Contributions to Zoology, 85 (1) – 2016V. Deepak; Varad B. Giri; Mohammad Asif; Sushil Kumar Dutta; Raju Vyas; Amod M. Zambre; Harshal Bhosale; K. Praveen Karanth: Systematics and phylogeny of Sitana (Reptilia: Agamidae) of Peninsular India, with the description of one new genus and five new species

To refer to this article use this url: http://ctoz.nl/vol85/nr01/a04

Results

Delimitation of species

next section

Morphological characters including dewlap morphology, dewlap colour, body size ratios, and variations of the scales in flank and around head were used in species delimitation. Based on detailed morphological examination of museum specimens and fresh samples collected during this study we describe one new genus Sarada gen. nov. and five new species of Fan-throated lizards: Sarada gen. nov., Sarada darwini sp. nov., Sarada superba sp. nov., Sitana visiri sp. nov., Sitana laticeps sp. nov. and Sitana spinaecephalus sp. nov.

Phylogenetic analysis

The Maximum Likelihood (ML) and Bayesian analyses generated similar trees, and we have opted to base subsequent analysis and interpretation on the ML tree shown in Fig. 2. In both the ML and Bayesian trees, three major clades were recovered. Clade 1 is designated as the Sarada deccanensis clade, Clade 2 is the Sitana spinaecephalus clade, and Clade 3 is the Sitana ponticeriana clade. Additionally, when all the four genes where analyzed separately both ML and Bayesian methods recovered a well-supported monophyly of the Sarada deccanensis clade. The Sarada (gen. nov.) clade is sister to the two Sitana clades, and exhibited high mtDNA divergence from them (uncorrected p-distance = 21-24%). The uncorrected p-distances between the Sitana ponticeriana and Sitana spinaecephalus clades was also quite high (14-16%). Each species was found to be monophyletic and well-supported with high bootstrap and Bayesian posterior probability support values despite that each was represented by specimens collected from across the species’ range. The uncorrected p-distance in mtDNA among species varied from 5-11% for Sarada and 4-16% for Sitana (Fig. 2, S4). Closely related species exhibit similar dewlap morphologies and colouration (Fig. 3). Hemipenial morphology, in contrast, exhibits substantial variation even among closely related species, with each species characterized by distinct morphologies. For example, both Sitana and Sarada include species exhibiting single and bilobed hemipenes (Fig. 4A-F).

FIG2

Fig. 2. Maximum likelihood phylogeny of Fan-throated lizards, showing the three different clades (Refer Fig. 3 for dewlap morphology and Fig. 4 for hemipenial morphology). Nodal support is indicated by the coloration of the circles at the nodes, with blue and white fill corresponding to ML bootstrap ≥75 and pure blue fill indicating bootstrap support > 95%.

FIG2

Fig. 3. Dewlap morphology and colouration of Fan-throated lizards. Clade 1: A. Sitana darwini sp. nov., B. Sarada deccanensis comb. nov., C. Sitana superba sp. nov.; Clade 2: D. Sitana spinaece­phalus sp. nov., E. Sitana laticeps sp. nov.; Clade 3: F. Sitana ponticeriana, G. Sitana visiri sp. nov., H. Sitana cf. bahiri. Scale bar = 10 mm

FIG2

Fig. 4. Drawings of sulcal view of a representative hemipenis for different species of Fan-throated lizards. Clade 1: A. Sarada darwini sp. nov. B. Sarada superba sp. nov.; Clade 2: C. Sitana spinaecephalus sp. nov., D. Sitana laticeps sp. nov.; Clade 3: E. Sitana ponticeriana, F. Sitana visiri sp. nov. Scale bar = 2mm.

Morphological analyses

Measurements were taken from 28 individual males representing the Sarada deccanensis clade, 25 individual males representing the Sitana ponticeriana clade, and 35 individual males representing the Sitana spinaecephalus clade. Twenty out of the 22 morphometric variables were highly correlated (Pearsons correlation > 0.70). Only TD, TW, DWLT and NE were normally distributed (Shapiro-Wilk test of normality P < 0.01), while the remaining variables did not follow the assumptions of normality. The PC1 axis explained 86% of total variance and PC2 explained 4% of the variance. Sitana and Sarada are similar in their general body shape because there was low variation within loadings in PC1 (Table 1). There are higher negative loadings on SVL, LAL, HL and UAL in the first component (PC1) and TD in second component (PC2). There are higher positive loadings on DWLT, 4TL and HFL in the second component (PC2) (Table 1). The species in the genus Sarada are large lizards and separated in morphospace with minor overlap with that of the one large specimen in Sitana spinaecephalus sp. nov. and one Sitana visiri sp. nov. (Fig. 5). Sarada darwini sp. nov., Sarada superba sp. nov. and Sarada deccanensis comb. nov. have some amount of overlap in their body size (Fig. 5). The two species within the Sitana ponticeriana and Sitana spinaecephalus clades are clearly distinct in body size ratios (Fig. 5). In essence head length, snout-vent length, limb lengths and size of dewlap are some of the morphometric characters which separate different genera and species of Fan-throated lizards (Table 1), of which some are evident in our raw data (see Tables 2 and 3).

FIG2

Fig. 5. PCA plot for Fan-throated lizards showing clear separation of Sitana from Sarada gen. nov.

FIG2

Table 1. Factor loadings on the principal components (PC) for each morphometric variable derived from PCA

FIG2

Table 2. Range of select measurements in adult Sarada specimens including the type series, mean in brackets. Abbreviations, all measurements in mm; “_” = absent.

FIG2

Table 3. Summary of select measurements in adult Sitana specimens including the type series. Abbreviations, all measurements in mm; “_” = absent.

External morphology

Presence of enlarged scales on the body and thigh of Sitana is one of the key character in differentiating Sarada gen. nov. which lacks both these characters (Table 4, Fig. 6). There were no visible external morphological scale characters to differentiate species in the genus Sarada. Presence of enlarged spine like scales on the back of the head is a key character to differentiate Sitana spinaecephalus sp. nov. from the rest of the species in Sitana. Number of lamallae on the fourth toe, supralabials, infralabials, midbody scales, enlarged scales on the lateral side of body were all overlapping in Sitana (Table 5). Number of enlarged scale rows on the dewlap was relatively low in S. laticeps sp. nov. with marginal overlap with scale counts of S. visiri sp. nov. (Table 5). Ventral scale counts were also overlapping in S. ponticeriana, S. visiri sp. nov., Sitana spinaecephalus sp. nov. and S. laticeps sp. nov. (Table 5). However ventral counts were useful in differentiating the species from Sri Lanka and India. (See key to the species). Details on the characters which were used to distinguish each species and genus are given in the taxonomy and systematics section below.

FIG2

Fig. 6. The position and structure of the enlarged projecting scale on the thigh of Sitana. Note the complete absence of this scale in Sarada gen. nov. Clade 1: A. Sarada darwini sp. nov., B. Sarada deccanensis comb. nov., C. Sarada superba sp. nov.; Clade 2: D. Sitana spinaecephalus sp. nov., E. Sitana laticeps sp. nov.; Clade 3: F. Sitana ponticeriana, G. Sitana visiri sp. nov., H. Sitana cf. bahiri. Image not to scale.

FIG2

Table 4. Range of select scale counts Sarada gen. nov. specimens including the type series. Abbreviations, all measurements in mm, “_” = absent. Lam4 = Lamallae on fourth toe; SL = Supralabials; IL = infralabial; MBS = Mid body scales; ENL = enlarged scales on the flank region; ESD = enlarged scale rows in dewlap; VEN = number of ventral scales.

FIG2

Table 5. Range of select scale counts of Sitana specimens including the type series, sample size in brackets.. Abbreviations, all measurements in mm, “_” = absent. Lam4 = Lamellae on fourth toe; SL = Supralabials; IL = infralabial; MBS = Mid body scales; ENL = enlarged scales on the flank region; ESD = enlarged scale rows in dewlap; VEN = number of ventral scales.

Hemipenis morphology. Sarada superba sp. nov., S. ponticeriana, S. spinaecephalus sp. nov. have a bilobed hemipenis (Fig. 4). However, they differ in their microornementations; S. superba sp. nov. have large calyces at the base of the sulcal side of the hemipenis compared to S. spinaecephalus sp. nov. and S. ponticeria­na. Sitana spinaecephalus sp. nov. have a prominent grove on the asulcal side of the hemipenis which is absent in all other species (S2). Sarada darwini sp. nov., S. laticeps sp. nov. and S. visiri sp. nov. have a single lobed hemipenis (Fig. 4). They also differ considerably in their microornementation; Sarada darwini sp. nov. have relatively less calyces on the asulcal side (only on the apex) compared to S. laticeps and S. visiri in which the calyces extend further down (S2). S. visiri have several smaller calyces on the center of the apical region on the asucal side whereas in S. laticeps and S. darwini they are more or less uniform (S2). More detailed descriptions of the hemipenies of each species is given in the species description section (appendix).

Osteology. In Sarada darwini sp. nov., S. decca­nensis, Sitana ponticeriana and S. spinaecephalus the skull is subpentagonal in outline. Width of supratemporal fossa less than or equal to 60% of length. Splenial fused with the surrounding mandibular. The hyoid appears to be mostly bony, except the tip of the entoglossal process, tip of ceratobranchial I, tip of epibranchial I and the ceratohyals which are cartilaginous. The ceratobranchial is evidently longer in male than the female owing to the large dewlap in males. Several skeletal characters overlap between the four species (Table 6) except a few; Sitana ponticeriana and S. spinaecephalus n.sp. have 19 vertebrae excluding atlas and axis whereas S. darwini and S. decca­nensis only have 18. Sitana ponticeriana and S. spinaecephalus n.sp. have four phalanges on the fourth toe whereas S. darwini and S. deccanensis have five.

FIG2

Table 6. Select osteological characters & measurements of representative species from the genus Sitana & Sarada gen. nov.

Taxonomy and systematics

Generic level taxonomy. We describe the new genus Sarada gen. nov. based on specimens collected from throughout its distributional range. The absence of enlarged scales on the thigh and lateral body in combination with body ratios and dewlap color pattern can be used to differentiate Sarada gen. nov. from genus Sitana (Fig. 3, 6). The enlarged scales on the lateral body are shown in the description section (Figs 10b-13b, Figs 15b-17b). Both Sitana and Sarada gen. nov. differ from Otocryptis (Wagler, 1830) in lacking the fifth toe (Smith, 1935). The skeletal characters of Sitana and Sarada gen. nov. are overall similar except two striking differences, an additional phalange on the fourth toe of manus and one reduced trunk vertebra in Sarada gen. nov., which is also observed in the other closely related sister genus and species Otocryptis beddomei (Boulenger, 1885) (Table 6, Fig. 7). Based on the body ratios and body size we identified three distinct morphotypes in the new genus Sarada gen. nov., which includes Sarada deccanensis comb. nov. and two undescribed species that are described here. Thus the nomen Sitana deccanensis become Sarada deccanensis comb. nov. Our molecular phylogenetic analysis suggests monophyly of the proposed genus containing the above mentioned three morphotypes collected from various localities in Maharashtra and Karnataka (Fig. 2, S1). Molecular analysis suggests that the Sarada gen. nov. is sister to Sitana with a deep divergence between these genera. Sarada gen. nov. is distributed predominantly in Maharashtra and north Karnataka (Fig. 8).

FIG2

Fig. 7. Clear stained images of left forelimbs showing finger formulae in: (A) Otocryptis beddomei, (B) Sarada darwini sp. nov. (C) Sarada deccanensis comb. nov. (D) Sitana ponticeriana (E) Sitana spinaecephalus sp. nov. Scale bar = 10 mm. Note the differing numbers of phalanges on the fourth digit of Sarada (6) versus Sitana (5).

FIG2

Fig. 8. Map showing our sampling in Peninsular India and the distribution of the previously known species of Sitana from India and Sri Lanka. Type localities of different species are marked with arrows. Major rivers are marked by a blue line and capitalized bold text. Based on Johnson (1867) and Colton and Colton (1855), the approximate limits of Berar and Deccan during the 1870s is indicated. Chanda is now called Chandrapur in Eastern Maharashtra.

Species level taxonomy. Based on detailed examination of multiple specimens from different populations we provide detailed diagnostic characters for the species in the genus Sarada gen. nov. and Sitana (see key to the genus and species). Sarada deccanensis comb. nov. can be differentiated from Sarada superba sp. nov. in having relatively long hindlimb length (HLL) which exceeds SVL (df=1, F=32.37, p < 0.001; Table 2), relatively longer heel length (HFL) to SVL ratio (df=1, F =38.95, p < 0.001; Table 2) and smaller body size (SVL) (df=1, F=6.52, p < 0.05; Table 2). Sarada deccanensis comb. nov. can be differentiated from Sarada darwini sp. nov. in having relatively long hindlimb length (HLL) which exceeds SVL (df=1, F=19.88, p < 0.001; Table 2), relatively longer heel length (HFL) to SVL ratio (df=1, F=9.38, p < 0.01; Table 2) and larger body size (SVL) (df=1, F=5.52, p < 0.05; Table. 3). Sarada superba sp. nov. can be differentiated from Sarada darwini sp. nov. in having shorter heel length (HFL) to SVL ratio (df=1, F=9.72, p < 0.01; Table 2) and larger body size (SVL) (df=1, F=19.31, p < 0.001; Table 2). All the above variables (HLL to SVL, HFL to SVL ratios and SVL) used in the analysis were normally distributed for different species (Shapiro-Wilk test of normality P < 0.01). There are several overlaps in the meristic characters between different species of Sarada gen. nov. (Table 4). All the scale counts except midbody scale counts are within the range of the two male syntypes (BMNH 1946.8.27.39 and BMNH 1946.8.27.40) of Sarada deccanensis comb. nov. There are minor variations in range of body measurements (SVL, TRL, DWL and DWLT) of our male specimens and the syntype (BMNH 1946.8.27.39). The type locality of Sarada deccanensis comb. nov. is “Deccan”, which is a large area during British India. We assume Jerdon’s (1853 and 1870) observations are from (Jalna, Maharashtra), where he was posted (Elliot, 1873), hence we assign the species name Sarada deccanensis comb. nov. to this population.

Sitana spinaecephalus sp. nov., and Sitana laticeps sp. nov. can be differentiated from all other Sitana species in having only a blue line on the throat extending in the dewlap. Sitana spinaecephalus sp. nov. can be differentiated from Sitana laticeps sp. nov. in having a very large dewlap which extends far posteriorly along the trunk (up to 45%) compared to dewlap up to 29%. Sitana ponticeriana, Sitana visiri sp. nov., Sitana bahiri and Sitana devakai can be differentiated from all other Sitana species in having a strongly serrated dewlap margin. Sitana visiri sp. nov. can be differentiated from Sitana ponticeriana, Sitana bahiri and Sitana devakai in having a dewlap which extends far posteriorly along the trunk (up to 56%) compared to dewlap up to 37%, 29% and 33% respectively. Sitana ponticeriana can be differentiated from Sitana devakai and Sitana bahiri in having a dewlap which extends far posteriorly along the trunk (up to 37%). Sitana devakai can be differentiated from Sitana bahiri in having higher number of ventral scales (100-108) compared to 87-89, Sitana ponticeriana has a much lower ventral scale count (64-76) which is an additional character for identification. The three subadult male specimens we collected from Rameshwaram Island and Karur in Tamil Nadu have overlapping characters with other species of Sitana spp in clade 3. They are similar to Sitana bahiri and Sitana ponticeriana in having 47 and 51 around the body scales. They have a lower number (69 and 73) of ventral scales that falls within the range of Sitana ponticeriana and Sitana visiri sp. nov. counts, but they are lower than Sitana bahiri and Sitana devakai. Since we only have three specimens from this locality and the fact that they are all subadults it is difficult to identify the species. Due to their close genetic similarity to the sequence of the specimen from the type locality of Sitana bahiri (2% uncorrected p-distance in ND2 mtDNA; Fig. 2) from here onwards we will refer them as Sitana cf. bahiri. There are several overlaps in the meristic characters (scale counts) between all the species of Sitana examined in this study (Table 5).