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) and specifically approved as part of obtaining the following field permits for research and collection: MAE-DNB-CM-2015-0017 (granted to Universidad Tecnológica Indoamérica), MAE-DNB-CM-2018-0105 and MAATE-DBI-CM-2022-0245 (granted to Universidad San Francisco de Quito), and 004-AIC-DPC-B-MAE-18 (granted to Universidad del Azuay). 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 Tecnológica Indoamérica (MZUTI), Museo de Zoología de la Universidad San Francisco de Quito (ZSFQ), Museo de Zoología de la Universidad del Azuay (MZUA), and the herpetology collection at Bioparque Amaru (AMARU). Specimens labeled JMG were also deposited at ZSFQ.

Criteria for common name designation are as proposed by Caramaschi et al. (2006) and Coloma and Guayasamin (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.).

Our terminology for Atractus cephalic shields follows Savage (1960), diagnoses and descriptions generally follow Passos et al. (2009a), and ventral and subcaudal counts follow Dowling (1951). We examined comparative alcohol-preserved specimens from the herpetology collections at MZUTI, MZUA, ZSFQ, American Museum of Natural History (AMNH), Museo de Zoología de la Pontificia Universidad Católica del Ecuador (QCAZ), and Muséum National d’Histoire Naturelle (MNHN) (Table 1). 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). Sex was determined by establishing the presence/absence of hemipenes through a subcaudal incision at the base of the tail unless hemipenes were everted.

Tissue samples from 12 individuals representing seven species (including the three new species described here) were obtained in Ecuador. All specimens included in the genetic analyses were morphologically identified according to Savage (1960), Arteaga et al. (2017), Melo-Sampaio et al. (2021), and Arteaga et al. (2022). We generated sequence data for samples marked with an asterisk under Appendix I, which includes museum vouchers at MZUTI, MZUA, and ZSFQ.

Genomic DNA was extracted from 96% ethanol-preserved tissue samples (liver, muscle tissue, or scales) using either a guanidinium isothiocyanate extraction protocol (Peñafiel 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 II. 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 I).

A total of 274 DNA sequences were used to build a phylogenetic tree of the genus Atractus, of which 32 were generated during this work and 242 were downloaded from GenBank, most of which were produced by Arteaga et al. (2017), Melo-Sampaio et al. (2021), and Passos et al. (2022). Of these, 85 sequences are 367–516 bp long fragments of the 16S gene, 66 are 578–1,079 bp long fragments of the CYTB gene, 69 are 567–849 bp long fragments of the ND4 gene, 18 are 513–573 bp long fragments of the C-MOS gene, 19 are 386–516 bp long fragments of the NT3 gene, and 17 are 736 bp long fragments of the RAG-1 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 I) 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 16 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 PartitionFinder 2.1.1 (Lanfear et al. 2016) under the Bayesian information criterion.

Phylogenetic relationships were assessed under both a Bayesian inference (BI) approach in MrBayes 3.2.0 (Ronquist and Huelsenbeck 2013) and a maximum likelihood (ML) approach in RAxML-NG v. 1.1.0 (Kozlov et al. 2019). For the ML analysis, nodal support was assessed using the standard bootstrapping algorithm with 1000 non-parametric bootstraps. For the BI analysis, four independent analyses were performed to reduce the chance of converging on a local optimum. Each analysis consisted of 6,666,667 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 5,000 saved trees per analysis were used to calculate posterior probabilities (PP) for each bipartition in a 50% majority-rule consensus tree. We used Tracer 1.7.2 (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. Genetic distances between Atractus roulei Despax, 1910 and its sister species were calculated using the uncorrected distance matrix in Geneious ProTM 2021.1.1. GenBank accession numbers are listed in Appendix I.

We present ranges of occurrence for five species of Atractus, including the three new species described here. Presence localities are derived from museum vouchers (Table 1), photographic records (iNaturalist), and the literature (all summarized under Suppl. material 1: Table S1). For three of the five 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 under Suppl. material 1: Table S1, 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 WorldClim 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 (Suppl. material 1: Table S1) 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”.

Selected partitions and models of evolution are presented in Table 2. We consider strong support for a clade when Bayesian analyses yield posterior probability values > 95%, following Felsenstein (2004), or when bootstrap values are greater than 70%. The overall topology and support of the BI (Fig. 1) and ML (Suppl. material 2: Figure S1) analyses are similar to that of Arteaga et al. (2017) and Passos et al. (2022). Species of the Atractus roulei species group are sister to all other sampled Atractus in the BI analysis, a view contrary to the ML analysis and to Murphy et al. (2019), in which A. trilineatus Wagler, 1928 and A. boimirim Passos et al., 2016, respectively are recovered as sister to all other Atractus. Below, we outline some differences between our analysis and those published in Murphy et al. (2019) and Passos et al. (2022).

Figure 1. 

Phylogenetic relationships within Atractus inferred using a Bayesian inference and derived from analysis of 3,985 bp of DNA (gene fragments 16S, CYTB, ND4, C-MOS, NT3, and RAG1). Support values on intra-specific branches are not shown for clarity. Voucher numbers for sequences are indicated for each terminal. Black dots indicate clades with posterior probability values from 95–100%. Grey dots indicate values from 70–94%. White dots indicate values from 50–69% (values < 50% not shown). Colored clades correspond to the species’ distribution presented in the map of Fig. 2. New or resurrected species are indicated in bold type.

Atractus roulei is the strongly supported sister species of A. carrioni Parker, 1930, a relationship recovered in previous studies, but we found additional geographically structured genetic divergence within the former species (Figs 1, 2). We found moderate support for the placement of A. trilineatus as sister to A. major sensu Schargel et al. (2013), but strong support for the reciprocal monophyly between snakes assignable to A. arangoi, previously subsumed under A. major, and all other samples of A. major, including samples from throughout the species’ area of distribution. Samples labeled A. arangoi in our phylogeny are not closely related to A. torquatus (Duméril, Bibron, & Duméril, 1854), a name that has been applied to Ecuadorian specimens of the former (see Maynard et al. 2017). Our sample of A. touzeti Schargel et al., 2013 from the type locality is strongly supported as sister to the sample of A. atlas Passos et al., 2018. We found strong support for the relationship between A. resplendens Werner, 1901 from near the type locality and a new species from southeastern Ecuador. Our included samples of A. orcesi Savage, 1955 form a strongly supported sister clade to A. duboisi (Boulenger, 1880). A new species previously confused with A. ecuadorensis Savage, 1955, A. orcesi, and A. resplendens is not closely related to any of these species, but is recovered as the strongly supported sister species to a clade that contains A. ukupacha Melo-Sampaio et al., 2021, A. pachacamac Melo-Sampaio et al., 2021, A. snethlageae da Cunha & do Nascimento, 1983, A. dapsilis Melo-Sampaio et al., 2019, A. schach (Boie, 1827), and A. trefauti Melo-Sampaio et al., 2019. The latter two are sister species and their topological distance is smaller than intraspecific distances in other Atractus species sampled.

Figure 2. 

Distribution of Atractus arangoi, A. roulei, A. michaelsabini sp. nov., A. zgap sp. nov., and A. discovery sp. nov. in Ecuador and adjacent Colombia. White dots represent localities listed under Suppl. material 1. Each colored area is a geographic representation of the suitable environmental conditions for one of the clades recovered in the phylogeny of Fig. 1.

We find strong support for the relationship between members of the Atractus iridescens species group, which mirrors the results of Arteaga et al. (2017) and Murphy et al. (2019), and even those of Passos et al. (2022), although in the latter work some the terminals have been renamed. However, in the ML analysis (Suppl. material 2: Figure S1), A. dunni Savage, 1955 is weakly nested within A. microrhynchus Cope, 1868. Finally, we excluded A. imperfectus Myers, 2003 (voucher CH 9399) from the analyses as the short sequence available for comparison in GenBank (gene fragment 16S) represented a rogue taxon that assumed varying phylogenetic positions in the tree collection used to build the consensus tree.

We name or provide redescriptions 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), we describe three new species of Atractus.

 Atractus discovery sp. nov.

https://zoobank.org/0343A95C-BC4B-4654-8333-55D8A34CD2EF

Figs 3, 4, 5d Proposed standard english name: Discovery Ground Snake. Proposed standard spanish name: Culebra tierrera de Discovery.

Holotype

ZSFQ 4937 (Figs 3, 4), adult male collected by Alejandro Arteaga and Amanda Quezada at Amaluza, Azuay province, Ecuador (S2.61582, W78.56537; 2002 m).

Figure 3. 

Adult male holotype of Atractus discovery sp. nov. ZSFQ 4937 in a dorsal and b ventral view.

Figure 4. 

Head of the adult male holotype of Atractus discovery sp. nov. ZSFQ 4937 in a dorsal b ventral c lateral right, and d lateral left view.

Paratypes

ZSFQ 4936 (Fig. 5d), adult female collected by Alejandro Arteaga and Amanda Quezada at the type locality. MZUA.Re.466, adult female collected on 16 November 2018 at Campamento Arenales, Morona Santiago province, Ecuador (S2.59253, W78.56507; 2057 m).

Figure 5. 

Photographs of living specimens of brown-colored Atractus occurring along the Amazonian slopes of the Andes in Ecuador a A. arangoi ZSFQ 4948 from Jatun Sacha Biological Reserve, Napo province, Ecuador b A. resplendens ZSFQ 4953 from Montañas de San Antonio, Tungurahua province, Ecuador c A. duboisi from Orito Yacu, Napo province, Ecuador d A. discovery sp. nov. ZSFQ 4936 from Amaluza, Azuay province, Ecuador e A. orcesi ZSFQ 2234 from El Higuerón, Sucumbíos province, Ecuador f A. pachacamac from Nangaritza, Zamora Chinchipe province, Ecuador g A. zgap sp. nov. ZSFQ 4946 from Santa Rosa, Napo province, Ecuador h A. occipitoalbus JMG-2077 from Macas, Morona Santiago province, Ecuador i A. major from Jatun Sacha Biological Reserve, Napo province, Ecuador; and j A. major from Reserva Natural Palmarí, Amazonas state, Brazil (photo by Sebastián Di Doménico).

Diagnosis

Atractus discovery sp. nov. is placed in the genus Atractus, as diagnosed by Savage (1960), based on phylogenetic evidence (Fig. 1). The species is diagnosed based on the following combination of characters: (1) 17/17/17 smooth dorsals; (2) one postocular; (3) loreal 2.5–3 × longer than high; (4) temporals 1+2; (5) eight supralabials, fourth and fifth contacting orbit; (6) seven infralabials, first four contacting chinshields; (7) six or seven maxillary teeth; (8) one row of gular scales; (9) three preventrals; (10) 168 ventrals in the male holotype (Fig. 3b) and 170–172 ventrals in females; (11) 27 subcaudals in the male holotype and 17–18 subcaudals in females; (12) dorsal ground color light brown with faint stippling of a darker shade (Figs 3a, 5d); (13) venter yellow with a brown ventral stripe (Fig. 3b); (14) 284 mm SVL in the male holotype and 308–328 mm SVL in females; (15) 28 mm TL in the male holotype and 19–24 mm TL in females.

Comparisons

Atractus discovery sp. nov. differs from most of its congeners by having a bright yellow belly with a conspicuous dark brown longitudinal stripe. This species is compared to other small brownish congeneric ground snakes distributed along the Amazonian slopes of the Andes (most of these are pictured in Fig. 5): Atractus avernus Passos et al., 2009b, A. duboisi, A. ecuadorensis, A. zgap sp. nov., A. occipitoalbus (Jan, 1862), A. orcesi, and A. resplendens. From A. avernus, A. duboisi, A. occipitoalbus, and A. orcesi, the new species differs in having 17/17/17 (instead of 15/15/15) dorsal scale rows. From A. ecuadorensis, A. zgap sp. nov., and A. resplendens, it differs in having a bright yellow belly with a conspicuous dark brown longitudinal stripe. From A. ecuadorensis and A. zgap sp. nov., it further differs by having one (instead of two) postocular scale (Fig. 4c).

Description of holotype

Adult male, SVL 284 mm, tail length 28 mm (9.9% SVL); body diameter 7.8 mm; head length 8.8 mm (3.1% SVL); head width 5.6 mm (2.0% SVL); interocular distance 3.4 mm; head slightly distinct from body; snout-orbit distance 3.4 mm; rostral 1.6 mm wide, ca. as broad as high; internasals 0.9 mm wide; prefrontals 2.1 mm wide; frontal 2.9 mm wide, with a curvilinear triangular shape in dorsal view; parietals 2.2 mm wide, ~ 2 × as long as wide; nasal divided; loreal 2.0 mm long, ~ 3 × longer than high; eye diameter 1.1 mm; pupil round; supraoculars 1.3 mm wide; one postocular; temporals 1+2, upper posterior temporal elongate; eight supralabials, fourth and fifth contacting orbit; symphysial 1.0 mm wide, ~ 2 × as broad as long and separated from chinshields by first pair of infralabials; seven infralabials, first four contacting chinshields; chinshields ~ 2 × as long as broad, posterior chinshields absent; four rows of gular scales; dorsal scales arranged in 17/17/17 rows, smooth without apical pits; two preventrals; ventrals 168; anal plate single; 27 paired subcaudals.

Natural history

The three known specimens of Atractus discovery sp. nov. were found in open areas adjacent to cloud forest border. MZUA.Re.466 was crawling at ground level at around 7:30 pm. It was crossing a series of cement stairs. ZSFQ 4936 and ZSFQ 4937 were found during a cloudy day, buried 15–40 cm under soft soil at the border between the clearing of a graveyard, pastures, and remnants of native vegetation.

Distribution

Atractus discovery sp. nov. is known only from two localities (Arena­les and Amaluza, listed under Suppl. material 1: Table S1) on each side of the Río Paute, in the Ecuadorian provinces Azuay and Morona Santiago, at elevations 2002–2057 m a.s.l. The airline distance between the two localities is 2.6 km (Fig. 2).

Etymology

The specific epithet discovery is used as a noun in apposition and honors ‘The Explorers Club Discovery Expedition Grants’ (https://www.explorers.org/grants) initiative, a program seeking to foster scientific understanding for the betterment of humanity and all life on Earth and beyond. The grant program supports researchers and explorers from around the world in their quest to mitigate climate change, prevent the extinction of species and cultures, and ensure the health of the Earth and its inhabitants. ‘The Explorers Club Discovery Expedition Grants’ program funded the expedition that resulted in the discovery of this new species of snake.

Conservation status

We consider Atractus discovery sp. nov. to be Data Deficient, following IUCN Red List criteria, because the species belongs to a poorly studied genus of snakes and is known only from three specimens collected recently in a single river valley (Río Paute) in the Amazonian slopes of the Ecuadorian Andes. In addition to the presence of a system of major hydroelectric dams in this valley, most of the native cloud forest habitat in the segment between Amaluza and Arenales has been converted to pastures. However, we consider there is insufficient data to estimate whether this new snake species is restricted to the immediate environs of the type locality or if it is widely distributed along the unexplored cloud forests of the adjacent Sangay National Park.

 Atractus zgap sp. nov.

https://zoobank.org/A9A58D40-CF58-4267-A691-B5E776B43C1B

Figs 5g, 6, 7 Proposed standard English name: ZGAP Ground Snake. Proposed standard Spanish name: Culebra tierrera de ZGAP.

Holotype

ZSFQ 4946 (Figs 5g, 6, 7), adult female collected by Diego Piñán at Santa Rosa, Napo province, Ecuador (S0.31004, W77.78591; 1500 m).

Figure 6. 

Adult female holotype of Atractus zgap sp. nov. ZSFQ 4946 in a dorsal and b ventral view.

Figure 7. 

Head of the adult female holotype of Atractus zgap sp. nov. ZSFQ 4946 in a dorsal b ventral c lateral right, and d lateral left view.

Paratypes

MZUTI 5311, adult female collected by Diego Piñán in February 2017 at El Chaco, Napo Province, Ecuador (S0.31004, W77.78591; 1500 m). QCAZ 12666, a juvenile collected by Pablo Medrano on 16 May 2014 at San Francisco de Borja, Napo province, Ecuador (S0.40953, W77.84005; 1703 m). QCAZ 5183, a juvenile collected by Patricia Bejarano on 13 November 2011 at Bosque Protector “La Cascada,” Napo province, Ecuador (S0.14572, W77.49593; 1460 m).

Diagnosis

Atractus zgap sp. nov. is placed in the genus Atractus, as diagnosed by Savage (1960), based on phylogenetic evidence (Fig. 1). The species is diagnosed based on the following combination of characters: (1) 17/17/17 smooth dorsals; (2) two postoculars; (3) loreal 2 × longer than high; (4) temporals 1+2; (5) seven supralabials, third and fourth contacting orbit; (6) seven infralabials, first three contacting chinshields; (7) seven maxillary teeth; (8) three rows of gular scales; (9) two or three preventrals; (10) 173–177 ventrals in females; (11) 31 subcaudals in an uncollected male and 25–27 subcaudals in females; (12) dorsal ground color brown with faint dark longitudinal lines (Figs 5g, 6a); (13) venter yellow with fine brown stippling (Fig. 6b); (14) 376 mm SVL in the female holotype; (15) 37 mm TL in the female holotype.

Comparisons

Atractus zgap sp. nov. is compared to other small brownish congeneric ground snakes distributed along the Amazonian slopes of the Andes (most of these are illustrated in Fig. 5): Atractus avernus, A. duboisi, A. discovery sp. nov., A. ecuadorensis, A. occipitoalbus, A. orcesi, and A. resplendens. From A. avernus, A. duboisi, A. occipitoalbus, and A. orcesi, the new species differs in having 17/17/17 dorsal scale rows. From A. discovery sp. nov., the new species differs in having two postocular scales (Fig. 7c) and no dark ventral stripe. From A. ecuadorensis, the new species differs in having fewer (31 instead of 41) subcaudals in males, seven (instead of five or six) infralabials, a shorter (2 × instead of 3 × longer than high) loreal, frontal longer than prefrontals, and five faint (instead of six or seven clearly defined) longitudinal black lines (Figs 5g, 6). From A. resplendens, the new species differs in having a shorter (2 × instead of 3 × longer than high) loreal, two (instead of one) postoculars, and a brownish dorsum with faint longitudinal black lines, whereas in A. resplendens the dorsum is dark gray with fine yellow stippling (Fig. 5b).

Description of holotype

Adult female, SVL 376 mm, tail length 37 mm (9.8% SVL); body diameter 9.1 mm; head length 11.7 mm (3.1% SVL); head width 6.4 mm (1.7% SVL); interocular distance 4.3 mm; head slightly distinct from body; snout-orbit distance 3.8 mm; rostral 2.5 mm wide, ca. as broad as high; internasals 1.3 mm wide; prefrontals 2.5 mm wide; frontal 3.1 mm wide, with a curvilinear triangular shape in dorsal view; parietals 2.4 mm wide (56% length); nasal divided; loreal 1.6 mm long, ~ 2 × longer than high; eye diameter 1.7 mm; pupil round; supraoculars 1.2 mm wide; two postoculars; temporals 1+2; seven supralabials, third and fourth contacting orbit; symphysial 1.7 mm wide, ~ 2 × as broad as long, separated from chinshields by first pair of infralabials; seven infralabials, first three contacting chin shields; chinshields ~ 2 × as long as broad, posterior chinshields absent; dorsal scales arranged in 17/17/17 rows, smooth without apical pits; two preventrals; ventrals 173; anal plate single; 25 paired subcaudals.

Natural history

Most individuals of Atractus zgap sp. nov. have been found during the day hidden under rocks, among herbs, or buried under soft soil in plantations and rural gardens close to remnants of native forest. At night, they have been seen crossing rural roads. Occasionally, during sunny days right after a rain, individuals have been seen crawling on the pavement or on gravel roads (Diego Piñán, pers. comm.).

Distribution

Atractus zgap sp. nov. is known only from five localities (See Suppl. material 1: Table S1) along the valley of the Río Quijos, Napo province, in the Amazonian slopes of the Andes in northeastern Ecuador, at elevations 1460–1703 m a.s.l. (Fig. 2).

Etymology

The specific epithet zgap is used as a noun in apposition and honors the ‘Zoological Society for the Conservation of Species and Populations’ (ZGAP) (https://www.zgap.de), a program seeking to conserve unknown but highly endangered species and their natural habitats throughout the world. The ZGAP grant program supports the fieldwork of young scientists who are eager to implement and start conservation projects in their home countries. Specifically, ZGAP has supported the work on endangered Andean reptiles in Ecuador conducted by AA and JV.

Conservation status

We consider Atractus zgap sp. nov. to be Endangered following the IUCN criteria B2a, b (i, iii) (IUCN 2001), because the species’ extent of occurrence is estimated to be less than 500 km² (Fig. 2) and its habitat is severely fragmented and declining in extent and quality due to deforestation. The valley of the Río Quijos formed the eastern frontier of the Incan Empire (1400–1532) and the cloud forest in the area suffered from intensive land-use even before European arrival (Loughlin et al. 2018). Today, this valley is one of the most important cattle farming areas along the eastern slopes of the Andes and the majority of the forest along the Quijos river plains has been destroyed. Although A. zgap occurs in one protected area (Bosque Protector “La Cascada”) and its presence is expected in adjacent Parque Nacional Cayambe-Coca and Parque Nacional Sumaco Napo-Galeras, it has so far not been recorded in major protected areas.

 Atractus michaelsabini sp. nov.

https://zoobank.org/E85C68A2-DAEF-4BC5-A6B3-6D1FEDEB9983

Figs 8, 9, 10f–h Proposed standard English name: Michael Sabin’s Ground Snake. Proposed standard Spanish name: Culebra tierrera de Michael Sabin.

Atractus roulei Savage, 1960: 68 (part).

Atractus lehmanni Arteaga et al., 2017: 97.

Holotype

ZSFQ 4938 (Figs 8, 9, 10g), adult male collected by Jorge Luis Romero at Corraleja, Azuay province, Ecuador (S3.3874, W79.22785; 2660 m).

Figure 8. 

Adult male holotype of Atractus michaelsabini sp. nov. ZSFQ 4938 in a dorsal and b ventral view.

Figure 9. 

Head of the adult male holotype of Atractus michaelsabini sp. nov. ZSFQ 4938 in a dorsal b ventral c lateral right, and d lateral left view.

Paratypes

MZUTI 5289, adult female collected by Jorge Luis Romero at the type locality. AMARU 002 (Fig. 10f), adult female collected by Jorge Luis Romero at the type locality. ZSFQ 4939 (Fig. 10h), juvenile female collected by Jose Vieira and Amanda Quezada at El Panecillo, El Oro province, Ecuador (S3.46753, W79.48248; 2750 m). QCAZ 7887 and 7902, adult male and female collected by Silvia Aldás in December 2006 at Guanazán, El Oro province, Ecuador (S3.44667, W79.49051; 2663 m). QCAZ 9643 and 9652, adult females collected by Silvia Aldás in August 2009 at El Panecillo, El Oro province, Ecuador (S3.46753, W79.48248; 2775 m). DHMECN 7644–45, adult males collected by Mario Yánez-Muñoz, Luis Oyagata, Patricia Bejarano, and Marco Altamirano in March 2010 at Reserva Biológica Yunguilla, Azuay province, Ecuador (S3.22684, W79.27520; 1748 m). AMNH 18325, adult female collected in July 1920 at El Chiral, El Oro province, Ecuador (S3.63825, W79.59723; 1841 m). AMNH 22110–11, collected in August 1921 at La Chonta, El Oro province, Ecuador (S3.56585, W79.85144; 1025 m).

Figure 10. 

Photographs of living specimens of Atractus roulei and A. michaelsabini sp. nov. a A. roulei ZSFQ 4942 from Tixán, Chimborazo province, Ecuador b A. roulei ZSFQ 4944 from Tixán, Chimborazo province, Ecuador c A. roulei ZSFQ 4941 from Tixán, Chimborazo province, Ecuador d A. roulei ZSFQ 4945 from Tixán, Chimborazo province, Ecuador e A. roulei from Tixán, Chimborazo province, Ecuador f A. michaelsabini sp. nov. AMARU 002 from Corraleja, Azuay province, Ecuador g A. michaelsabini sp. nov. holotype ZSFQ 4938 from Corraleja, Azuay province, Ecuador and h A. michaelsabini sp. nov. ZSFQ 4939 from El Panecillo, El Oro province, Ecuador.

Diagnosis

Atractus michaelsabini sp. nov. is placed in the genus Atractus, as diagnosed by Savage (1960), based on phylogenetic evidence (Fig. 1). The species is diagnosed based on the following combination of characters: (1) 15/15/15 smooth dorsals; (2) one postocular; (3) loreal 3 × longer than high; (4) temporals 1+2; (5) five or six supralabials, with (usually) third and fourth contacting orbit; (6) five or six infralabials, with (usually) first three contacting chinshields; (7) 9–13 maxillary teeth; (8) 1–3 rows of gular scales; (9) 1–3 preventrals; (10) 143–144 ventrals in males and 144–153 in females; (11) 24–31 subcaudals in males and 17–19 in females; (12) dorsal ground color golden yellow (Figs 8, 10f–g) to dark brown (Fig. 10h) with each scale outlined in black, forming a reticulation; (13) venter yellowish with various degrees of brown stippling (Fig. 8b); (14) 256–321 mm SVL in males and 201–392 mm SVL in females; (15) 35–42 mm TL in males and 21–37 mm TL in females.

Comparisons

Atractus michaelsabini sp. nov. is compared to other members of the A. roulei species group: Atractus carrioni and A. roulei. From A. carrioni, the new species differs in having a loreal scale (Fig. 9c) (absent in A. carrioni). From A. roulei (Figs 10a–e), the new species differs in having a dorsal pattern in which each scale is outlined in a thin black line, thus creating a reticulation, and by having the prefrontal scale in broad contact with the postnasal (Fig. 9c) (not in contact or barely in contact in A. roulei). Furthermore, the existence of the bright golden yellow morph in adult individuals has so far been recorded only in A. michaelsabini sp. nov.; not in A. roulei, where adults are dark brown dorsally (Fig. 10a–e). In A. roulei, there is a black spot at the base of each dorsal scale, whereas in A. michaelsabini sp. nov. the spot is at the tip of each dorsal scale and is connected to the black reticulum. Genetic divergence in a 578 bp long fragment of the mitochondrial CYTB gene between A. michaelsabini sp. nov. and A. roulei is 6.5–7.2%, whereas intraspecific distances are 0–4.5% in A. michaelsabini sp. nov. and 0–4.8% in A. roulei.

Description of holotype

Adult male, SVL 256 mm, tail length 39 mm (15.2% SVL); body diameter 7.4 mm; head length 10.7 mm (3.1% SVL); head width 6.4 mm (2.5% SVL); interocular distance 3.7 mm; head slightly distinct from body; snout-orbit distance 3.5 mm; rostral 1.9 mm wide, ca. as broad as high; internasals 1.0 mm wide; prefrontals 2.0 mm wide; frontal 3.0 mm wide, with a curvilinear triangular shape in dorsal view; parietals 2.9 mm wide (65% length); nasal divided; loreal 2.2 mm long, ~ 3 × longer than high; eye diameter 1.4 mm; pupil round; supraoculars 1.3 mm wide; one postocular; temporals 1+2; five supralabials, third contacting orbit; symphysial 1.7 mm wide, ~ 3 × as broad as long, separated from chinshields by first pair of infralabials; five infralabials, first three contacting chinshields; chinshields ~ 2 × as long as broad, posterior chinshields absent; dorsal scales arranged in 15/15/15 rows, smooth without apical pits; no preventrals; ventrals 143; anal plate single; 31 paired subcaudals.

Natural history

Most individuals of Atractus michaelsabini sp. nov. have been found during the day hidden under rocks, mats of rotten vegetation, or buried in soft soil in pastures and maize plantations close to remnants of native forest. At night, they have been seen crossing forest trails. At the type locality, clutches of three or four eggs have been found under soil (Jorge Luis Romero, pers. comm.). Anecdotal information suggests that these snakes are more active during the rainy months (February-May at the type locality; Jorge Luis Romero, pers. comm.).

Distribution

Atractus michaelsabini sp. nov. is endemic to an estimated 2,530 km² area along the Pacific slopes of the Andes in southwestern Ecuador. The species occurs in the xeric inter-Andean valley of the Río Jubones as well as on the slopes of the Cordillera de Chilla. Atractus michaelsabini sp. nov. is known from provinces Azuay, El Oro, and Loja, and has been recorded at elevations between 927 and 2922 a.s.l. (Fig. 2).

Etymology

The specific epithet michaelsabini is a patronym honoring a young nature lover, Michael Sabin, grandson of American philanthropist and conservationist Andrew “Andy” Sabin. The Sabin family is involved in conservation and field research of amphibians and reptiles and has protected over 264,365 acres of critical habitat throughout the world.

Conservation status

We consider Atractus michaelsabini sp. nov. to be Endangered following the IUCN criteria B1a, b (i, iii) (IUCN 2001), because the species’ extent of occurrence is estimated to be much less than 5,000 km² (Fig. 2) and its habitat is severely fragmented and declining in extent and quality due to deforestation. Although A. michaelsabini sp. nov. is present in two protected areas (private reserves Buenaventura and Yunguilla of Fundación Jocotoco), nine of the 14 localities where the species has been recorded (Suppl. material 1: Table S1) are in heavily human-modified areas. Based on maps of Ecuador’s vegetation cover (MAE 2012), we estimate that nearly 70% of the forest cover throughout the species’ potential distribution area has been destroyed, mostly due to the expansion of the agricultural frontier.

Our resulting distribution maps increase the number of known localities of occurrence for the studied taxa (listed under Suppl. material 1: Table S1) and show a distinct geographical separation between Atractus roulei and A. michaelsabini sp. nov. (Fig. 2). The predicted area of suitable habitat for A. michaelsabini sp. nov. includes the upper watershed of the Río Jubones (a xeric inter-Andean valley) as well as both slopes of the Cordillera de Chilla (an area having vegetation classified as evergreen montane forest; see Sierra 1999). Likewise, the predicted area of suitable habitat for A. roulei includes evergreen montane forests along the Pacific slopes of the Andes as well as the xeric inter-Andean valley of the upper Río Chanchán. The predicted area of suitable habitat for A. arangoi includes almost the entire extent of Pastaza province, although we did not find records of this species from this province. Although we did not build binary environmental niche models for A. discovery sp. nov. and A. zgap sp. nov. (only two and six localities are available for these species), they are both known only from their corresponding river valleys and occur on both sides of the Río Paute and Río Quijos, respectively.

Prado (1939) described Atractus arangoi from Colombia whereas Daniel (1949) reported this species in Puerto Asís, Putumayo department. Schargel et al. (2013) considered A. arangoi to be a junior synonym of A. major claiming that all the putative diagnostic characters for A. arangoi fall within the variation in A. major as defined in their work. In our phylogenetic tree of Atractus (Fig. 1), we included sequences of three snakes that fit the original description of A. arangoi. DHMECN 8343 (reported as A. major in Arteaga et al. 2017), ZSFQ 4947 (Fig. 11), and ZSFQ 4948 (Fig. 5a). These three specimens form a strongly supported clade sister to all other samples of A. major, which includes specimens from throughout the latter species’ area of distribution. Furthermore, we find that these specimens, in addition to others reported in the literature as A. torquatus and A. major (see Duellman 1978; Maynard et al. 2017) can easily be separated from A. major based on differences in coloration, body size (compare Figs 5a and 5i, j), and ventral and subcaudal counts (summarized in Table 3), as originally suggested by Prado (1939). Thus, we formally remove A. arangoi from the synonymy of A. major, include this species in the herpetofauna of Ecuador, and provide a distribution map for this species (Fig. 2).

Passos et al. (2010) reported Atractus gigas, a snake species previously considered to be endemic to the cloud forests of northwestern Ecuador (Myers and Schargel 2006), on the Amazonian slopes of the Andes in Peru. The identification of the Peruvian specimens as A. gigas was based on their large size and the partial overlap in some characters of lepidosis with the Ecuadorian samples. However, these Peruvian snakes have a smaller number of subcaudals (25 or 26 instead of 31–37 in Ecuadorian specimens), a shorter loreal scale, first four infralabials contacting chinshields (instead of first three in Ecuadorian specimens), and a completely different color pattern in both juveniles and adults (for a figure depicting the variation among Ecuadorian individuals see Arteaga 2022). Juveniles of “A. gigas” from Peru have a black dorsum with short (one scale wide) reddish brown bands whereas juveniles of Ecuadorian A. gigas have a contrasting pattern of dark-brown to black rounded bands or blotches on a rosy white background color. Adults of “A. gigas” from Peru have a dorsal pattern in which each scale is dark brown distally but cream towards the base, forming a reticulation. Adults of A. gigas from Ecuador are uniformly rich dark brown or glossy black, and the skin between the scales is whitish (Arteaga 2022). QCAZ 14946, a specimen identified as A. atlas in Melo-Sampaio et al. (2021) from Reserva Biológica Cerro Plateado, just 7 km from the Peruvian border on the southeastern slopes of the Ecuadorian Andes, resembles Peruvian “A. gigas” as depicted in Passos et al. (2010) in having a short loreal, dorsal scales with a cream base, first four infralabials contacting chinshields, and fewer than 30 subcaudals. This specimen was included in our phylogeny (Fig. 1) and was recovered as the strongly supported sister taxon to a new sample of A. touzeti from this species’ type locality. Based on this evidence, we suggest that Peruvian specimens CORBIDI 877 and ZFMK 89147, as well as other Atractus specimens from Cajamarca labeled as A. gigas, be reidentified as A. atlas, or at the very least, be considered as an undescribed species related to the latter. Thus, we suggest A. gigas be removed from the herpetofauna of Peru, a view that confirms this species as endemic to the cloud forests of northwestern Ecuador as originally suggested by Myers and Schargel (2006) and Arteaga et al. (2013).

Figure 11. 

Adult male of Atractus arangoi ZSFQ 4947 in a dorsal and b ventral view.

In his unpublished BSc thesis, Mejía Guerrero (2018) used species distribution models, a comprehensive (based on 88 specimens) comparison of scale counts, and species delimitation analysis based on a combination of novel DNA sequences and those provided in Arteaga et al. (2017) to test species limits within the Atractus iridescens species group. He proposed that A. occidentalis Savage, 1955 is a junior synonym of A. microrhynchus and that some individuals identified as A. dunni from Mindo are actually A. microrhynchus. The topology for the included members of the A. iridescens group in our BI phylogeny (Fig. 1) and that of Murphy et al. (2019), though not identical, agree with the proposal of Mejía Guerrero (2018). Based this evidence, we also consider A. occidentalis to be a junior synonym of A. microrhynchus. Recently, Passos et al. (2022) provided a list of reidentifications of 15 (not 17, because two are duplicates and MZUTI 4178 retained the same identification despite being listed in the table) Atractus specimens having sequences deposited in GenBank, notably among them the members of the A. iridescens species group deposited in MZUTI and DHMECN. In this work, one reidentification (that of the holotype of A. pyroni; MZUTI 5107) was backed up by ample evidence and two others (ANF 2390, now MZUTI 5409; and GFM 307, now MPEG 21582) were substantiated in Melo-Sampaio et al. (2021), but the remaining were proposed without providing any evidence, either in the form of new phylogenetic relationships, new scale counts, or previously unsampled morphological features. Since these specimens are deposited at MZUTI and DHMECN, as well as their corresponding photo vouchers available in Arteaga et al. (2017), and their DNA sequences on GenBank, their identity can be tested by anyone. Although the reidentification of the remaining specimens provided by Passos et al. (2022) was unsubstantiated, not all of them were unwarranted (see Table 4). We agree that DHMECN 7644 (identified as A. lehmanni Boettger, 1898 in Arteaga et al. 2017) and IBSP 71932 (identified as A. zebrinus Jan, 1862 in Grazziotin et al. 2012) are misidentified, but their new identifications provided by Passos et al. (2022) are not correct either (see Table 4). DHMECN 7644 is a paratype of A. michaelsabini sp. nov., as defined herein, and IBSP 71932 is probably an A. trihedrurus Amaral, 1926, not an “A. triherurus.” Although the latter probably represents a typo and is a minor error, the problems with the remaining reidentifications are not trivial. For example, Passos et al. (2022) reidentified the same specimen, MZUTI 3758, as A. iridescens Peracca, 1896 and also as A. cf. iridescens. Additionally, these authors completely reidentified the type series of both A. cerberus and A. esepe Arteaga et al., 2017, probably without much confidence since this action is not explained elsewhere in their work and is not trivial. Since MZUTI 4330 and MZUTI 3758 are name-bearing specimens, reidentification of these holotypes as A. iridescens, A. cf. iridescens, or anything other than their original identification presented in Arteaga et al. (2017) implies that these species are not valid. Surprisingly, the fact that the taxonomic validity of these two species is not questioned elsewhere in Passos et al. (2022) suggests that some of these reidentifications were proposed carelessly. Thus, in Table 4, we evaluate these reidentifications and mention whether they are substantiated or warranted or neither. Finally, we propose the reidentification of an additional six Atractus specimens (Table 5) having sequences deposited in GenBank based on the results presented in Fig. 1.