Morphological and molecular analyses reveal two new insular species of Cnemaspis Strauch, 1887 (Squamata, Gekkonidae) from Satun Province, southern Thailand

Abstract We describe two new insular gecko species of the genus Cnemaspis from Tarutao, Adang, and Rawi islands in Satun Province, southern Thailand. The new species are distinguished from their congeners in having a unique combination of morphological, scalation, and color pattern characters, and by genetic divergence in the mitochondrial NADH dehydrogenase subunit 2 (ND2) gene. Cnemaspistarutaoensissp. nov. was found to be a member of the C.kumpoli group, but is distinguished from all other species in that group by having 8–9 supralabials and 8 infralabials; 4–5 pore-bearing precloacal scales, pores rounded; 17–19 paravertebral tubercles randomly arranged; 27–29 subdigital lamellae under the fourth toe; subcaudal region yellowish, with smooth scales and a single enlarged median row; black gular markings in males and females; and 17.24–22.36% uncorrected pairwise sequence divergences. Cnemaspisadangrawisp. nov. was found to be a member of the C.siamensis group, but is distinguished from all other species in that group by having 10 supralabials and 9 infralabials; 6–8 pore-bearing precloacal scales, pores rounded and arranged in a chevron shape; 23–25 randomly arranged, separated paravertebral tubercle rows; 26–28 subdigital lamellae under the fourth toe; subcaudal scales keeled, without enlarged median row; gular region, abdomen, limbs and subcaudal region yellowish in males only; gular marking absent in males and females; and 8.30–26.38 % uncorrected pairwise sequence divergences. Cnemaspistarutaoensissp. nov. occurs in karst formations on Tarutao Island, while Cnemaspisadangrawisp. nov. is found near granitic, rocky streams on Adang and Rawi islands.


Introduction
Southeast Asia is a global biodiversity hotspot with extraordinary levels of species endemism (Myers et al. 2000).Southern Thailand serves as an important biogeographic transition zone between the Indochinese and Sundaic biotas, especially at the Isthmus of Kra and the Kangar-pattani line (Hughes et al. 2003;Woodruff and Turner 2009;Woodruff 2010;Parnell 2013).Southern Thailand has high levels of species diversity and endemism of reptiles (Sodhi et al. 2004;Grismer et al. 2010;Das and van Dijk 2013;Wood et al. 2017).
During recent fieldwork in 2017-2018 on Tarutao, Adang, and Rawi islands, Satun Province, southern Thailand, specimens of Cnemaspis were collected that differed from all other named species.Herein, we evaluate the morphological and molecular distinctiveness of these specimens.

Sampling
Specimens of Cnemaspis were collected from Tarutao, Adang, and Rawi islands in Tarutao National Park, Mueang Satun District, Satun Province, Thailand (Fig. 1) between November 2017-April 2018.Specimens were collected by hand during the day (1000-1800 h) and at night (1900-2200 h).Liver or muscle samples for genetic analysis were collected and preserved in 95% ethanol after euthanasia.Specimens were fixed in 10% formalin and later transferred to 70% ethanol for permanent storage.Specimens and tissue samples were deposited in the herpetological collection at the Zoological Museum of Kasetsart University, Bangkok, Thailand (ZMKU) and the Thailand Natural History Museum, Pathum Thani, Thailand (THNHM).

Morphology
Only adult individuals were used in the morphological analysis, as determined by the presence of hemipenes or precloacal pores in males, and the presence of calcium glands or eggs in females.Measurements were taken by the first author on the left side of preserved specimens to the nearest 0.1 mm using digital calipers under a Nikon SMZ 445 dissecting microscope.Sixteen measurements were taken following Grismer et al. (2014) and Wood et al. (2017): snout-vent length (SVL), taken from tip of snout to the anterior margin of vent; tail width (TW) at the base of the tail immediately posterior to the postcloacal swelling; tail length (TL), as distance from the vent to the tip of the tail, whether original or regenerated; forearm length (FL), taken on the dorsal surface from the posterior margin of the elbow while flexed 90° to the inflection of the flexed wrist; tibia length (TBL), taken on the ventral surface from the posterior surface of the knee while flexed 90° to the base of the heel; head length (HL), as distance from the posterior margin of the retroarticular process of the lower jaw to the tip of the snout; head width (HW) at the angle of the jaws; head depth (HD), as the maximum height of head from the occiput to the throat; axilla-groin length (AG), taken from the posterior margin of the forelimb at its insertion point on the body to the anterior margin of the hind limb at its insertion point on the body; eye diameter (ED), as the maximum horizontal diameter of the eyeball; eye-snout distance (ES), measured from the anterior margin of the eyeball to the tip of snout; eye-ear distance (EE), measured from the anterior edge of the ear opening to the posterior edge of the eyeball; eyenostril distance (EN), measured from the anterior most margin of the eyeball to the posterior margin of the external nares; inner orbital distance (IO), as the width of the frontal bone at the level of the anterior edges of the orbit; internarial distance (IN), measured between the medial margins of the nares across the rostrum; and ear length (EL), taken from the greatest vertical distance of the ear opening.
Meristic characters of scale counts and external observations of morphology were taken following Grismer et al. (2014) and Wood et al. (2017): number of supralabial and infralabial scales, counted from below the middle of the orbit to the rostral and mental scales, respectively; texture of scales on the anterior margin of the forearm; number of paravertebral tubercles between limb insertions, counted in a straight line immediately left of the vertebral column; presence or absence of a row of enlarged, widely spaced, tubercles along the ventrolateral edge of the body flank between limb insertions; number of subdigital lamellae beneath the fourth toe (=4 th toe lamellae), counted from the base of the first phalanx to the claw; general size (i.e., strong, moderate, weak) and arrangement (i.e., random or linear) of dorsal body tubercles; number, orientation and shape of precloacal pores; relative size of subcaudal and subtibial scales; and number of postcloacal tubercles on each side of tail base.
Comparative material was examined in the holdings of THNHM (Appendix 1), and comparative data were obtained from the original descriptions of other Thai species of Cnemaspis (Grismer et al. 2009;Grismer and Chan 2010;Grismer et al. 2010;Wood et al. 2017).

Phylogenetic analyses
Homologous sequences of 56 Cnemaspis and the outgroups Cyrtodactylus intermedius and Hemidactylus garnotii (following Bauer et al. 2008;Grismer et al. 2015b) were downloaded from GenBank and aligned to the eight newly generated Cnemaspis sequences using Geneious v. 5.6.3(Biomatters, Auckland, New Zealand).The aligned dataset was partitioned into four partitions consisting of ND2 first, second and third codon positions, and tRNAs.Phylogenies were reconstructed with the maximum likelihood (ML) criterion using IQ-TREE v. 1.6.7 (Nguyen et al. 2014) on the IQ-TREE web server (Trifinopoulos et al. 2016).The best-fit model of substitution for each partition was estimated using IQ-TREE's ModelFinder function (Kalyaanamoorthy et al. 2017) under the Akaike Information Criterion (AIC).The selected models were TIM+F+R4 for each ND2 codon position partition, and HKY+F+R4 for the tRNA partition.Bootstrap analysis was performed using the ultrafast bootstrap approximation (Minh et al. 2013) with 1,000 replicates and 0.95 minimum correlation coefficient.
Phylogenies were also reconstructed with Bayesian Inference (BI) using MrBayes v. 3.2 on XSEDE on the Cyberinfrastructure for Phylogenetic Research (CIPRES; Miller et al. 2010) computer cluster.The best-fit model of substitution was estimated for each partition using jModelTest 2.1.10(Posada 2008) under AIC.The selected models were GTR+ I+Γ for each ND2 codon position partition, and HKY+ I+Γ for the tRNA partition.Two simultaneous runs, each with three heated and one cold chain, were performed using the default priors for 10 × 10 6 generations, with trees sampled every 1,000 generations from the Markov Chain Monte Carlo (MCMC).Runs were halted after the average standard deviation of split frequencies was below 0.01 and convergence was assumed.The first 25% of the trees were discarded as burn-in using the sumt command.The convergence of the two simultaneous runs, and stationary state of each parameter, were evaluated using Tracer v. 1.6 (Rambaut et al. 2014).Runs were terminated when the effective sample sizes (ESS) of all parameters was greater than or equal to 200.
The most likely tree in the ML analysis, and the 50% majority-rule consensus of the sampled trees from the BI analysis, were visualized using FigTree v. 1.4.3 (Rambaut 2009).Nodes having bootstrap support (BS) of ≥70 and posterior probabilities (PP) of ≥0.95 were considered to be well-supported (Huelsenbeck and Ronquist 2001;Wilcox et al. 2002).Uncorrected pairwise sequence divergences were calculated using MEGA v. 7.0.26(Kumar et al. 2016).

Molecular analyses
The aligned dataset contained 1,296 characters of 64 individuals of Cnemaspis and two individuals of the outgroup species.The standard deviation of split frequencies among the two simultaneous BI runs was 0.001478.The ESS values were greater than or equal to 3,630 for all parameters.A single most likely tree resulted from the ML analysis.
The most likely ML tree and the 50% majority rule consensus tree from the BI analysis had similar topologies (Fig. 2).Cnemaspis samples from Tarutao Island represented a well-supported lineage (100 BS, 1.0 PP) within the kumpoli group, and was recovered as the sister species to C. monachorum from Pulau Langkawi, Malaysia (100 BS, 1.0 PP).The Tarutao samples differed from one another by uncorrected p-distances of 0.00-0.31%,but from other members of the kumpoli group by uncorrected p-distances of 17.24-22.36% (Table 2).
Cnemaspis samples from Adang and Rawi islands represented a well-supported lineage (100 BS, 1.0 PP) within the siamensis group, and was recovered as being closely related to a clade containing C. chanardi, C. phangngaensis, C. omari, and C. roticanai (Fig. 2).However, the exact sister taxon relationship of the Adang and Rawi islands was not resolved with strong support (Fig. 2).The Adang-Rawi samples differed from one another by uncorrected p-distances of 0.00-4.68%, but from other members of the siamensis group by uncorrected p-distances of 8.30-26.38 % (Table 3).

Taxonomic hypotheses
The Tarutao and Adang-Rawi samples of Cnemaspis differed from each other and all other congeners by being diagnosable in morphology and mitochondrial DNA.Owing to these independent lines of evidence, we hypothesize that the Tarutao and Adang-Rawi samples represent two new species to science, and are described below.
Body slender, elongate (AG/SVL 0.39); small, raised and equal in sized, dorsal scales throughout body intermixed with several large, multicarinate tubercles random-   ly arranged; 19 paravertebral tubercles; tubercles absent on lower flanks; tubercles extend from occiput to base of tail; dorsal scales slightly raised and keeled; pectoral and abdominal scales smooth and round, flat to concave, slightly larger than dorsal scales and not larger posteriorly; ventral scales of brachia raised, smooth and juxtaposed; four pores-bearing precloacal scales arranged in a chevron, separated; precloacal pore rounded; precloacal depression absent; femoral pores absent.Fore and hind limbs moderately long, slender; scales beneath forearm slightly raised, smooth and subimbricate; subtibial scales keeled; palmar scales smooth and juxtaposed; digits elongate, slender, inflected joint and bearing slightly recurved claws; subdigital lamellae unnotched; lamellae beneath first phalanges wide; lamellae beneath phalanx immediately following inflection granular; lamellae of distal phalanges wide; lamellae beneath inflection large; interdigital webbing absent; enlarge submetatarsal scales on 1 st toe absent; fingers increase in length from first to fourth with fourth and fifth nearly equal in length; relative length of fingers IV>V>III>II>I; toes increase in length from first to fifth with fourth and fifth nearly equal in length; relative length of toes IV>V>III>II>I; total number of subdigital lamellae on 4 th toe 28, 28 (right, left).Coloration in life (Figs 3, 4).Dorsal ground color of head light brown; top of the head bearing small black, sage and yellowish marking; snout yellowish; dorsal ground color of body, limbs and tail light brown with dark brown to black irregular blotches; ground color of ventral surfaces grayish white intermixed with light yellowish blotches; gular and throat regions are beige and light yellow; anterior gular region yellowish; midgular region with faint, dark lineate marking; thin, faint black postorbital stripe; light sage vertebral blotches extending from the nape to tail; flanks with irregular incomplete sage to yellowish blotches becoming smaller posteriorly; limbs yellowish brown with dark brown incomplete irregular spots subcaudal region yellowish; wide dark brown to black and yellow bands on tail.Coloration in preservative (Fig. 5).Color pattern similar to that in life with some fading.Dorsal ground color of head, body, limbs and tail brown with vertebral blotches indistinct; irregular pale marking; top of head with indistinct darker marking; all yellow markings faded to whitish gray; dorsal surfaces of limbs with irregular light and dark blotches; entire ventral surface whitish gray; gular region with faint dark lineate marking.
Variation.Most paratypes approximate the holotype in general aspects of color pattern (Fig. 6), with most differences found in the degree of vertebral blotches.ZMKU R 00761 (adult male) has dark spots in gular region.ZMKU R 00762 and ZMKU R 00765 (two adult males) have lighter gular markings than the holotype.THNHM 28201 and ZMKU R 00760 (two adult females) have lighter dorsal markings than the holotype.ZMKU R 00762 and THNHM 28205 (two adult males) have a pattern that resembles transverse bands rather than paravertebral blotches.ZMKU R 00762 and ZMKU R 00765 (two adult males) have regenerated tails of uniform  tan coloration.THNHM 28202 (adult female) and THNHM 28203 and THNHM 28205 (two adult males) have broken tails.THNHM 28205 (adult male) is an adult male with five continuous precloacal pores.Meristic and mensural variation within the type series are presented in Table 5.
Distribution and natural history.Cnemaspis tarutaoensis sp.nov. is known only from the type locality on Tarutao Island, approximately 40 km off the coast of Thailand.All specimens were found in karst forest near mangroves and karst outcrops near a stream (Fig. 7).Nine specimens (ZMKU R 00759-00760, ZMKU R 00762-00763, ZMKU R 00765-00766, and THNHM 28202-28204) were collected during the day (1100-1805 h) and five specimens (ZMKU R 00758, ZMKU R 00761, ZMKU R 00764, THNHM 28201 and THNHM 28205) were collected during the night (1920-2106 h).The male holotype was found during the day (1724 h) upside down on the interior surface of the karst formation.
Paratypes found during the day (ZMKU R 00759 and 00760, ZMKU R 00762 and 00763, ZMKU R 00765-00766, and THNHM 28202-28204) were in shaded areas, cracks, and crevices of rock boulders.When disturbed, some individuals would retreat into cracks and crevices, or hide in shaded areas of the rock boulder.Paratypes found at night (ZMKU R 00758, ZMKU R 00761, ZMKU R 00764, THNHM 28201 and THNHM 28205) were in deep crevices, within cracks on the shaded (by day) surfaces of boulders, or perched on vegetation near karst.Three gravid females (ZMKU R 00758, ZMKU R 00760, and THNHM 28202) contained two eggs during November 2017.THNHM 28204 (juvenile) was observed on vegetation near a rock boulder on 5 April 2018.At night, Cyrtodactylus cf.astrum was found in syntopy on rock boulders and karst formations with C. tarutaoensis sp.nov.
Etymology.The specific epithet refers to the type locality of the new species.
Fore and hind limbs moderately long, slender; scales beneath forearm slightly raised, smooth and subimbricate; subtibial scales keeled; palmar scales smooth and juxtaposed; digits elongate, slender, inflected joint and bearing slightly recurved claws; subdigital lamellae unnotched; lamellae beneath first phalanges wide; lamellae beneath phalanx immediately following inflection granular; lamellae of distal phalanges wide; lamellae beneath inflection large; interdigital webbing absent; enlarged submetatarsal scales on 1 st toe absent; fingers increase in length from first to fourth with fourth and fifth nearly equal in length; relative length of fingers IV>V>III>II>I; toes increase in length from first to fifth with fourth and fifth nearly equal in length; relative length of toes IV>V>III>II>I; total subdigital lamellae on 4 th toe 28, 28 (right, left).
Caudal and subcaudal scales keeled, similar to dorsal scale size; lateral caudal furrow present; enlarge caudal tubercles arranged in segmented whorls, not encircling tail; enlarge median subcaudal scales row absent; caudal tubercles present on lateral furrow; tail length (TL) 58.3 mm with regenerated tail; enlarge, flat, postcloacal tubercle 1, 1 (right, left) on lateral surface of hemipenial swellings at the base of tail.
Coloration in life (Figs 8,9).Dorsal ground color of head light brown, top of head bearing small, faint black and yellowish markings; thin, black postorbital stripes extending to nape; light-colored prescapular cresent; dorsal ground color of body, limbs and tail light brown with black irregular blotches; ground color of ventral surfaces grayish-white intermixed with yellowish blotches; ventral pattern sexually dimorphic, anterior gular region, abdominal region, and caudal region yellowish in males; two dark blotches on nape form a bipartite pattern; light sage vertebral blotches extending from the nape to tail; flanks with irregular incomplete brown to yellowish blotches becoming smaller posteriorly; tubercles on anterior and posterior of the body were white or yellow; widely separated, white or yellow tubercles occur on flanks; limbs beige with dark brown mottling; tail faintly marked with dark brown.
Coloration in preservative (Fig. 10).Color pattern similar to that in life with some fading of markings.Dorsal ground color of head, body, limbs and tail brown, darker with indistinct, irregular markings.All yellow coloration in gular region, ventral surfaces, flanks and tail faded to creamy white.
Variation.Most paratypes approximate the holotype in general aspects of morphology (Figs 11,12), with most differences found in the degree of vertebral blotches.All adult female paratypes lack yellowish coloration in the gular, abdominal, and caudal regions.ZMKU R 00767, THNHM 28208, THNHM 28210, and ZMKU R 00776 (four adult males) have regenerated tails of uniform tan coloration.THNHM 28207-28209, ZMKU R 00773, and ZMKU R 00775 (five adult males) have lighter dorsal markings that appear more as transverse bands than as paravertebral blotches.THNHM 28211 (one adult female) has a broken tail.Differences in meristic and morphometrics within the type series are presented in Table 7.
Distribution and natural history.Cnemaspis adangrawi sp.nov. is known only from Adang and Rawi islands, 60 and 61 km off the coast of Thailand, respectively (Fig. 1).All Adang specimens were found in a granitic rocky stream (Fig. 13A).Rawi Island specimens were found in rock outcrops along a stream (Fig. 13B) and along a forest stream near mangroves (Fig. 13C).Sixteen specimens (ZMKU R 00767-00768, ZMKU R 00770-00772, ZMKU R 00775-00776, THNHM 28206-28209, and THNHM 28211-28215) were collected during the day (1047-1823 h) and four specimens (ZMKU R 00769, ZMKU R 00773-00774, and THNHM 28210) were collected at night (1927-2024 h).The male holotype was found during the day (1047  EL 0.9 0.9 0.8 0.8 0.9 0.9 0.9 0.8 0.9 0.8 0.9 0.8 0.9 0.9 0.8 IN 0.9 0.9 0.9 0.9 0.9 0.9 0.9 0.9 h) on the base of a rock boulder with holes formed by the expansive soil between the ground and rock interface of a nearby stream.Paratypes found during the day (ZMKU R 00767-00768, ZMKU R 00770-00772, ZMKU R 00775-00776, THNHM 28206-28209, and THNHM 28211-28215) were in crevices of boulders, shaded areas with holes in the soil at the base of a rock wall near a stream, and on boulder outcrops near streams.When disturbed, some individuals would retreat into rock crevices or into holes in the soil at the base of a rock wall.Paratypes found at night (ZMKU R 00769, ZMKU R 00773-00774 and THNHM 28210) were in shaded areas (by day), deep at the base of boulders, or perched on vegetation near a rocky stream.Two gravid females (ZMKU R 00771 and THNHM 28211) contained one or two eggs during November 2017.Some juveniles (not collected) were found in holes in the soil and perched on vegetation near a stream at Rawi Island on 8 April 2018.At night, Cyrtodactylus macrotuberculatus was found in syntopy on the rock wall and vegetation near a stream at Jonsalad Waterfall, Adang Island, with Cnemaspis adangrawi sp.nov.
Etymology.The specific epithet refers to Adang and Rawi islands where the new species is found, and is a noun in apposition.

Discussion
Studies on the taxonomy and systematics of Cnemaspis in Southeast Asia have increased in the past two decades (Bauer and Das 1998;Das 2005;Bauer et al. 2007;Grismer and Chan 2010;Wood et al. 2013;Grismer et al. 2014;Iskandar et al. 2017;Wood et al. 2017).Integrative taxonomic approaches that incorporated both morphological and molecular data have been especially useful in uncovering cryptic diversity of Thai Cnemaspis (Grismer et al. 2014;Wood et al. 2017).Our descriptions of C. tarutaoensis sp.nov.and C. adangrawi sp.nov.bring the total number of Cnemaspis to 59 species, of which 18 occur in Thailand.Previously, the reported geographic distribution of Cnemaspis in Thailand was mostly restricted to the mainland (Smith 1925;Taylor 1963;Bauer and Das 1998;Grismer et al. 2010;Wood et al. 2017), with insular populations of Cnemaspis known only from four localities in Thailand, including Samui, Phangan and Ko Tao islands, Surat Thani Province (Gulf of Thailand; approximately 85 km offshore the mainland of Mueang Chumphon District, Chumphon Province; Grismer et al. 2010Grismer et al. , 2014) ) and Phuket Island, Phuket Province (Andaman Sea; approximately 30 km offshore the mainland of Takua Thung District, Phangnga Province; Das and Leong 2004).The descriptions of C. tarutaoensis sp.nov.and C. adangrawi sp.nov.double the number of Cnemaspis species known from islands in Thailand from two (C.chanardi and C. siamensis) to four.These two new species occur on Tarutao, Adang and Rawi islands in the Andaman Sea, offshore of the Thai mainland of Mueang Satun District, Satun Province (at approximately 40 km, 60 km, and 61 km, respectively).These islands were connected to the mainland during the last glacial maximum as recently as 21,000 years before present (Voris 2000;Sathiamurthy and Voris 2006), but the timing of their divergence from other Cnemaspis species remains untested.
The complex geological history in Thailand created a large number of limestone and granitic formations in southern Thailand (Day and Urich 2000;Morley et al. 2011).The karst regions and granitic rocky streams of southern Thailand are proving to harbor a high diversity of range-restricted species of geckos (Smith 1925;Taylor 1963;Grismer et al. 2010Grismer et al. , 2014;;Wood et al. 2017).Further research and additional field surveys in unexplored karst regions on islands and the mainland are needed to better understand the taxonomy, ecology, distribution, biogeography, and conservation of Cnemaspis in southern Thailand.

Figure 2 .
Figure 2. The single best maximum likelihood tree of the mitochondrial NADH dehydrogenase subunit 2 (ND2) gene and flanking tRNAs from geckos of the genera Cnemaspis, Cyrtodactylus and Hemidactylus, shown in full view (above) and close-up view of relevant clades (below).Support values at nodes are bootstrap values from a maximum likelihood analysis of the same dataset followed by posterior probabilities of the Bayesian analysis.
scales smooth, similar to dorsal scale size; lateral caudal furrow present; lateral caudal tubercle row absent; enlarge caudal tubercles at the base of tail not encircling tail; enlarged median subcaudal scales row present; tail length (TL) 34.3 mm with broken at tail tip; enlarged postcloacal tubercle 1, 1 (right, left) on lateral surface of hemipenial swellings at the base of tail.

Figure 7 .
Figure 7. Habitats of Cnemaspis tarutaoensis sp.nov.A Pha Toe Boo karst formation at type locality B habitat of paratypes in the exterior surface of karst cave at Tham Chorakae C habitat of paratypes in karst outcropped at Tarutao Island, Mueang Satun district, Satun Province, Thailand.

Figure 13 .
Figure 13.Habitats of Cnemaspis adangrawi sp.nov.A Jonsalad Waterfall at type locality of Adang Island B habitat of paratypes in outcropped near stream at Rawi Island C habitat of paratypes in forest stream near mangrove at Rawi Island, Mueang Satun district, Satun Province, Thailand.

Table 1 .
Samples used in this study, including catalogue numbers, Genbank accession numbers and localities of voucher specimens.Voucher abbreviations are as follows: Monte L. Bean Life Science Museum at Brigham Young University (BYU), California Academy of Sciences (CAS), the Field Museum of Natural History, Chicago, Illinois, USA (FMNH), La Sierra University Herpetological Collection (LSUHC), Universiti Sains Malaysia Herpetological Collection at the Universiti Sains Malaysia, Penang, Malaysia (USMHC), and Zoological Museum of Kasetsart University (ZMKU).

Table 2 .
Mean (minimum-maximum)uncorrected p-distances (%) within the Cnemaspis kumpoli group based on 1,296 bp of the mitochondrial ND2 gene and flanking tRNAs.Numbers in bold are within species divergence.n = number of individuals.

Table 3 .
Mean (minimum-maximum) uncorrected p-distances (%) within the Cnemaspsis siamensis group based on 1,296 bp of the mitochondrial ND2 gene and flanking tRNAs.Numbers in bold are within species divergence.n = number of individuals.

Table 4 .
Meristic character states and color patterns of species in the Cnemaspis kumpoli group.Measurements are taken in millimeters and measurement abbreviations are defined in the text.var = character variable; -= data unavailable, ant = anterior.
Cnemaspis tarutaoensis sp.nov. is further distinguished from C. kumpoli, C. monachorum and C. niyomwanae by having yellow coloration in the subcaudal region and wide black

Table 6 .
Meristic character state and color pattern of species in the Cnemapsis siamensis group.Measurements are taken in millimeters and measurement abbreviations are defined in the text.-= data unavailable, w = weak.