A new species of the Asian leaf litter toad genus Leptobrachella Smith, 1925 (Anura, Megophryidae) from northwest Guizhou Province, China

Abstract A new species of the Asian leaf litter toad genus Leptobrachella is described from Guizhou Province, China. Molecular phylogenetic analyses support the new species as an independent lineage deeply nested in the Leptobrachella clade. The new species is distinguished from its congeners by a combination of the following morphological characters: body size medium (SVL 29.7–31.2 mm in five adult males); dorsal skin shagreened, some of the granules forming longitudinal short skin ridges; tympanum distinctly discernible, slightly concave; supra-axillary, femoral, pectoral and ventrolateral glands distinctly visible; absence of webbing and lateral fringes on fingers; toes with narrow lateral fringes but without webbing; heels overlapping when thighs are positioned at right angles to the body; tibia-tarsal articulation reaching the middle of eye when leg stretched forward. The discovery highlighted the underestimated species diversity in the Leptobrachella toads in southwestern China.


Introduction
The Asian leaf litter toads of the genus Leptobrachella Smith, 1925 (Anura, Megophryidae) are widely distributed from southern China west to northeastern India and Myanmar, through mainland Indochina to peninsular Malaysia and the island of Borneo (Frost 2020). Many species in this genus have been classified into Leptolalax Dubois, 1983(e.g., Fei et al. 2009, 2012, but Chen et al. (2018) placed Leptolalax as a junior synonym of Leptobrachella based on large-scale molecular analyses. Currently, the genus Leptobrachella contains 82 species (Frost 2020) but a series of cryptic species is still suggested by molecular phylogenetic analyses . To date, 25 species of this genus have been recorded in China, i.e., L. alpina (Fei, Ye & Li, 1990) and L. bourreti (Dubois, 1983) from Yunnan and Guangxi; L. eos , L. nyx , L. pelodytoides (Boulenger, in 95% ethanol prior to fixation. Specimens were deposited in Chengdu Institute of Biology, Chinese Academy of Sciences (CIB, CAS).

Molecular phylogenetic analyses
All five adult male specimens of the new species collected in this work were included in the molecular phylogenetic analyses (Table 1). For phylogenetic analyses, the corresponding gene sequences for all those related species for which comparable sequences were available were also downloaded from GenBank (Table 1). Corresponding sequences of one Leptobrachium tengchongensis, one Leptobrachium huashen, and one Megophrys major were downloaded (Table 1) and used as outgroups based on previous studies Li et al. 2020a).
Total DNA was extracted using a standard phenol-chloroform extraction protocol (Sambrook et al. 1989). The mitochondrial 16S rRNA genes were amplified, and the primers P7 (5'-CGCCTGTTTACCAAAAACAT-3') and P8 (5'-CCGGTCT-GAACTCAGATCACGT-3') were used following Simon et al. (1994). Gene fragments were amplified under the following conditions: an initial denaturing step at 95 °C for 4 min; 36 cycles of denaturing at 95 °C for 30 sec, annealing at 51 °C for 30 sec and extending at 72 °C for 70 sec. Sequencing was conducted using an ABI3730 automated   Table 1). Sequences were assembled and aligned using the Clustalw module in BioEdit v. 7.0.9.0 (Hall 1999) with default settings. Alignments were checked by eye and revised manually if necessary. Phylogenetic analyses were conducted using maximum likelihood (ML) and Bayesian Inference (BI) methods, implemented in PhyML v. 3.0 (Guindon et al. 2010) and MrBayes v. 3.12 (Ronquist and Huelsenbeck 2003), respectively. We ran Jmodeltest v. 2.1.2 (Darriba et al. 2012) with Akaike and Bayesian information criteria on the alignment, resulting in the best-fitting nucleotide substitution models of GTR + I + G for the data. For the ML tree, branch supports were drawn from 10,000 nonparametric bootstrap replicates. In BI analyses, the parameters for each partition were unlinked, and branch lengths were allowed to vary proportionately across partitions. Two runs each with four Markov chains were simultaneously run for 50 million generations with sampling every 1,000 generations. The first 25% trees were removed as the "burn-in" stage followed by calculations of Bayesian posterior probabilities and the 50% majority-rule consensus of the post burn-in trees sampled at stationarity. Finally, genetic distance between Leptobrachella species based on uncorrected p-distance model was estimated on 16S gene using MEGA v. 6.06 (Tamura et al. 2013).

Morphological comparisons
All five adult male specimens of the new species were measured ( Table 2). The terminology and methods followed Fei et al. (2005), Mahony et al. (2011), andWang et al. (2019). Measurements were made with a dial caliper to the nearest 0.1 mm (Watters et al. 2016) with digital calipers. Corresponding measurements of L. bijie and L. chishuiensis were retrieved from Wang et al. (2019) and Li et al. (2020a). Nineteen morphometric characters of adult specimens were measured:

ED
eye diameter (distance from the anterior corner to the posterior corner of the eye); FL foot length (distance from tarsus to the tip of the fourth toe); HDL head length (distance from the tip of the snout to the articulation of jaw); HDW head width (greatest width between the left and right articulations of jaw); HLL hindlimb length (distance from tip of fourth toe to vent); IND internasal distance (minimum distance between the inner margins of the external nares); IOD interorbital distance (minimum distance between the inner edges of the upper eyelids); LAL length of lower arm and hand (distance from the elbow to the distal end of the Finger IV); LW lower arm width (maximum width of the lower arm); ML manus length (distance from tip of third digit to proximal edge of inner palmar tubercle); SL snout length (distance from the tip of the snout to the anterior corner of the eye); SVL snout-vent length (distance from the tip of the snout to the posterior edge of the vent); TEY tympanum-eye distance (distance from anterior edge of tympanum to posterior corner of eye); TFL length of foot and tarsus (distance from the tibiotarsal articulation to the distal end of the toe IV); THL thigh length (distance from vent to knee); TL tibia length (distance from knee to tarsus); TW maximal tibia width; TYD maximal tympanum diameter; UEW upper eyelid width (greatest width of the upper eyelid margins measured perpendicular to the anterior-posterior axis).
In order to reduce the impact of allometry, the correct value from the ratio of each character to SVL was calculated and then all of the data were log-transformed for the following morphometric analyses. Mann-Whitney U tests were conducted to test the significance of differences on morphometric characters between Leptobrachella jinshaensis sp. nov., L. bijie and L. chishuiensis. The significance level was set at 0.05. Furthermore, principal component analyses (PCA) were conducted to highlight whether the different species were separated in morphometric space. Due to only the measurements SVL, HDL, HDW, SL, IND, IOD, ED, TYD, TEY, LAL, ML, TL, HLL, and FL of male L. bijie being available from Wang et al. (2019), the morphometric analyses were conducted only based on these 14 morphometric characters for male group.
Leptobrachella jinshaensis sp. nov. was also compared with all other congeners of Leptobrachella based on morphological characters. Comparative morphological data were obtained from literatures (Table 3).

Bioacoustics data
The advertisement calls of L. jinshaensis sp. nov. were recorded from the holotype specimen CIBJS20200516004 in the field on 16 May 2020 in Lengshuihe Nature Reserve, Jinsha County, Guizhou Province, China. The advertisement call of L. jinshaensis sp. nov. was recorded in the stream at ambient air temperature of 20 °C and air humidity of 87%. A SONY PCM-D50 digital sound recorder was used to record within 20 cm of the calling individual. The sound files in wave format were resampled at 48 kHz with sampling depth 24 bits. Calls were recorded and examined as described by Wijayathilaka and Meegaskumbura (2016). Call recordings were visualised and edited with SoundRuler 0.9.6.0 (Gridi-Papp 2003-2007 and Raven Pro 1.5 software (Cornell Laboratory of Ornithology, Ithaca, NY, USA). Ambient temperature of the type locality was taken by a digital hygrothermograph. For comparison, bioacoustics data for the related species L. bijie and L. chishuiensis were obtained from Li et al. (2020a).

Results
Aligned sequence matrix of 16S rRNA gene contained 537 bps. ML and BI analyses resulted in essentially identical topologies (Fig. 2). All samples of L. jinshaensis sp. nov. were clustered into one independent clade nested into the Leptobrachella clade. The relationships between L. jinshaensis sp. nov. and its congeners are not resolved though it is likely sister to a clade in comprising of L. bijie and L. chishuiensis (Fig. 2). The smallest pairwise genetic divergence between L. jinshaensis sp. nov. and all other species of the genus Leptobrachella is 2.6% (vs. L. niveimontis or vs. L. purpurus), being at the same level with or higher than that between some pairs of substantial species, such as L. bijie vs. L. chishuiensis (2.1%), and L. chishuiensis vs. L. alpina (2.6%; Suppl. material 1: Table S1).
For the male group, PCA extracted five principal component axes with eigenvalues greater than one, and the percentage of variance of the first five principal components are 37.7%, 15.7%, 13.0%, 9.0% and 8.1%, with percentage of cumulative is 83.5% (Suppl. material 2: Table S2). There were 14 morphological features with major contributions in the first five principal components, and these morphological features were distributed in the anterior, middle, and posterior parts of the body (Suppl. material 2: Table S2). The total variation of the first two principal components was 53.4% (Suppl. material 2: Table S2). On the PCA plot (PC1 vs. PC2), the first principal component axis could separate L. jinshaensis sp. nov. from L. bijie and L. chishuiensis (Fig. 3) mainly based on SVL, HDL, HDW, SL, ED, IND, TEY, and FL, and the second component axis mainly based on ML, FL, and LAL. Mann-Whitney U tests indicated that L. jinshaensis sp. nov. was significantly different from L. bijie on HDW, SL, IOD, TYD, TEY, LW, and FL, and from L. chishuiensis on SVL, TYD, and TL (p-values < 0.05; Table 4).

No.
Leptobrachella species In total, 109 advertisement calls of Leptobrachella jinshaensis sp. nov. were recorded in Lengshuihe Nature Reserve, Jinsa County, Guizhou Province, China on 16 May 2020 between 21:00-22:00. The call description is based on recordings of the holotype CIB-JS20200516004 under a stone nearby a stream, and the ambient air temperature was 20 °C. The call characters of L. jinshaensis sp. nov. were demonstrated in the following section for describing it. There were some differences in sonograms and waveforms of calls between L. jinshaensis sp. nov., L. bijie, and L. chishuiensis (Suppl. material 3: Table S3). Leptobranchella jinshaensis sp. nov. has longer call interval (132.7 ± 8.6, N = 109) than L. bijie (101.9 ± 6.4, N = 33), and has lower dominant frequency (4525 ± 0.065 Hz) than L. bijie (4780.4 ± 76.5 Hz) and L. chishuiensis (6064-6284 H). Each call of L. jinshaensis sp. nov.has two kinds of notes, while each call of L. chishuiensis only has one kind of note.  Paratypes. Four adult males from the same place as holotype. Two adult males CIBJS20200516001 and CIBJS20200516002 collected by Shi-Ze LI, and two adult males CIBJS20200516003 and CIBJS20200516005 collected by Jing LIU, all of them were collected on 16 May 2020.
Diagnosis. Leptobrachella jinshaensis sp. nov. is assigned to the genus Leptobrachella based on molecular phylogenetic analyses and the following morphological characters: medium size, rounded finger tips, the presence of an elevated inner palmar tubercle not continuous to the thumb, the presence of macroglands on body (including supraaxillary, pectoral, and femoral glands), vomerine teeth absent, tubercles on eyelids, and the anterior tip of snout with a vertical white bar.
Leptobrachella jinshaensis sp. nov. can be distinguished from its congeners by a combination of the following characters: body of medium size 2 mm in five adult males); dorsal skin shagreened, some of the granules forming longitudi- nal short skin ridges; tympanum distinctly discernible, slightly concave; supra-axillary, femoral, pectoral, and ventrolateral glands distinctly visible; absence of webbing and lateral fringes on fingers; toes with narrow lateral fringes and without webbing; heels overlapping when thighs positioned at right angles to the body; tibia-tarsal articulation reaching the middle eye when leg stretched forward.
Description of holotype (Figs 4, 5). Adult male. SVL in 31.1 mm. Head length slightly longer than head width (HDL/HDW 1.02); snout slightly protruding, projecting slightly beyond margin of the lower jaw; nostril closer to snout than eye; canthus rostralis gently rounded; loreal region slightly concave; interorbital space flat, interorbital distance slightly longer than internarial distance; pineal ocellus absent; vertical pupil; eye diameter slightly shorter than snout length; tympanum distinct, rounded, and slightly concave, diameter smaller than that of the eye (TMP/ED 0.61); upper margin of tympanum in contact with supratympanic ridge; vomerine teeth absent; tongue notched behind; supratympanic ridge distinct, extending from posterior corner of eye to supra-axillary gland.
Forelimbs slender, 48.9% of snout-vent length; tips of fingers rounded, slightly swollen; relative finger lengths I < II <= IV < III; absence of webbing; nuptial pad and subarticular tubercles absent; inner palmar tubercle large, rounded separated from the smaller, round outer palmar tubercle.
Hindlimbs slender, tibia slightly longer than thigh length and 48.4% of snoutvent length; heels overlapping when thighs are positioned at right angles to the body, tibiotarsal articulation reaching middle eye when leg stretched forward; relative toe lengths I < II < V < III < IV; tips of toes round, slightly dilated; subarticular tubercle at the articulations of the toes absent; toes without webbing; lateral fringes narrow on all toes; inner metatarsal tubercle present, large, oval, outer metatarsal tubercle absent.
Dorsal surface shagreened and granular, some of the granules forming short longitudinal folds dorsally on the flank; ventral skin smooth; dense tiny granules present on ventral surface of thigh and tibia; pectoral gland and femoral gland white, oval, distinctly visible. Ventrolateral gland distinctly visible and forming an incomplete line.
Colouration of holotype in life. Dorsum brown, with small, distinct darker brown markings and spots, and irregularly dispersed light orange speckles. A dark brown inverted triangular pattern between anterior corners of eyes. Tympanum brown, a dark brown bar above tympanum, and a dark brown bar under the eye, distinct black supratympanic line present; transverse dark brown bars on dorsal surface of limbs; distinct dark brown blotches on flanks from groin to axilla, longitudinally in two rows; elbow and upper arms with dark bars and distinct coppery orange colouration; fingers and toes with distinct dark bars. Ventral surface of throat cream white, chest, and belly cream yellow with purple speckling, and on flanks presence of distinct nebulous greyish speckling; ventral surface of limbs grey purple. Supra-axillary gland, femoral, pectoral, and ventrolateral glands white (Fig. 5).
Colouration of holotype in preservation. Dorsum of body and limbs fade to brown copper; transverse bars on limbs become more distinct. Ventral surface of body and limbs fade to cream white. Supra-axillary, femoral, and pectoral glands fade to creamy yellow (Fig. 4).
Variation. Measurements of adult specimens were presented in Tables 2 and 4. All specimens were similar but some individuals different from the holotype in colour pattern. In CIBJS20200516002, the tympana are dark brown (Fig. 6A); in CIB-JS20200516005, the dorsum is olive grey (Fig. 6B) and the pectoral glands on the left side not obviously (Fig. 6D); in CIBJS20200516003 ventrolateral glands scattered and unlined (Fig. 6C).
Seven species (L. liui, L. oshanensis, L. purpuraventra, L. bijie, L. suiyangensis, L. chishuiensis, and L. ventripunctata) of the genus occur in Guizhou Province, China (Fei et al. 2012;Wang et al. 2019;Luo et al. 2020;Li et al. 2020a). The new species differs from L. liui by having narrow lateral fringes on toes (vs. wide in the latter), dorsal surface shagreened with small granules, and lacking enlarge tubercles or warts (vs. dorsum with round tubercles in the latter); differs from L. oshanensis by having narrow lateral fringes on toes (vs. lacking in the latter); differs from L. purpuraventra and L. suiyangensis by heels overlapping when thighs are positioned at right angles to the body (vs. just meeting in the latter); differs from L. purpuraventra by tibia-tarsal articulation reaches the middle eye when leg stretched forward (vs. only reaches the level between tympanum to eye in the latter).
In mitochondrial DNA trees, Leptobrachella jinshaensis sp. nov. clustered as an independent clade and appears to be sister to a clade in comprising of L. bijie and L. chishuiensis. The latter two species also occur near the type locality of the new species. The new species differs from L. bijie by the following characters: webbing on toes absent (vs. webbing rudimentary in the latter), heels overlapping when thighs are positioned at right angles to the body (vs. just meeting in the latter), having longer call interval (132.7 ± 8.6, N = 109 in the new species vs.101.9 ± 6.4, N = 33 in the latter), having lower dominant frequency of 4525 ± 0.065 Hz vs. 4780.4 ± 76.5 Hz in the latter, having significantly higher value of SVL in males, and having significantly higher value of TYD and TL to SVL in males. Leptobrachella jinshaensis sp. nov. differs from L. chishuiensis by webbing on toes absent (vs. webbing rudimentary in the latter), tibiatarsal articulation reaches the middle of eye when leg stretched forward (vs. reaches the tympanum or the level between tympanum to eye in the latter), the lower dominant frequency of calls 4500-4688 Hz (mean 4525 ± 0.065, 20 °C) vs. 6064-6284 Hz (6140. 15 ± 69.35, 20 °C) in the latter, each call with two kinds of notes vs. only one kind of note in the latter, and having significantly higher value of HDW, SL, IOD, TYD, TEY and FL to SVL in males (all p-values < 0.05; Table 4).
Ecology. Leptobrachella jinshaensis sp. nov. is known from the type locality, Lengshuihe Nature Reserve, Jinsha County, Guizhou Province, China. Specimens of the new species are frequently found from stream covered with reeds, and under the rocks (Fig. 8).
Etymology. The specific name jinshaensis refers to the distribution of this species, Jinsha County, Guizhou Province, China. We suggest its English common name "Jinsha leaf litter toads" and Chinese name "Jin Sha Zhang Tu Chan (金沙掌突蟾)".

Discussion
Molecular phylogenetic analyses, detailed morphological comparisons, and advertisement call data all supported the new species distinctly separated from its congeners especially the superficially-morphological-similar species, L. bijie and L. chishuiensis. Although the relationships between the new species and other closely related species were not resolved, the new species appears to be phylogenetically closer to L. bijie and L. chishuiensis, corresponding to their high similarity on morphology. However, the new species appears to have lower dominant frequency on calling than the two closely related species. Moreover, they could be separated by morphometric analyses on contributions of some characters, for example, on PC1 of PCA, several characters of head, SVL and FL, which might be associated the calling behaviours, breeding behaviours, and jumping behaviours. We need future work to detect the function of the characters of these species to explore the ecological differences between them.
The large-scale molecular phylogenetic analyses in Chen et al. (2018) revealed many cryptic species in the genus Leptobrachella but did not included samples of Leptobrachella jinshaensis sp. nov. Similarly, this large phylogenetic framework likely included a few population samples in Guizhou Province, China. However, the phylogenetic framework indicated that Guizhou Province might be the biogeographical zone of transition for western-to-eastern or southwestern-to-northeastern clades . The findings of series of new species (Leptobrachella jinshaensis sp. nov., L. chishuiensis, L. suiyangensis, L. bijie, and L. purpuraventra) obviously supply important supplemental materials for detecting detailed evolutionary and biogeographical models of the genus. Moreover, the findings of the new species also indicated a high degree of localised diversification and micro-endemism for the species in the genus Leptobrachella because in Guizhou Province, China, the five recent-described Leptobrachella species are just known only from their type localities or nearby areas. In addition, in recent years, large number of discoveries have been made from Guizhou, dramatically raising the number of frog species known from the region (Zhang et al. 2017;Li et al. 2018aLi et al. , b, 2019aLi et al. , b, 2020aLyu et al. 2019;Wang et al. 2019;Luo et al. 2020;Su et al. 2020;Xu et al. 2020;Wei et al. 2020). This further indicated that more investigations should be conducted in Guizhou Province to define more precisely distribution area of the new species and detect more cryptic species especially in the poorly-investigated areas.