On the systematics of the harlequin frogs (Amphibia: Bufonidae: Atelopus) from Amazonia. II: Redescription of Atelopus pulcher 1882) (ßoULENGER, from the eastern Andean versant in Peru STEFAN LöTTERS, WINFRIED HAAs, SUSANNE So-rrcK & WOLFGANG BöHME Zusammenfassung Zur Systematik der Harlekinfrösche (Amphibia: Bufonidae: Atelopus) aus Amazonien. II: Wiederbeschreibung von Atelopus pulcher (BouLENGER, 1882) vom östlichen Andenabhang in Peru. Die Systematik der neotropischen Bufoniden-GattungAtelopus ist kompliziert. Die Arten sind zum einen sehr merkmalsarm und können sich daher sehr ähnlich sehen; zugleich kann innerhalb mancher Arten die Merkmalsausprägung auch sehr variabel sein. Atelopus spumarius umschließt einen Artenkomplex, der im Amazonas-Becken sowie am Ost-Abhang der Anden und in der Guyana-Region vorkommt. In Peru können neben A. ,pumarius sensu stricto mindestens zwei weitere Arten unterschieden werden. Sie unterscheiden sich in der Lebendfärbung, Zeichnung, Adult-Größe, Hauttextur, Larvalmorphologie und im innerartlichen Kommunikationsverhalten (Winken mit dem Vorderfuß, Rufe). In dieser Arbeit wird A. pulcher vom östlichen Anden Abhang (Departamentos Loreto und San Martin) aus der Synonymie von A. spumarius heraus genommen. Es werden gepulste Rufe sowie die Larve von A. pulcher, deren oberer Homschnabel kürzer ist als der untere, beschrieben. Zudem wird ein alphataxonomischer Überblick über die Gattung Atelopus in Amazonien sowie den angrenzenden Gebieten gegeben. Schlagwörter: Amphibia: Bufonidae: Atelopus pulcher bona sp.; A. spumarius-Artenkomplex, Alphataxonomie, Bioakustik, Larve, Peru. Abstract Systematics ofthe Neotropical bufonid genus Atelopus are complicated. On the one hand, species are poor in characters and hence similar to each other; on the other hand, within some species characters are very variable. Atelopus spumarius comprises a complex of species, distributed in the Amazon basin as weil as on the eastem versant of the Andes and in the Guianan region. In Peru, besides A. spumarius sensu stricto, at least two species can be distinguished. They differ in life colour, pattem, adult size, skin texture, larval morphology and intraspecific communication behaviour (forefoot waving, calls). In this paper, A. pulcher from the eastem Andean versant (Departamentos L'oreto and San Martin) is removed from the synonymy of A. spumarius. Pulsed calls and the larva, which is characterised by having the upper beak shorter than the lower, of A. pulcher are described. In addition, an alphataxonomic overview of the genus Atelopus in Amazonia and adjacent areas is provided. Key words: Amphibia: Bufonidae: Atelopus pulcher bona sp.; A. spumarius species complex, alphataxonomy, bioacoustics, larva, Peru. 1 Introduction According to recent data, the wide-spread Amazonian Atelopus spumarius CoPE, 1871 comprises a complex of species. Allocation of populations has been difficult and confusion about the use of scientific names has occurred. Revisionary action is needed (cf. CocROFT et al. 1990, LöTTERS 1996). Unfortunately, many Atelopus populations are poor in external characters and we know that both interspecific variation can be limited and intraspecific variation can be considerable. As a result, alphataxonomic and phylogenetic relationships remain difficult. CoLOMA et al. (2000) suggested to study morphological, osteological, behavioural and bioacoustical data as well as adult Salamandra, Rheinbach, 30.09.2002, 38(3): 165-184. © 2002 Deutsche Gesellschaft für Herpetologie und Terrarienkunde e.V. (DGHT) 165 STEFAN LöTTERS, WTNFRJED RAAS, SUSANNE SCHICK & WOLFGANG BöHME Species/ Art RangeN erbreitung Comment(s )/Kommentar( e) A. andinus Pre-cordilleras, E Peru Recently elevated to the species (Deptos. San Martin level by LöTTERS & DE LA R1v A and Loreto) (1998), originally described as a subspecies of A. spumarius A. boulengeri Andes and pre-cordilleras Thoroughly redescribed by PETERS of SE Ecuador (Provs. (1973), placing A. bicolor as junior Morona-Santiago and synonym Loja), 900-2000 m A. erythropus Andes of SE Peru (Cordillera Status ofthe single known individu- Carabaya, Depto. Puno), al uncertain(cf. LöTTERS &DELA RNA ca. 1800 m 1998) A. flavescens Coastal French Guiana and Type species of Atelopus; adjacent Brazil (Estado thoroughly redescribed by LESCURE Amapa), < 100 111 1973) except intraspecific variation; possible junior synonyms are A. spumarius barbotini1 cf. KoK ( 2000) and A. vermiculatus (c f. LESCURE 1976); tadpole described by LESCURE (1981 b) A. franciscus Coastal French Guiana, Thoroughly described by LEscuRE < 50m ( 1973), but because intraspecific variation in A. flavescens requires further studies, the status of the similar A. franciscus from the same general area needs to be reanalysed (R. BOISTEL, pers. comm.) A. hahlihelos Cordillera de Cutucü, E Thoroughly described by PETERS Ecuador (Prov. Morona (1973) Santiago) , ca. 1900 m A. minutulus E versant of the Cordillera Thoroughly described by Oriental, Colo111bia (Depto. Ruiz-CARRANZA et al. (1988) Meta), ca. 1500 m A. nepiozomus Andes of SE Ecuador (Prov. Thoroughly described by PETERS Morona-Santiago), > 2000 m (1973) A. palmatus E Andean versant of Ecuador Thoroughly redescribed by PETERS (Provs. Napo and Pastaza), (1973), but CüLOMA (1997: 56) was ca. 1000-17 40 111 unable to distinguish it from A. planispina, which probably is a senior synonym A. planispina Cordillera de Cutucu and Thoroughly redescribed by PETERS Volcan Sumaco, E Ecuador (1973); see comment under A. (Provs. Morona-Santiago palmatus and Napo), ca. 500-3900 m A. pulcher E Andean versant, Peru See this paper (Depto. Loreto) and maybe ranging into adjacent Ecuador, ca. 600-900 m 166 SALAMANDRA, Rheinbach, 38(3), 2002 On the systematics of harlequin frogs. II: Redescription of Atelopus pulcher A. seminiferus Upper Amazon basin, E Peru The original description is poor; (Deptos. San Martin and/or GRAY & CANNATELLA (1985) Loreto) provided data on the holotype A. reticulatus Cordillera Azul, E Peru Thoroughly described by LöTIERS (Depto. Ucayali), ca. 1600 m et al. (2002) A. siranus Cordillera de Sira, E Peru Thoroughly described by LöTIERS (Depto. Hminuco) & HENZL (2000) A. spumarius Almost entire Amazon basin See this paper; including A. sensu lato into the Guianan region, spumarius hoogmoedi1 two ; < 100 m descriptions of different tadpoles are available (DUELLMAN & LYNCH 1969, ÜASCON 1989) A. spumarius Upper Amazon basin, E Peru See this paper sensu stricto (Depto. Loreto ), probably ranging into adjacent Brazil, Colombia and Ecuador, < 50m A. tricolor E Andean versant of SE Peru Thoroughly redescribed by (Depto. Cuzco) and adjacent LöTIERS & DE LA RlvA (1998), NE Bolivia (Deptos. La Paz placing A. rugulosus and A. and Cochabamba), ca. willimani as junior synonyms; 1250-2500 m tadpole described by LAVILLA et al. (1997) Table. 1. List of species of Atelopus currently recognized from the eastem versant of the Andes, the Amazon basin and the Guianan region. Auflistung der derzeit anerkannten Atelopus-Arten vom Osthang der Anden, dem Amazonas Becken sowie der Guyana-Region. 1 In a recent publication, LESCURE & MARTY (2000) refer to A. spumarius barbotini and A. spumarius hoogmoedi as valid names without discussion. life colour pattem and tadpole characters to assess Atelopus systernatics. In a recent contribution, LöTIERS et al. (2002) supported the former authors and described a new species out of the A. spumarius complex from Amazonian Peru: Atelopus reticulatus LöTTERS, HAAS, SCHICK & BöHME, 2002. lt is, besides morphological aspects, character ised by its specific communication behaviour, i.e. forefoot waving and pure tone calls. Ethological observations were made in captivity. Captive data on behaviour and reproduction also are available for a second Peruvian form out of the A. spumarius complex ( cf. HAAS 1995). lt originates from a site close to the type locality ofA . pule her (BouLENGER, 1882), currently a junior synonym of A. spumarius (e.g. FROST 1985). Specimens examined coincide with the type material of A. pulcher rather than with that of A. spumarius. The purpose of this paper is (i) to briefly review the alphataxo nomy of the genus Atelopus in the Amazon basin and adjacent areas, especially the A. spumarius complex, and (ii) to redescribe A. pulcher including a description of vocalisations and of its tadpole. SALAMANDRA, Rheinbach, 38(3), 2002 167 -" STEFAN LöTIERS, WTNFRJED HAAS, SUSANNE SCHICK & WOLFGANG BöHME 2 Briefreview of Atelopus alphataxonomy in the Amazon basin and adjacent areas Table 1 lists all species currently recognized (including those, treated in this paper). The known geographic ranges of Peruvian species are shown in Figure 1. The taxonomic status of the following taxa appear to be sufficiently cleared: Atelopus andinus RrvERO, 1968; A. boulengeri PERRACA, 1904; A. halihelos PETERS, 1973; A. minutulus Ruiz-CARRANZA & HERNANDEZ-CAMACHO & ARDILA, 1988; A. nepiozomus PETERS, 1973; A. planispina JrMENEZ DE LA EsPADA, 1875; A. reticulatus; A. siranus LöTTERS & HENZL, 2000; A. tricolor BouLENGER, 1902. Atelopus erythropus BouLENGER, 1903 and A. seminiferus CüPE, 1874 are not comprehensively described in comparison with modern standards (cf. CoLOMA et al. 2000) but seem tobe distinguishable species. The taxonomic status of A. palmatus ANDERSSON, 1945 remains to be clarified. Atelopus jlavescens DuMERIL & BIBRON, 1841 and A. franciscus LESCURE, 1973 "1972" are relatively weil defined but intraspecific variation in the former needs to be enlight ened. KoK (2000) provided evidence that A. jlavescens is a variable taxon. As a result, at present, validity of A.franciscus (as weil as probably ofthe current junior synonym of A. flavescens, A. vermiculatus McDIARMID, 1973) remains to be confinned (R. BmsTEL, pers. comm.). Three names listed in Table 1 have not been mentioned so far: Atelopus pulcher; A. spumarius sensu lato; A. spumarius sensu stricto. In the current context, they deserve detailed discussion. Atelopus spumarius was originally described from the upper Amazon basin in Peru (Fig. 1) . Several authors treated it as identical to the later-described A. pulcher, supposed to originate from a locality about 600 km southwest and more close to the Andean massif (Fig. 1) . However, confusion occurred, because the original material of A. spumarius is lost (e.g. RrVERO 1968, PETERS 1973, FROST 1985). LESCURE (1981 a) designated a neotype for A. spumarius. Accordingly, A. pulcher was considered a junior synonym of A. spumarius. The latter, LEscuRE (e.g. 1981 a) suggested tobe a wide-spread species ranging from the Andean foothills ofColombia, Ecuador and Peru over the Amazon basin into the Guianan region (cf. FROST 1985). Two subspecies from the north-eastem portion of the range were named: Atelopus spumarius barbotini LESCURE, 1981 which has recently been shown to be more related to ( or probably conspecific with) A. jlavescens (KoK 2000); Atelopus spumarius hoogmoedi LEscuRE 1974 "1973" (Fig. 2) which was synonymised with the nominotypical form by LESCURE & ÜASC (1986). It is our opinion and that of other authors that A. spumarius sensu LESCURE & GAsc ( 1986) consists of several species ( cf. LöTTERS 1996). CocROFT et al. (1990) provided bioacoustic evidence for a complex of species. GAscoN (1989) described an Atelopus tadpole from the Manaus region which is strikingly different from an Ecuadorian larva assigned to A. spumarius by DuELLMAN & LYNCH (1969). The population studied by ÜASCON (1989), he assigned to A. pulcher and thus revalidated this name. However, according to LöTTERS (1996), the validity of A. pulcher is not warranted. Allocations of tadpoles to either A. spumarius and A. pulcher have to be treated with care because adults have not been compared with type specimens or topotypical material of both nominal species (see description of tadpole below). For a comprehensive revision of the A. spumarius complex, additional material and data are necessary, especially from areas currently appearing as collection gaps. Our preliminary results suggest that A. spumarius sensu stricto is a relatively small species with snout-vent length 25.2 ± 1.94 mm (range 23.0-29.7 mm) in six adult females (see Appendix for specimens examined) and of about 19.3 mm in adult males (cf. AsounH & ALTTG 1987: 32), small warts behind the eye, greenish yellow to yellow 168 SALAMANDRA, Rheinbach, 38(3), 2002 On the systematics of harlequin frogs. II: Redescription of Atelopus pulcher dorsolateral bands, with reticulation or incorporated annuli, on a dark brown to black ground as weil as partly reddish venter in life, soles and palms (for neotype illustration see LöTTERS et al. 2002: Fig. 1; for illustration of living specimen see RoDRiGuEz & DuELLMAN 1994: Plate 1). Collections of specimens are very limited. At present, A. spumarius sensu stricto seems tobe restricted to the upper Amazon basin of Peru (Fig. 1) and apparently of adjacent Brazil and Colombia ( e.g. LEscuRE 1981 a, RooRiGUEZ & DuELLMAN 1994) as well as probably ofadjacent Ecuador (L.A. CoLOMA, pers. comm.). There are also populations in which individuals have !arger adult size, smooth skin, more robust bodies and life colour pattern different to A. spumarius sensu stricto ( cf. LEscuRE 1981 a; Figs. 2-3). According to our current knowledge, these populations fall into two geographic groups, one from the central Amazon basin (probably, in pati, in sympatry with A. spumarius sensu stricto) plus the Guianan region and another one from the lower eastern Andean versant of Peru and Ecuador. The fonner, we suggest to call A. spumarius sensu lato. Ventral sides of most populations included are yellowish or pinkish (versus partly reddish venter in life). The soles and palms are rarely red. Several species may be involved. One may be called A. hoogmoedi (SVL of female holotype 34.9 mm; Fig. 2), while others might have to be named as new to science. The tadpole described by GAscoN (1989) and pati of the vocalisations discussed by LESCURE (1981 a) and CocROFT et al. (1990) will have to be considered within this context. Populations of the second mentioned group of individuals from the lower eastern Andean versant differ from A. spumarius sensu lato in pattern and/or colour in life (see below ). As mentioned above, they are different from A. spumarius sensu stricto in adult Fig. 1. Map of Peru with major river systems and areas above 3000 m above sea level showing known distributions of species of Atelopus from the Amazon basin and the eastem Andean versant. Legend: 1 = Atelopus andinus; 2 = A. erythropus; 3 = A. seminiferus; 4 = A. siranus; oo 5 = A. tricolor; 6 = A. spumarius sensu stricto; 7 = A. pulcher; 8 = A. reticulatus. Localities are as listed in the Appendix and correspond to 20 unpublished data in the case of A. spumarius sensu stricto. Localities given by RrvERO (1968) 40 and allocable to A. pulcher (see text) are not considered. Type localities are surrounded by 60 squares. Karte von Peru mit den Haupt-Flusssystemen 80 und Bereichen über 3000 m NN sowie den bekannten Verbreitungen der Atelopus-Arten aus Amazonien und vom östlichen Andenab 10" hang. Legende: 1 = Atelopus andinus; 2 = A. erythropus; 3 = A. seminiferus; 4 = A. siranus; 12° 5 = A. tricolor; 6 = A. spumarius sensu stricto; 7 = A. pulcher; 8 = A. reticulatus. Fundorte sind 14° wie im Appendix aufgeführt und, im Falle von 0 200 A. spumarius sensu stricto, unpubliziert. Fund '--'----' 16" Kilometers orte, die RIVERO (1968) angibt und A. pulcher zugeschrieben werden könnten (s. Text) werden nicht berücksichtigt. Typuslokalitäten sind mit 18° Quadraten umrandet. 81° 79° 770 75° 73° 710 69° SALAMANDRA, Rheinbach, 38(3), 2002 169 STEFAN LöTTERS, W!NFRJED HAAS, SUSANNE SCHICK & WOLFGANG BöHME Fig. 2. Dorsal and ventral views of female holotype of Atelopus spumarius hoogmoedi (MNHNP A 522). Photo J. KöHLER. Dorsal-und Ventral ansieht vom Holo typus von Atelopus spumarius hoogmoedi, Weibchen (MNHNP A 522). Foto J. KöHLER. Fig. 3. Dorsal- and ventral views of two female types of Atelopus pulcher, with lectotype left (BM 1947.2.1480, 1947.2.1482). Photo J. KöHLER. Dorsal- und Ventral ansieht von zwei Typen von Atelopus pulcher (Weibchen) mit dem Lectotypus links (BM 1947.2.1480, 1947.2.1482). Foto J. KöHLER. 170 SALAMANDRA, Rheinbach, 38(3), 2002 On the systematics of harlequin frogs. II: Redescription of Atelopus pulcher size and skin texture (cf. R.!VERO 1968, PETERS 1973, LESCURE 1981 a). In our opinion they represent a distinct species. Comparisons with type specimens of A. pulcher (Fig. 3), which are supposed to originate from the Andean foothiJls of Peru (RrvERO 1968: 19; see Fig. 1), revealed that this name is applicable to those populations. Due to the lack of information on life history, tadpole morphology etc. for the many populations, we do not know how many species actuaJly are involved and therefore use for the revalidation of A. pulcher only part of the original material (BM 1947.2.1480, 1947.2.1482) as well as specimens from a nearby Peruvian locality (KU 211676-683, 212530, ZFMK 48573, 50680-685, 76243-244) (cf. Fig. 1). To avoid further confu sion, we designate BM 1947.2.1480 as a lectotype out of the five syntypes of A. pufcher. We are able to add reproductive data as well as descriptions of vocalisations and tadpoles, all obtained under laboratory conditions (based on ZFMK 76243-245; cf. HAAS 1995). For a better understanding of the A. spumarius complex, material from additional localities should be considered, accompanied by life history, tadpole data etc. Moreover, information on osteology and DNA sequencing may help to enlighten the complex, too. Alphataxonomy shall focus on (i) the status ofEcuadorian and Peruvian populations similar to A. pufcher, (ii) a fonnal redescription of A. spumarius sensu stJicto along with (iii) an analysis of the relationships of both A. pulcher and A. spumarius sensu stricto towards A. spumarius sensu lato. 3 Material and Methods Material examined, as listed in the Appendix, is harboured at AMNH (American Museum of Natural History, New York), BM (British Museum, London), CBF (Colecci6n Boliviana de Fauna, La Paz), KU (The University of Kansas, Natural History Museum, Lawrence), ICN (Instituto de Ciencias Naturales, Museo de Historia Natural, Universidad Nacional de Colombia, Santafä de Bogota), MNHNP (Museum Fig. 4. Sole and palm ofmale Atelopus pulcher from the vicinity ofTarapoto (ZFMK 50682). Line equals 2.0 mm. Fuß- und Handunterseite von einem Atelopus pulcher Männchen aus der Umgebung von Tarapoto (ZFMK 50682). Die Linie entspricht 2,0mm. SALAMANDRA, Rheinbach, 38(3), 2002 171 STEFAN LöTTERS, WfNFRIED RAAS, SUSANNE SCHICK & WOLFGANG BöHME national d'Histoire Naturelle, Paris), NHMW (Naturhistorisches Museum Wien) and ZFMK (Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn). Description scheme for adults follows LörrERS et al. (2002). We describe the webbing formula in the manner of SAVAGE & HEYER (1969) as modified by MYERS & DuELLMAN (1982) and SAYAGE & HEYER (1997). Sex determination of adults was by extemal characters used by previous authors ( e.g. PETERS 1973). Description scheme for tadpoles follows LöTIERS (2001 ). Terminology of larval features is as proposed by ALTIG & JOHNSTON (1989); developmental stages are those of GosNER (1960). Morpho metric data to the nearest 0.1 mm were obtained by measuring both adult specimens and tadpoles with dial callipers, when necessary under a dissection microscope, by the senior author. Definitions of measurements for adults follow GRA v & CANNATELLA (1985) and CoLOMA (1997); these are: SVL (snout-vent length), HDWD (head width), HLSQ (head length from the squamosal), EYDM (eye diameter), ITNA (intemarial distance), EYNO (eye to nostril distance), SW (sacrum width at widest), TIBL (tibia Jength), FOOT (foot length), HAND (hand length), THBL (thumb length). Vocalisations were recorded in captivity using a Sony WM D6C walkman and a Sony ECM 957 microphone. For analysis, Avisoft, SASLab Pro (Berlin), was used. To exclude presumed echo effects in pulsed calls, signals subsequent to the first pulse of a call with an amplitude < 40 m V were not considered. Captive conditions were described by HAAS (1995). 4 Systematics Atelopus pulcher (BouLENGER, 1882) (Figs. 3-10) Phryniscus pulcher BouLENGER, 1882: 154. Lectotype designated herein: BM 1947.2.14.80 (formerly BM 1867.6.13.19); paralectotypes: BM 1947.2.14.81-83 (formerly BM 1867.6.13.20-22), USNM (United States National Museum, Wash ington) 193574 (formerly BM 1867.6.13.23), from "Chyavetas. E. Peru". Atelopus pulcher - NIEDEN 1926: 80; RlVERO 1963: 107. Atelopus pulcher pulcher - PETERS 1973: 41 (partim?); LESCURE 1974 "1973": 997. Atelopus spumarius (non CorE, 1871)- R1vERO 1968: 19 (partim); LESCURE 1981 a: 894 (partim); HARDTNG 1983: 58 (partim); FROST 1985: 33 (partim); RoDJUGUEZ et al. 1993: 6 (partim); LörrERS 1996: 47 (partim). Atelopus spumarius spumarius (non CorE, 1871) - HAAS 1995: 2. Atelopus spumarius pulcher - HENLE 1992: 96. Lectotype: BM 1947.2.14.80, an adult female from Chyavetas, eastem Peru, which is considered tobe a misspelling for Chayahuitas ( ca. 5°50'S, 76°1 O'W), Departamento Loreto, Peru (cf. RlvERO 1968: 19); leg. Mr. HTGGTNS. Paralectotypes: BM 1947.2.14.81-83, USNM 193574, same locality and collec tion data as for lectotype. Diagnosis: A medium-sized to IargeAtelopus (SVL ofseven adult females 32.0-35.1 mm and 13 adult males 25.2-29.3 mm) that can be distinguished from all other known species of the genus by the following combination of characters: ( 1) body slender (SW/ SVL 0.23-0.34; n = 20), snout acurninate with tip gently rounded to slightly pointed; 172 SALAMANDRA, Rheinbach, 38(3), 2002 On the systematics of harlequin frogs. II: Redescription of Atelopus pulcher (2) neural spines extemally not or weakly visible; (3) hind limbs long, tibiotarsal articulation reaching at least to posterior corner of eye when leg adpressed forward along body (TlBL/SVL 0.43-0.49; n = 20); (4) foot shorter than tibia (FOOT/TIBL 0.76-0.95; n = 20); (5) tympanic membrane absent; (6) warts, spiculae or coni 1-uo completely absent; (7) foot webbing formula IO to 1-- 0 to to 1-- 1 to 2IIIO to 1+ - 2- to nv2+ to 3+ - 1 to 2-v; (8) thumb short (THBL/HAND 0.37-0.44; n = 20); (9) plantar and palmar surfaces almost smooth, with ill-defined subarticular tubercles on some phalanges; (10) in preservative and in life, dorsal body uniform brownish black with green dorsolateral band and irregular dorsal spotting or marks and similar pattern on limbs; ventral sides in life entirely reddish in females, cream with reddish postventral area in males (usually with brownish black markings in both sexes); sole and palm in life red (reddish pink or cream in preservative). Atelopus pulcher is most similar to A. spumarius sensu lato. They can usually be distinguished by dorsal pattern (c f. Figs. 2-3; LEscuRE 1981 a: Figs. 4-6) and colour in life (above usually orange, yellow or tan versus green in A. pulcher and below yellowish or pinkish in most A. spumarius sensu lato versus red); "morph C" of A. spumarius hoogmoedi (= here included in A. spumarius sensu lato) from French Guiana ( cf. LESCVRE 1981 a: 906) has a pattem similar to that of A. pulcher but is in life dorsally and ventrally yellowish versus above green and below red; green and red rarely occurs in A. spumarius sensu lato, either, but the pattem is different to A. pulcher (unpubl. observ.). Also, A. spumarius sensu stricto has similar colours (i.e. above greenish yellow to yellow and below red) but is smaller than A. pulcher (adult female SVL is 25.2 ± 1.94 mm, n = 6, versus 34.l ± 1.12, n = 7, in A. pulcher) and possesses small warts behind the eye (absent in A. pulcher). Other species with more or less similar pattem and reddish venter (in part) and/or sole and palm are A. reticulatus, A. siranus and A. tricolor (including its junior synonyms A. rugulosus NOBLE, 1921 and A. willimani DONOSO-BARROS, 1969) - all known from the eastem versaut of the Andes or outlying serranias in Peru. As adults, these three species are all smaller than A. pule her ( cf. LöTIERS & HENZL 2000, LörrERS et al. 2002) and possess warty skin ( versus smooth skin in A. pulcher). Atelopus siranus and A. tricolor lack green colour in life (present in A. pulcher), whereas A. reticulatus has yellowish green reticulation (versus green dorsolateral bands in A. pulcher). Atelopus andinus from Amazonian Peru is similar in adult size and pattern to A. pulcher (although it is tan in preservative rather than green) but A. andinus possesses warty versus smooth skin (c f. RrvERO 1968, LöTIERS & DE LA R1v A 1998). The only other Atelopus described from eastern Peru are A. erythropus and A. seminiferus. The single known (a dult?) specimen of the fonner is smaller than adult A. pulcher (SVL of A. erythropus holotype 20.4 mm), lacks dorsal pattem (at least in preservative), has a more blunt snout and small warts between eye and forearm (cf. LöTTERS & DE LA RrvA 1998). Adult A. seminiferus are considerably !arger than A. pulcher (SVL of A. seminiferus holotype 40.0 mm; GRAY & CANNATELLA 1985). In addition, specimens tentatively referred to A. seminiferus and examined by us (see Appendix) have tubercular skin and are overall dark brown to black. Atelopus boulengeri, A. halihelos and A. nepiozomus from Andean Ecuador differ from A. pulcher in life colour and pattem (A. boulengeri is brown with yellow sides and venter; A. halihelos is dorsally light brown, scattered with irregular dark brown spots and ventrally whitish; A. halihelos is above dark olive with brown spots or marbling and below orange and yellow); moreover, A. boulengeri (including its junior synonym A. bicolor NOBLE, 1921) is ]arger than A. pulcher (SVL of A. boulengeri > 40.0 mm) and A. halihelos andA. nepiozomus have dorsal and lateral warts (cf. PETERS 1973). Atelopus SALAMANDRA, Rheinbach, 38(3), 2002 173 STEFAN LöTTERS, WINFRIED HAAS, SUSANNE SCHICK & WOLFGANG BöHME Fig. 5. Atelopus pulcher from the vi cinity of Tarapoto in amplexus (ZFMK 76243-244). Photo: w. HAAS. Atelopus pulcher aus der Umgebung von Tarapoto im Ample xus (ZFMK 76243- 244). Foto: W. HAAs. palmatus and A. planispina from Ecuador may resemble A. pulcher in life colour pattern (although still different.) But A. palmatus is smaller (adult female SYL < 31.5 mm) andA. planispina possesses weil visible lateral warts (cf.PETERS 1973, LöTTERS 1996). Atelopus minutulus from Colombia can be distinguished from A. pulcher by having dorsal reticulation and weil visible lateral warts ( cf. Ruiz-CARRANZA et al. 1988). The Guianan forms A. jlavescens (here provisionally included A. spumarius barbotini and A. vermiculatus) and A. franciscus differ from A. pulcher in life colour pattern ( cf. LESCURE 1973) - both lack red ventral sides (they are pinkish to violet) and black and green dorsal sides (A. jlavescens is unifonn yellow to dark orange or yellowish, dark or black with orange to violet vermiculation; A. franciscus is uniform olive to dark brown). Fig. 6. Ventral side offemale (left) and male of Atelopus pulcher (same specimens as in Fig. 5). Photo: W. HAAS. Ventralseite eines Weibchens (links) und Männchens von Atelopus pulcher (d ie selben Tiere wie in Fig. 5). Foto: W. HAAs. 174 SALAMANDRA, Rheinbach, 38(3), 2002