A new Cyrtodactylus (Squamata: Gekkonidae) from Binh Thuan Province, southern Vietnam

We describe a new species of Cyrtodactylus on the basis of two specimens collected from Ta Kou Nature Reserve, Binh Thuan Province, southern Vietnam. Cyrtodactylus chungi sp. nov. is distinguished from the remaining Indochinese bent-toed geckos by a combination of the following characters: relatively small body size (SVL up to 68.5 mm); a continuous neckband; 5 or 6 irregular transverse dorsal bands; 11 or 12 bands on original tail; keeled tubercles present on dorsum, posterior limbs and tail; 17 or 18 irregular dorsal tubercle rows; 30 or 31 ventral scale rows; ventrolateral skin European Journal of Taxonomy 731: 47–70 (2021) 48 folds indistinct; an angular series of seven precloacal pores in male and six pitted, enlarged precloacal scales in female, each series separated by a diastema of undifferentiated scales from 4–6 enlarged, poreless femoral scales; median subcaudals slightly enlarged; 17–20 subdigital lamellae under the fourth toe. Based on molecular analyses of the fragment of mitochondrial gene cytochrome c oxidase subunit I (COI), the new species is recovered as the sister taxon to Cyrtodactylus cattienensis s. str. with a genetic divergence of more than 9%. In phylogenetic analyses, the new species is recovered as a member of the Cyrtodactylus irregularis species group.


Introduction
The gekkonid genus Cyrtodactylus Gray, 1827 is by far the most speciose with many new species being described every year (Grismer et al. 2018;Murdoch et al. 2019). A total of 305 species have been recognized to date (Uetz et al. 2020). Vietnam has been a hotspot of new discoveries with 43 recorded species (Nazarov et al. 2012;Luu et al. 2017;Ostrowski et al. 2020;Uetz et al. 2020). Gymnodactylus peguensis var. irregularis was described by Smith (1921a) from an open pine forest at Camly in the Langbian Plateau, Vietnam. Nazarov et al. (2008) suggested that this and all other allied species belonged to a species complex (Nazarov et al. 2012;Pauwels et al. 2018). The C. irregularis complex has recently been the subject of numerous morphological and molecular studies. So far, 20 species have been described along with several additional populations whose taxonomic status is unresolved or still unknown (S.N. Pauwels et al. 2018;Ostrowski et al. 2020;Uetz et al. 2020). Thus, Neang et al. (2020) described C. phnomchiensis, the only member of the C. irregularis complex, which occurs west of the Mekong River in Kampong Thom Province, Cambodia.
S.N.  investigated the effectiveness of DNA barcoding based on the COI gene using 164 samples of Vietnamese bent-toed geckos within the C. irregularis complex. They discovered eleven unnamed lineages from different localities, including the population analyzed in this study. Based on an integrative taxonomic approach, viz. combining morphological and molecular data, we herein clarify the status of the Cyrtodactylus population from Ta Kou, Binh Thuan Province.

Sampling
Field surveys were conducted in Ta Kou Nature Reserve, Ham Thuan Nam District, Binh Thuan Province, southern Vietnam in June 2017 (Fig. 1). Specimens were euthanized in a closed vessel with a piece of cotton wool containing ethyl acetate (Simmons 2002), fi xed in 80% ethanol and subsequently stored in 70% ethanol. Specimens were deposited in the collections of the Institute of Ecology and Biological Resources (IEBR), Hanoi, Vietnam.

Molecular data and phylogenetic analyses
All taxa of the Cyrtodactylus irregularis species complex were included in our analyses. Available sequences of the species were obtained from GenBank. Two new samples from a population in Ta Kou Mountain, Binh Thuan Province of southern Vietnam (IEBR 4581 and 4582) were included in the analysis. Two species, C. spelaeus Nazarov, Poyakov, Orlov, Nguyen, Milto, Martynov (2021) 50 Chulisov, 2014 andC. wayakonei Nguyen, Kingsada, Rösler, Auer &Ziegler, 2010, were used as an outgroup based on the results from Luu et al. (2016).
DNA was extracted using a DNeasy Blood and Tissue kit (Qiagen, Germany) following the manufacturer's instruction. Extracted DNA was amplifi ed by HotStar Taq PCR mastermix (Qiagen, Germany) with 21 μl volume (10 μl of mastermix, 5 μl of water, 2 μl of each primer at 10 pmol/ml and 2 μl of DNA). PCR condition was: 95°C for 15 minutes to active the taq; with 40 cycles at 95°C for 30 s, 45°C for 45 s, 72°C for 60 s; and the fi nal extension at 72°C for 6 minutes. A fragment of the mitochondrial gene, cytochrome c oxidase subunit I (COI), was amplifi ed using the primer pair VF1-d (5'-TTCTCAACCAACCACAARGAYATYGG-3') and VR1-d (5'-TAGACTTCTGGGTGGCCRAARAAYCA-3)' (Ivanova et al. 2006). PCR products were visualized using electrophoresis through a 2% low melting-point agarose gel stained with ethidium bromide. Successful amplifi cations were purifi ed to eliminate PCR components using GeneJET TM PCR Purifi cation kit (ThermoFisher Scientifi c, Lithuania). Purifi ed PCR products were sent to FirstBase (Malaysia) for sequencing.
After sequences were aligned by Clustal X ver. 2 (Thompson et al. 1997), data were analyzed using maximum parsimony (MP) as implemented in PAUP*4.0b10 (Swofford 2002), maximum likelihood (ML) in IQ-TREE ver. 1.6.7.1 (Nguyen et al. 2015), and Bayesian inference (BI) in MrBayes ver. 3.2 (Ronquist et al. 2012). For MP analysis, heuristic analysis was conducted with 100 random taxon addition replicates using tree-bisection and reconnection (TBR) branch swapping algorithm, with no upper limit set for the maximum number of trees saved. Bootstrap support (BP) was calculated using 1000 pseudoreplicates and 100 random taxon addition replicates. BP ≥ 70 was considered strong support for the clade (Hillis & Bull 1993). All characters were equally weighted and unordered. For the maximum likelihood (ML) analysis, the analysis was run with a single model and 10 000 ultrafast bootstrap replications using the command iqtree -s matrix.nex -m TIM+I+G -bb 10000. Ultrafast bootstrap ≥ 95% was regarded as signifi cant (Nguyen et al. 2015) The optimal model for nucleotide evolution was determined using Modeltest ver. 3.7 (Posada & Crandall 1998).
For BI analysis, we used the optimal model on an unpartitioned data set determined by Modeltest with parameters estimated by MrBayes ver. 3.2.1. Two independent analyses with four Markov chains (one cold and three heated) were run simultaneously for 10 million generations with a random starting tree and sampled every 1000 generations. Log-likelihood scores of sample points were plotted against generation time to determine stationarity of Markov chains. Trees generated before log-likelihood scores reached stationarity were discarded from the fi nal analyses using the burn-in function. The posterior probability (PP) values for all clades in the fi nal majority rule consensus tree were provided. PP ≥ 95% were deemed strong support for a clade (Ronquist et al. 2012). The optimal model for nucleotide evolution was set to TIM+I+G for ML and combined Bayesian analyses as selected by Modeltest ver. 3.7. The cutoff point for the burn-in function was set to 30 in the Bayesian analysis, as -lnL scores reached stationarity after 30 000 generations in both runs. Uncorrected pairwise divergences were calculated in PAUP*4.0b10.

Morphological characters
Measurements were taken with a digital calliper to the nearest 0.1 mm.

AG
= trunk length or axilla-groin distance, from posterior edge of forelimb insertion to anterior edge of hindlimb insertion ED = ear diameter, measured vertically

Results
Based on the molecular data, Cyrtodactylus chungi sp. nov. is placed in a clade with C. cattienensis s. str. (Fig. 2). Besides the many similarities like the closed neck band, an X-shaped shoulder band, 5 or 6 irregular transverse dorsal bands (vs 4-6) and 11 or 12 transverse bands on the tail (vs 4-12) it was found that there are still distinct differences. The new species can be distinguished from C. cattienensis s. str. by a median row of slightly enlarged subcaudal scales, dark brown transversal bands of the tail narrower than the light brown spaces (vs wider transversal bands than the light spaces), by the presence of pitted, enlarged precloacal scales in the female (6 vs absence), by the higher number of tubercle rows on the tail (6 vs 4), which are strongly to slightly keeled (vs anteriorly keeled tubercle) and by the relatively strongly keeled tubercles on the hindlimbs (vs smooth tubercles).

Phylogenetic analyses
The alignment contained no gaps and the fi nal matrix consisted of 652 aligned characters. Of these, 244 were parsimony informative. The MP analysis of the dataset recovered a single most parsimonious tree with 1328 steps (CI = 0.35; RI = 0.65). The topology derived from BI was largely similar to those of Luu et al. (2017) and Pauwels et al. (2018). Although the phylogenetic tree supported by Pauwels et al. (2018) is better resolved, most of the deep nodes in this study are insignifi cantly supported. There are a few discrepancies between our phylogenetic hypothesis and that of Pauwels et al. (2018), in particular, the position of Cyrtodactylus phuocbinhensis Nguyen, Le, Tran, Orlov, Lathrop, Macculoch, Le, Jin, Nguyen, Nguyen, Hoang, Che, Murphy & Zhang, 2013. This species was supported as the sister species to C. sangi Pauwels, Nazarov, Bobrov & Poyarkov, 2018+ C. cattienensis Geissler, Nazarov, Orlov, Böhme, Phung, Nguyen & Ziegler, 2009 in our BI and ML analyses (Fig. 2 . In our study, however, we incorporated more species and used different outgroups than did in Pauwels et al. (2018), and these changes might have caused the differences.
The phylogenetic results show that the new species from Ta Kou Mountain, Binh Dinh Province was recovered as the sister taxon to C. cattienensis s. str. with a high statistical support (all equal to 100%) in all analyses (Fig. 2). In terms of genetic divergences, the new species is separated from C. cattienensis s. str. with uncorrected p-distance of 9.4-9.5%. It is also signifi cantly divergent from other species within the C. irregularis species group with the pairwise divergence of 14.2-18.3%. Genetic divergences between other species in the group are shown in the distance matrix (Supplementary File 1).

Diagnosis
The new species can be distinguished from other members of the genus Cyrtodactylus by a combination of the following characters (Table 1): relatively small body size (SVL up to 68.5 mm); a continuous neckband; 5 or 6 irregular transverse dorsal bands; 11 or 12 bands on original tail; keeled tubercles present on dorsum, posterior limbs and tail; 17 or 18 irregular dorsal tubercle rows; 30 or 31 ventral scale rows; ventrolateral skin folds indistinct; an angular series of seven precloacal pores in male and six pitted, enlarged precloacal scales in female, each series separated by a diastema of undifferentiated scales from 4-6 enlarged, poreless femoral scales; median subcaudals slightly enlarged; 17-20 subdigital lamellae under the fourth toe.

Etymology
Specifi c epithet 'chungi' is in honor of Prof. Dr Ngo Dac Chung, College of Education, Hue University, in recognition of his outstanding contributions to training of many herpetologist generations from Hue University. For the common names we suggest Chung's Bent-toed Gecko (English) and Thạch sùng ngón chứng (Vietnamese).

Description of holotype
Adult male, snout-vent length (SVL) 66.6 mm; tail length (TL) 62.7 mm (45.1 mm regenerated); head distinct from neck, relatively long (HL: 20.1 mm; HL/SVL ratio 0.32), HW 13.3 mm (HW/ HL ratio 0.66), HH 7.8 mm (HH/HL ratio 0.4); loreal region slightly bulged; snout elongated and rounded (SE: 7.6; SE/HL ratio 0.38); eye diameter 4.4 mm, more than half of snout-eye distance (OD/SE ratio 0.58); median snout scales small and rounded; eyes relatively small compared to head (OD/HL ratio 0.22); anterior supraciliaries large and rounded, posterior supraciliaries small and tapered; ear openings relatively small (ED: 0.75 mm; ED/HL ratio 0.04), vertical and oval in shape; eye to ear distance slightly longer than eye diameter (EyeEar/ED ratio 1.39); rostral wider than high (RH: 1.65 mm; RW: 3.3 mm; RW/RH ratio 2), with an inverse, Y-shaped median suture, surrounded by fi rst supralabial, two nostrils, two supranasals and internasal; lateral nostrils elongated, oval and surrounded by rostral, fi rst supralabial, three postnasals and one supranasal; supranasals separated by one internasal; eye separated from supralabials by 3 or 4 scale rows; mental triangular, wider than high (MW: 1.75 mm; MH 2.6 mm; MW/MH ratio 1.49), surrounded by fi rst two infralabials and posteriomedially by postmentals; supralabials 10/9; infralabials 9/9; body slender and short (AG/SVL ratio 0.41); ventrolateral skin folds not developed; dorsal scales granular, heterogeneous in shape and size; dorsal tubercles present in posterior part of head, extremities, dorsum and fi rst third of tail; postocular tubercles irregularly distributed, rounded, fl at and about three times size of surrounding granular scales; median dorsal tubercles about fi ve times as large as granular scales, oval, and not keeled; posterior dorsal tubercles keeled and on sacral region very strongly keeled up to base of tail; 18 irregular transversal dorsal tubercle rows; 24 paravertebral tubercles; each tubercle surrounded by 9 or 10 granular scales; ventrals smooth, oval and slightly imbricate, in 31 longitudinal rows; gular scales homogeneous; dorsal surfaces of upper and lower arms without tubercles, but occasionally enlarged granular scales; left thigh with series of 6 enlarged femoral scales separated by diastema of 6 smaller scales from precloacal scales; right thigh with series of 3 enlarged femoral scales separated by diastema of 8 smaller scales from precloacal scales and single enlarged femoral scale distally; no femoral pores; 45 enlarged precloacal scales arranged in diamond shape, with angled series of 7 pore-bearing scales (3 right, 1 angled, 3 left); 15 posteriorly enlarged scales and 21 anteriorly enlarged scales; two postcloacal tubercles on each side; scales on palms and soles rounded to oval, smooth and arched; fore-and hindlimbs moderately long (ForeaL/SVL ratio 0.15; Tibia/SVL ratio 0.17) relatively narrow; toes and fi ngers narrow and curved at interphalangeal joint; slightly transparent, pointed and curved claws, bordered ventrally by large lamella and dorsally by small, narrow scale; subdigital lamellae proximally larger, widened and merge basally into smaller scales of palm and sole surfaces; distal lamellae smaller, more elongated and distally increasingly imbricate; right fourth fi nger with 15 subdigital lamellae (6 proximal, 9 distal); left fourth fi nger with 16 subdigital lamellae (5 proximal, 11 distal); right and left fourth toe with 17 subdigital lamellae each (6 proximal, 11 distal); base of tail with small granular supracaudals and strongly keeled, rounded tubercles; last two thirds of tail regenerated, scales heterogeneously arranged and shaped; tubercles of original part arranged in 6 transverse rows (8/-/4/4/4/4) in a distance of 6-7 granular scale rows; tubercles slightly keeled; subcaudals of original part oval, rounded and imbricate; median subcaudals about two to three times larger than the lateral scales, but not transversely enlarged, or arranged in defi nable rows.

Coloration in preservative (Figs 5-8)
Ground color of head, neck, limbs and tail grey to light brown; dorsal surface of head brown with irregular, small, dark brown spots; two larger dark brown blotches located in posterior interorbital area and two other dark brown blotches in occiput area; skin above eyes dark grey, blue to black; rostral dark brown; supralabials and infralabials also brown, turning to cream at adjacent areas; a distinct dark brown stripe in loreal region, reaching to third supralabial; nuchal loop dark brown, in V-shape, extending from posterior corner of eye to the neck, edged in light grey; tubercles on head, limbs, dorsum and tail light to dark brown; tubercles on lateral skin fold white to gray; dorsal surfaces of limbs with small, irregularly distributed, dark brown spots; dorsal surface of toes and fi ngers dark brown spotted with light grey; dorsum with 5 irregular transverse dark brown bands; fi rst band over the shoulder region, X-shaped; dark and light brown spots present between transverse bands and on fl anks; ventral scales of body light grey to cream; lateral sides of head, below postocular stripe, and ventrolateral margins of fl anks slightly white to cream; anterior part of tail with two transverse, dark brown bands dorsally with dark brown spots in interspaces; dark, transverse dorsal bands separated from underside of tail by light brown, whitish spotted sides of tail; tail light grey to cream ventrally and marbled brown on the original part; regenerated part of tail grey to light brown, marbled.

Coloration in life
Ground color of dorsal surface of head, limbs and dorsum bright-yellow; transverse body and tail bands darker; lateral head and belly with bright-yellow spots; marbling of surface of limbs darker; ventral surface white to light blue (Figs 3-4).

Sexual dimorphism and variation
The morphological characteristics of the female IEBR 4582 correspond well with the holotype (IEBR 4581) and differ only in a few details. The female has some slightly keeled tubercles on the dorsal surface of the forelimbs. Dorsal tubercles are keeled from the posterior insertion of the forelimbs. The two supranasals are separated by two internasals. Another internasal is inserted between the rostral and the two internasals. Two inverse sutures are X-shaped from the lower edges of the internasal and do not form a median suture. The tail is original. The median subcaudals are about twice as large as the supracaudals and extend to the tip of the tail. The two dark brown spots of the occiput are longer and larger. The neck band is also V-shaped but pointed posteriorly. The dorsal surfaces of the forearms and hindlimbs with fi ne, irregular, transverse stripes. Dorsum with 6 transverse, irregular bands. The fi rst band, between the insertion of the forelimbs, is clearly X-shaped over the shoulder region. Between the neck band and the fi rst body band is an oval, dark brown spot. The dorsal tubercles are clearly lighter. The dorsal side of the tail bears 11-12 transverse, dark brown bands, which completely enclose only the last third. Between these bands are irregular, fi ne, light brown to brown stripes and dots. The fi rst two thirds of the underside of the tail are marbled light grey to brown. The last third resembles the dorsal side.

Comparison
We compared the new species with its 19 congeners from the Cyrtodactylus irregularis complex based on an examination of specimens (Table 2) and data obtained from the literature (Smith 1921a(Smith , 1921bZiegler et al. 2002Ziegler et al. , 2013Heidrich et al. 2007;Orlov et al. 2007;Nazarov et al. 2008 Table 3).

Distribution
Cyrtodactylus chungi sp. nov. is currently known only from the type locality in Ta Kou Nature Reserve, Binh Thuan Province, Vietnam (Fig. 1).

Natural history
Specimens were found at night between 20:00 and 22:00, on granite rock nearby a forest path, about 0.5-1.0 m above the ground, at elevations between 400 and 500 m a.s.l. The surrounding habitat was evergreen forest of medium and small hardwoods mixed with shrubs and vines. The humidity was approximately 68-72% and the air temperature ranged from 26.8 to 31.6 o C (Fig. 9).
The C. irregularis group has been split into a northern C. pseudoquadrivirgatus clade (A), including C. cryptus, C. pseudoquadrivirgatus, C. taynguyenensis, and a southern C. irregularis clade (B), consisting of the remaining taxa, based on two nuclear genes, RPL35 and Rag1 (S.N. . C. chungi was found within the distribution of clade B, which is characterized by enlarged femoral scales. However, group A lacks this character. Accordingly, this group is characterized by small femoral scales. Similar to the results based the single mitochondrial COI reported by S.N. , our analyses and previous studies based on the same gene have not been able to fully resolve the phylogenetic relationships within the C. irregularis group, especially the deep nodes of the phylogeny, although species in the group are highly diverged from each other in terms of genetic distance. In addition, substantial discrepancies between phylogenetic relationships derived from nuclear and mitochondrial markers (see S.N. ) emphasize the need to include additional molecular data to investigate the complicated evolutionary history of the group. Moreover, the C. irregularis species group still includes unsolved taxonomic issues, such as an obvious cryptic taxon sister to C. irregularis and the C. pseudoquadrivirgatus species group apparently containing more than one species.
The new species was discovered in Ta Kou Nature Reserve, Binh Thuan Province, from where C. takouensis was described (Ngo & Bauer 2008). The latter species, which distinctly differs in morphology (color pattern and scalation) and by genetic divergence (16.4%), also seems to inhabit a different microhabitat, because it was found in a cave in deciduous forest, whereas C. chungi was found on rocks in the evergreen forest.
The Ta Kou Nature Reserve is located in the Binh Thuan Province, coastal southern central of Vietnam. It is like an isolated island on land with endangered species, e.g., Trachypithecus germaini (Milne-Edwards, 1876) and Pygathrix nigripes (Milne-Edwards, 1871), and two new species of geckos were discovered, namely Cyrtodactylus takouensis Ngo & Baur, 2008 and Gekko takouensis Ngo & Gamble, 2010(Uetz et al. 2020. A major threat to the local fauna is habitat loss by agricultural extension and tourism development.