A new riparian Andean Potamites (Reptilia, Squamata, Gymnophtalmidae) from El Sira Mountains, central Peru, with comments on P. ecpleopus Cope 1875, and on the taxonomy and biogeography of Potamites

We describe a new species of riparian lizard from the foothills and submontane forest of El Sira Communal Reserve, Departamento de Huánuco, Peru, at elevations from 540 to 760 m. We name the new species Potamites hydroimperator sp. nov. because it is the only lizard living in streams in this isolated mountain. The new species is distinguishable from other congeners (except P. ecpleopus) by having longitudinal rows of keeled scales on dorsum, and males with 5–8 lateral ocelli on each fl ank. Potamites hydroimperator sp. nov. diff ers from P. ecpleopus by being smaller, having dorsal scales on forelimbs smooth or slightly keeled, subconical tubercles on forelimbs weakly present or absent, and four to fi ve superciliar scales. We comment on variation in the pattern of keeling of dorsal scales among Amazonian lowland, Andean foothills and montane species of Potamites, and the potential relevance of this scalation pattern for taxonomic and biogeographic studies. The new species increases our knowledge of the diverse and endemic herpetofauna of El Sira, and of diversity within the genus Potamites. Despite El Sira’s status as natural protected area, gold mining activities are destroying the streams where we found P. hydroimperator sp. nov., threatening their populations.


Introduction
Over the last decade, molecular studies have helped resolve some mysteries in the taxonomy of gymnophtalmid lizards (Goicoechea et al. 2012(Goicoechea et al. , 2013Torres-Carvajal et al. 2016;Kok et al. 2018;Moravec et al. 2018;Sánchez-Pacheco et al. 2018;Fang et al. 2020;Vasquez-Restrepo et al. 2020). These studies untied complicated knots in the systematics of Cercosaurinae Gray, 1838, especially in the genera Echinosaura Boulenger, 1890, Gelanesaurus Torres-Carvajal, Lobos, Venegas, Chávez, Aguirre-Peñafi el, Zurita & Echevarría, 2016, Neusticurus Duméril & Bibron, 1839, Proctoporus Tschudi, 1845, Potamites Doan & Castoe, 2005, and Riama Gray, 1858. Among these studies, Torres-Carvajal et al. (2016) clarifi ed the systematics of the genus Potamites, elevating Potamites trachodus Uzzell, 1966 (formerly a subspecies of P. strangulatus Cope, 1868) to a species level, and assigning P. cochranae Burt & Burt, 1931 and P. fl avogularis Altamirano-Benavides, Zaher, Lobo, Grazziotin, Sales Nunes & Rodrigues, 2013 to the new genus Gelanesaurus. Therefore, the genus Potamites currently contains seven species: P. ecpleopus Cope, 1875, P. erythrocularis Chávez & Catenazzi, 2014, P. juruazensis Avila-Pires & Vitt, 2001, P. montanicola Chávez & Vasquez, 2012, P. ocellatus Sinitsin, 1930, P. strangulatus Cope, 1868, and P. trachodus Uzzell, 1966. Despite these eff orts, delimitation of species of Potamites is still challenging (Moravec et al. 2018). The problems preventing the solution of the taxonomic knots of Potamites are lying mainly with its type species, P. ecpleopus. Uzzell (1966) reported the loss of the holotype of P. ecpleopus and the uncertainty regarding its type locality. Also, the original description is vague, and populations assigned to this species show a highly variable morphology (Uzzell 1966;Avila-Pires 1995). Indeed, Uzzell (1966) reported a high degree of morphological variation in populations from Colombia, Ecuador and Peru, and Avila-Pires (1995) mentioned that samples from western Brazil resemble externally those from the Ecuadorian and Peruvian Amazon. In the same study, Avila-Pires found that the southern Peruvian populations studied by Uzzell are similar to those from Carajás (Pará, southern Brazil), but diff erent from those from northern Brazil. Despite these taxonomically relevant fi ndings, researchers were unable to identify traits of P. ecpleopus associated with geographic patterns that could inform the biogeography of this species. Unsurprisingly, the only consensus between these studies is that the skin of P. ecpleopus sensu stricto bears many tubercles, with tubercles present on dorsum, neck, fl anks and tail (Uzzell 1966;Avila-Pires 1995). The paucity of reliable diagnostic traits complicated the taxonomy of P. ecpleopus, forcing scientists to be cautious in assigning a neotype.
Molecular phylogenies of the tropical Andean Cercosaurini revealed Potamites ecpleopus to be paraphyletic, with two geographically separated clades, one in the Ecuadorian Amazon and the other in the Peruvian and Brazilian Amazonia (Torres-Carvajal et al. 2016;Moravec et al. 2018). Because none of the sequenced P. ecpleopus correspond to specimens from the type locality or nearby areas, it was not possible to assign the name P. ecpleopus to either of these clades (Moravec et al. 2018). Indeed, one of the Peruvian samples included in these studies corresponds to a population of Potamites inhabiting the Andean foothills of El Sira Communal Reserve, an isolated chain of mountains about 450 km from the type locality (distance as the crow fl ies).
El Sira is a protected area located on the eastern slopes of the Andes in central Peru (see Fig. 1). The streams within the reserve drain into the middle basin of the Ucayali River, the longest tributary of the Amazon River in Peru. The landscape consists mainly of montane forest in the Andean foothills, interspersed with isolated and moderately high mountains (Duellman & Toft 1979), where the topography goes from 200 to 2200 m a.s.l. (SERNANP 2010). The accidented topography makes El Sira Mountains quite hard to reach for scientifi c expeditions, which likely explains the low number of fi eld expeditions to date. Strikingly, researchers discovered new species of amphibians and reptiles during each one of the previous expeditions (Duellman & Toft 1979;Torres-Carvajal et al. 2015;Brown et al. 2019 Finally, after performing morphological revisions and adding new sequences to the genetic data previously published, we conclude that Potamites lizards from El Sira, previously included in the analysis of Torres-Carvajal et al. (2016) and Moravec et al. (2018), belong to an undescribed taxon that is morphologically distinguishable from congeners. Here we describe this new species of Potamites.

Material and methods
For the description format of the new species we followed Uzzell (1966), Vanzolini (1995), and Avila-Pires & Vitt (1998). For comparisons, we used the descriptions of all known species of Potamites in the literature: data for most Neusticurus and Potamites were taken from Uzzell (1966), data for P. juruazensis from Avila-Pires & Vitt (1998), for P. ocellatus from Vanzolini (1995), and for P. erythrocularis, P. juruazensis, P. montanicola, and P. ocellatus from morphological revision of the respective type series (see Appendix 1). For morphological comparisons with P. ecpleopus we relied on data reported by Cope (1875), Uzzell (1966), and Avila-Pires (1995) as well as examined specimens (Appendix 1). We followed Uzell (1966) and Köhler & Lehr (2004) for nomenclature of scale characters. We measured scale sizes using precision callipers and rounding to the nearest 0.1 mm. For characters recorded on both sides, we presented the condition on the right side fi rst. We fi xed everted hemipenes with formalin 3.7% (Zaher 1999) and followed Sánchez-Pacheco et al. (2018) for hemipenial terminology. We collected specimens at night while conducting opportunistic visual surveys. We euthanized specimens with T61 (Embrutamide), fi xed them in 10% formalin, and stored them in 70% ethanol. Before fi xation, we obtained tissue samples (liver) from all individuals and stored them in absolute ethanol.
We downloaded 12S, 16S, c-mos, and ND4 sequences of six of seven species currently assigned to Potamites (on the basis of BLAST) from GenBank and added two novel sequences to the dataset (see Appendix 2). We estimated genetic distances for the 12S, 16S, c-mos, and ND4 rRNA mitochondrial fragment (see Supp. fi les 1, 2, 3, 4) to provide further support of species delimitation.
All specimens captured for this study are covered by the permit RJ N° 003-2014-S ERNANP-RCS-JEF029-2016-SERFOR-DGGSPFFS issued by Servicio Nacional de Areas Protegidas de Peru which includes scientifi c collection of live individuals. We deposited specimens and tissue samples in the Herpetology Collection of Centro de Ornitología y Biodiversidad (CORBIDI). We used high-resolution photographs taken in the fi eld to describe the coloration in life of the type series.

Repositories
The material upon which this study is based is located in the following institutions:  Torres-Carvajal et al. (2016) examined the phylogenetic relationships of the new species and showed that the female CORBIDI 14382 (Genbank accession code KU902066) is part of a lineage within Potamites.

Generic allocation and species delimitation criteria
We also add novel sequences to the genetic analysis (Appendix 2) which agreed with Torres-Carvajal et al. (2016) placing the new species into the Potamites as a separated clade. We estimated uncorrected p-distances with the R package ʻAPEʼ (Paradis et al. 2004) to support the specifi c delimitation of the new species from its congeners. Although there is not a set threshold for delimiting species, we use 2% distance for 12S and 16S which is even more than the standard cut-off value suggested by the BOLD (Barcode of Life Data) system as a reasonable criterion to identify putative new species (Ratnasingham & Hebert 2007). The benefi t of using 12S, 16S, c-mos, and ND4 sequences is that these fragments have been previously sequenced for Potamites lizards (Pellegrino et al. 2001;Torres-Carvajal et al. 2016;Moravec et al. 2018). Additionally, our morphological revision confi rmed that other specimens collected in the same locality belong to the same taxon. Therefore, following de Queiroz (1998,2007) and the unifi ed species concept, we assign our sample to a new taxon into the genus Potamites.

Diagnosis
A medium-sized Potamites (SVL = 50.5-59.6 mm, n = 5) characterized by the following combination of morphological features: 1) Body slender, slightly depressed laterally, maximum SVL in males 59.6 mm, 51.6 mm in females; 2) head acuminate from dorsal and lateral view, scales on the dorsal surface of the head smooth; 3) frontonasal divided, shorter than frontal, small rounded azygous scale usually present between frontonasal and prefrontals; 4) prefrontals present; 5) frontal present; 6) loreal present, romboid, not in contact with supralabials; 7) nasal partially divided, behind the nostril; 8) supraoculars four, anteriormost not fused with anteriormost superciliar; 9) superciliar series complete, usually fi ve, occasionally four; 10) supralabial-subocular fusion absent; 11) lower palpebral disc oval, transparent, not divided; 12) postoculars three; 13) postparietals 10-11; 14) hemipenis in two pairs, transverse sutures perpendicular with respect to midline of body; 15) dorsals rectangular, subimbricate, granular and keeled; 16) 34-36 longitudinal rows of dorsal keeled scales, 2-3 granular scales between the two paravertebral rows of keeled dorsals; 17) fl anks partially covered by subconical low tubercles, absent or poorly present above the insertion of forelimbs; 18) anterolateral and dorsal brachial scales of forelimbs smooth or slightly keeled; 19) 32-36 scales around midbody; 20) 21-22 longitudinal rows of ventral scales; 21) 32-33 total femoral pores in males, two in females, two scales between femoral pores; 22) 14-16 subdigital scales on 4 th fi nger; 23) 21-24 on 4 th toe; 24) forelimb reaching anteriorly the fourth supralabial; 25) tail slightly compressed with three rows of lateral scales per two ventral caudals; 26) hemipenis globose (Supp. fi le 5), lobes indistinct from hemipenial body; 27) hemipenial fl ounces chevron shaped, lacking calcifi ed spines, laterally oriented and forming two chevrons on distal half of hemipenis while its basal half is covered with four transverse fl ounces, separated by a small expansion pleat; sulcate fl ounces about as wide as asulcate fl ounces; isolated transversal fl ounces on proximal-central region of asulcate face absent; distal fi liform appendages on the hemipenial lobes absent, sulcus spermaticus single; 28) dorsum yellowish brown to dark brown with slightly darker irregular blotches; 5-8 lateral ocelli with a creamy yellow center on each fl ank in males, a single ocellus with a pale brown center at the level of the insertion of forelimbs on each fl ank in females; fl anks yellowish brown or partially pigmented of vermilion red in adult males; a creamy white to yellow diagonal stripe (continuous or discontinuous) going from 4 th or 5 th supralabials to 5 th infralabial; ventral coloration in males vermilion red to reddish pink on belly and base of the tail, yellowish pink on ventral surface on limbs, pale blue to grayish blue on ventral surface of the neck and throat, saff ron yellow on femoral pores; ventral coloration in females creamy yellow on throat, chest, belly, ventral surface of limbs and base of the tail, with black speckling on throat; ventral surface of the rest of the tail yellowish brown to dark brown in males and females; pupil black surrounded by a yellowish orange to copper ring, iris olive to yellowish brown.

Diff erential diagnosis
The presence of longitudinal dorsal rows of keeled scales makes Potamites hydroimperator sp. nov. similar in appearance to populations of P. ecpleopus, a species widely distributed along the eastern Andean foothills and the Amazon (Uzzell 1966;Avila-Pires 1995;Ribeiro-Júnior & Amaral 2016, 2017. Potamites hydroimperator sp. nov. can be distinguished from P. ecpleopus (characters for P. ecpleopus in parenthesis) by having low tubercles on fl anks and above forelimbs (vs fl anks strongly tuberculate, see

Etymology
The specifi c epithet ʻhydroimperatorʼ derives from the Greek word ʻhydroʼ (ʻwaterʼ), and the Latin name ʻimperatorʼ (ʻemperorʼ). The name alludes to the riparian habits of the new species, which is the only lizard in El Sira using the streams as part of its habitat.

Coloration in life
Dorsal and lateral surfaces of head brown to dark brown (Fig. 2C); rostral scale same colour as head; superior labium bluish grey with pale red irregular blotches on third, fourth, and fi fth supralabials, irregular vermilion red blotches on sixth and seventh supralabials, an irregular diagonal pale yellow stripe going from fourth supralabial to superior edge of fi fth infralabial; infralabials grayish blue, pale red faded blotches on fi fth infralabial, and a vermilion red patch on sixth infralabial; iris olive yellow, with a copper orange ring surrounding pupil; throat and gular region greyish blue, darker on ventral surface of neck. Dorsal surface of body yellowish brown, with some darker areas; coloration on fl anks yellowish brown with eight black ocelli on both sides, going from neck to base of tail (after insertion of forelimbs), each ocelli bearing a white center; tuberculate scales darker than granular scales; belly vermilion red. Dorsal surface of limbs similar to body, ventral surface of arms and legs pinkish cream; femoral pores saff ron yellow. Coloration of dorsal surface of tail like that of body, ventral surface of tail pinkish cream with irregular red blotches at base, and a darker fl ecking pattern on rest of tail.

Coloration in preservative
General coloration pattern is as described for the holotype in life. Dorsal coloration is mostly brown, with labial region greyish brown, throat yellowish brown, chest dark brown and venter creamy yellow. Flanks yellowish brown, lateral ocelli present. Ventral surfaces of forelimbs pale yellow with black blotches; ventral surfaces of limbs creamy yellow with dark speckling; ventral surface of tail brownish yellow with diff use dark speckling.

Taxonomic remarks
Females CORBIDI 14382 and 14470 (Fig. 3 E-F) have four superciliars. The frontonasal scale is partially divided in male CORBIDI 14469 (Fig. 3 C-D). Furthermore, the azygous scale between the frontonasal and prefrontals is absent in male CORBIDI 14469 (Fig. 3C-D) and female CORBIDI 14470 (Fig. 3E-F). We described colour pattern variation of the type series in the diagnosis. Other measurements and scutellation data of the type series are summarized in Table 1.

Dis tribution and ecology
We observed the new species in streams and surrounding vegetation in two nearby sites, in the eastern slopes of the mountains of El Sira Communal Reserve, from 550 to 750 m a.s.l. (Fig. 1). These streams drain into the Pachitea River, which is a tributary of the Ucayali River. The habitat is a montane forest (Fig. 6), with riparian vegetation consisting mainly of bushes, tree ferns and trees canopy ~ 20 m tall. Climbers are also present (lianas) as well as low epiphytes, ferns, mushrooms, and lichens on the ground and on rocks alongside the streams. We found lizards at night by lifting rocks or removing vegetation near stream banks. The male CORBIDI 14468 and the female CORBIDI 14470 used the stream to escape from the fi eld researchers. We observed the new species in sympatry with the lizards Anolis cf. fuscoauratus, Enyaloides laticeps Guichenot, 1855, Iphisa elegans Gray, 1851, Potamites cf. juruazensis and the snakes Bothrocophias microphtalmus Cope, 1875, Dipsas catesbyi Sentzen, 1796, Dipsas pavonina Schlegel, 1837, and Xenoxybelis argentea Daudin, 1803. None of these squamates were found in water. On the basis of known diet habits and previous reports (Dueñas & Valencia 2018), we regard B. microphtalmus as a potential predator of Potamites hydroimperator sp. nov. Torr es-Carvajal et al. (2016) examined the phylogenetic relationships of Potamites, and assigned the new species to a lineage distinct from P. ecpleopus from lowlands and foothills of southern Ecuador. Moreover, inclusion of novel sequences from the Peruvian Amazon and Andean foothills confi rms this hypothesis (Moravec et al. 2018;this study). In the phylogeny published by Torres-Carvajal et al. (2016), a Brazilian specimen appears to be closely related to the new species. The genetic sequences belong to a specimen from Apiacás, Mato Grosso, southern Brazil (Pellegrino et al. 2001), more than 400 km as the crow fl ies from the type locality of P. hydroimperator sp. nov. Because we were not able to examine this specimen and its divergent distribution, we do not include it into the new taxa described herein. We recommend further genetic and morphological analyses to confi rm the relationships of the specimens from Apiacás, Brazil with the new species and other lineages of the Potamites ecpleopus complex. Thus,  Moravec et al. 2018). Uzzell (1966) suggested that the type locality was somewhere between Rioja, Moyobamba and Balsapuerto (Peru), three sites less than 70 km apart lying on the Andean foothills of the Huallaga River drainage, with no geographical barriers separating them (Fig. 1). Therefore, we may assume that the type series of P. ecpleopus belongs to a population from the Andean foothills.

Disc ussion
In order to make proper comparisons, we have reviewed individuals recorded in the Andean foothills of Yurimaguas, Pongo de Caynarachi, and Cerros de Kampankis ( Fig. 8B-C), all sites near the ʻtype localityʼ area (Fig. 1). We concluded that these populations share the phenotypic pattern and scutellation described by Cope (1875), Uzzell (1966) and Avila-Pires (1995) recognizing P. ecpleopus as one of the most tuberculate species among the genus. Indeed, we noted fl anks and forelimbs themselves are highly keeled as described by Uzzell (1966). Otherwise, we have not observed similar condition in these traits in P. hydroimperator sp. nov. Additionally, our 16S genetic analysis shows a 2-7% distance between Kampankis samples and the new species (see Appendix 2 and Supp. fi les 1, 2, 3, 4). On the basis of this evidence, it is possible to assign specimens we examined from Yurimaguas, Pongo de Caynarachi and Cerros de Kampankis to P. ecpleopus sensu Uzzell (1966).
In his comprehensive study of the Potamites lizards, Uzzell (1966) considered P. ocellatus as a junior synonym of P. ecpleopus, after reviewing the paratypes from Chanchamayo, Central Peru. Later Vanzolini (1995) revalidated P. ocellatus based on the holotype from Rurrenabaque, El Beni, Bolivia. Also, Vanzolini (1995) noticed that the strongly keeled dorsum and the scattered arrangement of the tubercles on the body of the paratypes contrast with the holotype morphology. Thus, the paratypes of P. ocellatus from Chanchamayo were assigned to P. ecpleopus (Uzzell 1966;Vanzolini 1995). The scattered arrangement of dorsal keels or tubercles is a diagnostic character for the recently described and well-supported montane species P. erythrocularis and P. montanicola (Chávez & Vásquez 2012;Chávez & Catenazzi 2014;Torres-Carvajal et al. 2016;Moravec et al. 2018). A scattered arrangement of dorsal keels or tubercles is not seen in P. ecpleopus sensu lato. Therefore, we consider unlikely that the specimens from Chanchamayo could belong to any form of P. ecpleopus. Instead, we suggest that further genetic analysis should be performed to solve their taxonomic status.
As mentioned a bove, scattered keeled dorsals is a character previously reported for P. erythrocularis and P. montanicola ( Fig. 7C and Fig. 7E respectively), two species inhabiting streams between 1000-2600 m a.s.l., and for P. ocellatus (Fig. 7F), the other montane species of this genus. Otherwise, the other Andean foothills species of Potamites bear scattered keeled dorsals (i.e., P. trachodus) or aligned rows (i.e., P. ecpleopus and P. hydroimperator sp. nov.). Finally, Amazon lowlands species such as P. juruazensis and lowland populations of P. ecpleopus have longitudinal aligned rows of dorsal keeled scales (Fig. 7D). Ecologically, Andean populations experience colder environments (water temperature in streams range between 3 and 12°C) and stronger water velocity in streams than lowland Potamites. Ecological diff erences might have contributed to montane species evolving diff erent patterns of dorsal keels from lowland populations (scattered keels in montane species, longitudinally aligned in lowland species). Considering that environmental conditions are more similar between montane forests and foothills ecosystems, dorsal keeling in Andean foothill populations could be variable, as exemplifi ed by P. trachodus having scattered keeles, and P. hydroimperator sp. nov. having aligned rows.
The recent discovery of several montane species highlights the importance of the Andes for speciation of Potamites and reveals a contrast between montane Potamites being restricted to upper basins from Central Bolivia to Central Peru, and lowland species being widespread across the Amazon Basin in Ecuador, Peru, Brazil, and southern Colombia (Uzzell 1966;Avila-Pires 1995;Ribeiro-Júnior & Amaral 2016, 2017. Despite geographic barriers (i.e., Apurimac River for P. montanicola, Inambari River for P. erythrocularis) separating the distribution ranges of montane species, both montane species share a similar keeling phenotype, whereas populations of the ʻlowland keeling phenotypeʼ are widely distributed in the Amazon basin. We hypothesize that adaptation to stronger torrents through the convergent evolution of the dorsal keels pattern occurred repeatedly as lowland Potamites colonized montane environments.  Uzzell, 1966 (CORBIDI 15515). additional molecular analyses focused on the diversifi cation of Potamites are necessary to test alternative biogeographic hypotheses.
The type locali ty of P. hydroimperator sp. nov. lies on a mountain ridge that has been recognized as a likely harbour of unknown fauna (Duellman & Toft 1979). Every herpetological survey performed in these mountains has discovered new species (Duellman & Toft 1979;Torres-Carvajal et al. 2016;Brown et al. 2019). The accidented topography, and the presence of a single trail into the El Sira Communal Reserve, explains why only a handful of scientifi c expeditions have been able to reach the highest elevations of these mountains. We believe that our knowledge of this area is still far from being complete. Future fi eldwork should focus on opening new trails to the top of El Sira Mountains, in order to increase our understanding of the herpetofauna inhabiting these forests.
In spite of the type locality of P. hydroimperator sp. nov. lying inside a protected area, during our fi eldwork we observed illegal gold mining activity near one of the spots where we recorded the new species. Gold mining occurred in the lower reaches of the streams that are fl owing down to the Pachitea River, whereas the new species inhabits the upper reaches of these streams. Nonetheless, the establishment of mining camps causes large forest clearings, which could aff ect the habitat of the new species. Strikingly, expeditions to the eastern slopes of El Sira have not found P. hydroimperator sp. nov. (GC unpublished data). Following the IUCN Red List Categories and Criteria (IUCN 2001), considering that the aforementioned threat could eventually cause habitat fragmentation and degradation, the known distribution of P. hydroimperator sp. nov. includes only two very close localities in the same mountain (no more than 4 km far), and the possible Area of Occupancy (AOO) is not wider than 6000 km 2 (which is the entire extension of the El Sira Communal Reserve (SERNANP 2010), we provisionally assess this species as Vulnerable (VU) under criteria B2ab(iii).