A New Species of the Genus Microhyla (Amphibia: Anura: Microhylidae) from the Dabie Mountains, China

Simple Summary In China, new species of Microhylidae continue to be discovered at a rapid rate, often as a result of re-examining geographically widespread species populations using new molecular and bioacoustic tools. Here, we show that members of the genus Microhyla from the Dabie mountains in East China can be distinguished from morphologically similar species (M. beilunensis, M. fanjinshanensis and M. mixtura) with the analysis of phylogeny, species delimitation, bioacoustics and morphology. Based on the above multiple lines of evidence, we describe the population of the Dabie Mountains as a new species, Microhyla dabieshanensis sp. nov. This study enriches the diversity of Microhylidae, and clarifies the species of the genus Microhyla in the Dabie Mountains. Microhyla dabieshanensis sp. nov. is distributed in the Dabie Mountains above 500 m above sea level, and in some areas has sympatric distribution with M. fissipes and M. heymonsi. Abstract Species belonging to the genus Microhyla are small-sized frogs that are widely distributed in southern, eastern, and south-eastern Asia. In China, the genus harbors many cryptic species, on two of which—M. beilunensis and M. fanjinshanensis—studies were recently published. In this study, we collected specimens from the Dabie Mountain range, which is at the junction of Anhui, Henan and Hubei Provinces, East China; these specimens belonged to a species previously identified as M. mixtura. Based on phylogenetic analyses, species delimitation analyses, morphological comparisons and advertisement calls comparisons, we found they were significantly different from other known congeners, and thus we describe them as a new species. This study enriches the diversity of Microhylidae, and clarifies the species of the genus Microhyla in the Dabie Mountains.

In China, nine species (M. beilunensis, M. berdmorei, M. butler, M. fanjingshanensis, Microhyla fissipes, M. heymonsi, M. mixtura, M. mukhlesuri and M. pulchra) of Microhyla have been recorded so far, and are widely distributed in the south and southeast of China [11]. M. beilunensis, M. fanjingshanensis and M. mixtura are endemic to China [1]; M. mixtura, especially, has a wide distribution in China (Anhui, Henan, Guizhou, Hubei, Chongqing, Shaanxi, Sichuan and Zhejiang Provinces) [12,13]. Previous studies have found that M. mixtura in different geographical groups have differences in morphology such as skin texture, patterns and spots, supratympanic fold and canthus rostralis [12][13][14]. 2 of 18 From 2019 to 2021, we collected nineteen Microhyla specimens from the Dabie Mountains of Anhui Province, China, which superficially resembled M. mixtura. Based on phylogenetic analyses using DNA sequences of the mitochondria (12S rRNA, 16S rRNA and COI genes), the specimens more likely represent a cryptic species in this study. In addition, the subsequent morphological comparisons and bioacoustics studies all consistently suggested that the Dabie Mountains specimens were distinctly different from known Microhyla species, and we therefore describe these specimens as a new species here.

Sampling
Between May 2019 and September 2021, we collected 19 adult specimens, comprising 5 females and 14 males, from the Dabie Mountains of Anhui Province, China ( Figure 1). In addition, one adult specimen of M. fissipes from Dabie Mountains of Anhui Province was collected for comparison. The specimens were euthanized, photographed and fixed in 4% formaldehyde for two days, and then finally washed and preserved in 70% ethyl alcohol. Voucher specimens for this work were deposited at the Anhui University Biology Museum (AHUBM). Voucher specimens were deposited at the Museum of Anhui University. especially, has a wide distribution in China (Anhui, Henan, Guizhou, Hubei, Chongqing, Shaanxi, Sichuan and Zhejiang Provinces) [12,13]. Previous studies have found that M. mixtura in different geographical groups have differences in morphology such as skin texture, patterns and spots, supratympanic fold and canthus rostralis [12][13][14].
From 2019 to 2021, we collected nineteen Microhyla specimens from the Dabie Mountains of Anhui Province, China, which superficially resembled M. mixtura. Based on phylogenetic analyses using DNA sequences of the mitochondria (12S rRNA, 16S rRNA and COI genes), the specimens more likely represent a cryptic species in this study. In addition, the subsequent morphological comparisons and bioacoustics studies all consistently suggested that the Dabie Mountains specimens were distinctly different from known Microhyla species, and we therefore describe these specimens as a new species here.

Sampling
Between May 2019 and September 2021, we collected 19 adult specimens, comprising 5 females and 14 males, from the Dabie Mountains of Anhui Province, China ( Figure 1). In addition, one adult specimen of M. fissipes from Dabie Mountains of Anhui Province was collected for comparison. The specimens were euthanized, photographed and fixed in 4% formaldehyde for two days, and then finally washed and preserved in 70% ethyl alcohol. Voucher specimens for this work were deposited at the Anhui University Biology Museum (AHUBM). Voucher specimens were deposited at the Museum of Anhui University

Molecular Phylogenetic Analyses
Tissue samples from six individuals of the new taxon were used for the phylogenetic analyses. Total genomic DNA was extracted from muscle or liver tissue based on a standard phenol/chloroform extraction method [15]. For molecular analysis, the mitochondrial genes (12S rRNA, 16S rRNA and COI) of six samples were amplified and sequenced (see Table S1). The PCR procedures and primers we used were from previous studies [16][17][18]. Experimental primer pairs were synthesized and provided by General Biosystems (Anhui) Co., Ltd. (Anhui, China) The conditions for PCR amplification are as follows: initial Animals 2022, 12, 2894 3 of 18 denaturation step, 95 • C, 4 min; the second step is 35 cycles at 94 • C, 35 S, annealing at 46-52 points; then, expand at 72 • C for 1 min; finally, extend the steps for 10 min of 72 • C. PCR amplification products were sent to General Biosystems (Anhui) Co., Ltd. The final sequence was submitted to GenBank (GenBank Accession numbers shown in Table S1).
SeqMan software (version 7.1.0, Madison, WI, USA) was used to splice the sequences obtained by sequencing and check whether there were any errors, which were manually corrected. MEGA software (version 7.0, Mega Limited, Auckland, New Zealand) [19] was used to compare the sequenced and spliced sequences with the downloaded apochromat sequences and adjust the sequences appropriately.

Phylogenetic Analyses
Phylogenetic analyses, involving 12S rRNA-16S rRNA-COI sequences of 49 Microhyla species and 8 Nanophla species, were downloaded from GenBank and used for phylogenetic reconstruction. Phylogenetic trees were constructed using maximum likelihood (ML) and Bayesian inference (BI) analyses.
According to previous studies, we chose Dyscophus antongilii and Kaloula verrucosa as the outgroups [20]. Phylogenetic analysis was performed using maximum likelihood (ML) and Bayesian inference (BI) methods, which were implemented in PhyML (version 3.0, France Génomique) [21] and MrBayes (version 3.2.2, Stockholm, Sweden) [22], respectively. PAUP (version 4.9, Gainesville, FL, USA) [23] software was used to construct the Bayesian system analysis under the MRModeltest 2.0 [23] program, and determine the best model parameters (minimum-lnL) with the nucleotide alternative model and the replacement rate heterogeneity test based on the Akaike information criterion (AIC) parameter standard. The best model parameters were selected to modify the Bayes program running file format, and MrBayes3.2.2 software was used to construct a Bayes phylogenetic tree based on the Markov chain Monte Carlo (MCMC) algorithm [24]. The results showed that the GTR+I+G model was the best partition. In the BI analysis, four Markov chains were used for two runs over three million generations, sampling every 300 generations. The first 25% of generations were burned in, and the remaining trees were used to generate the 50% strictly consistent tree. The posterior probability (PP) of each node in the phylogenetic tree was also calculated. All trees had to have an error of less than 0.01. For ML trees, branch support was extracted from 10,000 nonparametric bootleg replicates. By default, tree nodes in the ML tree with bootstrap values of 75% or greater were considered sufficiently resolved [25,26].
Genetic distances between and within species were calculated using uncorrected p-distances with 16S rRNA implemented in MEGA v7.0.

Morphological Analyses
The terminology and methods followed Fei et al. (2009) [12]. Morphological measurements were taken with a digital caliper to the nearest 0.01 mm. A total of 16 morphological indicators were measured, and the abbreviations of morphological characteristics used in the text are as follows: We also compared the morphological characteristics of the collection species with all other Microhyla genera. Morphological data for comparisons were obtained from the relevant literatures [2][3][4][5][6][7][8][9][10][11][12][13].
The collection species, M. beilunensis and M. fanjingshanensis, were previously classified as M. mixtura (Wu et al., 1986;Fei et al., 2009Fei et al., , 2012. When comparing the population morphology characteristics of our collection species with other known species, it is important to compare these three species that once belonged to M. mixtura (M. mixtura, M. beilunensis and M. fanjingshanensis). Therefore, we focused on conducted morphological comparisons to explore the morphological differences between these four species. A total of 19 specimens of our collection taxon, comprising 14 males and 5 females from the Dabie Mountains, were measured. At the same time, data on 8 males of M. mixtura [4], 37 males and 4 females of M. beilunensis [4], and 14 males and 2 females of M. fanjingshanensis [5] were obtained from the available literature (voucher information: see Table S2). To reduce the influence of allometric growth in afterward morphometric analyses, we calculated the size-correction value of each morphological trait as the ratio to SVL. Wilcoxon rank sum tests were conducted to test the significance of the differences of the morphometric characteristics between the different sexes, as well as between different species. The significance level was set at 0.05. Furthermore, we then employed principal component analysis (PCA) to explore the morphological differences between the populations of the Dabie Mountains-M. beilunensis, M. fanjingshanensis and M. mixtura-and checked whether the different species were consistently separated in morphological characteristics. All above analyses were carried out in R 4.0.2. All morphometric analyses were conducted separately for the male and female groups. Because we did not collect females of M. mixtura, only the collection taxon and M. beilunensis and M. fanjingshanensis females were included.  Figure 1). Recording the subjects' calls was conducted using a Sennheiser ME66 microphone (Sennheiser, Wedemark, Germany). The microphone must be placed 0.2 m from the subject for the best recording quality and mounted on a long bamboo rod to keep it at the right height. The microphone was connected to a laptop (Thinkpad X201; Lenovo, Beijing, China) with a sampling rate of 44.1 KHz and a 16-bit resolution. Adobe Audition 3.0 software (San Jose, CA, USA) was used to gather relevant data, such as amplitude-modulated waveforms (oscillograms) and audio spectrograms of male advertisement calls. The methods for data analysis were similar to those described previously [67]. The ambient temperature of the type locality was taken by a digital hygrothermograph.

Phylogeny Analysis and Species Delimitation
The aligned dataset contained 39 individuals of the Microhyla species, 8 individuals of the Nanohyla species and two individuals of the outgroup species (Table 1). The aligned matrix of the mitochondrial 12S + 16S + COI genes contained 1412 bp, of which 933 sites were conserved, 473 sites exhibited variation, and 379 were found to be potentially parsimony-informative. Maximum likelihood (ML) and Bayesian inference (BI) analyses obtained similar topologies ( Figure 2) that differed only at some poorly supported basal nodes. This ML tree could be divided into two major branches based on topology display, corresponding to Microhyla and Nanohyla, respectively. In the phylogenetic tree, all six new samples were aggregated into a monophyletic group, which was sister to M. mixtura and M. okinavensis with high nodal support values (1.00/99). Overall, the phylogenetic relationship of Microhyla species obtained in this study is consistent with the results of previous related studies [2,4,5,7].
The genetic distances (uncorrected p-distance based on the 16S rRNA genes) between the new samples and their congeners ranged from 2.0% (vs. M. okinavensis) to 11.4% (vs. M. zeylanica) ( Table S3) new samples were aggregated into a monophyletic group, which was sister to M. mixtura and M. okinavensis with high nodal support values (1.00/99). Overall, the phylogenetic relationship of Microhyla species obtained in this study is consistent with the results of previous related studies [2,4,5,7].   Table S1.

Morphology
In morphological comparison, the results of the Wilcoxon rank sum test indicated that the new taxon male group was significantly different from those of M. beilunensis, M. fanjingshanensis and M. mixtura in many morphometric characteristics (all p-values < 0.05; Table S4). Unfortunately, the new taxon's female group was not significantly different from those of M. beilunensis and M. fanjingshanensis according to the Wilcoxon rank sum test, in many morphometric characteristics (Table S4), which may be due to an insufficient number of samples. PCA extracted principal component factors with Eigenvalues greater than four in males and two in females. The first four principal components explained 81.72% of the total variation in the males, and the first two principal components explained 86.08% of the total variation in the females ( Table 2). These results were mainly affected by morphological characters, including SVL, HDL, HDW, ED, LW, TL, TW and TFL (the corrected relevant values shown in Table S4). Both the male and female groups of the four aforementioned species can be clearly separated on the two-dimensional graphs of PC1 and PC2 (Figure 3).

Bioacoustics
Ten advertisement calls of the new species were recorded in the Qingfengling Village, Huoshan County, Luan City, Anhui Province (Figure 1), on 22 April 2021 between 20:00 and 22:00. The advertisement description is based on recordings of AHU2021QFL01 ( Figure 4) from a field near the streamlet at an ambient air temperature of 13 • C and air humidity of 90.8%. Each call had 6-9 notes (mean 7.5 ± 0.849, n = 10). The call duration was 0.598-0.863 s (mean 0.744 ± 0.083, n = 10), with intervals ranging from 1.231 to 2.253 s (mean 1.656 ± 0.347, n = 10). Amplitude modulation within each note was apparent, beginning with moderately high energy pulses, increasing to the maximum and then decreasing towards the end. The average dominant frequency was 2764 ± 102.96 (2670-2971 Hz, n = 10). At the same time, we also collected the acoustic data of the new specimen's close relative species (M. beilunensis, M. fanjingshanensis and M. mixtura) distributed in China.
There were three differences in the sonograms and call waveforms among the new species and its close relative species, M. beilunensis, M. fanjingshanensis and M. mixtura (Figure 4; Table 3). The number of notes called by the new species (mean 7.5 ± 0.8) was lower than those of M. mixtura (mean 8.5 ± 0.8) and M. fanjingshanensis (mean 10.4 ± 0.8), but was higher than that of M. beilunensis (mean 6.0 ± 1.

Bioacoustics
Ten advertisement calls of the new species were recorded in the Qingfengling Village, Huoshan County, Luan City, Anhui Province (Figure 1  was 0.598-0.863 s (mean 0.744 ± 0.083, n = 10), with intervals ranging from 1.231 to 2.253 s (mean 1.656 ± 0.347, n = 10). Amplitude modulation within each note was apparent, beginning with moderately high energy pulses, increasing to the maximum and then decreasing towards the end. The average dominant frequency was 2764 ± 102.96 (2670-2971 Hz, n = 10). At the same time, we also collected the acoustic data of the new specimen's close relative species (M. beilunensis, M. fanjingshanensis and M. mixtura) distributed in China. There were three differences in the sonograms and call waveforms among the new species and its close relative species, M. beilunensis, M. fanjingshanensis and M. mixtura ( Figure 4; Table 3). The number of notes called by the new species (mean 7.5 ± 0.8) was lower than those of M. mixtura (mean 8.5 ± 0.8) and M. fanjingshanensis (mean 10.4 ± 0.8), but was higher than that of M. beilunensis (mean 6.0 ± 1.  These three main differences in bioacoustics can be distinguished from the three close relative species, M. beilunensis, M. fanjingshanensis and M. mixtura, and support the new specimen as a new species.

Taxonomic Account
The results of the molecular phylogenetic analyses, morphological comparison and bioacoustics all indicated that the new taxon of the genus Microhyla from the Dabie Mountains is significantly different from other known species of the same genus. Therefore, we describe it here as a new species.

Holotype Description
Adult males have a medium body size (SVL 20.9 mm), with a head wider than its length (HDL/HDW 83.6%), a snout short that is rounded, projecting beyond the lower jaw, and distinct rounded nostrils that are obviously closer to the tip of the snout than the eye. The interorbital distance (2.4 mm) is greater than the internasal distance (1.7 mm), and the upper eyelid width is 1.6 mm with an eye diameter of 2.2 mm. The eyes are small and slightly protuberant; their diameter makes up 43.1% of head length, and the pupils are round. This species also has hidden tympanum, with inconspicuous canthus rostralis and supratympanic fold. Their tongue is posteriorly oval and not notched behind, and they do not have maxillary or vomerine teeth.
Their forelimbs are relatively short and thin-about four times shorter than the snoutvent length (25.8% of SVL), and their fingers are short and slender, distally rounded, unwebbed, and have disks but absent dorsal longitudinal grooves. The relative finger lengths are: I < II < IV < III. There are distinct subarticular tubercles that are roughly circular. The supernumerary tubercles below the base of the finger are absent, although there are two palmar tubercles, with an elliptical inner palmar tubercle and round outer palmar tubercle slightly larger than the inner palmar tubercle ( Figure 5H). The nuptial pads are absent.
The hindlimbs are slender and comparatively long (HLL/SVL 189.5%), the tibia length is longer than half of the snout-vent length (TL/SVL 54.8%) and one-third of the hindlimb length (TL/HLL 29.0%). The heels overlap when the hindlimbs are at a right angle to the body, and the tibiotarsal articulation of the adpressed limb reaches the level between the eye and nostril when the leg is stretched forward. The foot length is equal to that of the tibia (FL/TL 1), and the relative toe lengths are: I < II < V < III < IV. The toes are distally rounded, equipped with disks and longitudinal grooves except for toe I, and the toes with have rudimentary webbing with absent supernumerary tubercles.
The dorsal skin surfaces of the head and body are slightly smooth with sparse tiny tubercles in the cloacal region. There are also thigh granules, although the throat, chest, and ventral part of the thigh and tibia are smooth.

Holotype Coloration
The dorsal body is reddish brown or grayish, with obvious dark brown markings and a light brown outline. A thick, dark brown inverted triangle marking sits between the eyes and on the dorsum, along with three pairs of symmetrical figures or stripes where the anterior pair of figures are shorter and extend laterally to the middle of the upper eyelid; three pairs connect to each other, with the middle pair being relatively long and extending laterally to the shorter. The posterior pair of stripes are the longest, running obliquely from above the shoulder to the hip; the dorsum of the body's posterior part, including the legs, is brownish, and scattered with dense, rounded dark brown irregular spots, as well as bar-shaped patterns and streaks with light margins; the supratympanic fold is dark brown; there are more continuous dark stripes on the side; the limbs are indefinitely barred in dark brown; the throat is dark purple, and the chest and belly are white-mottled with black.

Intraspecific Morphological Variations
The 14 adult male and 5 adult female specimens had nearly similar morphological features. The basic statistics of morphological measurement are shown in Table S2. However, the dorsum coloration and the color and shape of the stripes were considerably different between individual specimens ( Figure 6). In some specimens, the stripes on the dorsum were shallow or narrow, and asymmetrical or discontinuous in the posterior pair of stripes. Light-colored spots were unclear on the venter in some individuals. Many adult male individuals have a white-mottled brown venter and a vaguely visible V-shaped white stripe on the upper midsection. Some individuals have a round light spot in the middle of the stripe on the dorsum.

Sexual Dimorphism
Adult female snout-vent length is significantly larger than that of adult males. Adult males have slit-like openings to a median subgular vocal sac, and absent nupital pads and spines.

Distribution and Habit
To date, Microhyla dabieshanensis sp. nov. is only known from its type locality in the Dabie Mountains, China (Figure 1), and this species' distribution is in the water ditches or rice fields or adjacent grass thickets at elevations of 535-1300 m a. s. l., sympatric with M. fissipes and M. heymonsi.

Morphological Comparisons
Microhyla dabieshanensis sp. nov. is morphologically most like M. beilunensis, M. fanjingshanensis, M. kuramotoi, M. mixtura and M. okinavensis. However, Microhyla dabieshanensis sp. nov. could be distinguished from these species via smoother skin with few tubercles and granules, and the canthus rostralis and supratympanic fold were not obvious.
Microhyla dabieshanensis sp. nov. differs from M. beilunensis in that it has a vaguely visible V-shaped stripe in the upper middle of the chest (vs. without stripe), higher ratios of HDL, SL, LAL, LW, HAL, HLL, TFL and FL to SVL (p-values < 0.05; Table S4).
Microhyla dabieshanensis sp. nov. significantly differs from M. kuramotoi as follows: Microhyla dabieshanensis sp. nov. is distinct from M. mixtura, as it has: (1) the vaguely visible V-shaped stripe (vs. without the stripe); (2) posterior pairs of stripes on the dorsum with the longest on the belly, which is white-mottled mixed brown (vs. white); (3) higher ratios of HDL, SL, UEW, LW and HLL to SVL, and lower ratios of ED and TL to SVL in males (all p-values < 0.05; Table S4).
Microhyla dabieshanensis sp. nov. differs from M. okinavensis in the following ways:

Discussion
In this study, a new species, Microhyla dabieshanensis sp. nov., was described in the Dabie Mountains area, China, which had previously been identified as M. mixtura [14]. In fact, M. dabieshanensis sp. nov. should represent an independent lineage distinctly separate from M. mixtura, M. okinavensis and other congeners in the phylogeny tree ( Figure 2). In addition, it owns distinctive morphological characteristics and advertisement calls (Figures 3 and 4).
The discovery of Microhyla dabieshanensis sp. nov. has brought the total number of known species in the genus Microhyla to 49 and the number of species in China to 10 [1,11]. Therefore, based on our results and those of previous studies [4,5,8,12,13], we could further complete distribution conclusions on these six phylogenetically closely related species. According to the distribution information of the collection sites of the 19 specimens, it is speculated that M. dabieshanensis sp. nov. is distributed throughout the Dabie Mountains above 500 m above sea level. According to our current survey results, three species of Microhyla, M. dabieshanensis sp.nov., M. fissipes and M. heymonsi are sympatric in the Dabie Mountains area. Therefore, further research is needed to elucidate the true distribution range and ecological niche of the new species, which is crucial for studying the interspecific relationships and sympatric distributions of the three Microhyla species in this area.

Conclusions
We described a new species of the microhylid frogs, Microhyla dabieshanensis sp. nov., from the Dabie mountains in East China, using an integrative taxonomic approach.  Table S3. Uncorrected p-distance on 16S rRNA gene between Microhyla species included in the phylogenetic analyses (below the diagonal), and standard error estimates (above the diagonal). Table S4