Description and phylogeny of three new species of Synophis (Colubridae, Dipsadinae) from the tropical Andes in Ecuador and Peru

Abstract The discovery of three new species of Synophis snakes from the eastern slopes of the tropical Andes in Ecuador and Peru is reported. All previous records of Synophis bicolor from eastern Ecuador correspond to Synophis bogerti sp. n., which occurs between 1000–1750 m along a large part of the Amazonian slopes of the Ecuadorian Andes. In contrast, Synophis zamora sp. n. is restricted to southeastern Ecuador, including Cordillera del Cóndor, between 1543–1843 m. Synophis insulomontanus sp. n. is from the eastern slopes of the Andes in central and northern Peru, between 1122–1798 m, and represents the first record of Synophis from this country. All three new species share in common a large lateral spine at the base of the hemipenial body. A molecular phylogenetic tree based on three mitochondrial genes is presented, including samples of Diaphorolepis wagneri. Our tree strongly supports Synophis and Diaphorolepis as sister taxa, as well as monophyly of the three new species described here and Synophis calamitus. Inclusion of Synophis and Diaphorolepis within Dipsadinae as sister to a clade containing Imantodes, Dipsas, Ninia, Hypsiglena and Pseudoleptodeira is also supported.


Introduction
With only four recognized species, Synophis is among the least speciose snake groups formally recognized as genera in South America. Species of Synophis are known to occur in the Andes of Colombia and Ecuador between approximately 460-2200 m (Hillis 1990). Whereas S. plectovertebralis and S. calamitus are endemic to Colombia and Ecuador, respectively (Hillis 1990;Sheil and Grant 2001), S. bicolor and S. lasallei have been reported in both countries (Bogert 1964;Nicéforo-María 1970).
The taxonomic identity of specimens currently assigned to Synophis bicolor (Peracca 1896) has been problematic for two reasons. First, the type locality of this species is ambiguous ('America meridionale') preventing the collection of topotypes for comparison. Second, there is significant morphological variation between specimens of S. bicolor from Colombia and Ecuador. In his taxonomic review of Synophis and Diaphorolepis Bogert (1964) noted some differences between specimens from Ecuador and the holotype of S. bicolor (in parentheses): 10-11 infralabials (9), 160-166 ventrals (180), 100-118 subcaudals (136), 24-27 maxillary teeth (16), 14 palatine teeth (9-10), 32-34 pterygoid teeth (21)(22). Based on this variation, Bogert (1964) recognized that "specimens tentatively referred to S. bicolor might not be conspecific". Subsequently, Niceforo-María (1970) reported the first specimen of S. bicolor from Colombia and noted that the numbers of ventrals and subcaudals (184 and 127, respectively) are more similar to the holotype than the specimens from Ecuador. This suggests that at least some populations from Ecuador currently assigned to S. bicolor represent one or more similar undescribed species instead.
The study of Synophis has been hampered by the paucity of specimens in collections, possibly because of low densities or semifossorial habits (Sheil and Grant 2001). Recent collections in poorly explored areas of the Amazonian slopes of the Andes from Ecuador and Peru yielded a few specimens of Synophis that are similar in morphology to specimens of S. bicolor previously reported from eastern Ecuador (Bogert 1964). Based on these recent collections, including the first specimens of Synophis from Peru, we combine evidence from morphology and phylogenetic analyses of DNA sequence data to describe three new species of Synophis.

Morphological data
All type specimens of the new species are deposited at Museo de Zoología, Pontificia Universidad Católica del Ecuador, Quito (QCAZ); and Centro de Ornitología y Biodiversidad (CORBIDI), Lima, Peru. Other specimens used for comparisons are listed in Appendix I. Sex was determined by observation of hemipenes from X-ray images or by noting the presence of everted hemipenes. Snout-vent length (SVL) and tail length were measured with a ruler and recorded to the nearest millimeter. Other measurements were taken with digital calipers (±0.01 mm). We prepared partially everted hemipenes following Zaher and Prudente (2003), and immersed them for 6 h in an alcoholic solution of Alizarin Red to dye the calcareous ornaments (e.g., spines). Terminology for hemipenis description follows Dowling and Savage (1960), as augmented by Zaher (1999). Data on the hemipenes of Synophis calamitus and S. lasallei were taken from the literature (Zaher 1999).

Molecular 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.
Additionally, we obtained from GenBank sequences of 12 Dipsadinae taxa and Natrix natrix, which was used to root the tree following the phylogenetic hypothesis presented by Pyron et al. (2013). We only selected those Dipsadinae species that had sequence data for all three genes included in our analyses. Gene regions of outgroup taxa included in phylogenetic analyses along with their GenBank accession numbers are shown in Table 2.

Phylogenetic analyses
Data were assembled and aligned in Geneious v7.1.7 (Kearse et al. 2012) under default settings for MAFFT (Katoh and Toh 2010). ND4 and cyt-b sequences were translated into amino acids for confirmation of alignment. The best-fit nucleotide substitution models and partitioning scheme were chosen simultaneously using PartitionFinder v1.1.1 (Lanfear et al. 2012) under the Bayesian Information Criterion (BIC). The "greedy" algorithm was used with branch lengths of alternative partitions "linked" to search for the best-fit scheme, which consisted of three partitions: (i) 16S, 3 rd codon positions of both cyt-b and ND4 [GTR + I + G]; (ii) 2 nd codon positions of both cytb and ND4 [K81uf+G]; and (iii) 1 st codon positions of both cyt-b and ND4 [HKY + I + G]. Bayesian inference was used to obtain a phylogenetic tree of the combined dataset using the program MrBayes v3.2.1 (Ronquist et al. 2012). All parameters were unlinked between partitions (except topology and branch lengths), and rate variation (prset ratepr = variable) was invoked. Four independent runs, each with four MCMC chains, were run for five million generations, sampling every 1,000 generations. Results were analyzed in Tracer to assess convergence and effective sample sizes (ESS) for all parameters. Additionally, we verified that the average standard deviation of split frequencies between chains and the potential scale reduction factor (Gelman and Rubin 1992) of all the estimated parameters approached values of ≤ 0.01 and 1, respectively. Of the 5,000 trees resulting per run, 25% were arbitrarily discarded as "burn-in". The remaining trees were used to calculate posterior probabilities (PP) for each bipartition in a 50% majority-rule consensus tree. The phylogenetic tree was visualized and edited using FigTree v1.4.2 (Rambaut 2014).

Results
The taxonomic conclusions of this study are based on the observation of morphological features and color patterns, as well as inferred phylogenetic relationships. We consider this information as species delimitation criteria following a general lineage or unified species concept (de Queiroz 1998;2007).

Taxon
Genbank number a Sequences extracted from whole mitochondrial genomes.  Table 3.
Description of the holotype. Adult male (Figs 1, 2), SVL 367 mm; tail length 184 mm; eye diameter 1.17 mm; pupil round; head width 6.32 mm at level of supralabial 6; and head length 11.7 mm from snout to posterior margin of jaw; width at midbody 5.19 mm; head distinct from neck.
Prefrontals fused in a rectangular scale, wider than long; frontal single, with an incomplete suture from anterior margin to the middle of the scale, heptagonal, slightly wider than long; parietals large, paired, longer than wide; loreal trapezoidal, almost two times longer than high; preocular single, bordering anterior margin of orbit; supraocular single, bordering dorsal margin of orbit; temporals 1+2; anterior temporal more than two times longer than high; posterior temporals two times longer than high, approximately one half the length of anterior temporal; internasals in contact medially, distinctly wider than long; nasals not in contact; rostral visible from above, concave, nearly two times wider than long, in contact with first supralabials, nasals, and internasals; mental triangular, in contact with first pair of infralabials; infralabials 10/11; supralabials 8/8 (fourth and fifth entering orbit on both sides); anterior genials three times longer than wide, bordered laterally by infralabials 1-5 on right side, 1-6 on left side; posterior genials two times longer than wide, in contact anteromedially and separated by two gulars posteriorly, and bordered laterally by infralabials 5-6 on right side and 6-7 on left side; dorsal scale rows 19-19-17, first dorsal row weakly keeled from ventral 118, other rows strongly keeled; anal single; ventrals 163; subcaudals 115, paired. Type specimen data from Hillis (1990); 3 Data from Hillis (1990); 4 Data from Sheil and Grant (2001);  Hemipenial morphology. The following description is based on the right hemipenis of the holotype ( Fig. 3; QCAZ 12791). The fully everted and maximally expanded organ is bilobed, semicalyculate, semicapitate, and extends to the sixth subcaudal. Capitular grooves are on the asulcate side; capitula are ornamented with calcified papillae, larger on the asulcate side. Numerous larger papillae meet on the asulcate side of the lobular crotch. On the sulcate side, the capitula extend along the sides of the branches of the sulcus spermaticus, far down the hemipenial body. The sulcus spermaticus bifurcates on the proximal half of the body and its branches extend centrolineally to the tip of each lobe. The hemipenial body is ornamented with large calcified spines, except on the medial region of the asulcate side, where the spines are small. The spines increase in length towards the base of the hemipenial body, with one spine on the left side (sulcate view) being considerably longer than the others. The base of the hemipenial body bears much smaller and scattered spines.
Color in preservative of the holotype (Figs 1, 2). Dorsal surface of head, body and tail uniformly dark grey; skin among scales on flanks cream, visible on anterior half of body; ventrals mostly cream on anterior end of body (ventrals 1-10), becoming progressively pigmented with light grey posteriorly except on their margins; anal plate cream medially and grey laterally; subcaudals with cream margins and same tone of grey as posterior ventrals; sides of head same tone of grey as dorsal surface, except for labials, which are mostly cream ventrally; chin cream with light grey anterior margin (most of mental and first three pairs of infralabials).
Variation. Intraspecific variation in scale counts and measurements in Synophis bogerti is presented in Table 3. Keeling on the first row of dorsals starts on ventrals 5, 10, 87, 98 and 114 in paratypes QCAZ 5072, 13323, 3511, 13585 and 11070, respectively. Besides the holotype, male paratype QCAZ 5072 is the only specimen with an incomplete medial suture on the frontal scale. This condition was also reported and illustrated by Bogert (1964) in a specimen (UMMZ 91550) from eastern Ecuador, referred by him as S. bicolor and recognized by us as S. bogerti (Fig. 4).
Distribution and natural history. Synophis bogerti occurs along the Amazonian slopes of the Andes in central Ecuador at elevations between 1000-1750 m (Fig. 6). The type locality is part of Wildsumaco Wildlife Sanctuary, a 400 ha reserve consisting of primary and secondary forests in a matrix of agricultural land. Most localities where S. bogerti was collected lie within protected areas including two large national parks (Llanganates and Sangay), indicating that at least some populations of this species are protected. All specimens were found active at night (20h45-00h00), mostly on the ground or on shrubs 0.5 m above ground.
Etymology. The specific epithet bogerti is a noun in the genitive case and is a patronym for Charles M. Bogert (1908Bogert ( -1992, an American herpetologist and former curator of the American Museum of Natural History. Among his many contributions, Bogert published a systematic revision of Diaphorolepis and Synophis, in which he recognized that "It is also possible, of course, that specimens tentatively referred to S. bicolor are not actually conspecific" (Bogert 1964: 517). Specimens of "S. bicolor" from eastern Ecuador examined by Bogert (1964)  Diagnosis. Synophis zamora can be distinguished from other species of Synophis by having a noncapitate, bilobed hemipenis with a large lateral spine at the base of the hemipenial body (Fig. 9); 19 longitudinal rows of dorsals at midbody; strongly keeled dorsals except for first row, which is weakly keeled (at least posteriorly); and 147-153 ventrals in males. Scutellational characters of all recognized species of Synophis are presented in Table 3.
Description of the holotype. Adult male (Figs 7, 8); SVL 349 mm; tail length 185 mm; eye diameter 1.34 mm; pupil round; head width 5.15 mm at level of supralabials 6 and 7; head length 11.05 mm from snout to posterior margin of jaw; width at midbody 5.06 mm; head distinct from neck. Prefrontals fused in a rectangular scale, wider than long; frontal single, heptagonal, slightly wider than long; parietals large, paired, longer than wide; loreal trapezoidal, two times longer than high; preocular single, bordering anterior margin of orbit; supraocular single, bordering dorsal margin of orbit; temporals 1+2; anterior temporal more than two times longer than high; posterior temporals longer than high, approximately one half the length of anterior temporal; internasals in contact medially, distinctly wider than long; nasals not in contact; rostral visible from above, concave, two times wider than long, in contact with first supralabials, nasals, and internasals; mental triangular, in contact with first pair of infralabials; infralabials 10/10; supralabials 9/9 (fourth, fifth and sixth entering orbit on both sides); anterior genials almost three times longer than wide, bordered laterally by infralabials 1-5; posterior genials three times longer than wide, in contact anteromedially and separated by three gulars posteriorly, and bordered laterally by infralabials 5-6; dorsal scale rows 19-19-17, first row weakly keeled from 15 th ventral, other rows strongly keeled; anal single; ventrals 147; subcaudals 103, paired.
Hemipenial morphology. The following description is based on the right hemipenis of the holotype ( Fig. 9; QCAZ 9174). The fully everted and maximally expanded organ is bilobed, semicalyculate, noncapitate, and extends to the sixth subcaudal. Each lobe is ornamented with small calcified papillae, slightly larger on the asulcate and lateral sides and more scattered on the sulcate side. Some larger lobular papillae meet medially at the lobular crotch on the asulcate side. The sulcus spermaticus bifurcates on the proximal half of the body and its branches extend centrolineally to the tip of each lobe. The hemipenial body is ornamented with medium-sized calcified spines, except on the medial region of the asulcate side, where the spines are small. The spines increase in length towards the base of the hemipenial body, with one spine on the left side (sulcate view) being considerably longer than the others. The base of the hemipenial body bears much smaller and scattered spines.  (Figs 7, 8). Dorsal surface of head, body and tail uniformly dark grey; skin among dorsal scales cream, visible on anterior half of body; ventrals cream on anterior end of body (ventrals 1-5), becoming progressively pigmented with light grey posteriorly except on their margins; anal plate cream posteriorly and grey anteriorly; subcaudals with cream margins and same tone of grey as posterior ventrals; sides of head same tone of grey as dorsal surface, except for labials, which are mostly cream; chin cream with light grey anterior margin (most of mental and first two pairs of infralabials).

Color in preservative of the holotype
Variation. Intraspecific variation in scale counts and measurements in Synophis zamora is presented in Table 3. Keeling on the first row of dorsals starts on ventrals 9, 10, and 105 in paratypes 9175, 13854, and 12773, respectively. No major differences were found between the hemipenis of the holotype and those of paratypes QCAZ 12773 and 13854. Coloration in life (QCAZ 13854; Fig. 5) is the same as that described for the holotype above, except that the cream color has a light yellow tint.
Distribution and natural history. Synophis zamora occurs in the southeastern portion of the northern Andes in Cordillera del Cóndor and the Amazonian slopes of the Andes at elevations between 1543-1843 m (Fig. 6). It is known from localities close to the Bombuscaro and Nangaritza rivers, which are tributaries of the Zamora river. These localities lie in Ecuador within protected areas, such as Podocarpus National Park and Numbami Ecological Reserve, indicating that at least some populations of S. zamora are protected. All specimens were found active at night (20h30-00h00), mostly on the ground or on shrubs 1-1.5 m above ground. One specimen was found on a boulder covered with moss.
Etymology. The epithet zamora is a noun in apposition and refers to both the Zamora river and the province of Zamora Chinchipe. All type specimens were collected in this province along the upper basin of Zamora river.  Diagnosis. Synophis insulomontanus can be distinguished from other species of Synophis by having a semicapitate, bilobed hemipenis with a large lateral spine at the base of the hemipenial body, and the sulcus spermaticus bifurcating on the center of the hemipenial body (Fig. 12); 19 longitudinal rows of dorsals at midbody; strongly keeled dorsals except for first row, which is keeled to a lesser extent; 151-152 ventrals in males, 147-149 in females; 108-109 subcaudals in males, 103 in females. Scutellational characters of all recognized species of Synophis are presented in Table 3.

Synophis insulomontanus
Description of the holotype. Adult male (Figs 5, 10, 11), SVL 335.3 mm; tail length 180.9 mm; eye diameter 1.46 mm; pupil round; head width 6.7 mm at level of supralabial 6; head length 11.07 mm from snout to posterior margin of jaw; width at midbody 6.48 mm; head distinct from neck.
Prefrontals fused in a roughly pentagonal scale, wider than long; frontal single, pentagonal, posterior suture angular with apex directed posteriorly, wider than long; parietals large, paired, longer than wide; loreal trapezoidal, almost two times longer than high; preocular single, bordering anterior margin of orbit; supraocular single, bordering dorsal margin of orbit; temporals 1+3+3; anterior temporal more than two times longer than high; posterior temporals two times longer than high, approximately one half the length of anterior temporal; internasals in contact medially, wider than long; nasals not in contact; rostral visible from above, concave, nearly two times wider than long, in contact with first supralabials, nasals, and internasals; mental triangular, in contact with first pair of infralabials; infralabials 11/11; supralabials 8/8 (fourth and fifth entering orbit on both sides); anterior genials three times longer than wide, bordered laterally by infralabials 1-6 on both sides; posterior genials two times longer than wide, separated by gulars, and bordered laterally by infralabials 6-7 on both sides; dorsal scale rows 20-19-19, first dorsal row moderately keeled from ventral 7, other rows strongly keeled; anal single; ventrals 151; subcaudals 108, paired.
Hemipenial morphology. The following description is based on the left hemipenis of the holotype ( Fig. 12; CORBIDI 13940). The fully everted and maximally expanded organ is bilobed, semicalyculate, semicapitate, and extends to the fifth subcaudal. Capitular grooves are present on the asulcate side; capitula are ornamented with calcified papillae, remarkably larger on the asulcate side. A few papillae meet on the asulcate side of the lobular crotch. The sulcus spermaticus bifurcates on the center of the hemipenial body and its branches extend centrolineally to the tip of each lobe. Papillae are relatively small on the sulcate side of the hemipenial body between the bifurcating branches of the sulcus spermaticus. The hemipenial body is ornamented with large calcified spines, except on the medial region of the asulcate side and near the sulcus spermaticus, where the spines are very small. One spine on the left side (sulcate view) is considerably longer than the others. Very small spines cover the base of the hemipenial body.
Color in life of the holotype (Fig. 5). Dorsal surface of head, body and tail uniformly dark grey; skin among scales on flanks cream, visible on anterior half of body; first five ventrals cream, becoming progressively pigmented with grey, except on their posterior margin where cream pigmentation is always present; anal plate grey with cream posterior border; subcaudals grey, with the porsterior borders weakly pigmented with cream in some scales; sides of head and 1 st supralabial same tone of grey as dorsal surface, other supralabials mostly cream; first three infralabials mostly grey, others mostly cream; scales on throat with a pale blue tone.
Variation. Intraspecific variation in scale counts and measurements in Synophis insulomontanus is presented in Table 3. Two or three scales can be present on second row of temporals, three in the holotype and CORBIDI 9223, and two in CORBIDI 10418 and CORBIDI 13705. Paratype CORBIDI 10418 has 21 dorsals at midbody. No major differences were found between the hemipenis of the holotype and that of paratype CORBIDI 13705, except that the latter has more papillae between the bifurcating branches of the sulcus spermaticus on the sulcate side of the hemipenial body.
Specimen CORBIDI 10418 has a dense cream pigmentation on ventrals from anterior end of body to midbody.
Distribution and natural history. Synophis insulomontanus is known to occur between 1122-1798 m on the Amazonian slopes of the Andes in northern and central Peru (Fig. 6). Two localities within Departamento Huánuco, Cordillera Azul and Cordillera El Sira, correspond to sub-Andean mountain ridges, whereas the locality of Miraflores lies on the Amazonian slopes next to the Huallaga River.
The holotype was found at night, coiling inside a bromeliad, 1 m above the ground in primary premontane forest. Other specimens were found active at night, moving through leaf litter. Specimens from Cordillera Azul (CORBIDI 9223 and 10418) were found in primary premontane forest, whereas specimen CORBIDI 13705 from Miraflores, Huánuco, was found in secondary montane forest.
Etymology. The epithet insulomontanus is a noun that derives from the Latin words insulo (= isolated) and montanus (= mountain). It refers to the isolated mountain ridges in Departamento Huánuco, where the new species was discovered.
Within Synophis there is a basal split into two clades, one (PP = 1) containing the trans-Andean taxon S. calamitus, and the other (PP = 0.88) including the three cis-Andean species described in this paper (S. bogerti, S. insulomontanus and S. zamora). Within the cis-Andean clade, S. bogerti and S. zamora are recovered as sister species with maximum support (PP=1), forming a clade sister to S. insulomontanus.

Phylogeny of Synophis and Diaphorolepis
In spite of recent efforts to resolve the phylogenetic relationships of dipsadid snakes using DNA sequence data (e.g., Grazziotin et al. 2012;Pyron et al. 2013;Zaher et al. 2009), Synophis and Diaphorolepis have remained unsampled. Consequently, they have been considered as Dipsadidae incertae sedis (Zaher et al. 2009). In order to have a general idea of the phylogenetic position of both Synophis and Diaphorolepis, we included in our analysis 12 additional dipsadinae taxa used in previous phylogenetic studies (e.g., Pyron et al. 2013). We did not attempt to perform a taxonomically extensive phylogenetic analysis of Dipsadidae or Dipsadinae (sensu Pyron et al. 2013); instead, we preferred to include in our analysis only those species of Dipsadinae, for  which all gene regions used in this study were available in GenBank. Our phylogeny strongly supports (1) inclusion of Diaphorolepis and Synophis within Dipsadinae; (2) a close relationship between these genera and a clade including Imantodes, Dipsas, Ninia, Hypsiglena and Pseudoleptodeira; and (3) a close relationship between Diaphorolepis and Synophis, as has been hypothesized using morphological evidence (Hillis 1990).
All species of Synophis are known to occur on Andean slopes in Colombia and Ecuador, with S. insulomontanus sp. n. representing the first record from Peru. Along with S. lasallei, the three species described in this paper are restricted to Amazonian slopes of the Andes, except for one record of S. lasallei from the western slopes of the eastern Cordillera in Colombia (Fig. 6). Thus, the Andes represent a major geographic barrier separating species of Synophis. Recent studies on other reptile taxa (e.g., Alopoglossus, Enyalioides) with similar distributions suggest that the uplift of the northern Andes represents a major vicariant event explaining their radiation and present distribution (Torres-Carvajal and de Queiroz 2009; Torres-Carvajal and Lobos 2014). We could not test this hypothesis with Synophis because we had no access to tissue samples of S. bicolor, S. plectovertebralis and S. lasallei. Nonetheless, based on morphological similarity (e.g., strongly keeled dorsals, first row of dorsals keeled; Table 3), it is likely that S. lasallei is nested in the same clade with the three eastern-Andean species described in this paper. Vertebral morphology, not examined in most species of Synophis, seems to support this idea. Bogert (1964) noted that the vertebrae of S. "bicolor" (= S. bogerti sp. n.) and S. lasallei were similar in morphology, which is different from at least S. plectovertebralis (Sheil and Grant 2001). We examined superficially trunk vertebrae of S. bogerti and S. zamora using digital X-rays, and found that the vertebrae of both species are very similar and agree with the description presented by Bogert (1964) in that "vast expansions of the prezygapophyses and postzygapophyses are coalesced as projections with relatively straight margins parallel to the main axis of each vertebra". In addition, the zygapophyseal foramen is largely ossified as opposed to the same foramen in S. plectovertebralis (Sheil and Grant 2001). In conclusion, external and internal anatomy supports both the hypothesis presented above and the idea of a radiation of Synophis east of the Andes. Hillis (1990) described Synophis calamitus based on two specimens. Among other characters, he proposed that the number of postoculars and whether the internasals are in contact or not were useful taxonomic characters. According to Hillis (1990), S. calamitus differed from other species of Synophis in having one postocular (two in other species) and internasals separated by rostral and prefrontal (internasals in contact medially in other species). Among 12 specimens of S. calamitus examined in this study (Appendix I), nine have two postoculars on each side, two have one postocular on one side and two on the opposite side, and only one specimen (QCAZ 11931) has one postocular on each side. Moreover, we were able to examine the paratype of S. calamitus (KU 164208), a juvenile, badly-crushed roadkill, and found out that this specimen has one postocular on the left side and two on the right side, the ventral one difficult to observe because of the condition of the specimen. Thus, the number of postoculars is variable in S. calamitus and, therefore, it is not a useful taxonomic character. Regarding the contact between internasals, all specimens examined except for one (QCAZ 5847) had internasals in contact medially, as opposed to the condition described for both the holotype and paratype (internasals separated ;Hillis 1990). Specimen QCAZ 5847 is a roadkill collected in the northern province of Carchi, and does not seem to have other differences with the remaining specimens of S. calamitus examined in this study. However, given the large branch separating this specimen from all other specimens of S. calamitus in the phylogenetic tree (Fig. 13), as well as its disjunct distribution (Fig.  6), we believe that the taxonomic status of northern (Carchi) populations should be addressed in more detail.