A new species of Music frog (Anura, Ranidae, Nidirana) from Mt Daming, Guangxi, China

Abstract Nidiranaguangxiensissp. nov., a new music frog species, is proposed, based on a series of specimens collected from Mt Daming, Guangxi, southern China. The new species is close to N.yeae, N.daunchina, N.yaoica, and N.chapaensis from southwestern and south-central China and northern Indochina, while the relationships among these species remain unresolved. Nidiranaguangxiensis sp. nov. can be distinguished from all known congeners by the genetic divergences in the mitochondrial 16S and COI genes, the behavior of nest construction, the advertisement call containing 6–11 rapidly repeated regular notes, and a combination of morphological characteristics. Furthermore, the Nidirana populations recorded in Guangxi are clarified in this work, providing valuable new information on the knowledge of the genus Nidirana.


Introduction
The music frog genus Nidirana Dubois, 1992 was originally proposed as a subgenus of Rana Linnaeus, 1758. Later, Nidirana was controversially recognized as a full genus or a synonym of Babina Thompson, 1912(Chen et al. 2005Frost et al. 2006). Recently, comprehensive morphological, molecular, bioacoustic, and biogeographical evidence has resurrected Nidirana as a distinct genus (Lyu et al. 2017). The frogs of this genus usually inhabit the natural or artificial swamps, ponds, and paddy fields in the hilly regions of subtropical eastern and southeastern Asia, with some species having nest construction behavior when courting (Fei et al. 2009;Lyu et al. 2017). The known diversity of Nidirana increased dramatically from seven to 15 species since 2017 (Lyu et al. 2017(Lyu et al. , 2019(Lyu et al. , 2020a. Most of the newly described species were previously misidentified as other congeners, due to their conservative phenotypes (Lyu et al. 2019(Lyu et al. , 2020a(Lyu et al. , 2020b. For instance, Lyu et al. (2020b) revised multiple populations historically recorded as Nidirana adenopleura (Boulenger, 1909) from China. They suggested that only the populations from Taiwan, Jiangxi, Fujian, and southern Zhejiang are the true N. adenopleura, and nominated some other populations as three new species: N. guangdongensis Lyu, Wan & Wang, 2020, N. mangveni Lyu, Qi & Wang, 2020, and N. xiangica Lyu & Wang, 2020. Lyu et al.'s (2020b work did not clarify all historic records of N. adenopleura, and the taxonomic status for the records not involved in their study remains unresolved. The Nidirana populations in Guangxi Zhuang Autonomous Region, southern China, were previously recorded as N. adenopleura (Liu and Hu 1962;Zhang and Wen 2000;Fei et al. 2009;Mo et al. 2014). Fei et al. (2009) suspected this identification was not correct, but still tentatively followed it and suggested additional study. Recently, the population from Mt Dayao, eastern Guangxi, has been revealed as a new species, N. yaoica Lyu, Mo, Wan, Li, Pang & Wang, 2019, and the population from Mt Dupangling, northeastern Guangxi was assigned to N. xiangica (Lyu et al. 2019(Lyu et al. , 2020b. During our recent surveys in Guangxi, we collected a series of Nidirana specimens from Mt Daming (MDM), central Guangxi, and Mt Jiuwan (MJW), northern Guangxi ( Fig. 1). After comprehensive analyses, the specimens from MJW are identified as N. leishanensis Li, Wei, Xu, Cui, Fei, Jiang, Liu & Wang, 2019, while the specimens from MDM are herein proposed as a new species.

Phylogenetic analysis
Nine muscular samples of the unnamed species from Guangxi were used for molecular analysis, encompassing five samples from MDM and four from MJW. All samples were obtained from euthanized specimens and then preserved in 95% ethanol and stored at −40 °C. In addition, 33 sequences from all known Nidirana species and two sequences from the outgroup, Babina holsti (Boulenger, 1892) and B. subaspera (Barbour, 1908) (following Lyu et al. 2017), were obtained from Gen-Bank and incorporated into our dataset. Detailed information on these materials is shown in Table 1 and Figure 1.
Two mitochondrial genes, namely partial 16S ribosomal RNA gene (16S) and partial cytochrome c oxidase I gene (COI), were used for phylogenetic analysis. DNA extraction, PCR amplification, and sequencing conducted on the newly collected samples followed Lyu et al. (2019). Two gene segments, 1042 base pairs (bp) of 16S and 639 bp of COI, were concatenated seriatim into a 1681-bp matrix. The final alignment was partitioned by gene and COI was further partitioned by codon position. The partitions were tested in jmodeltest v. 2.1.2, resulting in the best-fitting nucleotide substitution models as GTR+I+G. Sequenced data were analyzed using maximum likelihood (ML) in RaxmlGUI v. 1.3 (Silvestro and Michalak 2012). The bootstrap consensus tree inferred from 1000 replicates was used to represent the evolutionary history of the taxa analyzed.

Morphological examination
Seventeen male and two female unnamed specimens collected from MDM were examined and measured, collection information is given in the taxonomic proposal. All specimens were fixed in 10% buffered formalin, transferred to 70% ethanol, and deposited in the Natural History Museum of Guangxi (NHMG) and the Museum of Biology, Sun Yat-sen University (SYS), China.
Morphological descriptions follow the consistent definition by Lyu et al. (2017Lyu et al. ( , 2019Lyu et al. ( , 2020aLyu et al. ( , 2020b. External measurements of specimens were made with digital calipers (Neiko 01407A stainless steel 6-inch digital calipers) to the nearest 0.1 mm. These measurements are as follows: SVL snout-vent length (from tip of snout to posterior margin of vent); HDL head length (from tip of snout to the articulation of the jaw); HDW head width (head width at the commissure of the jaws); SNT snout length (from tip of snout to the anterior corner of the eye); IND internasal distance (distance between nares); IOD interorbital distance (minimum distance between upper eyelids); ED eye diameter (from the anterior corner of the eye to posterior corner of the eye); TD tympanum diameter (horizontal diameter of tympanum); TED tympanum-eye distance (from anterior edge of tympanum to posterior corner of the eye); HND hand length (from the proximal border of the outer palmar tubercle to the tip of digit III); RAD radio-ulna length (from the flexed elbow to the proximal border of the outer palmar tubercle); FTL foot length (from distal end of shank to the tip of digit IV); TIB tibial length (from the outer surface of the flexed knee to the heel).
Sex and age were determined by examining the gonads. Webbing formula follows Savage (1975).
Particularly, since the new Nidirana species from MDM is geographically and phylogenetically close to N. yaoica, and phylogenetically close to N. yeae Wei, Li, Liu, Cheng, Xu & Wang, 2020, enhanced morphometric data of these three species were used for statistical analyses in R v. 4.0.0. Due to the limited number of females collected, only male specimens were used. Data of the MDM specimens were newly measured in this work; meanwhile data of N. yaoica and N. yeae were obtained from the literature (Lyu et al. 2019;Wei et al. 2020). All measurements were ln-transformed to normalize and reduce the variance. The t-test was conducted with statistically similar variances (p > 0.05 in the Levene's test) using car R package. Boxplots were visualized with the "ggplot2" R packages. For t-test and boxplots, measurements were scaled to remove allometric effects of body size in morphological analysis, using the following equation: X a = X lnβ • (SVL ln -SVL m ), where X a = adjusted value; X ln = ln-transformed measurements; β = unstandardized regression coefficient for each species; SVL ln = lntransformed SVL; and SVL m = overall average SVL ln of all samples. Principal component analysis (PCA) was performed to reduce the dimensionality of variation in the data to find whether morphological variation form the basis of detectable group structure, using the "prcomp" function and "ggplot2" package.

Bioacoustic analysis
Advertisement calls of the Nidirana population from MDM were recorded in the field at the air temperature of 18 °C by a Sony PCM D100 digital sound recorder on 20 April 2021. The recorded individuals were observed to ensure as the correct species but were not captured for conservation reasons. The sound files in wave format were sampled at 44.1 kHz with 24 bits in depth. Praat v. 6.0.27 (Boersma 2001) was used to obtain the oscillogram, sonogram, and power spectrum (window length = 0.005 s). Raven Pro v. 1.5 (Cornell Lab of Ornithology 2003-2014) was used to quantify the acoustic properties (window size = 256 points, fast Fourier transform, Hanning window with no overlap). The call duration (the time between onset of the first note and offset of the last note in a call) and call PF (peak frequency; the frequency at which max power occurs within the call) were measured for each call, and the note duration (the time between onset and offset of a note) and note interval (the time between adjacent notes in a call) were measured for each note.

Phylogeny
The result of ML analysis was given in Figure 2, in which the supportive nodes with the bootstrap supports (BS) > 90 were shown. This mitochondrial result is consistent with the phylogenic relationship from previous studies (e.g. Lyu et al. 2020a), with two species groups and four clades revealed. The Nidirana populations from MDM (ID 1-5) and MJW (ID 9-12) are both inserted in the Clade C (clade names following Lyu et al. 2017) of the N. adenopleura group, which are distant from the true N. adenopleura in Clade D in phylogeny. Within Clade C, the Nidirana population from MJW (ID 9-12) is clustered with samples of N. leishanensis from Mt Leigong and Mt Fanjing, Guizhou, with strong supports (BS = 100) and small divergences, which indicates the MJW population should be clarified as N. leishanensis. The Nidirana population from MDM (ID 1-5) forms an independent lineage with strong supports (BS = 100) and almost no divergence, which is close to but diverse from the lineages of N. yeae, N. daunchina (Chang, 1933), N. yaoica, and N. chapaensis (Bourret, 1937). The relationship among these five lineages remains unresolved, even though the MDM population seems closer to N. yeae with medium support (BS = 92).

Morphology
Detailed comparisons among all Nidirana species are listed in Table 2, which shows the distinct differences of the Nidirana specimens from MDM (detailed comparisons presented in the Taxonomic proposal below). The results of t-test and boxplots of morphometrics (Table 3; Fig. 3) show that the Nidirana specimens from MDM are significantly different from N. yeae from northern Guizhou, especially in the characteristics of HDL, HDW, IND, TD, RAD, FTL, and TIB, and different from N. yaoica from eastern Guangxi in the characteristics of HDL, HDW, SNT, ED, TD, and RAD. In the PCA analyses (Fig. 4), the extracted components PC1, PC2, PC3, and PC4 eigenvectors account for 46.4%, 17.5%, 11.7%, and 8.3% of the variance, respectively, or 83.9% cumulatively. As illustrated in the scatter plots of PC1 and PC2, samples of each species cluster together and do not overlap with each other.

Nidirana guangxiensis
Etymology. The specific name guangxiensis refers to the type locality of the new species in Guangxi Zhuang Autonomous Region. The Zhuang language, one of the official languages of Guangxi Zhuang Autonomous Region, is based on the dialect of Wuming, from where the new species was collected.
Diagnosis. Nidirana guangxiensis sp. nov. is placed in the genus Nidirana based on the morphological characteristics of the absence of the thumb-like structure on finger I, presence of well-developed dorsolateral folds, and the presence of suprabrachial glands in breeding males (Lyu et al. 2017). It is further assigned to the N. adenopleura group by the presence of lateroventral grooves on all toes (Dubois 1992;Lyu et al. 2019).
Particularly, Nidirana guangxiensis sp. nov. is relatively close in phylogeny to N. yeae from northern Guizhou, but it can be distinguished by: the relative fingers length II < I < IV < III [vs II < IV < I < III in N Description of holotype. NHMG 202007003 (Figs 6, 7A, B), adult male. Body medium-sized, SVL 43.8 mm; head relatively long and wide (HDL/SVL 0.42, HDW/ SVL 0.36), longer than wide (HDW/HDL 0.86), flat above; snout rounded in dorsal and lateral views, slightly protruding beyond lower jaw, longer than horizontal diameter of eye (SNT/ED 1.57); canthus rostralis distinct, slightly curved inwards on the nostril; loreal region concave; nostril round, closer to the snout than to the eye; a longitudinal swollen mandibular ridge extending from below nostril through lower edges of eye and tympanum to above insertion of arm, where the ridge is intermittent, forming a maxillary gland and shoulder gland; supratympanic fold absent; interorbital space flat, narrower than internasal distance (IND/IOD 1.31), pineal ocellus invisible; pupil elliptical, horizontal; tympanum distinct, round, relatively large, TD/ED 0.98; vomerine ridge present, bearing small teeth; tongue cordiform, margin of the tongue notched; a pair of subgular vocal sacs present.
Forelimbs moderately robust, lower arm 0.17 of SVL and hand 0.27 of SVL; fingers thin, relative finger lengths II < I < IV < III; tip of each finger slightly dilated, forming rounded disks; lateroventral grooves on fingers III and IV, not meeting at the tip of disks; fingers free of webbing; lateral fringes present and distinct on inner and outer sides of fingers II, III and IV, but absent on finger I; subarticular tubercles prominent and rounded; supernumerary tubercles present below the base of fingers III and IV; palmar tubercles three, elliptic, large, prominent and distinct; a single nuptial pad on the dorsal surface of finger I, nuptial spinules invisible.
Hindlimbs robust, tibia 0.53 of SVL, and foot 0.76 of SVL; heels overlapping when hindlimbs flexed at right angles to axis of body; tibio-tarsal articulation reaching at the nostril when hindlimb is stretched along the side of the body; toes relatively long and thin, relative lengths I < II < V < III < IV; tip of each toe slightly dilated with remarkable elongated ventral callous pad, forming long and pointed disk; lateroventral grooves well developed on each toe, not meeting at the tip of disks; webbing moderate, formula: I 1⅓-2 II 1⅓-2⅓ III 1⅔-3 IV 3⅓-1⅓ V; lateral fringes present on inner and outer sides of each toe, forming distinct dermal flap on the lateral edges of toes I and V; subarticular tubercles rounded, prominent; inner metatarsal tubercle elliptic, length triple width; outer metatarsal tubercle indistinct, small and rounded; tarsal folds present and tarsal tubercle absent.
Dorsal skin rough with dense granules, several tubercles on the posterior part, flanks, and dorsal hindlimbs, not bearing spinules on the skin; developed dorsolateral fold from posterior margin of upper eyelid to above groin but intermittent posteriorly; a large and smooth suprabrachial gland behind base of forelimb, prominent; weak longitudinal ridges on upper arms and slightly extending to lower arm; the dorsal surfaces of thigh and tibia relatively rough with tubercles, forming several longitudinal ridges. Ventral surface of throat, body, and limbs smooth; large flattened tubercles densely arranged on the rear of thigh and around vent.
Color of holotype. In life (Fig. 7A, B), dorsal surface of head and body brown; a longitudinal light brown mid-dorsal stripe faintly beginning from interorbital area, extending posteriorly to vent and become more distinct; several black spots on posterior dorsum of body; dorsolateral fold brown; upper flank brown with small black spots; lower flank light brown; suprabrachial gland yellowish brown. Dorsal forelimbs brown; dorsal hindlimbs brown, two olive crossbars on the thigh, three on the tibia, and three on the tarsus; irregular olive marks on dorsal toes. Loreal and temporal regions dark brown, tympanum pink; upper ⅓ iris brownish white and lower ⅔ iris reddish brown; maxillary gland and shoulder gland creamy white. Lips, throat, ventral surface of body and limbs creamy white; rear thigh tinged with pink and pale grey patches; ventral hand and foot pale white.
In preservative (Fig. 6), dorsal surface becomes dark brown with the mid-dorsal stripe and black spots more distinct; flank surface and the suprabrachial gland become pale; crossbars and marks on limbs dark brown; loreal and temporal regions dark brown; maxillary gland and shoulder gland more distinct; ventral surface pale grey; rear thigh and ventral foot become dark grey.
Variations. Measurements of type series are given in Table 4. All specimens were similar in morphology. Females are significantly larger than males, with relatively smoother skin and fewer tubercles on dorsum and flanks. The colorations vary from pale brown to reddish brown in individuals (Fig. 7C-F). The patterns of mid-dorsal stripes are also variable but always present.
Male secondary sexual characteristics. A pair of subgular vocal sacs, a pair of slitlike openings at posterior of jaw; a single light brown nuptial pad on the dorsal surface of finger I, nuptial spinules invisible; suprabrachial gland present.
Tadpole. Body length 19.1 mm and tail length 43.1 mm in the 37 th stage tadpole SYS a008814 (Fig. 8); body oval, flattened above; snout rounded in dorsal aspect and profile; eyes lateral; labial tooth row formula: 1:1+1/1+1:2; spiracle on left side of body, directed dorsoposteriorly; tail depth larger than body depth; dorsal fin arising just before origin of tail, maximum depth near mid-length, tapering gradually to narrow pointed tip. Distribution and ecology. Currently, Nidirana guangxiensis sp. nov. is known only from the type locality, Mt Daming, which is located between Wuming District and Shanglin County, Nanning, Guangxi (Fig. 1). This species of frog can only be found in the alpine swamp and neighboring brushwood on the peak of Mt Daming. The estimated extent of occurrence is less than 500 km 2 , and the estimated area of occupancy is less than 50 km 2 . The swamp was surrounded by subtropical evergreen broadleaf forests (Fig. 9A). Sympatric frog species observed in the swamp are Duttaphrynus melanostictus (Schneider, 1799), Gracixalus jinxiuensis (Hu, 1978), Kurixalus odontotarsus (Ye & Fei, 1993), and Polypedates mutus (Smith, 1940).
Nidirana guangxiensis sp. nov. was observed to have nest construction behavior. The nest is in the form of a mud burrow ca 25-30 mm in diameter and near the roots of plants. The top of the nest is open and may fill with water during the rainy season (Fig. 9B). From April to August, males call from dusk to midnight in the nest. In late April, tadpoles at the 26 th -42 nd stages can be observed, with the majority at the 33 rd -37 th stages.

Discussion
With this work, the historically recorded populations of Nidirana adenopleura from Guangxi, are all reassigned to other recently described species, namely N. yaoica from Mt Dayao in the east, N. xiangica from Mt Dupangling in the northeast, N. leishanensis from Mt Jiuwan in the north, and Nidirana guangxiensis sp. nov. from Mt Daming, central Guangxi (Fig. 1). Among them, Nidirana guangxiensis sp. nov. is phylogenetically close to N. yaoica, while N. xiangica and N. leishanensis are sister species (Fig. 2). The complex rivers and mountainous systems in Guangxi may play as important barriers to the speciation of these species pairs.
As indicated by the etymology of the generic epithet (Dubois 1992), some species of Nidirana were observed with the behavior of nest construction (Nidirana guangxiensis sp. nov., N. okinavana, N. nankunensis, N. hainanensis, N. chapaensis, and N. daunchina). According to our field observations, these nest-constructing species usually live in natural swamps and ponds with muddy bottoms (Fei et al. 2009(Fei et al. , 2012Lyu et al. 2017;this work). Such habitats obtain seasonal rainfall and unpredictable water accumulation, which implies that constructing a nest would be helpful for the growth of the eggs and tadpoles. In contrast, the congeners without such behavior (N. adenopleura,N. guangdongensis,N. leishanensis,N. lini,N. mangveni,Figure 8. Tadpole SYS a008814 of Nidirana guangxiensis sp. nov. Photos by Shuo Qi. N. occidentalis, N. pleuraden, and N. xiangica) usually inhabit natural or artificial ponds and paddies with perennial water, which allows them to directly oviposit into the water (Fei et al. 2009(Fei et al. , 2012Lyu et al. 2020aLyu et al. , 2020b. Additionally, the nest construction behaviors of two other congeners are still unknown (N. yaoica and N. yeae; Table 2), but to roughly illustrate and compare the reported ecological data which is correlated to such courtship behavior, N. yaoica living in seasonal swamps (Lyu et al. 2019) is likely to construct nests, and N. yeae inhabiting paddy field with tadpoles observed at the water surface (Wei et al. 2020) may not possess such behavior. Regarded as important for breeding, this behavior was used for the species-group divisions (Fei et al. 2009;Chuaynkern et al. 2010). Nevertheless, Lyu et al. (2019) revised the species groups based on phylogenetic results and found that the behavior of nest construction seems to evolve independently in different clades. As an infrequent habit in the family Ranidae, the evolution of nest construction behavior in Nidirana species would be a topic worth studying and requires more ethological and ecological work and the application of genomic data.
Based on the phylogenetic relationships, Lyu et al. (2017) partitioned the genus Nidirana into four robustly supported clades (Fig. 2). Clade A corresponds to the N. pleuraden group with two recognized species (Lyu et al. 2019(Lyu et al. , 2020a, while the other clades belong to the N. adenopleura group (Lyu et al. 2019). Clade B is monotypic and includes only N. lini, clade D is comprised of four species, and clade C includes nine species which are more than half the members of the genus. By bringing the phylogenetic analyses from this work and previous studies (Lyu et al. 2019(Lyu et al. , 2020a, the interspecies relationships within clade C are unclear due to the relatively lower supported values in mitochondrial genes. Species of clade C are mostly distributed in the hilly regions throughout southwestern and south-central China and northern Indochina (Fig. 1), at the edge of the Indo-Burma biodiversity hotspot. In view of the extensiveness of these hilly areas and the unclear relationship within this clade, Nidirana diversity in these areas seems still underestimated, which suggests that further surveys are required.