A new rock gecko in the Cnemaspis siamensis group (Reptilia, Gekkonidae) from Kanchanaburi Province, western Thailand

We describe a new species of the gekkonid genus Cnemaspi s from Erawan National Park in Kanchanaburi Province of western Thailand. Molecular phylogenetic analyses, based on the mitochondrial NADH dehydrogenase subunit 2 gene and flanking tRNAs, revealed that Cnemaspis auriventralis sp. nov. is nested within the C. siamensis group and is closely related to C. huaseesom , but has uncorrected pairwise genetic divergences of 12.12–27.92% from all other species in the C. siamensis group. The new species is also distinguished from other species in the C. siamensis group by having the combination of snout-vent length 36.7–38.6 mm in males ( N = 3), 32.9–36.9 mm in females ( N = 2); eight to ten supralabials; seven to nine infralabials; ventral scales smooth; six or seven precloacal pores in males; 16–17 paravertebral tubercles linearly arranged; tubercles on the lower flanks present; lateral caudal furrows present; no caudal tubercles in the lateral furrows; ventrolateral caudal tubercles present anteriorly; caudal tubercles not encircling tail; subcaudals smooth bearing a single median row of enlarged smooth scales; two postcloacal tubercles on each side; no shield-like subtibial scales; subtibial scales smooth; no enlarged submetatarsal scales; 23–27 subdigital lamellae on the fourth toe; sexually dimorphic for dorsal and ventral colour pattern; prescapular marking absent; gular marking absent; and yellow colouration in life of all ventral surfaces of head, body and tail in adult males. The new species is currently known only from upland karst habitat at its type locality.


Introduction
The Southeast Asian Rock Gecko genus Cnemaspis Strauch, 1887 comprises a monophyletic clade of approximately 66 recognised species that are distributed from Laos, southern Vietnam westwards through southern Indochina, southwards through the Thai-Malay Peninsula, Sumatra, Java and eastwards to Borneo (Grismer 2010;Grismer et al. 2014Grismer et al. , 2020Wood et al. 2017;Riyato et al. 2019;Ampai et al. 2020;Quah et al. 2020;Nashriq et al. 2022;Uetz et al. 2022). Based on molecular and morphological data, Southeast Asian Cnemaspis are recovered in four major monophyletic clades that contain six species groups (Grismer et al. 2014). The Cnemaspis siamensis group is distributed across the Thai-Malay Peninsula and northwards to Kanchanaburi Province, western Thailand (Grismer et al. 2014Ampai et al. 2019). This group currently contains 13 named species (Grismer et al. , 2014Wood et al. 2017;Ampai et al. 2019Ampai et al. , 2020 including C. adangrawi Ampai, Materials and methods Field collection of specimens Field surveys were conducted at Erawan National Park, Tha Kradan Subdistrict, Si Sawat District, Kanchanaburi Province, Thailand, in November 2019 and November 2021 (Fig. 1). Direct observations were made during the day (09:00-17:00 h) and at night (19:00-22:00 h) and Cnemaspis specimens were collected by hand. Captured specimens were humanely euthanised using tricaine methanesulphonate (MS-222) within 24 hours of collection (Simmons 2015). Liver or muscle tissues were immediately removed from euthanised individuals, preserved in 95% ethyl alcohol and stored at -20 °C for molecular analysis. Euthanised specimens were fixed in 10% formalin and later transferred to 70% ethyl alcohol for permanent storage. Specimens and tissues were deposited in the herpetological collection of the Zoological Museum, Kasetsart University, Thailand (ZMKU). Geographic coordinates and elevations were recorded using a Garmin GPSMAP 64s with WGS84 datum. Ambient air temperature and relative humidity were collected with a Kestrel 4000 Weather Meter. Live animals and preserved specimens were photographed using a Nikon D700 or Z50 digital camera with an AF-S Micro Nikkor 60-mm f/2.8G ED lens and external flashes.

Morphology
Morphological characters taken and their abbreviations were modified from recent studies of the genus Cnemaspis (Wood et al. 2017;Ampai et al. 2020;Grismer et al. 2020). Morphological measurements were taken with digital calipers to the nearest 0.1 mm. Scalation and other aspects of external morphology were examined using a Nikon SMZ745 stereomicroscope. Measurements were taken on the left side of the body, while scale counts were taken on both right and left sides (R/L) when possible. Measurements and meristic characters are shown in Table  1 and qualitative observations of external morphology evaluated are described below.
Additional character states evaluated were the general size (i.e. strong, moderate, weak) and arrangement (i.e. random or linear) of the dorsal body tubercles; the orientation and shape of precloacal pores; the number of precloacal scales lacking pores separating the left and right series of pore-bearing precloacal scales; the degree and arrangement of body and tail tuberculation; the relative size and morphology of the subcaudal scales, subtibial scales and submetatarsal scales beneath the first metatarsal. Sex and maturity were determined by the presence of secondary sexual characteristics, such as the presence of hemipenes or pore-bearing precloacal scales in males, the presence of calcium glands or eggs in females or sexually dimorphic colour patterns. Morphological data for comparisons were obtained from the original and expanded descriptions of other species in the C. siamensis group (Smith 1925;Grismer and Chan 2010;Grismer et al. , 2014Grismer et al. , 2020Wood et al. 2017;Ampai et al. 2019Ampai et al. , 2020.

DNA extraction and PCR amplification
We extracted genomic DNA from the liver tissue of five individuals of Cnemaspis from Erawan National Park, Kanchanaburi Province (Table 2) using the DNeasy Blood and Tissue Kit (Qiagen, Germany) according to the manufacturer's protocol. A portion of the mitochondrial NADH dehydrogenase subunit 2 gene (ND2) and its flanking tRNAs was amplified via a double-stranded polymerase chain reaction (PCR), using the light strand primer L4437b (5'-AAGCAGTTGGGCCCATACC-3'; Macey et al. 1997) and heavy strand primer H5934 (5' AGRGTGCCAATGTCTTTGTGRTT-3'; Macey et al. 1997). PCR reactions were executed in an Eppendorf Mastercycler gradient thermocycler under the following conditions: initial denaturation at 95 °C for 2 min, followed by a second denaturation at 95 °C for 35 s, annealing at 55 °C for 35 s, followed by a cycle extension at 72 °C for 35 s, for 33-40 cycles with a final extension at 72 °C for 10 min. PCR products were purified using a QIAquick PCR Purification Kit (Qiagen, Germany). PCR products were sequenced in both forward and reverse directions using the same amplifying primers at Biobasic Asia Inc. (Singapore) on an ABI 3730XL automatic sequencer (Applied Biosystems, CA, USA). Sequences were visually checked and edited in Geneious R11 (Biomatters Ltd, Auckland, New Zealand). The protein-coding region of ND2 was translated to amino acids and checked to confirm the lack of premature stop codons. All new sequences were deposited in GenBank under accession numbers OP093974-OP093978 (Table 2).

Phylogenetic analyses
Additional homologous sequences of 68 individuals belonging to the C. affinis group, C. boulengerii group, the C. argus group, the C. chanthaburiensis group, the Figure 1. Map illustrating the type locality (yellow star) of Cnemaspis auriventralis sp. nov. at Erawan National Park, Si Sawat District, Kanchanaburi Province, Thailand and the Cnemaspis siamensis group samples used in the molecular analyses (asterisk) given in Table 2. Stars indicate type localities and circles represent additional localities.
Maximum Likelihood (ML) and Bayesian Inference (BI) were used to estimate phylogenetic relationships. Best-fit models of evolution for each partition was determined using the Bayesian information criterion (BIC) implemented in ModelFinder (Kalyaanamoorthy et al. 2017). The best-fit evolutionary models were TPM2u+F+I+G4 for tRNAs and TVM+F+I+G4, TIM3+F+G4 and GTR+F+G4 for ND2 codon positions 1, 2 and 3, respectively. The ML analysis was performed using the IQ-TREE webserver 1.6.12 (Trifinopoulos et al. 2016) with 1,000 bootstrap pseudo-replicates using the ultrafast bootstrap analysis (Minh et al. 2013;Hoang et al. 2018). The BI analysis was implemented in MrBayes v.3.2 (Ronquist et al. 2012) on the CIPRES Science Gateway V.3.3 (Miller et al. 2010) using default priors and models of evolution that were selected by the BIC and used in the ML analysis. Two independent runs, each with three heated and one cold chain, were performed using Markov Chain Monte Carlo (MCMC). The MCMC chains were run for 10,000,000 generations and trees sampled every 1,000 generations with the first 25% of each run discarded as burn-in. Stationarity was evaluated by ensuring effective sample sizes (ESS) were above 200 for all parameters in Tracer v. 1.7 (Rambaut et al. 2018). The phylogenetic trees from the ML and BI analyses were visualised using FigTree v. 1.4.4 (http://tree.bio. ed.ac.uk/software/figtree/). Nodes having ultrafast bootstrap support values (UFB) ≥ 95 and Bayesian posterior probabilities (BPP) ≥ 0.95 were considered highly supported (Huelsenbeck and Ronquist 2001;Wilcox et al. 2002;Minh et al. 2013). Uncorrected pairwise sequence divergences (p-distances) were calculated in MEGA 11 (Tamura et al. 2021) using the pairwise deletion option to remove gaps and missing data from the alignment prior to analysis.

Results
The final alignment of ND2 and flanking tRNAs contained 1,327 characters of 71 individuals of Cnemaspis and two individuals of outgroup species (Table 2). The average standard deviation of split frequencies was 0.000732 and the ESS of all parameters were ≥ 5,153 for all parameters in the BI analysis. The best tree in the ML analysis had a Maximum Likelihood value (lnL) of -22,848.995. The ML and BI analyses recovered trees with topologies similar to each other and to those recovered by Ampai et al. (2020) (Fig. 2). The five samples from Erawan National Park formed a strongly supported monophyletic lineage (≥ 95 UFB, ≥ 0.95 BPP) within the C. siamensis group. The Erawan National Park population was strongly supported (≥ 95 UFB, ≥ 0.95 BPP) to be the sister taxon of C. huaseesom from Sai Yok National Park, Sai Yok District, Kanchanaburi Province. Uncorrected pairwise genetic divergences (p-distances) within the Erawan

4TL
The number of subdigital lamellae beneath the fourth toe, counted from the base of the first phalanx to the claw

PP
The total number of pore-bearing precloacal scales in males

PPS
The number of postcloacal tubercles on each side of tail base National Park population were 0.00-0.65%. The Erawan National Park population had uncorrected p-distances of 12.12-12.55% from C. huaseesom and 15.15-27.92% from the other species in the siamensis group. The p-distances amongst species in the C. siamensis group ranged from 8.23-29.00% (Table 3).

Taxonomic hypotheses
Cnemaspis samples from Erawan National Park, Tha Kradan Subdistrict, Si Sawat District, Kanchanaburi Province differed from congeners in mtDNA analyses and diagnostic morphological characters (see "Comparisons"). Based on these corroborating lines of evidence, we hypothesise that the Erawan National Park population represents a previously unnamed species, which is described below. Diagnosis. Cnemaspis auriventralis sp. nov. can be distinguished from all other species in the C. siamensis group by having the following combination of morphological and colour pattern characters: SVL 36.7-38.6 mm in adult males (N = 3), 32.9-36.9 mm in adult females (N = 2); eight to ten supralabials; seven to nine infralabials; ventral scales smooth; six or seven precloacal pores in males; 16-17 paravertebral tubercles linearly arranged; tubercles on the lower flanks present; lateral caudal furrows present; no caudal tubercles in the lateral furrows; ventrolateral caudal tubercles present anteriorly; caudal tubercles not encircling tail; subcaudals smooth bearing a single median row of enlarged smooth scales; two post-cloacal tubercles on each side; no shield-like subtibial scales; subtibial scales smooth; no enlarged submetatarsal scales; 23-27 subdigital lamellae on the fourth toe; sexually dimorphic in dorsal and ventral colour pattern; prescapular marking absent; gular marking absent; and yellow colouration in life on all ventral surfaces of head, body and tail in adult males.
Body slender, elongate (AG/SVL 0.42); small, raised, weakly keeled, dorsal scales generally equal in size throughout body, intermixed with numerous, large, multi-keeled, linearly arranged tubercles; enlarged, multi-keeled, conical tubercles on flanks; tubercles extend from the occiput to base of the tail and continue on tail in whorls; body tubercles slightly smaller anteriorly; 17 paravertebral tubercles; pectoral and abdominal scales smooth, flat, imbricate; abdominal scales larger than pectoral and dorsal scales; seven contiguous, pore-bearing, precloacal scales; precloacal pores round to elongate.
Tail original (broken at tip), long, slender, TL = 45.9 mm (TL/SVL 1.21); dorsal, caudal scales arranged in segmented whorls; caudal scales keeled, raised, juxtaposed; mid-dorsal and lateral, caudal furrows present; subcaudals smooth; median row of enlarged subcaudal scales present; paravertebral, dorsolateral and lateral rows of large, keeled, caudal tubercles extend length of tail; ventrolateral rows of tubercles present only anteriorly; caudal tubercle rows do not en- Table 3. Mean (minimum-maximum) percentages of uncorrected pairwise sequence divergences (p-distances) of Cnemaspis species in the C. siamensis group compared to C. auriventralis sp. nov., based on 1,327 aligned characters of the mitochondrial NADH dehydrogenase subunit 2 gene and flanking tRNAs. Intraspecific p-distances are in bold font.     (Figs 3-5). Dorsal ground colour of head, nape and fore-limbs grey; dorsal ground colour of trunk and hind-limbs yellowish-grey; dorsal ground colour of tail yellow; top of head bearing small, diffuse, faint, dark and light markings; dark postorbital stripes faint extending to occiput; pair of dark, diffuse, blotches on nape; large, light, irregularly-shaped, vertebral blotches extend from nape to base of tail, continuing on to tail as light yellow caudal bands; small, light, irregularly-shaped blotches in shoulder regions and flanks; limbs mottled with small, diffuse light and dark blotches; digits light grey bearing thin, dark bands; all ventral surfaces of head, body and tail yellow; ventral surfaces of limbs light grey with yellow speckling. Colouration in preservative. (Fig. 6). Dorsal and lateral surfaces of head, body, limbs and tail darker grey than in life, with some fading of markings. Ventral surfaces of head, body, limbs and tail creamy-white, with minute dark speckling on gular region, limbs and tail regions.
Variation. Cnemaspis auriventralis sp. nov. shows significant sexual dimorphism in colour pattern. All female paratypes lack yellow colouration on the tail and ventral surfaces. Female paratype ZMKU R 01003 was darker coloured in life than other members of the type series. Two male paratypes ZMKU R 00999-01000 were lighter coloured than the holotype in life. Dark markings on the dorsum of all paratypes are more prominent than the holotype. The female paratypes (ZMKU R 01002-01003) lack precloacal pores and have postcloacal tubercles that are relatively smaller than those in males. One male paratype (ZMKU R 01000) has six (2R/4L) pore-bearing pre-  cloacal scales separated by a single scale lacking pore. Variation in morphometric and meristic data amongst specimens in the type series are presented in Table 4.
Distribution. This species is known only from the type locality. The type series was collected from a karst formation that is part of the Tenasserim Mountain Range in Erawan National Park, Tha Kradan Subdistrict, Si Sawat District, Kanchanaburi Province, Thailand (Figs 1, 9). Natural history. The type locality of Cnemaspis auriventralis sp. nov. is a karst forest at 747 m elevation. The type series was collected in November 2019 at 14:00-14:30 h with temperature 25.4 °C and relative humidity 62.1% and in November 2021 at 16:00-20:00 h with temperature 25.3-26.8 °C and relative humidity 79.9-88.0%. All specimens were found on karst boulders along the nature trail to the Tham (= cave) Phra That. Most were observed clinging upside down to the undersides of large boulders (around 1-3 m 2 ) or in narrow crevices of the boulders. Four individuals (ZMKU R 00999-01002) were found during the daytime (14:00-16:30 h) and one individual (ZMKU R 01003) was found at night (20:30 h). Most observed individuals were found during the daytime and they were active and rapidly escaped from disturbances. The one individual found at night was relatively inactive (slow-moving). At night, the new species was found to co-occur with five other species of gekkonid lizards, Cyrtodactylus monilatus Yodthong, Rujirawan, Stuart, Grismer, Etymology. The species epithet auriventralis is derived from aurum (L.) for gold and ventralis (L.) for venter in reference to the new species having distinct yellow colouration on all ventral surfaces of the head, body and tail in adult males. We propose "Erawan Rock Gecko" for the common English name and "จิ ้ งจกนิ ้ วยาวเอราวั ณ" (Jing Jok Niew Yao Erawan) for the common Thai name of the new species.
Comparisons. Cnemaspis auriventralis sp. nov. is distinguishable from all other members of the C. siamensis group by a combination of morphological and colour pattern characteristics (see Table 5 for additional comparisons).

Discussion
Our phylogenetic analyses indicated that Cnemaspis auriventralis sp. nov. belongs to the C. siamensis group and is closely related to C. huaseesom from Sai Yok National Park, Kanchanaburi Province, approximately 25 km to the west. Ecologically, the new species and its close relative C. huaseesom occur in similar habitats and substrates (karst associated areas). However, the new species was found at 747 m elevation, whereas C. huaseesom was found in lowland areas (Grismer et al. , 2014. Therefore, the geographic boundaries of these two species could be separated by elevation zonation (upland [> 600 m] and lowland [< 600 m] species; Grismer et al. 2014). The description of C. auriventralis sp. nov. brings the total number of Thai Cnemaspis to 21 species (Grismer et al. 2014;Uetz et al. 2022). Cnemaspis auriventralis sp. nov. is the seventh new Thai Cnemaspis species described in the last five years (Ampai et al. 2020;Grismer et al. 2020;Uetz et al. 2022). The number of known Thai Cnemaspis species continues to increase, likely as a result of new field research in poorly known areas and the use of integrative taxonomic approaches to delimit species in this genus (Wood et al. 2017;Ampai et al. 2019Ampai et al. , 2020Grismer et al. 2020). Additional field surveys and further taxonomic investigations using multiple lines of evidence in western Thailand are needed to determine the extent of the geographic range of the new species and to improve documentation of the herpetofaunal diversity in Thailand.