A new species of Liolaemus related to L. nigroviridis from the Andean highlands of Central Chile (Iguania, Liolaemidae)

Abstract The Liolaemus nigroviridis group is a clade of highland lizards endemic to Chile. These species are distributed from northern to central Chile, and currently there are no cases of sympatric distribution. This study describes a new species, Liolaemus uniformis sp. n., from this group, and provides a detailed morphological characterization and mitochondrial phylogeny using cytochrome-b. Liolaemus uniformis was found in sympatry with Liolaemus nigroviridis but noticeably differed in size, scalation, and markedly in the color pattern, without sexual dichromatism. This new species has probably been confused with Liolaemus monticola and Liolaemus bellii, both of which do not belong to the nigroviridis group. The taxonomic issues of this group that remain uncertain are also discussed.


Introduction
and morphological comparisons determined that this population represents a new species that belongs to the nigroviridis group. This new species occurred in sympatry with L. nigroviridis, constituting the first case of sympatry within this group of lizards.
The current study describes this new species and provides a full diagnosis in regards to other species of the nigroviridis group. Although the color pattern of this new species resembles L. juanortizi and L. lorenzmuelleri, the scalation is markedly different and the distribution is allopatric (> 240 km of separation). Moreover, various taxonomical aspects of the nigroviridis group that require attention are discussed.

Morphological data and analyses
Specimens of all species currently considered within the nigroviridis group were examined. Morphological characteristics were examined according to Etheridge (1995) and Lobo (2005). Body measurements were taken with a digital Vernier caliper (0.02 mm precision) and given as the mean ± standard deviation (x ± SD). We applied a Kolmogorov-Smirnov test to verify data normality, a subsequent t-test or Mann-Whitney U test was used if data passed or failed the normality test, respectively, to compare scale count (midbody, dorsal and ventral) and size (snout vent length, SVL) of the new species against some related species (Liolaemus constanzae, L. juanortizi, L. lorenzmuelleri and L. nigroviridis). Only significant results are presented. Scales were observed with different magnifying lenses. Scalation and measurements were recorded on the right side of the specimen. Dorsal scales were counted between the occiput and the anterior border of the hind limbs. Ventral scales were counted from the mental scale to the anterior margin of the cloacal opening. Stomach and intestinal contents were analyzed under a binocular stereoscope for one specimen of the new species. Data for the midbody scales of Liolaemus juanortizi were taken from one revised specimen and six reported in Young-Downey and Moreno (1991). Classification was carried out considering species currently assigned to the nigroviridis group (Troncoso-Palacios 2013). Liolaemus isabelae is included in the comparison but the relationship of this species with the nigroviridis group is uncertain (see Discussion). The examined specimens are listed in Appendix I. Some mapping data were taken from existing literature or field observations without specimen collection: 1) L. nigroviridis from Manque (Mella 2005), El Arpa and El Roble (Cianferoni et al. 2013), Riecillo (Núñez et al. 2010), Campana (Hellmich 1050), Chepical and Juncal (field observations, 32°16'S -70°30'W and 32°53'S -70°07'W respectively); 2) L. maldonadae from Los Molles (Núñez et al. 1991). Acronyms used are: Museo Nacional de Historia Natural de Chile (MNHNCL), Museo de Zoología de la Universidad de Concepción (MZUC) and Colección de Flora y Fauna, Profesor Patricio Sánchez Reyes of the Pontificia Universidad Católica de Chile (SSUC).

DNA purification, PCR amplification, and sequencing
Samples from liver and thigh muscle were obtained from ethanol-fixed lizards which were subject to a rehydration process according to Coura (2005). Samples were washed twice in distilled water for 5 min at 55 °C to remove the fixative and then rehydrated with 1x Tris/EDTA for 5 min at 55 °C and then 1M Tris pH 7.5, at 55 °C overnight. Right after, samples were digested with proteinase K (20 mg/ml) at 55 °C overnight. Genomic DNA isolation (mitochondrial and nuclear) was done with the Wizard® Genomic DNA Purification kit (Cat # A1120, Promega, USA) following manufacturer´s instructions. The mitochondrial gene Cyt-b was amplified from total DNA through two phase conventional PCR with the primers GLUDGL (5´-TGA CTT GAA RAA CCA YCG TTG-3´) and CB3 (5´-GGC AAA TAG GAA RTA TCA TTC-3´), reported in Torres-Pérez et al. (2009), to generate a 700bp amplicon. PCR reactions were performed with the SapphireAmp® Fast PCR Master Mix (Cat # RR350A, Takara Clontech, USA) using 100 ng of total genomic DNA as a template and following the instruction manual. Two-phase PCR cycling was as follows: Phase 1, initial 98 °C denaturation for 3 min, then 5 cycles of 98 °C denaturation for 30 s, 47 °C annealing for 45 s and 72 °C extension for 45 s. The Phase 2, next 40 cycles of 98 °C denaturation for 30 s, 58 °C annealing for 45 s and 72 °C extension for 45 s. A final 72 °C extension step for 5 min was added to finish the PCR. The 700 bp PCR amplicon was checked by DNA electrophoresis on a 1% agarose gel in 1x Tris-Acetate-EDTA (TAE) buffer. The amplicons were purified with the E.Z.N.A.® Cycle-Pure Kit (Cat # D6492-02, Omega Biotek, USA) and sent for capillary sequencing to Macrogen, Korea.

Phylogenetic reconstruction
The accession numbers of the Cyt-b mitochondrial loci sequences generated in this study and the sequences obtained from GenBank are indicated in Appendix II. Forty three nucleotide sequences involved in the analysis were aligned using MUSCLE (Edgar 2004). We used the JModelTest v2.1.7 (Darriba et al. 2012, Guidon andGascuel 2003) to select an appropriate substitution model (HKY + G + I), with a BIC index. We performed a Bayesian inference (BI) analyses with MrBayes v3.1.5 (Ronquist and Huelsenbeck 2003). Two independent analyses, each consisting of two groups of four chains that run independently, that were run for 15.0 × 10 6 generation and a at sample frequency = 1000. Priors were let by default. Phymaturus vociferator Pincheira-Donoso 2004, was selected as the outgroup. The 25% of samples were discarded as burnin when calculating the convergence diagnostic, assessed examining values of average standard deviation of the Potential Scale Reduction Factor (PSRF) for all parameters.

Results
The genetic tree constructed from mitochondrial DNA (mtDNA) (Fig. 1) placed the newly identified Liolaemus species as a sister taxon of L. nigroviridis (posterior probability pp = 1). However, no data are available for most of the species in the nigroviridis group as sample collection is hampered by the high altitudes where these species inhabit. Therefore, the discovered topology should be considered preliminary (see Discussion). Liolaemus monticola is nested with strong support (pp = 1) in the monticola group, the sister clade of the nigroviridis group. Liolaemus bellii is not closely related to the new Liolaemus or L. monticola, and is nested in a node with polytomy. Liolaemus uniformis sp. n. http://zoobank.org/B412BEF2-C337-4472-A4CE-9AFD73876B07 Fig. 2A Etymology. The species name "uniformis" (Latin) refers to the lack of dorsal pattern and uniform color found for both males and females.
Diagnosis. Liolaemus uniformis is larger than L. constanzae (Mann-Whitney U = 0.5, P < 0.01, Table 1). Liolaemus constanzae has sexual dichromatism, a feature absent in L. uniformis. Males of L. constanzae have a black vertebral line and black spots on the paravertebral fields (Fig. 3A), whereas L. uniformis has no dorsal pattern. Additionally, the southern distributional limit of L. constanzae in Agua Verde, Antofagasta Region, Chile (Ortiz 1975), is more than 750 km north of the type locality recorded for L. uniformis.
Liolaemus uniformis differs from L. isabelae (Fig. 3C), because in the latter the nasal and the rostral scales are in contact only in 25% of specimens, whereas in L. uniformis, these scales are always in contact. Males of L. isabelae have black ventral coloration, a yellow dorsal color with a black vertebral line, black bars in the paravertebral fields, and a black lateral band, or some males have a completely black dorsal color; all traits that are absent in L. uniformis. Additionally, the southern distributional limit of L. isabelae in Salar de Pedernales, Atacama Region, Chile (Pincheira-Donoso and Núñez 2005) is more than 650 km north of the type locality recorded for L. uniformis.
Liolaemus uniformis resembles L. lorenzmuelleri (Fig. 3E) and L. juanortizi (Fig.  3D), species suggested as conspecific (Pincheira-Donoso and Núñez 2005). However, the dorsal scales in L. lorenzmuelleri and L. juanortizi are noticeably larger than those of L. uniformis, and have a distinct "ovoid" shape. Liolaemus uniformis has more dorsal scales (60.0 ± 2.9) than L. lorenzmuelleri (48.4 ± 4.2) (t = -5.4, P < 0.01). On the other hand, while only one specimen of L. juanortizi was examined, this one has 52 dorsal scales, which is below of the range for L. uniformis (Table 1). Liolaemus uniformis has more midbody scales (60.4 ± 1.7) than L. lorenzmuelleri (54.9 ± 4.5) (t = 2.6, P < 0.05) and L. juanortizi (56.7 ± 2.1) (t = 3.2, P < 0.05). Liolaemus lorenzmuelleri has a dark vertebral line and dark transversal lines running from the paravertebral fields to the flanks, whereas L. uniformis has no dorsal pattern. The dorsal pattern of L. juanortizi is very similar to L. lorenzmuelleri, but some specimens have a black ventral coloration, a black lateral band, and the lack of a dark vertebral line, whereas L. uniformis has no black ventral color or black lateral band. Additionally, the southern distributional limit of L. lorenzmuelleri (Embalse La Laguna, Coquimbo Region, Chile) is more than 240 km north of the type locality recorded for L. uniformis; and the southern distributional limit of L. juanortizi in Quebrada Contrabando, Atacama Region, Chile (MNHNCL collection catalog, unpublished) is more than 520 km north of the type locality recorded for L. uniformis.
Liolaemus uniformis differs from L. melanopleurus (a species with only three known specimens from an undetermined location, Fig. 3B) in that the latter has a blue-gray dorsal coloration (Philippi 1860) and a black lateral band running from the axilla to the midbody, features absent in L. uniformis. Although the type locality of L. melanopleurus is undetermined, the syntypes were collected by Philippi in his journey through the Atacama Desert, between the vicinities of Copiapó (27°23'S) and San Pedro de Atacama (22°54'S), more than 530 km north of the type locality recorded for L. uniformis.
Liolaemus uniformis differs from L. maldonadae (Fig. 3F), because males of the latter have a yellowish or reddish dorsal color with black transverse dorsal and ventral bars and black lateral band, whereas L. uniformis has no dorsal pattern or black trans- verse ventral bars. Dorsal scales in L. maldonadae are noticeably larger than found in L. uniformis, and they have an "ovoid" shape. Dorsal and ventral scale counts in L. maldonadae do not overlap with the same scale counts in L. uniformis (Table 1). Finally, the southern distributional limit of L. maldonadae in Los Molles (Núñez et al. 1991) is more than 150 km north of the type locality of L. uniformis.
(*) Taken from Navarro and Núñez (1993). (**) Examined specimen plus Young-Downey and Moreno (1991)    with black reticulation, and females have a brown dorsal color with a black lateral band, black vertebral line, and black paravertebral spots. In contrast, L. uniformis has a brown dorsal color without any pattern. Molecular data show that Liolaemus uniformis is not closely related to L. monticola (Fig. 1). Moreover, L. monticola is smaller (maximum SVL = 65.6 mm) than L. uniformis (max. SVL = 89.1 mm) (t = 3.9, P < 0.01) according to our samples, and although Pincheira-Donoso and Núñez (2005) recorded a max. SVL = 67.3 mm for L. monticola, the difference between both species is marked. Moreover, L. monticola exhibit a characteristic black lateral band between the axilla and midbody (diffuse in females), and males have white dots dispersed on the dorsum and a series of small black spots on the dorsum (Fig. 6). All these traits are absent in L. uniformis. The upper altitudinal limit of Liolaemus monticola distributions is 2000 m a.s.l. (Espinoza et al. 2004, Fuentes andIpinza 1979), whereas L. uniformis has a lower altitudinal distribution limit of 2820 m a.s.l.
Moreover, L. bellii exhibit a characteristic series of black dorsal "W" o "V" shaped spots (Fig. 6) Two postrostrals. Four internasals. Heptagonal interparietal, with a central, small, and whitish central spot marking the position of the parietal eye. Interparietal smaller than the parietals, surrounded by seven scales. Seven scales between the interparietal and rostral. Thirteen scales between the occiput and the rostral. Orbital semicircle incomplete on the right side and complete on the left (formed by thirteen scales). Three supraoculars on the left side and four on the right. Six superciliary scales. Frontal area divided into three scales (1 posterior and 2 anterior). Preocular separated from the lorilabials by one loreal scale. Two scales between nasal and canthal. Nasal in contact with the rostral, surrounded by six scales. One row of lorilabials between the supralabials and subocular. Four lorilabials in contact with the subocular. Six supralabials, the fourth is curved upward without contacting the subocular. Four infralabials scales. Pentagonal mental scale, in contact with four scales. Four pairs of post-mental shields, the second is separated by two scales. Temporal scales smooth or slightly keeled, imbricated. Six temporal scales between the level of superciliary scales and the rictal level. Four scales on the anterior edge of the ear, which do not cover the auditory meatus. Poorly differ-  entiated auricular scale, pentagonal and located at the upper part of the meatus. Thirty gulars between the auditory meatus. Lateral neck fold is "Y" shaped. Ventrolateral fold running from the neck to the groin. Dorsolateral fold slightly developed, running from the ear to the base of the tail. Midbody scales: 60. Dorsal scales are lanceolated, imbricated, keeled (without mucrons), with few interstitial granules. Dorsal smaller than the ventrals. Dorsal scales: 58. Ventrals scales are polymorphic (rounded, rhomboidal, pentagonal or hexagonal) smooth, imbricated, without interstitial granules. Ventrals: 91. Three precloacal pores. Supra-femoral scales lanceolate, imbricated, smooth or keeled. Infra-femoral scales lanceolate or rounded, smooth and imbricated. Supra-antebrachials scales are rounded or lanceolated, imbricated and smooth or keeled. Infra-antebrachials are rounded, imbricated and smooth. Dorsal scales of tail are pentagonal or rhomboidal, imbricated and keeled. Ventral tail scales are rounded or rhomboidal, smooth and imbricated. Lamellae of the fingers: I: 9, II: 13, III: 20, IV: 20 and V: 13. Lamellae of the toes: I: 11, II: 15, III: 21, VI: 27 and V: 17.
Color of the holotype in life. The specimen is notable for its lack of pattern and uniform color. The head is brown and darker than the body. There are several white dots dispersed over the head and cheeks. The dorsum is coppery brown and has a few white-spotted scales that did not form a pattern. The subocular is brown and crossed by three white, vertical lines. The dorsal surface of the tail is light brown and without a pattern. The limbs are a dorsal-brown, similar to the dorsal surface, with white dots dispersed on the forelimbs and white transversal lines on the hindlimbs. The flanks are whitish with abundant dark brown scales. Ventrally, the hands, feet, thighs, vent, and tail are yellowish. The belly is whitish with dark dispersed spots and a dark ventral stripe. The throat is whitish with a dark thick reticulation. The precloacal pores are orange.
In general, all specimens have the pattern and color described for the holotype, with slight variations in shade. The male paratype has a dark brown throat. Two females have inconspicuous dark rings and an inconspicuous vertebral stripe on the dorsal surface of the tail. Also, two females have an olive hue on the snout. One female has a very inconspicuous series of dark crossbars on the paravertebral fields, which, while difficult to count, approximated eight. The juvenile has a similar pattern and color as the holotype, but it has an inconspicuous and fragmented dark vertebral line and inconspicuous dark spots on the paravertebral fields.
Distribution and natural history. This species is currently only known from the type locality in the surroundings of the Chepical Lagoon, approximately 30 km NE of Alicahue, in the San Felipe de Aconcagua Province, Valparaíso Region, Chile (Fig.  7). Specimens were collected on the west shore of the Chepical Lagoon (32°15'S -70°30'W, 3050 m a.s.l.). This new species was found inhabiting rocky areas with little shrubby vegetation composed mainly of high-Andean forbs, such as Chuquiraga oppositifolia and Azorella sp. (Fig. 8). This lizard was found in abundance and was observed to have saxicolous habits. It was active between 9:00 h and 18:00 h and took refuge under rocks. Moreover, this species was found in syntopy with Phymaturus alicahuense Núñez, Veloso, Espejo, Veloso, Cortés & Araya 2010. Specimens were also observed at lower altitudes (32°16'S -70°30'W, 2820 m a.s.l.) in similar environments, altitudes at which this species was found in sympatry with a few specimens of L. nigroviridis.
One of the collected specimens had a yellow flower inside of its mouth. An analysis of intestinal contents showed that L. uniformis is omnivorous; plant and Hymenoptera remains were found. A large quantity of nematodes from an unidentified species was found in the intestines. While the reproductive mode is yet unknown, at the time of sampling (December) no evidence of embryos was found but one female had several small oocytes. Comparisons with the reproductive modes of other species in the nigroviridis group would not be helpful as there is little available data. It is known that L. nigroviridis is viviparous (Donoso-Barros 1966) and L. lorenzmuelleri is oviparous (Cortés et al. 1995). Pincheira-Donoso and Núñez (2005) reported that L. maldonadae and L. isabelae are viviparous, but the source of this information is unclear (see Lobo et al. 2010:4) since the reproductive mode was not mentioned in the original descriptions Núñez 1993, Núñez et al. 1991).

Discussion
Almost no molecular data are currently available for the nigroviridis group, probably due to the great difficulties of obtaining samples since all of these species inhabit high altitude mountainous areas (Pincheira-Donoso and Núñez 2005), with only L. constanzae (Ortiz 1975) and L. nigroviridis (Espinoza et al. 2004) recorded below 2000 m a.s.l. (1400 m a.s.l. and 500 m a.s.l., respectively). Moreover, most specimens from the MNHNCL and MZUC collections (the two major herpetological collections in Chile) are fixed with formaldehyde, making DNA extraction and amplification challenging (Lin et al. 2009). In regards to previous works, Torres-Pérez et al. (2009) performed three phylogenetic analysis (Bayesian inference, ML and maximum parsimony) and found that L. nigroviridis is the basalmost species of a clade also composed of L. pseudolemniscatus + L. nigromaculatus + L. platei and that this clade is closely related to L. monticola + L. nitidus clade. Our results are very similar with the nigroviridis and monticola clades as sister groups, but we did not want to include "L. nigromaculatus" from GenBank (Torres-Pérez et al. 2009) because the true identity of this species was only recently clarified (Troncoso-Palacios and Garín 2013) and although a specimen voucher is indicated , no locality data is provided. Since we have not seen this specimen we are not sure if it belongs to the true L. nigromaculatus or to L. atacamensis. We also did not include "L. platei" from GenBank (Torres-Pérez et al. 2009) because the specimen voucher (MZUC-30556) was collected in Laja Lagoon, Chile (according to MZUC Book catalog, unpublished) out of the known range for L. platei (Troncoso-Palacios and Marambio-Alfaro 2011), so it could be misidentified. In a recently mitochondrial ML phylogeny performed for a region spanning ND1-COI, Troncoso-Palacios et al. (2015b) found that the L. nigroviridis + L. fuscus clade is the sister group of the monticola clade (L. monticola + L. nitidus + L. confusus). This is also very similar to our result, but since there are not Cyt-b data for L. fuscus, it could not be included in the present analysis.
We recognize that one limitation to our work is that it is based in a phylogenetic analysis of only one mtDNA gene and that a wider phylogenetic DNA analysis (including nuclear genes) should be conducted in the future. This is also true for most of the 21 species of Liolaemus (sensu stricto) described in the last five years, which have been classified through different methodologies in regards to DNA comparisons. For example, three species (L. chavin, L. pachacutec and L. wari) include data from two mtDNA genes and shared data in GenBank (Aguilar et al. 2013). As our work, five species (L. antumalguen, L. burmeisteri, L. cyaneinotatus, L. lonquimayensis and L. ubaghsi) have been described with only Cyt-b data, and one species has been described with two mtDNA genes (L. crandalli). However, DNA data from all these have not been shared in GenBank or other online databases (Avila et al. 2010, Escobar-Huerta et al. 2015, Esquerré et al. 2014, Martínez et al. 2011) which does not allow the replication of the provided phylogenies or genetic distances. Two described species (Quinteros 2012, Troncoso-Palacios et al. 2015a, L. abdalai and L. zabalai, are supported in regards to DNA features by previously published phylogenetic works. Nine species (L. aparicioi, L. carlosgarini, L. choique, L. chungara, L. nigrocoeruleus, L. pyriphlogos, L. riodamas, L. scorialis and L. smaug) have been described without the support of molecular data (Abdala et al. 2010, Esquerré et al. 2013, Marambio-Alfaro and Troncoso-Palacios 2014, Ocampo et al. 2012, Quinteros 2012, Quinteros et al. 2014, Troncoso-Palacios et al. 2015a). Finally, one species, L. shitan, was described (Abdala et al. 2010) despite that no molecular differentiation was previously noted (Morando et al. 2003). No description in the last five year had included nuclear genes or more than two mtDNA genes and in most cases when DNA phylogeny is provided no data are shared in GenBank or other online databases. It is evident that Liolaemus researchers should put emphasis on trying to improve this situation in the future.
Although L. uniformis is strongly supported as a sister species of L. nigroviridis (pp = 1), a comprehensive phylogenetic study with more species of this group is needed. For example, L. isabelae was not placed within the nigroviridis group in a mitochondrial phylogenetic study that included one specimen (Schulte and Moreno-Roark 2010), despite that this species has been determined to be a member of this group in cladistic (Lobo 2005) and phenetic studies (Pincheira-Donoso and Núñez 2005) based on morphology. We included this species in our comparisons but for the time being, this should not be considered part of the nigroviridis group. Although the morphological cladistic analysis (Lobo 2005) found five apomorphies for the nigroviridis group (range of scale organs on postrostral scales, fourth supralabial -subocular not in contact, range of lamellae on the fourth finger, intraspecific female pattern and the relationship between the subocular length and the eye diameter), this study does not include all species currently accepted as part of the nigroviridis group and does not indicate the specific variation ranges of variation for these features in this group. On the other hand, the phenetic analysis of Pincheira-Donoso and Núñez (2005) does not provide supporting data for the features that were included in the matrix, so it cannot be replicated (see Lobo et al. 2010).
Liolaemus uniformis resembles L. lorenzmuelleri and L. juanortizi in that the three species share a similar background dorsal coloration. Although no molecular data exists to compare L. uniformis with these two species, we propose that the marked differences in scalation and the strongly allopatric distribution (> 240 km of separation), which is quiet considerable for lizards, support classifying L. uniformis as a new taxon. Liolaemus uniformis has probably been misidentified as L. monticola by Núñez et al. (2010), who noted L. monticola as the only lizard species to inhabit in syntopy with Phymaturus alicahuense (no specimen collection indicated). However, the present study found P. alicahuense residing at over 2900 m a.s.l, whereas the upper altitude limit for L. monticola is 2000 m a.s.l. (Espinoza et al. 2004, Fuentes andIpinza 1979). Therefore, the present data indicates that the only lizards occurring in syntopy with P. alicahuense are L. uniformis and L. nigroviridis. Moreover, L. uniformis and L. monticola shows deep morphological and molecular differences. Liolaemus uniformis has probably also been confused with L. bellii (formerly L. altissimus altissimus) by Mella (2005), who found presence of the latter species in the highlands of Putaendo (no specimen collection indicated). However, a field expedition to the highlands of Putaendo by the authors of the present study found no specimens of L. bellii, and no additional records of L. bellii in this zone are known. Taking into account these details, in addition to both species having a similar background dorsal color, we think that L. uniformis might have been confused with L. bellii.
Several aspects of the nigroviridis group remain uncertain. For example, L. nigroviridis possibly contains at least two species, the nominal species from the Andean highlands and populations from Coastal highlands, formerly L. n. campanae (Cianferoni et al. 2013). Liolaemus juanortizi might be a junior synonym of L. lorenzmuelleri (Pincheira-Donoso and Núñez 2005), and although both are certainly very similar, it is difficult to carry out a study on this matter because the type series of L. juanortizi is lost (Valladares 2011) and there are very few samples of this species (Pincheira-Donoso and Núñez 2005). On the other hand, L. melanopleurus remains a problematic species in terms of identification as the type locality is imprecise and no additional specimens have been found in more than 100 years (Troncoso-Palacios 2012).
The present work contributes to the existing taxonomical knowledge, but the nigroviridis group of Liolaemus lizards remains poorly studied, and new samples are required to better investigate its challenging taxonomy.