Morphology, Molecular Genetics, and Bioacoustics Support Two New Sympatric Xenophrys Toads (Amphibia: Anura: Megophryidae) in Southeast China

Given their recent worldwide declines and extinctions, characterization of species-level diversity is of critical importance for large-scale biodiversity assessments and conservation of amphibians. This task is made difficult by the existence of cryptic species complexes, species groups comprising closely related and morphologically analogous species. The combination of morphology, genetic, and bioacoustic analyses permits robust and accurate species identification. Using these methods, we discovered two undescribed Xenophrys species, namely Xenophrys lini sp. nov. and Xenophrys cheni sp. nov. from the middle range of Luoxiao Mountains, southeast China. These two new species can be reliably distinguished from other known congeners by morphological and morphometric differences, distinctness in male advertisement calls, and substantial genetic distances (>3.6%) based on the mitochondrial 16s and 12s rRNA genes. The two new species, together with X. jinggangensis, are sympatric in the middle range of Luoxiao Mountains but may be isolated altitudinally and ecologically. Our study provides a first step to help resolve previously unrecognized cryptic biodiversity and provides insights into the understanding of Xenophrys diversification in the mountain complexes of southeast China.


Introduction
Accurate taxonomic recognition is a prerequisite for preserving amphibian biodiversity, especially in the context of amphibian declines and extinctions worldwide [1]. This fundamental task is challenged by the existence of cryptic species complexes [2], a group consisting of two or more species that are reproductively isolated from each other but virtually identical in morphology [3]. Frogs and other groups of Amphibians are known to harbor substantially underestimated cryptic species diversity [4]. Hence unambiguous species delineation may be difficult in some frog groups based on morphological characteristics exclusively [5], but it is very important to provide a solid basis for conservation management, as well as deeper understanding of the macroevolutionary patterns in amphibians [6].
The horned toads, Megophrys Kuhl & Van Hasselt, 1822 and Xenophrys Günther, 1864, in the family Megophryidae, are an exemplary group with high cryptic species diversity [7][8][9][10], making their systematics and taxonomies poorly understood and considerablly debated, even though herpetologists have employed various taxonomic methods [8,[11][12][13][14]. Pending comprehensive phylogenetic and morphological research, we followed the recommendations of Li & Wang [13] and Pyron and Wiens [15] that Xenophrys is distinguished from Megophrys and all previously known Megophrys species in China and should be assigned to the genus Xenophrys. Currently, the genus Xenophrys contains 46 species and is distributed in Southeast Asia from the southern and eastern Himalayan regions to Borneo [16]. There are 31 species of Xenophrys recognized from China, among which only five have a body length less than 50 mm in both males and females in southeast China [10]. Species in this group include X. boettgeri and X. kuatunensis in the Wuyi Mountains, X. huangshanensis in the Huangshan Mountains, and the recently described X. jinggangensis from Mount Jinggang (26u139-26u529N, 113u599-114u189E), situated in the border between the Jiangxi and Hunan provinces [10]. Noticeably, three new species of Megophryid toads described from northeast India very recently [9]. Together with X. jinggangensis, these discoveries raise the possibility of new cryptic species that might be discovered in Southeast Asia and China. However, due to the morphological similarity of Megophryid toads, only morphological characters may not be sufficient to diagnose taxonomies. To implement unbiased species delineation in cryptic amphibians, integrating evidence from morphology, DNA sequence data, and behavior may be critically necessary [17].
In particular, molecular genetic approaches enable us to decipher phylogenetic relationships and thus the evolutionary history of a large number of species, thereby solving taxonomic uncertainties [18]. In addition, bioacoustic analysis is a very useful approach in species diagnosis because many male frogs and toads have species-specific advertising vocalization during breeding seasons [19]. Like birds, these sounds are courtship signals used for mate choice by females and male-male competition, and thus presents an important behavioral pre-mating reproductive barrier under sexual selection [20]. Indeed, combined multiple lines of evidence from different methods assist in discovering unrecognized amphibian species regularly and this substantially inflates amphibian species diversity across the world, e.g. [21][22][23][24].
To understand the distribution and ecology of the newly described X. jinggangensis, we carried out extensive herpetological surveys during 2011-2013 around the middle range of Luoxiao Mountains, SE China. Interestingly, we also found two small and unknown Megophryid toads. Both of these species are small in body length (,45 mm), they can be assigned to the genus Xenophrys based on the following characteristics: head broad and depressed, tympanum distinct, tubercles on the outer edge of the upper eyelids short, tubercles on the snout absent, no mid-dorsal fold, no black horny spines on dorsum, hind limbs long, and heels overlap [13]. Furthermore, we also noticed that the vocalizations and phylogenetic relationships of these two unknown Xenophrys toads seemed to be distinct from that of X. jinggangensis. Thus, we conducted morphological, bioacoustics, and molecular genetic analyses to resolve the taxonomic status and affinities of these two taxa. Based on all these evidence, we describe two new species from southeast China.

Ethics Statement
Permissions to visit the study sites were issued by the management administration of the reserves. We obtained permissions for specimen and tissue collection from Jiangxi Provincial Forestry Bureau. This study did not involve endangered or protected species. All the animal operations were approved by the Institutional Ethical Committee of Animal Experimentation of Sun Yat-sen University and strictly complied with the ethical conditions by the Chinese Animal Welfare Act (20090606).

DNA sample collection
To reconstruct the phylogenetic relationships among Xenophrys species in southern China, we collected samples of X. jinggangensis, X. cheni sp. nov., and X. lini sp. nov. from the middle range of Luoxiao Mountains, situated in the border between the Jiangxi and Hunan Provinces, X. brachykolos from Hong Kong, X. boettgeri from Mt. Yangjifeng and Mt. Tongbo, Jiangxi Province, X. kuatunensis from Guadun (Kuatun) Village, Fujian Province, situated in the Wuyi Mountains, X. huangshanensis from Wuyuan County, Jiangxi Province, situated in the Huangshan Mountains, X. mangshanensis from Mt. Nanling, Guangdong Province, and X. minor from central Sichuan province (Figure 1). An additional 16s rRNA sequence of X. minor deposited in GenBank was incorporated into our dataset. The data of all voucher specimens of the above species are shown in Table 1. All muscle tissue was preserved in 95% ethanol and stored in 280uC.

DNA extraction and sequencing
Genomic DNA was extracted from the muscle tissue using a standard phenol-chloroform extraction protocol [25]. We amplified a fragment of the mitochondrial 16s rRNA gene from Xenophrys species using the primer pairs L3975 and H4551 [26], and mitochondrial 12s rRNA gene using the primer pair Fphe40L and 12S600H [27]. PCR amplifications were performed in a reaction volume of 25 ml containing 100 ng of template DNA, 0.3 mM of each PCR primer and 10 ml Premix EX Taq TM (Takara, Dalian, China). The PCR condition for 16s rRNA was, an initial denaturation step at 94uC for 1. 5 min; 33 cycles of  denaturation at 94uC for 45 s, annealing at 55uC for 45 s, and  extending at 72uC for 90 s, and a final extension step of 72uC for  10 min. The PCR condition for 12s rRNA was, an initial  denaturation step at 96uC for 2 min; 35 cycles of denaturation  at 94uC for 15 s, annealing at 55uC for 1 min, and extending at  72uC for 1 min, and a final extension step of 72uC for 10 min. PCR products were purified with the GenElute TM PCR clean-up kit (Sigma-Aldrich, Dorset, UK). The purified products were sequenced with both forward and reverse primers using a BigDye Terminator v 3.1 Cycle Sequencing Kit (Applied Biosystems, Carlsbad, CA, USA) according to the manufacturer's instructions. The products were sequenced on an ABI Prism 3730 automated DNA sequencer (Applied Biosystems) in the Beijing Genomics Institute. All sequences have been deposited in GenBank (Table 1).

Phylogenetic analyses
The resulting sequences were first aligned using the Clustal W algorithm [28] in BioEdit 7.0 [29], with default parameters and the alignment being checked and manually revised, if necessary. The degree of polymorphism of our sequences was assessed using DnaSP 5.10.1 [30]. The General Time-Reversible model [31] assuming a gamma-shaped distribution across sites [32] for both 16s rRNA and 12s rRNA were selected as the best-fit nucleotide substitution model using Akaike's Information Criterion [33] in jModelTest 1.0 [34]. The aligned sequences of 16s rRNA and 12s rRNA dataset were analyzed separately, using both maximum likelihood (ML) implemented in PhyML [35] on the ATGC online server (http://www.atgc-montpellier.fr/), and Bayesian inference (BI) using MrBayes 3.12 [36]. For ML analysis, the bootstrap consensus tree inferred from 1000 replicates was used to estimate nodal supports of inferred relationships on phylogenetic trees. Branches corresponding to partitions reproduced in less than 50% of the bootstrap replicates were collapsed. For the tree searching and optimization, we applied strategies described in Liang et al. [37]. For BI analysis, two independent runs, each comprising four Markov Chain Monte Carlo simulations were performed for two million iterations and sampled every 1000 th step. The first 25% of the samples were discarded as burn-in. Convergence of the Markov Chain Monte Carlo simulations was assessed by checking the average standard deviation of split frequencies between two runs using Tracer v.1.4 (http://tree.bio.ed.ac.uk/software/tracer/ ). We further applied ML and BI for the joint dataset of 16s and 12s rRNA using the above settings. For both ML and BI, the homologous sequences of Paramegophrys oshanensis and Megophrys nasuta from GenBank were chosen as outgroup (Table 1). Apart from phylogenetic tree-based methods, we also calculated 'net between groups means distance' between Xenophrys species at both genes using the mentioned nucleotide substitution models in MEGA 5.2 [38]. This method takes within-group variations (individual level) into account when calculating the average distances between taxa. Furthermore, to incorporate both 16s rRNA and 12s rRNA datasets, we generated Neighbor-Net (NN) networks [39] of Xenophrys samples using uncorrected p-distance as implemented in SplitsTree 4.10 [40].

Bioacoustics analysis
We recorded the advertisement calls of male Xenophrys species (X. jinggangensis, X. cheni sp. nov., X. lini sp. nov., X. boettgeri, X. huangshanensis, and X. kuatunensis) by SONY ICD-MX20 IC at our sampling localities in southern China during 2011-2012. The calls of these species were frequently heard from June to September, temperatures varied in a small range between 15-18uC during the sound sampling period. Stereotypical male calls, which lasted between 20 s and 2 min, of each species were recorded in SONY MSV format and converted to 16-bit mono PCM format with resampling of 22 kHz. The spectrograms of male calls were generated in the Avisoft-SAS lab Lite software. We define a continuous vocalization with a pause less than 1 second as a call, and define the smallest non-split syllable as a note. The duration and frequency parameters of vocalization, such as call duration, notes per call and per second, note duration, inter-note interval, high and low frequency and frequency band-width were measured from the spectrograms. Differences between these Xenophrys species were tested with one-way ANOVA and Kruskal-Wallis tests in IBM SPSS Statistics 21.

Morphological analysis
Specimens for morphometric analysis were fixed in 10% buffered formalin and then transferred to 70% ethanol for longterm preservation in The Museum of Biology, Sun Yat-sen University (SYS), Guangzhou, Guangdong Province, China. Measurements were made with digital calipers (Neiko 01407A Stainless Steel 6-Inch Digital Caliper, USA) to the nearest 0.1 mm. Abbreviations used are SVL = snout-vent length; HDL = head length, from the tip of the snout to the articulation of the jaw; HDW = head width, between left and right articulations of the quadratojugal and maxilla; SNT = snout length, from the tip the of snout to the anterior corner of the eye; EYE = eye diameter, from the anterior to the posterior corner of the eye; IND = internasal distance; IOD = interorbital distance; TMP = tympanum diameter; TEY = tympanum-eye distance, from the anterior edge of the tympanum to the posterior corner of the eye; HND = hand length, from the distal end of the radioulna to the tip of the distal phalanx of III; RAD = radioulna length; FTL = foot length, from the distal end of the tibia to the tip of the distal phalanx of III; TIB = tibia length; TaL = tail length in tadpoles, from the tip of the tail fin to the vent. Differences between these parameters in the two new species (males only) were further analyzed with Mann-Whitney U test in IBM SPSS Statistics 21.

Nomenclatural Acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available

Molecular phylogenetic analyses
While we obtained a 422 bp of mitochondrial 16s rRNA sequence alignment from 52 Xenophrys samples, and we got a 498 bp of mitochondrial 12s rRNA sequence alignment from a smaller number of samples (N = 25) due to the lack of DNA or amplification difficulties (Table 1). The 16s rRNA alignment yielded 115 variable sites and 98 were parsimony-informative with five insertions. The 12s rRNA yielded 84 variable sites and 74 were parsimony-informative with four insertions. Indels were removed before phylogenetic analyses. For 16s rRNA, the ML and BI phylogenetic approaches resulted in virtually identical topology and all terminal clades had relatively high-supporting values. Both bootstrap supports and posterior probabilities for the clades representing X. cheni sp. nov. and X. lini sp. nov. were high (. 80% for bootstrap proportions and 1.0 for Bayesian posterior probability, respectively; Figure 2). The X. cheni sp. nov., X. lini sp. nov., X. jinggangensis, and X. brachykolos might form a clade. The large-bodied X. mangshanensis was basal to the rest of the eight small-size congeners on the phylogenetic tree and showed large net average distances to others (12%-16%, Table 2). Phylogenetic analyses based on 12s rRNA also revealed well support for the clades representing the two new species ( Figure S1), so as the results inferred by ML and BI based on the joint dataset of two mitochondrial locus ( Figure S2). The net average genetic distance between the two new species was 3.6% (16s rRNA) and 5.7% (12s rRNA), respectively ( Table 2). This differentiation was comparable to the divergence between X. jingangensis to X. brachykolos. Interestingly, the net average genetic distance between X. boettgeri and X. huangshanensis was only 0.005, in contrast to the values between the remaining comparisons (Table 2). Furthermore, in consistent with the phylogenetic gene trees, the inferred multilocus network based on concatenated sequences of 16s and 12s rRNA (910 bp) strongly supported the genetic distinctness of X. cheni sp. nov. and X. lini sp. nov. (Figure 3).

Acoustic analysis of advertisement calls
We recorded the male calls of 13 individuals from three sympatric Xenophrys species in Mt. Jinggang, X. jinggangensis, X. cheni sp. nov., and X. lini sp. nov. (four, three, and six individuals, respectively). We also recorded the vocalizations of X. kuatunensis (N = 2), X. huangshanensis (N = 3), and X. boettgeri (N = 3) permitting additional comparisons (the typical calls of six analyzed Xenophrys species are available in Appendix S1).
The male call of X. cheni sp. nov. has a much slower pace than X. jinggangensis and X. lini sp. nov. (notes per second, see Table 3 and Figure 4). Compared with X. jinggangensis and X. cheni sp. nov., X. lini sp. nov. has a unique pattern of calls, in each single call the inter-note intervals gradually increase as the call comes to an end (Table 3 and Figure 4).
The Kruskal-Wallis test of the three individual-independent vocalization parameters (call duration, notes per call, and notes per second) showed significant differences both between the first three sympatric species in Luoxiao Mountains or across all six Xenophrys species studied (p,0.05 for each parameter of the groups). The test on the five call-independent vocalization parameters (note duration, inter-notes interval, high frequency, low frequency, and frequency band-width) also revealed significant differences within the three sympatric species and across the six Xenophrys species recorded (Kruskal-Wallis test, p,0.05 for each parameter and groups)

Morphological comparisons
Specimens examined for morphometric analysis are listed in Table 4 and other specimens for morphological comparisons are listed in Appendix S2. We compared the two new Xenophrys species with their 46 known congeners. Of these 46 species, 22 (bodylength .45 mm) were larger than the two new species (bodylength ,40 mm). Of the remaining 23 species, the two new species can be distinguished from 11 of them (X. ancrae, X. jinggangensis, X. daweimontis, X. oropedion, X. pachyproctus, X. palpebralespinosa, X. parallela, X. parva, X. serchhipii, X. zhangi, and X. zunhebotoensis) by the absence of vomerine teeth; the two new species differ from nine of them (absent horn-like tubercle in X. binchuanensis, X. pachyproctus, X. wawuensis, X. wuliangshanensis, X. wushanensis and X. zhangi; present large horn-like tubercle in X. jinggangensis, X. palpebralespinosa and X. parallela) because they have a small horn-like tubercle at edge on the upper eyelid; of the remaining eight species, the two new species differ from six of them by having wide lateral fringes  Diagnosis: Xenophrys lini sp. nov. is characterized by the combination of the following characters: (1) a small-sized species with 34.1-39.7 mm SVL in adult males, 37.0-39.9 mm SVL in adult females; (2) head length approximately equal to head width (HDL/HDW ratio 1); (3) snout short, obtusely pointed in dorsal view, almost truncate and sloping backward to the mouth in profile, protruding well beyond the margin of the lower jaw; (4) vomerine teeth absent; (5) margin of the tongue smooth, not notched behind; (6) hind limbs elongated, heels overlapping and tibio-tarsal articulation reaching the anterior corner of the eye; (7) relative finger length II#I,IV,III; (8) lateral fringes on the digits wide, toes with rudimentary webbing at their bases; (9) subarticular tubercle on each digit distinct; (10) dorsal skin smooth with scattered granules, usually a few curved weak ridges on back, several tubercles on flanks; (11) ventral surface smooth; (12) a small horn-like tubercle at the edge of the eyelid; (13) supratympanic fold narrow, light colored, curving from the posterior corner of the eye to a level above the insertion of the arm; (14) light brown or olive above, a dark interorbital triangular marking and X-shaped dorsal marking bordered with a light edge; (15) scattered, tiny, black nuptial spines covering the middle of the dorsal surface of the first finger; (16) single vocal sac in males; (17) gravid females bear pure yellowish eggs.
Holotype description: Adult male, SVL 39.1 mm. Head length approximately equal to head width (HDL/HDW ratio 1.0); snout short (SNT/HDL ratio 0.4, SNT/SVL ratio 0.1), obtusely pointed in dorsal view, almost truncated and sloping backward to the mouth in profile, protruding well beyond the margin of the lower jaw; loreal region vertical, not concave; canthus rostralis well-developed; top of head flat; eye large and convex, eye diameter 35% of head length; pupil vertical; nostril oblique ovoid with low flap of skin laterally; internasal distance larger than interorbital distance; tympanum distinct, TMP/EYE ratio 0.53; tympanum-eye distance great, TEY 2.3 mm, TEY/TMP ratio 0.96; choanae large, ovoid, partly concealed by the maxillary shelves; two vomerine ridges weakly, oblique, posteromedial to choanae, no vomerine teeth; margin of tongue smooth, not notched behind.
Forelimbs moderately slender; radioulna length 23% SVL, hands without webbing, moderately long, 23% of SVL; fingers slender, relative finger length II#I,IV,III; tips of digits round, slightly dilated; subarticular tubercle distinct at the base of each finger; slight lateral fringes from bases of each fingers to terminal phalanges; two metacarpal tubercles, substantially enlarged. Hind limbs relatively elongated and moderately robust; heels overlapping when the flexed legs are held at right angles to the body axis; tibio-tarsal articulation reaching the anterior corner of the eye, when leg stretched along the side of the body; tibia length 49% of SVL; foot length 66% of SVL; relative toe lengths I,II,V,III, IV; tips of toes round, slightly dilated; subarticular tubercle distinct at the base of each toe; toes with rudimentary webbing at their bases, lateral fringes wide; tarsal fold absent; but as outer lateral fringes on toe V from hough to terminal phalanges; inner metatarsal tubercle ovoid; no outer metatarsal tubercle.
Skin of all upper surfaces and flanks smooth with scattered granules; back with a few curved weak, discontinuous ridges; several tubercles on the flanks and dorsal surface of thighs and tibias; a curved ridge on the upper eyelid, anteriorly starting near canthus rostralis (not in contact), extending backward to the middle of the upper eyelid and bending to the inside, where there is a slightly large, horn-like tubercle; supratympanic fold distinct, narrow, curving posteroventrally from posterior corner of the eye to a level above insertion of the arm; ventral surface smooth; pectoral gland large, round, prominently elevated relative to the ventral surface, closer to the axilla than to the mid-ventral line; single larger femoral gland on rear of thigh; distinct granules on Table 2. Net average genetic distances between Xenophrys species in southeast China.  Live holotype coloration: Olive-brown above with distinct dark brown markings bordered with light edge and obscure markings; a distinct dark triangular marking between the eyes, apex of triangle over occiput; a distinct X-shaped dark marking on the back; a small longitudinal dark stripe on the dorsum of the snout; dorsal surface of the limbs and digits with dark brown transverse bands; side of head with dark brown vertical bars, one from the tip of the snout to behind the nares; one under the eye, one along the supratympanic fold, coving tympanum; supratympanic fold light colored; lower lip black with six vertical white spots; lateral surface of trunk and anterior surface of thighs pinkish near the groin; ventral surface reddish brown, with an obscure longitudinal black streak down the center of the throat, with several white blotches on the belly; ventral surface of the limbs reddish brown with pale gray wormlike marks; palms and soles uniform reddish brown, tip of digits pale white; inner metatarsal tubercle and two metacarpal tubercles pinkish; pectoral and femoral glands white; pupils black; iris dark grey.
Preserved holotype coloration: Blackish green above with a black triangle and an X-shaped marking bordered with a distinct light edge line; dorsal surface of limbs and digits with black transverse bands; ventral surface darker brown with white blotches; creamy white replaces the pinkish color in the anterior surface of the thighs and lateral surface of the trunk.
Tadpole description: Body slender, oval, flattened above; tail depth slightly greater than body depth, dorsal fin arising behind the origin of the tail, maximum depth near mid-length, tapering gradually to narrow, pointed tip; tail 2.3-2.5 times as long as body length, tail depth 18-20% of tail length in the 28 th -34 th stages; maximum body width 37% of body length in the 34 th stage, 35% in the 33 rd stage, 30-33% in the 28 th -31 st stages; body depth 30% of body length in the 33 rd and 34 th stages, 29% in the 32 nd stage, 24-25% in the 28 th -31 st stages; eyes large, lateral; nostril dorsolateral, slightly closer to the eyes than to the umbelliform oral disk, rim raised; internasal wider than interorbital; spiracle on left side of the body, closer to the eye than to the end of the body; anal tube extends backward above the ventral fin, opening medial; oral disk terminal, lips expanded and directed upwardly into a typical Xenophrys umbelliform oral disk; transverse width of expanded funnel 38-42% of body length in 28 th -34 th stages.
Coloration in preservative: All upper and lateral surfaces brown grey with black marks; ventral surface of head red-brown, belly black with pale grey marks, tail and hind limbs creamy white. Variation: Measurements and body proportions of type series are given in Table 4. Color patterns in paratypes are more similar to the holotype, but SYS a001419, 001423, 001424, 002369, 002373, 002375, 002379, 002380, 002382, 002383 and 002385 had a dark triangular marking between the eyes with a light center between the eyes; five female paratypes were light brown above; lower lip black with eight white bands; ventral surface with pinkish, brown, white, and black markings; one longitudinal black streak down the center of the throat, surface of the posterior abdomen near the groin white; several large, black spots on ventral surface of the hind limb, forearm, and wrist.
Secondary sexual characteristics: Single vocal sac; scattered, tiny, black nuptial spines cover a circular area at the middle of the dorsal surface of the first finger in two male paratypes and holotype; five gravid female paratypes bear pure yellowish eggs in the oviducts.
Distribution and biological ecology. Currently, X. lini sp. nov. is known only from the Bamianshan, Jingzhushan, Nanfengmian Nature Reserve and Dabali, within the range of Mt. Jinggang, Jiangxi Province, and from adjacent Taoyuandong Nature Reserve, Hunan Province which are located in the middle of Luoxiao Mountains, running along the border between the Jiangxi and Hunan Provinces, China. All individuals were found in rushing mountain streams surrounded by moist subtropical evergreen broadleaved forests between elevations of 1100-1610 m (Figure 1-VIII: b, c, e, g and i).
All adult specimens were collected on the 13 th and 19 th September, 2011 and 5-6 th October 2013; males were heard calling day and night during the survey. The male paratype SYS a001421 has mature spermaries in the abdominal cavity, measuring 4.962.1 mm in the major and minor axes, respectively. All female paratypes bear pure yellowish mature eggs and atrophic ovary fat. According tadpole stages defined by Gosner [41], the individuals in 28-34 th stages were found under rocks in the stream on the 5 th December, 2011. Juveniles were collected on 21 th , May, 2013. Thus, we assume the breeding season of this species likely begins September-October.
Xenophrys cheni Wang and Liu sp. nov. Diagnosis: Xenophrys cheni sp. nov. is characterized by the combination of the following characteristics: (1) a small-size species with 31.8-34.1 mm SVL in adult females, 26.2-29.5 mm SVL in adult males; (2) head length approximately equal to head width (HDL/HDW ratio 1.00-1.06); (3) snout short, obtusely rounded in dorsal view, almost truncate, and sloping backward to the mouth in profile, protruding well beyond the margin of the lower jaw; (4) vomerine teeth absent; (5) margin of tongue notched behind; (6) tympanum distinct or indistinct, usually its upper part hidden under the supratympanic fold; (7) hind limbs elongated, the heels more overlapping and tibio-tarsal articulation reaching the region between the nostril and tip of snout; (8) relative finger length I, Table 3. Vocalization parameters of six Xenophrys species in southeast China.  Holotype description: Adult male, SVL 27.2 mm. Head length approximately equal to head width (HDL/HDW ratio 1.0); snout short (SNT/HDL ratio 0.44, SNT/SVL ratio 0.14), obtusely rounded in dorsal view, almost truncate and sloping backward to the mouth in profile, protruding well beyond the margin of the lower jaw; loreal region vertical, not concave; canthus rostralis well-developed; top of head flattened; eye large and convex, eye diameter 38% of head length; pupil vertical; nostril oblique ovoid with low flap of skin laterally; internasal distance larger than the interorbital distance; tympanum (TMP) distinctly rounded, TMP/ EYE ratio 0.42; tympanum-eye distance (TEY) 1.5 mm, TEY/ TMP ratio 0.94; choanae large, ovoid, partly concealed by the maxillary shelves; two vomerine ridges weakly, oblique, posteromedial to choanae, no vomerine teeth; margin of tongue notched behind.
Forelimbs moderately slender; radioulna length 25% of SVL, hands without webbing, moderately elongated, 25% of SVL; fingers slender, relative finger length I,II,IV,III; tips of fingers round, slightly dilated, narrower than width of terminal phalanges; subarticular tubercles indistinct; markedly enlarged lateral fringes from the bases of fingers to the terminal phalanges; two metacarpal tubercles. Hind limbs relatively elongated and moderately robust; heels overlapped, when the flexed legs are held at right angles to the body axis; tibio-tarsal articulation reaching the region between the nostril and tip of snout, when leg stretched along the side of the body; tibia length 52% of SVL; foot length 71% of SVL; relative toe lengths I,II,V,III,IV; tips of toes round, slightly dilated, narrower than the width of the terminal phalanges; toes with fleshy webs at their bases; subarticular tubercle indistinct; lateral fringes wide; tarsal fold absent; inner metatarsal tubercle ovoid; outer metatarsal tubercle absent.
Skin of all upper surfaces and flanks smooth with tubercles, usually forming a dorsolateral tubercle row of parallel to contralateral row, and an X-shaped weak ridge on the dorsum of body; five large transverse tubercle rows on the dorsal surface of the shanks; ventral surface smooth; a weak horn-like tubercle at the edge of the upper eyelid; supratympanic fold swollen, curving poster-oventrally from the posterior corner of the eye to a level above the insertion of the arm; pectoral gland small, closer to the Live holotype coloration: Olive-brown above with dark reticular markings, including a triangular marking bordered with a light edge between the eyes, apex of triangle over occiput, an Xshaped marking bordered with a light edge on the dorsum of body, five dark transverse bands on the dorsal surface of the thigh, four dark transverse bands on the dorsal surface of the shank; side of head with dark brown vertical bars, one from the tip of the snout to behind the nares; one under the eye, one along the  supratympanic fold, coving tympanum; supratympanic fold light colored bordered by a black lower edge; lower lip black with six white bars; lateral surface of trunk of body and anterior surface of the thighs near the groin pinkish; ventral surface of body olive with pinkish and white spots; an obscure longitudinal darker streak down the center of the throat; several white blotches on the belly; the ventrolateral regions covered with olive-green bordered black zigzag lines; ventral surface of limbs darker brown with white spots; palms and soles uniform darker brown, tip of digits pale grey; inner metatarsal tubercle and two metacarpal tubercles pinkish; pectoral and femoral glands white; pupils black; iris dark brown. Preserved holotype coloration: Dorsal surface sallow with darker brown reticular markings; triangular and X-shaped markings on dorsum and transverse bands on limbs and digits distinct; ventral surface yellowish with grey spots; black zigzag lines distinct; creamy-white substitutes the pinkish in the anterior surface of the thighs and lateral surface of the trunk.
Variation: Measurements and body proportions of type series are given in Table 4.
Color patterns in 14 male paratypes are more similar to the holotype. In the male paratypes SYS a002124, 002126, 002140, 002141 and 002145, tympanum or its edge is indistinct; in SYS a002126 and 002127, upper part of tympanum hidden under the supratympanic fold; in the female paratypes, red-brown above with brown reticular markings; lower lip black with white spots; gular region and chest black with red and white blotches; posteriorly black fades and becomes marbled with light and dark All specimens were collected from April to September, during which all male individuals were calling and bearing dilated spermaries in specimens collected July and September, lacking nuptial spines on the dorsal surface of the first finger; mature ovaries bearing pure yellow eggs and dilated oviducts in the female paratype SYS a001429. Thus, the breeding season of this species is likely from April to September, but no tadpoles were found.

Discussion
Most cryptic congeners in the genus Xenophrys are difficult to distinguish from each other due to the superficial similarities in morphologies: drab colorations, complicated markings and even changeable colorations and skin markings of the same individual under different environmental conditions [9][10][11]. These result in considerable challenges in field identification, which in turn cause ambiguities in taxonomy and distributions [11]. For example, in Luoxiao Mountains covered by our surveys, small sized Xenophrys toads occurred were misidentified as either X. boettgeri, X. kuatunensis or X. minor and were documented in the local amphibian checklists [42,43]. This issue seems ubiquitous throughout the geographical range of Xenophrys toads [9]. To solve these problems, extensive sampling with careful and robust diagnosis is necessary in order to unveil the cryptic species diversity of Xenophrys toads.
In this study, we characterized the cryptic diversity of Xenophrys toads with intensive surveys in the middle range of Luoxiao Mountains, with the realm about 250 km 2 in southeast China. We discovered two undescribed Xenophrys species, namely Xenophrys lini sp.nov. and Xenophrys cheni sp.nov. using morphology, molecular genetics and bioacoustics. The two new species, together with X. jinggangensis, are sympatric in Luoxiao Mountains but altitudinally and perhaps ecologically isolated (see discussion below). Morphologically these two new species can be reliably distinguished from other known congeners. Although the present genetic analyses are based on only two mitochondrial genes, the genetic differences between the two new species are of a comparable magnitude as other known Xenophrys species in southeast China (Table 2). Interestingly, the close phylogenetic relationships between the three sympatric species in Luoxiao Mountains and other known species in southeast China may indicate sign of evolutionary radiation in the region. However, our phylogenies in this study were just partial without including Xenophrys species in western China and Himalayas. This encourages further comprehensive phylogenetic analyses with extensive taxa coverage in order to resolve the systematics in Xenophrys. Furthermore, although our bioacoustics analysis shows that the advertisement calls of male Xenophrys are very similar, consisting of several rather short and repeated notes, the call styles, especially in the aspects of the note frequency ranges, note durations and inter-note intervals show significantly differences between all three sympatric species in Luoxiao Mountains and other three compared congeners.
Based on the present knowledge, the geographic ranges of the two new species as well as X. jinggangensis are endemic to several sites in Luoxiao Mountains between Jiangxi and Hunan provinces, China. Luoxiao Mountains are situated in the middle of southeast Chinese subtropical mountain ranges with complex topography and biogeography [44]. It is connected to Nanling Mountains in the south, which is a stronghold for X. mangshanensis [7] indicating potential parapatric. It is further paralleled with Huangshan-Tianmu and Yandang-Wuyi-Daiyun Mountains in the east, which harbors X. kuatunensis, X. huangshanensis and X. boettgeri [7], indicating potential allopatric. In a finer scale, we found that three sympatric Xenophrys species in Mt. Jinggang were distributed in different microhabitats, where the altitudes and the characters of the water bodies varied subtly. X. jinggangensis was found in slowmoving streams between 700-850 m a.s.l. [10] and X. lini sp. nov. in rushing streams between 1100-1610 m a.s.l. In contrast, X. cheni sp. nov. was restricted to swamps at forest edges around 1200-1530 m a.s.l. (Figure 1). Whether this pattern of spatial and ecological segregation in Xenophrys toads in Luoxiao Mountains is associated with local adaptation to divergent environments requires further investigation on the diversification mechanisms using ecological [45] and genetic (genomic) approaches [46].
This study further provides a few fresh insights into the taxonomy of Xenophrys toads. Most importantly, the identification of two new species may indicate previously underestimated diversity and endemism Xenophrys in southeast China [47]. However, the potential spatial or ecological limits of these species are still poorly known. Moreover, the closely related X. huangshanensis and X. boettgeri show moderate differences in morphology and vocalization and a small genetic distance (0.005) in contrast to the other species studied. The species validation between these two, X. huangshanensis and X. boettgeri, needs to be revisited. Finally, though we used only the mitochondrial 16s and 12s rRNA as genetic markers for its universal application in molecular analysis in amphibian systematic studies [48], the integrative manner of this study using multiple approaches merits robust species delineation. Regardless, our results provide a first step in the right direction to help resolve previously unrecognized amphibian biodiversity. More spatially extensive sampling ideally combined with habitat characteristics and bioacoustic recordings will be necessary to further understanding the cryptic diversity and diversification of Xenophrys toads in the mountain complexes of southeast China.
In conclusion, we show that the two sympatric Xenophrys species, X. lini sp. nov. and X. cheni sp. nov. have congruent differences in morphology, bioacoustics, genetic and habitats. We are reasonable to treat them as separate species based on the 'biological species concept' [49,50]. Nevertheless, more studies are needed on the distributions, ecology and life history of these locally endemic species, as well as on their conservation status. These efforts are very important given the ongoing declines of amphibians both regionally and globally [47,51,52].

Supporting Information
Appendix S1 Recordings of typical male advertisement calls of six Xenophrys species.