Between 2021 and 2022, fourteen specimens were collected at the Shiwandashan National Nature Reserve (SWDS), Shangsi County, Guangxi, China (permission no. SWDS20210501). For comparison, nine specimens of Leptobrachella ventripunctata (Fei, Ye & Li, 1990) were collected at the Jinzhongshan National Nature Reserve on 22 June 2021 (JZS) (permission no. JZS20210605). Additionally, L. sungi specimens were collected at the SWDS (n = 16) on 4 July 2021, and the Sishuihe Nature Reserve (SSH) (n = 3) on 20 June 2020, located in Lingyun County, Guangxi, China (permission no. SSH20200615) (Fig. 1). After euthanasia using isoflurane, all specimens were fixed in 10% formalin for 48 h and finally stored in 75% ethanol. Muscle samples were taken prior to fixation and stored in 100% ethanol for subsequent molecular analyses. All specimens and muscle samples are deposited in the collection of Nanning Normal University (NNU) (see Table 1 for details). Specimens were measured with a digital caliper to the nearest 0.1 mm. The following measurements were taken:

Figure 1. 

Localities of the new species and its sister taxa. Sample ID corresponding to those provided in Table 1.

SVL snout-vent length;

HL head length from the tip of snout to rear of jaws;

HW head width at commissure of jaws;

SNT snout length from the tip of snout to the anterior eye corner;

ED diameter of the exposed portion of eyeball;

IOD interorbital distance, the shortest distance between the anterior corners of the orbits;

IN internarial space distance;

EN distance from the eye to nostril, measured from the anterior corner of the eye to the posterior margin of the nostril;

TD horizontal diameter of tympanum;

TED distance from anterior edge of the tympanum to posterior eye corner;

TIB tibia length with flexed hindlimb;

FLL forelimb length from elbow to the tip of third finger;

THL thigh length from vent to knee;

ML manus length from the tip of third digit to proximal edge of the inner palmar tubercle;

PL pes length from the tip of fourth toe to the proximal edge of inner metatarsal tubercle;

FEM maximum diameter of femoral gland.

Sex was determined either directly through observation of calling males, presence of vocal sacs in males, or the presence of eggs in the abdomen of females. The webbing formula was determined following Savage (1975). Morphological data were obtained from the collected vouches specimens (Suppl. material 1: table S1) as well as other museum specimens (Suppl. material 1: table S2).

DNA was isolated from muscle samples using Tiangen Biotech Co. Ltd. tissue extraction kits (Beijing, China). The mitochondrial fragments of 16S (~530 bp) were amplified and sequenced using the primer pairs 16Sar_L (5’–CGCCTGTTTACCAA AAACAT–3’) and 16Sbr_H (5’–CCGGTCTGAACTCAGATCACGT–3’). Polymerase chain reaction (PCR) amplification followed the method described by Chen et al. (2021). The 16S fragments were sequenced on an ABI Prism 3730 automated DNA sequencer, and the new sequences were deposited in GenBank (OP548561–OP548567, OP548569–OP548579). Phylogenetic trees were reconstructed using the new sequences and homologous sequences of the genus Leptobrachella downloaded from GenBank (Table 1). The molecular data included topotypic sequences of L. ventripunctata (GenBank no. MH055831, KM014811, and MG520361) from Xishuangbanna, Yunnan, China (Sung et al. 2014; Chen et al. 2018; Yang et al. 2018). Bayesian inference (BI) and maximum likelihood (ML) methods were used to construct the phylogenetic trees. BI was performed using MrBayes v. 3.1.2 (Ronquist and Huelsenbeck 2003). The best-fit evolution model (GTR+I+G) was tested in JMODELTEST v. 2.1.7 (Posada 2008). Two independent runs with four Markov Chain Monte Carlo simulations were performed for 30 million iterations, and trees were sampled every 1,000^th generation. The first 25% of trees were discarded as burn-in. ML analysis was carried out on the CIPRES science gateway with 100 rapid bootstrap replicates (Miller et al. 2010) (https://www.phylo.org/portal2). Uncorrected p-distances of the 16S gene were estimated using Mega v. 7 (Kumar et al. 2016) with the default settings.

Advertisement calls were recorded using a SONY PCM-A10 recorder, and ambient temperature was measured using a digital hygrothermograph. The call recordings were analysed using the software Raven Pro v.1.6 (Cornell Laboratory of Ornithology, Ithaca, NY, USA). Audio-spectrograms were generated using Hanning windows, fast-Fourier transform (FFT) of 512 points, 50% overlap, and 172 Hz grid-spacing. Acoustic parameters were defined following Köhler et al. (2017) and Emmrich et al. (2020). Thus, we refer to a call refer as a group of notes, and the call duration is the time from the beginning of the first note to the end of the last note in a call. Call interval is defined as the time from the end of the last note of a call to the beginning of the first note of the subsequent call. Calls are often divided into two or more notes, which are smaller subunits that are usually separated by short intervals of silence relative to the note duration. The dominant frequency of a call is determined as the frequency with the highest energy concentration within the entire power spectrum.

According to our results of the phylogenetic analyses, the new species is closely related to L. ventripunctata. Consequently Mann-Whitney U tests were conducted to determine the significance of differences in morphometric characters between the new species and L. ventripunctata (from JZS). Differences were considered significant below a threshold of 0.05. Principal component analysis (PCA) was performed to examine the distribution of the two species based on their morphometric parameters. Prior to the analysis, morphometric parameters were adjusted by calculating the ratio of each parameter to SVL, and then log-transformed to minimise the impact of allometry. All statistical analyses were carried out using IBM SPSS v. 20.

BI and ML analyses yielded nearly identical phylogenetic trees (Fig. 2). The preliminary phylogenetic trees revealed that all SWDS specimens were classified into four distinct lineages, corresponding to L. shangsiensis, L. shiwandashanensis, L. sungi, and an unidentified Leptobrachella lineage (Fig. 2). The newly collected specimens from SWDS formed a monophyletic group that is closely related to L. ventripunctata. The JZS specimens and L. ventripunctata from the type locality clustered together. The genetic divergences between the newly collected specimens and three sympatric species (L. shangsiensis, L. shiwandashanensis, and L. sungi) exceeded 8.2% (Suppl. material 1: table S3). The genetic divergences between the topotypic samples of L. ventripunctata and the newly collected specimens ranged from 1.6% to 2.4%, while those between the newly collected specimens and L. ventripunctata from JZS ranged from 1.7% to 1.9% (Suppl. material 1: table S3).

Figure 2. 

BI trees based on the part of the 16S gene. Node support is indicated on branches as maximum likelihood support (upper half; > 70% < 90% = grey, > 90% = black) and Bayesian posterior probabilities (lower half; > 0.95 = grey, 1 = black).

The diagnostic characters for the new species of the genus Leptobrachella occurring north of the Isthmus of Kra are presented in Table 2, indicating that the newly collected specimens differ significantly from their congeners. The results of Mann-Whitney U tests revealed significant differences between the new specimens and L. ventripunctata from JZS in various measurements, including SVL, ED, IN, and FLL for males, and SVL, HL, HW, EN, TED, TIB, and PL for females (Table 3). There is no overlap in measurements between the new species and L. ventripunctata from JZS or the paratypes in terms of the measured parameters, including SVL, HL, HW, SNT, ED, TIB, FLL, ML, and PL (Table 3). Thus, PCA results showed clear differentiation between the new species and L. ventripunctata (Fig. 3). Furthermore, the newly described species exhibited distinct differences from L. ventripunctata in terms of body size (males being larger: SVL 30.5–32.5 mm vs 25.5–28.0 mm), ventral texture (ventral surface creamy white without dark brown spots vs chest and belly with dark brown spots), presence of a wide lateral fringe on toes (vs absence), presence of 1/3 toe webbing (vs absence), and presence of distinct dermal ridges under the toes (vs absence).

Figure 3. 

Scatter plot of PC1 and PC2 of PCA based on the morphometric measurements, distinguishing L. guinanensis sp. nov. and L. ventripunctata A male comparison B female comparison.

The calls of four individuals (NNU 00560–561, NNU 00875–876) were recorded. The main features of these calls are summarised in Table 4. The call of the newly described species consists of four notes (Fig. 4). The duration of the calls ranged from 23 milliseconds (ms) to 31 ms (mean 25.5 ± 1.4, n = 4), while the intervals between the calls were 55–133 ms (mean 91.2 ± 17.5, n = 4). The dominant frequencies of the calls were found to be in the range of 7.3–8.3 kHz. These characters were distinct from those of the sympatric species (L. shangsiensis, L. shiwandashanensis, and L. sungi) and L. ventripunctata (Fig. 4, Suppl. material 2: fig. S1, Table 4; Yang et al. 2018; Chen et al. 2019, 2021). Furthermore, the calls of the newly described species can also be differentiated from the known 40 species in the genus Leptobrachella (Suppl. material 1: table S4), as they possess the highest dominant frequencies ever recorded.

Figure 4. 

Advertisement calls of L. guinanensis sp. nov. A and L. ventripunctata B including waveforms and spectrograms.

Based on the evidence from morphology, phylogeny, and bioacoustics, it is evident that the collected specimens represent a distinct, previously undescribed species within the genus Leptobrachella. Therefore, we describe these specimens as a new species of the genus Leptobrachella.

The authors have declared that no competing interests exist.

No ethical statement was reported.

This work was supported by the National Natural Science Foundation of China (grant numbers 32360128 and 32060116) and Guangxi Natural Science Foundation, China (grant number 2020GXNSFDA238022).

CWC conceived and designed the study and prepared the manuscript. CWC measured the specimens, performed the molecular experiments, and analysed the data. CWC, LP, LYJ and HY conducted field surveys. PWX drew the figures. All authors read and approved the final version of the manuscript.

Wei-Cai Chen https://orcid.org/0000-0002-2398-4079

Peng Li https://orcid.org/0000-0001-8311-0544

Wan-Xiao Peng https://orcid.org/0000-0001-5635-9061

You-Jun Liu https://orcid.org/0000-0001-7285-4943

Yong Huang https://orcid.org/0000-0002-3493-9468

All of the data that support the findings of this study are available in the main text or Supplementary Information.