A new species of blunt-headed vine snake (Colubridae, Imantodes) from the Chocó region of Ecuador

Abstract We describe a new species of Imantodes from the Chocó region of northwestern Ecuador. The new species differs most significantly from all other congeners in lacking a loreal scale. We analyze the phylogenetic relationships among species of Imantodes based on two mitochondrial genes, and postulate that the new species and Imantodes lentiferus are sister taxa. A key to the species of Imantodes from Ecuador is presented.


introduction
The New World colubrid snake clade Dipsadinae Bonaparte 1838 includes more than 400 extant species assigned to approximately 25 taxa traditionally ranked as genera (Daza et al. 2009;Zaher et al. 2009). Most members of the Dipsadinae have unilobed (or nearly unilobed), unicapitate hemipenes, with the sulcus spermaticus dividing distally (Zaher et al. 2009). One of the most remarkable dipsadine genera is Imantodes. Its long, thin body, disproportionately slender neck, and blunt head, makes easy to distinguish Imantodes from all other New World snakes. This genus includes six currently recognized species (I. cenchoa, I. gemmistratus, I. inornatus, I. lentiferus, I. phantasma, and I. tenuissimus) commonly known as blunt-headed vine snakes, occurring from Mexico to Argentina (Myers 1982).
Studies on phylogenetic relationships and species limits among dipsadines are scarce. However, recent work provides strong evidence from DNA sequence data for a clade containing Imantodes and Leptodeira, although monophyly of Imantodes remains controversial (Daza et al. 2009;Mulcahy 2007). Futhermore, I. gemmistratus as currently circumscribed appears to be paraphyletic (Daza et al. 2009;Mulcahy 2007). Future studies with increased taxon and character sampling will probably clarify the phylogenetic relationships and species limits within Imantodes.
Three species of blunt-headed vine snakes are known from Ecuador; I. inornatus and I. lentiferus occur west and east of the Andes, respectively, whereas I. cenchoa is known from both versants (Torres-Carvajal and Salazar-Valenzuela 2012). In this paper we describe a new species of Imantodes from northwestern Ecuador and infer its phylogenetic affinities to other species in the genus as currently circumscribed.

Morphological data
All type specimens of the new species described in this paper are listed in the type series below, and were deposited at the Museo de Zoología, Pontificia Universidad Católica del Ecuador, Quito (QCAZ), the Museo Ecuatoriano de Ciencias Naturales, Quito (DHMECN), and the Amphibian & Reptile Diversity Research Center at The University of Texas at Arlington, USA (UTA). Specimens of other species of Imantodes examined in this study are listed in the appendix. Snout-vent length (SVL) and tail length (tL) measurements were recorded to the nearest millimeter. All other measurements were made with digital calipers and recorded to the nearest 0.01 mm. Sex was determined by noting the presence of hemipenes, everted or by tail dissection. Partially everted hemipenes were prepared following standard techniques (Pesantes 1994;Zaher 1999). Differences in scale counts between the new species and other species of Imantodes were evaluated with t-tests for normally distributed variables (i.e., Shapiro-Wilk test, P > 0.05), all of which had equal variances (i.e., F-test, P > 0.001). We used the program PAST 2.15 (Hammer et al. 2001) for all statistical tests.

DNA Sequence Data
Total genomic DNA was digested and extracted from liver or muscle tissue using a guanidinium isothiocyanate extraction protocol. Tissue samples were first mixed with Proteinase K and lysis buffer and digested overnight prior to extraction. DNA samples were quantified using a Nanodrop® ND-1000 (NanoDrop Technologies, Inc), re-suspended and diluted to 25 ng/ul in ddH2O prior to amplification.
We amplified 1674 nucleotides (nt) encompassing two mitochondrial loci, NADH dehydrogenase subunit 4 (ND4, 651 nt) and cytochrome b (cyt-b, 1023 nt) from five individuals of I. cenchoa, three of I. lentiferus, three of the new species described herein and one of Leptodeira septentrionalis. Cyt-b was amplified using the primers Gludg, L14910, and H16064 (Burbrink et al. 2000;Parkinson et al. 2002), whereas ND4 was amplified using the primers ND4, LEU and ND412931L (Arévalo et al. 1994;Blair et al. 2009). Additionally, we used sequences of I. cenchoa, I. gemmistratus, I. inornatus, I. lentiferus and L. septentrionalis from GenBank. Although monophyly of Imantodes has not been rigorously tested yet (see Discussion), for the purposes of this study we assume that Imantodes forms a clade and root our tree with L. septentrionalis. Gene regions of taxa included in phylogenetic analyses along with their GenBank accession numbers and locality data are shown in Table 1. Amplification of genomic DNA consisted of an initial cycle at 94 C for 3.5 min, 42 C for 1 min, and 68 C for 1.5 min, followed by 40 cycles of a denaturation at 94 C for 30 s, annealing at 52 C for 30 s, and extension at 72 C for 60 s, as well as a final extension at 72 C for 15 min.

Phylogenetic analyses
Editing, assembly, and alignment of sequences were performed with Geneious ProTM 5.3 (Drummond et al., 2010). Phylogenetic relationships were assessed under a Bayesian approach in MrBayes 3.2. 0 (Ronquist and Huelsenbeck, 2003). The model of character evolution for each gene was obtained in JModeltest (Posada, 2008) under the Akaike information criterion. Genes were combined into a single dataset with two partitions, one per gene. Four independent analyses were performed to reduce the chance of converging on a local optimum. Each analysis consisted of five million generations and four Markov chains with default heating values. Trees were sampled every 1000 generations resulting in 5000 saved trees per analysis. Stationarity was confirmed by plotting the log-likelihood scores per generation in the program Tracer 1.2 (Rambaut and Drummond, 2003). 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 bipartitions and branch lengths, respectively. Adequacy of mixing was assessed by examining the acceptance rates for the parameters in MrBayes and the effective sample sizes (ESS) in Tracer. After analyzing convergence and mixing, 500 trees were discarded as "burn-in" from each run. We then confirmed that the four analyses reached stationarity at a similar likelihood score and that the topologies were similar, and used the resultant 18,000 trees to calculate posterior probabilities (PP) for each bipartition on a 50% majority rule consensus tree.
Short, blunt head 1.7 times longer than broad and 2.5 times longer than deep; head abruptly distinct from neck, three times wider than thinnest part of neck and also slightly wider than greatest width of body; eye large and protuberant, occupying 27% of length of head, with elliptical pupil visible from anterior, lateral, dorsal, and ventral aspects; rostral 1.6 times wider than high, concave in anterior view, and narrowly visible from above; paired prefrontals extending anteroventrally to level of center of eye, each in contact with its mate and with frontal, supraocular, preocular, nasal, and internasal; frontal pentagonal, 1.6 times longer than wide (greatest width), and about 1.2 times longer than distance from its anterior edge to tip of snout; supraocular anteriorly narrow and posteriorly nearly as wide as greatest frontal width; broad parietals, about 1.3 times longer than wide; interparietal suture 1.2 times longer than length of frontal, and 1.4 times longer than distance from frontal to tip of snout; nasal plate single, centrally pierced by large naris (0.79 mm in diameter), in contact with rostral anteriorly, internasal dorsally, prefrontal posterodorsally, preocular posteriorly, and first and second supralabials ventrally; loreal absent; one large and high preocular; two postoculars (an extra tiny scale on left side ventrally), the lower somewhat less than half the size of the upper; temporals 2+2+3; supralabials 9, first and second in contact with nasal, fourth in contact with preocular, and fourth to seventh bordering the orbit; infralabials 13, with first six in contact with anterior genial, and sixth to eighth touching posterior genial; first pair of infralabials in contact medially behind mental; anterior and posterior genials nearly equal in length; gular scales with posterolateral apical pit.
Color in preservative of holotype (Figs 1,2). Dorsal background light brown, with a longitudinal series of 63 dark brown middorsal blotches from head to cloaca; dark middorsal blotches longer anteriorly, 2-3 vertebral scales long, than posteriorly, 1-2 vertebral scales long, and extending laterally 1-3 (anteriorly) or more (posteriorly) dorsal scale rows, but never reaching ventral scales; each dark middorsal blotch irregularly bordered anteriorly and posteriorly by thin cream line; ventral aspect of body yellowish cream with dark brown spots and flecks; ventral aspect of tail yellowish cream with spots concentrating midventrally; dorsal surface of head light brown with several dark brown spots and two short dark stripes extending from posterior aspect of parietals to a point just posterior to head; ventral surface of head whitish cream.
Hemipenes (Fig. 4). The right hemipenis of the paratype DHMECN 6753 of Imantodes chocoensis was removed, fully everted and expanded (Fig. 4). The organ is bulbous and relatively long, 11.2 mm in length, and when adpressed to the outside of the tail it extends from the cloaca to the sixth subcaudal scale. The organ is longer than wide (width 46% of length), unilobed, symmetrical, calyculate, capitate, and arched towards the sulcal side. The sulcus spermaticus is simple, linear, semicentripetal, and thin, demarcated by thick bordering tissue at the base, particularly at the anterior border, and ending on the surface of the capitulum facing medially. The capitulum is ornamented with papillated calyces, spinulated proximally. The capitulum, approximately 45% the length of the hemipenis, is slightly demarcated by a groove, more prominent on the sulcal side and joining the sulcus spematicus. In the asulcate side the base of the capitulum has more prominent spines. Truncus covered by large spines, on the sulcate and asulcate and surfaces, 23 in total, 13 to the right of the sulcus spermaticus and 10 to the left, and has a few rows of small spines, two at the base on the asulcate side and two to three rows just right of the sulcus spermaticus.
Variation. Intraspecific variation in scale counts and measurements in Imantodes chocoensis sp. n. is presented in Table 2. Color in life of paratypes UTA R-60205 and DHMECN 6753 (Fig. 5) is similar to color in preservative of holotype; iris copper brown. Middorsal blotches from head to cloaca vary between 55-66; one specimen (UTA R-60205) had one middorsal blotch covering five vertebral scales. Distribution and ecology. Imantodes chocoensis inhabits Chocoan rainforests on the Pacific coast in northern Ecuador (Fig. 6). It occurs in lowland evergreen forest (Cerón et al. 1999) at elevations of 115-260 m in the provinces of Carchi and Esmeraldas. This new species has been collected in sympatry with I. cenchoa in Esmeraldas, Pseustes shropshirei, Sibon nebulatus, Synophis bicolor, Tantilla melanocephala, and Xenodon rabdocephalus. The known localities of I. chocoensis lie in close proximity to the Ecuador-Colombia border and we expect for it to be found in neighboring Colombia.
Etymology. The specific epithet chocoensis is an adjective derived from Chocó, the very humid tropical region comprising the Pacific coast of northern Ecuador, Colombia and Panama (Morrone 2001). This region is part of the 274,597 km 2 Tumbes-Chocó-Magdalena hotspot as defined by Conservation International, which includes more than 320 species of reptiles.
Phylogenetic relationships. Selected models of evolution for sampled fragments of ND4 and cyt-b genes were HKY+I+G and TPM2uf+I+G, respectively. The resulting 50% majority rule consensus tree (Fig. 7) supports strongly (PP=1) a sister taxon relationship between Imantodes chocoensis sp. n. and I. lentiferus, as well as the exclusivity (Rieppel 2010) of both species. Similarly, I. inornatus and I. cenchoa are recovered as exclusive clades with strong support (PP=1). Noteworthy the I. cenchoa clade includes samples from Guatemala, Costa Rica, Panama, Brasil, and Colombia, Ecuador and Peru on both sides of the Andes. In agreement with previous hypotheses, I. gemmistratus is paraphyletic; of three samples included in this study, one from Guatemala is sister to the I. cenchoa clade with strong support (PP=1), whereas the other two, from Mexico, are weakly supported as sister to the I. chocoensis sp. n. and I. lentiferus clade. Myers (1982) distinguished two monophyletic groups within Imantodes -lentiferus and cenchoa -based on hemipenial characters, maxillary dentition, relative tongue length, and coloration. According to Myers, the lentiferus group included I. lentiferus and I. phantasma as sister taxa, as well as I. inornatus, whereas the cenchoa group included I. cenchoa, I. gemmistratus and I. tenuissimus. Since the phylogenetic tree presented in this paper does not include all species of Imantodes, we cannot rigorously test Myers' hypothesis of phylogenetic relationships within Imantodes. Nonetheless, two major differences are worth noting. First, in our phylogenetic tree I. inornatus is sister to all other species of Imantodes (but see below). Second, in agreement with previous work (Daza et al. 2009), we recover a paraphyletic I. gemmistratus, with specimens from Guatemala closely related to I. cenchoa as postulated by Myers (1982), and specimens from Mexico in a clade with I. lentiferus and the I. chocoensis sp. n. (Fig. 7).

Discussion
Monophyly of Imantodes remains controversial, but we refrained from testing it without better taxon and character sampling. Previous phylogenetic studies based on DNA sequence data have failed to support the monophyly of Imantodes as currently circumscribed (Daza et al. 2009;Mulcahy 2007). Except for a tree including only two species of Imantodes (Fig. 6 in Daza et al. 2009) and a Maximum Parsimony tree (Fig. 5 in Mulcahy 2007), these studies suggest that I. inornatus is sister to a clade containing Imantodes and Leptodeira as sister taxa. Furthermore, the phylogenetic tree presented in this paper is congruent with this hypothesis (Fig. 7), suggesting that I. inornatus might belong to a clade different from Imantodes. In fact, this species differs from other Imantodes in several morphological (e.g., no prominent dorsal blotches, or conspicuously enlarged vertebral scales; Fig. 3) and behavioral (e.g., head-flaring) features (Mulcahy 2007;Myers 1982). Vertebral scales 2.5-4 times wider than adjacent dorsal scales; dorsal color pattern with conspicuous dark blotches (Fig. 3)  Financial support for part of the field and laboratory expenses was provided by the Instituto Bioclon (Mexico) to ENS.