae; of anterior labia edges posterior upper surfaces of of limbs brown with scat- labia with a single row round-tipped tered black papillae (Fig. 3B). Color photographs of a marbling; belly and ventral surfacesof free-living tadpole and a recently meta- limbs black with light blue mar- morphosed froglet are in Martins and Sa- bling; chest and throat black with pale marbling on the of zima (1989). margin throat (speci- mens from Presidente Comments.-Although the poor con- Figueiredo, AM, Paraopebas, PA, and Rio Branco, AC). dition of the type specimens of Hylaplesia flavopicta make comparison difficult, the Specimens from Rio Tejo, Cruzeirodo Sul, do not specimens that we analyzed agree well with AC, have dorsolateral and upper lip the original description by A. Lutz (1925). stripes. In preservative, all except the darker colors fade to white. The specimens of E. flavopictus from Pa- raopebas, PA, are somewhat distinctive Diagnosis.-Epipedobates hahneli is compared to specimens of E. flavopictus distinguished from the other three species collected near the type locality. The for- reviewed here by having slender body and mer specimens have larger eyes and red limbs, white dorsolateralstripes (instead of marks on the thighs and tibiae instead of yellowish or yellow), and smaller, yellow orange as found in specimens from MG flash markson the groin (instead of orange and GO. Field data on reproductive bi- or reddish orange). It is distinguished from E. braccatus ology, especially vocalizations, are needed also by having a black dorsum to ascertain the specific status of the pop- (insteadof dark brown), by the lack of light ulation from Paraopebas. spots on the dorsum (however, some E. braccatus lack dorsal spots), and by the presence of blueish marbling on the belly; Epipedobates hahneli from E. flavopictus also by the lack of (Boulenger, 1883) yellow spots on the dorsum, upper surfaces Dendrobates hahneli Boulenger, 1883:636, of limbs, and flanks, by the presence of pl. 57, Fig. 4. Seven syntypes: BM blueish marbling on the belly, by relatively 1947.2.15.14-20, Yurimaguas, Rio larger eyes and smaller tympana, and by Huallaga, (Loreto), Peru. smaller size; from E. pictus also by having Dendrobates pictus hahneli, B. Lutz, 1952: a black dorsum (instead of dark brown) 600-601, pl. II-IV, Figs. 1, 3, 4. and by slightly smaller size. September 1994] HERPETOLOGICA 291

FIG. 6. Dorsal and ventral views of Epipedobates hahneli (ZUEC9049) from Presidente Figueiredo, AM, Brazil.

Distribution.-Epipedobates hahneli is from 2.5-7.0 kHz (Fig. 5B). The duration known from the lowlands of Amazonian of individual notes was about 15 ms. Brazil and adjacent areas (Fig. 2). Tadpoles.-The following description Natural history.-At Paraopebas, PA, is based on a series of eight tadpoles in and Presidente Figueiredo, AM, E. hah- stage 25 that were found while being car- neli was found amidst leaf litter at margins ried by a male at Presidente Figueiredo, of forest creeks. At Paraopebas,males were AM, on 24 April 1987. not calling during the field work (July Head and body highly depressed (width 1988). At Presidente Figueiredo, males nearly two times depth), slightly convex were heard calling in two of three field above and flattened below (Fig. 3C); eyes trips (December 1986 and April 1987). and nostrils directed dorsolaterallyfrom a Males called within leaf litter and from dorsolateral position. Head-body length low (up to 30 cm) perches (on small vines, 3.2-3.9 mm (x = 3.7 ? 0.2 mm); total fallen logs, and branches). On 24 April length 10.5-12.2 mm (x = 11.8 ? 0.6 mm); 1987, two males were found carrying tad- tail length 67-70% of total length; greatest poles: one on a fallen tree over a creek had depth of the low-finned tail 13-15% of eight small tadpoles on the dorsum and total length. In preservative,head and body the other was amidst leaf litter about 40 brown, turning slightly paler on throat and m from a creek, and had two larger tad- belly; tail cream, sparsely speckled with poles on the dorsum. On the same day, a brown. Oral disc directed ventrally; labial male marked on the previous day was teeth not yet keratinized; jaw sheath shal- found a few centimeters from the site of lowly keratinized, finely serrated; poste- marking; a detailed investigation within rior jaw sheath U-shaped; oral disc later- the leaf litter revealed that this male was ally emarginate; edges of anterior labia guarding a clutch of eight eggs deposited without papillae; posteriorlabia with a sin- on a large dead leaf beneath several others. gle row of slightly pointed papillae (Fig. On one occasion, a female followed a male 3C). that emitted three-note calls (courtship Comments. -Although the analyzed call?) intermittently, but a heavy storm cotype (MZUSP 15874) of Dendrobates interrupted the observations. hahneli is in poor condition, making com- Vocalization.-The advertisement call parison difficult, the specimens analyzed of E. hahneli consisted of notes separated here conform well to the original descrip- by about 150-300 ms (x = 190 ms; SD = tion of Dendrobates hahneli Boulenger 40; n = 10) in a wide frequency range (1883), except for the flash marks that are 292 HERPETOLOGICA [Vol. 50, No. 3

4" ~~~~~~~~~~~~NA

FIG. 7. Dorsal and ventral views of Epipedobatespictus (ZUEC 9055) from the type locality, Santa Cruz de la Sierra, Santa Cruz, Bolivia. reddish orange in the original description Dendrobates eucnemis Steindachner, instead of yellow as in our specimens. A. 1864:258, pl. XIII, Fig. 2, 2a, 2b. Type Schliiter (personal communication) pho- locality: Rio Mamore, Rondonia, Brasil. tographed and W. Hodl (personal com- Dendrobates pictus eucnemis, B. Lutz, munication) filmed specimens of a species 1952:600-601, pl. II, Fig. 8. similar to ours from Panguana, Peru, with Dendrobates pictus pictus, B. Lutz, 1952: reddish orange flash marks. The differ- 599-600, pls. III and IV, Fig. 4. ences in the calls of specimens from Pan- Phyllobates pictus (part.), Silverstone, guana (Schliuter, 1980) from those de- 1976:38-42 (only the syntypes of Hy- scribed here (e.g., note duration and note laplesia picta Bibron in Tschudi, 1838 repetition rate) indicate that they may be and Dendrobates eucnemis Steindach- two species. However, we prefer to refer ner, 1864). our specimens to E. hahneli based on the Epipedobates pictus, Myers, 1987:302- original description until additional spec- 303. imens from the type locality are examined. Epipedobates pictus, Martinsand Sazima, The tadpoles of E. hahneli described here- 1989:34. in are similar to those described by Lescure Epipedobates pictus (part.?), De la Riva, (1976) under the name Phyllobates pictus, 1990:270 (at least the specimen in Fig. except for the higher number of papillae 14). on the posteriorlabia in Lescure's descrip- tion. Description. -A small Epipedobates (21.3-23.1 mm SVL, sexes grouped) with Epipedobates pictus slightly granularskin (Fig. 7). Small finger (Bibron in Tschudi, 1838) and toe discs (Fig. 7), first finger longer Hylaplesia picta Bibron in Tschudi, 1838: than second. Measurementsand mean ra- 28,71. Two syntypes MNHNP (Museum tios are given in Tables 1 and 2. In life, E. National d'Histoire Naturelle, Paris) pictus has a dark brown dorsum with two 4910, Santa Cruz (de la Sierra), Santa golden yellow dorsolateral stripes; deep Cruz, Bolivia. black flanks; a yellowish white stripe on September 1994] HERPETOLOGICA 293

the upper lip extending to the axilla; dor- Tadpoles.-The following description solateral stripes become red orange flash is based on a series of 14 tadpoles in stage marks that extend along the posterior sur- 25 that were found while being carried by face of thighs; red orange flash marks also a male at Santa Cruz de la Sierra, Bolivia, present on the proximal half of the lower on 24 January 1989. surface of tibiae and behind the axillae; Head and body depressed (depth nearly upper surfaces of limbs brown; belly and two-thirds width), convex above and ventral surfacesof limbs black with blueish slightly flattened below; eyes and nostrils white marbling (more blueish on limbs); directed dorsolaterallyfrom a dorsolateral chest and throat black with pale marbling position (Fig. 3D). Head-body length 4.0- on the margin of throat (specimen from 4.4 mm (* = 4.2 ? 0.1 mm); total length Santa Cruz de la Sierra, Bolivia). In pre- 11.2-12.2 mm (x 11.9 ? 0.3 mm); tail servative, all except the darker colors fade length 63-65% of total length; greatest to white. depth of medium-finned tail 16-19% of Diagnosis.-Epipedobates pictus is dis- total length. In preservative,head and body tinguished from E. braccatus by the lack cream, finely dotted with dark brown of spots on the dorsum (however, some above and below, except on belly; tail individuals of E. braccatus lack dorsal cream with sparsedark brown dots, except spots) and transversal dark stripes on the on ventral fin. Oral disc directed antero- upper surfaces of limbs, by the presence ventrally; tooth-row formula 2(2)/2, sec- of blueish marbling on the belly, and by ond anteriorrow broadly broken above the longer fingers and toes. It is distinguished jaw sheath; jaw sheath keratinized, finely from E. flavopictus by having a darkbrown serrated; posterior jaw sheath V-shaped; dorsum (instead of black), by the lack of oral disc laterally emarginate; edges of an- spots on the dorsum, upper surfaces of terior labia without papillae; edges of pos- limbs, and flanks, by the presence of blue- terior labia with a single row of slightly ish marbling on the belly, by relatively pointed papillae (Fig. 3D). smaller tympana, and by the smaller size; Comments.-De la Riva (1990) report- from E. hahneli by larger flash marks on ed the occurrence of Epipedobates pictus the groin, by having a dark brown dorsum at several localities in Bolivia and figured (instead of black), and by slightly larger an individual from Santa Cruz that per- size. fectly fits the live specimens that we ex- Distribution.-Epipedobates pictus is amined from this same locality. However, known from the type locality, Santa Cruz some of the specimens cited by De la Riva de la Sierra, in eastern Bolivia, and Co- (1990) may be E. hahneli, especially those rumba'and Xavantina, MS, in southwest- from Amazonian lowlands adjacent to ern Brazil (Fig. 2). Rondonia and Acre, Brazil, where E. hah- Natural history.-At Santa Cruz de la neli is known to occur; in fact, it will not Sierra, males of E. pictus were heard call- be surprising to find these two species in ing among dead tree branches on the leaf sympatry in these regions. We did not ex- litter of the Rio Pirai gallery forest on 24 amine specimens from the Rio Mamore January 1989 (during a light rain). One (type locality of Dendrobates eucnemis), male collected at the margins of a small but the syntype illustratedin Steindachner forest creek with sandy banks was carrying (1864, pl. XIII, 2, 2a, 2b) is quite similar 14 tadpoles on its dorsum. to what we regard here as E. pictus, es- Vocalization.-The advertisement call pecially the long flash mark on the thighs of E. pictus consists of frequency modu- and the ventral pattern. Silverstone (1976) lated notes separated by about 220-280 ms illustrated a tadpole of Phyllobates pictus (x = 258 ms; SD = 22; n = 6) in a narrow from Bolivia that is similar to those de- frequency range between 3.4-4.3 kHz (Fig. scribed here, except for the higher exten- 5C). The duration of individual notes is sion of the row of papillae on the anterior about 50 ms. labia and the presence of a third row of 294 HERPETOLOGICA [Vol. 50, No. 3 labial teeth on the posterior labia in the pecies com as outras correntemente former. colocadas no grupo de E. pictus, incluindo E. hahneli, e de dificil analise no presente DISCUSSION est'agio de conhecimento e merece revisoes Two groups are evident within the four futuras. species reviewed here: (1) Epipedobates Acknowledgments.-We thank W. C. A. Boker- hahneli, a slender-bodied Amazonian spe- mann, U. Caramaschi, A. J. Cardoso, J. Jim, C. Striiss- cies with small flash marks on the thighs mann, W. Tomas, M. Uetanabaro, and P. E. Vanzolini and a strongly depressed and low-finned for loaning and/or permitting analysis of material tadpole, and (2) E. pictus, E. braccatus, under their responsibility. U. Caramaschi, A. J. Car- and E. doso, I. Sazima, C. Striissmann, and M. Uetanabaro flavopictus, robust-bodied nearly kindly provided field data and allowed analysis of extra-Amazonian species with large flash material that they collected. J. P. Caldwell kindly marks on the thighs and a slightly de- provided field notes on E. braccatus.Invaluable help pressed and medium-finned tadpole. in the field was provided to M. Martins by S. Egler, Studiesbased only on morphologicaldata F. Marques, and J. Rocha at Presidente Figueiredo, by C. Striissmann at Carajias,and Kiko and F. Peralta seem to be of little value when dealing at Santa Cruz de la Sierra. I. Sazima critically read with dendrobatids; these studies often re- earlier drafts of the manuscript. J. Vielliard provided sulted in a large number of confusing and valuable help in producing the sonagrams. J. Somera probably erroneous synonymies (e.g., B. expertly made the drawings of the Epipedobates Lutz, 1952; Data on adults. The study at Paraopebas was done during field Silverstone, 1976). courses at the Serra dos Carajas with logistic support habitat, vocalization, and color in life, be- provided by Companhia Vale do Rio Doce (contract sides morphology,like those presentedhere 220/82 with CNPq). and by C. Myers and colleagues in several papers (e.g., Caldwell and Myers, 1990; LITERATURE CITED Myers and Burrowes, 1987; Myers and ALTIG, R. 1970. A key to the tadpoles of the con- Daly, 1979) are more useful to assureprop- tinental United States and Canada. Herpetologica er identifications.Furthermore, analysis of 26: 180-207. BOULENGER, G. A. 1883. On a collection of frogs specimens from type localities is highly from Yurimaguas, Huallaga River, northern Peru. recommended (Martins and Haddad, Proc. Zool. Soc. London 1883:635-638. 1990). CALDWELL, J. P., AND C. W. MYERS. 1990. A new poison from Amazonian Brazil, with further RESUMO revision of the quinquevittatus group of Dendro- bates. Am. Mus. Novit. 2988:1-21. Apresentamos dados morfologicos e de COCHRAN, D. M. 1955. Frogs of southeastern Brazil. historia natural de Epipedobates pictus e Bull. U.S. Nat. Mus. 206:1-423. mais tres especies correntemente consi- COPE, E. D. 1887. Synopsis of the Batrachia and deradas como sinonimos daquela: Epipe- Reptilia obtained by H. H. Smith, in the province of Mato Grosso, Brazil. Proc. Am. Philos. Soc. 24: dobates braccatus, Epipedobates flavopic- 44-60. tus e Epipedobates hahneli. Topotipos DE LA RIVA, I. 1990. Lista preliminar comentada e/ou material tipo das quatro especies fo- de los anfibios de Bolivia con datos sobre su distri- ram examinados. Dados morfologicos de bucion. Boll. Mus. Reg. Sci. Nat. Torino 8:261-319. girinos e adultos, alem dos padroes de dis- DUELLMAN, W. E. 1978. The biology of an equa- torial herpetofauna in Amazonian Ecuador. Misc. tribuigcao,indicam que os quatro taxons Publ. Mus. Nat. Hist. Univ. Kansas 65:1-352. aqui considerados representam especies FROST, D. R. (Ed.) 1985. Species of the distintas. Dois grupos sao evidentes entre World. Allen Press, Lawrence, Kansas. estas especies: (1) E. hahneli, amazonico, GOSNER, K. L. 1960. A simplified table for staging com corpo delicado, pequenas manchas anuran embryos and larvae with notes on identi- fication. Herpetologica 16:183-190. coloridas nas coxas e girinos com nada- HADDAD, C. F. B., G. V. ANDRADE, AND A. J. deiras pouco desenvolvidas, e (2) E. brac- CARDOSO. 1988. Anfibios anuros no Parque Na- catus, E. flavopictus, e E. pictus, prati- cional da Serra da Canastra, Estado de Minas Ge- camente extra-amazonicos, com corpo rais. Brasil Florestal 64:9-20. robusto,grandes manchas coloridasnas co- HEATWOLE, H., H. SOLANO, AND A. HEATWOLE. 1965. Notes on from the Venezuelan xas e girinos com nadadeiras moderada- Guayanas with description of two new forms. Acta mente desenvolvidas. A relacao destas es- Biol. Venez. 4:349-364. September 1994] HERPETOLOGICA 295

LESCURE, J. 1976. Contribution ai l'etude des Am- Anfibios do Instituto Nacional de Pesquisas da Ama- phibiens de Guyane fran,caise. VI. Liste prelimi- z6nia, Manaus, Amazonas, Brazil), JJ (Coleqao Jorge naire des Anoures. Bull. Mus. natn. Hist. Nat., Paris, Jim, deposited at Departamento de Zoologia, UNESP, 3e ser., no. 377, Zool. 265:475-525. Botucatu, Sao Paulo, Brazil), MN (Museu Nacional, LUTZ, A. 1925. Batraciens du Bresil. Comptes Ren- Rio de Janeiro, Brazil), MZUSP (Museu de Zoologia, dus Soc. Biol. Paris 93:137-139. Universidade de Sao Paulo, Sao Paulo, Brazil), and LUTZ, B. 1952. Anfibios anuros na colegco Adolfo ZUEC (Museu de Hist6ria Natural, Universidade Es- Lutz do Instituto Oswaldo Cruz. VII Dendrobates tadual de Campinas, Campinas, Sao Paulo, Brazil). pictus flavopictus (Lutz), 1925. Mem. Inst. Oswal- Current abbreviations of Brazilian States are as fol- do Cruz 50:597-607. lows: AC, Acre; AM, Amazonas; GO, Goias; MA, Ma- MARTINS, M., AND C. F. B. HADDAD. 1990. On the ranhao; MG, Minas Gerais; MS, Mato Grosso do Sul; identity of Dendrobates quinquevittatus (Anura, MT, Mato Grosso; PA, Pari; TO, Tocantins. Dendrobatidae). Mem. Inst. Butantan 52:53-56. Epipedobatesbraccatus. -Santa Rita do Araguaia, MARTINS, M., AND I. SAZIMA. 1989. Dendrobatideos: GO: MZUSP 66691, 66726. Aquidauana, MS: ZUEC Cores e venenos. Ciencia Hoje 9(53):34-38. 9014. Aguas Quentes, MT: ZUEC 5347-5350. Barra MIRANDA-RIBEIRo,A. 1926. Notas para servirem do Bugres, MT: MN 10096-10098. Caceres, MT: ao estudo dos Gymnobatrachios (Anura) brasileiros. ZUEC 9142-9143, 9145-9146, 9148. Chapada dos Arq. Mus. Nac. Rio de Janeiro 27:1-227. Guimaraes, MT (type locality): MZUSP 37482; ZUEC MORELLATO, L. P. C., AND N. A. ROSA. 1991. Car- 9147. Jangada, MT: ZUEC 9144. acterizac,ao de alguns tipos de vegetacao na regiao Epipedobatesflavopictus.-Chapada dos Veadei- amaz6nica, Serra dos Carajas, Para, Brasil. Revta ros, Alto Paraiso de Goias, GO: MN 2711. Formosa, brasil. Bot. 14:1-14. GO: ZUEC 3917. Rio Miguel, GO: MN 726, 7091- MYERS,C. W. 1987. New generic names for some 7095. Serra Dourada, GO: ZUEC 7501. Serra Negra, Neotropical poison frogs (Dendrobatidae). Pap. GO: MZUSP 69221-69222. Amaro Leite, MA: MN Avuls. Zool. 36:301-306. 2955, 10099-10102, one specimen without number. MYERS, C. W., AND P. A. BURROWES. 1987. A new Alpin6polis, MG: ZUEC 3824-3825, 4448-4452, 8329. poison frog (Dendrobates) from Andean Colombia, Belo Horizonte, MG (type locality): AL 853-854 (Syn- with notes on a lowland relative. Am. Mus. Novit. types); MN 10108-10109; MZUSP 295. Jaboticatubas, 2899:1-17. MG: MZUSP 10971. Santana do Riacho, MG: ZUEC MYERS,C. W., AND J. W. DALY. 1979. A name for 2068, 2194-2195, 2521, 3012-3018, 3020, 3342, 3641- the poison frog of Cordillera Azul, eastern Peru, 3642, 3899, 9013. Sao Roque de Minas, MG: ZUEC with notes on its biology and skin toxins (Dendro- 4456. Serra dos Carajas, Paraopebas, PA: ZUEC 6269, batidae). Am. Mus. Novit. 2674:1-24. 6283, 9020-9035. Parana, TO: MN 2995. SCHLUYTER,A. 1980. Bio-akustische Untersuchun- Epipedobateshahneli.-Brasileia, AC: ZUEC 5790. gen an Dendrobatiden in einem begrenzten Gebiet Cruzeiro do Sul, AC: ZUEC 5021. Restauracao, Medio des tropischen Regenwaldes von Peru (Amphibia: Rio Tejo, AC: ZUEC 8462-8464, 8484. Rio Branco, Salientia: Dendrobatidae). Salamandra 16:149-161. AC: ZUEC 5599. Taumaturgo, Rio Tejo, AC: ZUEC SILVERSTONE,P. A. 1976. A revision of the poison- 8452, 8504-8506. Alvaraes, AM: MZUSP 51122. Be- arrow frogs of the genus Phyllobates Bibron in ruri, AM: MZUSP 55666-55671. Eirunepe, AM: AL Sagra (Family Dendrobatidae). Nat. Hist. Mus. Los without number. Humaita, AM: JJ 25 specimens with- Angeles Co. Sci. Bull. 27:1-53. out numbers. Igarap6 Puruzinho, Rio Madeira, AM: STEBBINS,R., AND J. HENDRICKSON. 1959. Field MZUSP 51338, 51494-51497. Itacoai, tributary of Rio studies of the amphibians in Colombia, South Javari, AM: AL without number. Lago Amana, AM: America. Univ. California Publ. Zool. 56:497-540. MZUSP 58091. Lago Badaj6s, AM: MZUSP 53622- STEINDACHNER,F. 1864. Batrachologische Mitthei- 53623. Lago Miua, AM: MZUSP 30747. Mucuripe, lungen. Vehr. Zool. Bot. Ges. Wien 14:239-288. Rio Purus, AM: MZUSP 51186. Parani-Amani, AM: TSCHUDI,J. J. 1838. Classification der Batrachier. MZUSP 58121-58125. Sao Leopoldo do Rio Ama- Petitpierre, Neuchatel. zonas, AM: INPA 1066. Terra Preta, Boca do Manaca- VIGLE,G. O., AND K. MIYATA. 1980. A new species puru, AM: MZUSP 42198. Rio Uatuma, Presidente of Dendrobates (Anura: Dendrobatidae) from the Figueiredo, AM: INPA 724-725, 732, 890, 892-895, lowland rain forests of western Colombia. Breviora 898, 901, 903-907, 936-937, 939-941, 966-970, 983- 459: 1-7. 985, 1007-1010, 1028-1031. Jacareacanga, left bank of Rio Tapaj6s, PA: AL without number. Rio Cumina- Accepted: 25 May 1993 Miri, PA: MZUSP 30768. Serra dos Carajas, Paraope- Associate Editor: David Cannatella bas, PA: ZUEC 9037-9051. Monte Alegre, PA: AL without number. Yurimaguas, Loreto, Peru: MZUSP APPENDIX I 15874. Epipedobates pictus.-Buena Vista, Santa Cruz, Specimens Examined Bolivia: MN without number. Santa Cruz de la Sierra, Museum abbreviations for the specimens analyzed Santa Cruz, Bolivia (type locality): ZUEC 9052-9057. are: AL (Colec,ao Adolpho Lutz, deposited at Museu Xavantina, MS: MN 10105. Corumba, MT: ZUEC Nacional, Rio de Janeiro, Brazil), INPA (Colegao de 3498-3499, 9018-9019.