Description of a new species of Homonota ( Reptilia , Squamata , Phyllodactylidae ) from the central region of northern Paraguay

Homonota is a gecko distributed in central and southern South America with 12 species allocated in three groups. In this work, we performed molecular and morphological analyses of samples of Homonota from the central region of northern Paraguay, comparing the data with those of related species of the group: H. horrida and H. septentrionalis. We found strong molecular evidence (based on 16S, Cyt-b, and PRLR gene sequences) to distinguish this lineage as a new species. Morphological statistical analysis showed that females of the three species are different in metric characters (SVL and TL as the most contributing variables), whereas males are less differentiated. No robust differences were found in meristic characters. The most remarkable trait for the diagnosis of the new species is the presence of well-developed keeled tubercles on the sides of the neck, and lack of a white band (crescent-shaped) in the occipital area, which is present in H. horrida and H. septentrionalis. Nevertheless, in our sample, we found three specimens (one juvenile and two young adults) that exhibit the white occipital band. Thus, this character seems only reliable in adults of the new species. The new species is parapatric to H. septentrionalis, both inhabiting the Dry Chaco of Paraguay.


Introduction
Homonota is a gecko, inhabiting mainly xeric and rocky areas in central and southern South America (Cei 1993, Avila et al. 2012), with Homonota darwinii reaching the most austral distribution of the genus at 54° latitude south (Morando et al. 2014).Most of the species in the genus are nocturnal, although H. uruguayensis can be either diurnal or nocturnal (Carreira et al. 2005).Homonota horrida (Burmeister 1861), distributed in Argentina and Paraguay, was the second described species of the genus, after the controversial H. fasciata (Duméril and Bibron 1836).This latter species was described from "Martinique", a Caribbean island located out of the distribution of the southern cone Neotropical genus.Both species were considered synonyms by Abdala and Lavilla (Peracca, 1897), H. uruguayensis (Vaz-Ferreira & Sierra de Soriano, 1961), H. rupicola Cacciali, Ávila &Bauer, 2007, andH. taragui Cájade, Etchepare, Falcione, Barrasso &Álvarez, 2013; and the horrida group (indicated as fasciata group by Morando et al. 2014) which contains H. horrida (Burmeister, 1861), H. underwoodi Kluge, 1964, and H. septentrionalis Cacciali, Morando, Medina, Köhler, Motte & Avila, 2017. Cacciali et al. (2017) suggested that more revisions are needed to understand the true taxonomic status of H. fasciata because currently it is not possible to know to which group it belongs and it is considered incertae sedis.
Four species of Homonota are recorded in Paraguay: H. borellii, H. rupicola, H. horrida, and H. septentrionalis.The most commonly known species was Homonota horrida recorded for the "Chaco" (Kluge 1964, Talbot 1978, 1979).Even after the synonymy of H. horrida with H. fasciata (Abdala and Lavilla 1993) the name H. horrida was still used in Paraguayan reports (Aquino et al. 1996, Ziegler et al. 2002).Many specimens of H. septentrionalis were referred to as H. horrida (those from the westernmost part of the Paraguayan Chaco) according to Cacciali et al. (2017).Homonota rupicola is an endemic species found in a rocky hill, east of the Paraguay River; and H. borellii was recorded from a few specimens from "Defensores del Chaco" and "Médanos del Chaco" National Parks (Cacciali et al. 2016).Thus, most of the species of Homonota from Paraguay are present in the Chaco, which is part of the "Dry Diagonal" formed by Caatinga, Cerrado, and Chaco, characterized by dry seasonal woodlands (Prado and Gibbs 1993).In Paraguay the Chaco is divided in two ecoregions: Humid Chaco and Dry Chaco, and most of the Homonota samples are located in the latter (Cacciali et al. 2016).
After the description of H. septentrionalis, the same authors continued to study and analyze the taxonomy of Paraguayan samples of Homonota from the Chaco, within the framework of a barcoding initiative of the herpetofauna from Paraguay.We performed genetic and morphological analyses among different populations of Homonota from the central region of northern Paraguay.Based on genetic and morphological differences, and applying a species delimitation algorithm, we found enough differences to consider these new samples as a different taxonomic unit from those previously recorded for Paraguay.We present here a detailed analysis along with the description of this new species.

Methods
We extracted DNA from three samples of Homonota from the central area of northern Paraguay (Occidental Region), which were compared with available sequences of the remaining members of the genus (except H. fasciata) to assess its taxonomic relationships in the gene tree.We sequenced fragments of mitochondrial genes rRNA 16S and Cytochrome b (Cytb) and the nuclear gene prolactine receptor (PRLR).Samples used and GenBank accession numbers are specified in Table 1.Samples of 16S were available only for the horrida group.To root the tree we included two outgroups (Garthia gaudichaudii and Phyllopezus przewalskii (Table 1) based on Morando et al. (2014).
Tissue samples were first washed for 15 h with 50 ml Phosphate-buffered saline (PBS) (diluted of 1:9 PBS: H 2 O).The DNA extraction was carried out with the DNeasey kit of Qiagen.We used 25 μl of reaction mix for every sample for the PCR (except for PRLR where we used 20 μl).Reagents and concentrations for the PCR mix for the amplification of every gene, are provided in Suppl.material 1: Appendix S1.Primers (produced by Eurofins MWG Operon) used for amplification and sequencing, along with PCR conditions for each gene are detailed in Suppl.material 1: Appendix S2.
We used SeqTrace 0.9.0 (Stucky 2012) for examination of chromatograms and to generate the consensus sequences.We used MAFFT 7 (Katoh and Standley 2013) to automatically align the sequences through its webserver.For alignment of sequences of 16S, we included the Q-INS-i search strategy for corrections with the secondary structure of that gene (Katoh and Toh 2008).We used MSA Viewer (Yachdav et al. 2016) to visualize the alignments and export them to fasta format.We estimated the best substitution model for each gene (separately) with PartitionFind-er2 (Lanfear et al. 2016) using the PhyML 3.0 algorithm (Guindon et al. 2010).We used the corrected Akaike Information Criterion (AICc) (Burnham and Anderson 2002) to select the best substitution model, but under the premise that it is not correct to use models that include both +G and +I (Sullivan et al. 1999, Mayrose et al. 2005).Then we chose the subsequent model in the best partition schemes when both were suggested by the AICc.
We performed two phylogenetic analyses, first using a Maximum Likelihood (ML) approach, and then a Bayesian inference (BI) to compare the trees topologies.These analyses were made for each gene individually and for a concatenated dataset of the three genes together.For the ML analysis we used IQ-Tree (Nguyen et al. 2015) through its webserver (Trifinopoulos et al. 2016) using 10,000 non parametric bootstrap replicates plus 10,000 replicates of Shimodaira-Hasegawa approximate likelihood ratio (SH-aLRT) (Anisimova et al. 2011) and 10,000 ultrafast bootstrap (UFBoot) approximation replicates (Minh et al. 2013).We converted the alignment to nexus format in the online server Alter (Glez-Peña et al. 2010) available at http://sing.ei.uvigo.es/ALTER/, to be used in MrBayes v3.2 (Huelsenbeck and Ronquist 2001, Ronquist andHuelsenbeck 2003) for a BI.For this, we ran the analysis in independent duplicates, each with 1,000,000 generations for MCMC with a sampling frequency of 500 generations.We visualized the trees and exported them using FigTree v1.4.3 (available at http:// tree.bio.ed.ac.uk/software/figtree/).We considered convergence when the standard deviation of split frequencies was 0.015 or less and when the Potential Scale Reduction Factor approached 1.0 (Gelman and Rubin 1992).
When frequencies did not converge we continued adding 500,000 generations until convergence was achieved.
We assessed the degree of intraspecific divergence within the alignment (removing the outgroups) with the species delimitation test ABGD (Puillandre et al. 2012) through its webserver (http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html), using 10 steps of prior minimum and maximum simple genetic distance from 0.001 to 0.1 (default), and 0.5 of relative gap width, since higher (default) values tend to exceedingly split clades (Kekkonen et al. 2015, Yang et al. 2016).For this analysis we used only Cytb which was the mitochondrial gene better represented in our samples, and available for all the species within the genus.The last step using genetic data was the assessment of a species tree based on the clustering proposed by the species delimitation test.To do this we used *BEAST (Drummond et al. 2012) in BEAST 2.4.7 (Ogilvie et al. 2017) under 1,000,000 generations for the mcmc model, visualizing the posterior probability in DensiTree 2.2.6 (Bouckaert et al. 2014).
Additionally, we generated morphological data for 13 specimens (7 males and 6 females) of the new species and Measurements were taken with digital calipers (precision 0.01), but only the first decimal considered to limit discrepancies.For the morphological analyses only specimens of ~40 mm or larger were included.When paired structures exist, data are presented in left/right orientation, and only the left side was used for statistical analyses.In the color descriptions, the capitalized colors and the color codes (in parentheses) are those of Smithe (1981) for live animals and Köhler (2012) for preserved specimens.
We compared the morphological variation among species through a discriminant function analysis (DFA), testing the normality of the variables with a Shapiro-Wilk (W) test (Shapiro et al. 1968, Zar 1999), and for the DFA we only used variables that were normally distributed.We used PAST 3.14 (Hammer et al. 2001) to perform these tests.Meristic (discrete) and metric (continuous) data were analyzed separately.Examined specimens are detailed in Appendix 1.We present a table of localities of the specimens examined in Suppl.material 1: Appendix S3.

Results
The final alignments of 16S, Cytb, and PRLR consisted of 539, 793, and 457 bp, respectively.Alignments and trees are available at TreeBASE (ID: 22305).The best substitution model for 16S was GTR+G, for Cytb TVM+I(1 st pos)|TIM+G(2 nd pos)|SYM+G(3 rd pos), and for PRLRK81(1 st pos)|GTR+G(2 nd +3 rd pos).The complete table with scores is provided in Suppl.material 1: Appendix S4.The topology of the ML (Suppl.material 1: Fig. S1) and BI (Suppl.material 1: Fig. S2) trees using 16S coincide in recognizing three clusters, but the ML tree shows the new species as a sister clade to H. septentrionalis, whereas BI shows a trichotomy including H. horrida, H. septentrionalis, and the new species.Trees of ML and BI based on Cytb have the same topology (Suppl.material 1: Figs S3-S4), with strong support values.In these trees H. andicola and H. whitii are sister clades, as are H. darwinii and H. williamsii, and they are sister to the remaining Homonota species.The borellii group shows H. uruguayensis as sister to H. borellii + H. rupicola + H. taragui.Finally, within the horrida group, H. underwoodi appears as the sister to the remaining Homonota with banded coloration pattern.In this part of the tree H. horrida is rendered as sister to the clade H. septentrionalis plus the new taxon.The topologies of the ML and BI trees using PRLR are also the same (Suppl.material 1: Figs S5-S6).Species in the whitii group are clustered together, and the borellii group also shows monophyly but with a unresolved polytomy.In the horrida group H. underwoodi is also suggested as sister to the remaining species, with the new species and H. septentrionalis showing the most recent divergence.The trees using the concatenated dataset (with both ML and BI) show similar branch arrangement previously observed in trees of individual genes (Suppl.material 1: Figs S7-S8).Only two samples (UFRGS 2139 of H. uruguayensis and LJAMM-CNP 9813 of H. darwinii) are not allocated within their respective taxa, probably because some genes are lacking for some species.
The analysis of intraspecific genetic divergence with ABGD results in 12 groups (Suppl.material 1: Appendix S5), which represent nearly all described species (except H. fasciata) and the new species, providing evidence for its recognition as a distinct taxon.This is highly congruent with the clusters shown by the gene trees.The species tree shows consensus in the clusters of the three groups of the genus Homonota, with slight differences in the branch arrangements.The lower value (density of green lines in Figure 1) shows a trichotomy where the three groups are nested together, and with similar proba-bilities a species tree that clusters the whitii group as sister to the borellii group (density of red lines) and another where the whitii group is sister to the borellii group + horrida group (density of blue lines).Same as observed in the gene trees, H. underwoodi is presented as the sister clade to the remaining members of the group, and H. horrida sister to the new species of Homonota and H. septentrionalis with a rather deep divergence between these two taxa.
All continuous morphological variables had normal distributions (Table 2).The DFA for metric data showed that females of the three species are more differentiated than males (Fig. 2).The most contributing variables were SVL and TL for Axis 1, and SVL and TrL for Axis 2 and 3 (Table 3).Given the high eigenvalue of axes 1 and 2 (3.79 and 2.34 respectively, Fig. 2) suggests that the groups are significantly differentiated.For the meristic data, only DT, TVS, and LVS reached normality (Table 2), and DFA with these variables showed a high degree of overlapping without group discrimination and low eigenvalues (Fig. 3), and weak discrimination values (Table 3).Raw metric and meristic data are presented in Suppl.material 1: Tables S1 and S2.There is a strong molecular congruence in the recognition of 12 taxa within the genus Homonota (three of them with a banded coloration pattern), which added  to the significant differences among the three species with banded pattern based on the DFA, and the additional morphological distinctions discussed below are used to identify the new taxon described here.

Diagnosis.
A species of Homonota assigned to the horrida group given its relationship (based on molecular evidence) with H. horrida, and by the color pattern composed of a vertebral and five to seven transversal clear lines appearing as a banded Homonota smilar to H. horrida and H. septentrionalis.Homonota marthae has a robust body, and prominently keeled tubercles disposed in four to eight longitudinal rows on the dorsum.
Homonota marthae can be differentiated from all species in the genus, except H. fasciata, H. horrida, H. darwinii, and H. septentrionalis by the color pattern of transversal bands on the dorsum (reticulated pattern in the remaining species).Homonota marthae is further differ-  entiated from H. andicola, H. whitii, H. darwinii, and H. underwoodi by the keeled scales along the whole dorsum (vs.smooth dorsal scales in H. andicola, H. whitii, and H. underwoodi, keeled scales restricted to the posterior part of the dorsum in H. darwinii).It differs from H. fasciata by having a serrated edge of the auditory meatus (vs.smooth anterior margin in H. fasciata); presence of one or two enlarged tubercles on the upper edge of the auditory meatus (vs.no enlarged tubercles in H. fasciata); and a smaller size of the postmental scales (vs.postmentals of the size of the first infralabials in H. fasciata).Homono-ta marthae differs from H. horrida by the higher position of the ear opening in relation to the level of the mouth (vs.lower positioned in H. horrida); from H. septentrionalis by more developed keeled tubercles on the sides of the neck (Fig. 6) (vs.less developed tubercles in H. septentrionalis).Finally, adults of H. marthae differ from these both species by the lack of a white band (usually crescent-shaped) on the occipital area (vs.white occipital crescent-shaped band present in H. horrida and H. septentrionalis) (Fig. 7).An artificial key for identification of the species of the genus is presented at the end of the work.than high (1.5 mm) with a median groove covering the upper two thirds of the scale; nares surrounded by rostral, supranasal, and postnasal; SL 8/8; one elongated tubercular scale on the mouth commissure; muzzle slightly convex, covered by large homogeneous juxtaposed scales; head covered with big homogeneous juxtaposed scales on the dorsal area, intermixed with small granules; super-ciliary scales imbricated, associated to spiny-like scales on the posterior half of the orbit; scales on lateral surface of the head heterogeneously covered with strongly conical tubercles intermixed with small granules; auditory meatus oblique and with serrated edge, and one large elongated scale on the upper border; IL 6/6, the last less than half the size of the others; mental bell-shaped; two postmentals less than twice the size of the following posterior scales, contacting the mental, the first IL, and four posterior scales; scales under the head gradually reducing in size posteriorly; dorsal and lateral parts of the neck with granular juxtaposed scales mixed with tubercles; ventral side of the head covered by imbricate cycloid scales; body dorsally covered with 14-16 rows of strongly keeled scales, separated by one to two small granules in the pleural areas, and three to four granules in the vertebral area; ventral scales cycloid and imbricate arranged in 16 longitudinal rows at midbody; suprascapular, axillary, inguinal regions, and cloacal opening surrounded by small imbricate granules; anterior and dorsal surfaces of limbs covered by large imbricate scales, keeled on the dorsal surface; posterior region of limbs covered by small juxtaposed granules; ventral surface of forelimbs with juxtaposed granules, and ventral surface of hind limbs with large imbricate scales; subdigital lamellae of hands starting from pollex were recorded as follows: 8/8 -10/12 -15/13 -16/16 -11/11; subdigital lamellae of feet starting from hallux were recorded as follow: 15/13 -19/17 -15/16 -12/12 -9/9; tail with large imbricated and mucronate scales, 10-12 per caudal whorl.

Coloration of the holotype (in preservative).
After five years in preservative, the coloration was recorded as follows: Head Mikado Brown (42) with Warm Sepia (40) speckling on the dorsal surface; Warm Sepia (40) on the sides, with a Light Buff (2) line from nares to orbit, and continuing behind the orbit above the temporal region; supralabials and infralabials Medium Neutral Gray ( 298 (MNHNP 11793, 45 mm SVL;SMF 101438, 45 mm SVL) out of the 17 examined specimens of Homonota marthae have a trace of white crescent-shaped band on the occipital area (more visible in the SMF 101439, Fig. 7G), typical of H. horrida and H. septentrionalis.Nevertheless, many juveniles (such as SMF 101436) show the same coloration as adults (Fig. 7).The specimen MNHNP 7832 has a narrow occipital white band, joined to the postocular lines (Fig. 7C).Some specimens have a darkish coloration (MNHNP 2810, 10744, 11791, 11793) dorsally, and ventrally most of the specimens have a clearer color than the holotype, except for MNHNP 2798, 2810, and 10744.In some specimens (MNHNP 2795(MNHNP , 2798(MNHNP , 2810, 10744) , 10744) the dorsal color is diffused and the transversal bands are little visible.
Etymology.This species is named in honor of our indefatigable colleague Martha Motte, who is not only dedicated to safekeeping the herpetological collection of the "Museo Nacional de Historia Natural del Paraguay", but also does a great job in providing selfless support to scientists that are striving to improve the knowledge of the Paraguayan herpetofauna.
Habitat and distribution.Homonota marthae is known from the central area of the Paraguayan Dry Chaco in the Department of Boquerón (Fig. 5).The environment is a xeric forest with abundance of thorny vegetation and almost absence of a herbaceous stratum.Nevertheless, a more detailed analysis of museum collections is advisable for a better knowledge of the distribution of this species.
This species is a dry forest inhabitant, but it is also frequently found in human dwellings.Talbot (1978) recorded the use of logs of Drunken tree (Chorisia speciosa: Malvaceae) as shelter by Homonota in the Dry Chaco, since the wood of this tree keeps high water levels.Additionally, Cacciali et al. (2007a) demonstrated the use of subterranean caves (usually armadillo burrows) by Homonota in several areas of the Paraguayan Chaco.

Discussion
The diversity of species groups within the genus Homonota was explored in the last decade, and resulted in the description of H. williamsii (Avila et al. 2012) of the whitii group, and H. rupicola (Cacciali et al. 2007b) and H. taragui (Cajade et al. 2013) of the borellii group.However, the taxonomy of the horrida group (referred to as fasciata in Morando et al. 2014) was untouched for many years, and was comprised of two species (H.horrida and H. underwoodi).Recently, with the description of H. septentrionalis by Cacciali et al. (2017), and adding H. marthae described herein, the diversity of the horrida group currently includes four species.Morando et al. (2014) and Cacciali et al. (2017) presented species trees where the whitii group is sister to horrida and borellii groups.Our deep cluster arrangement is not completely resolved probably due to the use of fewer genes.Nevertheless, there is a strong consensus in the topology of the horrida group, where H. underwoodi appears as the sister of the remaining taxa (Fig. 1).
No obvious external synapomorphy is known to diagnose the horrida group.Three of the four species (H.horrida, H. septentrionalis, and H. marthae) have a pattern characterized by transversal body bands and the presence of a vertebral line.This coloration is different from the remaining species of the genus.The fourth species of the horrida group, H. underwoodi, has homogeneous body scalation and a completely different pattern, and therefore H. horrida and H. underwoodi were considered not to be in the same group (Kluge 1964).
The most obvious external difference between H. marthae and its presumed closest relative, H. septentrionalis, is the lack of a white occipital band in the former taxon, although we found some specimens (mainly juveniles or hatchlings) of H. marthae that do have the occipital band.Given that this white occipital band is also present in H. horrida, it could be a plesiomorphic character, and therefore the lack of it could be interpreted as the derived state.
Both species seem to inhabit in parapatry the Dry Chaco in Paraguay, and although a major revision of the whole distribution of the group is needed in order to know their actual ranges, H. septentrionalis is distributed in the north-westernmost part of the Dry Chaco, whereas H. marthae occurs in the central and easternmost areas of the Dry Chaco.Due to the lack of evident geographic barriers between these two species and considering their relatively low morphological variation (especially in males), they remained recognized as a single taxonomic unit until now.Parapatric speciation or breaks to gene flow without evident geographic barriers were observed and discussed by Irwin (2002), and also documented for other geckos in South America's Dry Diagonal, where Werneck et al. (2012) documented a high diversity in sympatric clades of Phyllopezus in Caatinga and Cerrado.
The degree of genetic differentiation between these two species is evident, and larger than the degree of morphological differences.Small morphological differentiation or even complete crypsis is common for many organisms, especially when they use the same ecological niche.Specifically for geckos, a recent study showed that it is difficult to find morphological diagnostic characters that match those observed by genetic evidence, as it is the case of the genera Garthia and Homonota, which are very similar morphologically (Daza et al. 2017).This is in agreement with previous studies that found that molecular genetic tools provided additional evidence for the interpretation of gecko's systematics in the Neotropics (Gamble et al. 2011, Gamble et al. 2012, Morando et al. 2014).The evolutionary processes that led to the molecular differentiation between H. septentrionalis and H. marthae remain unknown.
Finally, Homonota marthae is a common species that resists human perturbation and can be found in rural environments, and although its actual distribution limits are not yet known, and more revisions are needed to target this issue, probably the records of "Homonota fasciata" from Defensores del Chaco National Park referred by Cacciali et al. (2016) belong to H. marthae, and one of the records of H. marthae in Comunidad Ayoreo Tunucojai lays at ~70 km W from Yaguareté Porã Natural Reserve.Thus, we consider that H. marthae is not under extinction risk.

Key for identification of the species of the genus Homonota
Information to generate the key was based on Cacciali et al. (2007b), Avila et al. (2012), andCájade et al. (2013).Given that the holotype of H. fasciata is completely bleeched, we consider the information on its coloration from the original color description (Duméril and Bibron 1836).

Figure 1 .
Figure 1.Species tree (left) and density of trees (right) for the species of the genus Homonota, based on the genes 16S, Cytb, and PRLR.The intensity in the color of the species tree is proportional to the probability.

Figure 2 .
Figure 2. Discriminant Function Analysis scatter plot of individual scores for the three most informative axes for continuous variables of Homonota horrida, H. septentrionalis, and Homonota sp.n.Eigval: Eigenvalues.M: males.F: females.

Figure 3 .
Figure 3. Discriminant Function Analysis scatter plot of individual scores for the three most informative axes for discrete variables of Homonota horrida, H. septentrionalis, and Homonota sp.n.Eigval: Eigenvalues.M: males.F: females.

Figure 5 .
Figure 5. Occidental Region of Paraguay, indicating the political division, showing the known records for Homonota septentrionalis (white circles) and the analyzed records of Homonota marthae (black circles), and its type locality (star).Circle with a white cross, indicates origin of the genetic samples.High resolution elevation base map (30 seconds resolution) taken from Consortium for Spatial Information (CGIAR-CSI) available on http://www.diva-gis.org/gdata(Jarvis et al. 2008).

Figure 6 .
Figure 6.Plate showing the difference in scalation among individuals of similar sizes and same gender, of Homonota marthae (A, C, E) and H. septentrionalis (B, D, F).Note the more developed keeled tubercles on the sides of the neck in the former species.

Figure 7 .
Figure 7. Variation in color patterns of Homonota marthae (A-H).The lack of the white occipital crescent-shaped line (present in H. septentrionalis, I-L) is evident in most of the specimens.Scale bars: 5 mm.

Table 1 .
Köhler (2014)(2012)genetic analyses and GenBank accession numbers for every gene.Asterisks (*) indicate tissue samples without voucher.Numbers in bold are samples generated for this work.withsimilarpattern(relatedtaxa of the horrida group) looking for potential diagnostic characters.Thus, we used for comparison H. horrida (7 males and 5 females) and H. septentrionalis (10 males and 12 females), using standard variables (continuous data expressed in mm) already used byAvila et al. (2012)andCacciali et al. (2017):SVL snout-vent length, from tip of snout to vent.TrL trunk length, distance from axilla to groin from posterior edge of forelimb insertion to anterior edge of hind limb insertion.FL foot length, from the tip of the claw of the 4 th straightened toe to the back of the heel.TL tibial length, measured between the level of the knee and the level of the heel, as shown byKöhler (2014).AL arm length, from tip of claws of the 3 rd finger to elbow.HL head length, distance between anterior edge of auditory meatus and snout tip.HW head width, taken at the level of the temporal region, corresponding to the widest part of the head.HH head height, maximum height of head, at level of parietal area.END eye-nostril distance, from the anterior edge of the eye to the posterior edge of the nostril.ESD eye-snout distance, from the anterior edge of the eye to the tip of the snout.EMD eye-meatus distance, from the posterior edge of the eye to the anterior border of the ear opening. taxa

Table 2 .
Normality Shapiro-Wilk (W) values for metric (above) and meristic (below) characters showing the p value.Values shaded in gray did not reach normality.See Methods section for reference to the acronyms.