A new species of Pseudotrapelus (Reptilia: Squamata: Agamidae) from Central Arabia

A recent molecular phylogeny of the agamid genus Pseudotrapelus , distributed in the rocky areas of North Africa and the Arabian Peninsula, revealed the presence of a genetically distinct lineage around the city of Riyadh in central Saudi Arabia. With the inclu - sion of additional specimens, we were able to describe this lineage as a new species, P. tuwaiqensis sp. nov. , confined to the Tuwaiq Escarpment, thus endemic to central Saudi Arabia. Our results of morphological examinations and molecular analyses, using three mitochondrial (COI, 16S, ND4-tRNAs) and two nuclear (c-mos, MC1R) gene fragments, show the new species is genetically differ - entiated and phylogenetically close to P. sinaitus and P. chlodnickii .


Introduction
The Arabian Peninsula is comprised of diverse landscapes with massive deserts spanning across most of the interior, enclosed by mountain ranges on the margins of the peninsula (Edgell 2006), and is renowned for its harsh, hot, and arid climate.These diverse landscapes are often inhabited by specialized species/lineages displaying a wide array of unique adaptations to arid conditions.As the dominant inhabitants of arid areas, squamates became exemplary models to investigate biodiversity patterns, phylogeographic assessments, and ecological and evolutionary studies (Camargo et al. 2010).Squamate reptiles from the arid areas of Arabia have been recently shown to harbor more diversity than previously thought.In the past two decades, the growing number of taxonomic, biogeographic, and phylogenetic studies have drastically broadened our knowledge of the Arabian squamate fauna, often presenting many specialized deep lineages, as well as newly formed species.Some studies included revisions of geographically widely distributed genera (e.g., Carranza and Arnold 2012; Metallinou et al. 2012Metallinou et al. , 2015;;de Pous et al. 2016;Tamar et al. 2016aTamar et al. , 2016bTamar et al. , 2018;;Šmíd et al. 2013, 2015a;Simó-Riudalbas et al. 2019;Machado et al. 2021).Other studies focused on species complexes or identification and description of new taxa (e.g., Busais and Joger 2011;Metallinou and Carranza 2013;Vasconcelos and Carranza 2014;Machado et al. 2019;Tamar et al. 2019aTamar et al. , 2019bTamar et al. , 2019c;;Šmíd et al. 2015b, 2017, 2023;Burriel-Carranza et al. 2023).
Within the agamid fauna of Arabia, the genus Pseudotrapelus Fitzinger, 1843 is restricted to the hilly and mountainous areas enclosing the Arabian Peninsula and the Red Sea (Fig. 1; Arnold 1980;Disi et al. 2001;Baha El Din 2006;Gardner 2013;Tamar et al. 2016b; Bar et al. 2021;Carranza et al. 2021;Šmíd et al. 2021).In previous years, taxonomic work on Pseudotrapelus has been hampered by the conservative and homogeneous morphology of its members.For many years authors followed a conservative approach treating most variations as a single species, P. sinaitus (Heyden, 1827), with several mor-phological forms or as a species complex (e.g., Anderson 1896Anderson , 1901;;Arnold 1980;Fritz and Schütte 1988;Schätti and Gasperetti 1994;Baha El Din 2006;Sindaco and Jeremčenko 2008).Within the past decade, four new species have been described and one was resurrected (Melnikov and Pierson 2012;Melnikov et al. 2012Melnikov et al. , 2013aMelnikov et al. , 2015;;Uetz et al. 2023).Currently Pseudotrapelus consists of six recognized species (Uetz et al. 2023): (i) P. aqabensis Melnikov, Nazarov, Ananjeva & Disi, 2012, from north western Saudi Arabia, through southern Jordan westwards to the Sinai Peninsula; (ii) P. chlodnickii Melnikov et al., 2015, from Sudan and Egypt, including the Sinai Peninsula; (iii) P. dhofarensis Melnikov & Pierson, 2012, from the mountains of southern Oman and southern Yemen; (iv) P. jensvindumi Melnikov, Ananjeva & Papenfuss, 2013, endemic to the Hajar Mountains of Oman and the United Arab Emirates (UAE); (v) P. neumanni (Tornier, 1905), from the mountains of southern Yemen and southwestern Saudi Arabia; and (vi) P. sinaitus from the Sinai Peninsula in Egypt, eastwards through most of Jordan to southern Syria and northwestern Saudi Arabia (Fig. 1).
Unfortunately, the recent descriptions of four species were based on a small number of specimens, with no comprehensive comparisons among species, and genetic data that were based solely on the mitochondrial COI gene (Melnikov and Pierson 2012;Melnikov et al. 2012Melnikov et al. , 2013aMelnikov et al. , 2013bMelnikov et al. , 2015)).This resulted in insufficient diagnostic characters to identify specimens or describe new species (see Tamar et al. 2016b and references therein).These studies thus created much biogeographic uncertainty and taxonomic confusion within Pseudotrapelus.
The comprehensive molecular phylogenetic study of Pseudotrapelus by Tamar et al. (2016b) provided much needed information on the genetic structure of the genus, and the phylogenetic relationships among its members using multiple loci, as well as on the distribution ranges of its species.Tamar et al. (2019a) later revealed cryptic diversity within the genus, with an unnamed lineage from central Saudi Arabia that is distinct in all genetic analyses performed, including both mitochondrial and nuclear species delimitation methods, and is phylogenetically close to P. sinaitus and P. chlodnickii.The absence of detailed morphological comparisons among Pseudotrapelus members, caused by the lack of specimens of this cryptic lineage, hindered its taxonomic evaluation at that time.In this study, we aim to provide an integrative taxonomic assessment of this lineage, including the addition of new specimens from central Saudi Arabia.

Materials and Methods
In addition to the two specimens reported in Tamar et al. (2019a), we collected two additional Pseudotrapelus individuals from the vicinity of Riyadh in Saudi Arabia (Fig. 1) during a targeted fieldwork in 2019.The specimens were deposited in the National History Museum Prague, Czech Republic (NMP-P6V; see voucher codes in the taxonomic section below and in Appendices 1-3).

Molecular dataset and analyses
We assembled two Pseudotrapelus datasets for this study: (i) a concatenated dataset for investigating phylogenetic relationships, including sequences of mitochondrial and nuclear gene fragments for 40 specimens (Appendix 1); (ii) a barcoding dataset for comparison to type specimens of the genus, including sequences for 29 specimens (Appendix 2).Specimens of Acanthocercus, phylogenetically close to Pseudotrapelus (Pyron at al. 2013), were used as an outgroup.Sample codes, localities, and GenBank (https://www.ncbi.nlm.nih.gov/genbank)/BOLD(www.barcodinglife.org)accession numbers are in Appendices 1-2.
We extracted DNA from ethanol-preserved tissue samples using the Geneaid extraction kit following manufacturer's instructions.The concatenated dataset comprised four gene fragments (2,375 bp): two mitochondrial, the ribosomal 16S rRNA (16S; ~500 bp) and the NADH dehydrogenase subunit 4 (ND4; 681 bp) with the adja-cent histidine, serine, and leucine tRNA genes (tRNA; ~158 bp), and two nuclear, the oocyte maturation factor Mos (c-mos; 372 bp) and the melano-cortin 1 receptor (MC1R; 663 bp).These markers were used in previous phylogenetic studies on Pseudotrapelus (Tamar et al. 2016b(Tamar et al. , 2019a)), thus providing reliable and consistent comparison of phylogenetic relationships.Primers and Polymerase Chain Reaction (PCR) conditions used for the amplification of these markers are as detailed in Tamar et al. (2016b).We additionally amplified the barcoding mitochondrial gene fragment Cytochrome c oxidase subunit I (COI; 645 bp) to compare with the holotypes, paratypes, and topotypes of most species (following Melnikov et al. 2015;Melnikova et al. 2015).We amplified the barcoding region using RepCOI-F and RepCOI-R primers as detailed in Nagy et al. (2012).In all amplifications, we sequenced both strands of the PCR products at Macrogen (the Netherlands).We checked, assembled, and edited chromatographs using Geneious v.7.1.9(Biomatter Ltd.).We aligned the sequences for each marker using MAFFT v.7.3 (Katoh and Standley 2013).To remove difficult-to-align regions and poorly aligned positions of the 16S and tRNA fragments we used Gblocks (Castresana 2000) with low stringency options (Talavera and Castresana 2007).We translated the protein-coding genes into amino acids, and we detected no stop codons, suggesting that they were not pseudogenes.For the nuclear markers, we identified heterozygous positions and coded them according to the standard IUPAC ambiguity codes and resolved these sites, for each gene independently, by using the PHASE 2.1.1 algorithm (Stephens et al. 2001;Stephens and Donnelly 2003) implemented in DNASP v.6 (Rozas et al. 2017) with probability threshold=0.9.We tested the occurrence of recombination for the two phased nuclear-gene alignments using the Pairwise Homoplasy Index (PhiTest; Bruen et al. 2006) implemented in SplitsTree v.4.14.5 (Huson and Bryant 2006), and we detected no evidence of recombination.
For the phylogenetic analyses of both the concatenated dataset (16S, ND4, tRNA, c-mos, MC1R) and the COI dataset, we partitioned by gene and selected substitution models for each marker using JModelTest v.2.1.7 (Darriba et al. 2012;Guindon and Gascuel 2003).The best models were as follows: HKY+G for 16S, tRNA and MC1R, TrN+I+G for ND4 and COI, and JC for c-mos.We analyzed both datasets under Maximum Likelihood (ML) and Bayesian Inference (BI) frameworks.We treated alignment gaps as missing data and the nuclear gene sequences were not phased.We performed the ML analyses using IQ-TREE v. 1.6 (Nguyen et al. 2015) through the web interface (Trifinopoulos et al. 2016).Branch support was assessed with the Shimodaira-Hasegawa-like approximate likelihood ratio test (SH-aLRT; Guindon et al. 2010) and the ultrafast bootstrap (UFBoot; Hoang et al. 2018), both with 1,000 replicates.We conducted BI analyses using MrBayes v.3.2.7 (Ronquist et al. 2012) with nucleotide substitution model parameters unlinked across partitions.The different partitions were allowed to evolve at different rates.Two simultaneous parallel runs were performed with four chains per run for 10 6 genera-tions with sampling every 100 generations.We examined the standard deviation of the split frequencies between the two runs and the potential scale reduction factor (PSRF) diagnostic; convergence was assessed by confirming that all parameters had reached stationarity and had sufficient effective sample sizes (>200) using Tracer v.1.6(Rambaut et al. 2014).We conservatively discarded the first 25% of trees as burn-in.We explored patterns of intraspecific diversity and nuclear allele sharing within Pseudotrapelus by inferring statistical parsimony networks for the two individual nuclear phased genes with the program TCS v.1.21(Clement et al. 2000; connection limit of 95%), consisting of all sampled specimens for each marker.We used tcsBU (Múrias dos Santos et al. 2016) for visualization of the nuclear networks.We calculated inter-and intraspecific uncorrected p-distance for 16S, ND4, and COI with pairwise deletion in MEGA11 v.11.0.11(Tamura et al. 2021).

Morphological data and analysis
We examined the morphology of the four specimens from central Saudi Arabia and compared it with (i) data published in the descriptions of the recently described species and (ii) additional vouchered and photographs of unvouchered specimens of all the other Pseudotrapelus species.The examined specimens included four specimens of the new species described below, three specimens of P. aqabensis, two specimens of P. chlodnickii, three specimens of P. dhofarensis, one specimen of P. jensvindumi, two specimens of P. neumanni, and eight specimens of P. sinaitus.High quality photographs (479 in total) of all specimens examined morphologically have been deposited in the public database MorphoBank (http://www.morphobank.org) in their original resolution where they are freely available for download (project number 4714).The MorphoBank accessions along with locality details are provided in Appendix 3.
We recorded the following morphological data: total length; snout-vent length (SVL), measured from the tip of the snout to the anterior margin of the cloaca; number of upper and lower labials; shape and size of the ear opening; position and arrangement of head scales; the presence of enlarged scales in the occipital area of the head; character of dorsal lateral scales on the body and tail (homogeneous versus heterogeneous, keeled versus unkeeled); the length of fingers and toes; number of precloacal pores and their arrangement; body and tail coloration and patterning.

Results
The phylogenetic analyses of Pseudotrapelus using BI and ML methods based on the concatenated dataset yielded similar topologies, with most nodes well supported and all recognized species monophyletic (Fig. 2).The phylogenetic structure within Pseudotrapelus recovered two major clades.Clade I (Bayesian posterior probabilities/SH-aLRT/UFBoot support values 1/97.8/98;support values are given in the same order hereafter), though with unsupported internal topology in both the ML or BI analyses, included the recognized species P. chlodnickii and P. sinaitus, and the four samples from the vicinity of Riyadh as the new species described below.Clade II (support 1/98.9/99)included the four remaining species with a sister species relationship between P. jensvindumi and P. dhofarensis (support 0.95/21.3/55),and between P. neumanni and P. aqabensis (support 1/87.7/93),though the topological structure was not supported for the former relationship.In contrast to the concatenated dataset, the phylogenetic tree based on the single barcoding COI marker recovered mostly weak support for the topological structure, apart from the distinction of clade I with strong support (support 1/99.6/99).The BI and ML analyses of the COI dataset recovered similar unsupported topology, though in both methods all recognized species were monophyletic, as well as the new species described herein.The paratype of the new species was recovered in a distinct lineage and did not cluster with any type specimens of the other recognized species.The nuclear networks inferred for the phased c-mos and MC1R alignments (Fig. 3) exhibited no allele sharing between any of the species in the MC1R network, contrasting to the c-mos network, in which allele sharing was present among the phylogenetically close Arabian species of clade II, which may indicate incomplete lineage sorting rather than gene flow among species.The two allele networks yielded a similar pattern of distinct haplotypes for the species of clade I, including the new species described below.Mitochondrial uncorrected interspecific genetic distances among Pseudotrapelus species (Appendix 4) ranged between 2 and 8.6% in 16S and between 11-19.8% in ND4.The lowest distances are between P. aqabensis and P. neumanni (2.2% in 16S and 11% in ND4).The genetic distances between the new species described herein and the other recognized species of clade II ranged between 7.9 and 8.6% in 16S and between 18.2 and 19.8% in ND4, while the distance to the species of clade I, to which it belongs, ranged between 6.3 and 7.0% in 16S and between 15.3 and 15.6% in ND4.The uncorrected interspecific genetic distances among the species in the COI marker indicate that the new species described below presents mostly high distances from all other species (between 13.6% and 15.9%).

Taxonomic account
Based on the degree of genetic differentiation of the new lineage from central Saudi Arabia from all other Pseudotrapelus species at both mitochondrial and nuclear levels, we recognize this lineage as a new species that we formally describe herein.

Pseudotrapelus tuwaiqensis sp. nov.
https://zoobank.org/A076FAE7-9DA1-460A-8CDF-FA160002D67CMorphoBank, project 4714, M862812-M863071 Etymology.The species epithet tuwaiqensis is derived from the geographic feature the species is associated with, the Tuwaiq Escarpment, that cuts through central Saudi Arabia from the southwest of the country to slightly north and northwest of the city of Riyadh.

Diagnosis.
A Pseudotrapelus species forming a clade together with P. sinaitus and P. chlodnickii, with the follow- ing combination of morphological and genetic characters: (1) large size with a total length of 200-203 mm and SVL 70.7-76.6mm; (2) 14-19 upper and 14-18 lower labial scales; (3) ear opening very large, oval, rimmed anterodorsally by conical scales of different sizes that give it a serrated appearance; (4) scales in the occipital area predominantly not enlarged; (5) heterogeneous dorsal scales with the mid-dorsals being distinctly keeled and larger than the scales on the flanks; (6) one continuous row of 4-7 precloacal pores in both sexes; (7) 3 rd toe longer than the 4 th ; (8) tail scales not arranged in whorls; (9) body and tail beige-brown in life with dark brown or orange transverse bars, the first on the nape, the second and the most prominent one in the scapular region, the third at midbody, the fourth in front of the insertion of the hind limbs, the fifth at the tail base; and the tail with regular dark bars down its length; ( 10

Differential diagnosis.
The genus Pseudotrapelus is morphologically very conservative, and it is virtually impossible to phenotypically distinguish one species from another without knowing precise locality data.Pseudotrapelus tuwaiqensis sp.nov. is no exception to this.While available literature and recent descriptions of new species in-dicate some key features that allow species identification, these do not hold when confronted with additional material (Table 1).For example, one of the diagnostic characters of P. jensvindumi was a small gap in the middle of the precloacal pores row (Melnikov et al. 2013a).However, similar state is present in some other species including P. aqabensis (specimen CN15112, MorphoBank accessions: M862611-M862613), P. dhofarensis (Melnikov and Pierson 2012, fig. 7 therein; JIR107, MorphoBank accession: M862670), and P. sinaitus (JOR22_75, Mor-phoBank accession: M862793; JOR22_85, MorphoBank accession: M862799).Similarly, P. neumanni is supposedly the only species with enlarged scales in the occipital area of the head (Melnikov et al. 2013a), but our examination of specimens of the other species revealed that in fact all but P. aqabensis and P. sinaitus have these enlarged scales present, at least in some specimens.In line with the above-said, our morphological examinations of P. tuwaiqensis sp.nov.did not reveal any characters that would be unique for the species and allowed its unambiguous identification (Table 1).As a result, we advise caution and prudence when trying to key out Pseudotrapelus specimens on the basis of morphology alone.
With the currently available evidence, the safest and perhaps the only way to differentiate the individual Pseudotrapelus species is either based on the origin of the specimens, or by using DNA barcoding.The ranges of most species are allopatric or parapatric, with P. tuwaiqensis sp.nov.being the only species that occurs in central Saudi Arabia around the city of Riyadh and further to the north and south along the Tuwaiq Escarpment.The other species are distributed in the peripheral mountain ranges of Arabia and some of them overlap to a certain extent (e.g., P. neumanni and P. dhofarensis in Yemen, P. neumanni and P. aqabensis in southwestern Saudi Arabia, P. aqabensis and P. sinaitus in northwestern Arabia; Fig. 1).The most reliable species identification tool in Pseudotrapelus is thus DNA barcoding.All presently recognized species have multiple specimens sequenced for three mitochondrial genes, including the COI marker that is the most commonly used barcode for animal identification (Vences et al. 2012) and show marked genetic differentiation in all these markers across species (Appendix 4).Pseudotrapelus tuwaiqensis sp.nov.can be clearly differentiated from its congeners at the genetic level by p-distances of 6. .In addition, all Pseudotrapelus species have been sequenced for two nuclear genes, of which the MC1R has unique alleles for each species and the c-mos has unique alleles for P. tuwaiqensis sp.nov., P. sinaitus,and P. chlodnickii (Fig. 3;Appendices 5,6).Pseudotrapelus aqabensis, P. dhofarensis, P. jensvindumi, and P. neumanni are genetically indistinguishable in the c-mos sequences.
Description of the holotype.Total length 203.9 mm (SVL 70.7 mm; original tail 133.2 mm).Head and body depressed (Fig. 4A, B).Three internasal scales, the middle one is triangular and contacts the rostral dorsally.Nostril tubular directed laterally and slightly posterodorsally, pierced in the posterior part of a large convex, smooth, pear-shaped nasal scale, which is situated on the anterior edge of the canthus rostralis.The nasal scale is partially visible from above and directly in contact with the first canthal scale.The first six canthal scales not in contact with the eye.Scales on the head smooth, somewhat polyhedral, interorbital scales as large or larger than the supraorbital scales; imbrications of temporal scales directed ventrally.Occipital pierced by a visible pineal foramen posteriorly, surrounded by irregular scales.Occipital scales are somewhat smaller than other head scales (Fig. 4C).19 (left)-17 (right) upper and 18 (left)-17 (right) lower labial scales (Fig. 4D, E).Ear opening as large as the eye, surrounded at its border by several single short conical scales of different sizes, with one single conical scale at the lower anterior edge of ear opening (Fig. 4E).Gular fold absent.Dorsal scales heterogeneous with a medial longitudinal rows of enlarged scales present.Medial dorsal scales diamond shaped and distinctly keeled, with the keel extending along the entire scale and not mucronate.Lateral scales not keeled.Gular and ventral scales smooth, becoming feebly keeled toward the lateral side of the body.Dorsal scales on limbs keeled and imbricate.Hind limbs long with the 3 rd toe reaching to the nostril level when adpressed.The 3 rd toe is the longest, reaching 9.3 mm (Fig. 4F).Ten lamellae under the left 4 th finger, 17 lamellae under the left 4 th toe.Forelimbs long with digits reaching to the cloaca when adpressed.The 3 rd finger is the longest, reaching 6.5 mm.One row of six continuous precloacal pores each about the size of two to three other precloacal scales (Fig. 4G).Tail depressed at its base, with a small pit after the cloaca.Large hemipenial pockets are absent, but two small bulges on either side of the pit are present.Dorsal tail scales strongly keeled, slightly mucronate, somewhat larger than the body scales.Ventral tail scales keeled and slightly mucronate.Tail scales not arranged in whorls.

Coloration (ethanol preservation).
Upper parts of the body and head brownish, a pale vertebral stripe from just behind the occiput along the body and extending on the tail is visible (Fig. 4A).The neck and anterior part of the dorsum are a dark brownish-gray color.A dark transverse band posteriorly to the forelimbs is present.Tail with dark brown, transversal, hourglass-shaped (when viewed dorsally) stripes starting just behind the cloaca and crossing the tail in regular intervals; the stripes do not extend onto the ventral side; the first and second stripe separated by seven light scales and the distance between the stripes diminishes posteriorly; the hourglass shape clear for the first four stripes, from the fifth on they become uniformly broad.Throat dark with a pattern of regular white spots or ocelli (Fig. 4D).Belly and the underside of the limbs and tail white (Fig. 4B).

Coloration in life.
All specimens share the general coloration pattern (Figs 4H,5).Body is light brown with faint light speckles scattered randomly on the body dorsum.
Head is slightly darker than body and has a blue tint in females.The nape is dark.There are prominent dark brown transverse bars across the body, the first is behind the insertion of forelimbs and runs to about the middle of the flank.Another, shorter and less conspicuous dark band is situated before the hind limbs.In some specimens, another small transverse stripe is in the middle of the body, sometimes marked only as a small darker patch.The dark dorsal stripes are orange or have reddish outlines in females.All body parts are white from the ventral side except for the throat, which has a dark background with regular, white, and evenly spaced ocelli.In some specimens, the throat pattern is less discernible and the dark and light scales form a reticulated pattern.
Habitat and Distribution.Pseudotrapelus tuwaiqensis sp.nov. is a rock-dwelling species inhabiting rocky areas, outcrops, isolated rock mounds and even dry riverbeds with large boulders (Fig. 6).Like all other Pseudotrapelus species, it perches on top of stones and rocks during the day.Individuals were found sleeping at night tucked in rock crevices or laying on the ground in the open.Currently available distribution data indicate that the species is endemic to Saudi Arabia, where it is confined to central Saudi Arabia around the city of Riyadh (Fig. 1).Most records come from the Tuwaiq Escarpment that runs from southwestern Arabia northeastwards to Riyadh where it bends to the northwest.The escarpment is bordered by massive sand seas: Rub al Khali from the southeast, Ad Dahna from the east, and An Nafud from the north, which seem to form a barrier to dispersal for these rock-dwelling agamas.There are several records west of Tuwaiq, suggesting that the species' range extends more to the west and is not confined to the escarpment.More field data are however needed to delineate the geographic limits of its distribution and its possible contact with P. aqabensis in the west and north-west, P. sinaitus in the northwest, and possibly P. neumanni in the southwest.Other reptile species observed to live in syntopy with P. tuwaiqensis sp.nov.were Bunopus tuberculatus, Stenodactylus doriae, Tropiocolotes wolfgangboehmei (all Gekkonidae), Ptyodactylus hasselquistii (Phyllodactylidae), Acanthodactylus boskianus, Mesalina lineage 4 (sensu Sindaco et al. 2018;Lacertidae), and Echis coloratus (Viperidae).

Discussion
A previous phylogenetic study of the agamid genus Pseudotrapelus disclosed an old cryptic lineage in the Arabian Peninsula (Tamar et al. 2019a).That study, using species delimitation analyses and phylogenetic relationships among currently recognized species, highlighted the presence of genetic differentiation and existence of a yet undescribed taxon within this genus.In this study, we investigated this old diversification event that has resulted in the description of a new species within Pseudotrapelus.We used an integrative approach applying phylogenetic analyses with morphological comparisons to assess the interspecific diversity of these agamid species distributed in the mountainous regions of Arabia and northeast Africa.The inferred topology of the concatenated dataset was congruent across analyses and supports the current taxonomy (Tamar et al. 2016b(Tamar et al. , 2019a)).We support the differentiation of a lineage occurring in the mountains of central Saudi Arabia described herein as P. tuwaiqensis sp.nov., phylogenetically close to P. sinaitus and P. chlodnickii.According to our phylogenetic analyses and genetic distances, the genetic distinctiveness of P. tuwaiqensis sp.nov. is evident at both the mitochon- drial and nuclear levels, not clustering with any samples of other members of the genus, as well as owning private alleles in the two nuclear genes analyzed, suggesting no signal of gene flow.Tamar et al. (2019a) included two samples of P. tuwaiqensis sp.nov.(the holotype and one paratype) and showed their distinctiveness using concatenated and species tree analyses, and mitochondrial-and nuclear based species delimitation analyses.These results further support the species status of P. tuwaiqensis sp.nov.
The current distribution of P. tuwaiqensis sp.nov.(specimens genetically identified) is known from around the city of Riyadh and is confined to the Tuwaiq Escarpment in Central Saudi Arabia (sightings are also known from the Ibex Reserve Protected Area located ca. 150 km south of Riyadh).The divergence time estimates of the clade comprising P. tuwaiqensis sp.nov., P. sinaitus, and P. chlodnickii was approximately from the late Miocene to the Early Pliocene, around 6 million years ago (Tamar et al. 2019a).The aridification of Arabia, fluctuating climate, and the progression of sandy areas in the Arabian region during the Late Miocene (Edgell 2006;Preusser 2009) might have created distributional restrictions for the ancestral Pseudotrapelus populations, limiting them to the remaining rocky habitats.Although the current geographic distribution of P. tuwaiqensis sp.nov.seems to have been influenced by the formation of the sand seas of Arabia (Edgell 2006), it is not possible to draw any firm conclusions about the biogeographic history of the species, especially given the unsupported phylogenetic relationships and the complex and dynamic geological history of the area.
Thanks to the extensive DNA barcoding efforts across Arabia, we have fairly good knowledge on the distribution limits of the individual Pseudotrapelus species across the peninsula.Nonetheless, there are still regions that would require more attention both in terms of fieldwork and genotyping.One such region is eastern Yemen where P. dhofarensis and P. neumanni probably connect.Other such areas are the geographically intermediate, and somewhat isolated, populations in Hail Province in northwestern central Saudi Arabia and near Wadi ad-Dawasir close to the 'Uruq Bani Ma'arid protected area in the southwest of the country.Both are part of the Arabian Shield rocky desert that stretches from the western coast inland to the Arabian interior and is typical for rugged, rocky outcrops and isolated inselbergs (Edgell 2006), all of which represent an ideal habitat for Pseudotrapelus.It therefore seems likely that they belong to some of the western mountain species, probably P. aqabensis.However, the area around Wadi ad-Dawasir lies at the southern edge of the Tuwaiq Escarpment and it cannot be ruled out that P. tuwaiqensis sp.nov.occurs this far south along the ridge of the escarpment.These are, however, mere speculations that will need to be verified by means of DNA barcoding.
Interestingly, in the phylogenetic study of Pseudotrapelus by Melnikova et al. (2015), the authors hinted at the potential existence of an unidentified population in northern and central Saudi Arabia, that should be further taxonomically investigated.Their hypothetical pop-ulation from Saudi Arabia around Riyadh most likely represents P. tuwaiqensis sp.nov.They also implied a possible hybridization between populations of P. sinaitus with that of the central Arabian species in the area of Al Mudawwara in southern Jordan.Our findings indicate no evidence of previous hybridization events between the two species, and propose complete reproductive isolation between P. tuwaiqensis sp.nov.and its geographically neighboring species.However, we must bear in mind that both our and Melnikova et al. (2015) results were based on a rather limited number of specimens, and thus more data and denser sampling of specimens and loci is needed to properly assess the distribution ranges and phylogenetic relationships within clade I of Pseudotrapelus.
Regarding conservation implications, considering the probable endemicity of the new species to central Saudi Arabia, its conservation status should be evaluated at national and worldwide levels, taking into account its limited distribution range, population density and trends, potential hybridization with other species, and future urbanization that is planned in the area of its occurrence.Interestingly, our field observations indicate that population densities of P. tuwaiqensis sp.nov.are much lower than those of other Pseudotrapelus species (e.g., P. dhofarensis in southern Oman, P. sinaitus in Jordan).In March 2023, a comprehensive field survey was conducted for three weeks by a team of four herpetologists actively searching for reptiles.The survey took place in habitats suitable for Pseudotrapelus within the known range of P. tuwaiqensis sp.nov. in the area north of Riyadh.Despite the efforts, the survey yielded the observation of only two individuals.
Fieldwork was conducted as part of the project "Systematics and biodiversity of the reptiles of southwestern Saudi Arabia," supported by the Saudi Wildlife Authority (SWA).We thank the SWA staff, most notably the vice-president Hany Tatwany, for his support, encouragement, and field work and permit arrangements.We thank Jerome Gaugris, Marius

Figure 1 .
Figure 1.Geographic distribution of the genus Pseudotrapelus in Arabia and adjacent territories.The circles show distribution records, those in color have been confirmed genetically to belong to the respective species.The white circles can be identified with confidence only to the genus level.All known records for Saudi Arabia, Yemen, Oman, and the UAE are shown.The map is not exhaustive in terms of records of P. aqabensis, P. chlodnickii, and P. sinaitus from Jordan west to Africa.Locality details of the numbered localities of P. tuwaiqensis sp.nov.are given in Appendices 1-2.

Figure 2 .
Figure 2. Bayesian inference phylogenetic trees of Pseudotrapelus.The trees were reconstructed based on A the concatenated dataset (16S, ND4, tRNA, c-mos, MC1R) and B the barcoding COI marker dataset.Support values are indicated near the nodes (Bayesian posterior probabilities/SH-aLRT/UFBoot).Sample codes correspond to specimens in Appendices 1, 2. Letters in rectangles in the COI tree indicate type specimens -H: holotype, P: paratype, T: topotype.

Figure 3 .
Figure 3. Allele networks of Pseudotrapelus for the two nuclear markers.A c-mos and B MC1R. Circle sizes are proportional to the number of alleles.White circles represent mutational steps.
) three unique mutations in the MC1R alignment: position 264 C instead of T, position 508 G instead of A, position 562 G instead of C; (11) one unique mutation in the c-mos alignment in position 202 C instead of G (see Appendices 5, 6).