A new species of Liolaemus (Iguania: Liolaemidae) from the hot deserts of northern Patagonia, Argentina

. A new species of Liolaemus is described from southwest of the town of Añelo, Neuquén Province, Argentina. Integrative evidence methodology of external morphological characters and molecular phylogenetic analyses of mitochondrial DNA (cyt-b ) is used to place the new species to the species group of Liolaemus boulengeri . The new species is phenotypically close to L. mapuc he . The new Liolaemus is medium to large in size (males 77.64–83.98 mm, females 72.88–78.58 mm), with evident sexual dichromatism. Genetic distances of the mtDNA (cyt-b ) between the new species and its closest relative species are greater than 3% ( L. cuyanus 7.48–12.02%; L. josei 7.56–9.60%; L. puelche 8.23–9.93%; L. mapuche 8.51–9.79%). Molecular and morphological phylogenetic results show L. mapuche as the sister species of the new one. The new species is larger than L. mapuche . Dorsal and ventral scales are more numerous in the new species than in L. mapuche , precloacal pores in females are present in L. mapuche and absent in the new species. It has strict psammophilic habits, using sand mounds and sheltering, under Alpataco ( Neltuma alpataco ) bushes. The L. boulengeri group now contains 75 species distributed in Argentina, Bolivia, Brazil, Chile, Paraguay, Peru and Uruguay.


Introduction
The genus Liolaemus Wiegmann 1834 is the third most diverse genus of all living tetrapods with 283 valid species (Abdala et al. 2021a) and, thus, cons titutes a very attractive study group of lizards for different disciplines of biology.

Collection of specimens and preparation
Nine individuals of the new species were collected, southwest of Añelo, Neuquén Province. Specimens were captured using a two meter long telescopic 'herpetological rod', with a nylon thread loop at its end. The capture was made by slowly approaching the rod to the lizard's body then the loop is placed around the neck and pulled up and backwards. This technique is effective for capturing lizards of various sizes and elusive behavior, when they are momentarily perched in places within reach of a person, and without causing them any physical harm. Some basic morphological characters were examined from each captured individual, sex was determined and confi rmed, and photographs were taken from different angles and perspectives in their natural environment. Most of the captured lizards were georeferenced from a mobile phone GPS waypoints application. Numerous other photographic records of lizards were made without capturing them. The individuals were sacrifi ced by injecting 1 ml of 1% Pentothal sodium and were fi xed with 10% formaldehyde and preserved in 70% alcohol after having collected tissue samples stored in 96% ethanol. Permits for collections were obtained from the Department of Territorial Development and Environment of Neuquén Province (Exp. 8903-3148 / 21), managed and authorized in the name of Pablo Chafrat. The tissue samples were sent to the molecular genetics' laboratory of IBONE, Corrientes, Argentina. Genomic DNA was isolated from ethanol-preserved liver samples using the extraction protocol described in Aljanabi & Martinez (1997). We apply polymerase chain reactions (PCR) to amplifi ed target regions of the mtDNA cytochrome b (cyt-b) region for 705 bp, under the conditions described in Aguilar- Puntriano et al. (2013). The integrity and quali ty of each amplifi ed DNA were verifi ed by electrophoresis and spectrophotometry, respectively. The samples were sent to Macrogen Inc., Korea, for sequencing.

Morphological observations
The characters described by Laurent (1985), Etheridge (1995), Abdala (2007), Abdala & Juárez Heredia (2013), and Quinteros et al. (2020) that are used in the taxonomy of Liolaemus were also used here to carry out the diagnosis and variations of the new species (Appendix 1). Description of the colors in life was made from photographs taken when capturing the lizards. The terminology of the body coloring pattern follows Hellmich (1934), Lobo & Espinoza (1999), and Abdala (2007). Observations of scaling and body measurements were taken using a binocular magnifying glass (10-40 ×) and a precision caliper of 0.01 mm (Mitutoyo®).

Phylogenetic analysis
Phylogenetic analyses based on morphological and molecular characters were carried out separately. For the development of the morphological matrix, we used the one cited by Abdala (2007), later modifi ed by Abdala & Juárez Heredia (2013) that included 90% of all species of the L. boulengeri group. The species used as an external group (fi ve species) were the same as those used in the work by Abdala (2007). The classifi cation used for the diagnosis in the non-formal categories of the subgenus Eulaemus corresponds to the new proposal by Abdala et al. (2021c). This proposal integrates phylogenetic hypotheses from various scientifi c manuscripts and authors, and suggests the following categories: genus, subgenus, section, group, clade, subclade, complex. The morphological matrix comprises 73 terminal taxa and 155 characters. Of these 155 characters, 32 are continuous and 123 are discrete, which were classifi ed as nonpolymorphic binaries, polymorphic binaries, non-polymorphic multistate, and polymorphic multistate. Following Abdala (2007), the multistate characters were divided into additives and non-additives. Polymorphic binary characters (Wiens 1995) were treated as such. The polymorphic multistate was treated as such with the values found for each taxon. In the phylogenetic analysis, the parsimony criterion was used as the optimality criterion. The software used to search for phylogenetic hypotheses was TNT 1.5 (Tree Analysis Using New Technology, ver. 1.5, Goloboff & Catalano 2016), since it is the only program that allows the analysis of continuous characters without converting them, for their treatment, into discrete characters. The continuous characters were treated using the methodology proposed by Goloboff et al. (2006), where they are analyzed as such, avoiding their discretization. For each character the range formed by the mean ± standard deviation was used. As the continuous characters are taken under different scales, a 'standardization' or 'rescaling' was carried out to avoid using some characters with greater infl uence over others in the analysis, using a script (mkstandb.run) associated with the TNT software (Goloboff et al. 2003). With this script, the maximum transformation costs that can exist between two continuous characters are standardized. In this way, from the smallest state to the largest, in a continuous character, a specifi c value is taken with respect to what it costs to transform it into a discrete character. For this analysis, 2 was considered as the highest transformation cost. Heuristic tests were carried out to fi nd the most parsimonious trees. For each heuristic search, 1000 replicas were made and 50 trees were saved for each one. The matrix was treated by analyzing the characters with equal weights and low implied weights (Goloboff 1993). For the latter, values of the weighing constant "K" from 1 to 20 were used.
We performed a molecular phylogenetic analysis under Bayesian Inference based on a matrix size of 143 terminals of cyt-b sequences of the Liolaemus boulengeri group (appendix 2) and two terminal taxa as an external group (L. vulcanus Abdala, 2011 andL. multicolor Koslowski, 1898). These sequences were obtained from GenBank (https://www.ncbi.nlm.nih.gov/genbank/), and we also included sequences from the new taxon and other taxa obtained for this study. The sequences were aligned in MEGA X (Kumar et al. 2018), using the Muscle algorithm. We used jModel Test ver. 3.0.4 (Posada 2008) to select the best fi tting model (GTR + Γ + I). Th e analysis was carried out in BEAST2 ver. 2.6.6 (Drummond & Rambaut 2007). We ran two runs of 50 million generatio ns each. Adequate mixing and convergence of the chain to the stationary distribution were confi rmed by inspection of MCMC samples using Tracer ver. 1.6 (Rambaut et al. 2014). The fi rst 20% of generations were discarded as burn-in, after evaluating the stability and adequate 'mixing' of sampled log-likelihood values assessed from the parameter estimates across generations (ESS > 200) of both runs were combined.
We use Tree Annotator ver. 2.0 (Drummond & Rambaut 2007) to generate a tree of maximum credibility and calculate the posterior probabilities and substitution rates for each node. The topology was visualized with Fig Tree ver. 1.2 (Rambaut 2010). We calculated the average of uncorrected genetic d istances of the closest species of L. kulinko sp. nov., L. cuyanus, L. josei, L. mapuche and L. puelche in MEGA X (Kumar et al. 2018).  1D). Within the group of L. boulengeri it belongs to the clade of L. melanops, subclade of L. goetschi, and L. cuyanus complex because it has light blue scales on the fl anks of the body and tail, a black margin on the posterior border of the paravertebral spots, four to six scales in contact with mental scale (Fig. 1E), presence of a melanic gular ring, evident scapular spots, and the same body shape and similar lepidosis (Abdala 2007;Abdala et al. 2012bAbdala et al. , 2021c. It differs from the species of the clades of L. anomalus and L. darwinii by having posterior teeth with crowns of expanded edges and four to six scales in contact with mental scale. It also differs from the species of the clade L. anomalus (Abdala & Juárez Heredia, 2013) (Liolaemus acostai Abdala & Juárez-Heredia, 2013, L. anomalus, L. ditatadi Cei, 1983, L. lentus Gallardo, Mertens, 1938, L. multimaculatus (Duméril & Bibron, 1837, L. occipitalis Boulenger, 1885, L. rabinoi (Cei, 1974), L. riojanus (Cei, 1979), L. salinicola Laurent, 1986, L. scapularis Laurent, 1982, andL. wiegmannii) in having one row of loreolabial scales (never two or three  Abdala, 2003, andL. telsen Cei &Scolaro, 1999) in having a greater SVL, four to six scales in contact with mental scales and a clearly different dorsal coloration pattern.

Taxonomy
It differs from the L. rothi subclade ( Table 1 for more differences.)

Etymology
The specifi c epithet 'kulinko' means 'aguada' in the language of the Mapuche, a group of indigenous inhabitants of south-central Chile and southwestern Argentina, including parts of Patagonia, and refers to the place where the species lives, "Aguada Pichana".  (Fig. 1). Dorsal and temporal region of head, brown with few whitish scales. Supralabial and infralabial light yellow. Dark spot through the subocular, loreolabial and supralabial scales. Sides of neck with thick white line reaching forelimbs. Brown body slightly darker than head. Undefi ned vertebral region, with yellowish-white circular or irregular spots. Evident paravertebral spots on neck to middle of body that gradually fade until disappearing when reaching tail. Paravertebral black spots with white posterior border. Antehumeral arch defi ned, black, wide, short. Faint prescapular spot, and large, evident postscapular spot, of deep black color. Sides of body same color as back, with numerous white circular spots, which disappear before reaching groin. Limbs with same color as body, with irregular black and whitish spots. Tail dorsally reddish brown without spots at its proximal end and with faint black vertebral line at its distal end. Ventral is yellowish white; yellow color is more accentuated in gular region and thighs. Gular band incomplete and two black lines extend from gular band to forelimbs. Numerous black scales distributed irregularly across venter, forming melanic spot in pectoral and anterior abdominal area.
MORPHOLOGICAL VARIATION. Based on nine specimens (4 ♂♂ and 5 ♀♀). Dorsal surface of head smooth with 16-18 (mean = 16.22; SD = 0.67) scales between the rostral and occiput. Nasal surrounded by 8-9 (mean = 8.22; SD = 0.6) scales. Supralabials 8-10 (mean = 9.00; SD = 0.50); 8-10 (mean = 8.89; SD = 0.78). Lorilabials arranged in single row. Supraocular 5-7 (mean = 5.75; SD = 0.4). Parietals larger than or equal in size to interparietal, surrounded by 6-8 scales (mean = 6.89; SD = 0.78). Infralabials 6-11 (mean = 7.78; SD = 1.39). Gulars 22-25 (mean = 23.78; SD = 1.39). Temporals 8-11 (mean = 10.1; SD = 0.46) keeled. Auditory meatus higher (mean = 2.92 mm; SD = 0.33) than wide (mean = 1.63 mm; SD = 0.20). Developed antehumeral fold. Head longer (mean = 16.34 mm; SD =0.98) than wide (mean = 11.98 mm; SD = 1.18) and high (mean = 8.73 mm; SD = 1.09). Trunk length (mean = 40.19 mm; SD = 3.18). Maximum snout-vent length 83.98 mm. Arm length (mean = 11.45 mm; SD = 1.48). Forearm length (mean = 9.60 mm; SD = 0.95). Hand length (mean = 11.65 mm; SD = 0.93). Thigh length     (Fig. 4A, C, E, G; based on nine specimens in life and 18 specimens photographed; with moderate sexual dichromatism). Males with highly variable coloration pattern and more showy colors than females; color of head varies from light brown to gray. Subtle dark brown line running from eye to neck, crossing temporal region. This line absent or attenuated in some large adults. Subocular generally white or gray in color, with dark upper border. Supralabials and lorilabials always lighter in color, generally light gray or white. General body color light brown or light gray, some specimens with bluish color. Vertebral region thin and not marked, except in proximal area of neck. A wide, short, antehumeral arch, black or gray in color, not expanded towards forelimbs. In larger males the gular melanism continues. With pre-and post-scapular spots of variable size and intensity, from almost absent to black or dark brown. Post-scapular generally slightly larger than pre-scapular. Several specimens with spots posterior to post-scapular, and these always smaller, black, or dark brown. Paravertebral spots arranged in non-aligned pairs along body, generally sub-quadrangular, dark brown on front and black on back, with white or light blue back edge. In some specimens spots in shape of line or stripe across body, while in others in the shape of point or small circle. In larger males, paravertebral spots in posterior region of body fainter or absent. A few specimens with almost total absence of paravertebral and lateral spots. Lateral spots diffuse or absent, but with same shape and color as paravertebral spots. Numerous scales and white spots on sides of neck and body. Some males with light blue scales forming thin line in lateral-posterior region of body. Same design as on body continues onto tail, but with line of light blue and / or green scales, iridescent on sides of tail. In several individual this line reaches sides of body. Ventrally, males white, with melanism present in gular, pectoral and abdominal region. Females with more uniform and consistent pattern of coloration than males. Head gray, bordered by gray in degrade. Line from eye to neck generally more notable than in males. Body also varies from gray to brown in color. Vertebral region more delimited in males than females, reddish-brown on back of body and tail. Antehumeral arch more diffuse than in males. Pre-and post-scapular spots present. Generally, largest post-scapular always black or dark gray in color. With sub-quadrangular paravertebral spots that are black or brown in color and white or yellow along posterior border and may be pointed in shape. On paravertebral spots, some females have reddish-brown spot or reddish-brown anterior border. Dorsolateral bands yellow, orange or white color, which can be continuous or fragmented and usually join at base of tail. In larger females dorsolateral bands not distinguishable or very faint. With and without lateral spots. When present they have same shape and color as paravertebral spots. Paravertebral spots unite on tail, where they form single stripe that extends longitudinally. White underneath with some irregularly scattered gray spots and scales.

Geographic distribution
The new species is restricted to the type locality and surroundings, in the localities of Aguada Pichana and Bajo de Añelo, Neuquén Province, Añelo Department, Argentina (Fig. 5), at elevations between 286-433 m a.s.l.

Phylogenetic analysis (Figs 6-7)
Both molecular and morphological analyses indicate that Liolaemus kulinko sp. nov. belongs to the L. boulengeri group. According to molecular data, Liolaemus kulinko is recovered within the subclade of L. goetschi, complex of L. cuyanus, as a sister species of L. mapuche. This relationship is nested within a clade made up of (L. goetschi (L. cuyanus (L. josei + L. puelche)) + (L. mapuche + L. kulinko sp. nov.)) (Fig. 4). This relationship is also refl ected in the table of molecular distances ( Table 2).
The results of the morphological analysis are partially consistent with the molecular ones. In all the runs carried out, the species under study was recovered as a sister species of L. mapuche (Fig. 7). Nevertheless, the L. cuyanus complex was not always recovered as monophyletic, since the species related to L. donosobarrosi ((L. tiranti + L. calliston) + (L. donosobarrosi + L. hugoi)) are recovered basal throughout the clade of L. melanops, very distant from the rest of species of L. cuyanus complex.
In the hypothesis presented in this work, Liolaemus kulinko sp. nov. shows 28 autoapomorphies, of which nine correspond to continuous characters and 21 to discrete ones (12 of scaling and nine of coloration). The relationship between L. kulinko and L. mapuche (Fig. 7), was recovered based on fi ve characters, one continuous and four discrete (one of scales and three of coloring).

Natural history
Liolaemus kulinko sp. nov., is an endemic species of the Añelo Department in the center of Neuquén Province, and whose type locality is located 30 km from the eponymous city, in the area between provincial route number 7 and Neuquén River. We consider that the river is a biogeographical barrier separating it from its sister species L. mapuche (Fig. 5). The Bajo de Añelo comprises an area of 280 km² that constitutes an extensive closed basin, displaying a convergent radial network of temporary channels (Pérez et al. 2009) and whose lowest elevation is at 223 m a.s.l. Sediments from the Cretaceous Period predominate in the area and correspond to different geological formations. These are mainly from the Neuquén group and modern sedimentary deposits of various origins (fl uvial, alluvial and aeolian). The internal dune system is the environment in which most of the endemic species of the area are found.
The vegetation is part of the Monte Phytogeographic Region (Cabrera 1971), which constitutes an arid to semi-arid biome that extends from Salta to Chubut Province (Abraham et al. 2009). The appearance of the vegetation is as islands or patches of shrubs that alternate with areas of bare soil (Bulacios Arroyo et al. 2021). Each patch constitutes a microenvironment that is used in different ways by the species that inhabit it (Bertiller et al. 2009). This ecosystem is characterized by the presence of numerous endemic species of insects, birds and reptiles (Roig-Juñent et al. 2001;Roig et al. 2009 (Burkart 1964;Cabrera 1971;Roig 1987;Arbo 1999;Gandullo et al. 2004Gandullo et al. , 2016Riveros et al. 2011  It frequents dunes and sandy areas, although it also lives in higher stony areas, on the lower slopes of Añelo. In March 2021, at twilight time, females were identifi ed in patches of vegetation along with newborn juveniles that may be their progeny; this suggests possible maternal care (Procheret personal observation).

Discussion
The validity of the new species L. kulinko sp. nov. is supported by molecular data (monophyly and p-distances > 8% on cyt-b) and is characterized by a unique combination of morphological characters.
The main different morphological characteristics of L. kulinko that differentiate it from its sister species L. mapuche, are: The precloacal pores in females which are absent in L. mapuche, and present in L. kulinko. The new species also has a greater number of dorsal (87-97 vs 70-86 scales), more scales around body (69-82 vs 65-76 scales), and more gulars scales (34-38 vs 25-35 scales). Patterns of coloration differ between the species. For example, head color in males of L. mapuche, is blue, light blue, brown or gray, but only brown or gray in L. kulinko; and in L. mapuche the posterior paravertebral spots in male L. mapuche are prominent, but blurred or absent in L. kulinko. With the description of Liolaemus kulinko sp. nov., the Liolaemus boulengeri group reaches 75 nominal species. In this work we recover the new species as a member of the L. melanops clade, and within the L. cuyanus complex (Abdala 2007). This complex is similar to the L. donosobarrosi group proposed by Avila et al. (2006), with few exceptions. Avila et al. (2006) recovered L. donosobarrosi as the sister species of L. cuyanus, while for Abdala (2007), and in this work, these species are not recovered as sisters, but forming part of a more inclusive group (i.e., the L. melanops clade). Within this complex Liolaemus kulinko is phenotypically closest to L. mapuche. But we fi nd morphological evidence that differentiates these two species. We also found high genetic distance values between L. kulinko, and L. mapuche (> 8%; Table 2). Genetic distance values greater than 3% are considered suffi cient to establish a species limit within Liolaemus (Avila et al. 2009;Troncoso-Palacios et al. 2016;Quinteros et al. 2020) , 1837)) and one species of Teiidae Gray, 1827 (Aurivela longicauda (Bell, 1843)). There are only a few similar cases in the country: among these other high-diversity areas are Caviahue-Copahue (Neuquén Province), with an assemblage of ten species of lizards (Abdala personal observation); El Nihuil (Mendoza Province), with eleven species (Semhan 2015;Abdala et al. 2016); Los Colorados in the Chaco Salteño with thirteen species (Lavil la et al. 1995). Such a rich assemblage of species should be a very important factor to be taken into account in the monitoring and conservation plans for certain areas of Aguada Pichana and Bajo de Añelo. Several of these species of Liolaemus show strict endemism for the Bajo de Añelo basin: the new species, L. cuyumhue, L. calliston, and L. quinterosi. This elevated endemism results from the combination of a rich geological history of macro and micro environmental conditions, which have led to speciation in the populations of Liolaemus over time.

Conclusions
Morphological and molecular data are concordant in confi rming that the focal population of Liolaemus correspond to a new species, which we name and describe. Morphological studies indicate that Liolaemus kulinko sp. nov. has 28 autoapomorphies (unique characters in the group) and that it belongs to the L. boulengeri species group, and is positioned within the clade of L. melanops, within the complex of L. cuyanus and that its closest relative is L. mapuche. The inter-morphological differences with this latter species are important despite their spatial and phylogenetical proximity. The Bajo de Añelo locality represents an area with strikingly rich and high endemism of lizards, such as L. cuyumuhue and L. quinterosi to which L. kulinko is added. The new species is considered as a DD IUCN category.