Pseudophilautus dilmah, a new species of shrub frog (Amphibia: Anura: Rhacophoridae) from a threatened habitat Loolkandura in Sri Lanka

1,2,3,6 Herpetological Foundation of Sri Lanka, 31/5, Alwis Town, Hendala, Wattala, Sri Lanka 2 Ellangàwa Unity Care for Community and Nature, Hapugoda, Ambathenna, Sri Lanka 4,5,6 Institute of Biochemistry, Molecular Biology, Biotechnology, University of Colombo, Sri Lanka 1 boiga2000@gmail.com (corresponding author), 2 imeshnu1@gmail.com, 3 dulanrangavp@gmail.com, 4 kamani@ibmbb.cmb.ac.lk, 5 ranganath148@yahoo.com, 6 nemzy821@gmail.com

We have meticulously conducted explorations on amphibians in the last few years in the Central Hills of Sri Lanka including Peak Wilderness, Horton Plains and the Knuckles massif and subsequently describe a new species of Pseudophilautus from Loolcondera Estate ( Fig. 1), adjacent to the Galaha Forest Reserve in Deltota (07 0 17'N & 080 0 70'E) on the border of Kandy and Nuwara-Eliya districts of Central Province.
The British colonials pronounced the Sinhala native term "Loolkandura" as "Lool-conde-ra" (Jayasuriya et al. 1993), and the English spelling has remained unchanged till date, here we shall be using the former spelling for reasons discussed later. Loolkandura and Galaha forest reserve are located in the Central Hills, which lead to Piduruthalagala, bordering the two districts of Kandy and Nuwara Eliya. Today this forest area is under the Department of Forest Conservation. The first tea plant is said to have been planted here by James Taylor in 1867, and the first tea estate established after clearing large areas of pristine forests. The tea estate is surrounded by severely fragmented and disturbed patches of pristine lower montane rain forests belonging to the Galaha Forest Reserve. The Loolkandura forest is surrounded by mostly tea estates, and a thin strip of forest connects with the Galaha Forest Reserve, while an entirely separated, large natural forest patch is located within the Loolkandura Tea Estate. The vegetation within the Loolkandura forest area can be grouped into six types, viz: natural forest, secondary forest, natural grassland, tea plantation, abandoned tea plantation, and pinus plantation.
We here describe a new species of shrub frog Pseudophilautus dilmah, from the Loolkandura forest of the Central Province of Sri Lanka. The species is reported as a result of an ongoing island-wide survey documenting the diversity of herpetofauna in Sri Lanka.

MATERIALS AND METHODS
Sampling was done mostly nocturnally, and photographs of most specimens were taken in the wild to avoid any confusion of change in colour after captivity and to avoid over collection of the same species. The specimens were photographed in life (Canon EOS 7D, Canon 100mm F/2.8 IS USM Macro Lens, Canon MT-24EX Macro Twin Lite Flash with Vello Bounce Dome Diffuser). Geographical coordinates were determined from GPS readings (Gamin eTrex Vista) at the locality. Specimens collected in the field were first fixed in 90% ethanol for two hours and stored in 70% ethanol. The material referred to is deposited Forty-four external measurements (Table 1) of specimens were taken with a Mitutoyo digital vernier calliper to the nearest 0.1mm, and observations were made through a Leica M50 (10-40) microscope. Nomenclature of external anatomy abbreviated in the text and external measurements taken are listed alphabetically herein: maximum breadth of disk of the third finger (DB), taken from anterior to posterior edge of circum-marginal groove; distance between back of eyes (DBE), measured between posterior edge of eyes; distance between front of eyes (DFE), measured between anterior edges of eyes; length of disk of the third finger (DL), taken from anterior edge of circummarginal groove to posterior edge of disk; width of disk of the third finger (DW), measured across the inner and outer edges of circum-marginal groove; eye diameter (ED), horizontal diameter of eye; eye to nostril distance (EN), measured between anterior most point of eye and middle of nostril; eye to snout distance (ES), measured between anterior most point of eye and tip of snout; thigh (femur) length (FEL), distance between vent and knee with both thigh and shank flexed; first finger length (FL-1), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of first finger; second finger length (FL-2), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of second finger; third finger length (FL-3), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of third finger; fourth finger length , measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of fourth finger; foot length (FOL), distance between heel and tip of fourth toe with both foot and shank flexed; groin to knee distance (GK), measured between groin and the tip of knee; head depth (HD), distance between the apex of eye and the bottom of jaw; head length (HL), distance between angle of jaws and snout tip; head width (HW), measured across angle of jaws; inner metatarsal tubercle length (IML), distance between the distal and proximal edges of the inner metatarsal tubercle; internarial distance (IN), least distance between the inner margin of nares; interorbital width (IO), least distance between the upper margins of orbits; knee-angle length (KT), distance from knee-angle to tibio-tarsal articulation; lower arm length (LAL), taken from elbow to posterior-most margin of inner palmar tubercle; mandible-back of eye distance (MBE), distance between angle of jaws and posterior-most point of eye; mandible-front of eye distance (MFE), taken as distance between angle of jaws and anterior-most point of eye; mandible-nostril distance (MN), taken as distance between angle of jaws and middle of nostril; palm length (PAL), taken from posterior-most margin of inner palmar tubercle to tip of disk of third finger; snout-nostril distance (SN), taken as distances between middle of nostril and tip of snout; snout-vent length (SVL), measured from tip of snout to vent; tympanumback of eye distance (TBE), distance between anteriormost point of inner margin of inner rim of tympanum and posterior most point of eye; tarsal length (TAS), measured between the tibio-tarsal articulation and anterior edge of inner metatarsal tubercle; tibia length (TBL), distance between knee and heel with both shank and foot flexed; first toe length (TL-1), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of first toe; second toe length (TL-2), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of second toe; third toe length (TL-3), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of third toe; fourth toe length (TL-4), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of fourth toe; fifth toe length (TL-5), measured between proximal margin of the most proximal subarticular tubercle or crease of articulation and the tip of fifth toe; tympanum-nostril distance (TND), distance between anterior-most point of inner margin of inner rim of tympanum and middle point of nostril; tympanum-front of eye distance (TFE), taken as distance between anterior-most point of inner margin of inner rim of tympanum and anterior-most point of eye; tympanum height (TYH), vertical diameter of the inner rim; tympanum width (TYW), horizontal diameter of the inner rim; upper arm length (UAL), distance between axilla and elbow; upper eyelid width (UEW), measured from bony edge of supraorbital to outer edge of upper eyelid; vent to knee-angle length (VKL), distance measured from knee-angle to vent. Comparison was carried out amongst congeners which closely resembled the new species. In the comparison tables for morphological characters (Table  2), and characters in the skin (Table 3), an empty cell in the table denotes either the same character as that of the new species, or for few extinct species when there was no data available.
Additional data for comparison were collected

Molecular phylogentic analysis
The phylogenetic relationships of the newly described species to other species of the genus Pseudophilautus in Sri Lanka, for which data is available, were estimated using the mitochondrial 16S ribosomal RNA region. Whole cellular DNA was extracted from ethanol preserved muscle tissue sample from the specimen number DWC 2015.01.001 of the new species and NMSL 2013.01.01 NH of P. bambaradeniyai. The tissue sample was digested at 55 0 C using STE buffer (0.1 M NaCl, 0.01 M TRIS, EDTA 0.001 M and 25 μl of 10% SDS) with 10μl of 20 mg/ml Proteinase K. Final extraction was carried out using the phenol-chloroform method. Polymerase chain reaction (PCR) amplifications of the 16S ribosomal RNA region were carried out in 25 μl reactions using the forward primer-5'-GCCTGTTTATCAAAAACAT-3' and reverse primer 5'-CCGGTCTGAACTCAGATCACGT-3' (Frost et al. 2006) with Promega GoTaq DNA polymerase enzyme in an Eppendorf Master cycler. The thermal profile was an initial denaturation step for seven minutes at 94 0 C, and 35 cycles of denaturation for 40 seconds at 94 0 C, annealing for 30 seconds at 52 0 C and extension for 30 seconds at 72 0 C, followed by final extension for 10 min at 72 0 C. Amplified DNA fragments were purified using the 'Promega Wizard Gel and PCR clean up' system. Purified PCR products were sequenced using a BigDye Terminator sequencer v. 3.0 (Applied BioSystems, CA, USA) in 8 μl reactions. The sequence chromatograms were visualized in a Genetic Analyzer 3500 Dx (Applied Biosystems, CA, USA).
The newly generated sequences have been deposited in GenBank under the accession numbers KP272046 and KP272047. Additional sequences of the Sri Lankan Pseudophilautus species were obtained from GenBank (Appendix 2). Gene sequences were aligned using MUSCLE (Edgar 2004). Molecular phylogenetic analysis was performed using MEGA 6 (Tamura et al. 2013). The best fit model for nucleotide substitution was selected from 24 models using MEGA 6 based on the minimum Bayesian Information Criterion (BIC) value (Schwarz 1978;Nei & Kumar 2000). Best fit model was used for computing the pairwise distances between the sequences and for maximum likelihood analysis. Maximum likelihood tree was rooted with Raorchestes charius, an Indian species that represent Raorchestes, the sister group to the Sri Lankan genus Pseudophilautus. Tree was drawn using FigTree v1.4.2 (Morariu et al. 2009). Diagnosis: Body rather small size (SVL 19-20 mm). Snout rounded in lateral aspect, bluntly pointed in dorsal and ventral aspect. Canthus rostralis rounded. Vomerine teeth, lingual papilla and nuptial pads absent. Dermal fringe distinct on inside of fingers III and IV, small blunt tubercles on metacarpal and ulnar folds. Toes basally webbed. Interorbital area smooth. Upper eyelid prominent tubercles present. Anterior and posterior dorsum without horny spinules but tubercles present. Upper part of flank weakly granular. Supratympanic fold distinct. Prominent small calcar present at the distal end of the tibia. Throat granular, chest and belly coarsely granular.

Skin of dorsal and lateral body
Dorsal and lateral regions of snout smooth. Prominent tubercles between eyes and on upper eyelid. Side of head smooth. Median dermal ridge absent. Anterior and posterior part of back with tubercles, without horny spinules. Upper part of flank weakly granular. Lower part of flank granular. Latero-dorsal folds, lateral line system, fejervarya line, cephalic ridges, and co-ossified skin absent. Supratympanic fold distinct. Upper arm, lower arm, and foot smooth. Anterior, dorsal, and posterior part of thigh smooth. Leg, and tarsus smooth with few tubercles. Prominent small calcar present at the distal end of the tibia. Hand smooth.

Skin of ventral part
Throat granular, margin of throat smooth with prominent tubercles. Chest and belly coarsely granular. Upper arm granular. Lower arm smooth. Thigh smooth. Leg smooth. Tarsus smooth. Vent anteriorly smooth, either side weakly granular. Macroglands absent.

Colour in life
Dorsum cream with light brown patches, dark brown cross band between eyes, pair of dark brown dots placed behind the cross band, a pair of broad light brown longitudinal bands ending at the sacrum, dark brown blotches on groin, lateral body lighter with a light olive greenish tinge; limbs cream, fore limbs, hind limbs, fingers and toes with light brown cross bands; ventral side belly off white with light brown blotches, chest off white with few blotching, throat uniform off white, hands, feet and webbing lighter (Image 1).

Colour in alcohol
Colour pattern remains with a little darkening, ventral side off white with dark brown blotching.

Colour variation
Dorsal colour varies ranging to a lighter brownish tinge, and at the ventral side belly shows prominent dark blotches.

Etymology
The species epithet dilmah is named after Dilmah Conservation, for its dedicated efforts to biodiversity conservation on the Island. Dilmah is treated as an invariable noun in apposition to the generic name.

Suggested vernacular names
Dilmah panduru madiya, and Dilmah Shrub Frog in Sinhala, and English respectively.

Molecular analysis
The final alignment consisted of a total of 590 base pairs (bp). Model test identified Tamura & Nei (1993) nucleotide substitution model with gamma distribution (BIC = 7964.99, lnL = -3360.55, G = 0.26) as the best-fit model. Maximum likelihood analysis of the 16S rRNA fragment (Fig. 2) places Pseudophilautus dilmah sp. nov. in a monophyletic group along with P. schmarda (Kelaart 1854) and P. hankeni Meegaskumbura & Manamendra-Arachchi 2011. Pairwise genetic distance between P. dilmah sp. nov. and P. schmarda was 1.9% while it was 1.6% between the former and P. hankeni. There was a 10% divergence in the 16s rRNA gene between P. dilmah sp. nov. and P. bambaradeniyai Wickramasinghe et al. 2012, the morphologically most closely related species.

Comparison
A critical comparison was made amongst all congeners, and   Table 3 provides discriminating characters of the skin and it was evident that Pseudophilautus dilmah sp. nov., resembles P. bambaradeniyai (Image 5), due to similar colour pattern, comparable adult body size, and because both species are found in the Central Hills. Pseudophilautus dilmah, can be easily differentiated from P. bambaradeniyai by the absence of horny spinules on the anterior dorsum (vs present), snout rounded in lateral aspect (vs truncate), fringe on fingers present (vs absent), and calcar present (vs absent). Although P. dilmah sp. nov., is genetically most closest to P. hankeni (Image 6) and P. schmarda (Image 7), with genetic distances of 1.6% and 1.9% respectively, the species is morphologically distinct and can be P. mittermeieri (Meegaskumbura & Manamendra-Arachchi 2005), and P. singu (Meegaskumbura et al., 2009), because of similar morphological characters, hence these species were included at the beginning of the comparison tables.

DISCUSSION
Pseudophilautus dilmah sp. nov., appears to be more resembling to P. bambaradeniyai, (Fig. 2) than to any other species of the same genus. But considering molecular evidences P. bambaradeniyai has been placed in a well separated clade with P. frankensbergi (Meegaskumbura & Manamendra-Arachchi 2005) with a pairwise genetic distance of 10%, which suggests a species level divergence. Although they were both found from the Central Hills they are allopatric. Loolkandura the type locality of P. dilmah sp. nov., is positioned towards the northern tip of the Central Hills where as Peak Wilderness the type locality of P. bambaradeniyai, and the only locality it is found in, is positioned towards south-west of Loolkandura (Fig. 1).
Although P. dilmah sp. nov., is genetically most closest to P. hankeni and P. schmarda, with genetic distances of 1.6% and 1.9%, respectively, the species is morphologically distinct and can be separated by the characters mentioned in the comparison. Furthermore P. hankeni is distributed in the Knuckles massif which is geographically well separated, and although P. schmarda is distributed in the Central Hills they are allopatric (Fig.  1). Although 3% genetic distance is a good indication Vences et al. (2005) mention that interspecies genetic distances could be from 1% to 10% and allopatric species are known with less than 3% differences. The pairwise distance for P. pleurotaenia (Boulenger 1904) and P. hoipolloi (Manamendra-Arachchi & Pethiyagoda 2005), for P. asankai (Manamendra-Arachchi & Pethiyagoda 2005) and P. hoffmanni (Meegaskumbura & Manamendra-Arachchi 2005), for P. decoris and P. mittermeieri, were 0.5%, 1%, and 1.4%, respectively. All of which are considered valid morphologically yet have a lower genetic distance than 1.6% for P. dilmah and P. hankeni, and 1.9% for P. dilmah and P. schmarda, hence our current genetic distances from its sister taxa suggests species level divergence.
Although to a lesser extent considering morphological characteristics P. dilmah may be confused with P. decoris; distributed in the Rakwana range, pairwise genetic distance of 18.3%; with P. mittermeieri; distributed in the lowland rain forests to lower montane rain forests, pairwise genetic distance of 17.3%, and with P. singu; distributed in the lowland rain forests, pairwise genetic distance of 10.1%, considering their geographic distribution and molecular divergence the new species is well separated from all the above three species.
The natural forest cover in the type locality is fast declining mainly due to encroachment by tea plantations (Images 8-10), which has evidently led to forest fragmentation (Image 11). In order to protect the species from habitat loss, the few remaining areas which is home to the species needs legal protection. It is hoped with the discovery of the species in this locality the authorities will take necessary measures to protect the natural forest cover which is home to the newly described species P. dilmah.

Proposed conservation status
The species was recorded from an elevation of about 1300m, and commonly seen perched on 1-2m high bushes. The specimens were found in natural forest cover as well as in the disturbed areas, with no canopy cover and in areas with regenerated forest covers. The area which the species was reported is currently under severe anthropogenic pressures. Since the colonial period the area was subjected to clearing of pristine forest covers for coffee, cinchona and later tea and cardamom plantations. A few remnant patches of unique "mid elevation lower montane forest", which shows the mixed characteristics of montane forests and lower montane forests, exist in some areas. Especially, natural forests remain only in the tough terrains where colonials could not spread the plantations and most of the stream banks are also severely affected due to encroaching tea cultivations (Images 8-11).
The proposed conservation status for the species according to the IUCN Red List Criteria can be considered Critically Endangered (criteria B1ab(iii)). Since the extent of occurrence is <100km 2 , is recorded from a single location, and their habitats under sever threat.