﻿A consolidated phylogeny of snail-eating snakes (Serpentes, Dipsadini), with the description of five new species from Colombia, Ecuador, and Panama

﻿Abstract A molecular phylogeny of the Neotropical snail-eating snakes (tribe Dipsadini Bonaparte, 1838) is presented that includes 60 of the 133 species currently recognized. There is morphological and phylogenetic support for four new species of Sibon Fitzinger, 1826 and one of Dipsas Laurenti, 1768, which are described here based on their unique combination of molecular, meristic, and color pattern characteristics. PlesiodipsasHarvey et al., 2008 is designated as a junior synonym of Dipsas and additional evidence is presented to support the transfer of the genus Geophis Wagler, 1830 to the tribe Dipsadini. Two of the subspecies of S.nebulatus (Linnaeus, 1758) are elevated to full species status. Insight into additional undescribed cryptic diversity within the S.nebulatus species complex is provided. Evidence that supports the existence of an undescribed species previously confused with D.temporalis is provided, as well as the first country record of S.ayerbeorum Vera-Pérez, 2019 in Ecuador with a comment on the ontogenetic variation of the latter. Finally, photographs of Colombian, Ecuadorian, and Panamanian snail-eating snakes are provided.


Introduction
The snail-eating snake tribe Dipsadini is one of the most diverse yet taxonomically complex group of snakes in the Neotropics. Many authors (Peters 1960;Downs 1961;Hoge 1964;Peters and Orejas-Miranda 1970;Kofron 1982;Orcés and Almendáriz 1987;Porto and Fernandes 1996;Fernandes et al. 1998Fernandes et al. , 2002Cadle and Myers 2003;Passos et al. 2004Passos et al. , 2005Cadle 2005Cadle , 2007Harvey 2008;Harvey and Embert 2008;) have attempted to clarify the systematics of the group or its subgroups using morphological characters. However, the majority of these authors disagree about the number of genera included in the tribe as well as the allocation of species among genera. Fortunately, this lack of consensus is likely coming to an end with the use of molecular tools on Dipsadini systematics. Three independent groups of researchers (Sheehy 2012;Arteaga et al. 2018;and Grünwald et al. 2021) arrived at similar conclusions about the contents and limits of this snake tribe when molecular and morphological evidence were combined. Although progress is being made regarding the taxonomy of Dipsadini, more than half of the species of the group remain unsampled for DNA characters and additional diversity remains undescribed.
In his unpublished PhD thesis, Sheehy (2012) presented a large (194 taxa) phylogeny of the group using two mitochondrial ant two nuclear genes. Among his most important findings were that the genus Sibynomorphus Fitzinger, 1843 is paraphyletic with respect to Dipsas and that the genus Geophis Wagler, 1830 is deeply nested within Dipsadini. He also uncovered two major (one South American and one Central American) and eleven minor geographically structured clades within the widely distributed snake species Sibon nebulatus and found that Sibon annulatus (Günther, 1872) is paraphyletic with respect to Sibon lamari Solórzano, 2001 and Sibon perissostichon Köhler et al., 2010. Sheehy (2012 also presented phylogenetic evidence that Geophis sanniolus (Cope, 1866) (previously in the genus Sibon) and D. gaigeae (Oliver, 1937) do not belong to their nominal genera and their relationships with the remaining groups of snail-eating snakes are ambiguous. Since Sheehy's work was not published, his findings were not integrated into the taxonomy of the Dipsadini. However, his work provided a solid framework for comparison for two subsequent studies.
First, Arteaga et al. (2018) presented a phylogeny of the group based on novel taxon sampling. This work differed from Sheehy's in that it had an emphasis on Dipsas (rather than Sibon) and on South American (rather than Central American) Dipsadini in general. However, it confirmed some of the results of Sheehy (2012); most notably the paraphyly of Dipsas with respect to Sibynomorphus, which resulted in the latter being designated as a junior subjective synonym of Dipsas. Besides describing five new species of Dipsadini, Arteaga et al. (2018) uncovered, but did not explore further, high levels of intraspecific divergence within each of the nominal species D. vermiculata Peters, 1960 and Sibon annulatus, which suggested that further diversity remained to be described in this group. They also presented evidence that Sibon nebulatus is paraphyletic with respect to Sibon dunni Peters, 1957 and Sibon bevridgelyi Arteaga et al., 2018, and recognized that elevating the subspecies Sibon nebulatus leucomelas (Boulenger, 1896) and Sibon nebulatus hartwegi Peters, 1960 to full species status would help resolve this paraphyly. However, Arteaga et al. (2018) refrained from proposing further taxonomic arrangements because their sample size for D. vermiculata, Sibon annulatus, and Sibon nebulatus was insufficient. Later, Grünwald et al. (2021) combined some of the DNA sequences of Sheehy (2012) with novel sequences from Mexico into a phylogeny of Dipsadini that focused on the genus Tropidodipsas Günther, 1858. Based on the results of their phylogenetic analyses, these authors transferred T. annuliferus (Boulenger, 1894), T. sartorii (Cope, 1863), and Sibon sanniolus to Geophis, a genus that, according to Sheehy (2012), should be added to the tribe Dipsadini, a decision seconded by Grünwald et al. (2021) and also herein. Lastly, in a project seeking to create a large DNA barcode library of reptiles from the National Museum of Natural History tissue holdings, Mulcahy et al. (2022) provided mitochondrial DNA sequences (gene fragments COI and 16S) for ten species of Dipsadini. However, these have not been included in any phylogenetic studies so far.
Here, we combine the datasets of Sheehy (2012), Arteaga et al. (2018), Grünwald et al. (2021), and Mulcahy et al. (2022) with novel DNA sequences of Colombian, Panamanian, and Ecuadorian material into a consolidated phylogeny of the tribe Dipsadini. Notably, we include the recently described Sibon ayerbeorum and the monotypic Plesiodipsas perijanensis (Aleman, 1953) in the analysis. The combined molecular sampling, together with morphological analysis and species distribution models, supports the existence of at least five new species of Neotropical snail-eating snakes, which we describe here.

Ethics statement
This study was carried out in strict accordance with the guidelines for use of live amphibians and reptiles in field research (Beaupre et al. 2004) compiled by the American Society of Ichthyologists and Herpetologists (ASIH), the Herpetologists' League (HL) and the Society for the Study of Amphibians and Reptiles (SSAR). All procedures with animals (see below) were reviewed by the Ministerio del Ambiente, Agua y Transición Ecológica (MAATE), Ecuador and UNARGEN-Ministerio de Ambiente Panamá, and specifically approved as part of obtaining the following field permits for research and collection: MAE-DNB-CM-2018-0105 and MAATE-DBI-CM-2022-0245 (granted to Universidad San Francisco de Quito) and SC/A-8-09, SC/A-28-09, SC/A-37-11, SC/A-33-12, SE/A-60-16, and SE/A-33-18 (granted to Museo Herpetológico de Chiriquí). Specimens were euthanized with 20% benzocaine, fixed in 10% formalin or 90% ethanol, and stored in 70% ethanol. Museum vouchers were deposited at Museo de Zoología de la Universidad San Francisco de Quito (ZSFQ), Museo Herpetológico de Chiriquí (MHCH), and at the Senckenberg Forschungsinstitut Frankfurt (SMF). Specimens labeled TH, SC, and JMG were also deposited at ZSFQ.

Common names
Criteria for common name designation are as proposed by Caramaschi et al. (2006) and Coloma andGuayasamin (2011-2017), reviewed by Arteaga et al. (2019). These are as follows (in order of importance): (i) the etymological intention (implicit or explicit) that the authors used when naming the species (specific epithet); (ii) a common name that is already widely used in the scientific literature; (iii) a common name that has an important ancestral or cultural meaning; (iv) a common name based on any distinctive aspect of the species (distribution, morphology, behavior, etc).

Morphological data
Our terminology for Dipsadini cephalic shields follows proposals by Peters (1960) and Harvey and Embert (2008). Diagnoses and descriptions generally follow Fernandes et al. (2010) and ventral and subcaudal counts follow Dowling (1951). We physically examined comparative alcohol-preserved specimens from the herpetology collections at Colección de Prácticas Zoológicas de la Universidad del Valle (CPZ-UV), Colección Zoológica de la Universidad ICESI (CZI), División de Herpetología del Instituto Nacional de Biodiversidad (DHMECN), MHCH, Museum d'Histoire Naturelle de la Ville de Genève (MHNG), Museo de Zoología de la Universidad del Azuay (MZUA), Museo de Zoología de la Universidad Tecnológica Indoamérica (MZUTI), SMF, Colección de Anfibios y Reptiles de la Universidad del Valle (UV-C), and ZSFQ (Suppl. material 1). We also examined photographs of specimens housed at Museo de Zoología de la Pontificia Universidad Católica del Ecuador (QCAZ). Morphological measurements were taken with measuring tapes to the nearest 1 mm, or with digital calipers to the nearest 0.1 mm. Abbreviations are as follows: snout-vent length (SVL); tail length (TL); total length, TOL (SVL + TL). Sex was determined by establishing the presence/absence of hemipenes through a subcaudal incision at the base of the tail unless hemipenes were everted.

Laboratory techniques
Genomic DNA was extracted from 96% ethanol-preserved tissue samples (liver, muscle tissue, or scales) using either a guanidinium isothiocyanate extraction protocol (Peña-Five new species of snail-eating snakes from Colombia, Ecuador, and Panama fiel et al. 2020), or a modified salt precipitation method based on the Puregene DNA purification kit (Gentra Systems). The nucleotide sequences of the primers and the PCR conditions applied to each primer pair are detailed in Appendix 2. PCR products were cleaned with either ExoSAP-IT (Affymetrix, Cleveland, OH), or Exonuclease I and Alkaline Phosphatase (Illustra ExoProStar by GE Healthcare) before they were sent to Macrogen Inc (Seoul, South Korea) for sequencing. All PCR products were sequenced in both forward and reverse directions with the same primers that were used for amplification. The edited sequences were deposited in GenBank (Appendix 1).

DNA phylogenetic analyses
A total of 343 DNA sequences was used to build a phylogenetic tree of the tribe Dipsadini, of which 35 were generated during this work and 308 were downloaded from GenBank, most of which were produced by Sheehy (2012), Arteaga et al. (2018), and Grünwald et al. (2021). Of these, 20 sequences are 242-473 bp long fragments of the 12S gene, 65 are 201-422 bp long fragments of the 16S gene, 16 are 493-657 bp long fragments of the COI gene, 85 are 559-1,071 bp long fragments of the CYTB gene, 80 are 325-684 bp long fragments of the ND4 gene, 29 are 606-674 bp long fragments of the DNAH3 gene, and 48 are 456-470 bp long fragments of the NT3 gene. New sequences were edited and assembled using the program Geneious ProTM 2021.1.1 (Drummond et al. 2021) and aligned with those downloaded from GenBank (Appendix 1) using MAFFT v. 7 (Katoh and Standley 2013) under the default parameters in Geneious ProTM 2021.1.1. Genes were combined into a single matrix with 17 partitions, one per non-coding gene and three per protein coding gene corresponding to each codon position. The best partition strategies along with the best-fit models of evolution were obtained in Partition-Finder 2.1.1 (Lanfear et al. 2016) under the Bayesian information criterion.
Phylogenetic relationships were assessed under a Bayesian inference (BI) approach in MrBayes 3.2.0 (Ronquist and Huelsenbeck 2013). Four independent analyses were performed to reduce the chance of converging on a local optimum. Each analysis consisted of 20,000,000 generations and four Markov chains with default heating settings. Trees were sampled every 1,000 generations and 25% of them were arbitrarily discarded as ''burn-in." The resulting 15,000 saved trees per analysis were used to calculate posterior probabilities (PP) for each bipartition in a 50% majority-rule consensus tree. We used Tracer 1.6 (Rambaut et al. 2022) 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 (PSRF) of all the estimated parameters approached values of ≤ 0.01 and 1, respectively. GenBank accession numbers are listed in Appendix 1.

Distribution maps and ecological niche models
We present ranges of occurrence for eleven species of Dipsadini, including five new species described here. Presence localities are derived from museum vouchers (Suppl. materal 1), photographic records (iNaturalist), and the literature (all summarized in Suppl. materal 2). For each species, a binary environmental niche model (ENM) accompanies the dot maps. These models estimate potential areas of distribution on the basis of observed presences and a set of environmental predictors (Elith and Leathwick 2009). To delimit the occupancy areas and the potential species distribution, we used the BAM diagram proposal (Soberón and Peterson 2005;Peterson et al. 2011). To create the models, we used presence localities listed in Suppl. materal 2, 19 bioclimatic variables from Worldclim 1.4 (Hijmans et al. 2005), and Maxent 3.4.1k, an algorithm based on the principle of maximum entropy (Phillips et al. 2006;Elith et al. 2011;Renner and Warton 2013).
For the first explorative exercise, we used the 19 climate layers from the World-Clim project and assessed which variables were the most important for the model, according to the Jackknife test calculated in MaxEnt (Royle et al. 2012). Correlated environmental variables (r < 0.8) were identified using the PEARSON correlation test of PAST 3. In a second modelling exercise, we used the locality records for each species and the variables identified in the first approach to generate the species distribution. 5,000 iterations were specified to the program with clamping and no extrapolation. All other parameters in MaxEnt were maintained at default settings. To create the binary environmental niche models, suitable areas were distinguished from unsuitable areas by setting a minimum training presence threshold value. The logistic format was used to obtain the values for habitat suitability (continuous probability from 0 to 1), which were subsequently converted to binary presence-absence values on the basis of the established threshold value, defined herein as the minimum training presence. The convergence threshold was set to 10 -5 , maximum iterations to 500, and the regularization parameter to "auto."

Molecular phylogeny and taxonomic consequences
Selected partitions and models of evolution are presented in Table 1. We consider strong support for a clade when Bayesian analyses yield posterior probability values > 95%, following Felsenstein (2004). The topology and support ( Fig. 1) of our phylogenetic tree differs from that of Sheehy (2012), Arteaga et al. (2018), and Grünwald et al. (2021), primarily regarding the relationships between the included genera. Below, we outline these differences and comment on the phylogenetic position of new material included in this work. Tropidodipsas fischeri Boulenger, 1894 is recovered as sister to all other sampled Dipsadini, with the exception of Geophis sanniolus, a species that did not form a group  with the remaining Geophis. Neither of these or any other higher relationships within Dipsadini are strongly supported in our analysis, but all other sampled species of Dipsadini are included in their corresponding genera sensu Arteaga et al. (2018) and Grünwald et al. (2021). With the exception of Geophis sanniolus, relationships within Geophis are identical to those presented in Grünwald et al. (2021). The relationships within Tropidodipsas are similar to those presented in Sheehy (2012) and the newly described T. tricolor Grünwald et al., 2021 is recovered as the strongly supported sister species of T. papavericola Grünwald et al., 2021. Dipsas gaigeae is recovered as the sister species of all other Dipsas, albeit with low support, but not as sister to (see Sheehy 2012) or forming a polytomy with (see Grünwald et al. 2021) Tropidodipsas. The D. articulata group, as defined by Peters (1960) and modified by Harvey (2008) and Arteaga et al. (2018) is monophyletic.   Within it, D. viguieri (Bocourt, 1884), a species not included in previous phylogenetic analyses, is recovered as the moderately supported sister taxon of D. articulata (Cope, 1868). We found D. pavonina Schlegel, 1837 to be the moderately supported sister taxon of a clade formed by D. peruana (Boettger, 1898), D. palmeri (Boulenger, 1912), and D. klebbai Arteaga et al., 2018, a relationship not recovered in any of the previous phylogenies. Plesiodipsas perijanensis is nested within the genus Dipsas and is recovered as the moderately supported sister species of D. albifrons (Sauvage, 1884). There are two reciprocally monophyletic, deeply divergent, and geographically structured clades within D. vermiculata sensu lato. One is D. vermiculata sensu stricto and the other is a new species endemic to the Cordillera del Cóndor in southeastern Ecuador and northern Peru. This new species is described in this work. The D. oreas group, as defined by Harvey (2008) and modified by Arteaga et al. (2018) is monophyletic and includes D. nicholsi Dunn, 1933. Within it, D. elegans Boulenger, 1896 is recovered as the strongly supported sister species of D. ellipsifera (Boulenger, 1898), a relationship already uncovered in Arteaga et al. (2018). There are two reciprocally monophyletic, deeply divergent, and geographically structured clades within D. temporalis (Werner, 1909). One is D. temporalis sensu stricto and the other is a new species endemic to central Panama.
Relationships within Sibon are most similar to those presented in Sheehy (2012). The S. argus group is sister to all other members of the genus and it includes the newly described S. ayerbeorum, a species not previously sampled for molecular characters. Sibon annulatus is paraphyletic with respect to S. perissostichon, S. lamari, and three new species described in this work. We restrict the name S. annulatus to the red clade in Fig. 1a based on the type locality of this species (Cartago, Costa Rica) where only members of the red clade have been recorded, as well as on the original description of this species. Günther (1872) mentioned that the holotype has 164 ventrals and the body and tail are encircled by black rings. Members of the yellow clade have more than 170 ventrals and lack full body rings. Sibon annulatus, S. perissostichon, S. lamari, and the three new species form a monophyletic unit exclusive of all other species of the paraphyletic S. annulatus species group (see Arteaga et al. 2018 for a list of species included in this group). There are eleven monophyletic, deeply divergent, and geographically structured clades within S. nebulatus sensu lato. Two of these correspond to species already described (S. bevridgelyi and S. dunni), three correspond to subspecies of S. nebulatus (nebulatus, leucomelas, and hartwegi), one corresponds to a new species described here, and the remaining clades are deemed putative new species. The allocation of subspecies names to each clade was based on direct examination of museum vouchers and whether these agree in coloration and lepidosis with the corresponding holotype, as well as on the geographic range of the included samples.
Finally, we excluded Sibon noalamina Lotzkat et al., 2012 (voucher SMF 91539) from the analyses as the short sequence available in GenBank (gene fragment 16S) represented a rogue taxon that assumed varying phylogenetic positions in the tree collection used to build the consensus tree.

Systematic accounts
We name and provide descriptions only for species that are monophyletic in our molecular phylogeny and share diagnostic features of their coloration pattern and lepidosis. Based on these species delimitation criteria, which follow the general species concept of de Queiroz (2007) Diagnosis. Sibon irmelindicaprioae sp. nov. is placed in the genus Sibon based on phylogenetic evidence (Fig. 1a) and on having the penultimate supralabial conspicuously higher than all other supralabials. The species is diagnosed based on the following combination of characters: (1) 15/15/15 smooth dorsals with enlarged vertebral row (1.5× as wide as adjacent rows); (2) loreal and prefrontal in contact with orbit; (3) 7-9 supralabials with, usually, 5 th and 6 th contacting orbit; (4) 8-10 infralabials with 3 rd -7 th in contact with chinshields, first pair of infralabials not in contact behind symphysial due to presence of postmentals; (5) 187-196 ventrals in males, 174 in the single female; (6) 110-128 divided subcaudals in males, 117 in the single female; (7) dorsal background color olive with maroon lateral body blotches or irregular bands (2-6 dorsal scales long) and a reddish tint along the vertebral line (Figs 2a,b,4b), ventral surfaces yellowish white with encroachment from the dorsal maroon blotches and with smaller blackish speckles and marks in-between the blotches, dorsal aspect of head variegated with a mixture of pinkish to maroon and pale olive yellow speckles (Fig. 5c), throat yellowish white with brownish blotches and spots, iris pale olive   Comparisons. Sibon irmelindicaprioae sp. nov. is compared to other species of Sibon previously subsumed under S. annulatus sensu lato (differences summarized in Table 2). From S. annulatus sensu stricto, the new species differs in having the dorsal body bands faint and broken along the vertebral line ( Description of holotype. Adult male, SVL 387 mm, tail length 193 mm (49% SVL); head length 14.3 mm (3.7% SVL) from tip of snout to angle of jaw; head width 9.0 mm (88% head length) taken at broadest point; snout-orbit distance 2.3 mm; head distinct from neck; snout short, blunt in dorsal outline and rounded in profile; rostral 1.8 mm wide, higher than broad; internasals 1.8 mm wide, broader than long; prefrontals 2.3 mm wide, longer than broad, entering orbit; supraocular 3.6 mm long, longer than broad; frontal 3.7 mm long, pentagonal and with an inward-bent anterior  border, in contact with prefrontals, supraoculars, and parietals; parietals 5.8 mm long, longer than broad; nasal divided, in contact with first two supralabials, loreal, prefrontal, internasal, and rostral; loreal 1.4 mm long, longer than high, entering the orbit; eye diameter 3.7 mm; pupil semi-elliptical; no preocular; two postoculars; temporals 1+3 on the right side, 2+3 on the left side; eight supralabials with 5 th and 6 th contacting orbit on the right side, eight supralabials with 5 th and 6 th contacting orbit on the left side; symphysial precluded from contacting chinshields by a pair of postmentals; ten infralabials, 3 rd -7 th contacting chinshields; two pair of chinshields longer than wide; dorsal scales in 15/15/15 rows, smooth, without apical pits; 193 ventrals; 128 paired subcaudals; cloacal plate single.
Natural history. Specimens of Sibon irmelindicaprioae sp. nov. have been found at night foraging on shrubs, trees, and palm fronds 200-300 cm above the ground in old-growth to moderately disturbed evergreen lowland/foothill forests. Snakes of this species are docile and never attempt to bite. When threatened, individuals may hide the head among body coils and produce a musky and distasteful odor.
Distribution. Sibon irmelindicaprioae sp. nov. is known from 16 localities (listed in Suppl. material 2) in the Chocó region of eastern Panama and northwestern Colombia, with an isolated population on the western slopes of the Cordillera Oriental of Colombia. The species occurs over an estimated area of 62,241 km 2 and has been recorded at elevations 346-1295 m above sea level (Fig. 6). Since the population on the Cordillera Oriental is isolated from the remaining populations and individuals in this area occur at higher elevations and have a different dorsal color pattern, we provisionally assign them to S. irmelindicaprioae sp. nov. pending more comprehensive genetic analyses.
Etymology. The specific epithet irmelindicaprioae is a patronym honoring Irmelin DiCaprio (1945-present), mother of Leonardo DiCaprio, long-time advocate and supporter of biodiversity conservation around the world. Conservation status. We consider Sibon irmelindicaprioae sp. nov. to be included in the Near Threatened category following IUCN Red List criteria (IUCN 2001) because the species is distributed over a region that holds large areas of continuous unspoiled forest. Based on the species distribution model presented in Fig. 6 in combination with maps of vegetation cover of Colombia (IDEAM 2014) and Panama (CATHALAC 2011), we estimate that more than half (~ 54%) of the species' forest habitat is still standing. Unfortunately, vast areas of the Chocó rainforest in northern Colombia and towards central Panama have already been converted to pastures (Myers et al. 2000). However, S. irmelindicaprioae sp. nov. occurs over an area greater than 50,000 km 2 and is presumably not declining fast enough to qualify for a threatened category.  Diagnosis. Sibon canopy sp. nov. is placed in the genus Sibon based on phylogenetic evidence (Fig. 1a) and on having the penultimate supralabial conspicuously higher than all other supralabials. The species is diagnosed based on the following combination of characters: (1) 15/15/15 smooth dorsals with enlarged vertebral row (1.4× as wide as adjacent rows); (2) loreal and prefrontal in contact with orbit; (3) 7-8 supralabials with, usually, 4 th , 5 th , and occasionally 6 th contacting orbit; (4) usually 7-8 infralabials with 2 nd -6 th in contact with chinshields, first pair of infralabials not in contact behind symphysial due to presence of a postmental; (5) 180-189 ventrals in males, 170-185 in females; (6) 113-130 divided subcaudals in males, 107-124 in females; (7) dorsal background color olive with maroon bands (1-2 dorsal scales long mid-dorsally and 3-5 dorsal scales long on the lower flanks) and a reddish tint along the vertebral line (Fig. 2c), ventral surfaces white with encroachment from the dorsal maroon blotches, dorsal aspect of head composed of broad irregular maroon to blackish blotches interspersed with olive to red blotches (Fig. 5b), throat white with brownish blotches, iris dark reddish brown; (8) 336-427 mm SVL in males, 318-357 mm in females; (9) 160-221 mm TL in males, 157-185 mm in females.
Comparisons. Sibon canopy sp. nov. is compared to other species of Sibon previously subsumed under S. annulatus sensu lato (differences summarized in Table 2). From S. annulatus sensu stricto, the new species differs in having a single postmental scale, olive spaces among dorsal bands enclosing maroon blotches (Figs 2c, 4a), and by having small irregular (rather than broad and symmetrical) markings on the dorsal surface of the head (Fig. 5). Sibon canopy sp. nov. differs from S. irmelindicaprioae sp. nov. by having one postmental scale (instead of two), a lower number of infralabials (6-10 vs. 8-10), a lower number of ventrals in males (180-189 vs. 187-196), a different pattern on the dorsal surface of the head (Fig. 5), and by having maroon spots enclosed in the dorsal olive interspaces (Figs 2, 4). Sibon canopy sp. nov. differs from S. marleyae sp. nov. by having one postmental scale (instead of two), olive spaces among dorsal bands enclosing maroon blotches (Figs 2, 4), and by having irregular (rather than symmetrical) markings on the dorsal surface of the head (Fig. 5).
Description of holotype. Adult female, SVL 321 mm, tail length 157 mm (48% SVL); head length 15.4 mm (4.7% SVL) from tip of snout to commissure of mouth; head width 8.0 mm (76% head length) taken at broadest point; snout-orbit distance 3.3 mm; head distinct from neck; snout short, blunt in dorsal outline and rounded in profile; rostral 2.1 mm wide, higher than broad; internasals 1.6 mm wide, broader than long; prefrontals 1.9 mm wide, longer than broad, entering orbit; supraocular 3.7 mm long, longer than broad; frontal 3.2 mm long, pentagonal and with a straight anterior border, in contact with prefrontals, supraoculars, and parietals; parietals 5.2 mm long, longer than broad; nasal divided, in contact with first three supralabials, loreal, prefrontal, internasal, and rostral; loreal 1.7 mm long, longer than high, entering the orbit; eye diameter 3.0 mm; pupil semi-elliptical; no preocular; two postoculars; temporals 1+2; eight supralabials with 5 th and 6 th contacting orbit on the right side, seven supralabials with 4 th and 5 th contacting orbit on the left side; symphysial in contact with chinshields; nine infralabials with 2 nd -5 th contacting chinshields; two pair of chinshields longer than wide; dorsal scales in 15/15/15 rows, smooth, without apical pits; 172 ventrals; 93+ divided subcaudals; cloacal plate entire.
Natural history. Lotzkat (2014) found specimens of Sibon canopy sp. nov. foraging at night on vegetation 50-300 cm above the ground in old-growth to moderately disturbed evergreen foothill/montane forests. At Cerro Gaital, Coclé province, we found two specimens moving on mossy branches and moist leaves 40-220 cm above the ground in primary forest during a drizzle. Ray et al. (2012) found this species to be more common in forest and along streams rather than around ponds. Only one individual (a juvenile) was seen crawling along a stream bed. Ray et al. (2012) found oligochaete and mollusk remains in fecal samples of 37 individuals of S. canopy sp. nov.
from El Copé and Altos del María, Panama. They also observed an individual feeding on a snail at El Copé.
Distribution. Sibon canopy sp. nov. is known from 25 localities (listed in Suppl. material 2) in both the Atlantic and Pacific slopes of the Cordillera Central in western Panama, with a population on the slopes of El Valle Volcano. The species occurs over an estimated area of 8,089 km 2 and has been recorded at elevations 543-1641 m above sea level (Fig. 6).
Etymology. The specific epithet canopy is used as a noun in apposition and honors the Canopy Family system of reserves, particularly its Canopy Lodge in Valle de Antón, Coclé province, Panama, where the new species occurs. Though best known for its world-class eco-tourism focused on birds, the Canopy Family also protects habitat that is critical for dozens of poorly studied Panamanian snakes such as S. canopy sp. nov. and S. irmelindicaprioae sp. nov. The project was founded in 1994 by Raúl Arias de Para and Denise Barakat de Arias, two champions of Panamanian conservation who are deeply intertwined with the Political history of the country. In 2019, the Canopy Family invited us to explore their system of reserves in order to discover their herpetofauna. As a result of this invitation, both S. canopy sp. nov. and a new species of Dipsas were discovered.
Conservation status. We consider Sibon canopy sp. nov. to be included in the Near Threatened category following IUCN Red List criteria (IUCN 2001) because, although the species' estimated extent of occurrence is less than 10,000 km 2 and nearly 40% of this area has already been deforested ( Fig. 6; CATHALAC 2011), the species occurs in at least four major national parks (Lotzkat 2014) and satellite images show that there is forest connectivity between populations. At Parque Nacional G. D. Omar Torríjos Table 2. Differences in coloration, scale counts, and size between Sibon annulatus, S. canopy sp. nov., S. irmelindicaprioae sp. nov., and S. marleyae sp. nov. The range of each continuous variable is from our own sample, Peters (1960), Savage andMcDiarmid 1992, Lewis et al. (2013), Lotzkat (2014), and Meneses-Pelayo et al. (2016). The numbers in parentheses represent the sample size. Herrera, the occurrence rates of S. canopy sp. nov. have actually increased by a factor of three in the period between 2006 and 2012 (Zipkin et al. 2020). However, the body condition of the individuals in this locality declined following the collapse of amphibian populations due to chytridiomycosis (Zipkin et al. 2020). The status and trend of other populations should be evaluated carefully given that S. canopy sp. nov. is endemic to Panama (but see Discussion) and probably highly dependent on old-growth forests. Diagnosis. Sibon marleyae sp. nov. is placed in the genus Sibon based on phylogenetic evidence (Fig. 1a) and on having the penultimate supralabial conspicuously higher than all other supralabials. The species is diagnosed based on the following combination of characters: (1) 15/15/15 smooth dorsals with enlarged vertebral row (up to 2× as wide as adjacent rows); (2) loreal and prefrontal in contact with orbit; (3) 7-8 supralabials with, usually, 4 th , 5 th , and occasionally 6 th contacting orbit; (4) usually 8-9 infralabials with 2 nd -6 th in contact with chinshields, first pair of infralabials not in contact behind symphysial due to presence of two postmentals; (5) 186-204 ventrals in males, 176-193 in females; (6) 130-143 divided subcaudals in males, 109-128 in females; (7) dorsal background color olive to yellow with maroon (black in juveniles) bands (1-2 dorsal scales long mid-dorsally and 3-5 dorsal scales long on the lower flanks) and a reddish tint along the vertebral line (Figs 2g-i, 4c), ventral surfaces white with encroachment from the dorsal maroon blotches (Fig. 8b), dorsal aspect of head composed of blackish symmetrical markings on a red background color (Fig. 5d), throat white with broad brownish blotches, iris rich reddish brown; (8) 308-464 mm SVL in males, 329-368 mm in females; (9) 167-233 mm TL in males, 175-183 mm in females.

Sibon annulatus
Comparisons. Sibon marleyae sp. nov. is compared to other species of Sibon previously subsumed under S. annulatus sensu lato (differences summarized in Table 2). From S. annulatus sensu stricto, the new species differs in having maroon bands usually broken along vertebral line rather than bands extending over the entire dorsal and lateral surfaces, a bright reddish coloration along the mid-dorsum and on the top of the head (Figs 4c, 5d, 8a), and a higher number of ventral scales in males and females (Table 2). Sibon marleyae sp. nov. differs from S. canopy sp. nov. by having two postmental scales (instead of only one), a higher number of ventrals in males and females, a pattern of symmetrical (rather than irregular and asymmetrical) markings on the dorsal surface of the head (Fig. 5), and by lacking maroon spots enclosed in the dorsal olive interspaces (Figs 2g-i, 4c). Sibon marleyae sp. nov. differs from S. irmelindicaprioae sp. nov. primarily by having a pattern of broad blackish markings on the head instead of a finely variegated pattern and by having a higher number of ventrals and subcaudals in both males and females ( Table 2).
Description of holotype. Adult male, SVL 335 mm, tail length 167 mm (49.8% SVL); head length 12.8 mm (3.8% SVL) from tip of snout to angle of jaw; head width 7.6 mm (59% head length) taken at broadest point; snout-orbit distance 3.1 mm; head distinct from neck; snout short, blunt in dorsal outline and rounded in profile; rostral 2.1 mm wide, higher than broad; internasals 1.3 mm wide, broader than long; prefrontals 1.6 mm wide, longer than broad, entering orbit; supraocular 2.8 mm long, longer than broad; frontal 2.9 mm long, pentagonal and with a straight anterior border, in contact with prefrontals, supraoculars, and parietals; parietals 4.3 mm long, longer than broad; nasal divided, in contact with two supralabials, loreal, prefrontal, internasal, and rostral; loreal 1.2 mm long, longer than high, entering the orbit; eye diameter 2.9 mm; pupil semi-elliptical; no preocular; two postoculars; temporals 1+2; seven supralabials with 4 th and 5 th contacting orbit; symphysial precluded from contacting chinshields by the presence of two small postmentals; eight infralabials with 2 nd -6 th contacting chinshields on the right side, nine infralabials with 2 nd -7 th contacting chinshields on the left side; two pairs of chinshields longer than wide; dorsal scales in 15/15/15 rows, smooth, without apical pits; 204 ventrals; 132 divided subcaudals; cloacal plate entire.
Natural history. Specimens of Sibon marleyae sp. nov. have been found at night foraging on shrubs and trees 1-6 m above the ground in old-growth evergreen lowland/foothill forests, particularly along streams and small rivers. Snakes of this species are docile and never attempt to bite. When threatened, individuals may hide the head among body coils and produce a musky and distasteful odor. One female (Fig. 2h) from the type locality laid two eggs in a terrarium. After an incubation period of 80 days, one of the eggs hatched (Fig. 2i).
Distribution. Sibon marleyae sp. nov. is known from 17 localities (listed in Suppl. material 2) along the Chocoan lowlands and adjacent foothills of the Andes in Ecuador and Colombia, with populations on the coastal mountain ranges Mache-Chindul and Cerro Pata de Pájaro in Ecuador. The species has been recorded at elevations 131-1338 m above sea level (Fig. 6).
Etymology. The specific epithet marleyae is a patronym honoring a young nature lover, Marley Sheth, the 11-year old daughter of Brian and Adria Sheth, both longtime supporters of biodiversity conservation around the world.
Conservation status. We consider Sibon marleyae sp. nov. to be included in the Least Concern category following IUCN Red List criteria (IUCN 2001) because the species is distributed over a region of the Chocó biome that holds large areas of continuous unspoiled forest. Based on the species distribution model presented in Fig. 6 in combination with maps of vegetation cover of Colombia (IDEAM 2014) and Ecuador (MAE 2012), we estimate that more than half (~ 55%) of the species' forest habitat is still standing. Unfortunately, vast areas of the Chocó rainforest in western Ecuador have already been converted to pastures (Myers et al. 2000). However, S. marleyae sp. nov. occurs over an area greater than 25,000 km 2 and is presumably not declining fast enough to qualify for a threatened category.
Comparisons. Sibon vieirai sp. nov. is most similar to S. leucomelas, from which it differs primarily on the basis of coloration (differences summarized under Table 3). In S. vieirai sp. nov. (Figs 4d, 10b-d), the complete black and pale dorsal bands typical of S. leucomelas (Fig. 10a) are usually absent. Instead, the white "bands" are formed by series of white spots and the black bands are faint and incomplete. The color of the pale "bands" also differs between species: rosy white in S. leucomelas and white in S. vieirai sp. nov. (Fig. 10). In S. vieirai sp. nov the throat has a checkerboard pattern of black and white markings whereas in S. leucomelas it is entirely black with fine white speckling (Fig. 11). Overall, specimens assignable to S. leucomelas have a greater number of ventral scales than Sibon vieirai sp. nov. in both males and females, although there is overlap in the counts (Table 3). Sibon vieirai sp. nov. differs from S. bevridgelyi by having white (instead of golden yellow) dorsal markings on a primarily gray (instead of rusty brown to deep maroon) background color. Arteaga at al. (2018) presented an in-depth comparison between S. bevridgelyi and Sibon vieirai sp. nov. (reported as S. nebulatus from Ecuador).
Natural history. Specimens of Sibon vieirai sp. nov. have been found in old growth to heavily disturbed evergreen lowland/foothill forests as well as in rural gardens and plantations. Active snakes have been seen at night foraging at ground level or on vegetation up to 3 m above the ground. One snake was spotted as it emerged from under a Table 3. Differences in coloration, scale counts, and size between Sibon leucomelas and S. vieirai sp. nov. The range of each continuous variable is from our own sample, Boulenger 1896, andFrazier et al. (2006). The numbers in parentheses represent the sample size. pile of logs at sunset. Based on our own field experience, individuals appear to be more active when it is raining or drizzling. In the field in Ecuador, specimens of S. vieirai sp. nov. have been observed feeding on slugs and snails. A female from Hostería Selva Virgen, Pichincha province, Ecuador laid a clutch of four eggs. Distribution. Sibon vieirai sp. nov. is known from at least 95 localities (listed in Suppl. material 2) along the Chocoan lowlands and adjacent foothills of the Andes in northwestern Ecuador and southwestern Colombia. Previous records of S. nebulatus from the rainforests of northwestern Ecuador as well as those of the Pacific lowlands of Colombia in Nariño department almost surely correspond to this new species. Sibon vieirai sp. nov. occurs over an estimated 58,551 km 2 area and has been recorded at elevations 5-1803 m above sea level (Fig. 12).
Etymology. The specific epithet vieirai is a patronym honoring Jose Vieira, a Venezuelan biologist and wildlife photographer who created the Ex-Situ project, a freeaccess photo bank depicting Latin American fauna on a white background. Jose Vieira's photos have been crucial in illustrating field guides about herpetofauna, educational posters, and research publications. Most of the images in this work were created by Jose Vieira. Additionally, after nearly six years of active collaboration with one of us (AA), it has become evident that Jose is one of the most tireless and focused young field biologists ever to sample the jungles of the tropics, a work ethic that has resulted in the generation of photo and museum vouchers for hundreds of poorly studied species of herpetofauna, including the holotype of this new Sibon.
Conservation status. We consider Sibon vieirai sp. nov. to be included in the Least Concern category following IUCN Red List criteria (IUCN 2001) because the species is distributed over a region of the Chocó biome that holds large areas of continuous unspoiled forest. Based on the species distribution model presented in Fig. 12 in combination with maps of vegetation cover of Colombia (IDEAM 2014) and Ecuador (MAE 2012), we estimate that more than half (~ 51%) of the species' forest habitat is still standing. Unfortunately, vast areas of the Chocó rainforest in western Ecuador have already been converted to pastures (Myers et al. 2000). However, S. vieirai sp. nov. occurs over an area greater than 50,000 km 2 and is presumably not declining fast enough to qualify for a threatened category.  Diagnosis. Dipsas welborni sp. nov. is placed in the genus Dipsas based on phylogenetic evidence (Fig. 1b) and the absence of a labial that is noticeably higher than other labials. The species is diagnosed based on the following combination of characters:
Comparisons. Dipsas welborni sp. nov. differs from the majority of its congeners by having dorsal scales arranged in 13/13/13 rows, loreal entering the orbit, and dorsum of head strongly vermiculated. The new species is most similar to D. vermiculata, from which it differs on the basis of the following characters of coloration and lepidosis ( Fig. 14; Table 4). In D. welborni sp. nov., there are two prefrontal scales (partially fused in ZSFQ 5060) whereas in all specimens of D. vermiculata examined (Suppl. material 1) as well as the four Ecuadorian specimens reported in Peters (1960), the prefrontals are fused into a single scale (Fig. 15). Females of Dipsas welborni sp. nov. have more ventrals (177-179) and subcaudals (105-106) than those of D. vermiculata (173-174  (Table 4), and males of the former also have more rows of spines on the asulcate surface of the hemipenis body (the hemipenis of D. vermiculata is depicted in Vera-Pérez 2020 whereas the organs of four males of D. welborni sp. nov. are depicted in Pazmiño-Otamendi et al. 2020). Finally, the two species further differ in the background color of the ventral surfaces: always white in D. welborni sp. nov. (Fig. 13b), and usually yellow or occasionally pale yellowish white in D. vermiculata.
Natural history. Specimens of Dipsas welborni sp. nov. have been found foraging on vegetation 20-350 cm above the ground in old-growth to moderately disturbed evergreen montane forests. Snakes of this species are docile and never attempt to bite. When threatened, individuals may flatten their body and expand their head to simulate a triangular shape as well as produce a musky and distasteful odor.
Distribution. Dipsas welborni sp. nov. is known from 26 localities (listed in Suppl. material 2) along the Cordillera del Cóndor in southeastern Ecuador (provinces Morona Santiago and Zamora Chinchipe) and northern Peru (Amazonas department). The species occurs over an estimated area of 10,521 km 2 and has been recorded at elevations 853-1843 m above sea level (Fig. 16). One locality, Etseketai Entse, Amazonas department, Peru, is in the Río Cenepa valley at ~ 245 m above sea level. Since this locality is much lower in elevation than other localities in the Cordillera del Cóndor, it is likely that the specimens collected there (USNM 316599-600) were actually found on the neighboring mountain ridges.
Etymology. The specific epithet welborni is a patronym honoring David Welborn, a lifelong champion of ecosystem and species conservation who supports and serves on several nonprofit boards dedicated to the environment. David retired from the board of Nature and Culture International in 2021 after 18 years of service, including four as board chairman. Nature and Culture International, a non-profit organization, has conserved more than 9 million hectares of tropical Latin American ecosystems, including key habitat in the Maycu Reserve of southeastern Ecuador, where Dipsas welborni sp. nov. was discovered. Table 4. Differences in coloration, scale counts, and size between Dipsas vermiculata and D. welborni sp. nov. The range of each continuous variable is from our own sample, Peters (1960), and Vera-Pérez (2020). The numbers in parentheses represent the sample size. Conservation status. We consider Dipsas welborni sp. nov. to be in the Near Threatened category following IUCN Red List criteria (IUCN 2001) because the species is distributed over a region of the Amazonian slopes of the Andes that holds large areas of continuous unspoiled forest. Based on the species distribution model presented in Fig. 16 in combination with the most recent maps of vegetation cover of the Amazon basin (MapBiomas Amazonía 2022), we estimate that the majority (~ 76%) of the species' forest habitat in Ecuador is still standing. Unfortunately, vast areas of the Cordillera del Cóndor, notably on the Ecuadorian part of the species' range, are being cleared to make room for large-scale opencast mining operations (Chicaiza 2010;Valencia et al. 2017). However, since D. welborni sp. nov. occurs over an area greater than 10,000 km 2 , the species does not qualify for a threatened category.

Presence of Sibon ayerbeorum in Ecuador and Valle del Cauca, Colombia
We expand the distribution of Sibon ayerbeorum, a species previously known only from departments Cauca (Vera-Pérez 2019), Chocó (Echavarría-Rentería and Medina-Rangel 2021), and Risaralda (Bonilla and Moya 2021) in Colombia. We examined three additional specimens (listed in Suppl. material 1) at Colección Zoológica de la Universidad ICESI (labeled CZI) and at ZSFQ that represent, respectively, new records for Valle del Cauca department in Colombia and Esmeraldas province in Ecuador. CZI-R063 is a juvenile male collected by Santiago Orozco on 17 August 2019 at La Loca, Valle del Cauca department, Colombia (3.57656,; 658 m a.s.l.). CZI-R067 is a juvenile male collected by Santiago Orozco on 30 August 2019 at La Riqueza, Valle del Cauca department, Colombia (3.59874,; 621 m a.s.l.). ZSFQ 5066 (Fig. 2f ) is a juvenile male collected by Jose Vieira, Daniela Franco, and Alex Mora on 4 December 2019 at Reserva Biológica Canandé, Esmeraldas Province, Ecuador (0.49531, -79.17832; 560 m a.s.l.). These specimens agree in coloration and lepidosis with the description of S. ayerbeorum presented in Vera-Pérez 2019 (expanded in Echavarría-Rentería and Medina-Rangel 2021), most notably in having a much lower (fewer than 160) number of ventral scales than any other sympatric Sibon species, vertebral scale row not noticeably wider than adjacent rows, postmentals absent, dorsal coloration green to grayish brown with black-bordered reddish markings, and ventral coloration consisting of a checkerboard pattern of yellowish white markings interspersed with blackish markings. We also report an individual of S. ayerbeorum photographed (https://www.inaturalist.org/photos/179847493; DHMECN 14936) by Mateo Vega on July 21, 2019 at Comunidad El Baboso, Carchi province, Ecuador (0.89972,; 803 m). We did not examine this specimen, but the photograph agrees in coloration with the variation reported for this species in Vera-Pérez (2019) and Echavarría-Rentería and Medina-Rangel (2021). The updated distribution of S. ayerbeorum is shown in Fig. 17 and includes both published records as well as new localities reported here (summarized in Suppl. material 2).

Discussion
This work marks the fourth attempt at elucidating the evolutionary relationships and limits between species within Dipsadini using molecular characters. It builds on the results of Sheehy (2012), Arteaga et al. (2018), and Grünwald et al. (2021) and seeks to combine the datasets of these monographic works into an updated phylogenetic framework to study the enigmatic Neotropical snail-eating snakes. Although a robust and stable taxonomy of the group is still far from being complete, the results presented here help solve some of the taxonomic issues left unaddressed in the aforementioned works, most notably the paraphyly of Dipsadini with respect to Geophis, the generic classification of Plesiodipsas perijanensis, the paraphyly of S. annulatus and S. nebulatus, the identity of unidentified Dipsas and Sibon from Ecuador, Colombia, and Panama, and the phylogenetic position of D. viguieri and S. ayerbeorum. Sheehy (2012) argued that the genus Geophis should be transferred to Dipsadini. Based on our results and those of Grünwald et al. (2021), we propose the same arrangement to maintain the monophyly of the tribe. Furthermore, we bring attention to the need to explore the limits within this genus given that the great majority of the included taxa, notably the type species G. chalybeus Wagler, 1830, have not been sampled for molecular characters.  erected the genus Plesiodipsas to accommodate the enigmatic D. perijanensis, primarily based on this species' unique osteology, musculature, and visceral morphology. We included a specimen of this species (UIS R-4180, an adult female collected by Elson Meneses at Santa Bárbara, Santander department, Colombia; pictured in https://www.santanderherps.com/dipsadidae) in the phylogeny. Although, this sample was recovered as the moderately supported sister species of D. albifrons, it is deeply nested in Dipsas (Fig. 1b). Based on these results, we synonymize Plesiodipsas with Dipsas.
The paraphyly of Sibon annulatus with respect to S. lamari and S. perissostichon was uncovered by Sheehy (2012). Here, we demonstrate that this is the result of previously unrecognized diversity within the group. The descriptions of S. irmelindicaprioae sp. nov., S. canopy sp. nov., and S. marleyae sp. nov. solve this paraphyly and increase our knowledge on the distribution, morphological variation, and distinct ecological requirements within species in this group. Samples of two specimens of S. irmelindicaprioae sp. nov. from the western slopes of the Cordillera Oriental of Colombia (UIS R-3701 and UIS R-3515) form a group with the remaining samples from eastern Panama. However, these specimens belong to an isolated population (see Fig. 6) having a greater number of ventral scales (Suppl. material 1) and a dorsal coloration consisting of bands extending over the entire dorsal and lateral surfaces (see images in Meneses-Pelayo et al. 2016 similar to those seen in specimens of S. annulatus sensu stricto. Since we only included sequences of the ND4 gene of these specimens and they do not overlap with the DNA sampling for the remaining specimens, we consider their phylogenetic position unresolved and provisional. We report Sibon canopy sp. nov. as being endemic to Panama and our distribution model for this species does not predict its presence in Costa Rica (Fig. 6). However, one of the records (Río Changena; Lotzkat et al. 2012) is 5.5 km from the border between the two countries. Thus, we anticipate the presence of this species in Costa Rica.
The paraphyly of Sibon nebulatus with respect to S. bevridgelyi and S. dunni was already uncovered by Arteaga et al. (2018). Here, we combined the samples generated in this work with our novel taxon sampling and those of Sheehy (2012) and discovered that even the subspecies leucomelas is not monophyletic. Instead, it comprises a northern species with a continuous distribution from Cauca department in Colombia to Veraguas province in Panama (Fig. 12) as well as a southern species ranging from Cauca department in Colombia to Los Ríos province in Ecuador. Curiously, the two are not sister species (see Fig. 1a) even despite being previously considered a single subspecies diagnosed by its distinctive coloration (Peters 1960). Instead, the southern species is nested within a strongly supported clade that also includes S. bevridgelyi and S. dunni. Although we only included DNA samples of Ecuadorian and Panamanian specimens of this subspecies in the analyses (Appendix 1), we restrict the name leucomelas to the northern clade based on examination of museum specimens from near the type locality (Buenaventura, Cauca department, Colombia; see Suppl. material 1), comparison with the description of Leptognathus leucomelas by Boulenger 1896, and a detailed study of the illustration of the holotype. The holotype of L. leucomelas can be allocated to the northern species based on its entirely black throat with fine white speckles (compare illustration in Boulenger 1896 with Fig. 11c, d) and the shape and width of the white bands (compare photo of holotype in Peters 1960 with Fig. 10a). Based on these results, we elevate S. leucomelas to full species status, provide a morphological diagnosis for this species (Table 3), an updated distribution map (Fig. 12), and erect a new name for the southern species, S. vieirai sp. nov., thus resolving the paraphyly of this group. We also examined photos of MHUA 14511 provided to us by Paúl Gutiérrez-Cárdenas. This specimen is included in our phylogenetic analysis and resembles the description of the subspecies S. nebulatus hartwegi provided by Peters (1960), particularly the presence of three wide and distinct dark dorsal blotches anteriorly followed by narrow and incomplete blotches throughout the rest of the dorsum. Both MHUA 14511 and SN 0001 form a strongly supported clade (Fig. 1a) and they were collected at the lowlands of the Río Magdalena valley (type locality of the subspecies S. nebulatus hartwegi). Based on these results, we elevate S. hartwegi to full species status as an action towards beginning to clarify the identity of the population-level genetic lineages of the S. nebulatus complex occurring in Colombia. Lotzkat (2014) already proposed that populations of Dipsas temporalis from Panama's Cordillera Central might be a new species, a hypothesis supported by the results of Sheehy (2012). Here, we provide additional phylogenetic evidence that supports this view. However, we refrain from providing a formal description for this species here since a more comprehensive study is being prepared by Ray et al. (in press).
Based on a personal communication with the collector of MHUA 14278 (labeled as Dipsas pratti [Boulenger, 1897] in Daza et al. 2009), Barros et al. (2012) suggested that this specimen is a Dipsas sanctijoannis (Boulenger, 1911). In our phylogeny, our included sample of MHUA 14278 clusters with the remaining three confirmed samples of D. temporalis, not with D. pratti (Sheehy 2012). Although we did not examine MHUA 14278, we have decided to provisionally label it in our phylogeny as D. temporalis following Sheehy (2012) and Arteaga et al. (2018) and not Barros et al. (2012). The rationale for this decision is based on the locality where the specimen was collected: the lowlands of northwestern Antioquia department at 233 m elevation, well below the elevation range of D. sanctijoannis. A more comprehensive review of the identity of this specimen is being prepared by Ray et al. (in press). Arteaga et al. (2018) uncovered deep intraspecific genetic divergence between samples of Dipsas vermiculata from the Cordillera del Cóndor and those from the remaining Amazon basin in Ecuador. Here, we show that this is the result of the existence of a new species of Dipsas endemic to this unique Amazonian mountain range. Peters (1960) diagnosed D. vermiculata by having fused prefrontal scales, a characteristic shared by the Colombian specimen reported by Vera-Pérez (2020). Populations previously identified as D. vermiculata from Cordillera del Cóndor usually have two prefrontal scales (Fig. 15) and this characteristic is used to diagnose the new species: D. welborni sp. nov.
The discovery of additional material of Sibon ayerbeorum in Ecuador and in Valle del Cauca, Colombia not only expands our understanding of the distribution of this newly described snake species, but it also provides additional insight into its ontogenetic variation. Based on the specimens we examined (Suppl. material 1) as well as photos presented in Vera-Pérez (2019) and Echavarría-Rentería and Medina-Rangel (2021), it appears like individuals of S. ayerbeorum undergo an ontogenetic shift in coloration from having mossy green dorsal surfaces in juveniles (Fig. 2f ) to grayish brown (see Vera-Pérez 2019) as adults.
Clearly, we are a long way from achieving a robust and stable Dipsadini taxonomy. Higher-level relationships within the tribe are far from being resolved and are not consistent across published phylogenies. This is probably the result of inadequate sampling of taxa (only 60 of 133 species are included) or characters (only five mtDNA and two nuclear loci were used). Also, no less than five putative new species remain to be described within the Sibon nebulatus species complex and the limits of S. nebulatus sensu stricto have yet to be clearly defined. Also, an overwhelming majority of the species of Geophis, including the type species of the genus, remain unsampled. We suggest that a more comprehensive collaborative work is needed in Dipsadini in general but particularly in the S. nebulatus species complex. Such work would gain much clarity by sampling cis-Andean populations, including a phylogenomic approach, and likewise by a denser geographical sampling especially around the hitherto depicted range limits. Until then, we hope that our results help bring attention to which branches of the Dipsadini tree of life warrant further study.
ZSFQ under the tutelage of Juan M. Guayasamin. Sequencing was made possible with support of the Inédita Program from the Ecuadorian Science Agency SENESCYT (Respuestas a la Crisis de Biodiversidad: La Descripción de Especies como Herramienta de Conservación; INEDITA PIC-20-INE-USFQ-001). Work by AB was supported by Sistema Nacional de Investigación (SNI) of the Secretaría Nacional de Ciencia, Tecnología e Innovación (SENACYT, Panamá). Table A1. GenBank accession numbers for loci and terminals of taxa and outgroups sampled in this study. Novel sequence data produced in this study are marked with an asterisk (*).