Complicated Histories: Three New Species Of Poison Frogs Of The Genus Ameerega (Anura: Dendrobatidae) From North-Central Peru

Abstract

We describe three new species of poison frogs discovered on recent expeditions throughout Central Peru. Two of these were discovered from the Serranía de Contamana, a small mountain range near the town of Contamana on the Río Ucayali in Central Peru. The first of these species is similar in appearance to Ameerega petersi, but the two species differ in call and size, have allopatric distributions, and are not close relatives according to a Bayesian phylogeny. This species is also similar morphologically to A. pongoensis, although the latter lacks flash marks above the groin and has a distinct advertisement call. Our phylogenetic data suggest that this species is closely related to A. bassleri, a species which is much larger and usually possesses a brightly-colored yellow or orange dorsum. The second new species has a red dorsum and is similar in appearance to A. bassleri, A. cainarachi, and the third new species described herein, but can be distinguished by its unusual advertisement call. The third new species in this paper was discovered in the upper Huallaga Valley near the towns of Tocache and Juanjui. This species resembles certain populations of A. bassleri and the second new species (mentioned above), but can be diagnosed by its advertisement call which is significantly slower than its closest relatives; approximately one-half the speed of A. bassleri and one-quarter the speed of the second new species. Finally, we discuss the biogeography of the A. bassleri clade in light of these new discoveries. The biogeography of this clade is consistent with the disturbance-vicariance hypothesis, and climatic fluctuations were likely a major driver in the divergence and speciation of this group. We also present a phylogeny with increased sampling in the A. petersi clade and discuss the taxonomy of the group.

Full text

Accepted by M. Vences: 9 Feb. 2009; published: 23 Mar. 2009 1
ZOOTAXA
ISSN 1175-5326 (print edition)
ISSN 1175-5334 (online edition)
Copyright © 2009 · Magnolia Press
Zootaxa 2049: 1–38 (2009)
www.mapress.com/zootaxa/Article
Complicated histories: three new species of poison frogs of the genus Ameerega
(Anura: Dendrobatidae) from north-central Peru
JASON L. BROWN1 & EVAN TWOMEY2
Department of Biology, East Carolina University, Greenville NC, U.S.A.
E –mail: 1jasonleebrown@hotmail.com, 2evan.twomey@gmail.com
Abstract
We describe three new species of poison frogs discovered on recent expeditions throughout Central Peru. Two of these
were discovered from the Serranía de Contamana, a small mountain range near the town of Contamana on the Río
Ucayali in Central Peru. The first of these species is similar in appearance to Ameerega petersi, but the two species differ
in call and size, have allopatric distributions, and are not close relatives according to a Bayesian phylogeny. This species
is also similar morphologically to A. pongoensis, although the latter lacks flash marks above the groin and has a distinct
advertisement call. Our phylogenetic data suggest that this species is closely related to A. bassleri, a species which is
much larger and usually possesses a brightly-colored yellow or orange dorsum. The second new species has a red dorsum
and is similar in appearance to A. bassleri, A. cainarachi, and the third new species described herein, but can be
distinguished by its unusual advertisement call. The third new species in this paper was discovered in the upper Huallaga
Valley near the towns of Tocache and Juanjui. This species resembles certain populations of A. bassleri and the second
new species (mentioned above), but can be diagnosed by its advertisement call which is significantly slower than its
closest relatives; approximately one-half the speed of A. bassleri and one-quarter the speed of the second new species.
Finally, we discuss the biogeography of the A. bassleri clade in light of these new discoveries. The biogeography of this
clade is consistent with the disturbance-vicariance hypothesis, and climatic fluctuations were likely a major driver in the
divergence and speciation of this group. We also present a phylogeny with increased sampling in the A. petersi clade and
discuss the taxonomy of the group.
Key words: Ameerega ignipedis sp. nov., Ameerega pepperi sp. nov., Ameerega yoshina sp. nov., coalescence,
Contamana, Dendrobatid frogs, disturbance-vicariance hypothesis, incomplete lineage sorting, introgressive
hybridization, new species, Peru, poison frogs, taxonomy
Resumen
Se describe tres especies nuevas de ranas venenosas que fueron descubiertos en expediciones recientes por el Perú
Central. Dos fueron descubiertos en el Serranía de Contamana, una cordillera pequeña cerca al pueblo de Contamana en
el Río Ucayali en el parte central del Perú. La primera de estas especies es similar en su aspecto a Ameerega petersi, pero
esas dos especies se distingue por sus cantos y tamaños, tienen distribuciones alopátricos, y no están cercanamente
emparentadas en base de un filogénio Bayesiano. También es similar morfológicamente a A. pongoensis, pero la última
especie le falta puntos sobre su ingle y tiene un canto distinto. Nuestros datos filogenéticos sugieren que esta especie está
cercanamente emparentada a A. bassleri, una especie mucho más grande y que usualmente tiene un dorso amarillo o
anaranjado brillante. La segunda especie nueva es similar en su aspecto a A. bassleri, A. cainarachi, y la tercera especie
descrito aquí, pero se puede distinguirse por su canto raro. La tercera especie nueva en este papel fue descubierto en la
parte alta del valle del Río Huallaga cerca los pueblos de Tocache y Juanjui. Esta especie es parecido a algunas
poblaciones de A. bassleri y la segunda especie (mencionado arriba), pero se puede distinguirse por su canto cual es
significamente más lento de sus parientes cercanas; aproximadamente la mitad de la velocidad de A. bassleri y un cuarto
de la velocidad de la segunda especie nueva. Finalmente, discutimos la biogeografía del clado A. bassleri siguiendo estos

BROWN & TWOMEY2 · Zootaxa 2049 © 2009 Magnolia Press
descubrimientos nuevos. La biogeografía de este clado concuerde con la hipótesis disturbio-vicariancia y fluctuaciones
climáticos probablemente eran un factor capital en la divergencia y especiación de ese grupo. También presentamos una
filogenia con muestreo aumentado en el clado de A. petersi y discutimos la taxonomía del grupo.
Introduction
The past two decades have witnessed a dramatic increase in the number of described species, in part due to
advances in molecular taxonomy. Prior to molecular phylogenetics (and other genetic-based techniques),
researchers relied primarily on morphological characters to diagnose species. Unfortunately, many taxonomic
groups have limited morphological variation and the available morphological characters are often more
indicative of species groups rather than single species. Many systematic relationships based solely on
morphological relationships have since been demonstrated to be incomplete or inaccurate (e.g. Stockman and
Bond 2007); and convergences have been relatively abundant (Myers 1982, Schulte 1986, Morales 1992,
Myers et al. 1998). Molecular phylogenetic estimates often provide robust species tree estimates and have
proven a useful tool for taxonomic studies; however they should not be used an absolute barometer of species
(Stockman & Bond 2007) as it is widely recognized that under certain circumstances, species trees and gene
trees can be discordant (Pamilio and Nei 1988, Doyle 1997, McGuire et al. 2007), particularly in analyses
based solely on mitochondrial data.
Two predominant processes affecting species tree and gene tree discordances are introgressive
hybridization and incomplete lineage sorting (Patton and Smith 1994, Maddison 1997, Funk and Omland
2003), and an increasing number of studies are detecting these processes (Weisrock et al. 2005, McGuire et al.
2007). Under certain circumstances, both processes can result in similar phylogenetic patterns (Avise and Ball
1990, Morando et al. 2004). However, in many situations, the phylogeographic patterns are able to distinguish
between these two phenomena. If the pattern is a result of incomplete lineage sorting, then alleles present in a
common ancestor have been maintained in descendant species and the alleles are expected be distributed
randomly among the descendant populations. Conversely, if the resulting pattern is a result of mitochondrial
introgression, then the incongruence between mitochondrial phylogenies and the estimated species trees are
concentrated geographically near species boundaries or observed hybrid zones (McGuire et al. 2007).
Gene tree polyphyly is common in initially-arisen species; as time progresses, lineage sorting typically
converts most gene trees to a status of paraphyly and then reciprocal monophyly (Avise 2000). The rate this
occurs depends on population size and reproduction rate, however there are also instances where stable
lineages fail to sort completely despite considerable time. For example, the Brown Bear, Ursus arctos, is
rendered paraphyletic by the Polar Bear, U. maritimus (Talbot and Shields 1996). Bufo marinus, the Cane
Toad, is paraphyletic to B. paracnemis where the two clades of B. marinus display a deep phylogenetic
separation based on the uplift of the Andes (Slade and Moritz 1998). The North American Deer Mouse,
Peromyscus maniculatus, is rendered paraphyletic by the Old-field Mouse, P. polionotus (Avise et al. 1983).
Under the strictest implementation of the phylogenetic species concept, each of the cases of paraphyly could
be reconciled by elevating unique phylogenetic lineages to species. In these cases the utility of multiple
species is questionable, particularly when gene flow continues to occur between members of the two (or
more) clades, or the clades have not diverged ecologically or morphologically.
Detailed investigations of Peruvian poison frogs have consistently revealed cases where a single,
widespread ‘species’ actually represents multiple cryptic species (Myers 1982, Schulte 1986, Morales 1992,
Myers et al. 1998, Schulte 1999, Brown et al. 2006, Twomey and Brown 2008a). While certain taxonomic
groups have undergone substantial revisions and restrictions (e.g. Dendrobates quinquevittatus sensu
Silverstone), others have persisted as ‘taxonomic dustbins’ (e.g. Ranitomeya ventrimaculata, Ameerega
hahneli, A. picta). Particularly confusing has been the Ameerega petersi group, a motley assortment of frogs
sharing a dark dorsum with greenish dorsolateral stripes, which was first referred to as Phyllobates petersi and
P. smaragdinus by Silverstone (1976). Subsequent studies (Myers et al. 1998, Schulte 1999) demonstrated

Zootaxa 2049 © 2009 Magnolia Press · 3
THREE NEW AMEEREGA FROM PERU
that Ameerega (= Phyllobates) petersi sensu Silverstone in fact included two additional species (A. simulans
and A. pongoensis) which were diagnosable from A. petersi mainly on the basis of size and advertisement call
parameters. Recent exploration in central Peru has led us to the discovery of three undescribed species of
Ameerega. We discovered two of these species during a 2006 reconnoiter of Contamana, an isolated town on
Río Ucayali 125 km NNW of Pucallpa. The area between Contamana and the Brazilian border is unique in
that it contains three distinct mountain ranges (collectively known as the Sierra del Divisor) which are isolated
from the Andes and largely isolated from each other. Few herpetological surveys have occurred in these
mountains until recently, although a rapid biological inventory in 2005 (Chicago Field Museum 2006)
identified 109 species of reptiles and amphibians, about a dozen of which were new to science. However, the
westernmost of these mountain chains, the Serranía de Contamana, was neglected in their survey. This 47 km-
long chain of hills runs in a NW-SE direction, rises up to a maximum elevation of 774 m above sea level, and
is characterized by intense geothermal activity, lush rainforest, and a rich dendrobatid community.
The first of these new species represents yet another cryptic species of Ameerega from upper Amazonia
which superficially resembles members of the petersi group. It is similar in appearance to A. petersi, A.
pongoensis, A. simulans, and A. smaragdina, can be diagnosed on the basis of call and color pattern, and is a
close relative of A. bassleri, a species occurring in the Andean foothills of north-central Peru. This species
represents the fourth to be teased apart from A. petersi sensu Silverstone (1976), and is now the fifth Peruvian
Ameerega species to possess a dark dorsum and greenish dorsolateral stripes.
The second new species does not resemble any other frogs in the vicinity of Contamana, having a red
dorsum and a relatively large SVL. It is similar in appearance to A. cainarachi and A. bassleri, two species
which occur throughout the Cordillera Oriental in northern San Martín. This species has a distinct
advertisement call which can be used to distinguish it from A. bassleri, its closest relative. We discovered an
additional population of this species in 2007 in the Huallaga Canyon east of Chazuta, approximately 130 km
NW of Contamana. This population occurs in close proximity (< 5 km) to populations of A. bassleri, yet these
frogs have an advertisement call identical to that of the Contamana population.
The third new species is found in the upper Huallaga river valley (approximately between Juanjui and
Tocache). It is very similar in appearance to A. cainarachi, A. bassleri, and the second new species
(mentioned above), none of which occur in the upper Huallaga river valley. This species has a distinct
advertisement call which can be used to diagnose it from its two sister taxa (the second new species and A.
bassleri).
The purpose of this paper is to describe these three new species of Ameerega on the basis of molecular,
morphological, and call data. Also, we describe one apparent instance of mitochondrial introgression and two
apparent cases of incomplete linage sorting in the Peruvian poison frog genus Ameerega. The two instances of
incomplete lineage sorting are likely the result of a speciation event involving a small population isolated
from an ancestral stock with a large geographic distribution. The resulting daughter species renders the
mother species paraphyletic, and despite considerable time, lineage sorting has not completed. In both our
examples, the mother species possesses a considerable amount of morphological variation (within and
between populations). In the one instance where the mother and daughter species are sympatric (or nearly
sympatric; the case regarding A. bassleri and the third new species, respectively), the daughter species
possesses dramatically different call characters that have likely reinforced the species boundaries.
We also include a phylogeny containing A. petersi and A. smaragdina individuals from several new
localities. Our phylogenetic analysis supports the notion that A. petersi and A. smaragdina may be distinct
species; however, this hypothesis is not supported by ecological or morphological data. Lastly, we discuss the
biogeography of the A. bassleri clade, as it appears now that the Serranía de Contamana may have been
colonized by members of this clade on two separate occasions.

BROWN & TWOMEY4 · Zootaxa 2049 © 2009 Magnolia Press
Materials and methods
The type series of the new species are deposited in 70% ethanol in the Museo de Historia Natural San Marcos,
Universidad Nacional Mayor de San Marcos, Lima, Peru (MUSM). The following measurements were made
with mechanical calipers and a micrometer to the nearest 0.1 mm, following Myers (1982) and Brown et al.
(2006): snout-vent length (SVL), femur length from vent to lateral edge of knee (FL), tibia length from medial
edge of heel to lateral edge of knee (TL), knee-knee distance with both legs extended straight (KK), foot
length from proximal edge of metatarsal tubercle to tip of toe IV (FoL), hand length from proximal edge of
metacarpal tubercle to tip of longest finger (HaL), head length from most exposed corner of occipitum to tip
of snout (HL), head width between tympana (HW), body width under axillae (BW), upper eyelid width
(UEW), interorbital distance (IOD), internarial distance (IND); horizontal tympanum diameter (TD),
horizontal eye diameter (ED), distance from outer corner of eye to tympanum (DET), length of finger I from
proximal edge of median palmar tubercle to tip of finger disc (L1F), length of finger II from proximal edge of
median palmar tubercle to tip of finger disc (L2F), width of disc of finger III (W3D), and width of finger III
just below disc (W3F). Sex was determined by checking for presence of vocal slits.
Advertisement calls for the new species and other species used in the analyses were recorded with a Sony
DCR-JC42 camera and Sony ECM-HGZ1 gun microphone or a Marantz PMD660 recorder and Sennheiser
microphone. Calls were analyzed in Raven 1.2 (Charif 2004) and compared to vocalizations of closely related
or morphologically similar species. We measured the following characteristics: notes per second, notes per
minute, note duration, space between notes, and dominant frequency. Notes per second were calculated by
counting the numbers of notes within the center of a call (excluding start-up and ending notes which tend to be
slower) and dividing that by the period of time; notes per minute were calculated by multiplying the notes per
second calculation by 60. The note duration and space between notes were the average time of the duration of
8 notes and the average of time between those notes, respectively. Results were analyzed in SPSS 15.0 (2008)
using one-way ANOVAs. If the assumption of homogeneity of variance was not met, data were ln
transformed. Tukey’s post hoc tests were used for pairwise comparisons since sample sizes were unbalanced.
Genetic data used for the phylogenetic analyses were adopted from previous studies (Grant et al. 2006,
Roberts et al. 2006, Twomey and Brown 2008a) with the exception of 45 individuals including: five
individuals of the first new species, six individuals from the second new species, nine individuals from the
third new species, 11 A. bassleri from four localities, two A. pongoensis from a previously unknown locality,
eight A. petersi from five localities, and four A. smaragdina from the type locality. Tissue collection, DNA
extraction, amplification, and sequence analysis of new samples followed the methods of Roberts et al.
(2006).
Phylogenetic analysis was done in MrBayes using methods outlined in Twomey and Brown (2008b).
Briefly, the final phylogenetic data set included 135 individuals from 67 localities and contained 1682 bp of
the following mitochondrial regions: 12s, 16s, cytochrome b (CytB). The dataset was partitioned into five
partitions: codons (1st, 2nd, and 3rd positions for CytB) and separate, single partitions each for 12s and 16s.
MrModeltest version 2.2 (Nylander 2004) was used to identify a substitution model, nucleotide frequencies,
and optimal priors for the gamma parameter and the proportion of invariant sites for each partition. We used
two species of Allobates, four species of Colostethus, five species of Epipedobates and two species of
Silverstoneia as outgroups. MrBayes version 3.1.2 (Huelsenbeck and Ronquist 2001) was run for twelve
million generations using the following mixed models indicated by MrModeltest 2.2: 12S (nst = 6, I +
Γ
), 16s
(nst = 6, I +
Γ
), CytB-1st (nst = 2, I +
Γ
), CytB-2nd (nst = 2, I +
Γ
), CytB-3rd (nst=6,
Γ
). Negative log-
likelihood (2ln) scores were used to identify the burn-in phase at 1.8 million generations, after which the
remaining trees were summed. The standard deviation of split frequencies for the Bayesian analysis reached
0.01 to ensure consistency of the resulting tree topology. All A. smaragdina and individuals from the second
new species were sequenced twice each direction (forward and reverse) to confirm resulting haplotypes.
The historical distributions of the bassleri and petersi clades were analyzed using dispersal-vicariance
analysis (DIVA; Ronquist 1996). In the analysis of the bassleri clade, we divided the species distributions into

Zootaxa 2049 © 2009 Magnolia Press · 5
THREE NEW AMEEREGA FROM PERU
six geographic groups: upper Huallaga river valley, Cordillera Azul, Cordillera Oriental, lower Huallaga
Canyon and Pampas del Sacramento, Serranía de Contamana, and Ecuador (for outgroup, see below). The two
river barriers depicted in Fig. 7 (The Ucayali River Valley and Huayabama River Valley) were not included
because none of the species currently occur in these areas. Each taxon in the bassleri clade (individuals in Fig.
6 and Ameerega bilinguis as an outgroup) were placed into one of the six groups. In the analysis of the petersi
clade, we divided the species distributions into four geographic groups: North Panguana (all localities of north
of Puerto Inca that are within the Pachitea and Aguaytia drainages), South Panguana (all localities south of
Puerto Inca), Cordillera Oriental (including Cordillera Azul), and Bolivia (for outgroup, see below). Each
individual in the petersi clade was placed into one of the four groups (individuals in Fig. 6 and A. yungicola as
outgroup). To simplify the analysis, clades composed of many individuals of the same species and same
geographic group were reduced to a single individual.
To evaluate the environmental interchangibility and ecological divergence of Ameerega bassleri and A.
sp. nov. 3, we used ecological niche modeling (Maxent version 3.2.1; Phillips et al. 2006; Stockman and Bond
2007). Maxent was run using the following parameters: random test percentage 25%, regularization multiplier
= 1, maximum iterations = 500, convergences threshold = 0.00001, maximum number of background points =
10000. We used 107 presence points of Ameerega bassleri and 16 presence points of A. sp. nov. 3 for each
species model. The models were built from the 19 standard bioclimatic environmental (Worldclim 1.4;
Hijmans et al. 2005). Environmental layers were reduced in geographic extent to an area about 50% larger
than the known distribution of each species. This is important for selection of realistic pseudo-absence points
by Maxent (R. P. Anderson pers. comm.). The resulting models were then projected into a larger geographic
area that encompassed the distributions of both species. The environmental variables were jackknifed and
each final model was run using the variables that contributed 5% or more to the initial jackknifed model. The
resulting model was converted to a binary absence-presence model using lowest habitat suitability value of
observed present points as a cutoff (Phillips et al. 2006).
All holotype and paratype specimens are deposited in the Museo de Historia Natural San Marcos,
Universidad Nacional Mayor de San Marcos, Lima, Peru (MUSM). Addition specimens mentioned were from
the following museums: the Field Museum of Natural History, Chicago, IL, USA (FMNH) and the American
Museum of Natural History, New York, NY, USA (AMNH).
Ameerega ignipedis sp. nov.
Figures 1, 11, 16
Phyllobates petersi: Silverstone 1976 p. 37 (FMNH 56248 collected in 1947 by J. M. Schunke at “Cerro Azul, 24 km E
Contamaná”).
Holotype. MUSM 24948, an adult female (Fig. 1) collected on 19 July 2006 by M. Pepper, E. Twomey, and J.
Brown in Departamento Loreto, Peru, 17.5 km NE Contamana at the western foot of the Serranía de
Contamana, 240 m elevation, 7° 11’ 55.46” S, 74° 57’ 35.28” W. Type locality near “El Unión”, a campsite
located at the confluence of a hot-water and cold-water stream.
Paratypes. Five adults (MUSM 24947, 24949–24952) collected on 19 July 2006, same locality as
holotype.
Etymology. The species name ignipedis is a Latin adjective meaning ‘fiery-footed’, referring to the fact
that the type locality is located alongside a geothermal stream. Our campsite in the Serranía de Contamana
was located at the confluence of two streams, one of which was fed by hot springs and reached temperatures
exceeding 90° C in some places. We found A. ignipedis along the cool-water stream in low abundances, but
along the hot-water stream they were much more common.
Definition and diagnosis. Assigned to the genus Ameerega on the basis of the following: first finger
longer than second, webbing absent between the toes, dorsal skin granular (Myers 1982, Grant et al. 2006).

BROWN & TWOMEY6 · Zootaxa 2049 © 2009 Magnolia Press
This is a small species of Ameerega with an adult SVL of approximately 20–24 mm. Dorsum granular and
brown medially, black laterally; pale yellowish-green dorsolateral stripes extending from loreal region to
groin. Pale yellowish-green labial stripe present starting behind nares and terminating above forelimb as a
yellow patch. Yellow spots present and distinct above groin, most individuals (~60%) also possess a yellow
spot on the medial face of the tibia. Venter sky-blue with black marbling. Teeth present. Appressed first finger
longer than second; finger discs weakly expanded; hands and feet lacking webbing between digits.
Vocalization consists of a series of regularly-spaced notes occurring at a rate of ~2 notes per second; note
length short (97 ms), dominant frequency of 4700 Hz.
FIGURE 1. (a) Holotype of Ameerega ignipedis sp. nov. (MUSM 24948); (b) A. pepperi sp. nov. (uncollected
individual from Polvora, San Martín, Peru); (c) Holotype of A. yoshina sp. nov. (MUSM 24945); (d) A. bassleri (MUSM
26965) from type locality (Roque, San Martín, Peru).
Ameerega ignipedis is the sister taxon to a larger clade containing Ameerega bassleri and the two
additional new species described in this paper. However, A. ignipedis bears little resemblance to any of these
species, as they are larger and usually have a yellow, orange, or red dorsum. Ameerega ignipedis bears a
pattern most similar to A. pongoensis, A. petersi, A. simulans, and A. smaragdina, all of which have a brown
or black dorsum with green or yellow dorsolateral stripes. For A. ignipedis individuals that possess a spot on
the medial face of the tibia, diagnosis can be made on this character alone, as none of the four similar species
possess such spots. Individuals that lack this shank spot can be diagnosed on the basis of body size,
advertisement call, and ventral coloration. Both A. petersi and A. smaragdina have a larger SVL than A.
ignipedis (Table 1). Furthermore, the dorsolateral stripes of A. smaragdina (and many populations of A.
petersi) are emerald-green as opposed to pale yellowish-green in A. ignipedis. Both A. petersi and A.
smaragdina have ventral coloration that is green and blue (Figures 11 & 12) (vs. uniform sky-blue in A.

Zootaxa 2049 © 2009 Magnolia Press · 7
THREE NEW AMEEREGA FROM PERU
ignipedis), and A. smaragdina lacks black marbling on the venter (vs. black marbling present in A. ignipedis).
Diagnosis against A. simulans can be made since A. simulans is smaller than A. ignipedis with respect to SVL
(Table 1). Ameerega simulans also lacks spots above the axillae and groin (as opposed spots above axillae and
groin present in A. ignipedis). Ameerega ignipedis can be distinguished from A. pongoensis in that the latter
species lacks distinct yellow spots above the groin (vs. large, distinct spots in A. ignipedis), and by differences
in advertisement call (Fig. 2). Ameerega pongoensis has a call consisting of a single- or double-peep which is
repeated irregularly at a rate of one note (or note couplet) every ~2 seconds (vs. notes repeated regularly at a
rate of two notes per second in A. ignipedis). Ameerega ignipedis can furthermore be diagnosed on the basis of
20 unambiguous nucleotide substitutions in the 16s and CytB regions of the mitochondrial genome.
Measurements (in mm) of holotype. The female holotype (Fig. 1) has SVL 23.8; FL 10.5; TL 11.1; KK
20.9; FoL 9.9; HaL 6.0; HL 6.0 HW 7.1; BW 7.0; UEW 4.6; IOD 3.4; IND 2.8; TD 2.6; ED 4.1; DET 1.4;
L1F 4.7; L2F 2.9; W3D 0.9; W3F 0.6. Measurements of paratypes are given in Table 3.
Description of holotype. Size medium, SVL 23.8 mm. Dorsal skin of head, body, and hind limbs
granular; skin smooth or nearly smooth on forelimbs and smooth on sides of head and body and ventral
surfaces. In life, the dorsum is brown medially with weak black marbling; laterally the dorsum is black. A pale
yellow dorsolateral stripe present on either side beginning at nares, passing over the eyelids, and continuing
posteriorly to terminate above the groin as a bright yellow spot. Stripe starts as pale yellow and blends to
bright yellow towards the groin. Dorsolateral lines are ~50% wider at groin than head. Pale yellow labial
stripe present beginning near nares, extending posteriorly, terminating above axillae as a weakly defined
yellow patch. Flanks black from groin to snout, ventral margin is pale yellow and fades to sky-blue ventrally.
Upper forelimbs are yellow-bronze dorsally and conspicuously yellow ventrally; lower part of forelimbs
brown dorsally and sky-blue with black reticulation ventrally. Hind limbs brown with irregular black
markings dorsally; sky-blue with black reticulation ventrally. A conspicuous yellow spot is present on the
medial surface of the tibia near the knee. Ventral surface of limbs, belly, and head sky-blue with a coarse
reticulum of irregular black lines. Hands and feet brown dorsally. Iris black.
Widest part of head between jaw articulations, head narrower than body; greatest head-width 29.9% of
SVL. Tongue medium sized, oval. Premaxillary and maxillary teeth present. Vocal slits absent. Snout narrow,
sloping from lateral view; bluntly rounded from dorsal view; truncate from ventral view. Nares situated and
directed posterolaterally to the tip of snout; nares visible from front and below but not from above. Canthus
rostralis sloped, slightly rounded; loreal region nearly vertical and slightly concave. Upper eyelid 1.4 times
wider than interorbital distance. Eye large and prominent with a maximum diameter of 17.2 % of the snout
vent length, pupil rounded and horizontally elliptical. Tympanum circular, partially concealed posterodorsally,
lacking tympanic annulus; its diameter less than 63.4 % of ED; supratympanic fold absent.
Hands relatively small, length being 25.2% of SVL. Relative length of appressed fingers: I
≈
II
≈
IV < III
(numbers starting interiorly). Discs moderately expanded on all fingers but finger II, which has a weakly
expanded disc. Disc on finger III is 1.5 times wider than distal end of adjacent phalanx. A large, circular outer
metacarpal tubercle on median base of palm; a smaller inner metacarpal tubercle on base of finger I; one well
developed and prominent subarticular tubercle on fingers I and II, two on fingers III and IV.
Hind limbs relatively short; heel reaches shoulder when appressed anteriorly. Tibia length 46.6% of SVL.
Relative lengths of appressed toes I < II < V < III < IV; first toe short (but conspicuously present), barely
reaching bottom of subarticular tubercle on base of second toe, with unexpanded disc; toes II and V barely
expanded (much smaller than finger discs); toe III and IV expanded (disc 1.5 times broader than adjacent
phalanx). Moderate-sized inner and small outer metatarsal tubercles, somewhat protuberant with rounded
surfaces. One protuberant subarticular tubercle on toes I and II, two on toes III and V, and three on IV,
however the proximal tubercle on toe IV is reduced. Hands and feet lacking supernumerary tubercles, lateral
fringes, and webbing. No basal webbing on toe fringes. Tarsal tubercles absent.
In 70% alcohol the color is almost identical to the living animal described above. The only differences are
that the brown on dorsum changed to grey, cream yellow dorsolateral stripes changed to silver-white. Flash
marks changed from yellow to cream yellow. Blue coloration on the venter changed to slate-blue.

BROWN & TWOMEY8 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 2. Advertisement calls of Ameerega ignipedis sp. nov. and A. pongoensis, spectrograms for calls with
matching waveforms. (A) A. ignipedis, field recording of a male in type locality, 24 C, 5 May 2007. Dominant frequency
4583 Hz. (B) A. pongoensis, captive recording of a male from type locality (Huallaga Canyon). Dominant frequency
4233 Hz.
TABLE 1. Comparison of body size and advertisement call parameters for Ameerega ignipedis and species similar in
appearance. Although this table and Fig. 10 would suggest that A. petersi and A. smaragdina have very different calls,
we emphasize that the calls in these species are variable and do not appear to consistently differ in any parameter we
have measured.
Body size Call parameters________
Species Mean SVL SVL Range Notes/sec Dominant frequency (Hz)
A. ignipedis 22.4 (N = 7) 20.3
ʹ
24.2 1.7 4583
A. petersi 26.2 (N = 28) 23.2–30.3 9.0 3570
A. pongoensis 22.9 (N = 5) 20.5–26.0 0.4 4233
A. simulans 20.0 (N = 13) 18.9–22.4 2.0 4250
A. smaragdina 26.5 (N = 2) 26.5–26.5 5.2 3594

Zootaxa 2049 © 2009 Magnolia Press · 9
THREE NEW AMEEREGA FROM PERU
Variation in type series. The most notable variation in the type series is the yellow shank-spot present in four
individuals and absent in two. The width of the oblique dorsolateral stripe and degree of reticulation varies
between individuals (Figures 1 & 11); in some individuals this stripe blends completely with ventral
reticulation. Within populations, the anterior terminus of dorsolateral lines varies between exterior edges of
the eyelids to the superior part of the snout.
FIGURE 3. Map of study areas. Inset in bottom right shows the two primary regions dealt with in this paper. Box B
corresponds to Fig. 8. Box A corresponds to the detailed map shown here. Black dots represent A. bassleri localities, red
dots represent A. yoshina localities, and orange dots represent A. pepperi localities. All localities are in Departamento
San Martín unless stated otherwise. Ameerega bassleri: (1) Chumia (6° 37' 5.10" S, 76° 10' 12.22" W); (2) Sauce (6° 41'
41.90" S, 76° 14' 11.87" W); (3) Cainarachi Valley (6° 27' 4.54" S, 76° 17' 30.84" W); (4) Huallaga Canyon (6° 34'
18.76" S, 75° 57' 51.12" W, north side of river); (5) Chazuta (6° 33' 30.68" S, 76° 8' 44.88" W); (6) Roque (6° 20' 48.54"
S, 76° 43' 38.60" W, type locality of A. bassleri); (7) Saposoa (6° 53' 46.85" S, 76° 49' 41.41" W); (8) Alto Shima (6° 46'
15.85" S, 76° 56' 28.32" W); (9) Sisa (6° 35' 37.10" S, 76° 34' 29.41" W). Ameerega yoshina: (1) Callanayacu (6° 35'
17.61" S, 75° 55' 2.47" W; south side of river); (2) El Unión, Loreto (7° 11' 7.43" S, 74° 57' 13.12" W, type locality of A.
yoshina). Ameerega pepperi: (1) Río Abiseo near Huicungo (7° 22' 2.88" S, 76° 48' 42.81" W); (2) Chumanza (7° 34'
7.8" S, 76° 41' 34.4" W); (3) Nuevo San Martín (7° 42' 13.64" S, 76° 40' 3.15" W; (4) San Francisco (8° 18' 30.3" S, 76°
40’40' 37.6" W, type locality of A. pepperi); (5) Campanilla (7° 25' 38.6" S, 76° 39' 53.3" W); (6) Polvora (7° 53' 28.75"
S, 76° 40' 21.49" W). Note: Shucushuyacu, Loreto (shown near the top of the map) is the locality of a newly-discovered
population of A. pongoensis (Fig. 11 c & d) and represents a substantial northward extension of the known range of this
species.
Vocalizations. The advertisement call for A. ignipedis (Fig. 2) can be characterized as a ‘retarded trill’
following Lötters et al. (2003). The call consists of regularly-spaced ‘peeps’ which are repeated at a rate of
about 1.7 notes per second. Note duration is 97 ms on average and notes are spaced 287–752 ms apart.
Dominant frequency is slightly modulated, starting at 4398 Hz at the beginning of the note and ending at 4730
Hz. We witnessed a single male calling in the type locality after a heavy rain. Calling lasted for several
minutes. The call sounds very similar to A. altamazonica (in fact their calls are nearly indistinguishable),
although we have never encountered A. altamazonica in Contamana, or in any other site east of Río Ucayali.

BROWN & TWOMEY10 · Zootaxa 2049 © 2009 Magnolia Press
In the type locality we also heard an unidentified species of Allobates calling whose call resembled that of A.
ignipedis, but differed by having notes about 1000 Hz higher in pitch and half the duration of A. ignipedis.
The call differs from that of A. pongoensis (Fig. 2), which consists of irregularly-spaced ‘peeps’ repeated at a
rate of 1 note every 2–3 seconds. These notes are sometimes given in couplets. Note duration 9–11 ms,
dominant frequency 4233 Hz.
Distribution and natural history. Ameerega ignipedis is known from two localities in the Serranía de
Contamana but probably occurs more widely (Fig. 3). Because these localities lie at the foot of low
mountains, we suspect that this species may occur widely throughout the foothills of the Serranía de
Contamana, as well as other parts of the Sierra del Divisor. Few dendrobatid species are known to occur on
both sides of Rio Ucayali, so it is probable that A. ignipedis occurs only to the east of this river.
This species occurs in lower montane rainforest. Although the elevation is relatively low (240 m), this
habitat is more reminiscent of higher elevation forests in the east Andean versant than the surrounding
lowlands. For example, at the foot of the Serranía de Contamana there is an abrupt change from typical
lowland forest to a lush habitat which is characterized by abundant epiphytic plants, large tree ferns, and a
comparatively moist leaf litter. This may be in part attributable to a slight Massenerhebung effect of these
isolated hills, trapping moisture as clouds pass across the lowland plain. These hills also have a surprisingly
diverse poison frog community. In addition to A. ignipedis, we found A. hahneli, A. trivittata, A. yoshina sp.
nov., Ranitomeya lamasi, and R. ventrimaculata.
Within this habitat, we found A. ignipedis near a small stream (Fig. 15h). Most individuals were observed
and collected in the forest near the stream, although we also saw a few individuals around the vegetation
growing in the sandy banks, one of which was a male transporting tadpoles (which were preserved but later
damaged). Since we found no other bodies of water nearby, we presume that he was in the process of
depositing tadpoles in this stream. The use of flowing water for tadpole deposition is rare in Ameerega
species, although has been observed in a few species such as A. cainarachi (pers. obsv.), A. macero
(Rodriguez and Myers 1993), and A. pongoensis (M. Pepper, pers. comm.). Additionally, we found tadpoles
of A. trivittata and what appeared to be a relatively large species of Leptodactylus. Fish were observed in the
deeper pools of this stream. Males were heard calling throughout the day, particularly after rain.
Conservation status. Following the IUCN Red List criteria (IUCN 2001), this species should be listed as
Data Deficient (DD). Being known from only two localities that are less than 3 km apart, its extent of
occurrence is unknown but probably occurs more widely. It is fairly common in the type locality and its
habitat appears to be remote enough that deforestation risk is low.
Ameerega yoshina sp. nov.
Figures 1, 14, 16
Holotype. MUSM 24945, an adult female (Fig. 1) collected on 20 July 2006 by J. Brown in Departamento
Loreto, Peru, 17.5 km NE Contamana at the western foot of the Serranía de Contamana, 310 m elevation, 7°
11' 7.43" S, 74° 57' 13.12" W. Found near El Unión, on the ground near a small creek flowing into the cold-
water stream.
Paratypes. All from Peru. Loreto: Two adult females (MUSM 26953-26954) and one adult male (MUSM
26955), collected 3 May 2007 by J. Brown and E. Twomey, 283 m elevation, approximately 1 km S of
holotype locality at 7° 11' 38.40" S, 74° 57' 7.20" W. San Martín: Three adult females (MUSM 26985–26987),
collected 19 July 2007 by J. Brown, E. Twomey, and K. Fieselman, 261 m elevation, 6° 35' 17.61" S, 75° 55'
2.47" W.
Etymology. The species name is formed as an adjective, derived from the Panoan word ‘Yoshín’ which
means ‘devil or evil spirit’, in reference to the cryptic nature of this species and its haunting, penetrating call.
The Panoan language is spoken by the Shipibo-Conibo people who are indigenous to the upper Río Ucayali
and the Sierra del Divisor.

Zootaxa 2049 © 2009 Magnolia Press · 11
THREE NEW AMEEREGA FROM PERU
Definition and diagnosis. Assigned to the genus Ameerega on the basis of the following: first finger
longer than second, webbing absent between the toes, dorsal skin granular (Myers 1982, Grant et al. 2006).
This is a medium-large species of Ameerega with an adult SVL of approximately 26–35 mm. Dorsum has a
black ground color, stippled brick-red to bright red or orange, stippling less dense near sacral region. Bright
yellow dorsolateral stripes extend from the eyelid to the groin; a yellow labial stripe is also present starting
near nares and extending posterior to the axillary region. Bright yellow-green spot present above groin in
some populations, shank spots and axillary spots absent. Limbs greenish-bronze or black; venter ranges from
greenish-gold to blue, black marbling weak or absent. Teeth present. Appressed first finger longer than
second; finger discs moderately expanded, hands and feet lacking webbing between digits. Vocalization
consists of a ‘burst’ of loud, whistle-like notes, each burst consisting of 4–31 notes, given at a rate of 4.0–4.5
notes per second; dominant frequency 3210– 3060 Hz (Fig. 4, Table 2).
Ameerega yoshina can be distinguished from most other species of Ameerega by its red dorsum. It is
similar in appearance to the following species: A. bassleri, A. bilinguis, A. cainarachi, A. macero, A. pepperi
sp. nov. (description below), and A. parvula, all of which have (or can have) a red dorsum. Ameerega parvula
and A. bilinguis both lack distinct dorsolateral stripes (vs. dorsolateral stripes present and conspicuous in A.
yoshina). Ameerega cainarachi is smaller, females up to 31.3 mm SVL (vs. 35 mm SVL in A. yoshina), and
has an advertisement call consisting of a regular chain of notes which are not frequency–modulated, repeated
at a rate of 9 notes per second (vs. ‘bursts’ of frequency-modulated notes, repeated 3.7–4.1 notes per second in
A. yoshina). Ameerega macero is smaller than A. yoshina, females up to 29.5 mm SVL, and has a call
consisting of a series of harsh ‘peeps’, repeated regularly at a rate of 10 notes per second. Ameerega bassleri
and A. pepperi have an advertisement call consisting of a series of regularly-spaced whistle-like notes given
continuously for several minutes at a rate of 1.6–2.1 notes per second in A. bassleri and 0.9–1.3 notes per
second in A. pepperi (vs. bursts of 4–31 notes, given at a rate of 3.7–4.1 notes per second, each burst less than
8 seconds in total duration in A. yoshina).
TABLE 2. Relationships of call structures of A. pepperi, A. bassleri and A. yoshina. Mean ± standard deviation (min,
max, pairwise significance).
Measurements (in mm) of holotype. The female holotype (Fig. 1) has SVL 34.8; FL 16.3; TL 17.6; KK
29.5; FoL 14.6; HaL 8.3; HL 8.1; HW 10.2; BW 12.1; UEW 7.3; IOD 4.9; TD 3.2; ED 5.3; DET 0.8; L1F 7.6;
L2F 4.5; W3D 1.0; W3F 0.5. Measurements of paratypes are given in Table 4.
Description of holotype. Size large, SVL 34.8 mm (Table 4). Dorsal skin coarsely granular, pigmented
black with brick red mottling densest near the snout. Thin light-yellow dorsolateral stripes extend from
eyelids to groin; posterior to eyes the stripes blend into the red dorsum. Dorsolateral stripes slightly wider at
groin. Flanks are solid black along the entire length of the body and head. Yellow labial stripe present
extending from nares to axillae and upper forelimb. Forelimbs are weakly granular and are metallic-brown in
color. Legs are granular, dark brown with faint olive green stippling. On each leg there is a large yellow spot
present on the anterodorsal surface of the thigh. Hands, feet, and digits colored as limbs. The ventral skin is
smooth and uniformly olive-gold in color on the belly and limbs; the head and chest are slightly darker than
Species Calling rate
(notes/min) Note Duration
(sec) Silence between
notes (sec) Dominant
Frequency (Hz) N
A. pepperi 62.1 ± 9.4
(51.0–75.9, a) 0.17 ± 0.05
(0.13–0.22, a) 0.80 ± 0.20
(0.56–1.03, a) 2970 ± 130
(2800–3160, a) 6 indiv.,
4 local.
A. bassleri 95.3 ± 10.0
(95.3–127.8, b) 0.15 ± 0.03
(0.09–0.19, a) 0.42 ± 0.08
(0.30–0.42, b) 2750 ± 190
(2400–3070, b) 18 indiv.,
6 local.
A. yoshina 235.2 ± 5.6
(227.3–243.2, c) 0.13 ± 0.02
(0.09–0.15, a) 0.13 ± 0.02
(0.11–0.15, c) 3140 ± 50
(3060–3210, a) 6 indiv.,
2 local.
ANOVA F2,27=597.02,
P < 0.001
F2,21=2.54,
P = 0.102
F2,21=56.97,
P < 0.001
F2,27 = 13.83,
P < 0.001

BROWN & TWOMEY12 · Zootaxa 2049 © 2009 Magnolia Press
the belly. No distinct black markings present on the venter. Iris black.
Widest part of head between jaw articulations and tympanum, head narrower than body; greatest head-
width 29.3% of SVL. Eyes protuberant. Tongue medium sized, oval. Premaxillary and maxillary teeth present.
Vocal slits absent. Snout sloping laterally; bluntly rounded dorsally; truncate ventrally. Nares situated and
directed posterolaterally to the tip of snout; nares visible from front and below but not from above. Canthus
rostralis sloped, slightly rounded; loreal region nearly vertical and slightly concave. Upper eyelid 1.5 times
wider than interorbital distance. Eye large and prominent with a maximum diameter of 9.3 % of the snout vent
length; pupil rounded and horizontally elliptical. Tympanum circular; without tympanic annulus its diameter
is less than 60.2 % of ED; supratympanic fold present.
FIGURE 4. Advertisement calls of Ameerega yoshina, A. pepperi, and A. bassleri. (A) A. yoshina standard
advertisement call, field recording from 1.6 km SE of Callanayacu, San Martín, 25.5 C, 19 July 2007. Dominant
frequency 3244 Hz. (B) A. yoshina short call, same location, temperature, and date as previous. Dominant frequency
3141 Hz. (C) A. pepperi advertisement call, field recording from near Polvora, San Martín, 24.5 C, 8 November 2008.
Dominant frequency 2749 Hz. (D) A. bassleri advertisement call, field recording from 5.7 km NE of Roque, San Martín,
26.5 C, 16 June 2007. Dominant frequency 2714 Hz. See Fig. 3 for map of localities.

Zootaxa 2049 © 2009 Magnolia Press · 13
THREE NEW AMEEREGA FROM PERU
Hands relatively small, length being 23.7% of SVL. Relative length of appressed fingers II
≈
IV < I < III
(with I being interior-most finger). Discs moderately expanded on all fingers. Disc on finger III is twice as
wide as distal end of adjacent phalanx. A large, circular outer metacarpal tubercle present on median base of
palm; a smaller inner metacarpal tubercle on base of finger I absent; one well developed subarticular tubercle
on fingers I and II, two on fingers III and IV.
Hind limbs short, with heel of appressed limb reaching the axillary region. Tibia length 50.7% of SVL.
Relative lengths of appressed toes I < II < V< III < IV; first toe extending past subarticular tubercle on base of
second toe; discs on toes I, II, III, IV barely expanded (much smaller than finger discs), and toe V expanded
(disc 1.2 times wider than adjacent phalanx). Moderate-sized inner and outer metatarsal tubercles, protuberant
with rounded surfaces. One protuberant subarticular tubercle on toes I and II, two on toes III, IV, and V. Hands
and feet lacking supernumerary tubercles and lateral fringes. No basal webbing on toe fringes. Tarsal tubercles
absent.
In 70% alcohol the color is similar to the living animal described above, except the red dorsum changed to
black, yellow dorsolateral stripes changed to silver-white, and gold coloration on venter changed to black and
grey-green.
FIGURE 5. Call variation between Ameerega bassleri, A. pepperi and A. yoshina. Box plots show period of silence
between notes (L) and the calling rate (notes per minute) of the three species (R).
Var ia ti on . Little variation has been observed in the two known populations of A. yoshina. Frogs from
near Contamana vary somewhat with respect to dorsal coloration; some individuals are brick-red while others
are more orange. These frogs also possess a venter which varies from greenish-gold to pale yellow-green,
although black ventral markings are absent or not well-defined. Frogs from Callanayacu tend to be more
orange as opposed to red on the dorsum (Fig. 14), and have blue venters that have diffuse and irregular black
markings. The limbs vary in color from black to pale olive.
Vocalizations. The advertisement call for A. yoshina (Fig. 4) can be classified as a ‘retarded trill’
following Lötters et al. (2003). The call consists of a series of musical notes resembling short ‘whistles’ which
are given in short bursts of 4–31 notes at a rate of 3.7–4.1 notes per second. Notes are 90–156 ms long, spaced
110–150 ms apart, and frequency-modulated by about 325 Hz (notes start at 2925 Hz, end at 3250 Hz, overall
dominant frequency 3140 Hz).

BROWN & TWOMEY14 · Zootaxa 2049 © 2009 Magnolia Press
The call of A. yoshina can be very easily distinguished from the advertisement call of A. bassleri, which is
a series of whistle-like notes repeated at a much slower rate and for much longer. Our recordings of A. bassleri
have notes repeated at a rate of 1.6–2.1 notes per second for 7.5 seconds or longer, although on several
occasions we have heard A. bassleri calling continuously for several minutes at a time. Ameerega bassleri also
has an aggressive call which consists of 3–5 notes given in quick succession. However, even this aggressive
call differs from the advertisement call of A. yoshina, as notes are repeated at a slower rate (1.6 notes per
second in A. bassleri vs. over 4 notes per second in A. yoshina) and are spaced further apart (notes spaced 420
ms in A. bassleri vs. 130 ms in A. yoshina). In the field, the call differences between the two species are
immediately recognizable and fixed in a given geographic region. For example, during several days in
Callanayacu and Contamana, we never heard a call which sounded like A. bassleri. Conversely, in areas
containing A. bassleri, we have never heard a call which sounded like A. yoshina.
Distribution and natural history. Ameerega yoshina is known from two localities, one in the Serranía de
Contamana and the other in the Huallaga Canyon1 in the northern Cordillera Azul (Fig. 3). These two
localities are 130 km apart and separated by the Ucayali floodplain, which may represent a significant barrier
between these two populations. We have not found A. yoshina in lowland habitats and expect that this species
is restricted to premontane forests of the Serranía de Contamana and northeastern Cordillera Azul. At their
closest points, the Cordillera Azul and Serranía de Contamana are separated by 26 km of lowland forest (and
Río Ucayali), and it is unlikely that there is currently gene flow across this barrier. In the lower Huallaga
Canyon, we have only found A. bassleri on the north-bank and A. yoshina on the south-bank of Río Huallaga.
Therefore, it is possible that A. bassleri is absent from the northeastern Cordillera Azul and that this area is
inhabited exclusively by A. yoshina.
Ameerega yoshina occurs in undisturbed premontane forests, particularly in habitats which are adjacent to
streams (Fig. 15h & j). This species appears to have extremely specific habitat requirements, as we were only
able to find them in isolated pockets near Contamana and Callanayacu. For example, in Callanayacu we hiked
for several kilometers along Quebrada Pacuyacu (a small stream), and only heard A. yoshina calling along a
~200-m stretch of habitat adjacent to the stream. In Contamana we were able to find this species alongside
three different streams, but only in small areas along these streams. This species is exceedingly cryptic and is
nearly impossible to detect if males are not calling. In Callanayacu we spent roughly 68 person-hours
searching within the 200-m stretch of habitat and were only able to capture three individuals, despite the fact
that numerous males were calling in the vicinity the entire time we were searching. In Contamana we searched
for 60 person-hours in 2006 and 48 person-hours in 2007, yet we were only able to encounter a total of six
frogs, one of which was a metamorph. Most adults were found under logs or under piles of leaf litter, although
two were found moving about the leaf litter after a heavy rain.
Reproduction appears to take place near small streams and forest pools. In the same spot where we found
A. ignipedis transporting tadpoles, we also found one A. yoshina metamorph at the edge of the water. There
were no other water sources nearby and we assume this metamorph emerged from the stream. We also found a
shallow forest pool at the base of a small waterfall, around which many A. yoshina males were calling. In this
pool were numerous tadpoles which were A. yoshina (confimed by DNA sequencing). Males appear to call
from the leaf litter throughout the day. In Callanayacu males called incessantly the entire day, but were well-
hidden.
1. The term “Huallaga Canyon” refers to a narrow section of the river which forms between the Cordillera Oriental
and Cordillera Azul, starting at Chumia and ending when the river reaches the lowlands. Callanayacu and Chazuta are
both towns located within this section of river.

Zootaxa 2049 © 2009 Magnolia Press · 15
THREE NEW AMEEREGA FROM PERU
FIGURE 6. Our Bayesian phylogenetic hypothesis. Thicker branches represent posterior probabilities greater than 80
(nodes of particular interest to this study are labeled with observed posterior probabilities). Ameerega ignipedis sp. nov.,
A. yoshina sp. nov., and A. pepperi sp. nov. are highlighted with green, orange and red boxes, respectively. Since this
study focused on the bassleri and petersi clades, to simplify visualization of the tree we reduced individuals of other
species of Ameerega and other genera to a single representative (reducing the number of terminals to 76 from 135, and in
every case, each clade comprised a monophyletic lineage).

BROWN & TWOMEY16 · Zootaxa 2049 © 2009 Magnolia Press
Conservation status. Following the IUCN Red List criteria (IUCN 2001), we tentatively suggest that A.
yoshina should be listed as Near Threatened (NT) based on the following: (1) we estimate the extent of
occurrence to be less than 4,500 km2, (2) populations appear to be fragmented, as we have only encountered
two disjunct populations which are restricted to small areas, and (3) there has been a small amount of potential
habitat loss in the Cordillera Azul near Río Huallaga, although this species probably occurs throughout the
northern part of the Cordillera Azul National Park. Further investigation may warrant a revised classification,
as the above estimate on extent of occurrence assumes that this species occurs at elevations from 280–600 m
and also occurs in the Ojo de Contaya (Fig. 3). While this species may range widely throughout the Cordillera
Azul and Sierra del Divisor, it may just as likely be restricted to the Serranía de Contamana and extreme
northern Cordillera Azul.
Ameerega pepperi sp. nov.
Figures 1, 14–16
Phyllobates bassleri: Silverstone 1976 p. 45 (AMNH 42327, 42867, 43402 collected in 1926 and 1928 by H.
Bassler at “Pachiza”).
Holotype. MUSM 26940 (Fig. 16), an adult male collected on 14 July 2006 by M. Pepper, Provincia Tocache,
Departamento San Martín, Peru, 2 km NE of San Francisco, 980 m elevation, 8° 18' 30.3" S, 76° 40' 37.6" W.
Found on the ground near a small waterfall.
Paratypes. All from San Martín, Peru. One adult female (MUSM 26942), collected 21 July 2006 by M.
Pepper and E. Twomey, and two adult females (MUSM 26975, 26976), collected 25 June 2007 by J. Brown,
E. Twomey, and K. Fieselman, 390 m elevation, 7.0 km SW of Huicungo at 7° 21' 58.6" S, 76° 48' 44.1" W:
Two adult females (MUSM 26979, 26980), collected on 25 June by J. Brown, E. Twomey, and K. Fieselman,
384 elevation, 6.6 km N of Campanilla at 7° 25' 38.6" S, 76° 39' 53.3" W: One adult female (MUSM 26968),
collected on 15 July 2006 by M. Pepper, 570 m elevation, SW of Chumanza (alternate spelling: Shumanza) at
7° 34' 7.8" S, 76° 41' 34.4" W: One adult male (MUSM 26941), collected at same locality of holotype, 980 m
elevation, 2 km NE of San Francisco at 8° 18' 30.3" S, 76° 40' 37.6" W.
Etymology. The specific epithet is a patronym for the discoverer of this species, Mark Pepper, an
enthusiastic Canadian explorer and conservationist whose discoveries have greatly contributed to the authors’
understanding of poison frog biogeography and taxonomy.
Definition and diagnosis. Assigned to the genus Ameerega on the basis of the following: first finger
longer than second, webbing absent between the toes, dorsal skin granular (Myers 1982, Grant et al. 2006).
This is a medium-large species of Ameerega with an adult SVL of approximately 28–34 mm. Dorsum has a
black ground color, finely stippled brick-red to bright red in southern populations (e.g. San Francisco) to
coarsely stippled orange or yellow northern populations (e.g. Campanilla), stippling concentrated near head
and is less dense near sacral region. Bright yellow dorsolateral stripes extend from the eyelid to the groin; a
yellow labial stripe is also present starting near nares and extending posterior to the axillary region. Faded
yellow-green to white spot present above groin in some populations, shank spots and axillary spots absent.
Limbs greenish-bronze or black; venter blue with strong black marbling. Teeth present. Appressed first finger
longer than second; finger discs moderately expanded, hands and feet lacking webbing between digits.
Vocalization consists of a loud retarded trill of whistle-like notes given at a rate of 0.9–1.3 notes per second;
dominant frequency 2800–3160 Hz (Fig. 4, Table 2).
Ameerega pepperi can be distinguished from most other species of Ameerega by its red, orange, or yellow
dorsum. It is similar in appearance to the following species: A. bassleri, A. bilinguis, A. cainarachi, A.
macero, A. parvula, and A. yoshina, all of which have (or can have) a solid red, orange, or yellow dorsum.
Ameerega parvula and A. bilinguis both lack distinct dorsolateral stripes (vs. dorsolateral stripes present and
conspicuous in A. pepperi). Ameerega cainarachi is smaller, females up to 31.3 mm SVL (vs. 34 mm SVL in

Zootaxa 2049 © 2009 Magnolia Press · 17
THREE NEW AMEEREGA FROM PERU
A. pepperi), and has an advertisement call consisting of a regular chain of notes which are not frequency
modulated, repeated at a rate of 9 notes per second (vs. ‘bursts’ of frequency-modulated notes, repeated
0.9–1.3 notes per second in A. pepperi). Ameerega macero is smaller than A. pepperi, females up to 29.5 mm
SVL, and has a call consisting of a series of harsh ‘peeps’, repeated regularly at a rate of 10 notes per second.
Ameerega bassleri has an advertisement call consisting of a series of regularly-spaced whistle-like notes given
continuously for several minutes at a rate of 1.6–2.1 notes per second (vs. notes given at a rate of 0.9–1.3
notes per second in A. pepperi). Ameerega yoshina has an advertisement call given in bursts of 4–31 notes,
given at a rate of 3.7–4.1 notes per second, each burst less than 8 seconds in total duration (vs. notes given at
a rate of 0.9–1.3 notes per second in A. pepperi).
Measurements (in mm) of holotype. The male holotype (Fig. 16) has SVL 28.6; FL 13.4; TL 16.1; KK
28.2; FoL 13.3; HaL 7.5; HL 7.7; HW 7.8; BW 8.3; UEW 5.5; IOD 3.9; IND 4.0; TD 3.0; ED 4.2; DET 1.6;
L1F 6.9; L2F 6.1; W3D 0.8; W3F 0.6. Measurements of paratypes are given in Table 5.
Description of holotype. Size medium, SVL 28.6 mm (Table 5). Dorsal skin is coarsely granular and is
pigmented black with brick red mottling densest near the snout. Thin yellow dorsolateral stripes extend from
groin, extending dorsally around snout; posterior to eyes the stripes blend into the red dorsum. Dorsolateral
stripe is sightly wider at groin. Flanks are black extending from groin to nares. Forelimbs lack granulation and
are colored tan and flecked in olive. Yellow coloration extends up upper arm surface, anteriorly around lip.
Legs are coarsely granulated; contain yellow femoral spots, dorsal surface colored olive green. The venter is
yellow-gold with olive marbling; marbling extends primarily on throat and margins of legs. Iris black.
Widest part of head between jaw articulations and tympanum, head narrower than body. Greatest head
width 27.0 % of SVL. Eyes very protuberant. Tongue medium sized, oval. Premaxillary and maxillary teeth
present. Vocal slits present. Dorsal skin of head and body is finely granular; skin smooth or nearly smooth on
limbs and smooth on sides of head and body and ventral surfaces. Snout sloping laterally; bluntly rounded
dorsally; truncate ventrally. Nares situated and directed posterolaterally to the tip of snout; nares visible from
front and below but not from above. Canthus rostralis sloped, slightly rounded; loreal region nearly vertical
and slightly concave. Upper eyelid 1.4 times wider than interorbital distance. Eye large and prominent with a
maximum diameter of 14.6 % of SVL, pupil rounded and horizontally elliptical. Tympanum circular, without
tympanic annulus its diameter less than 73.0 % of ED, supratympanic fold absent.
Hands relatively small, length being 26.2 % of SVL. Relative length of appressed fingers I
≈
II
≈
IV < III
(numbers starting interiorly). Discs moderately expanded on all fingers. In adults, disc on finger III is 1.25
times wider than distal end of adjacent phalanx A large, circular outer metacarpal tubercle present on median
base of palm; smaller inner metacarpal tubercle on base of finger I absent; one well developed subarticular
tubercle on fingers I and II, two on fingers III and IV.
Hind limbs short, small, with heel of appressed limb reaching the interior axilla. Tibia length 56.3% of
SVL. Relative lengths of appressed toes I < II < V < III < IV; first toe terminates at subarticular tubercle on
base of second toe; toes I, II barely expanded (much smaller than finger discs), and toe III, IV, and V expanded
(disc 1.5 times broader than adjacent phalanx). Moderate-sized inner and outer metatarsal tubercles,
protuberant with rounded surfaces. One protuberant subarticular tubercle on toes I and II, two on toes III, IV,
and V. Hands and feet lacking supernumerary tubercles and lateral fringes. Basal webbing absent. Tarsal
tubercles absent.
In 70% alcohol the color is similar to the living animal described above, except that the red on the dorsum
changed to black and the yellow dorsolateral lines changed to silver-white.
Var ia ti on . The dorsal coloration varies over a cline from the north to south. In southern populations, the
dorsum is finely stippled brick-red to bright red and gradually changes to coarsely stippled orange or yellow
in northern populations (Figures 14 & 15).
Vocalizations. The advertisement call for A. pepperi (Fig. 4) can be classified as a ‘retarded trill’
following Lötters et al. (2003). The call consists of a series of musical notes resembling short ‘whistles’ at a
rate of 0.9–1.3 notes per second. Notes are 130–220 ms long, spaced 560–1030 ms apart, overall dominant
frequency 2970 Hz. The call of A. pepperi can be distinguished from the advertisement call of A. bassleri by

BROWN & TWOMEY18 · Zootaxa 2049 © 2009 Magnolia Press
the rate the notes are repeated, with A. bassleri repeating notes at 1.6–2.1 notes per second (vs. 0.9–1.3 in A.
pepperi).
Distribution and natural history. Ameerega pepperi is known from throughout the upper Huallaga
Valley, south of Río Huayabama (near Huicungo) to the southern border of San Martín at elevations from 380
m to approximately 1000 m elevation. Most of the known localities for this species run along the main road to
Tingo Maria, which is primarily on the eastern side of Río Huallaga (i.e., the western slope of the Cordillera
Azul). This species also occurs on the western side of the Huallaga (i.e., the eastern slope of the Andes) in San
Francisco and Huicungo, and likely occurs throughout the area between these two sites based on the results of
our niche model (Fig. 17). This model also predicts suitable habitat throughout much of the Río Biabo valley
in the Cordillera Azul, although no surveys have taken place in this area. There appears to be very little
predicted overlap in the distributions of A. pepperi and A. bassleri in the niche model (Fig. 17), with the only
overlapping areas occurring near the town of Juanjui. In 2007 we spent a day in the mountains just west of
Juanjui and did not detect either species. The area near Juanjui appears to be somewhat drier than surrounding
areas, and this may function as a weak barrier that separates the distributions of these two species.
Ameerega pepperi occurs in undisturbed premontane forests and secondary premontane forest,
particularly in habitats which are adjacent to streams (Fig. 15i). Unlike A. bassleri, which we have often
observed far from streams (> 300 m), we have never observed A. pepperi greater than ~30 m from a stream.
This species can occur in moderately disturbed areas provided the streamside habitats are relatively shaded.
During the day, adults tend to hide amongst streamside boulders. Calling activity seems to peak during late
evening, just prior to dusk. Males call from elevated positions on boulders, although courtship and oviposition
appears to take place on the ground in the leaf litter. Clutches contain 22-44 eggs and are guarded by males
(M. Pepper pers. comm.). These eggs typically hatch in approximately 18 days. Tadpoles have been found in
shallow eddies or streamside pools which are sometimes left by receding water levels. Clay-bottomed streams
appear to be most commonly used, and tadpoles will often submerge themselves under the sediment,
particularly during the hotter hours in the afternoon. These pools also contain tadpoles of the sympatric A.
altamazonica and A. trivittata. Tadpoles of Ameerega pepperi complete development in as little as 5-6 weeks.
Conservation status. Following the IUCN Red List criteria (IUCN 2001), we tentatively suggest that A.
pepperi should be listed as Least Concern (LC) based on the following: (1) we estimate the extent of
occurrence to be at least 5,500 km2, (2) populations appear to be continuous and (3) this species occurs in
largely undisturbed forests, though there has been a small amount of potential habitat loss in the Cordillera
Azul and in the upper Huallaga Valley.
Discussion
Mitochondrial Introgression in A. bassleri
The observed phylogenetic relationship of A. pepperi and four A. bassleri individuals from Saposoa/Alto
Shima (individuals 19–22 in Fig. 6) does not reflect the geographical extent of A. pepperi. Further, other A.
bassleri (individuals 9–13) from the same localities are sister to nearby populations (Sisa). Since it is unlikely
that the two distinct Saposoa/Alto Shima clades represent two species that are identical in all regards (sharing
characteristics of A. bassleri), it is likely that the observed phylogenetic relationship is not reflective of the
“true” species tree. The most likely explanation for this pattern is the historic hybridization and introgression
of the mitochondrial genome of Ameerega pepperi into the adjacent population of the common ancestor to
both A. bassleri and A. yoshina.
Incomplete Lineage Sorting in A. yoshina
Given the dramatic acoustic differences, and their unlikely convergences, species with similar calls likely
share common ancestry; however our phylogeny does not reflect this relationship. The three A. yoshina
individuals from Callanayacu are not sister to the other individuals from Contamana, nor are they sister to the

Zootaxa 2049 © 2009 Magnolia Press · 19
THREE NEW AMEEREGA FROM PERU
individuals of A. bassleri from close proximity, such as individuals from the Huallaga Canyon (separated by 8
km) and Chazuta (separated by 25 km), a pattern predicted in mitochondrial introgression (McGuire et al.
2007). The three A. yoshina individuals are more closely related to an individual from near Roque (which is
separated by 100 km and three mountain ranges). This supports the pattern of incomplete lineage sorting, as
alleles present in a common ancestor have been retained in the Callanayacu populations of A. yoshina and are
randomly distributed in the A. bassleri clade.
FIGURE 7. The biogeography of the bassleri clade is likely the result of multiple disturbance-vicariance events that
resulted in periodic connections between the Cordillera Oriental and the Serranía de Contamana, followed by longer
periods of isolation. This scenario assumes three periods of connectivity between the Cordillera Oriental and Contamana
and would provide a mechanism for allopatric speciation in this group. Each species’ distribution (and their ancestors) is
depicted by the corresponding color. The bicolored pattern depicts the sympatry of two species.
Speciation in the bassleri clade: the importance of call characteristics
The divergence of call characteristics appears to be important factor in the speciation of members of the
bassleri clade. This group is distributed through a relatively small area (throughout San Martín and nearby
Loreto), but contains an incredible amount of diversity, being composed of the three new species plus A.
bassleri and A. pongoensis. The three larger, more colorful species (Ameerega bassleri, A. pepperi, and A.
yoshina) are very similar in appearance, suggesting divergences in behavior and acoustic characteristics have
played a major role in their speciation. We suspect a major factor reinforcing species boundaries in this group
is the divergence in call characteristics, particularly in calling rate. The individual notes that compose the calls
of A. bassleri, A. pepperi, and A. yoshina are very similar in pulse duration, structure, and frequency (Table 2,
Fig. 4); though between each species, there has been tremendous divergence in the rate that notes are given
(Table 2, Fig. 4). In the case of A. bassleri and A. yoshina, the divergence in call characteristics likely quickly
reinforced species boundaries, so fast that many other genes, including the mitochondrial genes sampled for
the phylogeny, have yet to coalesce. Preliminary call playback experiments on A. bassleri males (from
Chazuta) show that males will aggressively respond to recorded calls of the same species from different
localities (Saposoa and Cainarachi Valley). However, males do not respond to the calls of A. pepperi (JLB
unpub. data). This suggests that divergence in calls is recognized, at least, by A. bassleri males.
Although we are hesitant to describe a species without reciprocally monophyletic gene trees, we are faced
with a situation where we have acoustic data which strongly suggests the existence of three species.
Advertisement calls are thought to be important intraspecific signals (Boul et al. 2007) and have frequently

BROWN & TWOMEY20 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 8. Map of central Peru. Orange circles represent localities of Ameerega petersi, green circles represent A.
smaragdina localities. Localities are based on personal observations and Silverstone (1976). Localities labeled with
numbers are those which were included in the phylogenetic analysis; localities with letters are those for which
individuals have not been sequenced. (1) Miguel Grau, Ucayali (9° 11' 11.47" S, 75° 47' 4.13" W); (2) Aguaytia, Ucayali
(9° 0' 54.00" S, 75° 29' 2.40" W); (3) Tournavista, Huánuco (8° 54' 23.50" S, 74° 44' 7.20" W); (4) Codo del Pozuzo,
Huánuco (9° 44' 5.22" S, 75° 30' 37.59" W); (5) Puerto Inca, Huánuco (9° 23' 10.39" S, 74° 54' 37.21" W); (6,7)
Cordillera El Sira, Huánuco (9° 26' 50.86" S, 74° 48' 1.91" W); (8) Panguana, Huánuco (9° 33' 51.40" S, 74° 52' 16.29"
W); (9) Pan de Azucar, Pasco (10° 14' 56.34" S, 75° 13' 32.09" W, type locality of A. smaragdina); (A) Río Iscozacin,
Pasco (10° 24' 5.66" S, 75° 3' 47.29" W); (B) Cacazú, Pasco (10° 33' 16.29" S, 75° 5' 4.79" W); (C) Santa Isabel, Pasco
(approximate location; type locality of A. petersi); (D) Nevati, Pasco (approximate location); (E) Río Vitoc, Junín
(approximate location).

Zootaxa 2049 © 2009 Magnolia Press · 21
THREE NEW AMEEREGA FROM PERU
been used as fundamental characters in anuran systematics (Platz and Forester 1988, Platz 1989, Sullivan et
al. 1996, Gamble et al. 2008, Twomey and Brown 2008a). The distinct calls of these species are ecologically
relevant characters, which we believe justify full-species status, despite contradictory phylogenetic evidence.
FIGURE 9. Speciation and biogeography of the petersi clade. The series of maps depicts the putative ancestral
distributions of the petersi clade through time, with (A) being the oldest and (E) being present-day. This hypothetical
scenario could explain the discordance between our observed mitochondrial gene tree and the suggested species tree (in
which A. petersi and A. smaragdina are the same species). Each species’ distribution (and their ancestors’ distribution) is
depicted by the corresponding color. (F) Depiction of the known localities for each of these species.
The biogeography of bassleri clade
The results from our DIVA analysis of the bassleri clade are consistent with the disturbance-vicariance
hypothesis (Colinvaux 1993, Noonan and Chippindale 2006, Noonan and Gaucher 2006, Noonan and Wray
2006), which predicts that climatic oscillations throughout the quaternary period (1.8 MYA to present)
resulted in periods of connectivity and isolation between highland refuges. In our example, the biogeography
of the bassleri clade can be explained through multiple disturbance-vicariance events that resulted in
recurring connections between the Cordillera Oriental (for simplicity, this definition includes the Cordillera
Azul) and the Serranía de Contamana, followed by longer periods of isolation. One possible scenario, which is
congruent with the ancestral range estimates from our DIVA analysis (unpublished data), is that the common
ancestor to the bassleri clade (as defined in Fig. 6) was a widespread montane species whose distribution
extended into the adjacent lowlands and into Serranía de Contamana (Fig. 7a) (Roberts et al. 2006). Due to a
vicariant event, a lowland population became isolated and diverged into A. pongoensis (Fig. 7b). During this
time, the Contamana populations of the bassleri-ignipedis-yoshina-pepperi common ancestor could have
become isolated (Fig. 7b) and eventually diverged into A. ignipedis (Fig. 7c). This divergence may have been
facilitated by selection for aposematic coloration or sexual selection for brightly-colored frogs (Roberts et al.
2007), thus explaining the more brightly-colored frogs of the bassleri-yoshina-pepperi group. Soon after, the
Upper Huallaga population of the bassleri-yoshina-pepperi group became isolated from the Cordillera
Oriental populations (Fig. 7c) and eventually diverged into A. pepperi (Fig. 7d). Note that all A. pepperi
individuals are closely related, despite occupying a relatively large area. This suggests that gene flow
continues among sampled populations or that this species recently experienced a range expansion throughout
the Upper Huallaga Valley (see Fig. 6 for phylogenetic relationships). Later, a subsequent connection between
the Cordillera Oriental and the Upper Huallaga Valley allowed A. pepperi to re-colonize the Saposoa region
(where it possibly hybridized with A. bassleri). During the same time period, Cordillera Oriental and

BROWN & TWOMEY22 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 10. Advertisement calls of species in the Ameerega petersi group. (A) A. petersi, field recording in the
Cordillera Azul 20 km NE of Tingo Maria, near the village of Miguel Grau (#1 on Fig. 8), 24.5 C, 27 April 2006.
Dominant frequency 3570 Hz. (B) A. smaragdina, field recording from Pan de Azucar, near the town of Iscozacin (#9 on
Fig. 8), 24.5 C, 9 August 2007. Dominant frequency 3594 Hz. (C) A. cainarachi, field recording from Chazuta, San
Martin, Peru, 25 C, 7 July 2007, Dominant frequency 3590 Hz.
Contamana likely became connected and the common ancestor of A. bassleri and A. yoshina re-colonized
Contamana (Fig. 7e), became isolated (Fig. 7f), and diverged into A. yoshina (Fig. 7g). Finally, a relatively
recent connection between Contamana and the Cordillera Oriental may have allowed A. yoshina to disperse

Zootaxa 2049 © 2009 Magnolia Press · 23
THREE NEW AMEEREGA FROM PERU
back into the Cordillera Oriental (Fig. 7h), although this population may be currently isolated from the
Contamana population (Fig. 7i). This scenario requires at least three connections and subsequent dispersal
events between the Cordillera Oriental and Contamana. Support for this hypothesis is provided by our
phylogenetic results and the results of the DIVA analysis.
Remarks on the taxonomy of the petersi clade
A single frog collected from Contamana in 1947 was previously thought to represent an outlying
population of A. petersi (Silverstone 1976), a species which, sensu Silverstone (1976), ranged from northern
San Martín, Peru, south to Puno, Peru, near Bolivia. Subsequent revisions (Myers and Rodriguez 1998,
Schulte 1999) described these outlying northern and southern populations as A. pongoensis and A. simulans,
respectively. As currently defined, A. petersi is distributed throughout the Pachitea drainage in central Peru
(Fig. 8), where it occurs extensively throughout lowland forests and well into the montane forests of the east
Andean versant. Our record from the Cordillera Azul represents the highest-altitude locality for A. petersi by
at least 780 m. Here, we found several individuals at 1247 m (e.g. Fig. 11 i–k), and a single male was heard
calling at 1580 m in a site where A. silverstonei was present. Silverstone (1976) also referred specimens from
the upper Río Ene, Perené, and Chanchamayo drainages in Junín to A. petersi, although further phylogenetic
studies in this area are needed to confirm whether these frogs are, in fact, A. petersi, whether they represent
outlying populations of A. smaragdina, or whether they represent an undescribed species.
Characters originally used to diagnose A. smaragdina from A. petersi now appear to be present in
populations of both species. For example, Silverstone (1976) used the lack of ventral marbling in A.
smaragdina as the character to distinguish this species from A. petersi. In 2006, we found a population of A.
petersi (diagnosis based on genetic data) from Aguaytia, Huánuco (Fig. 11 l & m), that completely lacks
ventral marbling and is morphologically indistinguishable from A. smaragdina. Analysis of advertisement
calls has not proven to be useful in diagnosing these two species. Ameerega petersi is notorious for having
highly variable advertisement calls, which may consist of notes given as couplets, triplets, quadruplets (Myers
et al. 1998), or a continuous series (Fig. 10). During our field investigations we have not noticed any
particular trends with respect to call types, and in fact, a single male of either species will often make all
different types, typically “warming up” with couplets near the beginning of a calling bout, and eventually
progressing to a continuous chain of notes which may last for 5 seconds or more. We recorded a male A.
smaragdina from Pan de Azucar which had a call consisting of note couplets (Fig. 10), although other males
were heard calling with long chains of notes.
A potential case of incomplete lineage sorting: A. petersi and A. smaragdina
There is a distinct possibility, despite the observed high level of mitochondrial divergence, that A.
smaragdina and A. petersi represent a single paraphyletic species that underwent several disturbance-
vicariance events, and the observed phylogenetic pattern is a result of mitochondrial introgression between
two divergent populations of A.“petersi-smaragdina”. First, the two species are essentially sympatric, or at
least parapatric with no barrier separating them. Second, the two species appear to be indistinguishable
ecologically, morphologically, and acoustically, so we cannot envision a scenario where they would not freely
interbreed in nature. However, because phylogenetics support the current taxonomy, and because we have no
topotypic material for A. petersi (from nearby Santa Isabel), further investigation is needed before any
taxonomic revision is made in this group.
One possible scenario to explain our observed phylogeny could be that the common ancestor of A.
cainarachi, A. smaragdina, and A. petersi was widely distributed throughout the Pachitea drainage (Fig. 9a).
A primary disturbance-vicariance event separated the southern and northern populations, giving rise to A.
smaragdina and the common ancestor A. cainarachi and A. petersi (Fig. 9b). The northern population then
likely radiated into San Martín, but soon after, a second disturbance-vicariance event separated these
populations (Fig. 9c). Due to different selective regimes, the ancestor to A. cainarachi diverged from the
ancestor of A. petersi, giving rise to the distinct morphologies that we see today. Around the same time,

BROWN & TWOMEY24 · Zootaxa 2049 © 2009 Magnolia Press
populations of A. petersi (northern populations) and A. smaragdina (southern populations) came back into
contact with each other and due to the high morphological and high acoustic variability, “species” boundaries
were weak and the northern and southern populations introgressed (Fig. 9d and 9e). This example is just one
of many hypothetical biogeographic scenarios, although this scenario is consistent with the results from our
DIVA analysis on this group (unpublished data). If A. smaragdina and A. petersi are two divergent
populations of the same species, we expect there to be an area of introgression at the margins of both
“species” where A. petersi and A. smaragdina haplotypes are abundant and further away from this area, the
haplotypes will be dominated by the historic species’ haplotype. There is some support for this hypothesis as
one individual of A. smaragdina (classified as A. petersi by its collectors) was collected near Panguana
Research Station, an area largely composed of A. petersi haplotypes (Fig. 8 locality 8), however, more distant
haplotypes are entirely A. petersi (to the north) or A. smaragdina (to the south).
Acknowledgements
We thank Brian Sexton who generously donated money which helped finance our Contamana expeditions. We
thank Bryan Richardson and Wouter Olthof for their assistance in the field and Karen Siu Ting for her help in
the museum and help dissecting the type series of A. yoshina. We especially wish to thank Mark Pepper for
providing life history information on these frogs, a photograph of Ameerega pepperi, and for assistance in the
field. We also thank Justin Yeager and Marcos Ramírez Zárate for providing photos. We are grateful to Tim
Green who alerted us to the existence of a red-backed Ameerega from Contamana in 2005. We also thank
Manuel Sanchez Rodríguez for providing a Spanish translation of the abstract. Lastly, we are grateful to Kyle
Summers for his support of our taxonomic studies. Permits to conduct field work and to export materials were
issued by INRENA (N° 002765-AG-INRENA, N° 061-2003-INRENA-IFFS-DCB, N° 050-2006-INRENA-
IFFS-DCB, and N° 067-2007-INRENA-IFFS-DCB).
Literature cited
Avise, J.C. (2000) Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, MA,
U.S.A., 447 pp.
Avise, J.C. & Ball, R.M. (1990) Principles of genealogical concordance in species concepts and biological taxonomy.
Oxford Surveys in Evolutionary Biology, 7, 45–67.
Avise, J.C., Shapira, J.F., Daniel, S.W., Aquadro, C.F. & Lansman, R.A. (1983) The use of restriction endonucleases to
measure mitochondrial-DNA sequence relatedness in natural-populations. 1. Mitochondrial-DNA differentiation
during the speciation process in Peromyscus. Molecular Biology and Evolution, 1, 38–56.
Boul, K.E., Funk, W.C., Darst, C.R., Cannatella, D.C. & Ryan, M.J. (2007) Sexual selection drives speciation in an
Amazonian frog. Proceedings of the Royal Society of London Series B, 274, 399–406.
Brown, J.L., Schulte, R. & Summers, K. (2006) A new species of Dendrobates (Anura: Dendrobatidae) from the
Amazonian lowlands in Perú. Zootaxa, 1152, 45–58.
Charif, R.A., Clark, C.W. & Fristrup, K.M. (2004) Raven 1.2 User’s Manual. Cornell Laboratory of Ornithology, Ithaca,
New York, U.S.A, 205 pp.
Chicago Field Museum (2006) Rapid biological inventory: Peru: Sierra del Divisor. The Field Museum, Chicago,
Illinois, U.S.A., 96 pp.
Colinvaux, P. (1993) Pleistocene biogeography and diversity in tropical forests of South America. In Goldblatt, P. (ed),
Biological relationships between Africa and South America. Yale University Press, New Haven, Connecticut,
U.S.A., pp. 473–499.
Doyle, J.J. (1997) Trees within trees: Genes and species, molecules and morphology. Systematic Biology, 46, 537–553.
IUCN (2001) IUCN Red List Categories and Criteria: Version 3.1. IUCN Species Survival Commission, IUCN, Gland,
Switzerland and Cambridge, UK, 14 pp.
Funk, D.J. & Omland, K.E. (2003) Species-level paraphyly and polyphyly: Frequency, causes, and consequences, with
insights from animal mitochondrial DNA. Annual Review of Ecology Evolution and Systematics, 34, 397–423.
Gamble, T., Berendzen, P.B., Shaffer, H.B., Starkey, D.E. & Simons, A.M. (2008) Species limits and phylogeography of

Zootaxa 2049 © 2009 Magnolia Press · 25
THREE NEW AMEEREGA FROM PERU
North American cricket frogs (Acris, Hylidae). Molecular Phylogenetics and Evolution, 48, 112–125.
Graham, C.H., Ron, S.R., Santos, J.C., Schneider, C.J. & Moritz, C. (2004) Integrating phylogenetics and environmental
niche models to explore speciation mechanisms in dendrobatid frogs. Evolution, 58, 178–193.
Grant, T., Frost, D.R., Caldwell, J.P., Gagliardo, R., Haddad, C.F.B., Kok, P.J.R., Means, B.D., Noonan, B.P., Schargel,
W.C. & Wheeler, W.C. (2006) Phylogenetic systematics of dart-poison frogs and their relatives (Anura:
Athesphatanura: Dendrobatidae). Bulletin of the American Museum of Natural History, 299, 1–262.
Hijmans, R.J., Cameron, S.E., Parra, J.L., Jones, P.G. & Jarvis., A. (2005) Very high resolution interpolated climate
surfaces for global land areas. International Journal of Climatology, 25, 1965–1978.
Huelsenbeck, J.P. & Ronquist, F. (2001) MRBAYES: Bayesian inference of phylogeny, Version 3.1.2. Bioinformatics
(Oxford England), 17, 754–755.
Lötters, S., Reichle, S. & Jungfer, K.-H. (2003) Advertisement calls of neotropical poison frogs (Amphibia:
Dendrobatidae) of the genera Colostethus, Dendrobates and Epipedobates, with notes on dendrobatid call
classification. Journal of Natural History, 37, 1899–1911.
Lötters, S. & Vences, M. (2000) Bemerkungen zur Nomenklatur und taxonomie peruanischer Pfeilgiftfrösche (Anura:
Dendrobatidae: Dendrobates, Epipedobates) [Remarks on the nomenclature and taxonomy of Peruvian poison frogs
(Anura: Dendrobatidae: Dendrobates, Epipedobates)]. Salamandra, 36, 247–260.
Maddison, W.P. (1997) Gene trees in species trees. Systematic Biology, 46, 523–536.
McGuire, J.A., Linkem, C.W., Koo, M.S., Hutchison, D.W., Lappin, A.K., Orange, D.I., Lemos-Espinal, J., Riddle, B.R.
& Jaeger, J.R. (2007) Mitochondrial introgression and incomplete lineage sorting through space and time:
phylogenetics of Crotaphytid lizards. Evolution, 61, 2879–2897.
Morales, V.R. (1992) Dos especies nuevas de Dendrobates (Anura: Dendrobatidae) para Perú. Caribbean Journal of
Science, 28, 191–199.
Morando, M., Avila, L.J., Baker, J. & Sites, J.W. (2004) Phylogeny and phylogeography of the Liolaemus darwinii
complex (Squamata: Liolaemidae): Evidence for introgression and incomplete lineage sorting. Evolution, 58,
842–861.
Myers, C.W. (1982) Spotted poison frogs: Descriptions of three new Dendrobates from western Amazonia, and
resurrection of a lost species from "Chiriqui". American Museum Novitates, 272, 1–23.
Myers, C.W., Rodriguez, L.O. & Icochea, J. (1998) Epipedobates simulans, a new cryptic species of poison frog from
southeastern Peru, with notes on E. macero and E. petersi (Dendrobatidae). American Museum Novitates, 3238,
1–20.
Noonan, B.P. & Chippindale, P.T. (2006) Dispersal and vicariance: The complex evolutionary history of boid snakes.
Molecular Phylogenetics and Evolution, 40, 347–358.
Noonan, B.P. & Gaucher, P. (2006) Refugial isolation and secondary contact in the dyeing poison frog Dendrobates
tinctorius. Molecular Ecology, 15, 4425–4435.
Noonan, B.P. & Wray, K.P. (2006) Neotropical diversification: the effects of a complex history on diversity within the
poison frog genus Dendrobates. Journal of Biogeography, 33, 1007–1020.
Nylander, J.A.A. (2004) MrModeltest 2.2. Program distributed by author. Evolutionary Biology Centre, Uppsala
University.
Pamilo, P. & Nei, M. (1988) Relationships between gene trees and species trees. Molecular Biology and Evolution, 5,
568–583.
Patton, J.L. & Smith, M.F. (1994) Paraphyly, polyphyly, and the nature of species boundaries in pocket gophers (Genus
Thomomys). Systematic Biology, 43, 11–26.
Phillips, S.J., Anderson, R.P. & Schapire, R.E. (2006) Maximum entropy modeling of species geographic distributions.
Ecological Modelling, 190, 231–259.
Platz, J.E. (1989) Speciation within the chorus frog Pseudacris triseriata: morphometric and mating call analyses of the
boreal and western subspecies. Copeia, 1989, 704–712.
Platz, J.E. & Forester, D.C. (1988) Geographic variation in mating call among the four subspecies of the chorus frog:
Pseudacris triseriata (Wied). Copeia, 1988, 1062–1066.
Rodríguez, L. & Myers, C.W. (1993) A new poison frog from Manu National Park, southeastern Peru (Dendrobatidae,
Epipedobates). American Museum Novitates, 3068, 1–15.
Roberts, J.L., Brown, J.L., Arizabal, W., von May, R. Schulte, R. & Summers, K. (2006) Genetic divergence and
speciation in lowland and montane Peruvian poison frogs. Molecular Phylogenetics and Evolution, 41, 149–164.
Roberts, J.L., Brown, J.L., Schulte, R., Arizabal, W. & Summers, K. (2007) Rapid diversification of colouration among
populations of a poison frog isolated on sky-peninsulas in the central cordilleras of Peru. Journal of Biogeography,
34, 417–426.
Ronquist, F. (1996) DIVA version 1.1. Computer program and manual available by anonymous FTP from Uppsala
University (ftp.uu.se or ftp.systbot.uu.se). Downloaded on September 23, 2008.
Santos, J., Coloma, L. & Cannatella, D. (2003) Multiple, recurring origins of aposematism and diet specialization in
poison frogs. Proceedings of the National Academy of Sciences of the United States of America, 100, 12792–12797.

BROWN & TWOMEY26 · Zootaxa 2049 © 2009 Magnolia Press
Silverstone, P.A. (1976) A revision of the poison-arrow frogs of the genus Phyllobates Bibron in Sagra (Family
Dendrobatidae). Bulletin of the Natural History Museum of Los Angeles County, 27, 1–53.
Schulte, R. (1986) Ein neue Dendrobates- art aus Ostperu (Amphibia: Salientia: Dendrobatidae). Sauria, 8, 11–20.
Schulte, R. (1999) Pfeilgiftfrösche "Artenteil - Peru". Wailblingen, Germany, INBICO, 292 pp.
Slade, R.W. & Moritz, C. (1998) Phylogeography of Bufo marinus from its natural and introduced ranges. Proceedings of
the Royal Society of London Series B-Biological Sciences, 265, 769–777.
SPSS 15.0 (2008) SPSS for Windows, Rel. 15. 2008. Chicago: SPSS Inc.
Stockman, A.K. & Bond, J.E. (2007) Delimiting cohesion species: extreme population structuring and the role of
ecological interchangeability. Molecular Ecology, 16, 3374–3392.
Sullivan, B.K., Malmos, K.B. & Given, M.F. (1996) Systematics of the Bufo woodhousii complex (Anura: Bufonidae):
advertisement call variation. Copeia, 1996, 274–280.
Summers, K. & Clough, M. (2001) The evolution of coloration and toxicity in the poison frog family (Dendrobatidae).
Proceedings of the National Academy of Sciences of the United States of America, 98, 6227–6232.
Talbot, S.L. & Shields, G.F. (1996) Phylogeography of brown bears (Ursus arctos) of Alaska and paraphyly within the
Ursidae. Molecular Phylogenetics and Evolution, 5, 477–494.
Twomey, E. & Brown, J.L. (2008a) A partial revision of the Ameerega hahneli complex (Anura: Dendrobatidae) and a
new cryptic species from the east-Andean versant of central Peru. Zootaxa, 1757, 49–65.
Twomey, E. & Brown, J.L. (2008b) Spotted poison frogs: rediscovery of a lost species and a new genus (Anura:
Dendrobatidae) from northwestern Peru. Herpetologica, 64, 121–137.
Vences, M., Kosuch, J., Boistel, R., Haddad, C.F.B., La Marca, E., Lotters, S. & Veith, M. (2003) Convergent evolution
of aposematic coloration in Neotropical poison frogs: a molecular phylogenetic perspective. Organisms Diversity &
Evolution, 3, 215–226.
Weisrock, D.W., Kozak, K.H. & Larson, A. (2005) Phylogeographic analysis of mitochondrial gene flow and
introgression in the salamander, Plethodon shermani. Molecular Ecology, 14, 1457–1472.

Zootaxa 2049 © 2009 Magnolia Press · 27
THREE NEW AMEEREGA FROM PERU
Appendix figures
FIGURE 11. Ameerega ignipedis and similar species. (a,b) A. pongoensis from lower Cainarachi Valley, San Martin,
Peru; (c,d) A. pongoensis from Shucushuyacu, Loreto, Peru; (e,f) A. ignipedis from type locality (holotype and MUSM
24952, respectively); (g) Ameerega simulans from Mazuko, Madre de Díos, Peru (photo by K. Summers); (h) A. petersi
from Codo del Pozuzo, Huánuco, Peru; (i, k, j) A. petersi from Miguel Grau, Ucayali, Peru (photo k by J. Yeager); (l,m)
A. petersi from Aguaytia, Ucayali, Peru; (n) A. petersi from Puerto Inca, Huánuco, Peru.

BROWN & TWOMEY28 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 12. Ameerega petersi, A. smaragdina, and A. bassleri. (a) A. petersi from Puerto Inca, Huánuco, Peru; (b) A.
petersi from Tournavista, Huánuco, Peru; (c,d) A. smaragdina from Pan de Azucar, Pasco, Peru (photo ‘d’ by K.
Summers); (e-h) A. bassleri from 8.6 km E of San José de Sisa, San Martín, Peru; (i-o) A. bassleri from 7 km NW of
Saposoa, San Martín, Peru.

Zootaxa 2049 © 2009 Magnolia Press · 29
THREE NEW AMEEREGA FROM PERU
FIGURE 13. Ameerega bassleri. (a,b) Roque, San Martín, Peru (type locality); (c-e) Cainarachi Valley, San Martín,
Peru; (f-j) Chazuta, San Martín, Peru; (k) Sauce, San Martín, Peru; (l-m) Cordillera Azul, 14 km SE of Shamboyacu, San
Martín, Peru (photo by M. Ramírez Zárate); (n-p) eggs and tadpoles of A. bassleri.

BROWN & TWOMEY30 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 14. Ameerega yoshina, A. cainarachi, and A. pepperi. (a-c) A. yoshina from Callanayacu, San Martín, Peru
(MUSM 26985-26987); (d-f) A. yoshina from type locality (MUSM 24945, 26953, 26954), (g) A. yoshina (left) and
sympatric A. cainarachi (right) from Callanayacu, San Martín, Peru; (h) A. cainarachi from Cainarachi Valley, San
Martín, Peru; (i) A. cainarachi from Chazuta, San Martín, Peru; (j,k) A. pepperi from type locality, (l) A. pepperi from
Nuevo San Martín, San Martín, Peru.

Zootaxa 2049 © 2009 Magnolia Press · 31
THREE NEW AMEEREGA FROM PERU
FIGURE 15. Ameerega pepperi and habitat photos for the new species in this paper. (a,e) A. pepperi from Chumanza,
San Martín, Peru (photo ‘a’ by M. Pepper); (b-d) A. pepperi from Río Abiseo near Huicungo, San Martín, Peru; (f,g) A.
pepperi from 6.3 km N of Campanilla, San Martín, Peru; (h) type locality of A. ignipedis, a geothermal stream flowing
out of the Serranía de Contamana. The entire type series (as well as one A. yoshina metamorph) was found in the forest
adjacent to the stream or among the plants growing in the stream banks. (i) type locality of A. pepperi near San
Francisco, San Martín, Peru. (j) Quebrada Pacuyacu near Callanayacu. Three A. yoshina paratypes (MUSM 26985-
26987) were collected from amongst the rocks and vegetation alongside the stream.

BROWN & TWOMEY32 · Zootaxa 2049 © 2009 Magnolia Press
FIGURE 16. Type series of new species in this paper. (a) Ameerega ignipedis (L-R, MUSM 24947–24952); (b) A.
yoshina (L-R, MUSM 24945, 26953–26955, 26985–26987); (c) A. pepperi (L-R, MUSM 26940-26942, 26968, 26975,
26976, 26979, 26980).

Zootaxa 2049 © 2009 Magnolia Press · 33
THREE NEW AMEEREGA FROM PERU
FIGURE 17. Ecological niche models of Ameerega bassleri (red) and A. pepperi (blue). (A) Only 13 ”presence” pixels
overlapped between both species’ models (depicted in green). This is 0.17% (pixel count = 7790) of the “presence” area
in the model of A. bassleri and 0.28% (pixel count = 4622) of the “presence” area in the model of A. pepperi. These
results support the hypothesis that these two species have diverged ecologically and currently occupy separate niches.
(B) The model for A. pepperi also predicted this species’ occurrence at the type localities of A. yoshina and A. ignipedis
(near Contamana). This suggests that both A. yoshina and A. pepperi share similar niches, however due to limited locality
data on A. yoshina, we were unable to calculate this species’ distribution and therefore cannot make comparisons. Black
bar = 50 km; 1 pixel in niche model = 1 km2.

BROWN & TWOMEY34 · Zootaxa 2049 © 2009 Magnolia Press
Appendix tables
TABLE 3. Measurements (in mm) of Ameerega ignipedis type series. Averages (with standard deviation) were
calculated from the type series.
TABLE 4. Measurements (in mm) of Ameerega yoshina type series. Averages (with standard deviation) were calculated
from the type series, excluding MUSM 26955, which was a sub-adult male. An asterisk indicates individuals were
dissected to confirm sex.
Character MUSM MUSM MUSM MUSM MUSM MUSM Average
24947 24948 24949 24950 24951 24952 (N = 6)
SVL 20.3 23.8 20.8 21.7 24.2 23.5 22.4 ± 1.6
FL 8.2 10.5 9.3 10.6 8.9 10.5 9.7 ± 1.0
TL 9.6 11.1 9.6 11.0 11.1 11.0 10.6 ± 0.8
KK 17.2 20.9 18.4 20.1 18.6 20.7 19.3 ± 1.5
FoL 8.2 9.9 7.8 9.5 9.0 9.8 9.0 ± 0.9
HaL 4.8 6.0 5.2 5.8 5.9 6.2 5.7 ± 0.5
HL 4.8 6.0 5.6 5.4 5.5 5.3 5.4 ± 0.4
HW 6.1 7.1 6.1 6.6 6.1 6.8 6.5 ± 0.4
BW 5.9 7.0 6.0 5.5 6.1 6.2 6.1 ± 0.5
IND 2.1 2.8 2.3 2.6 2.2 2.3 2.4 ± 0.3
UEW 4.3 4.6 5.3 4.8 4.3 5.3 4.8 ± 0.5
IOD 3.0 3.4 4.0 4.2 4.0 4.6 3.9 ± 0.6
TD 2.1 2.2 2.0 1.9 1.8 2.2 2.0 ± 0.2
ED 3.6 4.1 4.3 3.8 4.0 4.2 4.0 ± 0.3
DET 0.9 1.4 1.2 1.0 1.3 1.1 1.1 ± 0.2
L1F 4.5 4.7 4.6 4.6 3.8 5.4 4.6 ± 0.5
L2F 2.1 2.9 2.5 3.0 2.2 2.6 2.6 ± 0.4
W3D 0.9 0.9 0.8 0.9 0.9 0.9 0.9 ± 0.1
W3F 0.5 0.6 0.6 0.6 0.5 0.5 0.5 ± 0.1
SEX M F F F M F
Character MUSM MUSM MUSM MUSM MUSM MUSM MUSM Average
24945 26953 26954 26955 26985 26986 26987 (N = 6)
SVL 34.8 29.2 33.1 20.8 31.5 33.9 26.2 31.4 ± 3.2
FL 16.3 13.4 15.2 11.5 14.0 14.6 14.4 14.7 ± 1.0
TL 17.6 15.9 17.0 11.6 16.5 17.5 17.8 17.1 ± 0.7
KK 29.5 28.7 30.6 11.3 29.7 29.6 31.0 28.3 ± 3.9
FoL 14.6 13.7 14.4 10.5 15.8 13.9 15.0 14.6 ± 0.8
HaL 8.3 8.5 8.2 6.1 9.0 8.8 8.8 8.6 ± 0.3
HL 8.1 9.2 9.7 7.3 8.5 9.1 9.8 9.1 ± 0.7
HW 10.2 8.8 9.8 6.7 9.3 10.3 10.7 9.9 ± 0.7
BW 12.1 9.5 11.1 5.7 8.7 11.8 12.8 11.0 ± 1.6
UEW 7.3 6.6 6.7 5.1 6.3 6.1 7.1 6.7 ± 0.5
IOD 4.9 3.9 4.7 3.2 4.0 4.8 5.6 4.6 ± 0.6
TD 3.2 1.9 1.9 1.8 1.9 1.9 2.6 2.2 ± 0.5

Zootaxa 2049 © 2009 Magnolia Press · 35
THREE NEW AMEEREGA FROM PERU
TABLE 5. Measurements (in mm) of Ameerega pepperi type series. Averages (with standard deviation) were calculated
from the type series.
TABLE 6. List of localities and GenBank accession numbers for individuals included in the phylogeny.
ED 5.3 5.8 5.8 4.5 5.8 4.5 5.1 5.4 ± 0.5
DET 0.8 1.4 1.6 1.1 1.6 1.4 1.6 1.4 ± 0.3
L1F 7.6 6.7 7.4 5.6 8.5 8.2 7.2 7.6 ± 0.6
L2F 4.5 6.7 7.2 5.0 8.2 7.2 5.6 6.6 ± 1.3
W3D 1.0 1.0 1.0 0.6 1.0 1.0 1.0 1.0 ± 0.1
W3F 0.5 0.5 0.8 0.5 0.5 0.5 0.6 0.6 ± 0.1
SEX FF* F* M* FF* F
Character MUSM MUSM MUSM MUSM MUSM MUSM MUSM MUSM Average
26976 26975 26980 26979 26940 26941 26942 26968 (N = 7)
SVL 32.5 32.8 31.5 30.8 28.6 28.6 34.4 34.1 31.7 ± 2.2
FL 14.8 15.7 16.1 16.0 13.9 13.4 15.6 15.8 15.2 ± 1.0
TL 15.7 16.6 16.5 16.3 15.0 16.1 16.8 16.8 16.2 ± 0.6
KK 30.1 30.5 30.4 29.7 27.7 28.2 31.4 30.8 29.9 ± 1.3
FoL 13.9 15.5 14.9 14.4 13.7 13.3 15.1 16.0 14.6 ± 0.9
HaL 8.9 8.6 8.6 8.2 7.8 7.5 8.0 9.3 8.4 ± 0.6
HL 9.8 9.8 9.3 8.5 8.2 7.7 9.1 8.6 8.9 ± 0.8
HW 9.2 9.7 9.5 9.1 8.6 7.8 9.1 9.7 9.1 ± 0.6
BW 8.6 9.8 8.7 8.9 8.9 8.3 11.4 9.9 9.3 ± 1.0
UEW 5.6 6.1 5.6 5.1 5.5 5.5 5.1 6.1 5.6 ± 0.4
IOD 3.9 3.7 4.0 3.9 4.2 3.9 4.2 3.5 3.9 ± 0.2
IND 4.7 5.0 4.0 4.2 4.7 4.0 4.8 4.8 4.5 ± 0.4
TD 2.7 1.8 2.6 2.1 1.8 3.0 1.6 2.1 2.2 ± 0.5
ED 4.2 4.0 4.5 4.2 3.7 4.2 4.5 4.8 4.3 ± 0.3
DET 1.6 1.8 1.6 1.6 1.5 1.6 1.6 1.6 1.6 ± 0.1
L1F 7.2 7.6 7.1 7.2 7.4 6.9 8.4 7.4 7.4 ± 0.4
L2F 7.2 7.1 6.9 6.4 6.9 6.1 7.2 7.4 6.9 ± 0.4
W3D 0.8 1.0 0.8 1.0 1.0 0.8 1.1 1.1 0.9 ± 0.1
W3F 0.5 0.5 0.3 0.5 0.6 0.6 0.6 0.6 0.5 ± 0.1
SEXFFFF MMFF
Species Locality 12s 16s CytB Reference
altamazonica 1 Chazuta, San Martín, PE DQ522980 DQ523051 DQ523121 Roberts et al. 2006
altamazonica 2 Tarapoto, San Martin, PE DQ522955 DQ523026 DQ523096 Roberts et al. 2006
altamazonica 3 San Jose de Sisa, San Martin, PE DQ523008 DQ523079 DQ523149 Roberts et al. 2006
altamazonica 4 Tarapoto, San Martin, PE DQ523007 DQ523078 DQ523148 Roberts et al. 2006
altamazonica 5 Chazuta, San Martin, PE DQ522966 DQ523037 DQ523107 Roberts et al. 2006
altamazonica 6 Tocache, San Martin, PE EU517662 EU517665 EU517671 Twomey & Brown 2008a
altamazonica 7 Tocache, San Martin, PE EU517661 EU517663 N/A Twomey & Brown 2008a
altamazonica 8 Pampa Hermosa, Loreto, PE EU517660 EU517664 EU517672 Twomey & Brown 2008a
altamazonica 9 Pampa Hermosa, Loreto, PE EU517659 EU517666 EU517677 Twomey & Brown 2008a
altamazonica 10 Saposoa, San Martin, PE DQ523015 DQ523086 DQ523156 Roberts et al. 2006

BROWN & TWOMEY36 · Zootaxa 2049 © 2009 Magnolia Press
altamazonica 11 Tarapoto, San Martin, PE DQ522951 DQ523022 DQ523092 Roberts et al. 2006
altamazonica 12 Tarapoto, San Martin, PE DQ522978 DQ523049 DQ523119 Roberts et al. 2006
bassleri 1 Chumia, San Martín, PE FJ752366 FJ752289 FJ752333 this study
bassleri 2 Chumia, San Martín, PE FJ752367 FJ752290 FJ752334 this study
bassleri 3 Sauce, San Martín, PE DQ522989 DQ523060 DQ523130 Roberts et al. 2006
bassleri 4 Cainarachi valley, San Martín, PE DQ522974 DQ523045 DQ523115 Roberts et al. 2006
bassleri 5 Cainarachi valley, San Martín, PE DQ523009 DQ523080 DQ523150 Roberts et al. 2006
bassleri 6 Huallaga Canyon, San Martín, PE DQ522986 DQ523057 DQ523127 Roberts et al. 2006
bassleri 7 Chazuta, San Martín, PE DQ523016 DQ523087 DQ523157 Roberts et al. 2006
bassleri 8 Roque, San Martín, PE FJ752352 FJ752273 FJ752318 this study
bassleri 9 Saposoa, San Martín, PE N/A FJ752266 FJ752310 this study
bassleri 10 Saposoa, San Martín, PE N/A FJ752267

FJ752311 this study
bassleri 11 Saposoa, San Martín, PE N/A FJ752268 FJ752312 this study
bassleri 12 Alto Shima, San Martín, PE N/A FJ752269 FJ752313 this study
bassleri 13 Alto Shima, San Martín, PE N/A FJ752270 FJ752314 this study
bassleri 14 San Jose de Sisa, San Martín, PE DQ522999 DQ523070 DQ523140 Roberts et al. 2006
bassleri 15 San Jose de Sisa, San Martín, PE DQ523002 DQ523073 DQ523143 Roberts et al. 2006
bassleri 16 San Jose de Sisa, San Martín, PE DQ523014 DQ523085 DQ523155 Roberts et al. 2006
bassleri 17 Shanao, San Martín, PE DQ522972 DQ523043 DQ523113 Roberts et al. 2006
bassleri 18 Roque, San Martín, PE FJ752351 FJ752272 FJ752317 this study
bassleri 19 Alto Shima, San Martín, PE DQ523012 DQ523083 DQ523153 Roberts et al. 2006
bassleri 20 Alto Shima, San Martín, PE N/A FJ752271 FJ752316 this study
bassleri 21 Saposoa, San Martín, PE DQ523017 DQ523088 DQ523158 Roberts et al. 2006
bassleri 22 Saposoa, San Martín, PE N/A NA FJ752315 this study
bilinguis 1 Primavera, Napo, EC DQ523003 DQ523074 DQ523144 Roberts et al. 2006
bilinguis 2 Cuyabeno, Sucumbios, EC DQ502095 DQ502095 DQ502527 Grant et al. 2006
bilinguis 3 Cuyabeno, Sucumbios, EC DQ502073 DQ502073 DQ502504 Grant et al. 2006
braccata Manso, Matto Grosso, BR DQ502125 DQ502125 N/A Grant et al. 2006
cainarachi 1 Cainarachi valley, San Martín, PE DQ522953 DQ523024 DQ523094 Roberts et al. 2006
cainarachi 2 Cainarachi valley, San Martín, PE DQ522982 DQ523053 DQ523123 Roberts et al. 2006
flavopicta Rio Tocantins, Paraná, BR DQ502124 DQ502124 N/A Grant et al. 2006
hahneli 1 Convento, San Martín, PE DQ522961 DQ523032 DQ523102 Roberts et al. 2006
hahneli 2 Puerto Inca, Huanuco, PE N/A EU517669 EU517676 Twomey & Brown 2008a
hahneli 3 Panguana, Huanuco, PE N/A AF282248 N/A Lötters & Vences 2000
hahneli 4 Puente Itaya, Loreto, PE DQ522990 DQ523061 DQ523131 Roberts et al. 2006
hahneli 5 Leticia, Amazonas, CO DQ502270 DQ502270 DQ502701 Grant et al. 2006
hahneli 6 Yasuni, Orellana, EC AY364573 AY364573 N/A Santos et al. 2003
hahneli 7 Rio Manati, Loreto, PE DQ523004 DQ523075 DQ523145 Roberts et al. 2006
hahneli 8 Alto Purus, Ucayali, PE DQ522970 DQ523041 DQ523111 Roberts et al. 2006
hahneli 9 Boca Manu, Cusco, PE DQ522956 DQ523027 DQ523097 Roberts et al. 2006
hahneli 10 Rio los Amigos, Madre de Dios, PE DQ522985 DQ523056 DQ523126 Roberts et al. 2006
hahneli 11 Cobija, Pando, BO N/A AF282246 N/A Lötters & Vences 2000
hahneli 12 Madre de Dios, PE DQ501997 DQ501997 DQ502422 Grant et al. 2006
hahneli 13 Amazonas, BR DQ522992 DQ523063 DQ523133 Roberts et al. 2006
hahneli 14 Amazonas, BR DQ522996 DQ523067 DQ523137 Roberts et al. 2006
hahneli 15 near Manaus, Amazonas, BR DQ502226 DQ502226 DQ502659 Grant et al. 2006
hahneli 16 near Manaus, Amazonas, BR DQ502226 DQ502226 DQ502659 Grant et al. 2006

Zootaxa 2049 © 2009 Magnolia Press · 37
THREE NEW AMEEREGA FROM PERU
hahneli 17 Reserve Trinite, Saint-Élie, FG N/A AY263247 N/A Vences et al. 2003
hahneli 18 Porto Walter, Acre, BR DQ502084 DQ502084 DQ502515 Grant et al. 2006
hahneli 19 Porto Walter, Acre, BR DQ502085 DQ502085 DQ502516 Grant et al. 2006
ignipedis 1 Contamana, Loreto, PE FJ752368 FJ752291 N/A this study
ignipedis 2 Contamana, Loreto, PE FJ752369 FJ752292 N/A this study
ignipedis 3 Contamana, Loreto, PE FJ752370 FJ752293 N/A this study
ignipedis 4 Contamana, Loreto, PE N/A FJ752294 FJ752335 this study
ignipedis 5 Contamana, Loreto, PE FJ752371 FJ752295 FJ752336 this study
macero 1 Ivochote, Cusco, PE DQ522968 DQ523039 DQ523109 Roberts et al. 2006
macero 2 near La Merced, Junin, PE N/A EU525852 EU525853 Twomey & Brown 2008a
macero 3 Manu, Madre de Dios, PE DQ502155 DQ502155 DQ502591 Grant et al. 2006
macero 4 Alto Purus, Ucayali, PE DQ523018 DQ523089 N/A Roberts et al. 2006
parvula Macas, Morona-Santiago, EC N/A AY364574 N/A Santos et al. 2003
pepperi 1 Abiseo, San Martin, PE FJ752357 FJ752280 FJ752324 this study
pepperi 2 Chumanza, San Martin, PE FJ752358 FJ752281 FJ752325 this study
pepperi 3 Nuevo San Martin, San Martin, PE FJ752359 FJ752282 FJ752326 this study
pepperi 4 San Francisco, San Martin, PE FJ752360 FJ752283 FJ752327 this study
pepperi 5 Abiseo, San Martin, PE FJ752361 FJ752284 FJ752328 this study
pepperi 6 Nuevo San Martin, San Martin, PE FJ752362 FJ752285 FJ752329 this study
pepperi 7 Chumanza, San Martin, PE FJ752363 FJ752286 FJ752330 this study
pepperi 8 San Francisco, San Martin, PE FJ752364 FJ752287 FJ752331 this study
pepperi 9 Campanilla, San Martin, PE FJ752365 FJ752288 FJ752332 this study
petersi 1 Aguaytia, Ucayali, PE N/A FJ752300 FJ752341 this study
petersi 2 Aguaytia, Ucayali, PE FJ752372 FJ752301 FJ752342 this study
petersi 3 Divisoria, Ucayali, PE FJ752373 FJ752302 FJ752343 this study
petersi 4 Divisoria, Ucayali, PE FJ752374 FJ752303 FJ752344 this study
petersi 5 Panguana, Huánuco, PE DQ502116 DQ502116 N/A Grant et al. 2006
petersi 6 Tournavista, Huánuco, PE FJ752375 FJ752304 FJ752345 this study
petersi 7 Cordillera El Sira, Huánuco, PE FJ752378 FJ752305 FJ752346 this study
petersi 8 Codo del Pozuzo, Huánuco, PE FJ752379 FJ752306 FJ752347 this study
petersi 9 Cordillera El Sira, Huánuco, PE FJ752380 FJ752307 FJ752348 this study
petersi 10 Panguana, Huánuco, PE DQ502114 DQ502114 N/A Grant et al. 2006
picta Kartabo Pt., Mazaruni-Potaro, GY DQ502252 DQ502252 N/A Grant et al. 2006
pongoensis 1 Shucushuyacu, Loreto, PE FJ752381 FJ752308 FJ752349 this study
pongoensis 2 Shucushuyacu, Loreto, PE FJ752382 FJ752309 FJ752350 this study
pongoensis 3 Huallaga Canyon, San Martín, PE DQ523005 DQ523076 DQ523146 Roberts et al. 2006
pongoensis 4 Convento, San Martín, PE DQ522973 DQ523044 DQ523114 Roberts et al. 2006
pulchripectus Serra do Navio, Amapá, BR DQ502033 DQ502033 N/A Grant et al. 2006
rubriventris 1 near Aguaytia, Ucayali, PE N/A EU517668 EU517674 Twomey & Brown 2008a
rubriventris 2 near Aguaytia, Ucayali, PE N/A EU517667 EU517673 Twomey & Brown 2008a
rubriventris 3 near Aguaytia, Ucayali, PE N/A AF282247 N/A Lötters & Vences 2000
silverstonei 1 Tingo Maria, Huánuco, PE DQ523013 DQ523084 DQ523154 Roberts et al. 2006
silverstonei 2 Captive bred, no data N/A N/A DQ502582 Grant et al. 2006
simulans 1 Quincemil, Cusco, PE DQ523020 DQ523090 DQ523160 Roberts et al. 2006
simulans 2 Mazuko, Madre de Dios, PE DQ523019 N/A DQ523159 Roberts et al. 2006
smaragdina 1 Pan de Azucar, Pasco, PE N/A FJ752296 FJ752337 this study
smaragdina 2 Pan de Azucar, Pasco, PE N/A FJ752297 FJ752338 this study

BROWN & TWOMEY38 · Zootaxa 2049 © 2009 Magnolia Press
smaragdina 3 Pan de Azucar, Pasco, PE FJ752376 FJ752298 FJ752339 this study
smaragdina 4 Iscozacin, Pasco, PE DQ522971 DQ523042 DQ533112 Roberts et al. 2006
smaragdina 5 Pan de Azucar, Pasco, PE FJ752377 FJ752299 FJ752340 this study
trivittata 1 Tarapoto, San Martín, PE DQ522950 DQ523021 DQ523091 Roberts et al. 2006
trivittata 2 Alto Purus, Ucayali, PE DQ522957 DQ523028 DQ523098 Roberts et al. 2006
trivittata 3 French Guiana DQ523006 DQ523077 DQ523147 Roberts et al. 2006
yoshina 1 Callanayacu, San Martín, PE FJ752353 FJ752274 FJ752319 this study
yoshina 2 Callanayacu, San Martín, PE FJ752354 FJ752275 FJ752320 this study
yoshina 3 Callanayacu, San Martín, PE N/A FJ752276 FJ752321 this study
yoshina 4 Contamana, Loreto, PE FJ752355 FJ752277 FJ752322 this study
yoshina 5 Contamana, Loreto, PE FJ752356 FJ752278 FJ752323 this study
yoshina 6 Contamana, Loreto, PE N/A FJ752279 N/A this study
yungicola Carnavi, La Paz, BO N/A AY263239 N/A Vences et al. 2003
Allobates femoralis 1 Boca Manu , Cusco, PE DQ523069 DQ523139 DQ522998 Roberts et al. 2006
A. femoralis 2 Rio Sucusari, Loreto, PE DQ522952 DQ523023 DQ523093 Roberts et al. 2006
A. marchesianus Bonilla, San Martín, PE DQ522977 DQ523048 DQ523118 Roberts et al. 2006
Colostethus inguinalis San Roque, Caldas, CO DQ502265 DQ502265 DQ502696 Grant et al. 2006
C. cf. pratti Cana, Darien, PA DQ502173 DQ502173 N/A Grant et al. 2006
C. fugax Morona-Santiago, EC AY364547 AY364547 N/A Santos et al. 2003
C. panamensis Cana, Darien, PA DQ502172 DQ502172 DQ502608 Grant et al. 2006
Epipedobates anthonyi Ecuador DQ502151 DQ502151 DQ502584 Grant et al. 2006
E. boulengeri 1 Ecuador N/A N/A DQ502447 Grant et al. 2006
E. boulengeri 2 Ecuador AF128555 AF128554 AF128556 Summers & Clough 2001
E. espinosai Santo Domingo, Pichincha, EC DQ502158 N/A DQ502594 Grant et al. 2006
E. sp. QCAZ16589 Mindo, Pichincha, EC AY364575 N/A N/A Santos et al. 2003
E. tricolor Moraspungo, Bolivar, EC AY395961 N/A N/A Graham et al. 2004
Silverstoneia flotator El Cope, Cocle, PA DQ502164 DQ502164 DQ502599 Grant et al. 2006
S. nubicola El Cope, Cocle, PA DQ502165 DQ502165 DQ502600 Grant et al. 2006