Indirana chiravasi , a new species of Leaping Frog ( Anura : Ranixalidae ) from Western Ghats of India

1 Department of Zoology, Abasaheb Garware College, Karve Road, Pune, Maharashtra 411004, India 2 Department of Biodiversity, Abasaheb Garware College, Karve Road, Pune, Maharashtra 411004, India 3 Indian Institute of Science Education and Research (IISER), G1 Block, Dr. Homi Bhabha Road, Pashan, Pune, Maharashtra 411008, India 3 Systematics, Ecology and Conservation Laboratory, Zoo Outreach Organization (ZOO), 96 Kumudham Nagar, Vilankurichi Road, Coimbatore, Tamil Nadu 641035, India 1 anand.padhye@mesagc.org (corresponding author), 2 nikhilsmodak@gmail.com, 3 n.dahanukar@iiserpune.ac.in

The true species diversity of Indirana within the Western Ghats, however, is poorly understood as several species are suggested to be in species complexes (Nair et al. 2012a,b;Modak et al. 2014).With the presence of undescribed species (Nair et al. 2012a,b;Modak et al. 2014) and poor knowledge on the distribution of known species (Modak et al. 2014), the knowledge on Indirana is plagued by both Linnean and Wallacean shortfalls (Bini et al. 2006).Because several species of Indirana are under threatened category (Stuart et al. 2008), taxonomic studies on this group are of immediate concern.
While studying the diversity and distribution of species under the genus Indirana, we came across a population showing marked difference in morphology from its congeners and genetically distinct from species for which genetic data is available.This population is described as a new species of Indirana from Amboli, in Western Ghats of southern Maharashtra, India.

Study site and specimen collection
Specimens of the new species were collected from Amboli (15.956 0 N & 73.997 0 E; 744m), Sindhudurg District, Maharashtra, India.Five male and two female specimens were collected and preserved in absolute alcohol for further analysis.Two specimen of Indirana gundia were collected from non-protected area of Gundia, Karnataka, for genetic analysis.Collection of specimens was kept at the minimum and all collections were made following the IUCN (2008) guidelines for use of threatened species in research.

Morphometry
Morphometry was done with the help of a digital caliper (Ocean Premium measuring instruments) to the nearest 0.1mm.Totally, 27 characters were selected following Biju et al. (2011) and Dubois & Ohler (1999) with some modifications for morphometry, viz.-SUL (Length of specimen from snout to the visible tip of urostyle); HL (head length: from the posterior border of tympanum to the tip of snout); HW (head width: width of head between to posterior borders of tympanum); SL (Snout Length: from the anterior orbital border to the tip of snout); EL (Eye Length: length of eye between two orbital borders); TYL (maximum tympanum length); UEW (upper eyelid width); SNL (snout to nostril distance); ENL (eye to nostril distance); INL (inter-narial distance); IOL (inter-orbital distance: minimum distance between two eyelids); UAL (Upper arm length); FoAL (Fore-arm Length); F1 to F4 (Finger 1 to Finger 4 length from the base of the sub-articular tubercle); F3D (finger three disc width); F3W (finger three width at the base of disc); THL (thigh length); TL (Tibia/shank length); ACL (Astragalocalcaneal length); FOL (Foot length: from the base of the inner metatarsal tubercle to the tip of the fourth toe); TFOL (Total foot length: from the tibio-tarsal articulation to the tip of fourth toe) and T1 to T5 (Toe1 to Toe5 length from the base of the respective sub-articular tubercle).Webbing formula was determined following the method provided by Savage & Heyer (1967) with modifications by Myers & Duellman (1982).

Statistical analysis
Statistical analysis of the morphometric data was performed on size adjusted measurements by taking all measurements as percent of SUL.The null hypothesis that the data is multivariate normal was checked using Doornik & Hansen (2008) omnibus.MANOVA/CVA was performed to understand whether related species of Indirana form significantly different clusters (Huberty & Olejnik 2006).Pillai's trace statistic was performed to find the significant difference between the clusters (Harris 2001).Statistical analysis was performed in PAST (Hammer et al. 2001).

Molecular analysis
Two specimens of the new species (BNHS 5890 and WILD-14-AMP-489) and two specimens of topotypic Indirana gundia (WILD-14-AMP-499 and WILD-14-AMP-500) were used for molecular work.Genomic DNA was extracted from thigh muscle tissue.The tissue was digested at 55 0 C using STE buffer (50 mM Tris-HCl, 20 mM EDTA and 50μl of 10%SDS) with 10μl of 20mg/ ml Proteinase K. RNase treatment was given for two hours at 37 0 C. Final extraction process was done using phenol-chloroform method.Polymerase Chain Reaction was performed for amplification of two mitochondrial genes (12S and 16S) and two nuclear genes (rho and rag 1) (Table 1).In addition to the DNA extracted in the current study, we also used DNA extracted in a previous study (Modak et al. 2014) for amplification of Rag1 gene from two topotypes of Indirana leithii BNHS 5590 and BNHS 5591.PCR reaction was performed in a 25μl reaction volume containing 5μl of template DNA (~200ng), 2.5μl of 10X reaction buffer (100 mM Tris pH 9.0, 500 mM KCl, 15 mM MgCl2, 0.1% Gelatin), 2μl of 25 mM MgCl2, 1μl of 10 mM dNTPs, 1μl of each primer, 1μl Taq polymerase and 16.5μl nuclease free water.The thermal profile was 10 minutes at 95 0 C, and 35 cycles of 1 minute at 94 0 C, 1 minute at respective annealing temperature for 12S, 16S, rho and rag1 primers (Table 1) and 2 minutes at 72 0 C, followed by extension of 10 minutes at 72 0 C. Amplified DNA fragments were purified using the 'Promega Wizard Gel and PCR clean up' system and sequenced.The purified PCR products were sequenced using ABI prism 3730 sequencer (Applied Biosystems, USA) and Big dye terminator sequencing kit (ABI Prism, USA).Sequences were analyzed by BLAST tool (Altschul et al. 1990).These sequences have been deposited in GenBank (Accession numbers KM386526-KM386543).
Additional sequences of related species were retrieved from NCBI GenBank database (http://www.ncbi.nlm.nih.gov/).GenBank accession numbers for the sequences used for the analysis are provided in Appendix A. Gene sequences were aligned separately using MUSCLE (Edgar 2004) implemented in MEGA 6 (Tamura et al. 2013) and were concatenated to make a combined matrix of 1342 nucleotides.Best fit model for nucleotide substitution was selected in TOPALi v2.5 (Milne et al. 2004) based on minimum Akaike Information Criterion (AIC) value (Posada & Crandall 2001).The phylogenetic relationships were inferred by maximum likelihood analysis using RAxML (Stamatakis 2006) implemented in TOPALi v2.5 (Milne et al. 2004).Reliability of the phylogenetic tree was estimated using bootstrap values run for 1000 iterations.

Diagnosis
Indirana chiravasi sp.nov.differs from all other congeners based on the following combination of characters: medium-sized (24.7-39.2mm SUL) frog, with median single internal vocal sac, head longer than wide, distinct canthus rostralis, tympanum 3/4 th the diameter of eye in males and 2/3 rd in females, vomerine teeth in two oblique series at the posterior border of choanae, long midventral lingual papilla, first finger longer than or equal to second, presence of double outer palmer tubercle, thin and elongated inner metatarsal tubercle present, outer metatarsal tubercle absent, webbing moderate and differs in male (I1-2II1-3III1¼-3IV3-1¼V) and female (I1-2II1-2½III1¼-3IV3-1¼V) by having reduced webbing on the third toe in males as compared to females, dorsally skin with glandular folds but without warts, ventrally skin smooth without mottling and palms and soles dark brown.

Description
General appearance of holotype as in Image 1 and of female paratype as in Image 2. Morphometric details as in Table 2.

Description of the holotype (BNHS 5888; male) (all measurements in mm)
Medium-sized (SUL 27.3); head longer than wide (HL 11.4 > HW 9.8); snout longer than horizontal diameter of eye (SL 4.8 > EL 3.7); pupil horizontal; outline of snout rounded in shape dorsally (Image 1a), truncated laterally (Image 3a); ventrally snout protruding beyond the mouth (Image 3c); nostrils nearer to snout than to the eye (SNL 2.0 < ENL 2.7); tympanum about 3/4 th the diameter of eye, very close to eye; supra-tympanic fold distinct; upper eyelid width slightly more than half the horizontal diameter of eye; upper eyelids bear very few granulations; inter-narial distance almost equal to the inter-orbital distance (INL 2.6 ≈ IOL 2.6); canthus rostralis obtuse; loreal region slightly concave and oblique; vomerine teeth in two slightly oblique rows at the posterior border of choanae; tongue thin, bifid; bear a long mid ventral papilla.
Dorsal and ventral skin smooth; few longitudinal folds on dorsal side; lateral side granular.

Description of female paratype (BNHS 5889) (all measurements in mm)
Medium-sized (SUL 39.2); with head longer than wide (HL 14.9 > HW 13.9); snout longer than eye (SL 6.5 > EL 5.0); outline of snout rounded in shape dorsally (Image 2a); truncated laterally (Image 3b); ventrally protruding beyond the mouth (Image 2b); nostrils nearer to snout than to the eye (SNL 2.6 < ENL 3.9); tympanum about 2/3 rd the diameter of eye, slightly more apart from eye than in male; supra-tympanic fold distinct; upper eyelid width slightly more than half the horizontal diameter of eye; upper eyelid bearing very few granulations; inter-narial width slightly narrower than inter-orbital distance (INL 4.1 < IOL 4.2); canthus rostralis obtuse; loreal region slightly concave and oblique; vomerine teeth in two slightly oblique rows at the posterior border of choanae; tongue thin, bifid; bearing a long mid-ventral papilla.
Dorsal and ventral skin smooth; few longitudinal folds on dorsal side; lateral side granular.

Colouration
Male: Dorsal colour grey to brown in preservation (Image 1) and olive brown with scattered yellow markings in living specimen (Image 6) with W-shaped marking comprising of densely organized black spots; black strip running from tip of snout to shoulder through eye and tympanum; creamy white band across head between upper eyelids usually present followed posteriorly by a dark band; in life, olive green to brown band running from above shoulder to groin (Image 6); forelimb bearing transverse bands also on fingers; palm dark brown in colour at least in live condition; brown bars on thigh, tibia, outer side of foot and dorsal surface of toes; ventral and inner side of foot dark brown in colour up to tibiotarsal articulation; ventrally white with few melanophores visible only under magnification.
Female: General appearance as in male.Dorsally more pale than male without W-shaped marking (Image 2); darker flanks; limbs with dark cross bars in life, faded in preservation.

Sexual Dimorphism
Tympanum about 3/4 th the diameter of eye in male and 2/3 rd the diameter of eye in female.Tympanum very close to eye in male and slightly farther apart in female

Etymology
The specific epithet, a combination of words 'chir' (singular) or 'chira' (plural) which means crevice or crevices in Marathi and 'vasi' in Sanskrit means 'inhabitant of', which emphasizes crevice dwelling habit of this species.The specific name is noun in apposition.

Distribution
The species is currently known only from its type locality at Amboli (15.956 0 N & 73.997 0 E; 744m), which is a small hill station in the southwestern Maharashtra, India (Fig. 1).

Habitat
The species occupies lateritic rocky outcrops (Image 7a).It is often found in the crevices of the laterite boulders (Image 7b).Males were mostly seen while calling from the wet rocks or boulders covered with mosses.Females were collected from under the log in the forest and from under the roadside stone.

Natural history and description of tadpoles
Eggs were seen laid under the mosses on lateritic wet rocks and boulders (Image 8).Hatchlings remain at the egg laying site (Image 9a).Embryos, hatchlings and tadpoles of two different stages were observed in the same habitat.Unhatched eggs (Images 8a,b) show the embryos with external gills, parallel to stage 20 (Gosner 1960).Tadpoles were seen on wet boulders feeding on algal matter (Image 9).Image 9a shows hatchlings, parallel to stage 25 (Gosner 1960).Image 9b shows tadpole in its terrestrial habitat with fully developed hind limbs (without forelimb), parallel to stage 40 (Gosner 1960).Image 9c shows terrestrial tadpoles with long, finless tail which is not under resorption, oral apparatus and fully developed forelimbs as well as hind limbs, parallel to stage 42 (Gosner 1960).Image 9d shows stage 44 (Gosner 1960) with mouth beneath eye and greatly reduced tail, while image 9e shows stage 46 (Gosner 1960) -a completely metamorphosed froglet.Tadpoles showed semi-condensed individual keratodont formula (Dubois 1994) as 4[A 1 -A 4 ]/4[P 1 -P 2 ].The oral apparatus is divided into two lateral parts by large horny beak.The first anterior keratodont ridge A 1 is divided while three succeeding anterior keratodont ridges A 2 -A 4 are placed lateral to the horny beak.On the posterior labia first keratodont ridge P 1 is marginal, present on the either sides, while the second one -P 2 is placed lateral to the horny beak.Third and fourth keratodont ridges -P 3 and P 4 are continuous.Although, the ridge P 3 appears to be divided into four subunits the keratodont rows are continuous on it (Image 10).and I. phrynoderma (Fig. 2a).Indirana chiravasi sp.nov.was separated from I. brachytarsus, I. gundia and I. semipalmata on the third canonical axis (Fig. 2b).CVA loadings of morphometric characters on the first three canonical axes are shown in Table 3. Characters such as ENL, INL, F3, F4, TL, ACL, FOL, T1 and T2 separated Indirana chiravasi sp.nov.from other related species.

Genetic analysis
Concatenated genetic sequences of mitochondrial 12S and 16S genes and nuclear rho and rag1 genes had total 1342 bases.Best fit model for the nucleotide substitution was general time reversal model with gamma distribution and invariant sites (GTR+G+I, AIC = 10516.66,lnL = -5157.33,G = 0.39, I = 0.34, df = 101, n = 1342).Maximum likelihood analysis of the genetic data (Fig. 3) suggested that Indirana chiravasi sp.nov.formed a monophyletic group genetically distinct from the other Indirana species for which genetic data are available.
© Hemant Ghate 3/4 th (in male) the diameter of eye (vs.tympanum half the diameter of eye in I. longicrus).Raw genetic distance between I. chiravasi and I. leptodactyla (as identified by Nair et al. 2012b) is 23.9% for 16S gene and 12.3-12.7 % for all the genes together; between I. chiravasi and I. diplosticta (as identified by Nair et al. 2012b) is 9.1-11.0% for 16S gene and 4.2-4.7 % for all the genes together; and between I. chiravasi and I. leithii (topotypic material from Modak et al. 2014) is 12.5-12.8% for 16S gene and 6.9-7.3 % for all the genes together.
The whereabouts of the type specimen of Indirana tenuilingua described by are not known and is suggested to be lost (Dubois 1984).Therefore, for the comparison of the new species with I. tenuilingua, we have relied on the original description by Rao (1937).Indirana chiravasi differs from I. tenuilingua in having head longer than broad (vs.head slightly wider than long), inter-orbital

Percent variation Eigenvalue
Indirana chiravasi sp.nov.distance equal to or wider than inter-narial distance (vs.interorbital width more than twice the distance between the nostrils) and toes and fingers with deep semicircular groove (vs.semicircular groove in front of the toes and fingers absent, faint or indistinct).
Indirana chiravasi can be distinguished from I. brachytarsus in having few longitudinal folds on dorsal side (vs.numerous longitudinal folds on dorsal side), thin, shovel-shaped inner metatarsal tubercle (vs.long and stout inner metatarsal tubercle) moderate webbing (vs.extensive webbing, webbing formula, I1-2II1-21/2III1-3IV3-1V) and having broader head .Furthermore, if the identification of I. brachytarsus by Nair et al. (2012b) is correct then the two species are also genetically quite distinct (Fig. 3).Raw genetic distance between I. chiravasi sp.nov.and I. brachytarsus (as identified by Nair et al. 2012b) is 11.6% for 16S gene and 8.0-8.3 % for all the genes together.
Indirana chiravasi sp.nov.differs from I. gundia in the most prominent feature of having a single internal vocal sac (vs.bilateral vocal sacs).The presence of bilateral vocal sac in I. gundia not only appears in the original description (Dubois 1986) but they can also be seen in two patches on the ventral side of the head (Image 3f).Furthermore, I. chiravasi sp.nov.differs from I. gundia in having the webbing formula I1-2II1-3III1¼-3IV3-1¼V in male and I1-2II1-2½III1¼-3IV3-1¼V in female (vs.webbing formula, I1-2II1-2⅟2III1-3IV3-1V, in both males and females) and inner metatarsal tubercle thin shovelshaped (vs.distinct and stout).Morphometrically (Fig. 2) both the species form significantly distinct clusters (Fisher's distance = 3.142, P = 0.004).Raw genetic distance between I. chiravasi sp.nov.and I. gundia is 5.2-5.5 % for 16S gene and 2.8-3.1% for all the genes together.

DISCUSSION
Indirana chiravasi sp.nov.forms the eleventh species in the Western Ghats endemic genera Indirana and monotypic family Ranixalidae.Phylogenetic analysis of Indirana species based on two mitochondrial and two nuclear genes suggests that, I. chiravasi forms a monophyletic group with topotypic I. gundia and five specimens (IND/AA/DD/CC 200, 220, 227, 230 and 231) of I. beddomii species complex (as identified by Nair et al. 2012b).
Both morphologically and genetically, Indirana gundia is one of the close congeners of I. chiravasi.However, apart from the morphological variations, there was a significant difference in the multivariate morphometric analysis of the two species.Further, the two species are separated by a genetic difference of 5.2-5.5 % for 16S gene.Based on the suggestions of Vences et al. (2005) the high genetic divergence in 16S gene validates that I. chiravasi is a distinct species.Further, based on the current distribution records, these two species are separated by a geographical distance of more than 400km.Genetic data for I. gundia is provided for the first time in the current study.This data is based on topotypic material (Image 11), for which the species identity is confirmed by its morphological comparison with type series as well as morphometric analysis.The two specimens of I. gundia did not form a monophyletic group, although there was only 0.3% genetic distance considering all genes together and 0.2% genetic distance in 16S gene.This could be attributed to small sample size (only two specimens) in the current study and further genetic analysis with additional samples may reveal within species variation in I. gundia.
Based on the phylogenetic analysis (Fig. 3) five specimens (IND/AA/DD/ CC 200,220,227,230 and 231) of Indirana beddomii species complex (as identified by Nair et al. 2012b) form a monophyletic group and fall between I. chiravasi and I. gundia.This cluster is separated from I. chiravasi by raw genetic distance of 3.5-3.9% and from I. gundia by 3.6-3.9% in 16S gene.However, because of the lack of information Image 11. Topotypic Indirana gundia male (a-e) and female (f-j) in life.WILD-14-AMP-499 (Male) and WILD-14-AMP-500 (female), Gundia, Karnataka.Recent documentation of primitive breeding in Indirana species from Amboli by Gaitonde & Giri (2014) actually refers to the breeding biology of I. chiravasi based on the photographs provided by Gaitonde & Giri (2014).
Further, according to our extensive survey records, there is no other species of Indirana present in Amboli.Gaitonde & Giri (2014) have provided a detailed account of courtship and breeding bahaviour of the species along with the fertilization success.However, they do not provide the information on developmental stages.
Our observations, therefore, complete the information on breeding and developmental cycle of the species.Although we have provided developmental stages of this species parallel to the stages suggested by Gosner (1960), it is essential to note that the development and metamorphosis of the genus Indirana is different from the general ranid pattern, as the tadpoles do not have tail fins and are semi-terrestrial (Dubois 1986;Gaitonde & Giri 2014;Modak et al. 2014).The developmental stages of Indirana chiravasi sp.nov.differs from the stages in Gosner (1960) as follows.The unhatched eggs with embryo of I. chiravasi sp.nov.showed presence of external gills.This stage is parallel to stage 20 of Gosner (1960).However, hatching does not occur at this stage.It should be noted that in I. chiravasi, hatching does not occur until the development of operculum, which is a stage parallel to stage 25 of Gosner (1960).Further, the semi terrestrial tadpole with oral apparatus, full tail without any regression and fully developed forelimb (stage 42 of Gosner 1960) persists for a long duration.In I. chiravasi, tadpole of this stage continues to feed in its terrestrial habitat for around a month after which metamorphosis completes.This is unlike the classical metamorphosis described by Gosner (1960) where in on complete development of fore limb there is onset of metamorphosis with changes such as the beginning of tail regression, oral apparatus starts disappearing leading to formation of adult like mouth and metamorphosis is essentially completed within a short duration.These subtle differences in the development of I. chiravasi from the pattern provided by Gosner (1960), calls for the detailed study of developmental patterns in Western Ghats endemic genera such as Indirana.
Description of the new species Indirana chiravasi, and previous suggestions that there are species complexes and undescribed species (Nair et al. 2012a(Nair et al. , 2012b;;Modak et al. 2014), suggests that especially for an endemic and monotypic family such as Ranixalidae, the Western Ghats are subject to Linnean shortfall (Bini et al. 2006).Out of the 10 known species of this genus, Indirana gundia and I. phrynoderma are Critically Endangered; I. brachytarsus, I. leptodactyla and I. diplosticta are Endangered; I. leithii is Vulnerable; I. longicrus and I. tenuilingua are Data Deficient; and I. beddomii and I. semipalmata are Least Concern (IUCN 2014).With high proportion of threatened species within this endemic family, there is immediate concern for conservation initiatives.Therefore, further studies to resolve taxonomic status of species complexes and understanding their distribution patters are essential.Moreover, additional information on ecology and natural history would help in designing conservation strategies.

Image 5 .
Foot of (a) Indirana chiravasi sp.nov.holotype and (b) I. chiravasi sp.nov.female paratype.© Neelesh Dahanukar.a b (Image 3).Inner metatarsal tubercle thin and shovelshaped in male while slightly thicker and long in female (Image 5).Webbing formula I1-2II1-3III1¼-3IV3-1¼V in male and I1-2II1-2½III1¼-3IV3-1¼V in female.Breeding males show single internal vocal sac that is visible only during calling (Appendix B); bearing nuptial pad on the outer side of first finger and femoral glands on thighs; mature eggs visible from transparent latero-ventral skin of breeding females.

Figure 2 .Figure 3 .
Figure 2. Multivariate Analysis of Variance -Canonical Variates Analysis (MANOVA-CVA) of morphometric data of Indirana species.(a) CVA on first two canonical axes, scree plot is shown in the inset, and (b) CVA on first three canonical axes for morphometrically closely related species to Indirana chiravasi sp.nov.
locality as well as details of museum deposition, we could not compare these specimens with I. chiravasi.Nevertheless, it should be noted that both I. chiravasi and I. gundia are morphologically distinct from any of the syntypes of I. beddomii.Despite of the above mentioned lacunae,Nair et al. (2012b) have provided genetic data for a wide distribution of Indirana populations in southern Karnataka, Kerala and Tamil Nadu.Further,Modak et al. (2014) have provided details for Indirana leithii based on topotypic material and a wide distribution in the northern Western Ghats.As a result genetic data fromNair et al. (2012b) andModak et al. (2014) forms a good comparative material for assessing the affinities between I. chiravasi and other populations of Indirana sp.nov.from a wide geographic range within the Western Ghats.