A new species of Pristimantis (Amphibia, Anura, Craugastoridae) from the foothills of the Andes in Manu National Park, southeastern Peru

Abstract We describe a new species of Pristimantis from the humid sub-montane forest of the Región Cusco in Peru. Pristimantis pluvialis sp. n. was collected in the Kosñipata and Entoro valleys at elevations from 740 to 1110 m a.s.l., near the borders of Manu National Park and within the Huachiperi Haramba Queros Conservation Concession. The new species can be distinguished from other members of the genus Pristimantis by its rostral tubercle, smooth dorsal skin, and by its advertisement call. Pristimantis lacrimosus and Pristimantis waoranii superficially most resemble the new species, but Pristimantis pluvialis sp. n. differs from both species by having a rostral tubercle (absent in Pristimantis waoranii and variable in Pristimantis lacrimosus) and larger size, from Pristimantis lacrimosus by its call emitted at a lower frequency, and from Pristimantis waoranii for its dorsal coloration with dark markings. Two other species have partially overlapping distributions and resemble the new species, Pristimantis mendax and Pristimantis olivaceus, but they produce advertisement calls with much higher dominant frequencies than the advertisement call of the new species. Furthermore, Pristimantis mendax differs from the new species by lacking a rostral tubercle and by having a sigmoid inner tarsal fold, whereas Pristimantis olivaceus differs by being smaller and by having dorsal skin shagreen with scattered tubercles. The new species has snout-vent length of 21.8–26.9 mm in males (n = 12) and 28.8 mm in a single female.


Introduction
The wet tropics are a region of incredibly high biodiversity. The combination of historical, climatic and geographic characteristics foster high speciation rates. In particular, Manu National Park and its surrounding areas have one of the highest herpetofaunal diversity in the world (Catenazzi et al. 2013). Over 155 amphibian species are known from this region, comprising over 2% of known amphibians (Catenazzi et al. 2013). Despite intensive survey efforts, new amphibian species are frequently discovered (Catenazzi et al. 2012;Chaparro et al. 2015;De la Riva et al. 2012).
Manu NP is particularly rich in members of the genus Pristimantis (Craugastoridae), as are other regions in the upper Amazon Basin and the eastern slopes of the Andes. This is one of the largest genera of all vertebrates, and is incredibly understudied. It contains nearly 500 species distributed throughout the New World (AmphibiaWeb 2016; Hedges et al. 2008). Pristimantis and most members of the Craugastoridae are primarily terrestrial and are generally assumed to be direct-developing, lacking an aquatic tadpole stage (Duellman and Lehr 2009).
A relatively recent divergence and morphological similarities among species may indicate remarkable cryptic diversity within Pristimantis (Ortega-Andrade et al. 2015). It can be difficult to discern new species without genetic information, particularly in Peru where this genus is diverse (Aguilar et al. 2010) and counts up to 125 species (AmphibiaWeb 2016). However, surveys sometimes reveal species with unique morphological traits, such is the case of a new, relatively large Pristimantis species bearing a rostral tubercle, related to P. lacrimosus, that we discovered during surveys in the Kosñipata Valley near Manu NP and within the Huachiperi Haramba Queros Conservation Concession. Here we describe this new species.

Methods
The format of the diagnosis, measurements and description follows Duellman and Lehr (2009). Taxonomy follows Hedges et al. (2008), except that we followed Pyron and Wiens (2011) for family placement. Specimens were fixed in 10% formalin and preserved in 70% ethanol. Sex and maturity of specimens were determined by observing sexual characters and gonads through dissections. We measured the following variables to the nearest 0.1 mm with digital calipers under a stereomicroscope: snout-vent length (SVL), tibia length (TL), foot length (FL, distance from proximal margin of inner metatarsal tubercle to tip of Toe IV), head length (HL, from angle of jaw to tip of snout), head width (HW, at level of angle of jaw), eye diameter (ED), tympanum diameter (TY), interorbital distance (IOD), upper eyelid width (EW), internarial distance (IND), and eye-nostril distance (E-N, straight line distance between anterior corner of orbit and posterior margin of external nares). Fingers and toes are numbered preaxially to postaxially from I-IV and I-V respectively. We compared the lengths of toes III and V by adpressing both toes against Toe IV; lengths of fingers I and II were determined by adpressing the fingers against each other. Photographs taken by A. Shepack and A. Catenazzi in the field were used for descriptions of coloration in life, and have been deposited at the Calphotos online database (http://calphotos.berkeley.edu).
We used two sets of recordings to describe the advertisement call. Holotype, COR-BIDI 16510 (SVL 24.6 mm; recording #9846 deposited at the Fonoteca Zoológica, Museo Nacional de Ciencias Naturales, Madrid, www.fonozoo.com) was recorded at the type locality at 20:30 on 16 January 2015 (T air = 21.4 °C), along with several unvouchered males. We used a digital recorder (Zoom H2; WAV format, 44 KHz, 24 bit) to record these advertisement calls in 2015. In a second set of recordings, we recorded paratype MUSM 35217 at 21:00 on 2 September 1999 (SVL =22.5 mm; T air = 20.2 °C; recording #9847 deposited at the Fonoteca Zoológica, Museo Nacional de Ciencias Naturales, Madrid), along with several uncaptured males, using a portable cassette recorder (Aiwa HS-F150), a small microphone, and audiotapes.
We used Raven, version 1.4 (Cornell Laboratory of Ornithology) to analyze call length, peak frequency, and calling rate. The Hamming window function for the spectrogram was set at 256 bands. We report means ± SD. We analyzed a total of 380 calls.
We determined the phylogenetic position of the new species with respect to other Pristimantis species through analysis of DNA sequence data. Our analysis included the 16S rRNA mitochondrial fragment and the protein-coding gene cytochrome c oxidase subunit I (COI). We used tissue samples from specimens collected in southern Peru (Cusco and Madre de Dios regions) to obtain DNA sequences for the new species and another undescribed Pristimantis from the foothills of Manu National Park (Appendix I). Additionally, we downloaded sequences from Genbank (Appendix I) of morphologically similar species (P. bromeliaceus, P. galdi, P. mendax, P. mindo, P. moro, P. omeviridis, P. schultei, P. subsigillatus) in the putative P. lacrimosus group (sensu (Arteaga et al. 2013); but see (Padial et al. 2014). Sequences for several other species of Pristimantis that possess a rostral tubercle were not available in Genbank. We included the distantly related Pristimantis ridens as outgroup (Padial et al. 2014).
We used Geneious R6, v. 6.1.8 (Biomatters, http://www.geneious.com/) to align the sequences with MAFFT, v. 7.017 (Katoh and Standley 2013) alignment program. Prior conducting phylogenetic analysis, we used PartitionFinder, v. 1.1.1 (Lanfear et al. 2012) to select the appropriate models of nucleotide evolution and determined the best partitioning scheme and substitution model for each gene with a Bayesian information criterion (BIC) to. We employed a Maximum Likelihood (ML) approach using RaxML, v. 8.2.4 (Stamatakis 2006) to infer a molecular phylogeny. We used the "f-a" function to conduct a bootstrap analysis and search for the optimal likelihood tree. Our analysis included 20 terminals and a 1064 bp alignment for the concatenated dataset. We used the GTR + Γ model of nucleotide substitution, performed 200 tree searches, and assessed node support using 1000 bootstrap replicates. Additionally, we used the R package APE (Paradis et al. 2004) to estimate uncorrected p-distances (i.e., the proportion of nucleotide sites at which any two sequences are different).
We quantified infection by Batrachochytrium dendrobatidis (Bd) by swabbing frogs with a synthetic dry swab (Medical Wire & Equipment, #113) using a standardized swabbing protocol. Swabs were stroked across the skin a total of 30 times: 5 strokes on each side of the abdominal midline, 5 strokes on the inner thighs of each hind leg, and 5 strokes on the foot webbing of each hind leg (total of 30 strokes/frog). We used a real-time Polymerase Chain Reaction (PCR) assay on material collected on swabs to quantify the level of infection (Boyle et al. 2004). After extraction using PrepMan Ultra, we analyzed DNA amplification in a Life Technologies StepOne Plus qPCR instrument following the protocol outlined in Hyatt et al. (2007) and Boyle et al. (2004), except that extracts were analyzed once. We calculated ZE, the genomic equivalent for Bd zoospores by comparing the qPCR results to a set of standards, and considered any sample with ZE > 1 to be infected or Bd-positive.
Specimens Generic placement. We assign this species to Pristimantis on the basis of general morphological similarity to other members of the genus and molecular data. The genus Pristimantis lacks any diagnostic morphological synapomorphies (Hedges et al. 2008), but molecular phylogenetic analyses support the placement of the new species within the genus (Fig. 6).
The two species that superficially most resemble P. pluvialis are P. lacrimosus and P. waoranii. However, P. pluvialis differs from both species by having a rostral tubercle (absent in P. waoranii and variable in P. lacrimosus), and by its larger size. Furthermore, it differs from P. lacrimosus by its call with lower dominant frequency (~2500 Hz). Calls of P. lacrimosus available at AmphibiaWeb Ecuador (Read 2012; Ron et al. 2016) have higher dominant frequency ranging from 3100-3273 Hz (n = 6). Furthermore, the     new species differs from P. waoranii by having dark bands or markings on the dorsum (absent in P. waoranii). Another morphologically similar species, P. schultei, has an acuminate snout in dorsal view (subacuminate in P. pluvialis), skin on dorsum shagreen (generally smooth), and heel and outer edge of tarsus bearing many low tubercles (tubercles absent). Furthermore, P. schultei occurs in northern Peru at elevations above 2400 m (below 1110 m for P. pluvialis), and its advertisement call consists of a double note, "ping-ping" (Duellman 1990), in contrast with the single, low frequency note produced by P. pluvialis. Two species related to P. lacrimosus, P. mendax and P. olivaceus, occur near the type locality of P. pluvialis in Manu NP and surrounding areas in southern Peru (Catenazzi et al. 2013;Duellman 1978b;Duellman and Lehr 2009;Köhler et al. 1998). In addition to the characters listed in the previous paragraph, P. olivaceus further differs from P. pluvialis by being smaller (17.7-22.1 mm in males, and up to 25.5 mm in females; Duellman and Lehr 2009) and by having dorsal skin shagreen with scattered tubercles and dorsal coloration brownish green or olive green. Both species produce advertisement calls with higher dominant frequencies (4000-5320 Hz; see FonoZoo recording #875 for P. mendax, and Köhler et al. 1998 andMárquez et al. 2002 for P. olivaceus) than the advertisement call of the new species (~2500 Hz). Pristimantis mendax further differs from P. pluvialis by lacking a rostral tubercle, by possessing a sigmoid inner tarsal fold and by having dorsal skin shagreen with scattered spicules. Furthermore, P. mendax occurs in montane cloud forests above 1400 m (Duellman and Lehr 2009), an elevational distribution range that does not appear to overlap with that of P. pluvialis.
Description of holotype. Adult male (24.6 mm SVL); head narrower than body, its length 36.3% of SVL; head slightly longer than wide; head width 33.6% of SVL; snout short, squared in dorsal view, subtruncate in lateral view (Fig. 2); eye large, 33.9% of head length, its diameter 0.97 times its distance from the nostril; nostrils slightly protuberant, situated close to snout; canthus rostralis weakly concave in dorsal view, rounded in profile; loreal area flat; lips rounded; dorsal surface of head smooth and upper eyelids with minute tubercles; upper eyelid width 65.7% of interorbital distance; supratympanic fold absent; tympanic membrane not differentiated, tympanic annulus visible; postrictal ridges or tubercles absent. Choanae round, very small, positioned far anterior and laterally, widely separated from each other, not concealed by palatal shelf of maxilla; dentigerous processes of vomer and vomerine teeth barely noticeable.
Coloration of holotype in life. Dorsum orange-brown with faint brown markings (Fig. 2). Interorbital bar dark brown, forming a triangular shape posteriorly; canthus rostralis dark brown; light green on upper eyelids. Triangular brown patterning on back, not extending to venter. Hind legs with broad brown barring. Forelimbs with faint brown barring. Throat yellowish-cream; venter cream.

Coloration of holotype in alcohol.
Dorsal surfaces of head, body, and limbs grayish-tan, with dark brown regions around scapulae (see Fig. 2). Interorbital as a dark blotch that extends posteriorly; canthus rostralis dark brown. Dorsal surfaces of hind limbs with dark flecking. Iris dark gray. Throat pale white; chest and belly pale white to cream; ventral surfaces of thighs the same color as chest and belly; plantar and palmar surfaces and tips of digits pale, tubercles darker gray.
Variation. Coloration in life is based on field notes and photographs taken by A. Shepack and A. Catenazzi of 13 type specimens. The dorsum is beige to reddish-brown with or without faint dark-brown markings (Figs 3-4). A dark brown interorbital bar is present in most specimens (barely visible in some individuals). The iris is bronze with dark-brown to red reticulations. Some individuals possess faint brown barring on hind legs. The throat is cream to yellowish-white while the belly is predominantly cream to white. Dorsal skin is generally smooth, but CORBIDI 16695, 17014 (Fig. 5), and MHNG 2607.13, have minute, scattered tubercles, indicating that skin texture might be a variable trait. Some individuals have small tubercles on the outer edge of tarsus.
Vocalization and reproduction. Males call from grasses, shrubs, and trees in the understory of the submontane forest, during crepuscular hours and at night, conspicuously after heavy rains. Holotype CORBIDI 16510 was calling from a broad fern leaf at 150 cm above the ground, along a trail at ~30 m from a stream (T air = 21.4°C). The advertisement call consisted of a note 28.7 ± 0.7 ms (range 23.0-35.0 ms, n = 20) in duration (Fig. 7). Pulses emitted at the highest amplitude had dominant frequencies of 2412-2584 Hz (average 2489 ± 20 Hz, n = 20) and were located in the first half of the note (Fig. 7). The calling rate was 0.70 calls/second at a temperature of 21.4°C. Male MUSM 35217 was perched on a shrub at 2 m, near a stream (T air = 20.2°C), and produced single note calls 36.0 ± 0.5 ms (range 24.0-58.0 ms, n = 102) in duration, with dominant frequencies of 2067-2756 Hz (average 2407 ± 19 Hz, n = 102), at a calling rate of 0.64 calls/s. At least three unvouchered males were recorded at the type locality near the holotype (T air = 21.4°C). Their calls were 40.7 ± 0.2 ms (range 26.0-47.0 ms, n = 220) in duration, with dominant frequencies of 2067-2756 Hz (average 2586 ± 20 Hz, n = 220); call rate could not be determined. Similarly, several unvouchered males recorded near MUSM 35217 (T air = 20.2°C) produced calls with dominant frequencies of 2067-2584 Hz (average 2312 ± 13 Hz, n = 104), but their note durations and call rates could not be determined. Overall, the call of the new species can be described as a single "tock" note, 23-58 ms in duration, emitted at a rate of 0.64-0.70 calls/s, and with peak frequency ranging from 2312-2756 Hz. The call has amplitude and frequency modulation (Fig. 7): a short, high energy pulse with frequency decreasing from dominant frequency (~2500 Hz, see above) to ~2000 Hz is followed by a low energy pulse with frequency increasing from ~2000 Hz to ~2500 Hz.
Etymology. The name of the new species is a Latin word and refers to the high rainfall recorded at the type locality, which averages ~6 m annually, and represents the peak rainfall amount along the elevational transition from the Amazon lowlands to the Andean peaks. Furthermore, males of P. pluvialis typically call during or immediately after heavy rains.  Distribution, natural history, and threats. The new species was found during surveys in the humid sub-montane forests of the Kosñipata and Entoro valleys (Fig.  8). Observers made intensive visual searches of vegetation and leaf litter during evenings (18h30-0h00). Individuals were found after rains, calling on vegetation up to 2 m above the ground. Male CORBIDI 16511 and female CORBIDI 16512 were captured while in amplexus, during one such choruses on 16 January 2015. Average individual mass was 0.97 g ± 0.06 for males (n = 12) and 2.1 g for one female. The oviducts of this female contained 22 unpigmented eggs, about 2.5 mm in diameter. Additionally, four out of ten individuals (MHNC 15490 and CORBIDI 11862, 16512 and 16695) tested positive for the amphibian chytrid fungus Batrachochytrium dendrobatidis. This fungus has been implicated in population declines of numerous other species in this region, although it is unknown what effect it has had on this species, and there is no evidence for declines in the populations of P. pluvialis (Catenazzi et al. 2011;). In addition to being found near, and likely within, Manu NP, P. pluvialis has been found within the Huachiperi Haramba Queros Conservation Concession, a protected area consisting of state-owned lands given in concession to private organizations with the goal of preserving biodiversity. The Huachiperi Haramba Queros concession, legally recognized in 2006 was the first concession to be granted to an indigenous community in Peru. Sympatric frog species at the type locality include Cochranella nola, Hypsiboas gladiator, Osteocephalus mimeticus, Pristimantis platydactylus, P. reichlei, and Rulyrana spiculata. Other species found around the type locality are Allobates alessandroi, Ameerega simulans, Dendropsophus parviceps, Hyalinobatrachium bergeri, Hyloscirtus phyllognathus, Hypsiboas lanciformis, Noblella sp., Oreobates granulosus, Pristimantis danae, P. fenestratus, P. mendax, P. toftae, Ranitomeya sirensis, Rhinella leptoscelis, R. margaritifera, R. tacana, Rulyrana spiculata and Scinax ruber.

Discussion
Our phylogenetic analysis indicates that P. pluvialis is most closely related to two undescribed Pristimantis species, one from the same valley near the type locality of P. pluvialis, and one from Guyana ( Fig. 6; Padial et al. 2014). The species from Guyana (Pristimantis sp. ROM 43978) had previously been identified as P. zeuctotylus (Hedges et al. 2008). Furthermore, four other species, P. omeviridis, P. galdi, P. mindo, and P. subsigillatus, are closely related to P. pluvialis (Fig. 6). The most comprehensive molecular phylogenetic study of terraranas to date also found that most of these taxa form one clade (Padial et al. 2014). Nevertheless, given that there was low resolution in some nodes, both in the study by Padial et al. (2014) and this study, analyses including additional genes are needed to further examine the relationships among species in this group.
The morphologically similar P. lacrimosus presumably is a complex of at least two species, formed by lowland populations of small-sized individuals around the type locality in Ecuador, and larger individuals from the Andean slopes (Duellman 1978a;Lynch and Duellman 1980), which possibly form a distinct and still unrecognized species. Pristimantis pluvialis is larger, calls at lower frequency, and occurs at higher elevations than P. lacrimosus. Furthermore, the locality of the neotype of P. lacrimosus, collected at Santa Cecilia (Heyer and Hardy 1991), is 1600 km NW of the type locality of P. pluvialis, a very long distance considering the degree of endemism in montane and sub-montane Pristimantis. We lack sequences for P. waoranii, but the absence of rostral tubercle and dark dorsal markings in this species differentiates it from all individuals of P. pluvialis in our type series.
Pristimantis pluvialis is morphologically similar to P. olivaceus, which has previously been reported from Manu NP (Duellman and Lehr 2009) and southeastern Peru (Köhler et al. 1998). In light of our present description, a closer examination of existing museum specimens, along with DNA sequence data and call recordings for new observations of putative P. olivaceus from southeastern Peru, are needed to confirm assignment of museum specimens to this species, and to properly identify newly collected specimens. Of the characters we have mentioned in the diagnosis and comparisons, skin texture should be used with caution, because previous work shows that this trait is variable in some Pristimantis (Guayasamin et al. 2015), and because three of our types, despite not displaying the shagreen skin texture seen in P. olivaceus, differ from other smooth-skin types by having minute, scattered tubercles on dorsal surfaces.
Pristimantis pluvialis and P. olivaceus are most easily distinguished by their overall coloration, which is reddish-brown in P. pluvialis and green in P. olivaceus, by the larger size of P. pluvialis (SVL of males <22 mm in P. olivaceus, and >22 mm in P. pluvialis), and most notably by differences in their advertisement calls. Both species produce calls consisting of short, single notes, but whereas P. pluvialis emits "tocks" with a dominant frequency at ~2500 Hz during an initial burst followed by downward frequency modulation in the middle of the call, P. olivaceus emits "chirps" with a dominant frequency at ~4900 Hz with increasing frequency during the call but little amplitude modulation (Köhler et al. 1998).
With the addition of P. pluvialis the genus Pristimantis now contains 474 known species (AmphibiaWeb 2016). This cryptically diverse group surely contains even more undescribed species (Ortega-Andrade et al. 2015). We discovered P. pluvialis in a region where multiple researchers, including the authors, have worked previously. This suggests that continued surveying efforts are necessary to achieve a full understanding of herpetological diversity in this area (Catenazzi et al. 2013;Catenazzi and von May 2014).