The geographic distribution of Atractus lehmanni (Boettger, 1898) (Serpentes, Colubridae, Dipsadinae) in Colombia, and clarification of its status and type locality in Ecuador

Atractus lehmanni (Boettger, 1898), Lehmann’s Ground Snake, (Serpentes, Colubridae, Dipsadinae), was described by Boettger (1898) from Cuenca, Ecuador. We examined records of snakes labeled as A. lehmanni to determine if they fit the original description of this species. The results of our record examinations, in conjunction with long-term field surveys and a review of Friedrich Carl Lehmann’s travel logs, indicate that A. lehmanni occurs in the Cordilleras Central and Occidental of Colombia. Conversely, this species is apparently absent from Ecuador, where records of this species are in error or based on misidentifications.


Introduction
Ground snakes (genus Atractus Wagler, 1828) comprise the most species-rich genus of modern-day snakes (class Reptilia, order Squamata, suborder Serpentes), consisting of 147 species (Uetz et al. 2021) collectively distributed from Panama (Köhler 2008) to Argentina (Nogueira et al. 2019) and Uruguay (Carreira et al. 2012), with the area of highest diversity in northwestern South America (Uetz et al. 2021). Although collectively widespread, species in this genus are largely fossorial and secretive. Most species, particularly those that inhabit montane areas, are known from limited or remote areas, and thus are rarely observed, collected, or studied. In fact, 64 species have been discovered or described since the turn of this century (Uetz et al. 2021). However, knowledge of the life histories and geographic distributions is incomplete even for some of the more widespread species described over a century ago, including A. lehmanni (Boettger, 1898), Lehmann's Ground Snake.
Atractus lehmanni was described by Boettger (1898) from a type series consisting of three female and three male specimens, from which a syntype was designated (MC 33513) and described in additional detail by Savage (1960). The type series was collected by Friedrich Carl Lehmann, reportedly from Cuenca in southern Ecuador. Although Lehmann visited Cuenca, he lived in and made most of his collections near Popayán, Colombia, over 600 km north of Cuenca, where specimens that fit Boettger's (1898) and Savage's (1960) descriptions of A. lehmanni have been found almost throughout the entire municipality. All specimens from Ecuador that Arteaga et al. (2017) initially identified as A. lehmanni and that Torres-Carvajal et al. (2019) subsequently examined are apparently misidentified. Therefore, we sought to investigate and clarify the status and distribution of this species.

Methods
We examined records labeled as A. lehmanni to determine those that fit Boettger's (1898) and Savage's (1960) descriptions of this species in scale characteristics and color pattern. We obtained vouchered records from the Instituto de Ciencias Naturales, Universidad Nacional de Colombia, Bogotá (ICN); Museo de Historia Natural, Universidad del Cauca (MHNUC); Museo de Herpetología, Universidad de Antioquia (MHUA); Museo de Historia Natural de La Salle, Bogotá (MLS); Universidad de Caldas (UCALDAS); Universidad del Quindío (ARUQ); and Universidad Valle del Cauca (UV-C). We reviewed unvouchered records from iNaturalist and also examined records of snakes that were identifiable as A. lehmanni from our own field surveys. We mapped confirmed records using ArcMap 10.8.1, provided they were ≥ 2 km from all other confirmed records and associated with specific coordinates, as well as the location of Cuenca from which the type series was reported (Boettger 1898). Besides the type series, we did not map records that we thought represented erroneous locations (e.g., far northern or eastern Colombia) because of georeferencing errors.

Results
We confirmed 66 records of A. lehmanni that were associated with specific coordinates and appeared to be properly georeferenced (Table 1). We mapped all 31 of these records that were ≥2 km from all other confirmed records  Table 1). Map inset illustrates the general location of confirmed records relative to the unconfirmed type locality of Cuenca, Ecuador. The red star indicates the location of the new record reported herein (UF Herp 192663). Table 1. Confirmed locality records ≥2 km apart of Atractus lehmanni in Colombia. Reference and voucher/evidence data on all specimens <2 km apart are listed within the same record. For records represented by specimens from multiple sites < 2 km apart, the specific specimen(s) to which the listed locality data pertain is (are) indicated with an asterisk (*). Only records for which scale characteristics and color pattern matched Boettger's (1898) and Savage's (1960)  Identification. Identification was made by comparing the scale characteristics and color pattern of the Quindío specimen ( Fig. 2A) to the descriptions by Boettger (1898) and Savage (1960). The Quindío specimen matched Boettger's (1898) and Savage's (1960) descriptions of scale characteristics as follows: 17 middorsal scale rows; a small rostral and internasals; long prefrontals; 2 postoculars; 3 (1 + 2) temporals; third, fourth, and seventh supralabials in contract with the orbit; third, fourth, and sixth infralabials in contract with the chin shields; 150 ventrals and 20 subcaudals. The Quindío specimen matched Boettger's (1898) and Savage's (1960) descriptions of color pattern as follows: dorsum dark brown; each dorsal scale above rows 1 and 2 darkest on lower and posterior margins, and with minute black markings; indistinct dark vertebral stripe present; light nuchal collar suffused with dark pigment; top of head mostly dark, but prefrontal-internasal region with some light areas; supralabials and postnasals mainly dark, but lower portions light; throat and chin light, but with large brown spots on infralabials, mental, and chin shields; lateral part of ventrals and dorsal scale rows 1 and 2 light; medial portion of ventrals dark brown; anal plate brown; underside of tail light, with irregular brown mottling on median portion. Similar identifications were made by comparing the other vouchered and unvouchered specimens that were labeled as A. lehmanni to the descriptions by Boettger (1898) and Savage (1960). The variation in scalation and color pattern among the specimens that we examined and confirmed as A. lehmanni falls well within the ranges reported in the descriptions by these authors (Fig.  2B-D). However, we did not confirm any specimens that exhibited variation outside these ranges as A. lehmanni. Moreover, no species of Atractus that has been described subsequent to Boettger (1898) and Savage (1960) exhibits the same suite of scale characteristics and color pattern as described for A. lehmanni by these authors.

Discussion
Several surveys and taxonomic studies on Atractus have recently been conducted across the cordilleras of Colombia and Ecuador, resulting in the documentation and description of numerous species to the north and south of where specimens that fit Boettger's (1898) and Savage's (1960) descriptions of A. lehmanni occur, exclusive of the reported type locality of Cuenca (e.g., Passos et al. 2009a;Passos and Lynch 2010;Arteaga et al. 2017;Torres-Carvajal et al. 2018). As a result, although still incomplete, knowledge of the distributions of Atractus across these areas has increased substantially. Therefore, the collective area over which A. lehmanni has been found in the Cordilleras Central and Occidental of western Colombia probably constitutes the vast majority of this species' actual extent of occurrence.
Although much of the area in which A. lehmanni occurs is remote and difficult to access, and thus its area of occupancy is still incompletely known, this species is common in urban areas and can survive in heavily disturbed habitat, as was readily evident at the locations where J.G. Himes, K.M. Enge, L.E.V. Pérez, and J.A. Rojas-Morales (2012; listed as "Atractus sp.") found it. Moreover, over 12 years of field surveys for snakes by A. Arteaga and colleagues have failed to find any specimens that fit the description of A. lehmanni in or near the reported type locality of Cuenca (Boettger 1898) or elsewhere in Ecuador. Instead, specimens from Ecuador identified as A. lehmanni (Arteaga et al. 2017) have subsequently been identified as A. roulei in a more recent work (Passos et al. 2022) that includes a more comprehensive taxon sampling. These specimens have dorsal scales arranged in 15 rows at mid-body, as opposed to the diagnostic 17 rows of A. lehmanni. Therefore, the occurrence of A. lehmanni in or near Cuenca, which is over 600 km south of all other records of this species, is highly improbable. Far more likely, the type locality of A. lehmanni is in or near Popayán, where Lehmann made most of his herpetological collections, and this species is confined to the portions of the Cordilleras Central and Occidental of western Colombia where all other specimens of A. lehmanni have been found.
Besides the fact that it is unlikely that A. lehmanni occurs in or near Cuenca or elsewhere in Ecuador, its presence in both the Cordillera Central and Occidental of Colombia is exceptional among other montane-inhabiting species of Atractus, most of which are restricted to a single side of one of the three cordilleras (Central, Occidental, or Oriental) (Passos et al. 2009a;Vanegas-Guerrero et al. 2014). Only six other species are definitively known from multiple cordilleras, with only A. nicefori being known from both the Cordilleras Central and Occidental; all others are known from the Cordilleras Central and Oriental (Passos and Lynch 2010;Passos et al. 2009b). Thus, although specimens from the Cordillera Occidental fit Boettger's (1898) and Savage's (1960) descriptions of A. lehmanni in terms of scale characteristics and color pattern, genetic analyses and comparisons between specimens from the Cordilleras Occidental and Central are necessary to further determine their taxonomic identities and relationships. In fact, the portion of the Cauca River valley that is located between the confirmed records of A. lehmanni in the Cordillera Central and Oriental drops to under 1000 m alt., whereas the lowest records of A. lehmanni were at 1620 and 1694 m alt. within these respective cordilleras. Therefore, gene flow has probably ceased between populations within the different cordilleras, and thus each population may be currently evolving along separate trajectories.