A new species of wood lizard (Hoplocercinae, Enyalioides) from the Río Huallaga Basin in Central Peru

We report the discovery of a new species of Enyalioides from the premontane forest of the Río Huallaga basin in central Peru. The most similar and phylogenetically related species are E. binzayedi and E. rudolfarndti. However, the new species differs from E. binzayedi (state of character in parentheses) by having dorsal scales strongly keeled on paravertebral region and feebly keeled or smooth elsewhere (prominent medial keel on each dorsal scale), more dorsals in transverse row between dorsolateral crests at midbody 26–39, x̄ = 30.44 (22–31, x̄ = 27.57), and a conspicuous posteromedial black patch in the gular region of males (absent). Contrarily, adult males of the new species and E. rudolfarndti are readily distinguished by having a conspicuous posteromedial black patch in the gular region, absent in E. rudolfarndti, and by lacking a conspicuous orange blotch (faint if present) on the antehumeral region, as in E. rudolfarndti. We also present an updated molecular phylogenetic tree of hoplocercines, which strongly supports both referral of the newly discovered species to Enyalioides and its status as distinct from those recognized previously.


Introduction
The Huallaga drainage flows north from the Cordillera Oriental of central Peru and has an extension of 1,138 km (SNE 2001), which makes it the largest tributary of the Marañon river. This basin harbours a great variety of ecosystems, including the Yugas, which are considered a shelter of endemic birds (Parker and O'neil 1976), mammals (Pacheco 2002) and amphibians (Lehr et al. 2002;Duellman et al. 2004). However, our knowledge on their biodiversity remains poor because the biological assessments in this area were disrupted in the eighties due to an internal war against terrorist legions and drug traffickers. Additionally, habitat destruction by coca plantations during the eighties and nineties caused dramatic forest fragmentation below 2,000 m (ONUDD 2006). Thus, it is not surprising that herpetological surveys along the Huallaga river basin in recent years have resulted in the discovery of eight new species of anurans including two species of Gastrotheca (Duellman et al. 2004), one of Nymphargus (Guayasamin et al. 2006), four of Phrynopus (Lehr et al. 2002;Lehr and Oroz 2012), and one of Pristimantis (Chávez and Catenazzi 2016), as well as five new species of lizards including one of Euspondylus (Chávez et al. 2017), one Petracola (Köhler and Lehr 2004), two Proctoporus (Köhler and Lehr 2004), and one Stenocercus (Venegas et al. 2013a).
In an effort to improve our knowledge on the herpetofauna of the Huallaga river basin, we carried out field surveys between 2011 and 2018, which led to the discovery of a new species of wood lizard described in this paper. Among hoplocercine lizards, wood lizards (Enyalioides) represent one of the South Ameri can lizard groups with the highest species discovery rate (corrected for clade size) of this century. Our discovery increases to 16 the number of known species of Enyalioides, of which only seven were known in previous centuries.

Field techniques
Most lizards were captured by hand during night. Coordinates and elevation were taken with a GPS (Garmin, WGS84). After the specimens were euthanized with T61 (Intervet), we took muscle tissue samples from the left thigh, and stored them in cryogenic vials (Axygen) at 96% ethanol. All collected specimens were fixed in 10% formalin for 48 hours and permanently stored in 70% ethanol. Voucher specimens were deposited at the herpetological collection of the Centro de Ornitología y Biodiversidad (CORBIDI) in Lima, Peru. Specimens were collected with the following permits:

Morphological data
Snout-vent length (SVL) and tail length (TL) were measured with a ruler and recorded to the nearest millimeter. All other measurements (i.e., head width, length and height; rostral and mental width and height) were made with digital calipers and recorded to the nearest 0.1 mm. Sex was determined by noting the presence of hemipenes or sexually dichromatic characters. We followed the format of Torres-Carvajal et al. (2011)

Phylogenetic analysis
Editing, assembly, and alignment of sequences were performed with Geneious v9.1.8 (Kearse et al. 2012) under default settings for MAFFT (Katoh and Standley 2013). Genes were combined into a single dataset with four partitions, three corresponding to each codon position in protein coding genes and one to all tRNAs. The best partition strategy along with the corresponding models of evolution were obtained in PartitionFinder v2 (Guindon et al. 2010;Lanfear et al. 2017) under both the Bayesian Information Criterion (BIC) and the "greedy" algorithm (Lanfear et al. 2012) with branch lengths of alternative partitions "linked" to search for the best-fit scheme.
Phylogenetic relationships were assessed under a Bayesian inference approach us ing MrBayes 3.2.7 (Ronquist et al. 2012) after partitioning the data as described above. To reduce the chance of converging on a local optimum, four runs were performed. Each consisted of ten million generations and four Markov chains with default heat ing values. Trees were sampled every 10,000 generations. Stationarity was confirmed by plotting the -ln L per generation in the program Tracer 1.7 (Rambaut et al. 2018). Additionally, the standard deviation of the partition frequencies and the potential scale reduction factor (Gelman and Rubin 1992) were used as convergence diagnostics for the posterior probabilities of biparti tions and branch lengths, respectively. Adequacy of mixing was assessed by examining the effective sample sizes (ESS) in Tracer, with ESS > 200 considered as satisfactory. After analyzing convergence, mixing, and sampling, the first 101 trees in the sample were discarded as "burnin" from each run. We used the resultant 3,600 trees to calculate posterior probabilities (PP) for each bipartition on a Maximum Clade Credibility Tree in TreeAnnotator (Rambaut and Drummond 2016). Additionally, uncorrected genetic distances among species of Enyalioides and Morunasaurus were calculated in DIVEIN (Deng et al. 2010) for ND2 and are presented as an Excel file in the Suppl. material 2: Table S1.

Results
The taxonomic conclusions of this study are based on the observation of morphologi cal features and color patterns, as well as the inferred phylogenetic relationships. This information is considered as species delimitation criteria following a general lineage or unified species concept (de Queiroz 1998;2007  Diagnosis. Enyalioides feiruzae can be distinguished from other species of Enyalioides, except E. binzayedi and E. rudolfarndti, by the combination of the following characters (see also Table 1): (1) scales posterior to the superciliaries enlarged and pointed (relative to adjacent scales), forming a well-defined longitudinal row of distinctly raised scales across the lateral edge of the head in juveniles and adults of both sexes; (2) vertebral crest strongly projected, continuous, and decreasing in size posteriorly, with vertebrals on neck at least four times higher than those between hind limbs; (3) projecting scales on body or limbs absent; (4) ventral scales strongly keeled; (5) caudal scales heterogeneous in size on each autotomic segment; and (6) superciliary triangular flap that projects posterolaterally over each eye absent. The most similar species to Enyalioides feiruzae is its sister species (see below) E. rudolfarndti (Fig. 4B). Both species are similar in dorsal coloration, scale counts, and size; however, E. feiruzae differs from E. rudolfarndti (character states in parentheses) in having a conspicuous posteromedial black patch, rarely dark brown, in the gular region in both sexes (absent); adult males usually without an orange round blotch on the antehumeral region or a faint blotch if present (conspicuous orange blotch on antehumeral region in adult males); scales on flanks homogeneous in size (distinctly heterogeneous); and fewer supralabials 8-10, x = 9.52 (11-12, x = 11.75). The new species differs from E. binzayedi in having dorsal scales strongly keeled on paravertebral region and feebly keeled or smooth elsewhere (prominent medial keel on each dorsal scale); fewer supralabials 8-10, x = 9.52 (11-15, x = 12); more dorsals in transverse row between dorsolateral crests at midbody 26-39, x = 30.44 (22-31, x = 27.57); and a conspicuous posteromedial black patch in the gular region (absent).
A longitudinal row of strongly projecting scales along the lateral edge of the skull posterior and continuous with the superciliaries is also present in Enyalioides altotambo and E. oshaughnessyi, which occur west of the Andes in Ecuador and Colombia and differ from E. feiruzae in having smooth or slightly keeled dorsals. Additionally, Enyalioides cofanorum and E. microlepis share with the new species the presence of strongly keeled ventral scales and caudal scales of heterogeneous size on each autotomic segment. However, E. cofanorum differs from E. feiruzae (characters states in parentheses) in having a smaller body size, maximum SVL 107 mm in males and 109 mm in females (maximum SVL = 122 mm in males and 119 mm in females); scattered, projecting, large dorsal scales (absent); and a black patch covering the gular region (patch only covering the gular region posteromedially). Enyalioides microlepis differs from E. feiruzae in having more than 40 dorsal scales in a transverse line between the dorsolateral crest at midbody (38 or fewer dorsal scales); a low vertebral crest (high); and a blue gular background (white or orange).
Description of holotype. Male (Fig. 1); SVL = 121 mm; TL = 174 mm; maximum head width = 22.9 mm; head length = 21.5 mm; head height = 22.1 mm; dorsal head scales uni-or multicarinate, projected dorsally; parietal eye present; five scales immediately posterior to superciliaries conical, dorsolaterally projected, and conspicuously larger than adjacent scales; temporal scales small, pyramidal, low; one enlarged, conical pretympanic scale; superciliaries 14; canthals three; postrostrals two; supralabials nine if counted to a point below middle of eye; rostral not divided, similar in size to adjacent supralabials; lorilabials in one longitudinal row between suboculars and supralabials at level of middle of eye, and two rows anterior to this point; loreal region with small, weakly keeled, and juxtaposed scales; nasal at level of supralabial III; infralabials nine if counted to a point below middle of eye; mental (2.48 mm wide × 1.51 mm high) wider and 1.5 times higher than adjacent infralabials; postmentals two; gulars ventrally projected, compressed laterally and  separated from each other by black skin covered with tiny granular scales; gular fold complete midventrally, extending dorsally and posteriorly to form an antehumeral fold; neck with some oblique folds, and a dorsolateral row of enlarged scales; ventral end of oblique fold immediately anterior to antehumeral fold with approximately five enlarged scales similar in size to adjacent gulars, but more than three or four times the size of adjacent fold scales. Vertebral crest strongly projected and decreasing in size posteriorly, with vertebrals on neck at least four times higher than those between hind limbs; crest bifurcates at a point approximately 15 mm posterior to the cloaca, and extends onto tail about 1/3 its length; body flanks between fore and hind limbs with slight dorsolateral and ventrolateral folds, as well as oblique folds; scales on dorsolateral folds strongly keeled, slightly larger than adjacent scales; dorsal scales in paravertebral region enlarged, strongly keeled, and mucronate, decreasing in size as they approach the dorsolateral fold; scales on flanks slightly smaller than dorsals, keeled or smooth, nearly homogeneous in size except for some scattered enlarged keeled scales near both the ventrolateral region and the groins; ventral scales imbricate, strongly keeled, rectangular or rhomboid, with a posteromedial mucron; ventrals more than twice the size of dorsals.
Limb scales strongly keeled dorsally and keeled or feebly keeled ventrally; scales on dorsal aspect of thighs more or less homogeneous in size; scales on posterior aspect of thighs strongly heterogeneous in size, with most scales less than half the size of those scales on anterior and ventral aspects; subdigitals on finger IV 21; subdigitals on toe IV 27; one femoral pore on each side; tail laterally compressed and gradually tapering posteriorly; caudal scales keeled at the base of tail, becoming strongly keeled and imbricate towards tip, gradually increasing in size posteriorly on lateral and dorsal aspects of each  caudal segment; caudals larger ventrally than dorsally; individual caudal segments three scales long ventrally and five scales long dorsally.
Color in life of holotype (Fig. 1): Dorsal surface of body, limbs and tail turquoise; dorsal surface of head with a greenish tone and scattered dark brown flecks; side of head orange on nasal, loreal, ocular, labial and mandibular regions; orange reticulation in temporal region and dorsal surface of neck; sides of neck with reddish brown reticulation and a faint postympanic white stripe; vertebral crest orange on the neck and turquoise with a reddish-brown base elsewhere; dorsum with a dorsolateral row of faint pale blotches and covered by reddish brown reticulations that are more defined on the flanks; limbs and tail with a greenish tone; limbs with dark brown or nearly black marks and flecks; chin white; throat orange with a black patch posteriorly; chest, venter, ventral surface of limbs and tail white with faint marks on ventral surface of thighs; iris reddish brown, darker in the center than peripherally, with a golden ring bordering the pupil.
Color in ethanol 70%: Dorsal surface of body, head, limbs and tail grayish turquoise; dorsal surface of head and sides of head with a dusty brown tone and dark brown flecks; labials gray; dorsal surface of neck with a dusty brown tone, sides of neck with dark reticulations and a faint white postympanic stripe; vertebral scales on neck with dusty brown tips; dorsum with dark reticulations; tail with a dusty brown tone; dorsal surface of limbs with dark flecks and marks, more abundant on hind limbs; chin and throat cream, covered by a black patch on the posteromedial gular region and ventral surface of neck; chest, belly, ventral surface of limbs and tail dirty cream with faint marks on ventral surface of thighs.
Intraspecific variation. Meristic and morphometric characters of Enyalioides feiruzae are summarized in Table 1. Scales on flanks are keeled, nearly homogeneous in size, slightly smaller or similar to dorsals in all specimens, but some individuals have a few scattered, enlarged scales near the insertion of hind limbs or along the ventrolateral region. Caudal segments are 5-6 scales long laterally. Color of the iris is nearly identical in all specimens, varying only from brown to reddish brown, darker in the center, with a golden ring bordering the pupil. Enyalioides feiruzae is a polymorphic species with more variation in color in males and juveniles than in adult females (Figs 2, 3). The dorsal background of adult males can be brownish turquoise (CORBIDI 14916) (Fig. 4A), gray or greenish brown with pale (CORBIDI 13366) or black (CORBIDI 9314; Fig.  2E) transverse bands along dorsum. The orange head coloration of the holotype is restricted to the labials and throat of other adult males. An adult male specimen (CORBIDI 14916) has two inconspicuous pale orange blotches, one on each side of the ventrolateral region of neck (Fig. 4A). One subadult male (CORBIDI 15469, SVL = 98.8 mm) has the dorsal background light green sprinkled with sky blue flecks; short pale transverse bands along dorsum and black reticulations on dorsum and flanks; dorsal surface of limbs with black flecks and transverse bands; tail with dark brown or black marks; chin gray and throat grayish brown, ventral surface of neck up to the neck fold dark; chest, venter, ventral surface of limbs and tail tan white with dark brown flecks on belly, thighs and tail ( Fig. 2A, B). Ventrally, specimen CORBIDI 9314 (Fig. 2F) has a white throat with black flecks, and a black patch posteromedially; the chest, venter, ventral surface of limbs and tail are white with a large grayish patch along the belly and black flecks on the ventral surface of thighs and base of tail.
Among adult females (Fig. 2), the dorsal background is greenish brown (CORBIDI 14933) or floury brown (CORBIDI 9317; Fig. 2C), with faint dark brown (COR-BIDI 14933), transverse bars (dorsum, limbs, and tail) or blotches (flanks); some specimens have dark subocular (broad) and supratemporal (narrow) stripes, whereas others have a faint postympanic stripe. Specimen CORBIDI 14933 has a distinct pale subocular stripe and a white blotch on rictal region. Specimen CORBIDI 14933 has a cinnamon brown venter without a dark patch, while specimen CORBIDI 9317 has a brown throat with dark brown flecks and a dark brown patch posteromedially (Fig. 2D), and a tan venter with scattered dark brown flecks on belly and dark brown marks on ventral surface of thighs.
Dorsal background in juveniles varies among different tones of brown and green (Fig. 3), with a paravertebral row of either dark brown blotches with pale interspaces (CORBIDI 9315, 9319 and 15470) or pale blotches (CORBIDI 9318). Flanks can have small pale blotches and black flecks (CORBIDI 9318), transverse rows of dark brown (nearly black) blotches and dark brown flecks (CORBIDI 15506), or faint transverse cream bands and dark brown marks (CORBIDI 9319). All juvenile specimens have dark marks or flecks on head, dark supratemporal stripe and postocular blotch, and postympanic pale stripe. Ventrally, CORBIDI 9315 is immaculate tan without a dark patch on gular region (Fig. 3B). Other specimens are grayish white, dirty white or white ventrally (CORBIDI 9318-19, 15470), with dark or faint brown marks and flecks on throat, belly and ventral surface of thighs and a dark brown or gray patch on the ventral surface of neck (Fig. 3D, F & H). Sexual dichromatism is not obvious in the dorsal background coloration due to the high variation presented in both sexes. The turquoise hue in the dorsum of some adult males is absent in the examined females. However, the orange coloration in the head and throat of some adult males, and the black patch on the posterior half of throat is absent in females.
Distribution and natural history observations. Enyalioides feiruzae is known from the Huallaga River basin (Huánuco Department) on the Amazonian slope of the Andes of central Peru at elevations between 830-1,614 m a. s. l. Specimens from Pachitea and Huánuco provinces were collected in patches of secondary forest surrounded by pasturelands and coffee, corn or fruit plantations, whereas specimens from Tingo María National Park in Leoncio Prado province were found in a steep primary forest area.
The general landscape along the Huallaga River basin between the Huánuco and Pachitea provinces consists of steep areas, which are impacted by human activities like agriculture and cattle ranching, with scattered patches of secondary forest and narrow fringes of riverine forest. In contrast, Tingo María National Park is an area of primary forest, where the ground is covered by leaf litter, fallen trunks and rocks. Streams with small, torrential waterfalls are common. Vegetation includes lichens, ferns, orchids, epiphytes, bushes, and trees. All localities where Enyalioides feiruzae was collected lie within Peruvian Yungas (Olson et al. 2001).
Most individuals reported here were collected at night sleeping on vertical and horizontal stems of bushes 20-150 cm above the ground. One juvenile was found sleeping in the middle of a big leaf 30 cm above the ground. Another juvenile was found sleeping in a vertical stem, and when disturbed by the light of the headlamp, it dropped and ran to hide in a burrow under a big rock. Male COR-BIDI 14916 was found at night sleeping on branches in a small patch of secondary forest surrounded by crops. Adults from Tingo María National Park were caught sleeping vertically on stems near streams and waterfalls; male CORBIDI 15516 was found 1.5 meters away from female CORBIDI 15517, both sleeping vertically.
Etymology. The specific name is a noun in the genitive case and is a patronym for Feiruz, a female green iguana, muse and lifelong friend, owned by Catherine Thomson from Farmington Hills, Michigan, USA, who supported our taxonomic research and nature conservation through the BIOPAT initiative.

Discussion
The phylogenetic tree presented here supports the position of the new species within the clade (E. azulae (E. palpebralis (E. sophiarothschildae (E. binzayedi (E. rudolfarndti, E. feiruzae sp. nov.))))), herein named "palpebralis group", with maximal support (Fig. 6). Species in this clade occur mostly along the Amazonian slopes of the Andes in Peru and, except for E. palpebralis, were discovered in the last decade (Venegas et al. 2011;Venegas et al. 2013b;Torres-Carvajal et al. 2015). Curiously, the lining of the throat in all species of this clade is black or dark purple, which in some cases extends to the floor of the mouth. Moreover, the lining of the mouth in some adult males of E. palpebralis is bright orange (Torres-Carvajal et al. 2011). Dark pigment on lining of throat or mouth is absent in E. anisolepis, E. heterolepis, E. laticeps, E. microlepis, E. oshaughnessyi, E. praestabilis, E. rubrigularis and E. touzeti; this condition has not been examined in other species of hoplocercines. This suggests that presence of dark pigment in the buccal mucosa is a synapomorphy of the "palpebralis group".
The current landscape of the Huallaga River basin is the result of tectonic activity and climatic oscillations that occurred from the Late Oligocene to the Early Miocene (Brown and Twomey 2009;Mora et al. 2010;Gonzales and Pfiffner 2012;Pfiffner and Gonzales 2013). Mountain uplifting and cooling periods have been pointed as key drivers of vicariance events and subsequent allopatric speciation in the central Andes, where species of the "palpebralis group" occur (Winger and Bates 2015;Lagomarsino et al. 2016, Wollenberg-Valero et al. 2019. As in other Andean sister species (Roberts et al. 2006;Twomey 2009, Winger andBates 2015;Winger 2017), vicariance is probably the main cause leading to the divergence between E. feiruzae and its sister species E. rudolfarndti. We acknowledge that the ND2 genetic distance between E. feiruzae and E. rudolfarndti is the lowest among species of Enyalioides (see Suppl. material 2: Table S1); however, we believe that differences in key color features (throat patch, antehumeral blotch) that are presumably used for species recognition represent strong evidence for assigning both taxa to different species. Figure 6. Phylogeny of Hoplocercinae. Maximum clade credibility tree based on a Bayesian analysis of mtDNA sequences. Posterior probabilities are indicated next to branches, except for short "intraspecific" branches (not shown for clarity). The new species described in this paper is shown in red. Outgroup taxa are not shown.
The upper basin of the Huallaga River has been highly disturbed by agriculture activities (Catenazzi and von May 2014). In fact, specimens of E. feiruzae were collected mostly in fragmented habitats surrounded by cacao and corn plantations. The only known protected populations are those from Tingo María National Park. On the basis of our limited knowledge on distribution, population status, and the resilience of E. feiruzae in a fragmented landscape, we recommend to place E. feiruzae in the Data Deficient category of the IUCN Red List of Threatened species (IUCN 2019).