Distribution and identification of the species in the genus Helicops Wagler, 1830 (Serpentes, Colubridae, Xenodontinae)

Abstract Background The aquatic snakes of the genus Helicops are widely distributed throughout northern South America, but understudied concerning some aspects, including morphological traits and distribution. The most recent publication that provided an identification key to all species of Helicops is over 50 years old. This key is of limited value today since it includes taxa no longer recognised and lacks 8 of the 19 species currently recognised. There never was a publication trying to summarise distributional and morphological information of all species of Helicops. Most knowledge of these species is distributed throughout many small publications, such as short observation notes. New information Here, we present distribution maps (point records), an identification key and comments on identification for all species in this genus. We base our results on a comprehensive literature review of over 300 scientific publications and own examinations. Our examinations comprise 190 specimens of 10 of the 19 currently recognised species and one Helicops sp. We report range extensions for the species H.danieli, H.infrataeniatus, H.leopardinus, H.pastazae and H.polylepis.


Introduction
Water snakes of the genus Helicops are widely distributed, mainly in the northern half of South America, but they also reach Uruguay and central Argentina. The genus currently comprises 19 species of aquatic snakes inhabiting nearly all kinds of water bodies within their distribution range, from small ponds and puddles to slow-flowing streams, also in urban areas (França et al. 2012, Hernández-Ruiz et al. 2014, Koski et al. 2016). However, for many species, not much more than their description is known. The genus Helicops belongs to the Hydropsini tribe together with the genera Hydrops and Pseudoeryx. The Hydropsini are part of the subfamily Dipsadinae and characterised by the wide origin of the superficialis muscle (Di Pietro et al. 2014). All 19 species of the genus Helicops share the combination of having eyes and nostrils in a dorsal position on the top of the head, a single internasal scale, a divided cloacal shield and at least some keeled dorsal scales (Costa et al. 2016). The first description of the species in this genus was Helicops angulatus, which was described as Coluber angulatus and C. alidras by Linnaeus (1758). In 1830, Wagler (1830) assigned C. angulatus and other species of the genera Coluber and Natrix to the newly-created genus Helicops. The last identification key was published by Peters and Orejas-Miranda (1970). In this work, the authors recognised 13 species, including the no longer recognised H. hogei (currently synonym of H. scalaris) and H. pictiventris (currently synonym of H. infrataeniatus) (Rossman 2002). Since then, H. infrataeniatus has been raised to species level again and seven new species have been described (H. apiaka, H. boitata, H. nentur, H. phantasma, H. petersi, H. tapajonicus and H. yacu). Of these, more than half were discovered in the past 10 years (H. apiaka, H. boitata, H. nentur and H. phantasma). For these seven new species, not much more than their description is known. This and the number of recently described new species show the missing taxonomic overview for this genus. Most of the knowledge is scattered across many publications. Especially distributional information is mostly presented in observation notes. Therefore, we aim to present a basis for further taxonomic studies, by providing point records, an annotated checklist and an identification key, based on a comprehensive literature review and own observations.

Materials and methods
We base our species assessment of the genus Helicops on the morphological examination of 190 specimens, representing 10 of the 19 currently recognised species in this genus and on a comprehensive literature review. The examined specimens are located in six herpetological museum collections in Germany: Senckenberg Research Institution Frankfurt (SMF); Senckenberg Naturhistorische Sammlungen Dresden (MTKD); Zoologisches Forschungsmuseum Alexander König in Bonn (ZFMK); Zoologische Staatssammlung München (ZSM); Staatliches Museum für Naturkunde Stuttgart (SMNS) and Naturkundemuseum Berlin (NMB).
The examined morphologic characters were: snout-vent length (SVL), tail length (TL), the ratio between tail length and snout-vent length ( For male and female specimens, we recorded the number of ventral scales, number of subcaudal scales, snout-vent length, tail length and the ratio between snout vent length and tail length separately. Measurements were taken using a millimetric tape measure and ventrals were counted as proposed by Dowling (1951). We determined the sex of each specimen by exterior examination of the shape of the tail base (tail base bulge caused by presence of hemipenes in males, bulging absent in females). The results of our examinations are available in Suppl. material 1. We did not take tail measurements in specimens with caudal damage. Head scutellation was recorded for each side separately.
We base the species' distribution summaries on locality data of the examined specimens and additionally on literature data. One further locality was detected by browsing through iNaturalist (https://www.inaturalist.org/observations/9053312). We could identify the species (H. polylepis) by its unique colouration. Only records were included for which a reliable description of the locality was available, i.e. a map which enables the extraction of the coordinates or they were provided directly. In Nogueira et al. (2019), we were not able to unambiguously identify some of the cited literature, even with the help of the authors. Therefore, we simply pass their citation to our supplement with a respective note. We treat distribution records with ambiguous references separately in the distribution maps and supplements. All details on literature used and references for the distribution records are listed in Suppl. material 2.
We created the distribution maps using QGIS 3.12.2 and maps freely available at naturalearthdata.com.
We created the identification key using the morphological data gathered by examining specimens and literature data. Literature references used for morphology are listed in Suppl. material 3.

Data resources
Suppl. materials 1, 2 contain tables in tab delimited text format. Suppl. material 1 contains the examination results for each specimen. It has columns with following headers: Species; Catalog numb.; Sex; Snout-venth length [mm]; Tail  Suppl. material 3 is a plain text file containing all the references used for the morphological assessment. Each reference is provided in a separate line.

Distribution
The only known specimens of H. apiaka are from northern Mato Grosso and southern Pará (Fig. 1b).

Morphology remark
Regarding the number of dorsal scale roles at mid-body in H. apiaka, there is contradictory information. Moraes-da-Silva et al. (2019)

Diagnosis
Helicops danieli is readily distinguished from its congeners by its unique colour pattern, namely a spotted dorsum in combination with a ventral pattern consisting of two rows of semi-lunar marks on a light background (for information on references see, Suppl. material 3).

Distribution
Helicops danieli is only occurring in Colombia, where it seems to be found mainly west of the Andes. There is a report from the lowland in the east near the Brazilian border (

Distribution
The distribution of H. hagmanni ranges from the Estuary of the Amazonas to the Brazilian Provinces Amazonas, Acre, Rondônia and the Venezuelan Province Amazonas. There is also one record from south-western Colombia (Rossman 1975, Fig. 2a).

Morphology remark
The examined specimens had smooth subcaudal scales on the anterior part of the tail, changing to weakly-keeled scales at the posterior tail, which contrasts with the examination results in Moraes-da-Silva et al.   (Fig. 2b).

Distribution
Helicops leopardinus records range from north-western Argentina to the Estuary of the Amazon and from Ecuador to the Brazilian State Bahia. There are nearly no records in the south-eastern provinces of Brazil (Fig. 2c).

Distribution
Helicops modestus is occurring from the Brazilian Province Bahia to the Province Paraná and seems to range from the east shore of Brazil to the south of Mato Grosso. There is also one literature record from Volta Grande do Xingu in the Brazilian Province Para, near its estuary into the Amazon (Vaz-Silva et al. 2015, Fig. 2d).

Distribution
Helicops polylepis is recorded from southern Bolivia to the Amazon Estuary and from Peru to the east of the Brazilian Province Para. There are also two reports from Colombia (Fig. 3c).

Distribution
Helicops scalaris is known only from the northern border area between Colombia and Venezuela, western and north of Lake Maracaibo in Venezuela (Fig. 3d).

Helicops tapajonicus da Frota, 2005
Diagnosis This species can be distinguished from all its congeners by its unique colour pattern, namely the combination of a uniform moss-green dorsum, laterally with a greenishyellow stripe and a black and greenish-yellow banded venter (for information on references, see Suppl. material 3).

Distribution
Helicops tapajonicus is only known from two localities at the River Tapajos close to its confluence with the Amazon in the Brazilian State Para (Fig. 3e).

Diagnosis
This species can be distinguished from all congeners by the unique colour pattern, namely a combination of five narrow light stripes on the dorsum and a light venter with black semi-lunar markings, which extend on to the tail (for information on references, see Suppl. material 3).

Distribution
Helicops trivittatus is present from the eastern part of the Brazilian Province Para to approximately its borders with Maranhao, Tocantins and northern Mato Grosso. There are no reports of this species from western Para (Fig. 3f).

Morphology remark
The presence of intergenials seems to be a reliable identification character in all other species of this genus, whereas in H. trivittatus, this character shows considerable variation. Intergenials are sometimes present. In our dataset, there were two specimens with and two without intergenials. There is no obvious biogeographical pattern perceiveable (pers. Comm. Antonio Moraes-da-Silva).

Helicops yacu Rossman & Dixon, 1975
Diagnosis Helicops yacu can be distinguished from all congeners by the combination of having an entire nasal scale and intergenials present (for information on references, see Suppl. material 3).

Distribution
Helicops yacu is known only from north-eastern of the Province Loreto, Peru and one locality in north-western Acre, Brazil (Rossman andDixon 1975, Nogueira et al. 2019, Fig. 4).

Taxonomic remark
In Rossman and Abe (1979) Table 3; for information on references, see Suppl. material 3).

Distribution
The specimen originates from the Province Pernambuco in Brazil, no exact locality is available.

Identification keys
Identification key to the species of Helicops Wagler, 1830 Dichotomous identification key, based on our own examinations and literature (listed in Suppl. material 3).
1 Dorsum uniform or with longitudinal stripes 2

Taxonomic discussion
The last published identification key by Peters and Orejas-Miranda (1970) is by now outdated and many taxonomic changes have since taken place. Therefore, we evaluated the suitability of identification characters proposed in literature. Based on this revision, we created a completely restructured identification key. Regarding the head scutellation, we found that only the presence of intergenial scales is a stable diagnostic character s in all species, except H. trivittatus. In this species, specimens with and without intergenials occur without a geographical pattern (pers. comm. Antonio Moraes-da-Silva).
Colouration seems to be a rather good character for distinguishing some species (e.g. H. trivttatus) and species groups, but this might change with further molecular studies and the possible identification of cryptic species. We could not find a pholidotic character in order to distinguish H. carinicaudus from H. infrataeniatus with neither head scalation showing clear differences nor the ratio TL/SVL. We found them to differ only in ventral colouration, which agrees with Peters and Orejas-Miranda (1970) In Murphy et al. (2020), the authors resurrected the species Helicops cyclops, which was previously considered to be a synonym of H. angulatus. However, the resurrection is only based on a photograph of a single specimen, which was not examined by the authors themselves. We think the observed differences to H. angulatus (short snout, shortened chin shields and a distinct dark band between the eyes) could also be explained by individual variation, especially in a species with such a large distribution range. The authors, furthermore, list a higher number of ventral scales (124 ventrals) as diagnostic character, which is well in the range we observed. Therefore, we do not include this species in our key. We encourage a re-evaluation including genetic data and a proper redescription of the species to clarify its taxonomic status.
The presence versus the absence of subcaudal keels seems to be a stable character in most species. However, we found conflicting reports for this character in H. infrataeniatus for which Kawashita-Ribeiro et al. (2013) reported subcaudal keels, whereas Moraes-da-Silva et al. (2019) stated that these are absent in this species. Amongst the 58 specimens of this species, only one (SMNS 3065) possessed subcaudal keels. This finding might be explained by variation throughout the large distribution range of this species. It could be a sign that distinct genetic lineages exist, as we experienced this character extremely stable in the other species.
We think it is possible that some species with supposedly large geographical distributions actually comprise species complexes. For example, Murphy et al. (2020) already showed cryptic diversity in what is currently recognised as H. angulatus. We expect future integrative studies including genetic data to reveal cryptic and, therefore, yet undiscovered species.

Geographical extensions
When comparing distribution data from our examined specimens with literature records, we discovered range extensions for five species. For H. carinicaudus, we report one specimen collected in 1935 in Porto Alegre, Rio Grande do Sul, Brazil. This would represent a distribution range extension of over 150 km from the closest literature record by Deiques and Cechin (1991). However, this species is not known to occur so far south (Nogueira et al. 2019). It might only have been shipped from Porto Alegre. For H. danieli, we report the first specimen for Brazil. The specimen ZMB 9490 has no precise locality data. The nearest literature report of this species to Brazil is only 20 km away from the Columbian-Brazilian border at Mitu, in the Columbian Province Vaupes (Yuki and Castano 1998). For H. infrataeniatus, we report an expansion of the distribution range approximately 50 km further west from the nearest record by Giraudo (2001)  Finally, we report this species from the Province of Bahia in Brazil. Specimen ZMB 17428 was collected there, but no exact locality data are available.
The number of range extensions we report shows that the distribution ranges of the species in this genus are not yet well known. In order to change that, a comprehensive examination of collected material at an international level and additional fieldwork are required.