A new species of Odorrana (Anura, Ranidae) from Hunan Province, China

Abstract A new species, Odorrana sangzhiensissp. nov., is described, based on five specimens from Sangzhi County, Zhangjiajie City, Hunan Province, China. Molecular phylogenetic analyses, based on mitochondrial 12S rRNA and 16S rRNA gene sequences, strongly support the new species as a monophyletic group nested into the O. schmackeri species complex. The new species can be distinguished from its congeners by a combination of the following characters: (1) body size medium (SVL: 42.1–45.1 mm in males, 83.3–92.7 mm in females); (2) dorsolateral folds absent; (3) tympanum diameter 1.53 times as long as the width of the disc of finger III in females; 2.3 times in males; (4) dorsal skin green with dense granules and sparse irregular brown spots; males with several large warts on dorsum; (5) two metacarpal tubercles; (6) relative finger lengths: I ≤ II < IV < III; (7) tibiotarsal articulation beyond the tip of the snout; (8) ventral surface smooth in females; throat and chest having pale spinules in adult males; (9) dorsal limbs green or yellow green with brown transverse bands; and (10) paired external vocal sacs located at corners of the throat, finger I with light yellow nuptial pad in males. This discovery increases the number of Odorrana species to 59 and those known from China to 37.


introduction
The odorous frogs of the genus Odorrana Fei, Ye & Huang, 1990 are distributed in tropical and subtropical Asia, including southern mainland China, Japan and Indochina west to north-eastern India, Myanmar and Thailand and south through Malaya and Sumatra to Borneo. There are 58 recognised species (Frost 2020). Most of the species inhabit mountain streams at elevations of 200-2000 m and can also be found on rocks or branches near the riverbed (Fei et al. 2009a;Frost 2020). Seven species from the genus Odorrana have recently been described using both molecular and morphological analyses (Chen et al. 2010a, b;Kuramoto et al. 2011;Mo et al. 2015;Wang et al. 2015;Pham et al. 2016;Li et al. 2018a). The phylogeny and diversity of Odorrana and the systematic status of taxa within the genus have been debated by taxonomists (Frost et al. 2006;Che et al. 2007;Fei et al. 2009a;Kurabayashi et al. 2010;Chen et al. 2013;Li et al. 2015). For example, the genus Odorrana was first recognised by Fei et al. (1990), while Dubois (1992) treated Odorrana as a subgenus of Rana and erected a new subgenus R. (Eburana). Frost et al. (2006) expanded the genus Huia to include both Odorrana and R. (Eburana). However, the monophyly of Odorrana was supported by subsequent analyses of mtDNA data, nuclear data and combinations of mtDNA and nuclear data and R. (Eburana) and Huia were rejected (Matsui et al. 2005;Stuart 2008;Pyron and Wiens 2011). Chen et al. (2013) used two mitochondrial genes to study the molecular phylogeny and diversity of Odorrana and they identified seven major clades.
Odorrana schmackeri Boettger, 1892 is a widely distributed species in subtropical and tropical regions in the south of the Qinling Mountains in China. Paleogeological events and climatic oscillation may have led to lineage divergence within O. schmackeri . A comprehensive comparison of the morphological and molecular biological characteristics of O. schmackeri has revealed several cryptic species, such as O. nanjiangensis Fei, Ye, Xie & Jiang, 2007, O. huanggangensis Chen, Zhou & Zheng, 2010a, O. tianmuii Chen, Zhou & Zheng, 2010band O. kweichowensis Li, Xu, Lv, Jiang, Wei & Wang, 2018. The phylogenetic relationships identified within the genus Odorrana has led to a proposed "Odorrana schmackeri species complex" Li et al. 2018a). The "O. schmackeri species complex" is characterised by species morphologically similar to O. schmackeri and contains several described species and several cryptic species: O. hejiangensis, O. huanggangensis, O. kweichowensis, O. schmackeri, O. tianmuii and several cryptic species Zhu 2016;Li et al. 2018a). Zhu (2016) considered O. nanjiangensis to be a synonym of O. hejiangensis indicating that the taxonomic status of O. nanjiangensis should be re-examined.
The Wuling Mountains are a priority area for biodiversity conservation in China (Ministry of Ecology and Environment of China 2015). During a survey of herpetological species diversity in Sangzhi County, Zhangjiajie City, Hunan Province of northeast Wuling Mountains, five specimens of Odorrana were collected in July 2019. These specimens were similar to species in the O. schmackeri species complex in morphology, but detailed taxonomic comparisons showed that these specimens differ from the other known species and appear to represent an independent evolutionary lineage of the genus Odorrana. The results indicate that these specimens represent a new species and that this species belongs to the O. schmackeri species complex.

Sampling
We conducted a herpetological survey in Zhangjiajie City, Hunan Province, from 23 July to 6 August 2019. A total of five specimens (two adult males and three adult females) were collected in mountain streams on 30 July 2019, in Sangzhi County, Zhangjiajie City, Hunan Province, China ( Figure 1). All of the specimens were stored in 75% ethanol (aq) for preservation and liver tissues were preserved in 95% ethanol (aq) for genetic analysis. All of the specimens were deposited at the Animal Museum of Central South University of Forestry and Technology (CSUFT).

Phylogenetic analysis
Genomic DNA was extracted from 95% ethanol-preserved liver tissues using a TSINGKE (https://www.tsingke.net) DNA extraction kit. Two fragments of the mitochondrial gene 12s rRNA and 16s rRNA from five samples were amplified using the primers in Kocher et al. (1989) and Simon et al. (1994). The amplification programme followed the sequence of 98 °C for 2 min; 35 cycles of 98 °C for 10 sec, 55 °C for 10 sec, 72 °C for 10 sec; and a final extending step at 72 °C for 2 min. PCR purification and sequencing were performed by Beijing Tsingke Biological Technology Co. Ltd. For phylogenetic analyses, the available sequence data for all related species of Odorrana and an outgroup were downloaded from GenBank especially for the topotypes of Odorrana species, based on previous studies (Chen et al. 2013;Li et al. 2018a;Suppl. material 1). Rana chensinensis David, 1875 was used as the outgroup following Chen et al. (2013). DNA sequences were aligned by the Clustal W algorithm with default parameters (Thompson et al. 1997). Alignments were checked by eye and revised manually, if necessary. To avoid bias in alignments, GBLOCKS v. 0.91.b with default settings was used to extract regions of defined sequence conservation from the length-variable 12S and 16S gene fragments. Non-sequenced fragments were regarded as missing data. Finally, 12S and 16S gene fragments were concatenated for phylogenetic analyses of mitochondrial DNA. Sequenced data were analysed using Bayesian Inference (BI) in MrBayes 3.2.4 (Ronquist et al. 2012) and Maximum Likelihood (ML) in RaxmlGUI 1.3 (Silvestro and Michalak 2012). Prior to phylogenetic analyses, 12S and 16S genes were defined as two partitions in the concatenated data and the partitions were tested in jmodeltest v.2.1.2 with Akaike and Bayesian information criteria, all resulting in the best-fitting nucleotide substitution models of GTR+I+G. Two independent runs were conducted in a BI analysis, each of which was performed for 10,000,000 generations and sampled every 1000 generations with the first 25% samples discarded as burn-in. This resulted in a potential scale reduction factor (PSRF) of < 0.005. In ML analysis, the bootstrap consensus tree inferred from 1000 replicates was utilised to represent the evolutionary history of the taxa analysed. Genetic distances between and within species were calculated in MEGA 7 using the uncorrected p-distance model, based on the 16S rRNA gene (Kumar et al. 2016).

Morphological analysis
Measurements were taken by a digital caliper (0.01 mm, rounded to the nearest 0.1 mm). Character measurements followed those of Fei et al. (2009b) and Li et al. (2018a): SVL snout-vent length (distance from tip of snout to posterior margin of vent); HDL head length (distance from the tip of the snout to the articulation of jaw); HDW maximum head width (greatest width between the left and right articulations of jaw); SL snout length (distance from the tip of the snout to the anterior corner of the eye); ED eye diameter (distance from the anterior corner to the posterior corner of the eye); IOD interorbital distance (minimum distance between the inner edges of the upper eyelids); IND internasal distance (minimum distance between the inner margins of the external nares); NED nasal to eye distance (distance between the nasal margin and the anterior corner of the eye); TYD maximal tympanum diameter; LAL length of lower arm and hand (distance from the elbow to the distal end of finger III); LW lower arm width (maximum width of the lower arm); THL thigh length (the distance from the vent to the knee); TL tibia length (the distance from the knee to the tarsus); TW maximal tibia width; TFL length of foot and tarsus (distance from the tibiotarsal articulation to the distal end of toe IV); FL foot length (distance from tarsus to the tip of toe IV); FDW width of finger disc (width at the widest part of the disc of finger III).
Sexual size dimorphism was represented by the size dimorphism index (SDI). SDI was calculated as the mean of SVL of the larger sex/mean of SVL in the smaller sex -1 (Kupfer 2009).
Comparative morphological data of four species (Odorrana hejiangensis, O. huanggangensis, O. kweichowensis and O. schmackeri) allocated to the Odorrana schmackeri species complex (currently containing five species) were obtained from examination of museum specimens (see Appendix 1) and from literature (Suppl. material 2) (Li et al. 2018a).
To reduce the impact of allometry, a corrected value from the ratio of each morphological measurement to SVL was calculated and was log-transformed for the following morphometric analyses. An independent sample t-test was used to test the significance of differences on morphometric characters between the new species and different species from the Odorrana schmackeri species complex. The sample size of male specimens was less than three, so we only analysed the morphometric data of females of the new species. Scores were considered significant at p < 0.05. To show the spatial distribution of different species on the morphometric characters, principal component analyses (PCA) were performed. We used one-way ANOVA to compare the morphometric character differences amongst the new species, O. hejiangensis, O. huanggangensis, O. kweichowensis and O. schmackeri and then selected the characters with significant differences amongst species for principal component analyses. All statistical analyses were conducted in R v. 3.5.1. Since the collection location of new species was about 855 km away from the distribution area of O. tianmuii and the distribution area of these two species was also separated by the distribution area of O. schmackeri (Li et al. 2018a), we did not use an independent samples t-test and PCA to compare the morphometric data between the new species and O. tianmuii.

Phylogenetic analyses
The ML and BI phylogenetic trees were constructed, based on DNA sequences of mitochondrial 12S and 16S genes with a total length of 1,960 bp ( Figure 2). The validity of the unnamed Odorrana specimens was affirmed by molecular phylogenetic analysis. All five samples of the new taxon occurring in Sangzhi County were densely clustered into a strongly-supported monophyletic group. The phylogenetic tree suggested that the new taxon is sister to O. hejiangensis and O. nanjiangensis and belongs to the O. schmackeri species complex. The uncorrected p-distances between the new taxon and all other species of the complex are 2.8%-9.8%, which were larger than the interspecific genetic distance between  Diagnosis. Odorrana sangzhiensis sp. nov. can be distinguished from all of its congeners by a combination of following characters: (1) medium size, SDI = 1.03 (SVL 83.3-92.7 mm in females and 42.1-45.1 mm in males); (2) head length greater than head width; (3) dorsolateral folds absent; (4) a small white or beige dot between eyes; (5) brown supratympanic fold slightly distinct; (6) the tympanum diameter in females 1.53 times as long as the width of the disc of finger III; in males 2.3 times; (7) green dorsal skin with dense granules and sparse, irregular brown spots; the males with several large warts on dorsum; (8) flanks with larger tubercles, 4-7 large tubercles usually arranged in a dorsolateral row of the males; (9) two metacarpal tubercles; (10) discs of all digits with circum-marginal grooves; (11) relative finger lengths: I ≤ II < IV < III; (12) feet fully webbed; (13) tibiotarsal articulation beyond the tip of the snout; (14) ventre smooth in females; pale spinules present on throat and chest of adult males; (15) dorsal digits light yellow or beige with light brown spots, dorsal limbs green or yellow green with brown transverse bands; (16) having paired external vocal sacs located at corners of the throat, light yellow nuptial pad on finger I in males.
Description of holotype. Adult male, SVL 42.1 mm; top of head flat; head length greater than maximum head width (HDL: HDW = 1.36); snout obtusely rounded in dorsal view (SL/HDL = 0.39), rounded in profile, projecting beyond lower jaw; large and protruding eyes (ED/SL = 0.79), deeply concave loreal region; IND = 5.3 mm, larger than IOD = 3.9 mm; nasal margin slightly closer to the tip of snout; tympanum circular, large and distinct, surrounded by many granules, tympanum diameter 0.65 times the eye diameter; vomerine teeth developed into mass on two oblique ridges on the inside of two internal nostrils; tongue deeply notched posteriorly.
Forelimbs sturdy (LW/LAL = 0.21); length of lower arm and hand just over half the body length (LAL/SVL > 0.50); relative finger lengths: I ≤ II < IV < III; FDW/TYD = 0.44; finger webbing absent, tips of fingers moderately expanded, presenting circular discs with slightly narrow top and circum-marginal grooves; subarticular tubercles prominent; supernumerary tubercle at the base of each finger smaller than subarticular tubercles; two oblong metacarpal tubercles; buff nuptial pad on the first finger.
Hind limbs relatively long, the heels overlapping obviously, tibiotarsal articulation beyond the tip of the snout; tibia length (TL) 0.54 times SVL; foot length (FL) 0.56 times SVL; toes, slender, relative toe lengths I < II < III < V < IV; tibia length slightly shorter than foot length; subarticular tubercles prominent; toes discs same as fingers; obvious horizontal groove in dorsal view of discs; feet fully webbed to discs; lateral fringes on free edges of toes I and V not obvious, metatarsal fold absent; inner metatarsal tubercle present, elliptical; outer metatarsal tubercle absent; inner tarsal fold absent.
Skin slightly rough with dense granules, several large warts on dorsum and flanks; dorsolateral folds absent; two glands behind the back edge of the lips; a small beige dot between anterior corners of the eyes; pale spinules present on throat and chest.
Colouration in life. In life, the yellow-green dorsum of head and body with scattered irregular dark brown spots; no light edges around spots; small area of dark brown spots on the dorsum of head; large dark black spots in the centre on dorsum, continuing on to light yellow flank with several brown spots; supratympanic fold slightly distinct, brown; upper and lower lip with vertical brown bars; arms green-yellow with brown transverse bands, thighs and tibias with four brown bands; different widths of brown striations on green-yellow limbs, different distances between striations (Figure 3).
Colouration in alcohol. On dorsum, colour fades to grey-blue with dark grey spots, brown bars on upper and lower lips change to dark grey; ventre variable from yellowish to creamy-white; beige dot between anterior corners of the eyes fades to white; underside of limbs varies from brown to beige (Figure 4).
Variation. Measurements of the five specimens are shown in Table 1. The specimens show obvious differences between females and males in each character measurement. The snout-vent length of females is approximately twice as long as the males (SVL mean 88.6 mm, range 83.3-92.7 mm in females; SVL mean 43.6 mm, range 42.1-45.1 mm in males). In the ratio of each character to SVL, HDL, HDW, ED, IOD, TYD and LW, these are all significantly smaller in females than in males. Some females have dense brown spots on the dorsum and most of the yellow green on their limbs is replaced by yellow white ( Figure 3E).
Etymology. The scientific name "sangzhiensis" is derived from its type locality Sangzhi County in Hunan Province. As common names, we suggest Sangzhi Odorous Frog (English) and Sang Zhi Chou Wa (Chinese).
Distribution and natural history. The new species is currently known only from its type locality, Sangzhi County, Zhangjiajie, Hunan Province, China (Figure 1). It is found in canyon streams and perches on the rocks beside the streams ( Figure 7A, B). They usually hide under the rocks during the day and are active at night. The vegetation on both sides of the stream is luxuriant and comprises a mixed forest of evergreen and deciduous flora. In their habitat, the dominant trees are Ulmus changii Cheng, 1936, Castanopsis carlesii Hayata, 1917and Sloanea hemsleyana Rehder & Wilson, 1916and the dominant shrubs are Boehmeria penduliflora Weddell & Long, 1982and Distylium myricoides Hemsl, 1907. The dominant herbs are Pilea sinofasciata Chen, 1982, Strobilanthes dimorphotricha Hance, 1883 and Miscanthus floridulus Labillardière, 1824. One sympatric amphibian species, Amolops ricketti Boulenger, 1899, was found. In the field investigation, the tadpoles and eggs of this new species were not found. However, we observed reproductive behaviour ( Figure 3F). In addition, one female (CSUFT 4308220047) collected on 30 July 2019, contained mature eggs ( Figure 3C). The eggs in preservative showed yellowish-white. Thus, the breeding period of this species may be in July and August.

Discussion
Odorous frog species are widely distributed and have many morphological characteristics of derivation or specialisation. Different species of Odorrana are distributed in the same region or in different niches within the same area, which makes species identification difficult and may also cause some cryptic species to be overlooked (Fei et al. 2009a;Li et al. 2015;Frost 2020). DNA sequence comparison and the genetic distance threshold is a highly effective method to identify amphibian species and eliminate misidentifications (Xiong et al. 2015). A cryptic species, Odorrana kweichowensis, has recently been described from Guizhou Province using morphological and molecular data and has morphological characteristics similar to O. schmackeri (Li et al. 2018a). The discovery of Odorrana sangzhiensis sp. nov. adds another species to the O. schmackeri species complex, indicating that the species diversity of Odorrana may be under-estimated.
The Yunnan-Guizhou Plateau is thought to be the centre of Odorrana origin . The uplift of the Qinghai-Tibet Plateau altered the water systems and affected the geographic pattern and genetic structure of Odorrana species . To clarify the species composition and distribution range of the Odorrana schmackeri species complex, Zhu (2016) analysed the genetic differentiation of the complex by sampling based on the partial mitochondrial 12S and 16S r RNA genes (1021 individuals from 85 populations in 15 Chinese Provinces were studied). The results showed that the Odorrana schmackeri species complex included four described species The nearest distance between the distribution area of Odorrana sangzhiensis sp. nov. and that of O. hejiangensis is about 140 km. It will be necessary to determine if the two species are sympatric or whether their reproduction period is staggered in this area. In the field investigation, we found that the habitat of this species was seriously threatened by a hydropower station ( Figure 7C, D). Further investigation in this area should be conducted to determine the population status and distributional range of this species and locate additional cryptic species of Odorrana.