Revealing two centuries of confusion: new insights on nomenclature and systematic position of Argyrogena fasciolata (Shaw, 1802) (auctt.), with description of a new species from India (Reptilia: Squamata: Colubridae)

Coluber fasciolatus Shaw, 1802 (today Argyrogena fasciolata auctt.) is the name used for a widespread terrestrial colubrid snake species inhabiting subtropical and tropical dry deciduous/thorn forests of South Asia from Pakistan to India, with unconfirmed records of distribution in Nepal, Bangladesh and Myanmar and a single doubtful record from the northern tip of Sri Lanka. During the past 200 years, A. fasciolata (common name Banded racer) has been placed in different genera, i.e. Tyria Fitzinger, 1826, Zamenis Wagler, 1830, Coryphodon Duméril, Bibron and Duméril, 1854, and Coluber Linnaeus, 1758 where it primarly remained until the mid 1960s and exceptionally until the year 2011. Three subsequently introduced names, viz. Coluber hebe Daudin, 1803, Coluber curvirostris Cantor, 1839, and Argyrogena rostrata Werner, 1924 were synonymized with C. fasciolatus shortly after its description. Based on a combination of characters including body pattern, external morphology and osteological differences Wilson (1967) reviewed the taxon fasciolatus Shaw and considered it as generically distinguishable, removed it from the then heterogeneous and undefined collective genus Coluber and assigned it to the resurrected genus Argyrogena Werner, 1924. Shaw’s (1802) description of C. fasciolatus was based exclusively on the information of Russell’s “Nooni Paragoodoo” published in 1796 in his “Account of Indian Serpents, collected on the coast of Coromandel; [...]”. Our analysis of the original data and the depicted type specimen in Russell (1796) revealed that the name fasciolata was initially established for a species distinct from that currently known as the “Banded racer”, and that Russell’s data have been used simultaneously but unwittingly, for more than 150 years, as original source for two valid species from two different genera. Specimens of Banded racer found in the southeastern part of peninsular India are morphologically and genetically distinct from populations of the rest of the distribution area. These populations from central and southern Tamil Nadu state represent a different Vertebrate Zoology 71, 2021, 253–316 | DOI 10.3897/vz.71.e64345 Copyright V. Deepak et al. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. V. Deepak et al.: new insights on nomenclature and systematic position of Argyrogena fasciolata 254 species, consequently described as a new species herein. Furthermore, examination of specimens of the Banded racer from different populations across its entire range, including the type specimen of the genus Argyrogena (A. rostrata Werner), reveals a similarity in morphology with the genus Platyceps Blyth, 1860. This was further supported by molecular data which demonstrates that the genus Argyrogena is nested within Platyceps.


Introduction
The Banded racer, presently named Argyrogena fasciolata (auctt.), is a colubrid snake species with an assumed wide but scattered distribution in the Indian subcontinent and has an unstable taxonomic history. This species was first mentioned by the Scottish surgeon and naturalist Patrick Russell (1796, p. 26 f., pl. 21) when he described and depicted a "Coluber", called "Nooni Paragoodoo" by the natives, in his monumental book about the "[…] Indian Serpents, collected on the Coast of Coromandel […]". Russell's enormous effect on the systematic herpetology of South and Southeast Asia is evidenced through numerous descriptions of Asian snakes by subsequent authors, including G. K. Shaw, F. P. Nodder, J. G. Schneider, P. A. Latreille, F. M. Daudin, M. Bechstein, B. Merrem, C. S. Rafinesque-Schmaltz, T. Bell, H. Boie, F. Boie, L. J. Fitzinger, and Ph. Schmidt who relied exclusively on Russell's text and images (see also . Based on  information, the English zoologist George Kearsley Shaw (1802) formally described Coluber fasciolatus and since then the specific epithet has been used in binomial combinations with different genera, first used by the following authors as: Coluber Shaw (1802), Coryphodon , Tyria Cope (1863), and Zamenis Günther (1864) (see list of synonyms and chresonyms). Werner (1924 ) described a new genus and species under the name Argyrogena rostrata but that taxon was synonymized with C. fasciolatus by Smith (1928). Wilson (1967) reviewed morphology and nomenclature of the C. fasciolatus and placed the taxon in the monotypic genus Argyrogena erected by Werner (1924), the combination predominantly used until today.
Our critical analysis of the original source  and the subsequent interpretation of his descriptions raises serious doubts regarding the present allocation of the name Argyrogena fasciolata to what is commonly regarded as "Banded racer" today. Russell's description of the "Nooni Paragoodoo" and the species fasciolatus Shaw and hebe Daudin based thereon, show remarkable differences in regard of colour, pattern and pholidosis in comparison to our own and published data on the Banded racer. The present study clarifies the taxonomic status of Coluber fasciolatus sensu Shaw. We review nomenclature, taxonomic allocation, morphology, and distribution of the Banded racer and describe a related but distinct new species on the basis of morphological and genetic differences. Furthermore, we provide genetic data for other related colubrid snakes found in India including Platyceps gracilis (Günther, 1862), P. ventromaculatus (Gray, 1834), P. bholanathi (Sharma, 1976), Lytorhynchus paradoxus (Günther, 1875), Wallaceophis gujaratensis Mirza, Vyas, Patel, Maheta and Sanap, 2016 and Wallophis brachyura (Günther, 1866).

Sampling
We generated DNA sequence data for eight specimens of Argyrogena (auctt.) collected from different parts of peninsular India (Appendix 1). Additionally, molecular data for some closely related species including Platyceps gracilis, Platyceps bholanathi, Platyceps ventromaculatus and the distantly related Lytorhynchus paradoxus, Wallaceophis gujaratensis and Wallophis brachyura from India were generated in this study (Appendix 1).

Molecular analysis
Genomic DNA was extracted from liver or tail tissue samples stored in absolute ethanol at -20°C. DNeasy (Qiagen) blood and tissue kits were used to extract DNA. We amplified partial sequences of two nuclear (nu) and three mitochondrial (mt) genes. Cytochrome b (cytb) and NADH dehydrogenase subunit 4 (ND4), 16S rRNA (16S) and the nu markers are oocyte maturation factor (cmos) and recombination activating gene 1 (rag1). PCR conditions followed previously reported protocols for the respective primer pairs (cytb, primers GluDG L: Palumbi 1996 andH16064: Burbrink et al. 2000;ND4, primers ND4 and Leu: Arévalo et al. 1994; 16S, primers 16Sar-L and 16Sbr-H: Palumbi et al. 1991;cmos, primers S77 and S78: Lawson et al. 2005; rag1, primers R13 and R18: Groth and Barrowclough 1999). In total 37 individual se-quences were generated in this study (Appendix 1). PCR amplifications were carried out in a S1000TM Thermal Cycler (Bio-Rad, USA). Amplified PCR products were run on a 2% agarose gel and viewed under an Essential V4 (UVITEC Cambridge, UK) gel documentation system. Sanger sequencing was carried out in both directions using the same primers as used for amplification. Lab work was carried out at Centre for Ecological Sci- for Platyceps spp. mentioned in this study. Green (triangle) denotes the type locality and green (diamonds) for records of P. plinii comb. nov. Yellow (star) denotes the type locality, and yellow (circles) for the records of P. josephi sp. nov. Historically relevant names and important physiographic features are labelled. ences, Bangalore and sequencing was done at Medauxin sequencing services (Bangalore, India). Sequence chromatograms were edited and assembled using snapgene Viewer (http://www.snapgene.com/products/snapgene_ vie wer). Protein-coding genes (cytb,ND4,cmos,rag1) were checked for stop codons in unexpected regions by translating nucleotide alignments to amino acids in mega7 (Kumar et al. 2016).
Apart from the five genes for which we generated data, we also included an additional gene 12S because sequences of this marker were available for 55% of the samples. Taxon sampling was based on the recent publications on global snake and old world racer phylogenies (Figueroa et al. 2016;Zaher et al. 2019;Rajabizadeh et al. 2020). Newly generated data were combined with data from GenBank for other old world racers and the outgroups (Appendix 1), and aligned using the ClustalW algorithm (Thompson et al. 1994) with default settings as implemented in mega7 (Kumar et al. 2016).
All phylogenetic analyses were carried out using the CIPRES Science Gateway portal v3.3 (Miller et al. 2010). PartitionFinder v2 (Lanfear et al. 2016) was used (default MrBayes settings) to find the best-fit partition scheme for the concatenated dataset and model of sequence evolution for each partition. The best-fit scheme comprised six partitions (Appendix 2). We estimated phylogenetic relationships using maximum likelihood (ML ;Felsenstein 1981) and Bayesian inference (BI; Huelsenbeck et al. 2001) as implemented in RAxML 1.8 (Stamatakis 2014) and MrBayes 3.2 (Ronquist et al. 2012), respectively. ML analysis in RAxML 1.8 accepts only one model (GTR) of sequence evolution, and PartitionFinder suggested Gamma (G) correction for all the partitions (Stamatakis 2006), therefore, GTR + G model was used for this analysis. The ML + rapid bootstrap method was used to search for best trees, and branch support quantified via 1000 non-parametric bootstrap. The BI analysis used two runs with four Markov chains initiated from random trees and was run for 10,000,000 generations, sampling every 1000 generations. When the BI analysis was terminated, the standard deviation of split frequencies was less than 0.005, and the first 25% of trees were discarded as "burn-in". Convergence was confirmed for all parameters using Tracer v1.6 (Rambaut et al. 2014). Uncorrected pairwise genetic distances were calculated in MEGA 7 (Kumar et al. 2016) and resulting data were summarized for different genera in R 3.2.0 (R Core Development Team 2017) using package tidyverse (Wickham 2017).

Morphology
In addition to the type series and referred material (n=11) of the new species described below, comparative material examined directly includes the holotype of Argyrogena rostrata (type species of the genus Argyrogena Werner, 1924) and additional preserved museum specimens so far determined as A. fasciolata (n=50) and representative specimens from the genera Lycodon, Hemorrhois, Platyceps and Spalerosophis (n=33). Examined mate-rial is listed in Appendix 10. Voucher specimens were studied or mentioned from the following museums and institutional collections: BMNH (Natural History Museum, London Total length, snout-vent length and tail length were measured with thread and a ruler to the nearest 1 mm. Other dimensions were recorded with dial callipers, to the nearest 0.1 mm. Identification of sex was performed by dissection (presence or absence of hemipenes, inspected through a subcaudal incision at the tail base).
The hemipenis of freshly euthanized specimens were everted by pressing the organ until it was everted to the maximum extent, after which the retractor muscle was cut at the level of the 10 th subcaudal scale. From preserved specimens in which the hemipenis was not everted, it was removed through a longitudinal incision at 15 th subcaudal and detached from the body by cutting the retractor muscle. The organ was separated by making an incision around its circumference at the cloacal region and was immersed in warm water (60°C) for about 5 minutes to soften the tissue. It was then slowly everted with a bluntend forceps by gently pushing the organ from the distal to proximal end. Afterwards, the hemipenis was softened by again immersing it in warm water for intervals of 5-10 minutes. While everting, petroleum jelly was applied to the inner edge for smooth sliding of the inner ornamented layer. Care was also taken to avoid puncturing the hemipenial wall due to the rubbing of the spines. After eversion, the organ was inflated with petroleum jelly and tied at the base with a thread. Later it was immersed in 1% alizarin red solution in 70% ethanol and removed from the solution after one hour of staining. The calcified spicules were stained using this process and excess stain was removed by immersion of the hemipenis in 70% alcohol for approximately 30 minutes. Detailed photographs of the stained hemipenes were taken using Nikon D 5100 with a 100 mm macro lens (Nikon). Descriptions of hemipenial characters and terminology follow Dowling and Savage (1960), Savage (1997), Myers and Campbell (1981), Keogh (1999) and Zaher (1999).
Length and width of head scales were measured at the longest and widest points of the respective scale(s). In-terocular width was measured on top of the head at the shortest distance between the eyes. The number of ventral scales was counted according to Dowling (1951a). Dorsal scale row reduction formulae were based on Dowling (1951b). The terminal scale was not included in the subcaudal count. Values for symmetric characters are provided in left/right order.
To evaluate character states of midbody vertebrae we created µCT-scans of postcranial skeletons for six "Banded racers", four Platyceps rhodorachis and one each for P. florulentus, P. ventromaculatus and the new species, respectively and also included literature data for these taxa provided by Wilson (1967), Schätti (1987) and Schätti et al. (2014) in our analysis. Hemorrhois ravergieri and Spalerosophis diadema have previously been µCTscanned and vertebrae data of Hemorrhois nummifer were taken from Schätti (1987) for comparative analysis (see Appendix 7 and 8). Values for the ratio of the total length to the maximum width of the nasal bone are abbreviated as TLn/TWn. Values for the ratio of the total length of dorsal laminae of nasal to the length of the nasal shield is given as TLln/Lns.
To obtain counts of teeth by a non-invasive procedure, heads of Platyceps species and comparative material were subjected to micro-tomographic analyses with GE Inspections Technologies computer tomographs, e.g. a Phoenix nanotom X-ray|s, at the Museum für Naturkunde Berlin, Germany, and a Phoenix v|tome|x M at Chakan, Pune, India. For body scans we used a FF35 (YXLON International, Hamburg, Germany) at the Museum für Naturkunde Berlin, Germany. The scans were performed with 60-95 kV and 135-1000 μA, generating 1000 or 1440 projections with 750 or 1000 ms per scan. The different kV and µA-settings depended on the respective specimen size. Effective voxel size, i.e. resolution in three-dimensional space, ranged from 10-71.16 μm. The cone beam reconstruction was performed using the datos|x-reconstruction software (GE Sensing and Inspection Technologies GMBH phoenix|x-ray datos|x 2.1) and the data were visualized in VGStudio Max 3.0.
Abbreviations of morphological characters used in text and tables are as follows: Measures and ratios: SVL -snout-vent length; TaL -tail length; TL -total length (SVL+TaL); TaL/TL -ratio tail length/total length; HW -head width, measured at its widest lateral extension; HL -head length, measured from tip of snout to posterior edge of mandible; ED -eye diameter, measured horizontally; RW -rostral width; FL -frontal length; FW -frontal width; PrefL -prefrontal length; InL -internasal length; E-S -distance from anterior border of the eye to tip of snout; ParL -parietal length; ParW -parietal width; DSR -dorsal scale rows: numbers are given at one head length behind head, at midbody (taken at half of the total number of ventral scales), and at one head length before vent, respectively; PreVen -number of preventrals; Ven -number of ventrals; TS -total number of ventral and subcaudal scales including preventrals and terminal scale; PrO -number of preoculars; PoO -number of pos-toculars; T -number of temporal scales; SupL -number of supralabials; SubL -number of sublabials.
The opensource software QGIS (http://www.qgis.org) was used to prepare the map. Some of the metrics used in the IUCN Red-Listing criteria (IUCN, 2012) were used to calculate the area of occupancy (AOO) and the Extent of occurrence (EOO) of the new species using the GeoCat online tool (Bachman et al. 2011).
On the same basis and almost simultaneously as Shaw (1802), François Marie Daudin described Coluber hebe in 1803. Merrem (1820, p. 96, footnote a) was the first to recognise the synonymy of fasciolatus and hebe but he preferred to use Daudin's younger name as valid and combined it with the generic name Natrix Laurenti, 1768. This synonymy went unnoticed by subsequent authors and while fasciolatus remained in the genus Coluber, Daudin's hebe was relegated in the newly erected genus Lycodon by H. Boie (1826). A few years later Gray (1831), and subsequently Schlegel (1837) and Duméril, Bibron and Duméril (1854) recognized the conspecificity and continued to use Lycodon as the preferred genus. Like Merrem (l.c.), Schlegel (1837) also disregarded the priority of the name fasciolatus Shaw and considered specimens from Java to be identical to the taxon described from India by Shaw (l.c.) and Daudin (l. c.). It has to be noted that some subsequent authors erroneously credited the name hebe to Schlegel (e.g Boie, H. in Boie, F., 1827 by Duméril, Bibron and Duméril (l.c.), two morphologically almost indistinguishable species with the same lepidosis and variable colour and pattern (see O'Shea et al. 2018).
In his "Catalogue of Colubrine snakes […] of the British Museum" Günther (1858, p. 203) listed a "Var. F." under the synonymy of Lycodon aulicus without voucher specimens. This refers to Russell's (1769, p. 34-35, pl. 29) account on a "Coluber" ("Patza Tutta") from "Casemcottah" which is undoubtedly a specimen of what is commonly regarded as a "Banded racer" today, but not a Lycodon (wolfsnake). With his comment "belongs perhaps to Coryphodon fasciolatus" [= Coluber fasciolatus Shaw] he justified the confusion of two valid colubrid taxa, which continues today. In 1864 (p. 316) Günther noted: "The synonymy [of Lycodon aulicus] in all preceding herpetological works is confused -C. striatus, C. malignus, C. hebe, C. fasciolatus, and probably C. capucinus belonging to other species". He ignored the previous generic allocation to Lycodon, excluded the aforementioned taxa from the synonymy of L. aulicus, relegated fasciolatus Shaw, and hebe Daudin together with the simultaneously synonymized Coluber curvirostris Cantor, 1839 to the genus Zamenis Wagler, 1830, and created a conglomerate of unrelated taxa of racers and wolfsnakes. From that point on, authorities used the name fasciolatus in combination with different genera as the scientific name for the Banded racer whereas C. hebe together with C. curvirostris were treated as junior synonyms (e.g. Theobald 1876;Boulenger 1890Wall 1921;Smith 1943;Wallach et al. 2014;Uetz et al. 2019). At the same time, Russell's "Nooni Paragoodoo" and C. hebe were also recognized as synonyms of Lycodon aulicus or L. capucinus (e.g. Boulenger 1893;Wall 1921;Iskandar and Colijn 2002;Wallach et al. 2014;Uetz et al. 2019). Smith (1943, p. 531-532) studied the sketches in Russell's (1796) 'Indian Serpents' and identified the specimens depicted on plate 21 "Nooni Paragoodoo" together with that on plate 29 "Patza Tutta" as Coluber fasciolatus and continued to confuse the two different taxa. Further misidentifications are done in the list of determinations made from the Thomas Hardwicke collection of water-colour sketches deposited in the Zoological Library of the Natural History Museum, London, where he determined that four snakes depicted in sketch 55 and 61-63 were examples of Coluber fasciolatus (Smith 1943, p. 529). In contrast, our re-examination of the illustrations revealed that sketch no. 55 depicts a wolfsnake, Lycodon sp., and no. 61 a Common ratsnake, Ptyas mucosa and only sketch 62 and 63 illustrates a Banded racer (see Fig. 8).
It is also to be noted that from the initial confusion caused by Günther (1864), the original scalation data of Shaw's Coluber fasciolatus (192 ventrals and 62 subcaudals) were no longer considered representative of the scalation variation in the Banded racer. Not a single subsequent author, even when citing the original source, has included the values of the "Nooni Paragoodoo" in their variation of fasciolatus; not even Wilson (1967) in his revision discusses the very low number of 62 subcaudals recorded for the type specimen. Bauer et al. (2015) investigated two collections of dried snake skins housed in the Natural History Museum, London. The first collection arrived in 1837, the second in 1904 and are preliminarily attributed to Patrick Russell (see also Campbell 2015). The skins are attached to paper and in addition to their inventory numbers, for some specimens, handwritten notes are added which refer to Russell's books of 'Indian Serpents' (see Fig. 4). It is remarkable that all specimens from the 1904 collection labelled with the vernacular name "Nooni Paragoodoo" have been identified as Lycodon aulicus by Bauer et al. (2015). Our reexamination of the Lycodon specimens amongst Russell's dried skin collections (including BMNH 1837.9.26.50-52;BMNH 1904.7.27.34, 43, 48, 52, 55, 72, 89 and 93) revealed that none of these specimens qualifies as the holotype of Coluber fasciolatus, as the measured lengths of these skins (ranging from 37 to 55 cm) are well below the 2 feet [60.96 cm] length reported by Russell for his "Nooni Paragoodoo".
In this work, we clarify the identity of C. fasciolatus, by analysing  original data and comparing it with specimens from various museums, data and photographs from the authors and literature concerning Indian snakes (e.g. Günther 1864;Boulenger 1890;Smith 1943;Whitaker and Captain 2004). We conclude that, in contrast to earlier determinations (Günther 1864;Boulenger 1890Smith 1943, to name a few authorities only), Russell's "Nooni Paragoodoo" and therefore Coluber fasciolatus Shaw (incl. C. hebe Daudin) is not a racer but a wolfsnake of the genus Lycodon. Wall (1914, p. 34) explained the meaning of the vernacular name "Nooni Paragoodoo" as: 'Nooni' = oil in . Lycodon fasciolatus comb. nov. Previously unpublished extract from a paper sheet with glued specimens from the 1904 collection of dried snake skins preliminarily attributed to Patrick Russell showing a specimen of the "Nooni Paragoodoo" kept in the herpetological collection of the Natural History Museum, London, U.K. (BMNH 1904.7.27.93).
Telugu language and 'Paragoodoo' stands for a runner or glistening, concerning the smooth and polished character of the scales in the snake. The extraordinary glossy character of dorsal scalation is well known for Lycodon species.
To determine the species affiliation we carefully compared Russell's description and the sketch on plate 21 with published data of Lycodon species from southern and eastern India (Table 1). In terms of the combination of pholidosis, colouration and pattern, Russell's specimens are generally comparable to a short banded morphotype of the Common Indian wolfsnake, Lycodon aulicus, a widespread species in the Indian subcontinent.
In a recent study on population systematics of the L. aulicus complex Ganesh and Vogel (2018) identified two different species-groups. Based on head dimensions, collar band pattern, hemipenal morphology, preocular-frontal and supraocular-prefrontal scale contact conditions, they reassessed the systematics of Lycodon aulicus, provided a redefined diagnosis for aulicus (sensu stricto) and revalidated Lycodon anamallensis Günther, 1864, a taxon so far regarded as conspecific with L. aulicus. Another result of Ganesh and Vogel's (l.c.) analysis is the synonymization of Lycodon osmanhilli Taylor, 1950 with L. anamallensis.
Our comparison of the original information of fasciolatus Shaw, with data from the two morphotypes identified by Ganesh and Vogel (l.c.), showed the closest match with L. anamallensis (see Table 1). Although ventral and subcaudal counts of female fasciolatus lie within the ranges of females from both anamallensis and aulicus, it is clearly distinguished from the latter by the combination of following diagnostic characters: (1) preoculars not in contact with frontal vs. in contact, (2) supraoculars in contact with prefrontals vs. not in contact, (3) collar absent vs. collar present and in contact with parietals, (4) upper labials yellow with dark brown center vs. upper labials whitish, (5) short whitish bands stippled with black, not extending into lateral edges of ventrals vs. elongated cream-white bands, extending onto lateral edges of ventrals.
Based on these consistencies of its morphological characters, colour and pattern, we consider Coluber fasciolatus Shaw, 1802 and Coluber hebe Daudin, 1803 as conspecific with Lycodon anamallensis Günther, 1864, including its synonym Lycodon osmanhilli Taylor, 1950. With the reallocation of the taxon fasciolatus Shaw to the genus Lycodon we follow the lead of H. Boie (1826) and regard Lycodon fasciolatus (Shaw, 1802) as a valid combination with priority over its younger synonyms L. hebe, L. anamallensis and L. osmanhilli.
Distribution.  does not provide specific information for the origin of his "Nooni Paragoodoo", Shaw (1802) only states "India" as the type locality for C. fasciolatus and Daudin (1803), referring to the title of Russell's publication, i.e. "Coromandel" as the origin of his C. hebe. As already noted by Bauer (2015, p. 35) the description of C. fasciolatus is based on a single specimen and the snake depicted on plate 21 in  is the holotype. This makes Wallach et al's (2014, p. 54) lectotype designation unnecessary and their type locality restriction "Vizagapatam, coast of Coromandel, India" [= Vishakhapatnam, Andhra Pradesh state, SE India, 17°41'N, 83°13'E, elevation 25 m] via lectotype selection" is unjustified because no such information was provided by Russell for the holotype. Even considering that Russell spent most of the time of his service for the East India Company in the Visakhapatnam district (Andhra Pradesh state of India), it is not clear whether he captured the individual specimen himself.  mentioned 11 persons by name who sent him material but does not name anybody in relation to the "Nooni Paragoodoo". Other specimens described by him without a named donor and therefore possibly captured by himself personally came from Lake Ankapilly, Bimblipatam, Boni, Masulupatam and Nerva in Andhra Pradesh, from Tanjore in Tamil Nadu, from Ganjam in Odisha and Hyderabad in Telangana. Strictly speaking, besides Vizagapatnam any of these places could be considered as the possible type locality of L. fasciolatus.
Lycodon fasciolatus seems to be widespread in southern peninsular India and recent records from the Coromandel region support the possible demarcation of the type locality in this area. According to Das and de Sil-  Whitaker and Captain (2004), Mukherjee and Bhupathy (2007), Purkayastha (2013), Ganesh and Vogel (2018), Vogel and David (2019), Ganesh et al. (2020, in part), andKalki et al. (2020). We did not include the character "Supraocular-prefrontal-contact" as used by Ganesh and Vogel (2018) for delimiting Lycodon anamallensis from L. aulicus because it is a consequence of the condition "Preocular-frontal-contact". Remarks. Blyth (1855b, p. 740, smallprint) described Coluber vittacaudatus (collector unknown, holotype not located) from the vicinity of Darjeeling (West Bengal state, India) and considered it as allied to Coluber fasciolatus Shaw. The taxon fell into oblivion until Das et al. (1998, p. 157) listed it again as a valid species based on a personal communication by Van Wallach. Whitaker and Captain (2004, footnote 6) were unsure where C. vittacaudatus should be placed and Schätti et al. (2014, p. 384, smallprint) regarded it as "possibly a senior synonym of the Oriental ratsnake Orthriophis taeniurus yunnanensis (Anderson, 1879), and of O. taeniurus (Cope, 1861) as well." The alleged relationship to C. fasciolata Shaw, as originally suggested by Blyth (1855b), prompted Wallach et al. (2014) and later Das and Das (2017) to place vittacaudatus into the previously monotypic genus Argyrogena Werner, unaware that fasciolatus Shaw is not a racer.

A.[rgyrogena] rostrata
Coluber curvirostris was based on a single specimen from "Bengal" described by Theodore Cantor (1839) in his "Specilegium Serpentium Indicorum". A coloured sketch of the holotype is deposited in the Bodleian Li-brary (Oxford, U.K.) and reproduced here for the first time ( Fig. 7).
Franz Werner (1924) described Argyrogena rostrata (genus et species nova) based on a specimen stored in the Natural History Museum, Vienna, which was originally listed as having come from "Argentinien" and for a few years it was considered a valid neotropical taxon (Dunn 1928;Amaral 1929). Malcolm A. Smith, who visited the herpetological collection of the Natural History Museum, Vienna in 1927, examined several of Werner's type specimens, including the holotype A. rostrata, recognized its true identity, synonymized it with fasciolatus and relegated Argyrogena to the synonymy of Coluber. This classification was largely maintained until Wilson (1967) dedicated a study to the Banded racer and in this context revalidated Werner's name Argyrogena. Subsequently, the combination Argyrogena fasciolata (Shaw, 1802) was established and has been used for the Banded racer until today (see list of synonyms and chresonyms and e.g. Williams and Wallach 1989;Smith and David 1999;Whitaker and Captain 2004;Khan 2006;Somaweera 2006;Wallach et al. 2014;Koushik, 2015;Das and Das 2017;Uetz et al. 2019, to name a few general publications). In a short overview of the taxonomy of Coluber, Amr and Disi (2011, p. 191) erroneously stated that Wilson (1967) reallocated the taxon fasciolatus to the genus Platyceps.
We also tracked down two older and previously overlooked or wrongly assigned names that clearly refer to what is currently understood to be the Banded racer, viz. Coluber pictus Daudin, 1803 and Natrix plinii Merrem, 1820. Again both names are entirely based on the same description in Russell (1796, p. 34-35, pl. 29: "Coluber" named "Patza Tutta" by natives).
From the mid of the 19 th century onwards both names fell into oblivion as potential synonyms of the Banded racer. This was probably also due to Schlegel (1837, pt. Descriptive p. 172) who regarded both pictus Daudin and plinii Merrem as possibly identical with Coluber triscalis Linnaeus, 1758 and Coluber corallinus Linnaeus, 1758. The latter two are currently regarded as synonyms of the Caribbean colubrid snake Liophis (now Erythrolamprus) triscalis (see Dundee 1994).
Coluber pictus Daudin, 1803 however, is threatened for reasons of homonymy by an older name, Coluber pictus Gmelin, 1789, the accepted original combination for the Painted bronzeback (now Dendrelaphis pictus, see ICZN 1958). Therefore, Coluber pictus Daudin, 1803 represents a younger primary homonym of Coluber pictus Gmelin, 1789 and is unavailable (ICZN 1999, Art. 57.2; see also above the list of synonyms and chresonyms for further homonymous names).
Merrem (1820) used Russell's (1796) account of the "Patza Tutta" and formally named it Natrix Plinii. His diagnosis was very short and besides some general features concerning habitus and characteristics of dorsal scales, he mentioned only the number of ventral (202) and subcaudal (91) scales. Additional information can be found in Russell's (1796, p. 34-35, pl. 29) original text and plate. He described the "Patza Tutta" with the following characters: 1) head small, very little broader than neck, ovate, depressed, 2) two internasals, 3) two prefrontals, larger than internasals but similar in form, 4) two supraoculars, purse-shaped, 5) frontal bell-shaped, 6) parietals semi heart-shaped, truncate, 7) mouth wide jaws unequal, lower considerably shorter, 8) teeth small sharp, reflex, two palatal rows, and one marginal [maxillar], in the upper jaw, 9) eyes distant, lateral, large, orbicular, 10) body round, covered with very small smooth, ovate scales, outer row orbicular, 11) length one foot eleven inches [58.4 cm], tail five inches [12.7 cm], very tapered and sharply pointed, 12) dorsal colour brown, forebody with narrow crossbands, composed of short, black and white lines, behind these, a few obscure reddish-brown bands interspaced by white colour of interstitial skin; bands fading at posterior part of body, 13) outermost dorsal scale row yellow, 14) ventral with pale-yellowish cast. The description resembles the specimen depicted on plate 29 and in comparison with our own and published data (see below) it corresponds to male specimens of the Banded racer.
To our best knowledge, Natrix plinii Merrem, 1820 is the oldest available name for the snake species which is commonly now regarded as Banded racer. Merrem (l.c.) simplified the type locality to "Bengal", but according to Russell (1796, p. 35) two specimens of the "Patza Tutta" were sent to him from "Casemcottah" by Captain Gowdie in 1788 and by Lieutenant Whyte without year specification (see Appendix Note 1).
The same applies to the type specimen of Coluber curvirostris Cantor, 1839 from "Bengal" which is not yet located. Although some of Cantor's original specimens are known to be stored in the Natural History Museum, London (Adler 2007(Adler , 2016, we were unable to identify the holotype of C. curvirostris among the preserved specimens in the collection. A possible candidate donated by the Zoological Society of London to the British Museum (Nat. Hist.) (BMNH 1851.9.13.260 from "India"), can be excluded as the holotype because it does not correspond to Cantor's original data and plate, in respect of the number of ventrals (226 vs. 220) and subcaudals (88 vs. 85). A transcript of Cantor's unpublished original manuscript for C. curvirostris, with an expanded description, is presented here in the Appendix (Note 2).
Description. Based on our own data and summarized information from literature Smith 1943;Wilson 1967;Whitaker and Captain 2004), Platyceps plinii is characterized as follows: (1) dorsal scales smooth in 23-25 (exceptionally 21 or 22) rows at forebody, 23 (exceptionally 21 or 25) at midbody and 17 (exceptionally 16 or 15) before the vent, (2) one to three preventrals followed by 190-234 ventrals, (3) 73-101 paired subcaudals, (4) a divided cloacal plate, (5) nasals divided, in con-tact with first and second supralabial, naris medium sized (6) internasals smaller than prefrontals, (7) frontal shorter than its distance to tip of snout, as long as or little shorter than parietals, (8) loreal usually slightly longer than high, rarely as long as high or higher than long, exceptionally fused with prefrontal, (9) usually two (80% of examined specimens) exceptionally three preoculars, one larger upper preocular usually in contact with frontal and one (exceptionally two) very small lower presuboculars wedged between third and fourth supralabial, (10) two (exceptionally three) postoculars, (11) supraocular not in contact with prefrontal, (12) temporals variable, in anterior two (rarely three or one) and secondary three (rarely two) rows, (13) eight (exceptionally seven, nine or ten) supralabials, fourth and fifth (rarely fifth and sixth or sixth and seventh) in contact with the eye, (14) usually ten, sometimes nine or eleven sublabials, first four to five in contact with anterior submaxillars, (15) head wide and moderately pointed, barely distinct from neck, body round, olive brown to reddish brown coloured, patternless in adults or with light vermicular pattern in juveniles, (16) iris dark grey or blackish, pupil black with light brown corona, (17) dorsal colour greyish or reddish brown, (18) collar absent, (19) whole body patternless or with indistinct small light bands on forebody in adults; in juveniles with distinct black and white strippled bands reaching lateral edges of ventrals, extending to near vent, (20) supralabials creme brown, patternless, or with small light spots in juveniles (21) throat, ventral body and tail yellowish cream without pattern, (22) hemipenis simple; proximal fourth smooth; distal three fourths covered with small spines gradually decreasing in size distally; distal area with irregular shaped calyces, denser towards the apex (23) total length up to 148 cm, (24) relative tail length 0.19-0.31.
Dorsal scale reduction formula summarized from 31 examined specimens (see Appendix 10; divergent data from literature [Wilson 1967] in square brackets). Only main reductions are given.   scale rows involving rows one to three between 89 and 99% of ventrals.
Hemipenis (based on ZSI-CZRC-V-6416, Pune, left organ, Fig. 11). The hemipenis reaches up to 13 th subcaudals and is 36.9 mm in length and 8.2 mm maximum width at midbody and apex. It is divided into three distinct areas; the proximal area (nearly 30% of the total hemipenial length) is smooth without any ornamentation, middle zone (about 25% of the total hemipenial length) is with evenly scattered spinules and the apical calyculated portion (45% of the total hemipenial length).
No enlarged spines are present, but the size of spinules at the proximal end are slightly larger than the ones at the distal end. The calyculated area can be further divided almost equally into proximal half with large calyces and distal half with smaller and denser calyces. The sulcus spermaticus is single, bounded with thick walls, runs straight across its length and opens into a delta (3 mm wide at apex, 1.9 mm in length), which is nude. The delta is subapical in position and the apex is calyculated, edged with fewer papillae. At the sulcate side, the calyces are larger proximally and gradually smaller and denser towards the distal end. The distal calyces are scalloped and edged with papillae. Along the asulcate side the spinous area starts more distally than the sulcate and lateral side. After the spinous zone, there are 2-3 rows of large calyces followed by densely packed calyces towards the apex. The proximal large calyces (9-10 in number) are of uneven size (mostly rounded or pentag-onal) and among them the distal ones are scalloped. The large scalloped calyces are of 2.2-2.7 mm in length and 1.9-2.5 mm in width. Wilson (1967) mentioned that the hemipenis is clavate, but in a fully inflated organ it starts widening from the middle and from the asulcate side there is a slight constriction between the midbody and apex. The ornamentation in ZSI-CZRC-6416 from Pune and ZSI-CZRC-6284 matches with that of LSUMZ 9425 as provided by Wilson (1967). However, the only available drawing of the organ as provided by the author is of the sulcate side, but apparently, differentiations in calyces are seen at the asulcate side.

Dentition (based on ZMB 8053, male, Bengalen).
15/15 maxillary teeth, the anterior 13/13 precranterian teeth are increasing in size posteriorly and are followed by a small diastema and 2/2 enlarged roundish cranterian teeth without groove. All maxillary teeth are slightly curved posteriorly, without significant interspaces. Medial to the precranterian teeth are single replacement tooth at different growth stages. One to two replacement teeth per cranterian tooth are found posteromedially to each tooth, showing different growth stages. 11/11 palatine teeth, decreasing in size posteriorly. All are curved posteriorly shortly above the base. Lateral to the palatine teeth there is a single replacement tooth at different growth stages. No significant interspace exists between the different palatine teeth. Posteromedial process of palatine large, expanding two pterygoid teeth. 13/12 ptery- goid teeth, decreasing in size posteriorly. All are curved posteriorly shortly above the base. Lateral to the pterygoid teeth there is a single replacement tooth at different growth stages. No significant interspace exists between the different palatine teeth. The posterior 57% of the pterygoid are without teeth. 17/17 mandibular teeth, increas- ing in size up to tooth 5/5 and from there decreasing in size posteriorly. All are slightly curved posteriorly. Medial to the mandibular teeth there are 1-2 replacement teeth at different growth stages.
Variation in dentition.  based on one skull from his collection described 13 subequal maxillary teeth which are followed by a diastema which is as large as one tooth and followed by two little larger posterior teeth. He described 11 palatine, 15 to 16 pterygoid teeth all decreasing in length posteriorly and 18 mandibular teeth which increasing in length up to the fifth tooth and from there decrease in length posteriorly. Smith (1943, p. 159, fig. 48 D) depicted an upper jaw with 13 precranterian teeth which are increasing very slightly in size posteriorly and are followed by a distinct diastema and a pair of enlarged cranterian teeth. Wilson (1967, p. 271) mentioned like Smith (1943, p. 170), 12 -14 maxillary teeth followed by a distinct diastema but in contrast to the latter, Wilson (l. c.) identified the following two cranterian teeth as unenlarged and counted 9-11 palatine teeth, 14-17 pterygoid teeth, and 14-18 mandibular teeth. The dentition of four Indian specimens (BMNH 37a, see Fig.  12; MCZ R-28645; NMW 25465:3 and ZMB 4786a) is characterized by 12-14 maxillary teeth followed by a diastema (except MCZ R-28645, without diastema) and two enlarged cranterian teeth, 9-11 palatine teeth with a posteriomedial process of palatine expanding two or three pterygoid teeth, 13-17 pterygoid teeth with 48-52% of the posterior pterygoid without teeth, and 16-19 mandibular teeth, increasing in size up to tooth 4-6 and from there decreasing in size posteriorly.

Dimensions and proportions.
The body is robust, moderately stout, roundish in cross-section at midbody. Head moderately pointed, barely distinct from neck. Eye large, 16-24% of head length. Male specimens grow slightly larger than females; the longest examined specimens are from Sind, Pakistan including a male (SMF 50446) with 1482 mm and a female (SMF 50409) with 1361 mm total length. The longest Indian specimen was reported by Gharpurey (1931) from Ahmednagar, Maharashtra with 1422 mm total length. The smallest male (NMW 25465: 3) with 336 mm and the smallest female (SMF 57311) with 553 mm total length came from "Calcutta" and Sind respectively. The tail versus body length ratio ranges from 0.21 to 0.31 (males 0.21-0.31, females 0.23-0.28). For ten specimens from Telangana state, India measured by Seetharamaraju (2014, p. 95 ff.) the tail versus body length ratio ranges from 0.19 to 0.23 (males 0.19, females 0.19-0.23). (Fig. 9 A-K and Fig. 10 A-J). Dorsal colour and pattern show a significant ontogenetic change. In juvenile and subadult specimens the dorsal ground colour is olive-brown to light brown, the head dorsally slightly darker with narrow elongated irregular shaped whitish markings. Along the body up to 70 narrow whitish/black or dark brown stippled crossbars, usually two scales wide and in contact with lateral edges of ventrals. The first band is separated from the posterior border of parietals, the interspace between bands is three to four dorsal scales wide. The intensity of bands decreases posteriorly and the contours become blurred near to the vent. Dorsocaudally the bands are dissolved in an irregular fine pattern of light and dark points. Temporals, preoculars and supralabials can show small whitish spots. From about 600 mm total length the colour and pattern of young specimens begin to fade and disappear gradually. Adult specimens are dorsally dark brown to dark reddish-brown without any markings on head, body or tail. The ventral side of head, body and tail in all age classes is patternless yellowish cream.

Colour and pattern
Distribution. Platyceps plinii is commonly believed to occur from Pakistan in the West to Bangladesh in the East with a North to South distribution from Nepal to Sri Lanka (see e.g. Smith 1943;Mahendra 1984;Welch 1988;Daniel 2002;Gruber 2002;Whitaker and Captain 2004;Wallach et al. 2014;Uetz et al. 2019). We analyzed pertinent literature and locality information from museum specimens to present a list of localities corrected and updated for each country (see below and Appendix 9 Gazetteer).
According to our current knowledge, Platyceps plinii is only known with certainty from India and Pakistan. In India it is recorded from the states of Andhra Pradesh (Kasimkota [Casemcottah, type locality], Eluru, Nallamala Hills, Pulicat Lake, Rajahmundry, Rollapenta, Sundipen- In Pakistan confirmed records are known from the Indus plain in Sindh province (nr. Badin, Hala, Jati, Lakarna, Makli Hills, Mohenjo-daro, Pir Patho, Raj Malk, Sonda, and Tatta) and some authors mentioned southern and central Punjab without specific locality (e.g. Khan 1980Khan , 1982Khan , 2002Khan , 2006. We found only one record with more precise information for Punjab from the Lal Suhanra National Park in Bahawalpur published by Khan et al. (2018Khan et al. ( , p. 1731. Unconfirmed or erroneous distribution records. Sporadically the Banded racer is mentioned for the Himalayan region in northwestern and northeastern India. Mahajan and Agrawal (1976) and Agrawal (1979) published records (as Coluber fasciolatus) from Saproon village and the Gambhar bridge from an elevation of 1513 m and 1485 m a.s.l. respectively. Both localities are situated in the Solan district of Himachal Pradesh state of India. From Uttarakhand state the Banded racer was listed by Sharma (2004) from Sokharak, Mandal Rest House and Mandal-Chopta road in Kedarnath Wildlife Sanctuary located between 1500 and 3300 m a.s.l. in the Chamoli and Rudraprayag districts, and by Bahuguna (2010) for the Corbett Tiger Reserve (now Jim Corbett National Park). Since these reports, no further specimens from the Himalayan range of Himachal Pradesh and Uttarakhand have become known and because of the fact that these localities are considerably above the average altitudinal distribution of this species (see below), we suspect that these records are based on incorrect determinations (Abhijit Das, pers. comm. 2020) and consider the records to be very doubtful. The Banded racer is also listed from the Eastern Himalayas by e.g. Jha and Thapa (2002), Chettri and Bhupathy (2007) and Chettri et al. (2011). Whereas the first authors consider it "rare", Chettri and Bhupathy (l.c.) regarded it surprisingly as "common" in Sikkim. Since none of the above-mentioned authors provided a single voucher specimen or a specified locality in that East Himalayan state, and no preserved material is known from any collection, so we consider its distribution in Sikkim to be unproven.
Based on water-colour sketch no. 61 from the Hardwicke collection, Smith (1943, p. 529) mentioned "Cawnpore" [Kanpur, Uttar Pradesh state of India] for Coluber fasciolatus. Although the Banded Racer is known to occur in this state (see above), the snake depicted on this drawing is in fact, a Common ratsnake, Ptyas mucosa. Wallach et al. (2014) added the Pakistani province of Balochistan to the distribution range, but the literature cited provided no such indications and we could not identify a single credible record for this state.
When it comes to the distribution of the Banded racer in Sri Lanka, there is a consistent reference to Haly (1886Haly ( , 1891, who allegedly provided the first evidence. But, as already noted by Somaweera (2006), Ferguson (1877) was the first who listed this species for "Ceylon". Ferguson (l.c.) (1886) the Banded racer is consistently listed as part of the Sri Lankan snake fauna (e.g. Wall 1921; Deranyiagala 1955; P. de Silva 1980;de Silva 1990de Silva , 2001Das and de Silva 2005;Somaweera 2006;de Silva and Jinasena 2009;Wallach et al. 2014;de Silva and Ukuwela 2017) although no additional specimen has been collected, no preserved voucher specimens are known from any collection, and recent fieldwork on Jaffna Peninsula revealed no further material (Abyerami and Sivashanthini 2008;Madawala et al. 2019). We follow Taylor (1950) and regard this species as doubtful for Sri Lanka. In several publications, especially in overview works and faunal lists, the Banded racer is mentioned for Bangladesh. This is probably because in older publications (e.g. Cantor 1839; Günther 1864; Theobald 1876; Boulenger 1890) "Bengal" is listed as a known distribution, a region currently divided between Bangladesh and the West Bengal state of India. After the separation of the central and eastern regions of Bengal from India in 1947 and the foundation of the state of Bangladesh in 1971, the majority of (Indian) authors changed the original information "Bengal" to "West Bengal"(India), but others still maintained the distribution for Bangladesh (e.g. Welch 1988;Daniel 1989Daniel , 2002Das 2002;Whitaker and Captain 2004;Wallach et al. 2014;Uetz et al. 2019), although  already clarified that the species is not found further east than Calcutta in West Bengal state of India. Authors who refer exclusively to the herpetofauna of Bangaladesh (e.g. Islam et al. 2000;Khan 1988Khan , 2004Kabir et al. 2009) do not present more detailed localities and while Hasan et al. (2014) listed the Banded racer as expected to occur in the Northeast and Southeast of the country, he did not provide any supporting evidence. We could not find any historical or current evidence that this snake species occurs in Bangladesh.
From Nepal, the Banded racer was first mentioned by Günther (1861) in a list of "[…] Cold-Blooded Verte-brata […]" collected in that country. Günther's source was a water-colour painting of a specimen from the "Central hilly region" (reproduced here in Fig. 6) probably made by Raj Man Sinh Chitrakar, a Nepalese draftsman employed by Brian H. Hodgson (McClelland and Griffith 1844;Cocker and Inskipp 1988). The locality most probably refers to the complex of hills and valleys surrounding the Kathmandu valley where the majority of specimens in the Hodgson Nepalese collection originated. Our examination of Hodgson's original drawing, in the Zoological Library of Natural History Museum, London, revealed that the specimen depicted is not a Banded racer but a natricid snake, Rhabdophis himalayanus (Günther, 1864), a species reported from that region by e.g. Gruber in Schleich et al. (2002). Shrestha (2001) stated that the Banded racer has been reported from the Royal Chitwan National Park, and "[…] lives in dense Sal forest as well as grassland of Chitwan valley in Nepal […]". We regard this statement as based on a misidentification because no voucher is known. Similar to the aforementioned record for Sri Lanka, the occurrence of the Banded racer in Nepal relies solely on an old and unreviewed original source, so the presence of this species in Nepal could not be confirmed. Although Kramer (1977) already considered the distribution in Nepal as questionable, and no further material has been collected (Karan B. Shah in litt. 2019), most subsequent authors continue to list the Banded racer in the Nepalese herpetofauna (e.g. Swan and Leviton 1962;Welch 1988;Schleich 1993;Shrestha 1998Shrestha , 2001Shrestha , 2003Gruber 2002;Shah and Tiwari 2004;Wallach et al. 2014;Uetz et al. 2019).
Based on records of the Banded racer in Bengal, Schleich and Kästle (2000) and Lenz (2012) suspected its possible occurrence in the Himalayan Kingdom of Bhutan which was adopted as questionable by Uetz et al. (2019). As explained above for the Himalayan regions of India and Nepal such a presumption lacks any basis and we refer to the list of verified records above (see also Appendix 9 Gazetteer).
Cantor (1847a, b) mentioned a specimen from the Province Wellesley (Seberang Perai, Penang state, West Malaysia) which corresponds with Russell's "Nooni Paragoodoo", and determined it as Coluber fasciolatus. Based on dentitional characters, which he regarded as similar to Russell's (1796) description, he concluded that C. fasciolatus "[…] cannot be placed in the genus Lycodon, to which it has been referred by M.M. H. Boie, -Wagler and Schlegel". Wall (1908, p. 21 footnote, 1913, p. 20 footnote, 1921, p. 195 footnote, 1924. 619) and Wilson (1967, p. 266) already doubted the authenticity of Cantor's record. Our re-examination of this juvenile specimen (BMNH 60.3.19.1121) confirmed its identity as P. plinii, thus the locality given by Cantor must be considered as an error. Mason (1861) listed "Coluber fasciolatus" for "Burmah" and Theobald (1868b) and Nicholson (1874) mentioned, "Zamenis fasciolatus Shaw" from "Mergui" [Mergui or Myeik Archipelago, Myanmar]. The latter quoted "Blyth" but gives no specific reference, that makes a reassessment impossible. Theobald (1880) added "Tenasserim, &c." to the distribution of the Banded racer, without reference to specific specimens or exact localities. As indicated above, the easternmost records are from West Bengal and we, therefore, consider the records for Myanmar to be based on misidentification.
Habitat and natural history. Platyceps plinii is widespread over the Indian Subcontinent (see above, Map 1 and Appendix 9 Gazetteer). It is reported as regionally common in e.g. Maharashtra and Karnataka (Murthy and Acharjyo 1987;Daniel 1989Daniel , 2002Deshmukh et al. 2015;Ingle et al. 2014) and abundant in parts of Goa (Murthy and Acharjyo 1987). It inhabits plains, plateaus and mid hills covered with mixed and dry deciduous forest, scrub forest as well as semi-arid and rocky areas (for habitat examples see Fig. 13 A and B).
The vertical distribution ranges from sea level to over 1000 m altitudes, with about 85% of all finds below 600 m a.s.l. A few unusually high sites are known from the states of Madhya Pradesh (Pachmarhi Biosphere Reserve, 1053 m a.s.l.) and Maharashtra (Kaas Plateau, 1191 m a.s.l., and Panchgani, 1258 m a.s.l.) (see Appendix 9 Gazetteer).
It is a mainly terrestrial species but also observed to be a good climber (Ingle 2008) and found in open places, tall grass, dense bushes, under roots of old trees, in spiles of stones and bricks, and rodent holes. The Banded racer is also reported from urban and cultivated areas where it was found in houses, village environs, agricultural open fields, gardens, parks and farms (Stoliczka 1871a ;Khaire 1996Khaire , 2014Ingle 2002;Whitaker and Captain 2004;Rao et al. 2005;Srinivasulu et al. 2006;Chandra et al. 2008;Chikane and Bhosale 2012;Charjan and Joshi 2015;Sayeswara et al. 2015;Dileepkumar et al. 2016;Ingle et al. 2019). In some places, e.g. Bhadrak, Odisha, 30-40 individuals were found hibernating together in the cob walls in human habitations (Mohalik pers. comm.).
It is a diurnal species that is only exceptionally observed active during the night and known for its defensive and pugnacious behaviour. If threatened or captured it erects itself, flattens the forebody and furiously attacks when further provoked which one can confuse with a cobra (Wall 1907(Wall , 1921Whitaker and Captain 2004). Its antipredator mechanism was summarized by Green (1988) as follows: (1) dorsoventral body compression, (2) dorsoventral neck compression, (3) lifting of head and neck, (4) S-coil striking posture, (5) lunging and striking, and (6) bite. One female caught while crossing a road immediately defecated and started attempting to bite. In other instances, individuals were observed climbing electric cables and visiting switch boxes, possibly looking for hiding geckos (Hemidactylus spp.).
The Banded racer is an active hunter that usually kills its prey by constriction or body pressure. Its prey consists of small mammals like rats, mice, shrews and bats, but also frogs, lizards, birds, worms, and for juveniles insects are part of the food spectrum Daniel 1989Daniel , 2002Sharma 1999a;Das and de Silva 2005;Satish 2008;Whitaker and Captain 2004).
Mating of the Banded racer was observed between late winter and mid-monsoon (Dileepkumar et al. 2016). The reproduction is oviparous with clutch sizes between 2 and 23 eggs and egg sizes between 26.8-34.7 mm length and 16.5-22.4 mm width, laid between January and September Khaire 1996Khaire , 2008Khaire , 2010Khaire , 2014   and July Daniel 1989Daniel , 2002 For Pakistani specimens,  and Wilson (1967) observed semifossorial behaviour and Khan (2002Khan ( , 2006 reported that reproduction in Banded racer occurs twice a year, during the months April to May and August to September, with clutch sizes from 23 to 30 eggs.
Platyceps josephi sp. nov. shows most similarities with its sister taxon P. plinii in regard to pholidosis and colour pattern but can be distinguished from the latter by its lower number of midbody dorsal scale rows (21 vs. 23), its lower mean value of ventral scales (202 vs. 214), its lower mean value of subcaudal scales (82 vs. 87) and its lower mean of the sum of ventral and subcaudal scales (285 vs. 304), the presence of the clearly demarcated two slanting roughly "∏" shaped white markings on the back of head vs. absence of such markings and distinct white bands on the dorsum in both juveniles and adults vs. ontogenetic change, i.e. bands present in juveniles only, but usually absent or faded in adults. Additionally, Platyceps josephi sp. nov. differs from P. plinii in variation of mitochondrial DNA sequences. With pairwise uncorrected p-distances varying 4-5% in cytb & ND4 and 3% in 16S. It is also clear from our thorough verification of distribution that these two species only have a minor range overlap in northeastern Tamil Nadu (Fig. 1 Map). Etymology. The specific epithet is a patronym of late Mr Naveen Joseph. Naveen was a naturalist from Tuticorin, well known for his research on reptiles, particularly snakes in that region. He was a friend of VD, SN, and GM and helped them collect specimens of the new species. Suggested English name: Joseph's racer. Vernacular name: In various parts of Tamil Nadu state this snake is called by the name "Odugali Pambu" " ஓடு காலிப் பாம் பு ". It is a portmanteau word in Tamil language "Odugali" is often used to address someone "who doesn't stay at home and elopes" and "Pambu" is the word for "snake".
Description of the holotype. Morphometric and merestic data are provided in Table 2. Female. Specimen in good condition with three incisions into coelom at 80 th , 106 th and 123 rd ventral respectively.
15/15 mandibular teeth, increasing in size up to tooth 6/6 and from there decreasing in size posteriorly. All are slightly curved posteriorly. Medial to the mandibular teeth there are 1-2 replacement teeth at different growth stages.
Hemipenis (based on ZSI-CZRC-6639, paratype, Karur, right organ, Fig. 19). The hemipenis reaches up to 12 th subcaudals and is 24.3 mm in length with a maximum width of 5.8 mm at 1/3 rd of the proximal end. It is divided into three distinct areas; the proximal area (nearly 1/4 th of the total hemipenial length) is smooth without any ornamentation, middle zone (about 28% of the total hemipenial length) is with spinulae and the apical calyculated portion (48% of the total hemipenial length). The spicules at the midbody of the hemipenis gradually reduce in size from proximal to the distal end. The calyculated area can be further divided almost equally into proximal half with large calyces and distal half with smaller and denser calyces. The sulcus spermaticus is single and runs straight across its length and opens into a wide and elongate naked area (3.3 mm) extending laterally from the proximal to distal end extending to the apex. The opening of the sulcus forms a triangular area and ends apically. The sulcus is bounded with thick walls. The calyces are larger proximally and gradually smaller and denser towards the distal end. The calyces distally are edged with papillae. Along the asulcate side, the spinous area starts more distally than the sulcate and lateral side. The proximal calyces are wider and elongated, with eight to nine well defined calyces extending towards the lateral side and the distal part of these calyces are scalloped. The calyces towards the distal part are with more of papiliated edges and are similar in micro-ornamentation with the sulcate side.
Variation among the paratypes. See Table 2   and NCBS-AU733 are mostly single or double in the dorsal and double (single in few scales) above sacral but from the anterior part of the body, it is consistently double in all the scales from the position of the 10 th ventral (single before that) until the tip of the tail for ZSI-CZRC-6639. NCBS-AU733 differs from the holotype in having a slightly darker frontal and suprocular scales, presence of irregular white spots in the temporal scales, and a small white line along the midline between the two prefrontals. The first band on the body is separated from the "∏" shaped markings by 2 scales on both the sides along the  lateral side. 46-48 bands on the dorsum from neck to vent, anterior 18 bands are prominent with white and black edges which fades and become wider towards the vent.
BNHS 3516 differs from the holotype by having dorsum darker, two prominent and two less prominent white spots parallel to each other in each parietal. The slanting "∏" mark behind the head, two corners starts from the first row of scales behind parietal but shifts to the next row of dorsal after 4 (right) and 5 (left) scales, the other two diagonal corners ends on the first dorsal scales near 4 th ventral. There are 16 subequal horizontal bands on 2-3 dorsal scales, first band on the body separated from the "∏" shaped markings by three scale on right and two scales on left and all the other bands laterally connected with a mix of broken pale and black markings. There are 34 bands on the dorsum from neck to vent, anterior 16 prominent which becomes less prominent but wider towards the vent. ZSI-CZRC-6639 is overall similar to the holotype in the dorsum colours but with irregular white spots on the dorsal side of the head, a faint white spot at the junction of preocular, supraocular, prefrontal and frontal on both the sides; a faint white stripe in the anterior end of the frontal, the white spot in the suture between frontal and suproculars on both the sides, faint; 8-10 irregular spots together on both parietals. The slanting "∏" marking in the back of the head starts from the second row after the parietals on both the sides and continues towards the body but irregular. The first band on the body is separated from the "∏" shaped markings by 2 scales on both the sides along the lateral side. There are 42-44 bands on the dorsum from neck to vent, 14 prominent (white with black edges on each scale) which becomes less prominent but wider towards the vent. NMW 24565: 2 differs from holotype in having much lighter dorsum, by the presence of a white spot on the preocular-anterior frontal on both sides, two white spots along the midline suture on both parietals, one pair adjoining to frontal edge and other in the middle. The slanting "∏" shaped markings are much shorter extending only up to 7 rows of dorsal scales. About 39 visible bands on the dorsum of which anterior 13 are distinctly visible (see . Variation in dentition. The male paratype from "Salem" (NMW 25465: 2) show 12/12 maxillary teeth followed by a distinct diastema which is 50% longer as the socket of the last precranterian tooth and two enlarged cranterian teeth, 11/11 palatine teeth with a posteriomedial process of palatine expanding three pterygoid teeth, 15/14 pterygoid teeth with 41% of the posterior pterygoid without teeth, and 15-16 mandibular teeth, increasing in size up to tooth 7/7 and from there decreasing in size posteriorly.
Pholidosis. Head 1.25-1.99 times longer than broad (male 1.25-1.67, females 1.35-1.99). Rostral twice as broad as high and visible from above. Internasals (in midline suture) usually smaller than prefrontals. Frontals 1.32-1.59 times longer than the maximum width (male 1.40-1.59, females 1.32-1.42). Posterior border of parietals uneven in shape at the median suture forming an acute angle in ZSI-CZRC-6639 and NCBS-AU733 or both parietals uneven forming a slightly acute angle in BNHS 3516 or more or less straight with an obtuse angle in NMW 25465:2 and NCBS-AU-732. Posterior edge of parietals less than half the maximum width. Loreals usually longer than high or higher than long in NCBS-AU732 (on both sides). Preoculars single with a small pre-subocular present below it. Preocular usually in contact with frontal or rarely separated in, e.g., NCBS-AU733. Predominantly eight supralabials, usually last three being larger (in length), fifth highest. Fourth and fifth supralabials in contact with eye. Postoculars often two with the upper scale slightly wider than the lower scale that is usually higher or somewhat subequal in e.g., ZSI-CZRC-6639 and NCBS-AU732 in left side or lower scale smaller in ZSI-CZRC-6639 on the right side but in one specimen NMW 25565:2 with three postoculars on the right side. Usually two anterior temporals (on both sides). Secondary temporals variable, normally three but sometimes two in, e.g., NMW 25465: 2 and ZSI-CZRC-6639 (on right side) and BNHS 3516, NMW 25465: 2 and NCBS-AU732 (on left side). Mostly ten sublabials, sometimes nine in, e.g., NCBS-AU732 or eleven in NMW 25465: 2 (on both sides); the anterior four, rarely five in contact with first inframaxillary, sixth, rarely seventh largest in NMW 25465: 2. Anterior inframaxillaries normally longer than wider than the posterior ones except in BNHS 3516 and ZSI-CZRC-6522. The posterior pair of inframaxillaries usually separated by two or three rows of gular scales of variable shape and size. Gulars in four to five oblique rows between the apical edge of the posterior inframaxillaries and first ventral. Ventrals in examined material 189-218 (males 192-197, females 189-218). Usually, one or two preventrals, rarely without preventrals as in NCBS-AU733. Anal scute divided right part overlapping left part. Subcaudals in 76-88 in pairs (males 82-88, females 76-88). Total body scales, including preventrals and terminal scale 275-297 (males 276-288, females 275-297).
Dorsal scales with single or paired apical pits. Supracaudal scales usually with one to three pits on the first three to five scales followed the supracaudal reduction. In male (ZSI-CZRC-6639), consistently two pits from the midbody to the last supracaudal scale.
Dorsal scale reduction formula summarized from five examined specimens (see Appendix 10). Only main reductions are given. Two examined specimens (NCBS-AU732, holotype, and NCBS-AU733) and one specimen (ZSI-CZRC-6521) show an additional lateral increase to 18 and reduction to 16 dorsal scales involving rows 1-3 between 95 and 99% of ventral scales respectively.

Dimensions and proportions.
The body robust, moderately stout, roundish in cross section at midbody. Head moderately pointed, barely distinct from neck or sometimes slightly distinct. Eye large with black round pupil, 16-25% of head length. The longest examined specimens are from Karur (ZSI-CZRC-6639), a male specimen with 740 mm and a female from Tuticorin (ZSI-CZRC-6521) with 951 mm total length. The smallest specimens are from Salem including a male (NMW 25465: 2) with 582 mm and female (NMW 25465:1) with 315 mm total length. The tail/body length ratio ranges from 0.15 to 0.32 (males 0.29-0.32, females 0.15-0.32).
Habitat and natural history. Platyceps josephi sp. nov. mostly inhabits open habitats with sandy or rocky patches in grasslands and scrublands in both inland and coastal areas of Tamil Nadu from elevations between 10 and 580 m a.s.l. (see Fig. 21 and Appendix 9). Most of the areas where they occur receive less than 500 mm annual rainfall except locations near the rain shadowed areas close to the Western Ghats receives higher rainfall (Anaikatti and Megamalai) (http://www.tnsccc.in/rainfall.php). Sagadevan et al. (2019) report a sighting from the dry evergreen forest in Auroville, Villupuram district, but it is pertinent to mention here that this region was originally a scrub jungle and the current dry evergreen forest is an anthropogenic habitat created a few years ago (Blanchflower 2005).
Platyceps josephi sp. nov. is usually observed under thorny bushes, rock boulders, paddy fields, heaps of dry coconut fronds and seen crossing roads (Rajendran 1986; NJ and MS personal observation). The holotype (NCBS-AU732) and the paratype (BNHS 3156) were both collected from human habitations. ZSI-CZRC-6639 was found under heaps of coconut fronds within a plantation surrounded by dry grasslands. NCBS-AU733 was collected from under a woodpile near a farmhouse surrounded by grassland. Two other referred specimens (ZSI-CZRC 6521 and 6522) from Tuticorin and one specimen (ZSI-CZRC-7358) from Anaikatti, Coimbatore were collected as roadkilled. It appears to be an uncommon snake in all known locations. However, in the past at least in Megamalai hills, it was recorded to be "fairly common" in the dry deciduous forests (Hutton 1949). The late Naveen Joseph and his team, who maintain a record for snake rescues in and around Tuticorin, had seen only 16 individuals in the 18 years between 2002 and February 2021.
Platyceps josephi sp. nov. is a diurnal snake, terrestrial and swift in locomotion. However, there is a report of arboreal behaviour under artificial conditions (Rajendran 1986), but this is not known in the wild. It is an aggressive snake, flattening its head as a mock hood display mimicking a cobra, when agitated and biting freely, but it becomes docile after a few days in captivity (Rajendran, 1986;Hutton, 1949). Similar behaviour was observed during this study for specimens from Tuticorin, Tamil Nadu (NCBS-AU732, NCBS-AU733 and BNHS 3516).
The diet of the P. josephi sp. nov. chiefly consists of geckos, lizards and small rodents. It is observed to kill the prey by constricting and/or crushing the prey against ground or tree trunk. In all captive observations, it swallows the prey from the head first. It is an oviparous snake with few records of clutch sizes, 7-12 eggs (Hutton 1949;Rajendran 1986;Rameshwaran 2008), during the month of March and June and were 40 mm in length (Rameshwaran 2008). Females are reported to lay their eggs in bunds (raised areas surrounding the paddy fields) of paddy fields (Rajendran 1986).
Platyceps josephi sp. nov. is reported only from one protected area (Megamalai) in its known range. The species faces a number of threats across its distributional range, including habitat destruction, because the grasslands in southern Tamil Nadu are being actively converted into plantations, farmlands and urbanisation. Although there are records from such converted plantations and human settlements,the species' ability to adapt and its reproductive success is not known. Rocky habitats in Madurai region are also highly affected by the mining activities and road traffic is another important threat to P. josephi sp. nov. Three out of the seven specimens collected in this study were roadkilled and six other uncollected roadkilled specimens were observed from various parts of Tamil Nadu between 2017 and 2020. The Area Of Occupancy (AOO) of P. josephi sp. nov. is 72.000 km 2 and Extent Of Occupancy (EOO) is 70,698 km 2 . Even though this is a relatively large area of distribution for a species, our field surveys and records suggest that this species has patchy distribution within its range. Also, much of the habitat in these regions where P. josephi sp. nov. is reported are under severe threats like conversion of grasslands to farmlands, widescale monoculture plantations (Eucalyptus sp.) and urbanization. Given this information, we suggest that P. josephi sp. nov. should be considered a species in the Vulnerable category according to the IUCN criteria.
Although local envenoming by congeners is reported for P. rhodorachis and P. najadum (Minton 1990;Kuch and Mebs 2002;Weinstein et al. 2011), bites from P. josephi sp. nov. seem to be harmless and without any local symptoms. Rajendran (1986) reported two bite cases, one being himself got bitten while trying to catch an individual in a paddy field and another for an adult male got bitten while harvesting the paddy. No local symptoms were observed after a bite from a later preserved specimen (BNHS 3516, paratype) to one of the authors (NJ) upon handling.
Juvenile Platyceps josephi sp. nov. can easily be confused with juveniles of Platyceps plinii but may be differentiated by the two slanting roughly "∏" shaped markings on the back of the head. Platyceps josephi sp. nov. is found in sympatry with the Common cobra (Naja naja) and the Common ratsnake, Ptyas mucosa, and can potentially also be confused with these two species.
Differences between Platyceps josephi sp. nov. and South Asian congeners. Platyceps josephi sp. nov. is distinguished clearly from P. bholanathi, P. mintonorum, P. noeli, P. rhodorachis, P. sindhensis and P. ventromaculatus by its higher number of midbody dorsal scale rows (21 vs. 19) and from P. gracilis by its lower mean value of ventral scales (202 vs. 214), its lower value of subcaudal scales without overlapping (88 maximum vs. 118 minimum), its different neck pattern (whitish ∏-shaped marking vs. a yellowish-cream and black edged V-shaped marking), and by differing dorsal body pattern (irregular whitish transverse bars stippled with black vs. dorsal yellowish-cream and black-edged transverse bars). Platyceps josephi sp. nov. shows similarities with P. plinii with regard to pholidosis and colour pattern but can be distinguished from the latter by its lower number of midbody dorsal scale rows (21 vs. 23), its lower mean value of ventral scales (202 vs. 214), its lower mean value of subcaudal scales (82 vs. 87), and its lower mean of the sum of ventral and subcaudal scales (285 vs. 304).
Comparision of Osteology. The first detailed description of osteological features of Platyceps plinii were based on a male and a female specimen . Later, Wilson (1967) studied 15 preserved specimens of the Banded racer, including two disarticulated skeletons, and compared it among other species formerly allocated to the genus Coluber with material of racer-like genera including Asian species currently placed in Platyceps (e.g. elegantissimus, florulentus, gracilis, karelini, najadum, rhodorachis, rogersi, thomasi, variabilis and ventromaculatus) as well as Dolichophis jugularis, Hemorrhois ravergieri and with Orientocoluber spinalis. He described several osteological characters which he used to delimit C. fasciolatus from other species of Coluber (sensu lato) and, because of differences from other racer-like genera he resurrected the genus Argyrogena Werner for the Banded racer.
In his comparison, Wilson (1967) stated that the premaxilla of Argyrogena is short and stout and its lateral processes are posteriorly orientated, whereas Coluber (sensu lato) has a longer premaxilla with more nearly transverse processes. In Argyrogena the nasals are stout, overlap the ascending process of premaxilla to some extent and posteriorly cover the nasal capsule more than in Coluber (sensu lato), there is nearly no overlap between nasals and premaxilla. He further states that the pterygoid of Argyrogena show a medial flange on the bone with the result that the teeth are situated near the center of the pterygoid rather than along its medial edge as in the other species of Coluber (sensu lato), which have no medial flange. Additionally, he writes that the tooth rows on the pterygoid bones converge posteriorly, whereas in Coluber (sensu lato) the rows diverge and furthermore that the pterygoid processes at the basioccipital are lacking in Argyrogena, but well developed in Coluber (sensu lato). Beside skull morphology, he also studied the postcranial skeleton and included characters of midbody and caudal vertebrae in his investigation. He stated that in Argyrogena at the midvertebrae the accessory process is shorter and more dorsolaterally orientated and the ventral aspect of the centrum is stouter and that the transverse processes of the caudal vertebrae are broad based and show a prominent notch between the posterior edge of each process and the centrum.
Some of the osteological characters used by Wilson (1967) were later specified and their phylogenetic significance assessed by Schätti (1987). The latter did not include Argyrogena in his analysis but indicated extensive sexual, ontogenetic, geographical or uncorrelated intraspecific variation in old world racers. Schätti (1987) refused to use dentition to separate distinct groups as "The total number of teeth, their absolute length, the presence of a diastema, as well as the enlargement of the teeth posterior to the diastema is subject to considerable variation." For this reason, he considered osteological characters as diagnostic to delimit Palearctic from Nearctic groups, but only the shape and size ratios of the vertebrae show distinct characters to separate Platyceps from other Palearctic genera because craniological features may show remarkable intraspecific variation. Schätti (1987) took into account (1) length of centrum/ least width of neural arch (LC/WN), (2) length of centrum/width across prezygapophyses between outer edges of articular facets (LC/WP), (3) length of neural spine/ least width of neural arch (LN/WN), but emphasizes that especially the ratio (1) is useful for delimitation, because it varies only within narrow limits (see Appendix 6 for the description of the measuring distances).
Recently, Das et al. (2019) provided a detailed skull description based on six specimens of the Banded racer and distinguished it from Platyceps mainly by its longer and elongated nasal horizontal lamina. The latter authors also mentioned a backward curved transverse process in contrast to a laterally directed narrow transverse process in Platyceps, but did not name the belonging bone. We assume they were describing the premaxillary transverse processes. Finally, they separated Argyrogena from Platyceps by the existence of a mesial transverse process at the pterygoid vs. no mesial transverse process.
In contrast to the former studies on Argyrogena mentioned above, material examined by us lead to different results (see also Appendix 7 and 8). In all examined Platyceps species lateral processes of the premaxilla are always curved posteriorly, but with variable length. This can result in short lateral processes as described by Wilson (1967) and  and can be found in P. plinii comb. nov., P. josephi sp. nov., and P. rhodorachis, but also in longer lateral processes as present in P. florulentus, P. rhodorachis, P. ventromaculatus and also in P. plinii comb. nov. Furthermore, we found that short lateral processes lead to a more pointed snout, thus a projected rostrum, a condition also commonly used as a key character to separate Argyrogena from Platyceps (see e.g. Wilson 1967;Rajabizadeh et al. 2020). However, a projected rostrum of variable expression and a countersunk lower jaw is also reported for different Platyceps species by Schätti et al. (2014; see also comments below).
We found differences within the distal ends of the lateral processes of the premaxilla, which either are tapering into a single tip in P. plinii comb. nov., P. josephi sp. nov., and P. rhodorachis, expanded and divided into two tips in P. florulentus or expanded into a stout end in P. rhodorachis as well as in P. ventromaculatus. Because of this considerable intraspecific variation, we regard the form of the lateral processes of the premaxilla as an inappropriate character to delimit Argyrogena from Platyceps. Das et al. (2019, p. 313) stated that the premaxilla in Argyrogena has no nasal process but in contrast to that, a short (ascending) process is shown in the sketch (Fig. 1 B,lateral view). In all examined specimens we found an ascending process of the premaxilla, which varies in length and can be clasp by the anterior process of the nasals. If the ascending process of the premaxilla is longer than the lateral processes, we define them as 'long', e.g. in P. rhodorachis and P. ventromaculatus otherwise as 'short' as in found in P. plinii comb. nov., P. josephi sp. nov. and P. florulentus. Wilson (1967) noted that the anterior processes of nasals overlapping the ascending process of premaxilla in Argyrogena vs. less overlap of nasals and ascending process of premaxilla in Coluber (sensu lato). We found this overlap in all examined species, whereby the nasals can be in contact with the ascending process of premaxilla in P. plinii comb. nov., P. josephi sp. nov., P. florulentus and P. ventromaculatus or just have a loose connection as in P. rhodorachis. Since this is a variable character and its presence in the Banded racer and Platyceps spp., makes this an inappropriate character for distinguishing genera.
Furthermore, Wilson (1967) described the nasals of Argyrogena as stout, as was verified by our analysis. The Banded racer, P. plinii comb. nov., has the most compact nasal bone of all Platyceps examined (TLn/TWn range 0.93-1.04), distinguished by the following values, viz. P. josephi sp. nov. (TLn/TWn range 1.16-1.18), P. florulentus (TLn/TWn 1.24), P. rhodorachis (TLn/TWn range 1.27-1.39), and P. ventromaculatus (TLn/TWn 1.28). Wilson (l. c.) mentioned that in Argyrogena the nasal shield posteriorly expands dorsolaterally to cover the nasal capsule more completely as in the other Coluber (sensu lato). Our measurements show that the nasal shield of the Banded racer tends to be shorter (TLln/Lns range 40%-50%) than in other examined Platyceps spp. (TLln/ Lns range 56%-77%) and therefore covers only a smaller part of the nasal capsule. Our comparison of the nasal shields revealed that it covers in P. plinii comb. nov. the posterior part of the nasal capsule completely, whereas in the other examined Platyceps spp. (excluding josephi) only the middle part of the nasal capsule is enclosed. This resulted in a larger gap between the nasal shield and the frontal. Platyceps josephi sp. nov. shows an intermediate state regarding this character as the nasal shield do enclose the posterior part of the nasal capsule but shows a larger gap than observed in P. plinii comb. nov. This character seems to be highly influenced by the lifestyle of the species and varies considerably interspecifically (Schätti 1987). Wilson (1967) and  identified the presence of a medial flange at the pterygoid bone (with teeth) are situated near the centre of the pterygoid rather than along its medial edge vs. no such medial flange as a main character to separate Argyrogena from Coluber (sensu lato) and Platyceps spp. This condition was not mentioned by  and in Banded racer skulls studied by us, this medial flange is present in 60% of the examined specimens only (see Fig. 12  Another character listed by Wilson (l. c.) as typical for Argyrogena deals with the pterygoid teeth row, which converges posteriorly in the Banded racer but diverges in Coluber (sensu lato). We cannot follow this observation as in all examined specimens the teeth row follows the medial edge of the pterygoid bone and therefore always converge posteriorly.
Wilson (l. c.) also described the pterygoid processes at the basioccipital which he mentioned as clearly visible in e.g. Platyceps karelini, P. ventromaculatus, Dolichophis jugularis and Hierophis viridiflavus but lacking in Argyrogena. In Platyceps spp. examined by us, we found a high variability of this character, ranging from a basioccipital without structures, with three or five small tips or sometimes with a high crest. Our comparison of different development stages of D. jugularis, H. gemonensis and H. viridiflavus show that in juveniles such structures are lacking while in adults, high crests and tips appear and the expression of this character is very likely age-dependent.
Differences in midbody vertebrae morphology of racer species is highlighted as a character of phylogenetic significance (Schätti 1987) and is compared here with the next related taxa. Vertebrae have been described and compared for Argyrogena and Coluber (sensu lato, including Platyceps) by Wilson (1967), for old world racers in general by Schätti (1987) and for Platyceps by Schätti et al. (2014). Wilson (1967) stated that in Argyrogena the accessory processes at the mid-vertebrae are shorter and more dorsolaterally orientated as in Coluber (sensu lato). Based on Auffenberg's (1963) definition, Schätti (1987) defined the same character as ratio of the length of centrum to the width across prezygapophyses between outer edges of articular facets (LC/WP) (see Appendix 6-8). Our analysis of this character revealed, that intraspecific variation can be very low as in P. ventromaculatus (LC/WP range 0.76-0.86), but also more variable, e.g. in P. rhodorachis (LC/ WP range 0.55-0.88). The total range of ratio LC/WP for Platyceps spp. examined ranges from 0.55-0.88 and included, with a wide overlap, the Banded racer, Platyceps plinii comb. nov. with a LC/WP ratio range from 0.73-0.95 (see Appendix 7 and 8). Furthermore, this character state cannot be used to delimit the Banded racer and other Platyceps spp. from next related genera because of overlapping values as determined for Hemorrhois nummifer (LC/WP range 0.63-0.71), H. ravergieri (LC/WP range 0.66-0.86) and Spalerosophis diadema (LC/WP 0.62).
According to Wilson (1967) the ventral aspect of the centrum of middorsal vertebrae in Argyrogena is stouter than in Coluber (sensu lato). Schätti (1987) used this character as quotient of the length of centrum and the least width of neural arch (LC/WN), based on Auffenberg's (1963) definition. Our analysis of this character shows that for the Banded racer the values vary from 1.21-1.59 with lower extremes below the range of other Platyceps species with an LC/WN range from 1.32-1.70. But also this character does not allow reliable distinction between Platyceps spp. (LC/WN range 1.21-1.70) and Hemorrhois (nummifer 1.22-1.39; ravergieri 1.20-1.61). Although Spalerosophis diadema is distinguished here by the lowest value (LC/WN 1.08) it is based on a single examined specimen only and the variation for this character is not sufficiently studied for this species (see Appendix 6-8). As further character Schätti (1987) defines the ratio of length of neural spine to least width of neural arch (LN/WN) as highly variable throughout the vertebral column in Holarctic racers. The analysis of this trait with respect to the position of the Banded racer to Platyceps and other related genera revealed that the Banded racer tend to lower values (LN/WN range 0.81-1.01) but overlap to some degree the other examined Platyceps spp. Wilson (1967) used also the structure of caudal vertebrae to delimit Argyrogena from Coluber (sensu lato) and described the transverse processes of it as broad-based proximally with a prominent notch between posterior edge of each process and the centrum. In contrast to that, he described for Platyceps karelini, P. najadum and P. rhodorachis the processes as rather broad-based proximally, but not as much as in Argyrogena and without a posterior notch. In contrast to Wilson (l. c.) we found in five examined specimens of the Banded racer and four other Platyceps spp., including P. josephi sp. nov., the transverse processes of the caudal vertebrae as broad based but with striking differences between the sexes. In all males the transverse processes are directed laterally, the base occupies ~ 75% of the centrum and show a posterior notch, the latter characterized by considerable variability in expression even within a single individual. Contrary to males, the processes in females are directed anterio-laterally with a much shorter base which occupies not more than 60% of the centrum and show a posterior notch with the same degree of variability as in males. For specimens studied by us, we conclude that this character is sex dependent and qualifies no differentiation between Argyrogena and Platyceps.
In summary, it is evident that Argyrogena shares most osteological characters with Platyceps. Previous use of skeletal traits to underpin the distinctness of the genus Argyrogena is shown invalid, with the exception of the form of the nasal shield as the only distinguishing character to the previous known Platyceps spp. But with respect to the latter character, Platyceps josephi sp. nov. show an intermediate state and linked the Banded racer with Platyceps. We therefore consider Argyrogena Werner, 1924 a junior subjective synonym of Platyceps Blyth, 1860 and relegate the Banded racer (auctt.) to the genus Platyceps in the new combination Platyceps plinii (Merrem, 1820) (see the also conclusions in previous chapters).

Discussion
The maximum likelihood (ML) and Bayesian (BI) phylogeny presented in this study were similar but not same, there were some unresolved nodes in the BI phylogeny. Furthermore, the phylogenetic relationships are different from that presented by Rajabizadeh et al. (2020) (e.g. position of Persiophis and Muhtarophis). This is probably due to the addition of new lineages and/or missing gene sequences or both.  reported a three gene Bayesian phylogeny for old world racers including one Platyceps plinii sample from India. In this phylogeny, Platyceps plinii (as A. fasciolata) was recovered as sister to other Platyceps species with moderate support. Our multilocus phylogeny shows them to be monophyletic with the highest support. Within species genetic differences is also very low for Platyceps plinii and Platyceps josephi sp. nov. across the sampled range. We were not able to acquire samples of Platyceps plinii from Pakistan, this population appears to be disconnected from the rest of the subcontinent and warrants further investigation. The taxonomic confusions between two common and widespread species identified and resolved in this study has not been recognized for over 150 years. This highlights the importance of thorough literature review and verification of historic specimens and original sources before describing new species. In particular areas where there were extensive collections in the past and preserved materials are available from various collections worldwide (e.g. Indian subcontinent). Platyceps josephi sp. nov. is so far recorded only from Tamil Nadu state and all except two records (Salem and Auroville, Villupuram district) are south of the Cauvery River. The Nilgiri Mountains, the Moyar gorge and Cauvery River have been proposed to be a biogeographic barrier for fan-throated lizards (Deepak and Karanth 2018) and similar observations were also made in ground dwelling Hemidactylus geckos (H. scabriceps, Srikanthan et al. 2018). The northwestern distribution limit of Platyceps josephi sp. nov. lie south of Nilgiri Hills and Platyceps plinii is found north of Nilgiris in the Mysore plateau but their ranges appear to overlap in the northeastern part of Tamil Nadu (see, Fig. 1). Conversely, the southernmost distribution of Platyceps plinii and P. bholanathi extends only till the northeastern and northcentral part of Tamil Nadu and north of Cauvery River (Melvinselvan et al. 2018;this study). This further highlights the importance of the lesser known biogeographic barriers in peninsular India which were overlooked in the past.
In their study on Western Palearctic colubrid snakes, Rajabizadeh et al. (2020) provided a comparative morphological table with eight characters for colubrine genera of India, Central and Western Asia and North Africa. All these characters were identical for Wallaceophis and Wallophis and for Argyrogena and Platyceps as well. However, for the latter pair, the rostral scale character was incorrectly mentioned as "normal" for Argyrogena. Furthermore, within the currently recognized species of the genus Platyceps, some species show a projected ros-trum as noted by Schätti et al. (2014) which is particularly distinct for examined P. r. rhodorachis (including SMF 67907 and ZFMK 92802), P. cf. k. karelini (including CAS 120540), P. cf. mintonorum (including CAS-SU 13251) and hybrid racer e.g. P. k. karelini x P. mintonorum (including BMNH 1886.9.21.101). The validity of the genus Wallaceophis needs further scrutiny through additional morphological examination.
Muhamed Jafer Palot (ZSI Calicut, India) shared with us information about the occurrence of the racer in different parts of India and Karan Shah (Kathmandu, Nepal) provided information on species distribution in Nepal and translated text fragments from historic original drawings.
Carsten Eckert (Stiftung Schloss Friedenstein, Gotha, Germany) helped us with the transcription of historical documents. We thank Ashok Captain (Pune, India) for discussion on status and identification of Russell's "Nooni Paragoodoo" and "Patza Tutta" and for sharing images of Banded racers (Fig. 6 B, F (Thanjavur, Tamil Nadu), was then sent as surgeon to the Vellore General Hospital and is listed for 1786 as surgeon at Tanjore hospital (Crawford 1916, Chakrabarti 2006. He is mentioned together with Patrick Russell as head surgeon for the Madras Medical Service at Trichonopoly (Tiruchirappalli, Tamil Nadu) for the year 1788 (Crawford 1914).
Earlier attempts by  to locate "Casemcottah" failed and our intensive reseach on historical maps for the Indian Subcontinent revealed only a single similar spelled location on the southern western coast. On a map dated 1792, printed 1793 in London by William Faden, a place "Caser cotta" (probably today Kasergode town, Kerala state, 12.51000°N, 74.98527°E) is situated about 32 km south of Mangalore (Mangaluru, Karnataka) and directely south of Chianderghery (Chandragiri Fort, Kerala) respectively. It is documented by Russell that some material reached him from locations outside of his personal collecting area in the Visakhapatnam region.  listed 27 different localities in India for Russell's material including some from the western part of the subcontinent like e.g. Maharashtra. Therefore, it is possible that the region was visited by the two donors of the "Nooni Paragoodoo" during their military service, especially also because the area was already under British East India Company sovereignty after the Treaty of Mangalore in 1784.
But "Caser cotta" lies on the northern part of the Malabar Coast and therefore opposite of the Coast of Coromandel where Russell spend most of the time during his medical service for the East Indian Company. Furthermore, Russell (1803, p. 49) mentioned that he himself collected a fish referred to as "Kora Motta" (= Channa gachua (Hamilton, 1822) fide Fricke et al. 2020) in the lake at "Casem Cottah". In at least two gazetteers of geographical place names (Marshall 1832, p. 226, right column, andWright 1834, p. 188) a locality "Cossimcotta" is mentioned situated southeast of Ankapilly in division Cicacole, Hindoostan. Marshall (l.c.)

Note 2
The description of Coluber curvirostris published 1839 in Theodore E. Cantor's "Spicilegium Serpentium Indicorum" is a very short bilingual (Latin/English) section of nine lines of text only. Cantor's unpublished original manuscript entitled "Indian Serpents-Innocuous-Collected, figured & described (1831-1837)" is in the possession of the Bodleian Library, Oxford, U.K. This manuscript was the basis for Cantor's (1839) publication which un-fortunately was published without plates and with much reduced descriptions only.
We here provide the transcript of the detailed description of C. curvirostris extracted from Cantor's manuscript. The description is accompanied by the coloured plate no. 7 [reproduced here as Fig. 7].
The title page contains following additional text: "A catalogue of the species will appear in the Proceedings of the Zool. Soc. London. April 29. 1839", and the following information about the content of the manuscript: "35 Drawings {28 Figures of Serpents. New Species., 7. Anatomical details of Do.", and " The descriptions occupy 240 pages".
The description of C. curvirostris occupies the manuscript pages 37 to 44 as follows: "Coluber curvirostris/ Fig: 7./ C. supra partim late/ olivaceo-viridis punctis/ & lineis obliquis al-/ bis nigrisquae, -par-/ tim aeneus; abdomine subfusco/ Scuta abdominalia …… 220./ Scutella subcaudalia 85./ Habitat: Bengal./ Above partly bright/ olive green with white/ and|<38>|black dots & oblique/ lines, -partly bron-/ ce-coloured; light/ yellow underneath./ Vernacular name:/ Tukkr-Bora./ The head a little broader,/ than the neck, subo-/ vate & strongly ar-/ ched, so that the/ profil of the rostrum/ is very oblique; the/ upper jaw extends/ about a line & a/ half beyond the low-/ er; |<39>| The nostrils moderate/ between two scuta./ The eyes, surrounded/ by two prae-& one post/ orbital scutum, large/ dark brilliant with a/ narrow reddish golden iris &/ a small round black/ pupil. The gape wide,/ straight with 8/8 labial/ scuta & with numerous little/ sharp teeth of equal size in the/ palaté & jaws. The/ tongue jet-black bi-/ furcate./ The shape of the body/ is elegant cylindrical;/ the tail more slender/ than |<40>| the trunk, gradually/ tapering into a small sharp point./ The scales of the body/ are oval-rhomboidal,/ very slightly imbricate,/ those of the tail shorter/ approaching a square/ form. The length of/ the abdominal scuta/ is about two lines;/ their breadth six lines./ The scutum of the/anus bipartite._/ The general colour/ is a shining bright/ olive green; on the an-/ terior part of the back,/ where the |<41>| sky-blue skin is visib-/ ble between the scales,/ a number of white & black dots, transversal-/ ly arranged, some of/ them forming rings./ About six inches from/ the head these dots change/ into white & bluish/ black oblique lines/ with points/ towards the head, ceasing/ on the posterior,/ third part of the body/ where the general color/ changes to a light/ brown |<42>| with a green metal-/ lic lustre./ The lips, the throat/ & the abdominal/ surface are of a light/ yellow./ *The upper surface of the/ head being strongly arched,/ the measurement is taken on/ the lower surface from the projecting muzzle to the anterior margin/ of the third scutum abdominal |<43>| This not very common/ species is found in jungle although/ it may occasionally/ enter houses in/ search of mice & small/ lizards. It will sponta-/ neously starve for a/ considerable length of/ time: a female specimen/ in my possession lived/ upwards of eight months/ without food, during/ which period the skin/ was changed three times/ Having been shut up/ in a cage for more,/ than|<44>| five months, the cold/ weather inclusively, she/ brought forth in the commencement of May 1836 an egg/ of a cylindrical form, /1 ½ inch long, & half/ and inch in diameter/ of a light reddish/ yellow color. It was/ left in the cage for/ about a forthnight,/ until putrefaction required/ it removed. On opening/ the soft external mem-/ brane, it was found/ contain nothing but yolk./ Only when about changing/ the skin she became irritable /& attempted to bite as repre/ sented in the sketch./ This serpent is seldom/ found beyond three feet/ or so in length."

Note 3
The alleged collector Sir William Crofton Twynam (1827-1922) also known as "Raja of the North" was born in Galle. He entered the Ceylon Civil Service in 1845 and held from 1869 until his retirement in 1896 the post as Government Agent on Jaffna (Martyn 1923