Systematics of treefrogs of the Hypsiboas calcaratus and Hypsiboas fasciatus species complex (Anura, Hylidae) with the description of four new species

Abstract We review the systematics of the Hypsiboas calcaratus species complex, a group of widely distributed Amazonian hylid frogs. A comprehensive analysis of genetic, morphological, and bioacoustic datasets uncovered the existence of eleven candidate species, six of which are confirmed. Two of them correspond to Hypsiboas fasciatus and Hypsiboas calcaratus and the remaining four are new species that we describe here. Hypsiboas fasciatus sensu stricto has a geographic range restricted to the eastern Andean foothills of southern Ecuador while Hypsiboas calcaratus sensu stricto has a wide distribution in the Amazon basin. Hypsiboas almendarizae sp. n. occurs at elevations between 500 and 1950 m in central and northern Ecuador; the other new species (H. maculateralis sp. n., H. alfaroi sp. n., and H. tetete sp. n.) occur at elevations below 500 m in Amazonian Ecuador and Peru. The new species differ from H. calcaratus and H. fasciatus in morphology, advertisement calls, and mitochondrial and nuclear DNA sequences. Five candidate species from the Guianan region, Peru, and Bolivia are left as unconfirmed. Examination of the type material of Hyla steinbachi, from Bolivia, shows that it is not conspecific with H. fasciatus and thus is removed from its synonymy.


introduction
Management and conservation planning of biodiversity requires proper species identifications and comprehensive inventories. Recent DNA-based studies of Amazonian amphibians have shown the existence of a large proportion of undescribed species that have escaped detection in morphology-based assessments (Elmer et al. 2013;Elmer and Cannatella 2008;Fouquet et al. 2012;Fouquet et al. 2007;Funk et al. 2012;Jungfer et al. 2013;Ron et al. 2012). The discovery of these taxonomic voids highlights the need of renewed and intensive efforts to discover and catalogue amphibians in the Amazon region. This is particularly urgent because recent evaluations of the potential impact of climate change have shown that Amazonian amphibians are highly vulnerable (Foden et al. 2013). Conservation measures could be ineffective without reliable baseline data on species identity and distribution.
Among the species complexes that have been proven to contain a large proportion of hidden species richness is the Hypsiboas calcaratus-fasciatus species complex (Funk et al. 2012). Both species are widely distributed in the Amazon basin (Azevedo-Ramos et al. 2010;Icochea et al. 2004) and are characterized by having brown dorsal coloration, basal hand webbing, and presence of dark marks or bars on the flanks and concealed surfaces of the thighs (Duellman 1973). According to most current accounts Hypsiboas calcaratus is a species with large calcars and vertical bars on the flanks and thighs while H. fasciatus has small calcars and irregular black spots on the flanks and thighs (e.g., De la Riva et al. 2000;Duellman 2005;Rodríguez and Duellman 1994). Presumably, these accounts are based on the taxonomic review by Duellman (1973), who included them in the Hyla geographica group. In their phylogeny of Hylidae, Faivovich et al. (2005) demonstrated that the "Hyla geographica" group was paraphyletic. Hypsiboas calcaratus and H. fasciatus were assigned to the Hypsiboas albopunctatus group, which was diagnosed by genetic characters. The H. albopunctatus group, according to the definition of Faivovich et al. (2005), contains nine formally described species: H. albopunctatus (Spix, 1824), H. calcaratus (Troschel, 1848), H. dentei (Bokermann, 1967), H. fasciatus (Günther, 1858), H. heilprini (Noble, 1923), H. lanciformis (Cope, 1871), H. leucocheilus (Caramaschi & Niemeyer, 2003), H. multifasciatus (Günther, 1859) and H. raniceps (Cope, 1862). Wiens et al. (2010) found strong support for the H. albopunctatus group with H. calcaratus and H. fasciatus being sister to each other. Pyron and Wiens (2011) Recently published genetic, morphological and bioacoustic evidence suggests that H. fasciatus and H. calcaratus are a species complex. A phylogenetic analysis based on mitochondrial DNA (mtDNA) sequences revealed the presence of unconfirmed candidate species (Fouquet et al. 2007a). More recently, Funk et al. (2012) analyzed variation in mitochondrial and nuclear DNA, acoustic and morphological characters of populations of both species and found the existence of nine candidate species. Here in we incorporate additional genetic and morphological data and identify 11 candidate species of which six are confirmed candidate species. We describe four of them including their advertisement calls and variation in external morphology.
Adult specimens were measured for the following variables: (1) snout-vent length; (2) head length; (3) head width; (4) femur length; (5) tibia length; (6) foot length; (7) tympanum diameter; and (8) calcar length. Measurements were made according to the methodology described in Duellman (1970) with digital calipers (nearest 0.01 mm) from specimens fixed in 10% formalin and preserved in 70% ethanol. Sex was determined by gonadal inspection or by the presence of prepollical spines and/or vocal sac folds in males.
Multivariate analyses of variance (MANOVA) and Principal Components Analysis (PCA) were used to assess the degree of morphometric differentiation among adult individuals of all species considered. To remove the effect of covariation in size, the MANOVA and PCA were applied to the residuals from the linear regressions between six morphometric variables and SVL; this procedure was performed separately for males (n = 136) and females (n = 34). For the PCA, only components with eigenvalues > 1 were retained. Finally, variables were compared between species with Student's ttests after ensuring their fit to a normal distribution.
In the Diagnosis sections, coloration refers to preserved specimens unless otherwise noted. Seven qualitative morphological characters were evaluated: (1) dorsal coloration, (2) ventral coloration, (3) iris coloration, (4) middorsal stripe [i. present, ii. absent], (5) black flecks on the neck and chest [i. present, ii. absent], (6) marks on flank and hidden surfaces of the thighs [i. dark transversal bars, ii. thin dark transversal bars, iii. dark blotches, iv. dark dots] and (7) size and shape of calcar [i. large and triangular, ii. large and conical, iii. small and conical, iv. small and tubercular, v. absent]. Color in life was obtained from color photographs.
Sound recordings were made with a Sennheiser K6-ME67 TM directional microphone with digital recorders Olympus LS-10 TM and Marantz PMD660 TM . Calls were analyzed using Raven 1.3 software (Cornell Lab of Ornithology 2003 at a sampling rate of 44.1 kHz and a frequency resolution of 10.8 Hz. Measured call variables are defined in Table 3. If available, several calls or notes were analyzed per individual to calculate an individual average. Temporal variables were measured on the oscillogram, spectral variables on the power spectrum. Five call variables were used to run a Principal Components Analysis (PCA) to assess the degree of acoustic differentiation between calls from five males of Hypsiboas fasciatus (from La Pradera and Zamora) The conservation status of each species was evaluated according to the IUCN Red List categories and criteria (IUCN 2001). The estimates of the distribution area were based on the minimum convex polygon (the smallest polygon in which no internal angle exceeds 180 degrees and which contains all known localities). Vegetation types are based on WWF Ecoregions (available at http://www.eoearth.org/view/article/151948) except for Ecuadorian localities, which are based in the more detailed classification of Sierra et al. (1999).

Phylogenetic analyses
Phylogenetic analyses are an expansion of those presented by Funk et al. (2012) to include additional sequences of H. calcaratus, H. dentei, H. fasciatus and H. raniceps published by Darst and Cannatella (2004), Faivovich et al. (2004, Salducci et al. (2002Salducci et al. ( , 2005, Jansen et al. (2011), Wiens et al. (2005, 2006 and 19 sequences of the nuclear gene RAG-1 (recombination activating gene 1). GenBank accession numbers for the sequences not included in Funk et al. (2012) are shown in Table 4. The same matrix and alignment of mitochondrial DNA sequence data (genes 12S rRNA, 16S rRNA and COI mtDNA) was employed. For the nuclear DNA, preliminary alignment of RAG-1 and POMC sequences were done with CLUSTAL X 2.0 (Larkin et al. 2007). Manual adjustments to the alignment were made using Mesquite version 2.75 (Maddison and Maddison 2009).  Because it is likely that each of our sampled genes (or codon positions in protein coding genes) evolved under different processes, we partitioned the matrices according to gene and codon position to analyze each partition under separate models of evolution. We used PartitionFinder v. 1.1.1 (Lanfear et al. 2012) to find the best-fit model for each partition and also the best partition strategy. In the mitochondrial matrix, we defined five a priori partitions (12S, 16S and one partition for each codon position of COI). In the nuclear matrix we defined six a priori partitions (corresponding to each codon position of our two genes).
Phylogenetic relationships were inferred separately for mitochondrial and nuclear genes using maximum-likelihood and Bayesian inference. Maximum-likelihood and Bayesian analyses were conducted in GARLI v. 2.0 (Zwickl 2006) and MRBAYES v. 3.2.1 (Ronquist et al. 2012) respectively using the same methodology described by Funk et al. (2012).

Phylogenetic relationships
The phylogenetic relationships recovered from the analysis of the mitochondrial DNA sequences (Fig. 1) were consistent with those reported by Funk et al. (2012). The following sections describe the relationships of the samples not included in Funk et al. (2012).
We found strong support for a clade that includes Hypsiboas dentei and the H. calcaratus-H. fasciatus species complex. This clade is sister to H. lanciformis + H. multifasciatus. There are eleven candidate species, two more than those reported by Funk et al. (2012). The population from Bolivia (clade J) is sister to clades I and H from French Guiana, Guyana, and Suriname. The branch lengths in the phylogeny and the genetic distances between J and I-H (uncorrected p-distance range 3.1-5.8% in gen 16S; Table 5) suggest that J is not conspecific to any of the candidate species reported by Funk et al. (2012). An additional unconfirmed candidate species (clade K) is represented by a single sample (NMP6V 71250; reported as Hypsiboas calcaratus by Faivovich et al. 2005) from 50 km W of Iquitos (Peru). The genetic distance relative to its closest relative (clade D) ranges from 4.8 to 5.9% (16S), strongly suggesting that it is a separate species. The branch lengths that separate clade K from clade D are longer than the lengths that separate pairs of some confirmed candidate species (e.g., Hypsiboas fasciatus vs. Hypsiboas almendarizae sp. n.) further implying that K is a valid undescribed species.
The phylogeny based on the nuclear genes (RAG-1 and POMC) is generally consistent with the mitochondrial phylogeny but is less resolved (Fig. 2). Similarly to the mitochondrial phylogeny, it shows strong support for two basal clades (Clades A-D and Clades E-F). At the species level, it provides strong support for clades A and D.
The only strongly supported disagreement between the nuclear and the mitochondrial phylogeny is the placement of individual QCAZ 44177 because in the nuclear tree it appears within a clade formed by individuals from the mitochondrial clades A and B, while in the mitochondrial tree it appears as a member of clade C. The morphological data shows that QCAZ 44177 is a member of clade C because it has its distinctive characters (e.g., large and triangular calcar).

taxonomic review
The available names for the populations sampled in our phylogenies are Hypsiboas calcaratus (Troschel, 1848), Hypsiboas fasciatus (Günther, 1858), Hyla leptoscelis Boulenger, 1918, andHyla steinbachi Boulenger, 1905. Examination of three of the holotypes and published descriptions from the literature allowed us to assign the available names to clades A and C (Fig. 1). We document those assignments in the following section.

Taxonomic status of H. fasciatus and H. calcaratus
The holotype of H. fasciatus is an adult female (SVL = 58.59 mm; BM 1858.7. 25.22, reported as "BM 58.4.25.22" by Duellman 1973;Fig (1858) the type locality is vaguely stated as "Anden von Ecuador", it is almost certain that the holotype was collected either in Gualaquiza or Zamora, Morona Santiago and Zamora Chinchipe provinces, Ecuador. We infer this from letters and records of the collections of the travels of Louis Fraser (Gardner 1983 and references therein). During his trip, which lasted from 1857 to 1859, the only collection sites within the distribution range of the H. fasciatus-H. calcaratus species complex are Gualaquiza (elevation 1100 m) and Zamora (1000 m). Both localities overlap with the distribution range of clade A exclusively. Clade A has an elevation range from 700 to 1600 m. Duellman (1973) characterized H. fasciatus as having, in the majority of the specimens, brown flecks on the throat and chest and black spots on the flanks and thighs. The holotype of H. fasciatus differs from this diagnosis because it has dark bars instead of dots on the flanks and thighs and lacks flecks on the throat and chest. Most of the specimens reported by Duellman (1973) as "H. fasciatus" are from localities below 700 m and resemble individuals from our clades D, E, and F (Figs 4D-E and 8D-F). Although Duellman (1973) noted the aforementioned differences between the holotype and the other individuals, he considered them conspecific.  The holotype of H. calcaratus could not be examined because it is lost (Duellman 1973;Frost 2013). Nevertheless, the description of the holotype of H. calcaratus by Troschel (1848) and the location of the type locality (British Guiana) suggest that the binomial H. calcaratus is conspecific to clade C (Fig. 1). The holotype was described as having a long calcar, perpendicular short black stripes on the flanks, and transversal bars on the legs, which are characteristic of clade C. This designation is also supported by the type locality "British Guiana" because it overlaps with the distribution of clade C. The presence of clade C in the Republic of Guyana (formerly known as British Guiana) is documented by photographs and the morphological description of "Hypsiboas calcaratus" by Kok and Kalamandeen (2008). Clade C is the only species with developed calcar and dark bars on the legs and flanks that occurs below 500 m (Fig. 9). Based on the available evidence, we assign the binomial Hypsiboas calcaratus to clade C.
Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in Figure 9. Background dorsal coloration varies from cream (e.g., QCAZ 40085) to pinkish white (e.g., QCAZ 44530), reddish brown (e.g., QCAZ 14957, 43256, 44422), pale reddish brown (e.g., QCAZ 43259) or pale grayish brown (e.g., QCAZ 48718). Irregular dorsal marks may be present in diverse patterns. A dark middorsal line extends from the tip of the snout to the mid-sacrum (e.g., QCAZ 43256), but in some specimens it only extends along the head (e.g., QCAZ 25514) or on the anterior half of the body (e.g., QCAZ 43131). There is variation in the number, size, and shape of dorsal marks. Some individuals (e.g., QCAZ 43256) have five to seven brown diffuse transversal bands (sometimes interconnected). Brown transversal bars are present on the dorsal surfaces of the limbs (one or two on the upper arm and forearm and three to five on the thigh, shank, and foot). In some individuals, the dorsum and dorsal surfaces of the forearms and shanks have large black stains (e.g., QCAZ 14957) or scattered brown or white dots (e.g., QCAZ 40085, 44178, 14971). The coloration of flanks and hidden surfaces of thighs vary from pale cream to creamy white or light gray, with dark brown transversal bars. The number of bars on the flank varies from 4 to 13; the number of bars on the thigh varies from 4 to 9. The extent of the area with bars varies from the groin to the mid flank (e.g., QCAZ 43259) to from the groin to the axilla (e.g., QCAZ 43256). In some individuals, the bars can also be present on the hidden surfaces of the shanks, ventral surfaces of the forelimbs, and dorsal surfaces of the feet (e.g., QCAZ 43256).  Table 8 for character loadings on each component. Ventral surfaces of preserved specimens vary from creamy white (e.g., QCAZ 44530) to yellowish white (e.g., QCAZ 43256). In some individuals, scattered minute pale brown blotches are present on the lips (e.g., QCAZ 31446, 44178). Coloration of webbing and discs vary from yellowish white to brown or gray. Coloration of bones is white or green.
In the examined adult series, the largest male has a SVL of 42.50 mm, and the largest female 56.29 mm; mean male SVL = 37.08 mm (n = 35; SD = 2.09), mean female SVL = 50.92 mm (n = 4; SD = 4.80). Females are significantly larger than males (t = -5.71, df = 3, P = 0.009). Inter-population variation in size and other morphometric variables is shown in Tables 1 and 2 Hypsiboas calcaratus occurs in Terra Firme forest, flooded forests (Várzea and Igapó), and swamps. It is generally found next to streams, ponds, and lakes. Individuals have been recorded at night perching on vegetation 15 to 200 cm above the ground. Their occurrence in secondary forests and artificial open areas suggest at least some tolerance of anthropogenic habitat disturbance.
Vegetation types at known localities include Southwest Amazon Moist Forest and Napo Moist Forest for the Peruvian and Ecuadorian localities, Guianan Moist Forest for the Guyana and French Guiana localities, and Madeira-Tapajós Moist Forest for the Brazilian locality (according to the World Wildlife Fund, 2012).
Conservation status. Its distribution polygon has 3'586,597 km 2 and overlaps with protected areas and large regions of pristine forest. Hypsiboas calcaratus is relatively frequent in scientific collections suggesting that, at least in part of its range, it is not a rare species. For these reasons we propose assigning H. calcaratus to the Red List category of Least Concern.   Table 6.
Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in Figure 5. Background dorsal coloration varies from cream (e.g., QCAZ 41488, 48584) to pale grayish brown (e.g., QCAZ 18271), grayish brown (e.g., QCAZ 27258), pinkish white (e.g., QCAZ 41575), pale brown (e.g., QCAZ 23148) or brown (e.g., QCAZ 26497). Irregular dorsal marks can be present in diverse patterns. A dark middorsal stripe extends from the tip of the snout to the mid-sacrum (e.g., QCAZ 26497) or to the vent (QCAZ 23144). In a few individuals, the middorsal line is absent (QCAZ 17123, 31040). In some individuals (e.g., QCAZ 41575) broad transversal bands are present on the dorsum; narrower brown transversal bars are present on the dorsal surfaces of the limbs (one or two on the upper arm and forearm and three to five on the thigh, shank, and foot). In some individuals, scattered minute brown dots may be present on dorsum (e.g., QCAZ 48584). The coloration of flanks and hidden surfaces of thighs vary from pale cream to creamy white, brown or gray, with dark brown vertical bars (4 to 12 on each flank and 6 to 12 on each thigh). The area with vertical bars extends from the groin to the axilla (e.g., QCAZ 41575) or to the mid flank (e.g., QCAZ 41576, 18271). In some individuals, similar transversal bars can be present on the hidden surfaces of the shanks, ventral surfaces of the forelimbs, and dorsal surfaces of the feet (e.g., QCAZ 41575). The shape of the calcar is small and conical and shows little variation among individuals.
Ventral areas vary from creamy white (e.g., QCAZ 48584) to yellowish white (e.g., QCAZ 41576, 41488). A narrow to wide brown stripe can be present on the outer edge of the hands, forearms, thighs, feet, and tarsal folds (e.g., QCAZ 48584, 41576, 41575, 26497). In some specimens, scattered minute pale brown blotches can be present on the lips (e.g., QCAZ 26497, 24866). Coloration of webbing and discs vary from yellowish white to brown or gray. Bones are white.
In the examined series of adults, the largest male has a SVL of 37.74 mm, and the largest female 54.84 mm; mean male SVL = 35.40 mm (n = 19; SD = 1.65), mean female SVL = 51.89 mm (n = 5; SD = 3.18). Females are significantly larger than males (t = -16.24, df = 22, P < 0.001). Inter-population variation in size and other morphometric variables is shown in Tables 1 and 2.
Advertisement call. We recorded calls of one male at Bombuscaro (Provincia Zamora Chinchipe) on 9 June 2010, three males at La Pradera (Provincia Morona Santiago) on 10 June 2010, and one male at Comunidad San Luis (Provincia Morona Santiago) on 13 June 2010. The advertisement call consists of three to five quack notes (Fig. 12A-B). Mean call duration is 0.52 s (SD = 0.24) and mean rise time is 0.24 s (SD = 0.06). Sound frequency increases from the beginning to the end of the note; mean dominant frequency is 1855.81 Hz (SD = 148.08) and mean fundamental frequency is 884.89 Hz (SD = 105.52). Other call parameters are listed in Table 7.
Description of the holotype. Adult male, 36.72 mm SVL, foot length 15.93 mm, head length 9.18 mm, head width 11.95 mm, eye diameter 4.26 mm, tympanum diameter 2.41 mm, tibia length 21.78 mm, femur length 20.03 mm, calcar length 1.10 mm, arm length 6.17 mm, eye-nostril distance 2.7 mm, head wider than long and wider than body; snout round in lateral view, truncate in dorsal view; distance from nostril to eye shorter than diameter of eye; canthus rostralis indistinct, rounded; loreal region concave; internarial area convex; nostrils not protuberant, directed laterally; interorbital area slightly convex; eye large, strongly protuberant; diameter of eye 1.8 times diameter of tympanic annulus; tympanum concealed beneath skin; tympanic annulus evident, ovoid, longer dorsoventrally, concealed dorsally by supratympanic fold, separated from eye by ca. 1.2 times its diameter; posterior end of supratympanic fold reaches anterior border of arm insertion. Arm slender, axillary membrane absent; indistinct low tubercles present along ventrolateral edge of forearm; relative length of fingers I<II<IV<III; fingers bearing large, oval discs, that of third finger about three fourths of tympanum diameter; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles present; palmar tubercle small, elongated; prepollical tubercle large, flat, elliptical; prepollex enlarged, claw shaped; nuptial excrescences absent; webbing absent between fingers. Large conical calcar on tibiotarsal articulation; ill defined, scattered tubercles on tarsus and along ventrolateral edge of foot; toes bearing discs slightly wider than long, smaller than those of fingers; relative length of toes I<II<V<III<IV; outer metatarsal tubercle ill defined, small, round; inner metatarsal tubercle large, elliptical; subarticular tubercles single, round, flat; supernumerary tubercles restricted to the soles; webbing formula of toes I2-2¾II1¾-3 -III2 --3 + IV3 + -2 -V. Skin on dorsum, head, and dorsal surfaces of limbs smooth; skin on flanks smooth with weak longitudinal wrinkles posterior to the arm; skin on venter coarsely granular; skin on ventral surfaces of head and thighs granular, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs; cloacal sheath short simple, covering cloacal opening; round tubercles below and on the sides of vent, larger proximally. Tongue cordiform, widely attached to mouth floor; vomerine odontophores triangular, narrowly separated, posteromedial to choanae, bearing six vomerine teeth on each side; choanae trapezoidal, oblique.
Color of holotype in preservative. Dorsum pale grayish brown with five to seven broad brown transversal bands and scattered minute black dots; few small cream dots on the posterior one third of the dorsum; dark brown middorsal line from tip of the snout to the mid-dorsum; brown transversal bars on dorsal surfaces of the limbs (two on the upper arm, five on the thigh, and three on the shank and foot); flanks pale cream with ten thin dark brown transversal bars; hidden surfaces of thighs pale cream with seven to nine thin dark brown transversal bars; narrow cream stripe present above the anus; venter yellowish white becoming creamy white on its posterior half; ventral surfaces of hindlimbs, forelimbs, and webbing yellowish white; faint narrow to wide pale brown stripe on the outer edge of the hands and forearms; minute brown blotches on lips; bones white.
Color of holotype in life. (Fig. 14E). Dorsum pale yellow tan with pale broad brown transversal marks and scattered minute black and white dots; dark brown middorsal line is present; pale brown transversal bars on the dorsal surfaces of the limbs; flanks white with faint thin dark brown vertical bars; hidden surfaces of thighs white with thin dark brown transversal bars; venter creamy white; ventral surfaces of hindlimbs and forelimbs translucent pinkish; discs and webbing yellowish; iris cream with upper yellow band; bones white.
Etymology. The specific name is a noun in the genitive case and is a patronym for Ana Almendáriz, Ecuadorian herpetologist who for more than three decades has contributed to the study of Ecuadorian amphibians and reptiles. Ana Almendáriz is curator of Herpetology in the Museo de Historia Natural Gustavo Orcés at Escuela Politécnica Nacional del Ecuador.
Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in Figure 6. Background dorsal coloration varies from creamy white (e.g., QCAZ 32638, 32645) to reddish brown, grayish brown (e.g., QCAZ 24394, 39647), pale grayish brown (e.g., QCAZ 24392, 39650) or pale brown (e.g., QCAZ 24386). A dark middorsal line extends from the tip of the snout to the middle of the sacrum (e.g., QCAZ 39647) or to the vent (e.g., QCAZ 32645), but in some specimens is restricted to the head (e.g., QCAZ 26429) or to the anterior half of the body (e.g., QCAZ 39650). In few specimens, the middorsal line is faint or absent (e.g., QCAZ 24392). Irregular dorsal marks may be present in varying number, size, and shape. Five to seven broad transversal bands (sometimes interconnected) can be present on the dorsum. Some individuals (e.g., QCAZ 24394, 39650, 39647) have narrow brown transversal bars on the dorsal surfaces of the limbs (one or two on the upper arm and forearm, and three to five on the thigh, shank, and foot). Faint to well-defined narrow longitudinal lines may be present on the dorsum (e.g., QCAZ 32638, 32645). In some individuals (e.g., QCAZ 32645), there are few scattered brown or white dots on the dorsum, and dorsal surfaces of forearms, shanks and tarsi. The coloration of the flanks and hidden surfaces of the thighs vary from pale cream to creamy white, with thin dark brown transversal bars. The number of bars on the flanks and thighs varies from six to ten on each flank and six to thirteen on each thigh; the area with vertical bars extends from the groin to the mid-flank. In few individuals, the dark transversal bars of the thighs and flanks are faint (e.g., QCAZ 32645).
Ventral surfaces of preserved specimens vary from creamy white (e.g., QCAZ 39650) to yellowish white (e.g., QCAZ 39647). A narrow to wide faint pale brown stripe is present in some individuals on the outer edge of the hands, forearms, feet, thighs, and tarsal folds (e.g., QCAZ 39647). In some individuals, scattered minute pale brown blotches may be present on the lips (e.g., QCAZ 32638, 39647). Coloration of webbing and discs vary from yellowish white to brown or gray. Vomerine odontophores are triangular (with arched base in some specimens). Bones white.
In the adult type series, the largest male has a SVL 44.56 mm, and the largest female 51.94 mm; mean male SVL = 37.64 mm (n = 23; SD = 2.01), mean female SVL = 48.11 mm (n = 4; SD = 6.88). Females were not significantly larger than males (t = -3.02, df = 3, P = 0.055). Inter-population variation in size and other morphometric variables is shown in Tables 1 and 2.
Advertisement call. We recorded calls from five males at Limón Indanza (Provincia Morona Santiago) on 17 January 2009 (Table 7). The advertisement call consists of three to four quack notes (Fig. 12C-D). Mean call duration is 0.48 s (SD = 0.007) and mean rise time is 0.30 s (SD = 0.10). Sound frequency increases from the beginning to the end of the note; mean dominant frequency is 1954.43 Hz (SD = 128.43) and mean fundamental frequency is 951.76 Hz (SD = 61.38). The advertisement calls of H. almendarizae and H. fasciatus are similar to each other ( Fig. 12 A-D). The distribution ranges of these species are parapatric with a small area of sympatry in the surroundings of Leonidas Plaza (Provincia Morona Santiago). The analyzed calls from H. fasciatus were from a region of allopatry. Future studies should aim to compare calls from the region of sympatry to explore the possibility of reproductive character displacement.
Distribution and ecology. Hypsiboas almendarizae occurs on the eastern Andean slopes of central and southern Ecuador (Morona Santiago, Napo, and Tungurahua provinces) (Fig. 15). Localities with known elevation range from 500 to 1950 m above sea level. The elevation at Río Hollín (1950 m) is the highest known locality for H. almendarizae, while Méndez (500 m) is the lowest.
Most specimens of H. almendarizae were collected at Río Napinaza, a river surrounded by secondary forest, pastures, and agricultural lands. Frogs were perching on vegetation 25 to 80 cm above the ground. Individuals from Limón Indanza, Río Hollín, Río Lagarto, Méndez, Río Pastaza, and Nueve de Octubre were found in flooded areas with pastures, on vegetation 30 to 70 cm above the ground. Few individuals were found near river Hollín, in small ponds in primary and secondary forest, roosting on branches and leaves 20 to 150 cm above the ground. All the specimens from Rosario and Plan de Milagro were found next to a highway, in swamps and streams with grass. Conservation status. Its distribution polygon has 14,983 km 2 . Within this area, 4,864 km 2 (32.4%) of its habitat has been degraded by human activities, especially agriculture and cattle raising (estimated from Ministerio de Ambiente del Ecuador 2013). Current habitat degradation within the range of H. almendarizae is extensive and its increase may threaten its survival in the future. Therefore, we propose that H. almendarizae is assigned to the Near Threatened category. Its distribution polygon overlaps with Parque Nacional Sangay, Parque Nacional Llanganates, and Parque Nacional Sumaco Napo-Galeras. (9) ventral coloration varying from creamy white to yellowish white with or without brown flecks on the neck and chest; (10) webbing on feet; (11) in life, iris cream silver, yellowish or cream with faint yellow to orange color on its upper quarter; (12) prepollical spine present in males.
Description of holotype. Adult male, 36.16 mm SVL, foot length 14.33 mm, head length 8.98 mm, head width 11.71 mm, eye diameter 4.07 mm, tympanum diameter 2.44 mm, tibia length 20.29 mm, femur length 18.44 mm, calcar length 1.44 mm, arm length 6.36 mm, eye-nostril distance 2.86 mm, head wider than long and wider than body; snout rounded in lateral view, truncate in dorsal view; distance from nostril to eye shorter than diameter of eye; canthus rostralis indistinct, rounded; loreal region concave; internarial area convex; nostrils not protuberant, directed laterally; interorbital area slightly convex; eye large, strongly protuberant; diameter of eye 1.7 times diameter of tympanic annulus; tympanum concealed beneath skin; tympanic annulus evident, ovoid, longer dorsoventrally and concealed dorsally by supratympanic fold, separated from eye by ca. 1.1 times its diameter; posterior end of supratympanic fold reaches anterior border of arm insertion. Arm slender, axillary membrane absent; indistinct low tubercles along ventrolateral edge of forearm; relative length of fingers I<II<IV<III; fingers bearing large, oval discs, that of third finger about three fourths of tympanum diameter; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles present; palmar tubercle small, elongated; prepollical tubercle large, flat, elliptical; prepollex enlarged, claw shaped; nuptial excrescences absent; webbing absent between fingers. Large triangular calcar on tibiotarsal articulation; ill defined, scattered tubercles on tarsus and along ventrolateral edge of foot; toes bearing discs slightly wider than long, smaller than those of fingers; relative length of toes I<II<V<III<IV; outer metatarsal tubercle ill defined, small, round; inner metatarsal tubercle large, elliptical; subarticular tubercles single, round, flat; supernumerary tubercles restricted to the soles; webbing formula of toes I2-2 + II1¾-3 -III2 --3IV3-1½V. Skin on dorsum, head, and dorsal surfaces of limbs smooth; skin on flanks smooth with weak longitudinal wrinkles posterior to the arm; skin on venter coarsely granular; skin on ventral surfaces of head and thighs granular, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs; cloacal sheath short and simple, covering cloacal opening; round tubercles below and on sides of vent, larger proximally. Tongue ovoid, widely attached to mouth floor; vomerine odontophores triangular with arched base, narrowly separated, posteromedial to choanae, each bearing 8-12 teeth; choanae ovoid.
Color of holotype in preservative. Dorsum pale reddish brown with two faint brown narrow longitudinal lines interconnected in the scapular region (Fig. 16); scattered min- ute black dots distributed on dorsum; dark brown middorsal line along snout; dorsal surfaces of limbs pale reddish brown with five pale brown narrow transversal bars on each thigh; flanks and hidden surfaces of thighs pale cream with dark brown blotches; narrow white stripe above vent; venter creamy white; ventral surfaces of limbs and webbing yellowish white; narrow to wide pale brown stripe on the outer edge of the hands, forearms, thighs, tarsal folds, and feet; minute brown blotches on lips; bones white.
Color of holotype in life. (Fig. 14F). Dorsum creamy white with six narrow pale brown longitudinal lines; a dark brown middorsal line; five pale brown faint transversal bars on dorsal surface of each thigh and two narrow pale brown longitudinal lines on dorsal surface of each shank; scattered minute black dots on dorsum; flanks and hidden surfaces of the thighs light-bluish white with dark brown blotches; discs and webbing pale yellowish tan; iris cream with faint yellow coloration on its upper margin.
Etymology. The specific name is derived from the Latin words macula = stain, and lateralis = lateral, in reference to the brown dark blotches on the flanks of these frogs. Figure 16. Background dorsal coloration varies from creamy white (e.g., QCAZ 44020) to reddish brown (e.g., QCAZ 28401), pale reddish brown (e.g., QCAZ 28004), pale pinkish white (e.g., QCAZ 44021, 43897), pinkish white (e.g., QCAZ 44531,44651,44184) or brown (e.g., QCAZ 43825). A dark middorsal line extends from the tip of the snout to the middle of dorsum (e.g., QCAZ 2156), but in some specimens is restricted to the snout (e.g., QCAZ 43825, 44651, 28004) or is absent (e.g., QCAZ 44531). There is variation in the number, size, and shape of the dorsal marks. Five to seven broad transversal bands on the dorsum (sometimes interconnected) and narrower brown transversal bars on the dorsal surfaces of the limbs (one or two each on upper arm and forearm and three to five each on thigh, shank and foot) may be present (e.g., QCAZ 28401). In some individuals, faint, narrow longitudinal dark lines may be present (e.g., QCAZ 43897). In a few specimens, the dorsum, dorsal surfaces of forearms, and shanks may present large black stains (e.g., QCAZ 44021) and/or few scattered brown dots (e.g., QCAZ 43897, 44020). The coloration of the flanks and hidden surfaces of thighs vary from pale cream to creamy white or gray, with dark brown blotches. There is variation in the distribution of blotches on flanks; the extent covered by the blotches ranges from groin to the mid-flank (e.g., QCAZ 28004) or to the axilla (e.g., QCAZ 55374). In some individuals, similar blotches can also be present on the hidden surfaces of the shanks, ventral surfaces of the forelimbs, and dorsal surfaces of the feet (e.g., QCAZ 43825). In few individuals, the dark blotches of the thighs and flanks are faint (e.g., QCAZ 43897).

Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in
Ventral surfaces of preserved specimens vary from creamy white (e.g., QCAZ 44020, 44184; Fig. 16) to yellowish white (e.g., QCAZ 28004). Some individuals have brown flecks on the gular region, chest, and lips (e.g., QCAZ 43825). A narrow to wide brown stripe is present in some individuals on the outer edge of the hands, forearms, thighs, tarsal folds and feet (e.g., QCAZ 28004, 28401, 44021). Coloration of webbing and disc varies from yellowish white to brown or gray. Vomerine odontophores are triangular (with arched base in some specimens). Coloration of bones is white.
In the adult type series, the largest male has a SVL of 39.17 mm, and the largest female 55.31 mm; mean male SVL = 36 mm (n = 22; SD = 1.92), mean female SVL = 45.18 mm (n = 6; SD = 9.39). Females are significantly larger than males (t = -3.43, df = 4, P = 0.025). Inter-population variation in size and other morphometric variables are shown in Tables 1 and 2.
Advertisement call. A call from a male (QCAZ 40082) at Comunidad Santa Rosa (Napo Provincia) recorded on 28 February 2009 (Table 7) consists of three to four notes that sound like a high-pitched chuckle. The mean call duration is 0.35 s (SD = 0.04) and the mean rise time is 0.19s (SD = 0.1). The dominant frequency is higher (mean = 2217.93 Hz, SD = 56.94) and the fundamental frequency lower (mean = 488.10 Hz, SD = 12.47) compared to H. fasciatus, H. almendarizae, H. calcaratus, H. alfaroi sp. n. and H. tetete sp. n.
Distribution and ecology. Hypsiboas maculateralis inhabits the Amazon basin of Ecuador (Napo, Orellana, Pastaza, and Sucumbíos provinces) and Peru (Región de Madre de Dios) (Fig. 17). Localities with known elevation range vary between 186 and 354 m of elevation. The elevation of Comunidad Santa Rosa (354 m) is the highest known and Santa Teresita (186 m) is the lowest.
All the specimens in Zábalo, Reserva Cuyabeno, Santa Teresita, Bataburo Lodge, and km 74 Pompeya-Iro road were found at night, in primary or secondary forests, perching on vegetation between 40 and 200 cm above the ground, in areas of pasture and flooded forest.
Vegetation types for Ecuadorian localities are: (1) Amazonian Lowland Evergreen Forest, characterized by high plant alpha-diversity and a canopy of 30 m with emergent trees that reach 40 m, (2)   fingers bearing large, oval discs, that of third finger about three fourths of tympanum diameter; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles present; palmar tubercle small, elongated; prepollical tubercle large, flat, elliptical; prepollex enlarged, claw shaped; nuptial excrescences absent; webbing absent between fingers. Small tubercle on tibiotarsal articulation; ill defined, scattered tuber-cles on tarsus and along ventrolateral edge of foot; toes bearing discs slightly wider than long, smaller than those of fingers; relative length of toes I<II<V<III<IV; outer metatarsal tubercle ill defined, small, round; inner metatarsal tubercle large, elongated, elliptical; subarticular tubercles single, round, flat; supernumerary tubercles restricted to the soles; webbing formula of toes I2-2II2-3III2-3IV3-1¾V. Skin on dorsum, head, and dorsal surfaces of limbs smooth; skin on flanks smooth with weak longitudinal wrinkles posterior to the arm; skin on venter coarsely granular; skin on ventral surfaces of head and thighs granular, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs; short simple cloacal sheath covering cloacal opening; round tubercles below and on the sides of vent. Tongue ovoid, widely attached to mouth floor; vomerine odontophores triangular, narrowly separated, posteromedial to choanae, bearing 10 vomerine teeth on each side; choanae trapezoidal, oblique.
Color of holotype in preservative. Dorsum grayish brown with five to six broad diffuse brown transversal bands; scattered minute black dots on dorsal surfaces; few small cream dots restricted to the posterior dorsum; faint brown narrow middorsal line restricted to the head; flanks creamy white with dark brown dots; dorsal surfaces of hindlimbs and forelimbs grayish brown with narrow transversal brown bars (one or two on each upper arm and forearm, and three or four on each thigh, shank, and tarsus); hidden surfaces of thighs grayish brown with dark brown dots; venter creamy white with brown flecks on head, neck, and chest; ventral surfaces of hindlimbs and forelimbs yellowish white with a narrow to wide brown stripe on the outer edge of the hands, forearms, tarsal folds, and feet; bones white.
Etymology. The specific name is a noun in the genitive case and is a patronym for Eloy Alfaro Delgado, former Ecuadorian president (1897-1901 and 1906-1911) and leader of the liberal revolution in Ecuador. His government promoted the separation between church and state and the modernization of Ecuador by supporting education and large-scale systems of transportation and communication.
Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in Figure 18. Background dorsal coloration varies from creamy white (e.g., QCAZ 44429) to pale grayish brown (e.g., QCAZ 19328), grayish brown (e.g., QCAZ 44180), pale brown (e.g., QCAZ 28398) or brown (e.g., QCAZ 28272). Irregular dorsal marks may be present in diverse patterns. A faint dark middorsal line extends from the tip of the snout to the mid-dorsum (e.g., QCAZ 43895) or the sacral region (QCAZ 44025), but in some specimens it is restricted to the head (e.g., QCAZ 55926) or is altogether absent (e.g., QCAZ 28272). There is variation in the number, size, and shape of the dorsal marks. Five to seven diffuse broad transversal bands on the dorsum (sometimes interconnected), and narrower brown transversal bars on the dorsal surfaces of the limbs (one or two each on the arm and forearm, and three to five each on the thigh, shank, and foot) may be present (e.g., QCAZ 28272). Faint, narrow longitudinal lines may be present on the dorsum (e.g., QCAZ 44858, 43263). The dorsum and dorsal surfaces of limbs can have scattered black or white dots (e.g., QCAZ 43263, 28398). The coloration of flanks varies from creamy white to gray with irregular dark brown spots distributed from the groin to the mid-flank. In some indi-viduals, similar spots can also be present in the hidden surfaces of the thighs, shanks, ventral surfaces of the forelimbs, and dorsal surfaces of the feet (e.g., QCAZ 28272). The hidden surfaces of thighs sometimes have dark brown blotches (QCAZ 28272).
Ventral areas of preserved specimens vary from creamy white (e.g., QCAZ 28398, 44180) to yellowish white (e.g., QCAZ 43263) with dark flecks on the head, neck, and chest (Fig. 18). In some individuals, the flecks are also present on hindlimbs, forelimbs, and belly (e.g., QCAZ 28272, 28398). A narrow to wide brown stripe is present in some individuals on the outer edge of the hands, forearms, thighs, tarsal folds, and feet (e.g., QCAZ 43263). Coloration of webbing and disc vary from brown to gray. Vomerine odontophores are triangular (with arched base in some specimens). Bones are white.
In the adult type series, the largest male has a SVL of 36.27 mm, and the largest female 49.21 mm; mean male SVL = 32.80 mm (n = 32; SD = 1.97), mean female SVL = 44.51 mm (n = 12; SD = 3.09). Females were significantly larger than males (t = -14.94, df = 42, P < 0.001). Inter-population variation in size and other morphometric variables are shown in Tables 1 and 2.
Advertisement call. Three males were recorded at Estación Científica Yasuní PUCE (Provincia Orellana) on 20 June 2009. The call (Fig. 13C-D) consists of four to five trill-like notes with a mean duration of 0.20 s (SD = 0.05) and mean rise time of 0.07 s (SD = 0.03). Other call parameters are listed in Table 7. 45.85; n = 2); (2) basal webbing on the fingers; (3) tubercle on the heel; (4) dorsal background coloration varying from grayish brown to pale brown, sometimes with dark markings (e.g., diffuse broad transversal bands); (5) faint brown middorsal line often present; (6) flanks creamy white or gray (in life, light blue in large females and white in males) with dark brown irregular spots; (7) hidden surfaces of thighs creamy white or brown with dark brown irregular spots or blotches; (8) ventral surfaces of thighs creamy white or yellowish white; (9) ventral areas creamy white or yellowish white with brown flecks on the head, neck, and chest; (10) webbing on the feet; (11) in life, iris yellow or cream with yellow on its anterior half; (12) prepollical spine present in males.
Hypsiboas tetete differs from H. fasciatus, H. almendarizae, H. calcaratus, and H. maculateralis in advertisement call (Figs 12A-F and 13A-H) and by having a tubercle on the heel instead of a calcar. Hypsiboas tetete is most similar to H. alfaroi from which differs in advertisement call (Fig. 13C-H). Although with overlapping values, Hypsiboas tetete has a statistically significant larger tympanum than H. alfaroi (see H. alfaroi diagnosis). Morphological characters useful to differentiate H. tetete from other species are shown in Table 6.
Description of holotype. Adult male, SVL 31.15 mm, foot length 12.01 mm, head length 7.48 mm, head width 10.31 mm, eye diameter 3.51 mm, tympanum diameter 2.25 mm, tibia length 17.76 mm, femur length 16.09 mm, arm length 5.50 mm, eye-nostril distance 1.97 mm, head wider than long and wider than body; snout rounded in lateral view, truncate in dorsal view; distance from nostril to eye shorter than diameter of eye; canthus rostralis indistinct, rounded; loreal region concave; internarial area convex; nostrils not protuberant, directed laterally; interorbital area slightly convex; eye large, strongly protuberant; diameter of eye 1.6 times diameter of tympanic annulus; tympanum concealed beneath skin; tympanic annulus evident, ovoid, longer dorsoventrally and concealed dorsally by supratympanic fold, separated from eye by ca. 1.03 times its diameter; posterior end of supratympanic fold reaches anterior border of arm insertion. Arm slender, axillary membrane absent; indistinct low tubercles present along ventrolateral edge of forearm; relative length of fingers I<II<IV<III; fingers bearing large, oval discs, that of third finger about three fourths of tympanum diameter; subarticular tubercles prominent, round to ovoid, single; supernumerary tubercles present; palmar tubercle small, elongated; prepollical tubercle large, flat, elliptical; prepollex enlarged, claw shaped; nuptial excrescences absent; webbing absent between fingers. Small tubercle on tibiotarsal articulation; ill defined, scattered tubercles on tarsus and along ventrolateral edge of foot; toes bearing discs slightly wider than long, smaller than those of fingers; relative length of toes I<II<V<III<IV; outer metatarsal tubercle ill defined, small, round; inner metatarsal tubercle large, elongated and elliptical; subarticular tubercles single, round, flat; supernumerary tubercles restricted to the soles; webbing formula of toes I2 --2½II2-3 + III2-3 + IV3 + -2 -V. Skin on dorsum, head, and dorsal surfaces of limbs smooth; skin on flanks smooth with weak longitudinal wrinkles posterior to the arm; skin on venter coarsely granular; skin on ventral surfaces of head and thighs granular, those of shanks smooth. Cloacal opening directed posteriorly at upper level of thighs; short simple cloacal sheath covering cloacal opening; round tubercles below and on the sides of vent. Tongue ovoid, widely attached to mouth floor; vomerine odontophores triangular with arched base, barely separated, posteromedial to choanae, bearing eight vomerine teeth on each side; choanae ovoid.
Color of holotype in preservative. Dorsum grayish brown with scattered minute black dots; faint brown narrow middorsal line extends from the tip of the snout to the mid-sacrum; dorsal surfaces of hindlimbs and forelimbs grayish brown with narrow transversal brown bars (one or two on each forearm and three or four on each thigh, shank, and tarsus); flanks creamy white with dark brown irregular spots; hidden surfaces of thighs grayish brown with dark brown irregular spots; venter creamy white with brown flecks on the neck, chest, and lips; ventral surfaces of hindlimbs and forelimbs yellowish white with a narrow to wide brown stripe on the outer edge of the hands, forearms, thighs, tarsal folds, and feet; bones white.
Color of holotype in life. (Fig. 14G). Dorsum pale yellowish tan with four narrow pale brown longitudinal lines; a dark brown middorsal line extends from the tip of snout to mid-sacrum; dorsal surfaces of hindlimbs yellowish tan with pale brown transversal bars; scattered minute dark brown dots on the dorsal surfaces of limbs and dorsum; flanks white with well-defined dark brown irregular spots; hidden surfaces of thighs pale yellowish tan with dark brown spots; venter creamy white with yellowish white belly; ventrally, scattered brown flecks on the chest, gular region, and jaw margin; ventral surfaces of hindlimbs and forelimbs creamy white; discs and webbing pale yellow tan; iris cream with faint yellow coloration on its upper half; bones white.
Etymology. The specific name is a noun and refers to the Tetete, a Western Tucanoan indigenous group that inhabited the Colombian and Ecuadorian Amazon. It was decimated by the rubber exploitation and became extinct during the 1970s (Wasserstrom et al. 2011). Its recent disappearance parallels the destruction of increasingly large areas of forest in the Ecuadorian Amazon with the ensuing decline of biodiversity.
Variation. Variation in dorsal and ventral coloration of preserved specimens is shown in Figure 19. Background dorsal coloration varies from grayish brown (e.g., QCAZ 48094) to pale grayish brown (e.g., QCAZ 48096) or pale brown (e.g., QCAZ 48094). Irregular dorsal marks may be present in diverse patterns. A faint middorsal line extends from the tip of snout to mid-dorsum (e.g., QCAZ 40079), mid-sacrum (e.g., QCAZ 40060) or to the cloaca (e.g., QCAZ 48095-96). There is variation in the number, size and shape of the dorsum marks. Five to seven broad transversal bands (sometimes interconnected) may be present on the dorsum; the dorsal surfaces of the limbs have brown transversal bars (one or two each on upper arm and forearm and three to five each on thigh, shank, and foot) (e.g., QCAZ 40080). Some individuals have scattered black or white dots on the dorsum (e.g., QCAZ 40060). The coloration of flanks varies from creamy white to gray with conspicuous dark brown irregular spots distributed from the groin to mid-flank. In some individuals, similar spots can also be present on the hidden surfaces of shanks and dorsal surfaces of feet (e.g., QCAZ 48096). The coloration of the hidden surfaces of thighs varies from creamy white to brown, with dark brown spots (e.g., QCAZ 40060) or blotches (QCAZ 48096).
Ventral areas of preserved specimens vary from creamy white (e.g., QCAZ 48094, 48096) to yellowish white (e.g., QCAZ 40079) with scattered flecks on the head and chest. In some individuals, the flecks are also present on hindlimbs, forelimbs, and belly (e.g., QCAZ 48095). Some individuals (e.g., QCAZ 48096) have a narrow to wide brown stripe on the outer edge of the hands, forearms, thighs and tarsal folds. Coloration of webbing and discs vary from brown to gray. Vomerine odontophores are triangular (with arched base in some specimens). Bones white.
In the adult type series, the largest male has a SVL of 32.24 mm, and the largest female 45.85 mm; mean male SVL = 31.72 mm (n = 5; SD = 0.42), female SVL range is 45.33 to 45.85 mm (n = 2). Inter-population variation in size and other morphometric variables are shown in Tables 1 and 2.
Calls. We recorded the calls of four males at Comunidad Santa Rosa (Provincia Napo) on 28 February 2009, in flooded areas of secondary forest. Call parameters are shown in Table 7. Two call types were recorded. Type one (Fig. 13 E-F) was the most common and consisted of a beep-like note with a mean duration of 0.10 s (SD = 0.02), average rise time 0.03 s (SD = 0.02) and average dominant frequency 1938.47 Hz (SD = 26.24). Type two (Fig 13 G-H) consisted of a single pulsed note with a mean duration of 0.11 s (SD = 0.02), mean rise time 0.05 s (SD = 0.02), and mean dominant frequency 1829.12 Hz (SD = 12.61). Call type two was alternated between calls of type one. Only males QCAZ 40060, 40080-81 produced this type of call.
Distribution and ecology. Hypsiboas tetete is distributed in the Ecuadorian (Provincia Napo) and Peruvian Amazon basin (Región Loreto) (Fig. 15). Known localities range in elevation from 180 m (San Jacinto) to 420 m (Jatun Sacha). It is likely to have a larger distribution. Unfortunately, the lack of distinctive morphological characters relative to H. alfaroi, preclude the unequivocal identification of museum specimens not associated with advertisement calls or genetic data. All specimens from Comunidad Santa Rosa and Jatun Sacha were found in flooded areas, in secondary forest, roosting on vegetation, 50 to 80 cm above ground.
Vegetation type for the Ecuadorian localities is Amazonian Lowland Evergreen Forest characterized by high plant alpha-diversity and a canopy of 30 m with emergent trees that reach 40 m.
Vegetation type at the Peruvian locality is Napo Moist Forest. Conservation status. The distribution polygon has 2,950 km 2 of which 106 (3.5%) have been degraded by human activities (estimated from Ministerio de Ambiente del Ecuador 2013). Because its known distribution range is small with less than five localities and habitat degradation is increasing, H. tetete is assigned to the Endangered category under criteria B1ab(iii).
Two components with eigenvalues > 1.0 were extracted from the PCA for 136 males. The two components accounted for 61.30% of the total variation. The highest loadings for PC I were tibia length, femur length, and foot length; for PC II it was tympanum diameter (Table 8). The morphometric space shows high overlap among species (Fig. 7B). The first PC shows a partial segregation between H. maculateralis (shorts limbs) and H. calcaratus (long limbs). Comparisons of PC I between males of the six species show significant differences between H. fasciatus and H. almendarizae relative to H. calcaratus and H. alfaroi (all P values for t tests < 0.02); H. calcaratus is significantly different from H. maculateralis (t = 4.56, df = 55, P < 0.001), H. alfaroi (t = 6.29, df = 65, P < 0.001), and H. tetete (t = 3.43, df = 20, P = 0.003); H. fasciatus is different from H. maculateralis (t = 2.14, df = 39, P = 0.038); sister species H. alfaroi and H. tetete are also significantly different along PC I (t = -3.35, df = 15, P = 0.004). Comparisons of PC II (mainly tympanum diameter) shows that H. tetete differs from all the others species (all P values for t tests < 0.05); while H. fasciatus and H. almendarizae are different relative to H. calcaratus, H. maculateralis, and H. alfaroi (all P values for t tests < 0.05).

Call comparisons between species
Two components with eigenvalues > 1.0 were extracted from the PCA of calls from 25 males. Both PCs accounted for 72.74% of the total variation. Principal Component I (46.27% of the variance) has high loadings on call duration and fundamental frequency; PC II (26.47% of the variance) has high loadings on number of notes and dominant frequency (Table 9). The acoustic space (as represented by PC I and PC II) shows segregation among the advertisement calls of all species except H. fasciatus and H. almendarizae, which are overlapping (Fig. 20). There are also qualitative differences among the advertisement calls. The closely related H. fasciatus, H. almendarizae, and H. calcaratus share calls consisting of varying numbers of quack-like notes. These calls are markedly different from those of H. maculateralis, H. alfaroi, and H. tetete. Although the calls of H. alfaroi and H. tetete appear close in acoustic space, they have a markedly different structure (Fig. 13C-H). Variation in recording temperature between species was low (< 6 C degrees; Table 7) and cannot explain the marked interspecific structural differences in advertisement calls.   (7), H. maculateralis (1), H. alfaroi (4), and H. tetete (4). See Table 9 for character loadings on each component.

Discussion
The use of genetic characters in the study of the systematics of Amazonian amphibians has resulted in the discovery of an unexpectedly large proportion of undescribed species. Most of those new species were previously considered populations of a single species with a large distribution range. Examples of these complexes include Bolitoglossa peruviana, Engystomops petersi, Osteocephalus buckleyi, Osteocephalus taurinus, Pristimantis ockendeni, Rhinella margaritifera, and Scinax ruber (Elmer et al. 2013;Elmer and Cannatella 2008;Fouquet et al. 2007;Funk et al. 2012;Jungfer et al. 2013;Ron et al. 2012). Hypsiboas calcaratus and Hypsiboas fasciatus show a similar pattern because both were considered to have a widespread distribution comprising most of the Amazon Basin (IUCN 2011). Our results and those from Funk et al. (2012) indicate that they form a complex of eleven candidate species. Available morphological and advertisement call data allowed us to confirm the species status of six of them. Two correspond to H. calcaratus and H. fasciatus and four are new species described herein. The five remaining candidate species remain unconfirmed although geographic distribution suggests that clade J from Buenavista, Bolivia (Fig. 1) could correspond to H. steinbachi (type locality Sara province, Departamento Santa Cruz, Bolivia). Ichilo province (where Buenavista is found, the source of genetic samples for this study) is adjacent to Sara province.

Implications of the discovery of hidden species richness in our current understanding of the biology of Amazonian amphibians
The discovery of hidden species richness could require a review of our current understanding of the biogeography, evolution and conservation status of Amazonian amphibians Angulo 2010, Peloso 2010). This need arises from the anticipated increase in the total number of described species and the availability of more accurate assessments of the geographic distribution of the species. Although the number of genetic-based studies on the systematics of Amazonian amphibians is still limited, our results and those from other species complexes (see above) suggest that the changes could be substantial. In the H. calcaratus-H. fasciatus complex the number of species increased three to five times and there was a sizeable decrease in the distribution range of Hypsiboas fasciatus. According to the distribution polygon of the IUCN Red List (Icochea et al. 2004), H. fasciatus has a distribution of 5'867,000 km 2 . Our results show that the distribution polygon of H. fasciatus, based on the localities shown in Fig.  15, does not overlap with the Red List polygon and has less than 10,000 km 2 , about 0.014% of Red List estimate. If these changes are typical among Amazonian amphibians, the reliability of large global databases, like those of the IUCN Red List, Global Amphibian Assessment (http://www.iucnredlist.org/initiatives/amphibians) could be compromised. The same problem would affect large scale analyses of biogeography, evolution, and conservation of amphibians that have relied those databases (e.g., Buckley and Jetz 2007;Pyron and Wiens 2013). The conservation status of Amazonian amphibians will also require a reassessment under the new taxonomy. The aggregated results for the H. fasciatus species complex show a decrease in the proportion of Least Concern species from 100% to 50%. If this amount of change is representative of Amazonian amphibians, we would expect an increase in the proportion of species under threatened categories.
The use of molecular tools is revolutionizing and reinvigorating the fields of taxonomy and systematics. By combining genetic data with other sets of independent characters, a practice that has been recently referred as "integrative taxonomy" (Dayrat 2005; Padial and De la Riva 2009), species boundaries can be defined objectively. These new tools have allowed documenting large numbers of undescribed Amazonian amphibians, which highlight the need for dense scientific collections (including genome samples) and comprehensive molecular-based taxonomic reviews. The completion of the inventory of species should be considered a priority because other scientific endeavors cannot succeed without reliable taxonomic data.