A large and unusually colored new snake species of the genus Tantilla (Squamata; Colubridae) from the Peruvian Andes

A new colubrid species of the genus Tantilla from the dry forest of the northern Peruvian Andes is described on the basis of two specimens, which exhibit a conspicuous sexual dimorphism. Tantilla tjiasmantoi sp. nov. represents the third species of the genus in Peru. The new species is easily distinguished from its congeners by the combination of scalation characteristics and the unusual transversely-banded color pattern on the dorsum. A detailed description of the skull morphology of the new species is given based on micro-computed tomography images. The habitat of this new species is gravely threatened due to human interventions. Conservation efforts are urgently needed in the inter-Andean valley of the Maranon River.


INTRODUCTION
Seasonally dry tropical forests (SDTF) are characterised by a distinct seasonality with several months of arid-like conditions in which many plants lose their leaves (Murphy & Lugo, 1986). In South America SDTFs are discontinuously distributed and can occupy large areas such as the Caatinga in northeastern Brazil or small fragments as being found in inter-Andean valleys of Peru or Ecuador (Werneck et al., 2011). Nevertheless, the different areas of South American SDTFs are very divers and the species compositions differ substantially (Linares-Palomino, 2006;Miles et al., 2006;DRYFLOR, 2016). However, SDTFs are considered as being one of the most threatened tropical ecosystems with a strong rate of annual deforestation (Janzen, 1988;Miles et al., 2006;Pennington et al., 2006;Linares-Palomino et al., 2011;DRYFLOR, 2016). The Equatorial dry forest is one representative of this forest type, which expands from southern Ecuador to the northern part of Peru (Brack, 1986;Särkinen et al., 2011;Venegas, 2005), where it extends southward in two small stripes. One stripe continues along the coast west of the Andes, whereas the other penetrates the valley of the Marañón River and its tributaries. This North-to-South oriented valley is located in the Central Andes and bordered to the West and East by the Cordillera Occidental and the Cordillera Central, respectively. The inter-Andean dry forest pattern. Only T. shawi Taylor, 1949from Mexico, T. semicincta (Duméril, Bibron & Duméril, 1854 from Colombia and Venezuela, and T. supracincta (Peters, 1863) from Colombia, Costa Rica, Ecuador, Nicaragua, and Panama have a transversely-banded color pattern on the dorsal part of the body (Wilson, 1976;Wilson & Mata-Silva, 2015).
Reviewing published information on morphological characteristics of all other species of the genus Tantilla (Wilson, 1987;Sawaya & Sazima, 2003;Townsend et al., 2013;Wilson & Mata-Silva, 2014; revealed that the collected specimens can be easily differentiated from others of the genus by its comparatively high number of ventral scales and other scalation characteristics, its relatively large size, and transversely-banded color pattern. Herein a detailed description of the new species is given.

Morphological analyses
Measurements of head and scales were taken with a digital caliper and rounded to the nearest 0.1 mm, snout-vent length and tail length were taken with a measuring tape and rounded to the nearest 1 mm. Morphometric and meristic characters are abbreviated as follows: SVL (snoutvent length, from tip of snout to cloaca); TL (tail length); HW (head width across supraoculars); HH (head height at highest part of head); HL (head length); DSN (distance from tip of snout to nostril); DNE (distance from nostril to anterior margin of eye); ED (eye diameter); MBD (body diameter at midbody); MTD (midtail diameter). The number of ventral scales was counted in longitudinal row from mental to anal plate, and the number of subcaudal scales was counted in longitudinal row from the cloaca to the tip of the tail (Dowling, 1951). The number of dorsal scales rows around the body was counted at three different points: (1) at a head's length behind the head; (2) at midbody; (3) at a head's length before the cloaca. A dissecting microscope was used to count and characterize small scales and to identify the number of teeth in the male paratype.
For obtaining information on skeletal morphology, specimens were X-rayed in 2D (Faxitron Xray LX60) and in 3D by use of a micro-CT scanner (Bruker Skyscan 1272). Terminology for the skull structures was adopted from Bulloch & Tanner (1966) and Cundall & Irish (2008). The The partially everted hemipenes of the male paratype were removed from the specimen and prepared following Zaher & Prudente (2003). Finally the organs were scanned with the micro-CT scanner. Left hemipenis was scanned dry, whereas right hemipenis was scanned in alcohol.

Phylogenetic analysis
Genomic DNA was extracted from the collected tissue samples at the Center for Molecular  Palumbi et al. (1991). For amplification of RAG1 the primer pair RAG1f2 (light chain) and RAG1r3 (heavy chain) of Schulte & Cartwright (2009) were used.
Amplification with the 12S primer pair started with an initial denaturation step for 90 s at 94°C, and 38 cycles were run with denaturing for 45 s at 94°C, annealing for 60 s at 50°C, elongation for 120 s at 74°C, the final elongation for 300 s at 74°C, and cooling at 10°C. Amplification with the 16S primer pair started with an initial denaturation step for 900 s at 95°C, followed by 15 cycles of denaturation for 35 s at 94°C, annealing for 90 s at 60°C, elongation for 90 s at 72°C, plus 25 cycles of denaturation for 35 s at 94°C, annealing for 90 s at 45°C, elongation for 90 s at 72°C, the final elongation for 600 s at 72°C, and cooling at 10°C. Amplification with the RAG1 primer pair started with an initial denaturation step for 900 s at 95°C, and 40 cycles were run with denaturing for 20 s at 94°C, annealing for 50 s at 60°C, elongation for 90 s at 72°C, the final elongation for 600 s at 72°C, and cooling at 10°C. After the PCR, each sample proving successful DNA amplification in an agarose gel electrophoresis, was purified for sequencing using the QIAquick PCR Purification Kit (Qiagen). Subsequently, the samples were sequenced by Macrogen Europe Laboratory (Amsterdam, Netherlands). Obtained sequences were checked with the original chromatograph data using BioEdit 7.5.2 (Hall, 1999). The 12S rRNA and 16S rRNA data was supplemented with sequences of 48 species representing 27 genera of American colubrid snakes obtained from GenBank. Accession numbers are provided in Table 1 Modeltest (Posada & Crandall, 1998) as implemented in the package 'phangorn' for Cran R.
Phylogenetic trees were inferred using MrBayes 3.2.6 (Ronquist et al., 2012), estimating model parameters separately for each gene by partitioning the data set. We used a random starting tree and four independent runs with a maximum of 10 million generations each, sampled every 1000.
Runs were stopped when the average standard deviation of split frequencies had reached 0.01.
Convergence of the Markov chains and effective sample sizes were checked with Tracer v1.6 (Rambaut et al., 2014) and the initial 25% of each run were discarded prior to building a consensus tree. In addition to the Bayesian inference (BI), phylogenies were also calculated with Maximum Likelihood (ML) via the RAxML BlackBox (Stamatakis, Hoover & Rougemont, 2008) using the partitioned data, the Gamma model of rate heterogeneity, and 100 bootstraps.

Morphological analyses
As typical for the genus Tantilla, the two specimens possess a number of 15 smooth dorsal scale rows throughout the body, one preocular, no loreal, no suboculars, 1+1 temporals, a divided cloacal shield, paired subcaudals (Table 2). Additionally, the skull of the new species is composed of similar bones and bone structures as other species of the genus Tantilla. A comparison with three other congeners (T. capistrata, T. melanocephala and T. relicta) reveals great similarity to our new species with only minor differences in the shape or size of some bones (Fig. 6). Scalation characteristics and dorsal color pattern are very similar in both specimens. Nevertheless, they show conspicuous differences in body size and ventral coloration which are most likely due to sexual dimorphism and/or age difference.

Phylogenetic analysis
Fragments of the mitochondrial gene 16S (526 bps) and nuclear gene RAG1 (1050 bps) of both specimens were compared. 16S showed no differences between specimens, and RAG1 revealed only a single base pair variation. The strong genetic similarity coupled with the weak morphological variation suggests these specimens are the same species.
The Bayesian consensus tree ( Fig. 1) obtained from 788 bp of mitochondrial DNA (12S and 16S rRNA) was based on 3600 sampled trees and effective sample sizes were > 1000, which indicated good mixing of the Markov chains. Although the topologies obtained from BI and ML (Fig. S1) differ and the nodes were generally not very well resolved in terms of posterior probabilities and bootstrap values, the four species of Tantilla formed a well-supported clade including our two specimens (CORBIDI 7726 and ZFMK 95238). However, node support is lacking between the different species of Tantilla, obscuring the exact relationship between these taxa. Description of holotype. an adult female with a SVL of 513 mm; TL 125 mm; HL 16.3 mm; HW Body robust, tail long, body and tail round in cross-section; dorsal scales in 15-15-15-rows, without reduction, rhomboid, smooth, lacking keels or apical pits; 182 ventrals; tail distinctly smaller in diameter than the body, long and tapering, tail spine pointed; 57 paired subcaudals; cloacal plate divided. Head (Fig. 2)  Manuscript to be reviewed times longer than posterior chinshields; posterior chinshields about 2.2 times longer than wide, laterally contacting fourth infralabial, dorsally separated from the ventrals by four gular scales.
Coloration. In live, the dorsal ground color of head, body and tail is orange-yellowish, slightly paler laterally, most scales on body and tail with reddish-brown outlines; there are about 27 blackish dorsal crossbars on the body that are four to seven scales in length and are stretched across all dorsal scale rows except the most lateral row, slightly longer than ground color interspaces, fused in some parts of the body to form a zigzag band, slightly mottled in some parts with yellow. There are 12 dark tail blotches, reaching to subcaudals on both sides, fused in median part of the tail along the midline to form a zigzag band. Head with a large dark dorsal tshirt-shaped blotch covering frontal, supraoculars, most of parietals except for the most posterior parts, and posterior part of prefrontals, the dark blotch is laterally extended at eye level, covering orbit, preocular, third supralabial, and adjacent parts of second and fourth supralabials, respectively. Infralabials, rostral and mental yellowish, except for blackish region surrounding the lingual groove. The ventral scales of head and body and subcaudal scales are cream-colored with dark dotted outlines in some parts. The coloration of the tongue is black to grayish-black (Fig. 3).
In preservative, the dorsal pattern on body and tail of a light ground color with dark crossbands remains unchanged, likewise the dark coloration of the head; the orange-yellowish dorsal ground color changed to cream and the darker outlines of most dorsal scales disappeared; the ventral coloration changed to grayish-white in some parts (Fig. 3). The coloration of the tongue changed to gray.
Variation. The single paratype is a small male with a conspicuous sexual dimorphism in body size and ventral coloration compared to the holotype. As the hemipenes ornamentation is as detailed as in other adult specimens of the genus Tantilla we assumed it to be already sexually mature. Intraspecific variation in scale counts and measurements is shown in Table 2

Manuscript to be reviewed
The dorsal ground coloration of the paratype (Fig. 3)  The anterolateral portion of the jaw is formed by the dentary, which contains a row of about 22, slightly recurved teeth. The longer proximal part of each jaw is without teeth. (Fig. 5). There are Distribution and natural history. This species is so far known from the southern portion of the seasonally dry forest along the Marañón River and its tributaries, from near Santa Rosa de Marcamachay at the Río Crisnejas, Province Cajabamba, and from near Laguna de Pías, Province Pataz, both Department of La Libertad, at elevations of 1154 m and 1726 m a.s.l, respectively ( Fig. 8 & 9). The female CORBIDI 7726 was detected on 7 th of January 2010 at 12.30 pm resting on a stone. The male ZFMK 95238 was detected on 12 th of October 2010 at 8.15 pm on pebblyclayey ground. Air temperature when animals were sighted was 33.3°C and 28.1°C, respectively.

DISCUSSION
Despite the comparatively close localities (> 70 km air distance) and similarities in scale counts and arrangement of scales, the conspicuous differences of both specimens in body size and ventral coloration created some doubt if they represent the same species. A pairwise analysis of 1576 bp derived from the mitochondrial gene 16S and the nuclear gene RAG1 showed only a single difference in one base position in the RAG1 fragment, thus strongly supporting the assumption that both specimens belong to the same species. The collection and examination of further specimens is needed to determine whether the differences in size and color pattern are a product of sexual dimorphism or are referred to a different cause (e.g. age dependence, geographic variation).
In order to get a general idea of the phylogenetic position of the new species described herein, we performed phylogenetic analyses based on 12S rRNA and 16S rRNA. We did not attempt to conduct a taxonomically extensive analysis of South American Colubridae, instead we preferred to include only those species, for which the two gene regions sequenced in this study were available in GenBank. Our phylogenetic tree based on mitochondrial DNA (Fig. 1) corroborates the assignment of our new species into the genus Tantilla. Both the monophyly of the sampled Tantilla and the conspecificity of our two dimorphic specimens are well supported by the analyses. However, intrageneric relationships remain dubious due to the few species with genetic data available. Moreover, scutellation characteristics and the comparison of the skull morphology via micro-CT scans (Fig. 6) strongly support this hypothesis.
With currently 62 species assigned to this genus it represents the second largest genus of New World colubrid snakes, after the genus Atractus. However, taxonomic approaches are limited by the fact that pretty much nothing concrete is known about the phylogenetic relationships within the genus Tantilla, and its monophyly has not yet been tested adequately. Only few specimens representing fewer than ten species of Tantilla have been included in previous DNA-based studies (e.g. Vidal et al., 2000;Lawson et al., 2005;Burbrink & Myers, 2015;Schrey et al., 2015;Chambers & Hebert, 2016). Thus, Tantilla could benefit from a thorough taxonomic treatment involving a stronger genomic sampling component. Such a revision would test species concepts, update their known distributions, reveal their genetic diversity and give some clues about their evolutionary history. Furthermore, with robust molecular data generic designation could be proposed with more confidence.
The habitat of this new species is gravely threatened due to human interventions, such as deforestation, mining activities, and intended dam constructions for hydroelectric projects. To date no protected area has been established in the Marañón river valley. We hope that this beautiful and untypically colored new Tantilla could serve as a flagship species, together with several other endemic species of reptiles and birds, for the establishment of conservation strategies in this region. Unless these strategies are implemented the biodiversity found in this unique habitat, including the new, endemic species described here, may fall into serious decline.  Manuscript to be reviewed