A new species of the genus Tylototriton (Urodela, Salamandridae) from western Thailand

Abstract We describe a new species of the newt genus Tylototriton from Umphang Wildlife Sanctuary, Tak Province, western Thailand based on molecular and morphological evidence and named here as Tylototritonumphangensissp. nov. The new species is assigned to the subgenus Tylototriton and differs from other species in having dark-brown to blackish-brown body and limbs, truncate snout, prominent antero-medial ends of the expansion of the dentary bones, laterally protruding quadrate regions, indistinct and small rib nodules, a well-segmented vertebral ridge, and rough dorsolateral bony ridges, which are steeper anterior, and curved medially at the posterior ends. The molecular data show that Tylototritonumphangensissp. nov. differs from T.uyenoi sensu stricto by a 5% genetic sequence divergence of the mitochondrial NADH dehydrogenase subunit 2 region gene. The new species and T.uyenoi are both endemic to Thailand, distributed along the Northwest Thai (Dawna) Uplands of Indochina. To clarify the species boundary between Tylototritonumphangensissp. nov. and T.uyenoi, additional field research is needed in adjacent areas. Tylototritonumphangensissp. nov. is restricted to evergreen hill forests in Umphang Wildlife Sanctuary. We suggest that the new species should be classified as Endangered (EN) in the IUCN Red List.


Introduction
The salamandrid genus Tylototriton Anderson, 1871, commonly known as crocodile newts, includes 32 nominal species ranging across eastern Himalaya, eastern Nepal, northern India, Bhutan, Myanmar, central to southern China (including Hainan island), and southwards through Laos to Thailand and Vietnam (Hernandez 2016;Bernardes et al. 2020;Pomchote et al. 2020aPomchote et al. , 2020bPoyarkov et al. 2021a). Outside of breeding season, all known species are terrestrial and micro-endemic, regarded mostly as niche specialists that are generally found at middle to high elevations in subtropical, moist, forested environments. These provide a relatively narrow thermal range (15−24.0 °C) with a high annual rainfall, especially during the monsoon season, which ensures favorable breeding conditions and survival of Tylototriton species (Hernandez et al. 2018).
The genus is subdivided into three subgenera, Tylototriton, Yaotriton, and Liangshantriton (e.g., Dubois and Raffaëlli 2009;Fei et al. 2012;Nishikawa et al. 2013aNishikawa et al. , 2013bPhimmachak et al. 2015;Wang et al. 2018;Poyarkov et al. 2021a) and includes several, as yet, unnamed taxa, which contain cryptic species that are morphologically difficult to distinguish (Hernandez 2016;Poyarkov et al. 2021a). Recent studies have provided a better understanding of the ecology, biology, taxonomy, phylogenetic relationships, and conservation of these endangered species that have been highly harvested in recent years throughout South-east Asia (Phimmachak et al. 2015;Hernandez et al. 2018;Wang et al. 2018;Bernardes et al. 2020;Pomchote et al. 2020aPomchote et al. , 2020bPoyarkov et al. 2021a). Several recent phylogenetic studies have also revealed the presence of undescribed cryptic lineages, which actually might represent independent species, in South-east Asia, especially in the Indochina region (Wang et al. 2018;Zaw et al. 2019;Bernardes et al. 2020;Pomchote et al. 2020aPomchote et al. , 2020bPoyarkov et al. 2021a).
To our knowledge, Thailand contains five Tylototriton species (Nishikawa et al. 2013a;Le et al. 2015;Pomchote et al. 2020aPomchote et al. , 2020b. They are distributed allopatrically in high mountainous areas at altitudes above 1,000 m mean sea level throughout the northern (T. verrucosus, T. uyenoi, T. anguliceps, and T. phukhaensis), northeastern (T. panhai), and western (T. uyenoi) regions (Hernandez and Pomchote 2020a, 2020cPomchote et al. 2020aPomchote et al. , 2020b. Recent field surveys recorded several new Tylototriton populations distributed in the western region of Thailand, where the southernmost record in Asia of the genus was recorded (Hernandez and Pomchote 2020c). These populations were previously identified as T. uyenoi according to their morphological appearance and distribution, and they range from the Daen Lao and Thanon Thong Chai Ranges, southwards to the Dawna Range (Hernandez 2016(Hernandez , 2017Hernandez et al. 2019).
However, according to Hernandez (2016), these newly found T. uyenoi populations show different phenotypes and an allopatric distribution in scattered and sepa-rated mountainous areas, resulting in the recent description of T. phukhaensis by Pomchote et al. (2020b). Moreover, the limited number of specimens examined in previous studies (Hernandez 2016(Hernandez , 2017Hernandez et al. 2019;Hernandez and Pomchote 2020c), lack any detailed morphological examination and molecular analysis leading to the question of the taxonomic status of these populations.
As polymorphic species provide an opportunity to examine the role of isolation in populations that may contribute to the process of divergence, we assessed the western populations of Tylototriton species in the Umphang Wildlife Sanctuary (UPWS), Tak Province, which is located through the Dawna Range in western Thailand. This divergent population was discovered several years ago (Hernandez 2016;Hernandez et al. 2019). The molecular and morphological evidence indicate that the crocodile newt specimens from UPWS belong to a lineage distinct from the known Tylototriton species. As a consequence, we describe the UPWS newts as a new species, Tylototriton umphangensis sp. nov., and discuss its taxonomic relationships, distribution, and implications for conservation.

Sampling
The field survey was performed on 19 June 2021 at UPWS, Tak Province, western Thailand ( Fig. 1) using the visual encounter survey method (Heyer et al. 1994). Four specimens were found in a small pond that drains into a stream. The pond was surrounded by hill evergreen forest and located on the mountain at an elevation of approximately 1,150 m a.m.s.l. (above mean sea level). Specimens of Tylototriton were caught by hand and kept in plastic boxes for examination. The biological and physical parameters of their habitats were recorded.
All four newts were checked for sex and maturity using the cloacal characters (Pomchote et al. 2008) and were subsequently sexed as breeding males. All specimens were used for molecular and morphological analyses.
Live specimens were anesthetized by immersion in a solution of tricaine methane sulfonate (MS-222; 5 g/L) for about 5 min , euthanized by a solution of chloretone (Heyer et al. 1994), and then measured for morphometrics and body weight (BW), as detailed below. The tissue samples (liver) of each individual were taken, and then stored in 95% (v/v) ethanol for molecular study prior to preservation. The voucher specimens were subsequently preserved in 70% (v/v) ethanol and deposited at the Chulalongkorn University Museum of Natural History (CUMZ).

Molecular analyses
Total DNA was extracted from the liver using a PureDireX TM genomic isolation kit (Bio-Helix, Taiwan). The mitochondrial NADH dehydrogenase 2 gene (ND2) was amplified using the polymerase chain reaction (PCR) with the SL-1 (5'-ATAGAG-GTTCAAACCCTCTC-3') and the SL-2 (5'-TTAAAGTGTCTGGGTTGC ATTCA G-3') primers (Wang et al. 2018). Each PCR reaction consisted of 15 µL of OnePCR TM Ultra (GeneDirex, Taiwan), which is a premixed solution, 1.5 µL of each primer (10 µM), 9 µL of UltraPure TM DNase/RNase-Free distilled water (Invitrogen, USA), and 3 µL of DNA template. The thermal cycling was performed at 94 °C for 4 min, followed by 35 cycles of 94 °C for 30 s, 55 °C for 1 min, and 72 °C for 90 s (Wang et al. 2018). The PCR products were checked by agarose gel electrophoresis to confirm their size and estimate the concentration. The desired PCR products were purified and commercially sequenced by Bioneer Inc. in South Korea.
We combined the four new sequences of the UPWS samples obtained in this study with those of the other related species available from GenBank ( Table 1). The optimum substitution models were selected using Kakusan 4 (Tanabe 2011). We then constructed phylogenetic trees by Bayesian inference (BI) and maximum likelihood (ML) analyses using MrBayes v. 3.1.2 (Huelsenbeck and Ronquist 2001) and RAxML v. 8 (Stamatakis 2014), respectively. The criterion used for model selection was AIC, with the codon-equal-rate model with the general time reversible model (GTR) + Gamma (G) being selected for ML and the codon-proportional model with the Hasegawa-Kishino-Yano-1985 (HKY85) model + G for each codon position for the BI. The BI analysis was performed as two independent runs of four Markov chains for 10 million generations, sampling one tree every 100 generations and calculating a consensus topology for 70,000 trees after discarding the first 30,001 trees (burnin = 3,000,000). For the BI, we considered posterior probabilities (bpp) of 95% or greater as significant support (Leaché and Reeder 2002). The robustness of the ML tree was tested using bootstrap analysis (Felsenstein 1985) with 2,000 replicates, and we accepted tree topologies with bootstrap values (bs) of ≥ 70% to be significantly supported (Huelsenbeck and Hillis 1993). Pairwise comparisons of uncorrected sequence divergences (p-distance by 1,013 base pairs; bp) were calculated using MEGA v. 7 (Kumar et al. 2016).

Morphological examination
The morphometric characters of the UPWS newts were compared with those of T. uyenoi because their appearances and color pattern are rather similar; moreover, previous studies identified the UPWS newts as T. uyenoi (Hernandez et al. 2019). Note that the other four Tylototriton species from Thailand (T. verrucosus, T. anguliceps, T. phukhaensis, and T. panhai) were not included in this morphometric study for two reasons. Firstly, the external morphology of T. verrucosus, T. anguliceps, and T. phukhaensis was clearly different from that of T. uyenoi (see Pomchote et al. 2020aPomchote et al. , 2020b, although morphological comparisons using the published literature were made (see comparisons below). Secondly, T. panhai has different color pattern from that of the other Thai Tylototriton species; moreover, T. panhai is a member of another lineage, the subgenus Yaotriton (Nishikawa et al. 2013a). We compared the reported morphometrics of a total of 12 specimens, the four Tylototriton sp. from UPWS (four males: CUMZ-A-8243 to -8246) of this study, and eight specimens of T. uyenoi obtained previously from the same localities of the holotype and paratypes of T. uyenoi (Nishikawa et al. 2013a). The specimens of T. uyenoi were loaned from the Natural History Museum, National Science Museum, Thailand (THNHM): topotypic specimens THNHM 10319-10320, 20170 (three males) from Doi (= Mountain in Thai language) Suthep-Doi Pui National Park (NP), Chiang Mai Province; and THNHM 13866, 13868, 13870-13871 (four males), and THNHM 13869 (one female) from Doi Inthanon NP, Chiang Mai Province, which is the same locality as the paratypes.

Statistical analysis
We compared the SVL, BW, and the other 25 ratio values to SVL (presented as % SVL) between Tylototriton sp. from UPWS and the other T. uyenoi specimens. Differences in morphological characters between the Tylototriton sp. from UPWS and T. uyenoi were analyzed by the Mann-Whitney U test. The relationships of all morphometric characters were examined using principal component analysis (PCA). Note that the vent length of the one T. uyenoi female (THNHM 13869) was excluded from the morphological comparison because this parameter is much longer in males than in females [RVL 7.4 vs 1.7 and 1.9; 9.3 vs 4.0 in T. uyenoi, data from Nishikawa et al. (2014) and the present study, respectively]. All statistical analyses were performed using the SPSS v. 22 for Windows. Statistical significance was accepted at the p < 0.05 level.

Molecular analyses
We obtained 452-1,039 bp sequences of the partial ND2 region for 20 specimens, including the outgroup (Table 1). The sequences of the four specimens from UPWS (this study) were the same, and of the 1,039 nucleotide sites, 340 were variable and 158 were parsimony informative within the ingroup (sequence statistics available upon request from the senior author). The mean likelihood score of the BI analyses for all trees sampled at stationary was −4033.667. The likelihood value of the ML tree was −3955.266.
Phylogenetic analyses employing the BI and ML criteria yielded nearly identical topologies and so we present only the BI tree in Figure 2. Monophyly of the subgenus Tylototriton (all samples except Samples 9 and 20) was fully supported in the BI and ML trees (bpp = 98% and bs = 96%). Within the subgenus, T. taliangensis was first separated from the remaining lineages. The latter group was further divided into two clades: one including T. shanorum, T. himalayanus, and T. kachinorum; the other included the remaining lineages. The newts from UPWS (same sequence) were nested in the latter clade and was first clustered with T. uyenoi with significant support.
The p-distances between each pair of a total 17 haplotypes recognized above ranged from 1.4% (between T. verrucosus and T. shanjing) to 18.8% (between Echinotriton and T. uyenoi and between Echinotriton and T. kachinorum) ( Table 2). The distance between the newts from UPWS and its sister species T. uyenoi was 5.0%, which was larger than the 24 heterospecific combinations in this study. Bayesian inference tree based on the partial ND2 gene for the samples examined. Numbers above branches represent the bpp/bs, and asterisks indicate nodes with bpp ≥ 0.95 and bs ≥ 70%. Numbers at branches tips are the sample numbers, as shown in Table 1. Scale bar = 0.04 substitutions/site.

Table 2.
Genetic uncorrected p-distance (%) of the ND2 region between samples examined in this study.
In life, the dorsal ground color was dark brown to black (Nishikawa et al. 2013a;Pomchote et al. 2020a), while in preservative, the background coloration of T. uyenoi was light brown to cream (Nishikawa et al. 2013a) or light brown to brown (this study), although the UPWS samples tended to be the blackish both in life and in preservative. In T. uyenoi, the dorsal and ventral head, parotoids, vertebral ridge, rib nodules, limbs, vent, and tail were orange to reddish-brown in life (Nishikawa et al. 2013a;Pomchote et al. 2020a) and light brown to orange-brown (Nishikawa et al. 2013a) or cream to orange-brown (this study) in preservative, although the UPWS samples had darker markings than T. uyenoi, both in life and in preservative (Figs 3-5).
The UPWS samples and T. uyenoi also showed a few similar morphological characteristics. For example, the sagittal ridge on head and the parotoids were distinct and projected posteriorly. However, morphological differences between the UPWS population and T. uyenoi were also present (Fig. 4).
The snout of the UPWS population was truncate, while that of T. uyenoi was almost rounded to blunt, except for two specimens (THNHM 10319 and 10320) that were relatively truncate. In lateral view, the degree that the snout projects beyond the lower jaw was more distinct in T. uyenoi than in the UPWS population that hardly projected beyond the lower jaw.
The dorsolateral bony ridges of the UPWS population were rough, steeper anteriorly, and curved medially at the posterior ends, while those of T. uyenoi were rough, especially from above the eye to above the anterior end of the parotoid, less steep anteriorly, and weakly or rather curved medially at the posterior ends.
In lateral view, the parotoids of the UPWS population were oriented rather parallel to the body axis and the posterior part curved upwards, while those of T. uyenoi were oriented obliquely downwards or rather parallel relative to the body axis.
In dorsal view, the quadrate regions of the UPWS population protruded laterally, while those of T. uyenoi were weakly curving.
In the urodeles, the dentaries are elongated, paired bones that curve medially. The left and right dentaries touch each other antero-medially on the lower jaw. At the antero-medial ends, some expansions are developed posteriorly in the dorsal view, while in the anterior view this expansion slightly develops in the ventral direction (e.g., Villa et al. 2014;Parra Olea et al. 2020;Ponssa and Abdala 2020). This expansion is prominently present in the UPWS population (Fig. A4), both in life and in preservative, whereas it was absent in T. uyenoi (Figs B4-C4).
The vertebral ridge of the UPWS population and T. uyenoi was segmented from the anterior end to the tail base but was less segmented in T. uyenoi than in the UPWS Table 3.

Morphometric comparisons of the examined specimens of Tylototriton
[mean ± SD of SVL (in mm), mean ± SD of BW (g), and median of ratios of characters (R: % SVL), with range in parentheses]. Character abbreviations refer to the text. samples, especially the two paratypes CUMZ-A-8245 and 8246, which were prominently segmented.
The rib nodules of T. uyenoi were isolated, rounded, distinct but small, and forming knob-like warts, with the number of rib nodules ranging from 12-15, but almost all the specimens had 14 warts on each side of body. In contrast, the rib nodules of the UPWS newts were indistinct and small in shape, ranging from 14-15 warts.
The overall morphological differences were examined using PCA for the UPWS population and T. uyenoi. The first two principal components (PCs) explained 49.0% of the total variation. The two-dimensional plots of PC1 vs PC2 showed that the UPWS population was clustered together and separated from T. uyenoi (Fig. 6).
Based on the molecular and morphological evidence, the Tylototriton sp. from UPWS, Tak Province, western Thailand is confirmed as an undescribed species. Therefore, we describe it as a new species, Tylototriton umphangensis sp. nov. Etymology. The specific epithet umphangensis refers to Umphang Wildlife Sanctuary, the type locality of the new species.

Systematics
Diagnosis. The new species is placed in the genus Tylototriton by having a combination of dorsal granules present, dorsolateral bony ridges on head present, knob-like warts (rib nodules) on dorsolateral body present, and quadrate spine absent. Tylototriton umphangensis sp. nov. differs from its congeners by having the following morphological characters: medium-sized, adult SVL 65.6-75.3 mm in males; skin rough with fine granules; snout truncate; quadrate regions laterally protruding; antero-medial ends of dentaries distinctly expanded; dorsolateral bony ridges on head prominent, steep, rough, narrow, and posterior ends curved medially; parotoids distinct, oriented rather parallel to the body axis and posterior part curved upwards in the lateral view; vertebral ridge distinct and segmented; rib nodules 14-15, small, and indistinct; limbs long and thin; tips of forelimbs and hindlimbs overlapping when adpressed along the body; tail thin.
Description of holotype. Body rather slim and long (RTRL 76.6%); skin rough; fine granules dense on dorsum, dense on sides of body and tail, and arranged in transverse striations on mid-ventrum; head longer than wide (HW/HL 0.97), hexagonal in shape, depressed, and slightly oblique in profile; snout truncate, hardly projecting beyond lower jaw; nostrils close to snout tip, not visible from dorsal view; quadrate regions protruding laterally from dorsal view; antero-medial ends of dentaries distinctly expanded; dorsolateral bony ridges on head narrow, rough, and posterior ends curved proximally; sagittal ridge on head short and weak; labial fold absent; tongue oval, attached to anterior floor of mouth, free laterally and posteriorly; vomerine tooth series in an inverted V-shape, converging anteriorly and reaching choanae; parotoids distinct, projecting posteriorly, posterior ends slightly curved medially, oriented rather parallel to body axis and curved upwards in lateral view; gular fold present; costal folds absent; vertebral ridge prominent, narrow, and slightly segmented from neck to groin, separated from sagittal ridge on head; two low and flat bony ridges on the dorsal head surface forming a "V" shape, connected with the anterior end of vertebral ridge; rib nodules small, indistinct, forming knob-like warts, 15 on each side of body from axilla to base  of tail; rib nodules slightly increasing in size from most anterior to forth nodule, then decreasing posteriorly; forelimbs (34.2% SVL) shorter than hindlimbs (40.0% SVL); tips of forelimb and hindlimb overlapping when adpressed along body; fingers and toes well developed, free of webbing; fingers four, comparative finger lengths 2 > 3 > 1 > 4; toes five, comparative toe lengths 4 > 3 > 2 > 5 > 1; tail laterally compressed, dorsal fin more distinct posteriorly, ventral edge smooth, tip pointed; tail short (91.9% SVL); cloaca slightly swollen; vent slit longitudinal.
Color of holotype. In life, dorsal ground coloration is dark-brown to blackishbrown, while the ventral color is slightly lighter than dorsum. Dorsal, ventral, and lateral of head, parotoids, vertebral ridge, rib nodules, limbs, vent region, and whole tail are orange-brown. Tip of tail is slightly lighter than dorsal and lateral sides of tail. Ven-tral side of head, part of pectoral and pubic region, limbs, and tail prominently lighter than dorsum. The lightest is the ventral edge of the tail. The lighter region between the ventral edge of the tail and the area of the vent is connected. Color of digit tips is dark brown. After a week in preservation, the color pattern is rather similar to that in life.  MTAH 8.7;FLL 24.9;HLL 29.1;2FL 5.3;3FL 4.1;3TL 6.5;and 5TL 3.5. Variation. Some differences in morphology were observed among the four specimens. The dorsolateral bony ridges on the head of one paratype (CUMZ-A-8245) are rougher than the holotype and the other paratypes. The sagittal ridge on the head of one paratype (CUMZ-A-8244) is smaller and weaker than the holotype and the other paratypes. Two paratypes (CUMZ-A-8245 and CUMZ-A-8246) have a more distinctly segmented vertebral ridge than the holotype and the other paratype. The holotype has much more distinct rib nodules than the three paratypes. Sizes of rib nodules varied from rounded anteriorly to irregularly shaped posteriorly among the paratypes. One paratype (CUMZ-A-8246) has an undulated surface of the dorsal fin, while the other type specimens have an even-surfaced dorsal fin. Type specimens are  generally similar in color pattern, but the coloration of the dorsal, ventral, and lateral head, parotoids, vertebral ridge, rib nodules, limbs, and whole tail is much lighter in the holotype than the three paratypes. The color of the digit tips of the holotype is dark brown, but those of the paratypes are black. Morphological variations between the specimens are shown in Figure 5.

Measurement of holotype (in mm
Comparisons. Tylototriton umphangensis sp. nov. differs from the other species of subgenus Tylototriton as follows: from T. taliangensis by having orange-brown markings on the head, trunk, limbs, and tail (vs uniformly black body except for distal fingers, toes, and posterior parotoids in T. taliangensis); from T. kweichowensis and T. pseudoverrucosus by having separated rib nodules (vs connected orange markings forming continuous dorsolateral lines in T. kweichowensis and T. pseudoverrucosus); from T. shanorum and T. anguliceps by having a sagittal ridge and rather steep dorsolateral bony ridges on the head (vs no sagittal ridge and rather flat dorsolateral bony ridges on head in T. shanorum, and prominent sagittal ridge and the posterior ends of dorsolateral bony ridges distinctly curved medially in T. anguliceps); from T. ngarsuensis by having truncate snout in dorsal view (vs rounded in T. ngarsuensis); from T. himalayanus by lacking grooves on either side at the basal tail (vs present in T. himalayanus); from T. yangi by having uniformly orange-brown parotoids (vs black coloration except for posterior end of parotoids with orange coloration in T. yangi); from T. kachinorum, T. pulcherrimus, and T. shanjing by having light orange-brown on part of pubic region (vs light yellowish-grey ventral surfaces in T. kachinorum, and yellowish-orange to bright yellow ventral trunk in T. pulcherrimus and T. shanjing); from T. verrucosus by having rough dorsolateral bony ridges (vs smooth in T. verrucosus); from T. podichthys and T. phukhaensis by having short and weak sagittal ridge on the head (vs indistinct sagittal ridge on head in T. podichthys, and narrow, long, and prominent sagittal ridge on head in T. phukhaensis); from T. panwaensis by having narrow vertebral ridge (vs wide in T. panwaensis).
Distribution. Umphang Wildlife Sanctuary, Tak Province, western Thailand (Fig.  1). The Umphang Wildlife Sanctuary is located along the Dawna Range, which is a mountain range in eastern Myanmar and northwestern Thailand. Thus, this species is expected to also occur in Myanmar and elsewhere in western Thailand.
Natural history. All specimens were found during the afternoon at around 14:30 h hidden under leaf litter and between stems of arrowroot plants (family Marantaceae) in a small ephemeral pond (Fig. 7) during the rainy season, which is the breeding season of Tylototriton species. The pond had clear water and the bottom was covered with dense leaf litter. The surrounding area was composed of hill evergreen forest. The pond size was approximately 520 cm long, 270 cm wide, and 17 cm in maximum depth. The water temperature was 23.1 °C. The water quality parameters were: pH 6.4; dissolved oxygen 4.13 mg/L; conductivity 23 µS/cm; total dissolved solid 15 mg/L; and turbidity 7.6 NTU. No fish were observed.

Discussion
Molecular and morphological evidence indicate that the newts found at UPWS, Tak Province, western Thailand are a distinct, new species described here. With our description of this new species, the number of Tylototriton species is now 33, with six of them present in Thailand: T. verrucosus, T. uyenoi, T. panhai, T. anguliceps, T. phukhaensis, and T. umphangensis sp. nov. Three of these six species are endemic to Thailand (T. uyenoi, T. phukhaensis, and T. umphangensis sp. nov.). Thus, Thailand has the third highest number of species of Tylototriton and the second highest in Indochina; the highest number of species is in China (17) followed by Vietnam (7) (Frost 2021).
Tylototriton umphangensis sp. nov. has been confused with T. uyenoi because of these species have morphological similarities and a similar distribution (Hernandez et al. 2019). In the present study, we compared T. umphangensis sp. nov. with T. uyenoi, the latter having been described and named based on a holotype and eight paratypes (total of nine adult males) collected from Phuping Rajanives Palace, Doi Suthep and the Royal Garden Siribhume, Doi Inthanon, Chiang Mai Province, respectively (Nishikawa et al. 2013a). According to Nishikawa et al. (2013a), T. uyenoi specimens are basically similar in morphology and color pattern, but show variations in the degree of segmented ver-tebral ridge, size of rib nodules, texture of dorsolateral bony ridges, and color markings. These morphological variations are also present in the specimens of T. umphangensis sp. nov. that we examined in our study. However, T. umphangensis sp. nov. can be distinguished from the most closely related species (T. uyenoi) and other congeners.
Geographic isolation may limit gene flow and promote genetic differentiation among populations which can result in the formation of new species (Eckert et al. 2008;Qian et al. 2017). Tak Province is located in the Northwest Thai (Dawna) Uplands of Indochina (Poyarkov et al. 2021b), which consists of several high-mountain areas in the three major mountain ranges: (i) Thanon Thong Chai and (ii) Daen Lao Ranges in northernmost Tak (Hernandez and Pomchote 2020c) and Umphang, Tak Province (Hernandez et al. 2019; this study) (Fig. 1). Hernandez et al. (2018) stated that Tylototriton species are niche specialists because they reside at high elevations with moist, and cool conditions, a narrow thermal range, and high rainfall during the breeding season. This is consistent with previous studies that Thai Tylototriton species are distributed in high mountainous areas at an altitude of more than 1,000 m a.m.s.l. (Pomchote et al. 2008. Following to previous studies (Pomchote et al. 2008, we defined lowland and highland areas according to the distribution of six Tylototriton species in Thailand: lowlands are areas below 1,000 m a.m.s.l. and uplands are areas above 1,000 m a.m.s.l. Thus, the lowland areas, located between each highland area, of the Northwest Thai Uplands may serve as a barrier restricting the gene flow between Tylototriton populations. Consequently, further morphological and molecular analyses, as well as field surveys in Tak Province and its nearby areas located along Thanon Thong Chai, Daen Lao, and Dawna Ranges, need to be done to clarify the species boundary between T. umphangensis sp. nov. and T. uyenoi. According to previous data (Watchara Sanguansombat and Chattraphas Pongcharoen, personal communication) and the check list of fauna diversity of UPWS (Department of National Parks, Wildlife and Plant Conservation), T. verrucosus (now named T. umphangensis sp. nov.) were first found near an artificial pond adjacent to a deserted hut near a road that was about 6 km from the Mae Klong Khi Forest Ranger Station. Our field survey was conducted there on 18 June 2021 at night, but we did not find any newts. Not only did this pond have fish (released by someone?), but this area is also under construction. Moreover, there are cattle that belong to local people roaming freely in UPWS that may cause damage to the forest, including breeding sites of the newts, as previously reported in other NPs, such as in Phu Suan Sai NP, Loei Prov-ince that harbors T. panhai (Hernandez and Pomchote 2020a); and in Doi Phu Kha NP, Nan Province that harbors T. phukhaensis ). Thus, effects from anthropogenic activities, including cattle, should be evaluated in detail. Tylototriton umphangensis sp. nov. is currently only known from the hill evergreen forests of UPWS. We suggest that the new species should be classified as Endangered (EN) in the IUCN Red List and that it needs further conservation management.