Abstract

Four new species of Hynobius are described from the Chugoku Mountains, Japan. The complex of Hynobius utsunomiyaorum, including Hynobius hidamontanus, was divided into six groups based on morphological and molecular analyses. The Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, and northeastern Okayama groups of H. utsunomiyaorum were described as Hynobius pseudoutsunomiyaorum sp. nov., Hynobius mengamemontanus sp. nov., Hynobius sakuhokumontanus sp. nov., and Hynobius ushiromontanus sp. nov., respectively. Molecular analyses of samples from across the entire distribution range of H. utsunomiyaorum revealed that H. utsunomiyaorum (including H. hidamontanus) has six genetic groups. Morphological analyses suggested that males of the four new species are morphologically different from H. utsunomiyaorum according to specific characters. The distribution of the new species, as well as those of H. utsunomiyaorum and H. hidamontanus, were determined and compared to indicate the isolation among the species. Although further sampling from the northern part of Okayama Prefecture is needed to determine the distribution ranges of H. utsunomiyaorum and H. sakuhokumontanus sp. nov., our results suggest that the distribution area of true H. utsunomiyaorum is mainly limited to the central area of the Chugoku Mountains.

1. Introduction

The genus Hynobius, is the most diversified within the family Hynobiidae with 66 described species and is found in eastern Asia ¹. Most species in the genus are found in Japan, with 48 species distributed in throughout the country (48/66 = 72.7%) ¹. Traditionally, Hynobius species have been identified based on morphological characters ^{2, 3}. However, Hynobius exhibits high number of cryptic species, making identification without molecular markers and collection site information difficult ⁴. Molecular markers have revealed the existence of cryptic species in Hynobius ^{5, 6}, and they are essential for classifying this genus and describing new species.

The Hiba salamander, Hynobius utsunomiyaorum, was described from Rokunohara, Mt. Hiba, Saijocho Yuki, Shobara City, Hiroshima Prefecture, and is found in Hyogo, Okayama, Hiroshima, Tottori, and Shimane Prefectures, Chugoku District, Japan ⁷. The Hakuba salamander, Hynobius hidamontanus, was described from Ochikura Swamp, Hokujo, Hakuba Village, Nagano Prefecture, and is found in Niigata, Toyama, and Gifu Prefectures ^{8, 9}. Hynobius utsunomiyaorum has four distinct groups with H. hidamontanus distributed around the Hokuriku District ⁷. The monophyly of the clade, including H. utsunomiyaorum and H. hidamontanus, was strongly supported by mitochondrial DNA analysis results [maximum likelihood (ML) estimation: bootstrap value = 99%; Bayesian inference (BI): posterior probability = 1.00] ⁷. However, the results of molecular analyses using more samples suggested that the monophyly is doubtful according to BI ¹⁰. Thus, based on the results of previous studies ^{7, 10} and the phylogenetic species concept, H. hidamontanus may be an invalid species because it is clearly included in the clade of H. utsunomiyaorum. Moreover, if H. hidamontanus is a valid species according to the phylogenetic species concept, H. utsunomiyaorum should be divided into at least three species based on the data of previous study ⁷: the Shimane (haplotype H9-H11), northern Okayama (haplotype H6-H8), and Hiroshima (haplotype H1-H5) groups (haplotype numbers follow those of the study ⁷). In addition, the northern Okayama group might be divisible into two groups, the northern central Okayama (haplotype H6) and northeastern Okayama (haplotype H7-H8) groups, although these are monophyletic groups ⁷. Matsui et al. ⁷ performed molecular analysis using only 18 individuals from eleven localities, and Sugawara et al. ¹⁰ did not collect samples from the entire distribution range of H. utsunomiyaorum. Thus, the phylogenetic relationship of H. utsunomiyaorum and H. hidamontanus should be reassessed using samples collected across the entire distribution range of H. utsunomiyaorum. From a morphological perspective, the commonality of the four groups (Shimane, northern central Okayama, northeastern Okayama, and Hiroshima groups) of H. utsunomiyaorum is unreliable because Matsui et al. ⁷ did not compare these groups. Although H. hidamontanus is included in the H. utsunomiyaorum clade, Matsui et al. ⁷ and Sugawara et al. ¹⁰ did not use H. hidamontanus in their morphological analyses; thus, the morphologies of the four groups of H. utsunomiyaorum and H. hidamontanus must also be compared. Additionally, H. utsunomiyaorum has a wide distribution range, so there is a need to determine whether another genetically distinct group exists. If such a group is found, it should be included in subsequent morphological analyses.

Frost and Hillis ¹¹ recommend using phylogenetic and evolutionary species concepts in herpetology. Additionally, Dubois ¹² mentioned that essential features must be directly derived from characteristics observed in specimens rather than indirectly inferred when a new species is described. The original description of Hynobius kunibiki before being included in H. utsunomiyaorum used three species concepts (i.e. phylogenetic, evolutionary, and morphological species concepts) to support their conceptions ¹⁰. Thus, we followed these three species concepts consistent with previous studies ^{4, 10}. First, we determined the number of candidate species included in H. utsunomiyaorum based on the phylogenetic and evolutionary species concepts. Second, we evaluated the species validity of each candidate species and H. hidamontanus based on the morphological species concept. Finally, we described several new species and revealed the distribution ranges of these species in detail based on the results of our study.

2. Materials and Methods

We sampled 49 individuals from 49 localities from April 2007 to June 2022 (Table 1; Figure 1) for phylogenetic analyses. Figure 1 indicates the sampling map of Hynobius species, according to the present study. The cross in the upper left area shows the type locality of H. hidamontanus (Figure 1). The enlarged area includes the central Chugoku District where the H. utsunomiyaorum complex is found (Figure 1). The closed symbols correspond to each of four groups sequenced in this study: Mt. Sanbe and Kitahiroshima group (closed circles), Mt. Mengame group (closed pentagons), northern central Okayama group (closed triangle), and widely distributed group (closed squares) (Figure 1). The open symbols correspond to each of four groups sequenced by Matsui et al. ⁷ or Sugawara et al. ¹⁰: Mt. Sanbe and Kitahiroshima group (open circles), northern central Okayama group (open triangle), northeastern Okayama group (open stars), and widely distributed group (open squares) (Figure 1). The open symbols including diagonal lines correspond to each of four groups described by Matsui et al. ⁷ or Hayashi and Ooi ¹³ for which the detailed locality is unclear: Mt. Sanbe and Kitahiroshima group (shaded circles), northern central Okayama group (shaded triangle), northeastern Okayama group (shaded star), and widely distributed group (shaded squares) (Figure 1). Dotted lines including BA (1-4) represent the blank area of distribution of the H. utsunomiyaorum complex (Figure 1). Table 1 shows the sample list of Hynobius species using this study. Population number corresponds to the localities in Figure 1 (Table 1). Asterisks after sampling localities show the type locality of each species (Table 1). In this study, the widely distributed group means the true H. utsunomiyaorum (sensu stricto), and four groups have another species name as follows: Mt. Sanbe and Kitahiroshima group = Hynobius pseudoutsunomiyaorum sp. nov.; Mt. Mengame group = Hynobius mengamemontanus sp. nov.; northern central Okayama group = Hynobius sakuhokumontanus sp. nov.; northeastern Okayama group = Hynobius ushiromontanus sp. nov.

Total genomic DNA was extracted using a DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). The mitochondrial cytochrome b gene (630-base-pairs) was amplified using ExTaq (TaKaRa, Tokyo, Japan) with primers L14010 (5′-TAHGGWGAHGGATTWGAWGCMACWGC-3′) and H14778 (5′-AARTAYGGGTGRAADGRRAYTTTRTCT-3′) ¹⁴. The PCR reaction mix (total volume: 10 µl) contained 1.0 µl of 10×Ex Taq Buffer, 0.8 µl of 25 mM dNTP mix, 0.5 µl of each of the forward and reverse primers (10 pM), 0.05 µl of Taq polymerase, 6.15 µl of distilled deionized water, and 1.0 µl of template DNA. The PCR protocols implemented using a T100™ thermal cycler (Bio-Rad, Hercules, CA, USA) were as follows: an initial 3-min denaturing step at 94°C; 40 cycles of 30 s at 94°C, 45 s at 56°C, and 90 s at 72°C; with a final 10-min extension at 72°C. The PCR products were purified with illustra™ Exo ProStar™ 1-Step (GE Healthcare, Buckinghamshire, UK) and sequenced using BigDye® Terminator ver. 3.1 (Applied Biosystems, Foster City, CA, USA) on an ABI 3130xl Genetic Analyzer (Applied Biosystems). Acquired DNA sequences were deposited to the DNA Data Bank of Japan (Table 1), and these sequences were aligned using MEGA X ¹⁵. Phylogenetic analyses were performed using Bayesian inference (BI) and maximum likelihood (ML) estimation with closely related species, including the highland and Aki groups ⁷, and Hynobius abuensis and Hynobius tosashimizuensis as the outgroup, given that they are the closest species of the highland and Aki groups ⁷. We estimated the best-fit nucleotide substitution model based on the Bayesian information criterion (BIC) ¹⁶ and corrected Akaike’s information criterion (AICc) ¹⁷ using jModelTest 2 ¹⁸. We selected the Hasegwa-Kishino-Yano (HKY) model with gamma distribution in BI and ML. We constructed Bayesian and maximum likelihood trees using MrBayes 3.2 ¹⁹ and MEGA X ¹⁵, respectively. For Bayesian analyses, we performed two independent MCMC runs for 2000000 generations; with a sample frequency of 100. In Bayesian analysis, we examined the stationarity of the likelihood scores of sampled trees using Tracer version 1.7 (https://tree.bio.ed.ac.uk/software/tracer/), and the first 25% of generations were discarded as burn-in. The assessment of monophyly was performed by values of posterior probability (PP) and bootstrap (BS) based on the criteria of Huelsenbeck and Rannala ²⁰ and Hillis and Bull ²¹: monophyletic group = PP ≥ 0.95 and BP ≥ 70.

Underlined localities in Figure 1 show the sampling points of individuals used in morphological comparisons. In this study, we used only male individuals for morphological analyses, consistent with Matsui et al. ⁷, because we could not collect sufficient female specimens to support statistical analyses, and the morphological differences between the sexes of H. utsunomiyaorum are limited ¹⁰. We sampled 67 males of H. utsunomiyaorum from April 2019 to May 2022: Mt. Sanbe and Kitahiroshima group, 14 individuals from four populations (Pops. 3, 4, 6, and 7); Mt. Mengame group, 10 individuals from two populations (Pops. 13 and 14); northern central Okayama group, 14 individuals from two populations (Pops. 15 and 16); northeastern Okayama group, 14 individuals from one population (Pop. 18); and widely distributed group, 15 individuals from four populations (Pops. 35, 56, 60, and 62) (Table 1; Figure 1). We also sampled seven males of H. hidamontanus in April 2021 at the type locality of this species (Table 1; Figure 1) after obtaining permission from the governments of Nagano Prefecture and Hakuba Village. The collected individuals were measured under anesthesia using ethyl 3-aminobenzoate methanesulfonate salt (Sigma-Aldrich®, St. Louis, MO, USA) diluted 1000 times with water ²². From a conservation perspective, measured individuals were returned to their site of capture except for the candidate individuals of type specimens or candidate individuals required for redescribing H. utsunomiyaorum. Before the individuals were returned, photos were taken of the dorsal, ventral, and lateral sides on a black back-ground, and tissue samples (preserved in 99.9% ethanol) were collected from the tail tips of all individuals as evidence of collection. For all examined individuals, a vernier caliper was used to take 21 measurements following Nishikawa et al. ²³: snout-vent length (SVL), trunk length (TRL), axilla-groin distance (AGD), head length (HL), tail length (TAL), median tail width (MTAW), median tail height (MTAH), vomerine teeth length (VTL), and vomerine teeth wide (VTW), head width (HW), forelimb length (FLL), hindlimb length (HLL), second finger length (2FL), third finger length (3FL), third toe length (3TL), internarial distance (IND), interorbital distance (IOD), upper eyelid length (UEL), snout length (SL), upper eyelid width (UEW), lower jaw length (LJL). We also checked the presence of mid–dorsal line between forelimb and hindlimb (MDL), the presence of distinct white spots on ventral side of the body (DWSV), the presence of distinct white spots on lateral sides of the body (DWSL), the presence of distinct yellowish–brown lines on the dorsal (DYBLDT) and ventral (DYBLVT) sides of the tail, and presence of distinct gular mottling (DGM) for each individual. The number of costal folds between the adpressed limbs (CFBALN) and the number of costal grooves (CGN) were counted. The counting method of CGN was followed by Matsui et al. ⁷. For assessments of the morphological differences among species, we performed the statistical approaches using R ²⁴ with significance level equal 0.05. Before the statistical tests, we tested for normality using Shapiro-Wilk test. When data did not follow a normal distribution and variances among populations were not equal, we performed Steel-Dwass tests. On the other hand, when data followed a normal distribution, we tested for homoscedasticity using Bartlett’s test. As the next step, when the variances among the populations were equal, we performed Tukey-Kramer tests, but when variances were not equal, we performed Games-Howel tests. In addition, we performed canonical discriminant analysis using standardized values for the 20 measurements and SVL to examine the overall morphological variation among six groups (including H. hidamontanus).

When the measurement of holotype and topotype specimens, we took 43 measurements referred to Nishikawa et al. ²³: SVL, TRL, AGD, HL, TAL, MTAW, MTAH, basal tail width (BTAW), basal tail height (BTAH), left vomerine teeth length (LVTL), right vomerine teeth length (RVTL), VTW, HW, left forelimb length (LFLL), left hindlimb length (LHLL), right forelimb length (RFLL), right hindlimb length (RHLL), left first finger length (L1FL), left second finger length (L2FL), left third finger length (L3FL), left fourth finger length (L4FL), right first finger length (R1FL), right second finger length (R2FL), right third finger length (R3FL), right fourth finger length (R4FL), left first toe length (L1TL), left second toe length (L2TL), left third toe length (L3TL), left fourth toe length (L4TL), right first toe length (R1TL), right second toe length (R2TL), right third toe length (R3TL), right fourth toe length (R4TL), IND, IOD, left upper eyelid length (LUEL), right upper eyelid length (RUEL), left snout length (LSL), right snout length (RSL), left upper eyelid width (LUEW), right upper eyelid width (RUEW), left lower jaw length (LLJL), and right lower jaw length (RLJL). All measurements and observations of holotype and topotype were carried out after fixation in 10% formalin and transferred to 70% ethanol. To avoid the over-collection of new species, details of type specimens are available only by contacting the museum that stored each specimen.

3. Results

Molecular phylogenetic trees based on BI and ML were nearly identical. Scale bars in Figure 2A and 2B show the genetic distance (expected changes per site). Numbers located near the nodes are posterior probabilities (PP) for Bayesian inference and bootstrap values (BS) for maximum likelihood estimation (Figures 2A and 2B). The nodes covered by closed circles and open circles indicate the support by both values or either value, respectively (Figures 2A and 2B). Nodes not covered by circles indicate that both values do not satisfy the conditions of statistical support (PP ≥ 0.95 and BS ≥ 70%) (Figures 2A and 2B). Values in parentheses after haplotype names correspond to population numbers, as indicated in Table 1 and Figure 1 (Figures 2A and 2B). Labels boxed by red lines show the type locality of six species (Figures 2A and 2B). Labels covered by shaded boxes show the Daisen type of H. utsunomiyaorum (Figure 2B). The monophyly of the highland and Aki groups suggested by Matsui et al. ⁷ was strongly supported by BI and ML, as in the previous study (Figure 2A). The monophyly of the H. utsunomiyaorum clade with H. hidamontanus was supported by PP and BS, consistent with Matsui et al. ⁷, and five monophyletic groups of H. utsunomiyaorum were confirmed in the clade (Figure 2): Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, northeastern Okayama, and widely distributed (H. utsunomiyaorum sensu stricto) groups. The monophyly of the northern central Okayama and northeastern Okayama groups was supported by BI and ML (Figure 2A). In the phylogeny of H. utsunomiyaorum sensu stricto, the monophyly of the Daisen type of H. utsunomiyaorum lacked support from BI and ML (Figure 2B).

Morphological measurements of the five groups and H. hidamontanus are presented in Table 2. Specifically, PSE, MEN, SAK, USH, and UTS correspond to the Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, northeastern Okayama, and widely distributed groups, respectively, with HID representing H. hidamontanus. Ranges are shown in parentheses (Table 2). The significant values of the measurements among these groups and H. hidamontanus are shown in Table 3. In the same way as Table 2, PSE, MEN, SAK, USH, UTS, and HID correspond to the Mt. Sanbe and Kitahiroshima group, Mt. Mengame group, northern central Okayama group, northeastern Okayama group, widely distributed group, and H. hidamontanus, respectively. Significant differences at P < 0.001 are shown in bold (Table 3). Definitions of morphological characters are described in the Materials and Methods section. Males showed no significant differences in 21 morphological characters in the following group comparisons: Mt. Sanbe and Kitahiroshima vs. Mt. Mengame, Mt. Sanbe and Kitahiroshima vs. northern central Okayama, Mt. Sanbe and Kitahiroshima vs. widely distributed, and Mt. Mengame vs. northern central Okayama (Table 3). A comparison of males between the Mt. Sanbe and Kitahiroshima and northeastern Okayama groups revealed significant differences in four morphological characters (Table 3). Further, a comparison of males between the Mt. Sanbe and Kitahiroshima group and H. hidamontanus showed significant differences in five morphological characters (Table 3). In a comparison between the Mt. Mengame and northeastern Okayama groups, males differed significantly in four morphological characters (Table 3), whereas in a comparison between the Mt. Mengame and widely distributed groups, males differed significantly in only one morphological character (Table 3). Notably, a comparison of males between the Mt. Mengame group and H. hidamontanus revealed significant differences in three morphological characters (Table 3). Compared with the northeastern Okayama group, widely distributed group, and H. hidamontanus, males from the northern central Okayama differed significantly in ten, two, and three morphological characters, respectively (Table 3). Similarly, compared with the widely distributed groups and H. hidamontanus, males from the north eastern Okayama differed significantly in three and ten morphological characters, respectively (Table 3). Finally, comparison of males between the widely distributed group and H. hidamontanus showed significant differences in six morphological characters (Table 3)

Canonical discriminant analysis results in males indicated distinctiveness among the six groups including H. hidamontanus, as evidenced by nonoverlapping distribution area scores. The only exception was observed between the Mt. Sanbe and Kitahiroshima and widely distributed groups (Figure 3: green, violet, yellow, blue, red, and pink circles indicate the scores of the Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, northeastern Okayama, and widely distributed groups, and H. hidamontanus, respectively).

Morphological observations among the five groups of H. utsunomiyaorum and H. hidamontanus are shown in Table 4. PSE, MEN, SAK, USH, and UTS correspond to the Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, northeastern Okayama, and widely distributed groups, respectively, with HID representing H. hidamontanus. The Mt. Sanbe and Kitahiroshima group almost always had 12 CGN (13/14 = 92.9%), almost always had no DYBLVT (13/14 = 92.9%) and DGM (13/14 = 92.9%), usually had ≥ 0.0 CFBALN (12/14 = 85.8%), usually had no MDL (12/14 = 85.7%), and frequently had DWSV (11/14 = 78.6%) and DYBLDT (10/14 = 71.4%) (Table 4). The Mt. Mengame group always had no more than −0.5 CFBALN (10/10 = 100%), always had no DYBLVT (10/10 = 100%), almost always had DWSV (9/10 = 90%) and DWSL (9/10 = 90%), usually had 12 CGN (8/10 = 80%), usually had no DYBLDT (8/10 = 80%), and frequently had no MDL (7/10 = 70%) (Table 4). The northern central Okayama group always had DWSV (14/14 = 100%) and DWSL (14/14 = 100%), always had no DYBLVT (14/14 = 100%) and DGM (14/14 = 100%), almost always had 12 CGN (13/14 = 92.9%), usually had MDL (12/14 = 85.7%), and frequently had ≥ 0.0 CFBALN (11/14 = 78.6%) (Table 4). The northeastern Okayama group always had DWSV (14/14 = 100%) and DWSL (14/14 = 100%), always had no DYBLVT (14/14 = 100%) and DGM (14/14 = 100%), frequently had 12 CGN (11/14 = 78.6%) and ≤ 0.0 CFBALN (11/14 = 78.6%), and frequently had no DYBLDT (11/14 = 78.6%) (Table 4). The widely distributed group always had no DYBLVT (15/15 = 100%), almost always had DYBLDT (14/15 = 93.3%), usually had DWSV (12/15 = 80%), DWSL (13/15 = 86.7%), 12 CGN (13/15 = 86.7%), and ≤ 0.0 CFBALN (12/15 = 80.0%), and usually had no DGM (12/15 = 80%) (Table 4). H. hidamontanus always had DWSV (7/7 = 100%), DWSL (7/7 = 100%), and ≥ 0.5 CFBALN (7/7 = 100%), always had no MDL (7/7 = 100%), DYBLVT (7/7 = 100%), and DGM (7/7 = 100%), usually had 12 CGN (6/7 = 85.7%), and usually had no DYBLDT (6/7 = 85.7%) (Table 4).

Based on the three species concepts, the Mt. Sanbe and Kitahiroshima, Mt. Mengame, northern central Okayama, and northeastern Okayama groups of H. utsunomiyaorum were described as new species.

4. Systematics

ZooBank LSID: urn:lsid:zoobank.org:act:C5145D84-2E02-4C4A-899F-396BEE860049 (Figures 4-5)

Synonymy: Hynobius utsunomiyaorum Matsui et al. (2019: clade B2b4, H9-H11).

Holotype. An adult male (specimen number: HMNH-AM-324) from Shinjo, Kitahiroshima Town, Hiroshima Prefecture, Chugoku, Japan (34° 46’ N, 132° 29’ E; elevation = 420 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Jun-ichi Naito on 17 March 2022. This specimen was deposited in the Shobara Municipal Hiwa Museum for Natural History: 1119-1, Hiwacho Hiwa, Shobara City, Hiroshima Prefecture, 727-0301, Japan.

Paratypes. An adult female (specimen number: YCM-RA598) from Yamaguchicho Yamaguchi, Oda City, Shimane Prefecture, Chugoku, Japan (35° 10’ N, 132° 38’ E; elevation = 380 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Takayuki Iwata on 26 February 2021. An adult male (specimen number: YCM-RA599) from the same locality, collected by Takayuki Iwata on 6 March 2020. The female and male specimens were deposited in the Yokosuka City Museum: 95 Fukadadai, Yokosuka City, Kanagawa Prefecture, 238-0016, Japan.

Diagnosis. A small species (mean snout-vent length = 52.6 mm in male) within the Japanese lentic Hynobius: mid-dorsal line usually absent; distinct white spots (or mottles) on venter frequently present; distinct yellowish-brown line on dorsal edge of the tail frequently present; distinct yellowish-brown line on ventral edge of the tail almost always absent; distinct gular mottling almost always absent; almost always 12 costal grooves (rarely 11); number of costal folds between adpressed limbs usually ≥ 0.0; fifth toe of hindlimb always absent (or rudimentary); dorsal color very variable; U (or V) shaped vomerine teeth series; coil-shaped egg sacs.

Description of holotype. A large individual: HL larger than HW; TAL shorter than SVL; body almost cylindrical; rounded snout; gular fold present; tail gradually compressed toward the tip; unexpanded cloaca; webbing between digits absent; four fingers on each forelimb, order of length II > III > IV > I in both sides; four toes on each hindlimb, order of length III > II > IV > I in both sides; U-shaped like vomerine teeth; skin smooth and shiny; distinct and scattered black spots on dorsum present; MDL absent; DWSV and DWSL present; DYBLDT present; DYBLVT absent; DGM absent. The holotype had the following measurements (in mm): SVL = 53.7, TRL = 42.4, AGD = 29.5, HL = 12.1, TAL = 36.3, MTAW = 3.2, MTAH = 5.3, BTAW = 5.8, BTAH = 5.3, LVTL = 2.7, RVTL = 2.8, VTW = 2.8, HW = 8.8, LFLL = 12.3, RFLL = 12.6, LHLL = 14.9, RHLL = 14.5, L1FL = 1.1, L2FL = 2.6, L3FL = 2.1, L4FL = 1.2, R1FL = 0.9, R2FL = 2.1, R3FL = 2.0, R4FL = 1.1, L1TL = 0.7, L2TL = 2.6, L3TL = 3.8, L4TL = 1.5, R1TL = 1.2, R2TL = 2.9, R3TL = 3.6, R4TL = 1.8, IND = 2.7, IOD = 3.0, LUEW = 1.3, RUEW = 1.1, LSL =3.7, RSL = 3.6, LUEL =2.7, RUEL = 2.5, LLJL = 6.4, RLJL = 5.9, CGN = 12.

Comparisons. The new species does not differ statistically from H. utsunomiyaorum (Table 3), but frequently has ≥ 0.5 CFBALN (10/14 = 71.4%) (Table 4). The new species differs statistically from H. hidamontanus in the following length measurements: RMTAH, RVTL, RFLL, RIND, and RIOD (Table 3). The lengths of these measurements, except for RVTL, are significantly shorter in H. pseudoutsunomiyaorum sp. nov. than in H. hidamontanus (Tables 2-3). DYBLDT is frequently present in H. pseudoutsunomiyaorum sp. nov. (10/14 = 71.4%) but usually absent in H. hidamontanus (6/7 = 85.7%) (Table 4). The distribution ranges of the two species are clearly separated (Figure 1). The new species is parapatrically distributed with Hynobius iwami, but it is distinguishable by the absence of a distinct yellow line on the ventral side of the tail ⁷. In addition, it is parapatrically distributed with Hynobius akiensis, has a distinct fifth toe, and usually lacks a distinct brownish line on the dorsal side of the tail ²⁵. The new species and Hynobius sematonotos and Hynobius kimurae are distributed in proximity, but the former and latter species have distinct silver mottles and distinct yellow mottles, respectively, on the dorsal and lateral sides of the body ^{26, 27}. Hynobius sematonotos and H. kimurae never have yellow or brown lines on the dorsal and ventral sides of the tail, although the latter species has longer vomerine teeth series ^{26, 27}.

Variation and coloration. Morphometric measurements and observations are presented in Tables 2 and 4, respectively. The new species rarely has DYBLVT (1/14 = 7.1%), DGM (1/14 = 7.1%), and 11 CGN (1/14 = 7.1%), rarely has MDL (2/14 = 14.3%) and ≤ −0.5 CFBALN (2/14 = 14.3%), rarely has no DWSV (3/14 = 21.4%) and DYBLDT (4/14 = 28.6%), and sometimes has DWSL (7/14 = 50.0%) (Table 4). Its dorsal coloration was variable and tends to fade to dark gray when preserved. Its ventral coloration was lighter than the dorsal coloration and tends to fade to a darker color when preserved.

Distribution. It is known from Oda City (including the former Oda City), Iinan (including the former Akagi and Tonbara Towns) and Misato (including the former Ochi Town) Towns in Shimane Prefecture and Shobara City (including the former Takano Town) and Kitahiroshima Town (including the former Oasa Town) in Hiroshima Prefecture. The distribution record of Shobara City was only recorded by Matsui et al. ⁷ based on DNA data. The dominant vegetation type of the type locality is mixed forest of Quercus crispula, Quercus serrata, and Cryptomeria japonica (Figure 9).

Larvae and egg sacs. The breeding season of this species is February to April. Breeding locations are pools or ditches at forest edges. The larval and egg sac morphologies of this new species are similar to those of H. utsunomiyaorum ⁷. Larvae usually have black dots on the body and tail (more densely on the tail), but body color is variable. Claws are absent from the fingers and toes of larvae. One pair of balancers are present during the early developmental stages of larvae. Larvae always metamorphose and land within the same year of hatching. Egg sacs are coiled without striations and attached to fallen branches or leaves.

Etymology. The specific name derives from the morphology of this species, which closely resembles that of H. utsunomiyaorum: pseudo (= false in Latin) + utsunomiyaorum (= specific name of Hiba salamander). The suggested Japanese common name is Nisehiba-sanshouo.

Remarks. The morphology of females is unknown; thus, further studies are needed to clarify the morphological characters of females.

ZooBank LSID: urn:lsid:zoobank.org:act:D042FD71-C291-45F3-9179-E2181BA0D316 (Figures 6-7)

Synonymy: Hynobius nebulosus Okawa et al. (1990: Pops. 59-60).

Holotype. An adult male (specimen number: HMNH-AM-325) from Sakugicho Okamibuchi, Miyoshi City, Hiroshima Prefecture, Chugoku, Japan (34° 56’ N, 132° 41’ E; elevation = 520 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Jun-ichi Naito on 14 December 2022. This specimen was deposited in the Shobara Municipal Hiwa Museum for Natural History.

Paratypes. An adult female (specimen number: YCM-RA600) and male (specimen number: YCM-RA601) from Kamiakana, Iinan Town, Shimane Prefecture, Chugoku, Japan (34° 57’ N, 132° 43’ E; elevation = 660 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Takayuki Iwata on 4 March 2021 and 21 March 2020, respectively. These specimens were deposited in the Yokosuka City Museum.

Diagnosis. A small species (mean snout-vent length = 52.4 mm in male) within the Japanese lentic Hynobius: mid-dorsal line frequently absent; distinct white spots (or mottles) on ventral and lateral sides of the body almost always present; distinct yellowish-brown line on dorsal edge of the tail usually absent; distinct yellowish-brown line on ventral edge of the tail always absent; usually 12 costal grooves (rarely 11 or 13); number of costal folds between adpressed limbs always ≤ –0.5; fifth toe of hindlimb always absent; dorsal color very variable; V (or U) shaped vomerine teeth series; coil-shaped egg sacs.

Description of holotype. A large individual: HL larger than HW; TAL shorter than SVL; body almost cylindrical; rounded snout; gular fold present; tail gradually compressed toward the tip; slightly expanded cloaca; webbing between digits absent; four fingers on each forelimb, order of length II > III > I > IV in left and II > I > III > IV in right; four toes on each hindlimb, order of length III > II > IV > I in both sides; V-shaped vomerine teeth; skin smooth and shiny; distinct black spots on dorsum present; MDL present on the latter half of dorsum; DWSV and DWSL present; DYBLDT present; DYBLVT absent; DGM absent. The holotype had the following measurements (in mm): SVL = 55.4, TRL = 43.4, AGD = 29.4, HL = 12.1, TAL = 38.1, MTAW = 3.0, MTAH = 5.4, BTAW = 6.3, BTAH = 6.0, LVTL = 2.4, RVTL = 2.3, VTW = 2.8, HW = 10.0, LFLL = 14.1, RFLL = 12.8, LHLL = 16.6, RHLL = 16.9, L1FL = 1.0, L2FL = 3.0, L3FL = 1.5, L4FL = 0.5, R1FL = 1.5, R2FL = 3.0, R3FL = 1.3, R4FL = 0.7, L1TL = 1.6, L2TL = 3.1, L3TL = 4.2, L4TL = 2.0, R1TL = 1.5, R2TL = 3.2, R3TL = 4.2, R4TL = 1.9, IND = 2.7, IOD = 3.0, LUEW = 1.3, RUEW = 1.1, LSL = 4.2, RSL = 4.1, LUEL = 2.3, RUEL = 2.3, LLJL = 7.4, RLJL = 7.2, CGN = 12.

Comparisons. The new species statistically differs from H. utsunomiyaorum in RHLL (Table 3). The length of the measurement is significantly shorter in H. mengamemontanus sp. nov. than in H. utsunomiyaorum (Tables 2-3). DYBLDT usually absent in H. mengamemontanus sp. nov. (8/10 = 80.0%), but almost always present in H. utsunomiyaorum (14/15 = 93.3%) (Table 4). The new species statistically differs from H. hidamontanus in the following length measurements: RFLL, RIND and RIOD (Table 3). The lengths of these measurements are significantly shorter in H. mengamemontanus sp. nov. than in H. hidamontanus (Tables 2-3). CFBALN always ≤ −0.5 in H. mengamemontanus sp. nov. (10/10 = 100%), but always ≥ 0.5 in H. hidamontanus (7/7 = 100%) (Table 4). Their distribution ranges are clearly separated (Figure 1). The new species statistically dose not differ from H. pseudoutsunomiyaorum sp. nov. (Table 3). DYBLDT usually absent in H. mengamemontanus sp. nov. (8/10 = 80%), but frequently present in H. pseudoutsunomiyaorum sp. nov. (10/14 = 71.4%) (Table 4). CFBALN always ≤ −0.5 in H. mengamemontanus sp. nov. (10/10 = 100%), but usually ≥ 0.0 in H. pseudoutsunomiyaorum sp. nov. (12/14 = 85.8%) (Table 4). The new species and H. sematonotos are geographically close, but the latter have distinct silver mottles on the dorsal and lateral sides of the body ²⁷.

Variation and coloration. Morphometric measurements and observations are presented in Tables 2 and 4, respectively. The new species had rarely 11 (1/10 = 10%) or 13 (1/10 = 10%) CGN and DYBLDT (2/10 = 20%), occasionally had MDL (3/10 = 30%), occasionally had no DGM (4/10 = 40%), and rarely had no DWSV (1/10 = 10.0%) and DWSL (1/10 = 10.0%) (Table 4). The dorsal coloration was very variable, and it tended to fade to dark gray when preserved.

Distribution. It is known from Iinan Town (only former Akagi Town) in Shimane Prefecture and Miyoshi City (only former Sakugi and Funo Villages) in Hiroshima Prefecture. The distribution record of Funo Village was suggested by Okawa et al. ²⁸. The dominant vegetation type in the type locality is Japanese cypress (Chamaecyparis obtusa).

Larvae and egg sacs. The breeding season of this species is March to April. Breeding places are pools or ditches at forest edges. Larvae always metamorphose and land within the same year of hatching. The larval and egg sacs morphologies of this new species are similar with H. pseudoutsunomiyaorum sp. nov. and H. utsunomiyaorum ⁷.

Etymology. The specific epithet “mengamemontanus” derives from Mt. Mengame located in the northeastern part of Miyoshi City (Hiroshima Prefecture) and southern part of Iinan Town (Shimane Prefecture); the new species only occurs around this mountain. Although the mountain is called Mt. Mengame in Hiroshima Prefecture, it is called Mt. Megame in Shimane Prefecture. Many years ago, the mountain was called Mt. Megami, which means goddess in Japanese. Over time, “Megami” and “Mengami” were corrupted into “Megame” and “Mengame” in Shimane Prefecture and Hiroshima Prefecture, respectively. As the name of the mountain originates from Megami (= goddess), the suggested standard Japanese name is Megame-sanshouo.

Remarks. The new species likely forms a monophyletic group with H. pseudoutsunomiyaorum sp. nov., but the monophyly was not supported statistically (Figure 2A). The morphology of females is unknown; thus, further studies are needed to clarify the morphological characters of females.

ZooBank LSID: urn:lsid:zoobank.org:act:2E4C0758-E62A-4A8E-A8D0-C2BBB66051A1 (Figures 8-9)

Synonymy: Hynobius utsunomiyaorum Matsui et al. (2019: clade B2b4, H6).

Holotype. An adult male (specimen number: TRPM-ARA-0000015) from Kamisaibara, Kagamino Town, Okayama Prefecture, Chugoku, Japan (35° 18’ N, 133° 54’ E; elevation = 780 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Takayuki Iwata on 3 May 2019. This specimen was deposited in the Tottori Prefectural Museum: 2-124, Higashimachi, Tottori City, Tottori Prefecture, 680-0011, Japan.

Paratypes. An adult female (specimen number: YCM-RA602) from the same locality of holotype, collected by Takayuki Iwata on 23 April 2021. An adult male (specimen number: YCM-RA603) from Hiruzenshitao, Maniwa City, Okayama Prefecture, Chugoku, Japan (35° 15’ N, 133° 46’ E; elevation = 490 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Masaru Yamada on 23 October 2021. These specimens were deposited in the Yokosuka City Museum.

Diagnosis. A small species (mean snout-vent length = 50.4 mm in male) within the Japanese lentic Hynobius: mid-dorsal line usually present; distinct white spots (or mottles) on ventral and lateral sides of the body always present; distinct yellowish-brown line on ventral edge of the tail always absent; distinct gular mottling always absent; almost always 12 costal grooves (rarely 11); number of costal folds between adpressed limbs frequently ≥ 0.0; fifth toe of hindlimb always absent (or rudimentary); dorsal color very variable; V or U-shaped vomerine teeth series; coil-shaped egg sacs.

Description of holotype. A large individual: HL larger than HW; TAL shorter than SVL; body almost cylindrical; rounded snout; gular fold present; tail gradually compressed toward the tip; unexpanded cloaca; webbing between digits absent; four fingers on each forelimb, order of length II > III > IV > I in both sides; four toes on each hindlimb, order of length III > II > IV > I in both sides; V-shaped vomerine teeth; skin smooth and shiny; distinct and scattered brownish-yellow spots on dorsum present; MDL present on the latter half of dorsum; DWSV and DWSL present; DYBLDT absent; DYBLVT absent; DGM absent. The holotype had the following measurements (in mm): SVL = 56.3, TRL = 43.1, AGD = 29.6, HL = 13.3, TAL = 42.9, MTAW = 2.9, MTAH = 5.1, BTAW = 6.5, BTAH = 5.5, LVTL = 2.5, RVTL = 2.6, VTW = 2.7, HW = 9.5, LFLL = 14.1, RFLL = 14.7, LHLL = 17.7, RHLL = 16.9, L1FL = 0.8, L2FL = 2.6, L3FL = 2.4, L4FL = 1.0, R1FL = 0.9, R2FL = 2.7, R3FL = 2.4, R4FL = 1.1, L1TL = 1.3, L2TL = 2.9, L3TL = 4.1, L4TL = 2.7, R1TL = 1.4, R2TL = 3.4, R3TL = 4.0, R4TL = 1.9, IND = 2.7, IOD = 3.5, LUEW = 1.7, RUEW = 1.6, LSL = 3.6, RSL = 3.4, LUEL =2.6, RUEL = 2.4, LLJL = 7.7, RLJL = 6.7, CGN = 12.

Comparisons. The new species statistically differs from H. utsunomiyaorum in the following length measurements: SVL and R2FL. The lengths of SVL and R2FL are significantly shorter and longer in H. sakuhokumontanus sp. nov. than in H. utsunomiyaorum, respectively (Tables 2-3). Their distribution ranges are slightly separated (Figure 1). The new species statistically differs from H. hidamontanus in the following length measurements: RHW, RIND, and RIOD (Table 3). The lengths of these measurements are significantly shorter in H. sakuhokumontanus sp. nov. than in H. hidamontanus (Tables 2-3). MDL usually present in H. sakuhokumontanus sp. nov. (12/14 = 85.7%), but always absent in H. hidamontanus (7/7 = 100%) (Table 4). Their distribution ranges are clearly separated (Figure 1). The new species statistically dose not differ from H. pseudoutsunomiyaorum sp. nov., but the new species usually has MDL (12/14 = 85.7%) towards H. pseudoutsunomiyaorum sp. nov. usually has no MDL (12/14 = 85.7%) (Table 4). The distribution ranges of two species are clearly separated (Figure 1). Although the new species statistically dose not differ from H. mengamemontanus sp. nov. (Table 3), CFBALN frequently ≥ 0.0 in H. sakuhokumontanus sp. nov. (11/14 = 78.6%) against always ≤–0.5 (10/10 = 100%) in H. mengamemontanus sp. nov. (Table 4). Their distribution ranges are clearly separated (Figure 1). The new species and H. sematonotos and H. kimurae are geographically close, but the former has distinct silver mottles, latter has distinct yellow mottles on the dorsal and lateral sides of the body ^{26, 27}. Hynobius sematonotos and H. kimurae never has yellow or brown lines on the dorsal and ventral sides of the tail, and latter has longer vomerine teeth series ^{26, 27}.

Variation and coloration. Morphometric measurements and observations are presented in Tables 2 and 4, respectively. The new species very rarely had 11 CGN (1/14 = 7.1%), rarely had ≤ –0.5 CFBALN (3/14 = 21.4%), rarely had no MDL (2/14 = 14.3%), occasionally had no DYBLDT (5/14 = 35.7%) (Table 4). The dorsal coloration was very variable, and it tended to fade to dark gray when preserved.

Distribution. This new species is endemic to Okayama Prefecture; it is known from Kagamino Town (only former Kamisaibara Village) and Maniwa City (only former Chuka Village). The dominant vegetation type in the type locality is Quercus acutissima.

Larvae and egg sacs. The breeding season of this species is usually April to May, but the breeding season of the population of former Chuka Village is December to January. Larvae usually metamorphose and land within the same year of hatching. However, there are confirmed instances of populations in which most individuals metamorphose and land in following year. Breeding places are pools or ditches at forest edges. The larval and egg sacs morphologies of this new species are similar with H. pseudoutsunomiyaorum sp. nov. and H. utsunomiyaorum ⁷.

Etymology. The specific epithet “sakuhokumontanus” derives from the Sakuhoku Mountains located in the northern part of Okayama Prefecture where the new species occurs. The suggested standard Japanese name is Sakuhoku-sanshouo.

Remarks. The distribution ranges of this new species and H. utsunomiyaorum are separated, but re-examination of the record of H. utsunomiyaorum from Kagamino Town lacking detailed information about sampling points ⁷ is essential. The morphology of female is unknown, so further studies to clarify the morphological characters of females are needed.

ZooBank LSID: urn:lsid:zoobank.org:act:BC0BEBC9-9E6E-467F-AE86-55256FD770DB (Figures 10-11)

Synonymy: Hynobius utsunomiyaorum Matsui et al. (2019: clade B2b4, H7-H8).

Holotype. An adult male (specimen number: TRPM-ARA-0000016) from Ogaya, Nishiawakura Village, Okayama Prefecture, Chugoku, Japan (35° 13’ N, 134° 23’ E; elevation = 1020 m above sea level [a.s.l.]; in all cases, datum = WGS84), collected by Takayuki Iwata on 2 May 2019. This specimen was deposited in the Tottori Prefectural Museum.

Paratypes. An adult female (specimen number: YCM-RA604) and male (specimen number: YCM-RA605) from the same locality of holotype, collected by Takayuki Iwata on 28 April 2021 and 2 May 2019. These specimens were deposited in the Yokosuka City Museum.

Diagnosis. A small species (mean snout-vent length = 55.5 mm in male) within the Japanese lentic Hynobius: distinct white spots (or mottles) on ventral and lateral sides of the body always present; distinct yellowish-brown line on dorsal edge of the tail frequently absent; distinct yellowish-brown line on ventral edge of the tail always absent; distinct gular mottling always absent; frequently 12 costal grooves (rarely 11 or 13); number of costal folds between adpressed limbs almost always ≥ –0.5; fifth toe of hindlimb always absent; dorsal color almost always brownish with black spots; V or U shaped vomerine teeth series; coil-shaped egg sacs.

Description of holotype. A moderately large individual: HL larger than HW; TAL shorter than SVL; body almost cylindrical; rounded snout; gular fold present; tail gradually compressed toward the tip; unexpanded cloaca; webbing between digits absent; four fingers on each forelimb, order of length II > III > I > IV in both sides; four toes on each hindlimb, order of length III > II > IV > I in both sides; V-shaped like vomerine teeth; skin smooth and shiny; distinct black spots on dorsum present; MDL present; DWSV and DWSL present; DYBLDT present; DYBLVT absent; DGM absent. The holotype had the following measurements (in mm): SVL =56.1, TRL = 43.5, AGD = 28.4, HL = 12.4, TAL = 40.9, MTAW = 3.2, MTAH = 5.2, BTAW = 6.0, BTAH = 4.9, LVTL = 2.6, RVTL = 2.5, VTW = 3.1, HW = 9.2, LFLL = 13.6, RFLL = 14.5, LHLL = 17.7, RHLL = 18.3, L1FL = 0.9, L2FL = 2.5, L3FL = 1.7, L4FL = 0.7, R1FL = 0.8, R2FL = 2.9, R3FL = 1.9, R4FL = 0.7, L1TL = 1.5, L2TL = 3.4, L3TL = 4.2, L4TL = 2.3, R1TL = 1.1, R2TL = 3.4, R3TL = 4.6, R4TL = 2.4, IND = 2.7, IOD = 3.6, LUEW = 1.0, RUEW = 0.9, LSL =3.9, RSL = 3.5, LUEL =2.5, RUEL = 2.3, LLJL = 7.4, RLJL = 7.2, CGN = 12.

Comparisons. The new species statistically differs from H. utsunomiyaorum in the following length measurements: RMTAH, RIOD, and RLJL (Table 3). Their lengths are significantly shorter in H. ushiromontanus sp. nov. than in H. utsunomiyaorum (Tables 2-3). DYBLDT frequently absent in H. ushiromontanus sp. nov. (11/14 = 78.6%), but almost always present in H. utsunomiyaorum (14/15 = 93.3%) (Table 4). Distribution ranges of two species are clearly separated (Figure 1). The new species statistically differs from H. hidamontanus in the following length measurements: SVL, RMTAW, RMTAH, RVTL, RHW, RFLL, R2FL, RIND, RIOD, and RSL. The lengths of these measurements excepting for SVL and RVTL are significantly shorter in H. ushiromontanus sp. nov. than in H. hidamontanus. CFBALN frequently ≤ 0.0 in H. ushiromontanus sp. nov. (11/14 = 78.6%), but always ≥ 0.5 in H. hidamontanus (7/7 = 100%) (Table 4). Distribution ranges of two species are clearly separated (Figure 1). The new species statistically differs from H. pseudoutsunomiyaorum sp. nov. in the following length measurements: RHL, RHW, RIOD, and RLJL (Table 3). Their lengths are significantly shorter in H. ushiromontanus sp. nov. than in H. pseudoutsunomiyaorum sp. nov. (Tables 2-3). DYBLDT frequently absent in H. ushiromontanus sp. nov. (11/14 = 78.6%), but frequently present in H. pseudoutsunomiyaorum sp. nov. (10/14 = 71.4%) (Table 4). CFBALN frequently ≤ 0.0 in H. ushiromontanus sp. nov. (11/14 = 78.6%), but frequently ≥ 0.5 in H. pseudoutsunomiyaorum sp. nov. (10/14 = 71.4%) (Table 4). Distribution ranges of two species are clearly separated (Figure 1). The new species statistically differs from H. mengamemontanus sp. nov. in the following length measurements: RMTAW, RMTAH, HW, and RIOD (Table 3). Their lengths are significantly shorter in H. ushiromontanus sp. nov. than in H. mengamemontanus sp. nov. (Tables 2-3). Distribution ranges of then are clearly separated (Figure 1). The new species statistically differs from H. sakuhokumontanus sp. nov. in the following length measurements: SVL, RHL, RMTAW, RMTAH, RFLL, R2FL, R3FL, RIOD, RUEW, and RLJL (Table 3). Their lengths excepting for SVL are significantly shorter in H. ushiromontanus sp. nov. than in H. sakuhokumontanus sp. nov. (Tables 2-3). Distribution ranges of then are clearly separated (Figure 1). The new species and H. sematonotos and H. kimurae are geographically close, but the former has distinct silver mottles, latter has distinct yellow mottles on the dorsal and lateral sides of the body ^{26, 27}. Hynobius sematonotos and H. kimurae never has yellow or brown lines on the dorsal and ventral sides of the tail, and latter has longer vomerine teeth series ^{26, 27}.

Variation and coloration. Morphometric measurements and observations are presented in Tables 2 and 4, respectively. The new species very rarely had 11 CGN (1/14 = 7.1%), rarely had 13 CGN (2/14 = 14.3%), rarely had DYBLDT (3/14 = 21.4%) and ≥ 0.5 CFBALN (3/14 = 21.4%), and sometimes had MDL (7/14 = 50%). The dorsal coloration is almost always brownish with black spots and it tended to fade to dark gray when preserved.

Distribution. It is known from Nishiawakura Village, Okayama Prefecture, and Shiso City (only former Chikusa Town), Hyogo Prefecture. The dominant vegetation type in the type locality is Chamaecyparis obtusa.

Larvae and egg sacs. The breeding season of this species is April to May. Breeding places are pools or ditches at forest edges. Larvae always metamorphose and land within the same year of hatching. The larval and egg sacs morphologies of this new species are similar with H. pseudoutsunomiyaorum sp. nov. and H. utsunomiyaorum ⁷.

Etymology. The specific epithet “ushiromontanus” derives from Mt. Ushiro located in the northeastern part of Okayama Prefecture and northwestern part of Hyogo Prefecture, and it occurs around the mountain. The suggested standard Japanese name is Ushiroyama-sanshouo.

Remarks. The new species forms a monophyletic group with H. sakuhokumontanus sp. nov., although they are clearly distinct in terms of morphology (Table 3; Figure 3). The morphology of female is unknown, so further studies to clarify the morphological characters of females are needed.

Holotype. An adult male (specimen number: KUHE 6491) from Rokunohara, Mt. Hiba, Saijocho Yuki, Shobara City, Hiroshima Prefecture (35°04'20"N, 133°04'18"E, alt. 850 m), collected by Masafumi Matsui on 3 May 1982 ⁷. This specimen was deposited in the Graduate School of Human and Environmental Studies, Kyoto University: Yoshidahonmachi, Sakyo Ward, Kyoto City, Kyoto Prefecture, 606-8501, Japan ⁷.

Diagnosis. A small species (mean snout-vent length = 54.7 mm in male) within the Japanese lentic Hynobius: distinct white spots (or mottles) on ventral and lateral sides of the body usually present; distinct yellowish-brown line on dorsal edge of the tail almost always present; distinct yellowish-brown line on ventral edge of the tail always absent; distinct gular mottling usually absent; usually 12 costal grooves (very rarely 11 or 13); number of costal folds between adpressed limbs usually ≤ 0.0; fifth toe of hindlimb always absent (or rudimentary) except the populations from Daisen (Tottori Prefecture) and around areas; dorsal color very variable; U or V-shaped vomerine teeth series; coil-shaped egg sacs.

Description of topotype. An adult male (specimen number: YCM-RA606) from near the type locality (35°04'20"N, 133°04'18"E), collected by Takayuki Iwata on 3 April 2020 (Figure 12). This specimen was deposited in Yokosuka City Museum. A moderately large individual: HL larger than HW; TAL shorter than SVL; body almost cylindrical; rounded snout; gular fold present; tail gradually compressed toward the tip; unexpanded cloaca; webbing between digits absent; four fingers on each forelimb, order of length II > III > I > IV in both sides; four toes on each hindlimb, order of length III > II > IV > I in left side and III > IV > II > I in right side; U-shaped like vomerine teeth; skin smooth and shiny; distinct and scattered black spots on dorsum absent; distinct yellowish-brown mottling present on dorsum; MDL present on the latter half of dorsum; DWSV and DWSL present; DYBLDT present; DYBLVT absent; DGM absent. The holotype had the following measurements (in mm): SVL =52.8, TRL = 41.1, AGD = 27.3, HL = 12.2, TAL = 39.7, MTAW = 2.7, MTAH = 4.9, BTAW = 5.6, BTAH = 4.8, LVTL = 2.2, RVTL = 2.0, VTW = 2.3, HW = 8.8, LFLL = 16.1, RFLL = 15.8, LHLL = 18.3, RHLL = 17.9, L1FL = 1.3, L2FL = 3.6, L3FL = 2.2, L4FL = 0.8, R1FL = 1.2, R2FL = 2.9, R3FL = 2.3, R4FL = 1.0, L1TL = 1.8, L2TL = 3.9, L3TL = 5.4, L4TL = 2.9, R1TL = 1.7, R2TL = 3.6, R3TL = 4.7, R4TL = 3.7, IND = 2.8, IOD = 3.0, LUEW = 1.0, RUEW = 1.1, LSL =3.5, RSL = 3.9, LUEL =2.2, RUEL = 2.0, LLJL = 6.4, RLJL = 6.2, CGN = 12.

Comparisons. This species statistically differs from H. hidamontanus in the following length measurements: SVL, RVTL, RHW, RFLL, R2FL, and RIOD (Table 3). The lengths of these measurements excepting for SVL and RVTL are significantly shorter in H. utsunomiyaorum than in H. hidamontanus (Tables 2-3). DYBLDT almost always present in H. utsunomiyaorum (14/15 = 93.3%), but usually absent in H. hidamontanus (6/7 = 85.7%) (Table 4). Distribution ranges of then are clearly separated (Figure 1). This species is parapatrically distributed with Hynobius kunibiki and Hynobius setoi, but it is distinguishable in terms of the absence of distinct fifth toe and the distinct yellow line on ventral side of the tail ¹⁰. Hynobius utsunomiyaorum is also distributed near the range of H. akiensis, but it has distinct fifth toe and usually lacks the distinct brownish line on dorsal side of the tail (at least in male) ^{7, 25}. The new species and H. sematonotos and H. kimurae are geographically close, but the former has distinct silver mottles, latter has distinct yellow mottles on the dorsal and lateral sides of the body ^{26, 27}. Hynobius sematonotos and H. kimurae never has yellow or brown lines on the dorsal and ventral sides of the tail, and latter has longer vomerine teeth series ^{26, 27}.

Variation and coloration. Morphometric measurements and observations are presented in Tables 2 and 4, respectively. The new species very rarely had 11 (1/15 = 6.7%) or 13 CGN (1/15 = 6.7%), rarely had DGM (3/15 = 20%) and ≥ 0.5 CFBALN (3/15 = 20%), very rarely had no DYBLDT (1/15 = 6.7%), rarely had no DWSL (2/15 = 13.3%) and DWSV (3/15 = 20%), and sometimes had no MDL (6/15 = 40.0%). The dorsal coloration is very variable, and it tended to fade to dark gray when preserved.

Distribution. Based on our field surveys and some articles, it is known from Tottori, Shimane, Okayama, and Hiroshima Prefectures: Kotoura (former Akasaki and/or Tohaku Towns), Daisen (only former Daisen Town), Hoki (former Kishimoto and Mizokuchi Towns), Nanbu (only former Aimi Town), Kofu, Hino, and Nichinan Towns, Tottori Prefecture, Matsue (only former Mihonoseki Town), Yasugi (only former Hakuta and Hirose Towns), and Unnan (only former Daito and Kisuki Towns and Yoshida Village) Cities, and Okuizumo (former Nita and Yokota Towns) and Iinan (only former Tonbara Town) Towns, Shimane Prefecture, Maniwa (only former Kawakami and Yatsuka Villages), Niimi (only former Tetta Town), and Takahashi (only former Bittchu Town) Cities, Okayama Prefecture, and Shobara City (former Shobara City, and Tojo, Saijo, Hiwa, Takano, Kuchiwa, and Soryo Towns) and Jinsekikogen Town (only former Jinseki, Yuki, and Sanwa Towns), Hiroshima Prefecture ^{28, 29, 30}. According to Matsui et al. ⁷, this species is also distributed in Kagamino Town (Pop. 71: Figure 1). Although the specific sampling points are unknown, the haplotype of this new species is not found in the area of former Kamisaibara Village based on our field surveys; therefore, its habitat area may be former Okutsu Town (or near the boundary of the Okutsu Town and Kamisaibara Village) located in the south side of Kamisaibara Village, i.e., an area that we were unable to survey in detail. This population shares a haplotype with Tetta Town (Pop. 64: Figures 1-2); thus, the new species may be distributed in the area located between these towns (i.e., former Niimi City and former Katsuyama and Kuse Towns of Maniwa City, located in BA2: Figure 1).

Larvae and egg sacs. The larval and egg sacs morphologies of this species are similar with H. pseudoutsunomiyaorum sp. nov. Larvae usually metamorphose and land within the same year of hatching. However, there are confirmed instances of populations in which most individuals metamorphose and land in following year. According to Matsui et al. ⁷, the breeding season of H. utsunomiyaorum is late March to late May, but our field surveys do not support this conclusion. The breeding season of H. utsunomiyaorum is revised as follows: from late December to late May. Breeding places are pools or ditches at forest edges.

Remarks. Based on the previous study ⁷, the lowest altitude record for known localities was 35 m. However, our field surveys contradict this finding, leading to a revision of the lowest altitude of the known locality to 5 m. The population of H. utsunomiyaorum from Daisen (Pops. 56, 69, and 72) differs from other H. utsunomiyaorum populations because it has a distinct fifth toe ^{7, 31}. However, salamanders in that population are clearly polyphyletic (Figure 2), and overall morphometric differences are unclear (Figure 3).

5. Discussion

The first taxonomic investigation of H. utsunomiyaorum was performed by Matsui et al. ³². According to this study, phylogenetic analyses using allozyme data did not support the monophyletic relationship of H. utsunomiyaorum, although sampling localities were limited ³². The second taxonomic investigation of H. utsunomiyaorum was performed by Matsui et al. ⁷. According to this study, H. utsunomiyaorum was a non-monophyletic group based on a mitochondrial phylogeny, but nuclear data indicated genetic similarity among the groups, although they were not monophyletic ⁷. The third taxonomic investigation of H. utsunomiyaorum was conducted by Sugawara et al. ¹⁰. In this study, the Izumo lineage of H. utsunomiyaorum suggested by Hayashi and Ooi ¹³ was described as H. kunibiki, and the polyphyly of H. utsunomiyaorum was partly resolved ¹⁰. However, the validity of H. hidamontanus included in the clade of H. utsunomiyaorum based on the morphological, phylogenetic, and evolutionary species concepts was unconfirmed. Matsui et al. ⁷ conducted nuclear data analysis rather than not phylogenetic analyses, and the nuclear data of H. hidamontanus was not included, despite this species clearly being included in the clade of H. utsunomiyaorum ^{7, 10}. Thus, the taxonomic validity of H. hidamontanus could not be assessed using the phylogenetic species concept. Based on the phylogenetic and evolutionary species concepts, H. utsunomiyaorum should be divided into five groups if H. hidamontanus is a valid species: H. pseudoutsunomiyaorum sp. nov., H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov. + H. ushiromontanus sp. nov., H. utsunomiyaorum, and H. hidamontanus. From a morphological perspective, the morphological differences of H. hidamontanus were clear when each species was compared (Table 3; Figure 3). Although H. pseudoutsunomiyaorum sp. nov. was not distinguishable from H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov., and H. utsunomiyaorum based on statistical tests (Table 3), the ranges of the discriminant scores of H. mengamemontanus sp. nov. and H. sakuhokumontanus sp. nov. did not overlap with the range of H. pseudoutsunomiyaorum sp. nov. (Figure 3). Additionally, H. mengamemontanus sp. nov. and H. sakuhokumontanus sp. nov. were distinguished from H. pseudoutsunomiyaorum sp. nov. based on CFBALN (Table 4) and the frequency of the presence of MDL (as well as geographical isolation), respectively. Furthermore, the results of canonical discriminant analysis supported the morphological differentiation between H. sakuhokumontanus sp. nov. and H. pseudoutsunomiyaorum sp. nov. (Figure 3). However, H. pseudoutsunomiyaorum sp. nov. and H. utsunomiyaorum were similar with distribution areas that were not clearly separated (Figure 1). In terms of morphology, the discriminant scores of these two species overlapped based on morphometric data (Figure 3). Our analysis mainly depended on individuals from Pop. 4, which were slightly isolated from the main distribution area of H. pseudoutsunomioyaorum sp. nov. If this population has coincidently similar morphological characteristics with H. utsunomiyaorum, the morphology of H. pseudoutsunomioyaorum sp. nov. may be similar to that of H. utsunomiyaroum. Thus, it is possible that morphometric differences between H. pseudoutsunomiyaorum sp. nov. and H. utsunomiyaorum will be detected if individuals from other populations (except for Pops. 4 and 5) are examined. In Japanese Hynobius, two species may still be considered distinct species in some cases if the range of scores overlap ³³. CFBALN tended to be larger in H. pseudoutsunomiyaorum sp. nov. than in H. utsunomiyaorum (Table 4). Hynobius sakuhokumontanus sp. nov. and H. ushiromontanus sp. nov. formed a monophyletic clade, consistent with the finding of Matsui et al. ⁷ (Figure 2), but they were clearly distinguishable based on the morphological species concept (Table 3; Figure 3). Furthermore, the distribution areas of the two species were clearly separated (Figure 1). Two groups may still be different species despite being monophyletic (e.g., Hynobius sengokui vs. Hynobius tokyoensis and Hynobius tsurugiensis vs. Hynobius guttatus) ^{33, 34}. Moreover, H. sakuhokumontanus sp. nov. and H. ushiromontanus sp. nov did not form a monophyletic clade based on the phylogeny using allozyme data ³². Finally, we described the four new species and recognized two described species based on three species concepts. Results from comparisons using morphometric data did not reveal significant differences between H. pseudoutsunomiyaroum sp. nov. vs. H. mengamemontanus sp. nov., H. pseudoutsunomiyaroum sp. nov. vs. H. sakuhokumontanus sp. nov., H. pseudoutsunomiyaroum sp. nov. vs. H. utsunomiyaorum, and H. mengamemontanus sp. nov. vs. H. sakuhokumontanus sp. nov. (Table 3), despite observed morphological differences being confirmed in these combinations (Table 4). To ascertain whether these combinations indeed exhibit significant differences in morphometric data, a more extensive study involving additional individuals and localities of the four species (H. pseudoutsunomiyaroum sp. nov., H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov., and H. utsunomiyaorum) is warranted. This study was conducted without female individuals. To accumulate the taxonomic perspectives of the six species using this study, it is necessary to perform morphological comparisons among them using female individuals in a future study.

The population of H. utsunomiyaorum from Daisen (Pops. 56, 69, and 72) exhibits morphological differences in some characters ³¹ and generally has a fifth toe ⁷; therefore, it was previously suggested that the Daisen populations may be a new species ³¹. The present study included the Daisen population, but it was polyphyletic (Figure 2B) without conspicuous morphological differences (Figure 3). Thus, the populations around Daisen are likely a morphological variation of H. utsunomiyaorum characterized by the presence of a fifth toe.

The phylogenetic problem of H. utsunomiyaorum was resolved in the present study, but the phylogenetic problem of H. hidamontanus remains. Hynobius tenuis, described by Nambu ³⁵, is a junior synonym of H. hidamontanus ³⁶. This synonymization was assigned based on the morphological and phylogenetic species concepts using skull characters and allozyme data ³⁶. Skull morphology is not absolutely needed for species description according to Matsui et al. ⁷; therefore, it is possible the two species are different species based on the morphological species concept if other as yet undetected morphological differences exist between H. hidamontanus and H. tenuis. However, phylogenetic analyses using allozyme data were performed previously using UPGMA and neighbor-joining methods ³⁶, and these methods are insufficient for constructing a robust molecular phylogenetic tree to assess the phylogenetic and evolutionary species concepts ³⁷. Moreover, the taxonomic perspectives of Hynobius nebulosus based on the allozyme data of Matsui et al. ³² were contradicted by Matsui et al. ⁷; thus, the synonymization of H. tenuis using allozyme data lacks credibility. Furthermore, Matsui et al. ³⁶ constructed a phylogeny without conducting genetic surveys across the distribution range of H. tenuis (i.e., using only two sampling points). Thus, detailed morphological comparisons and molecular phylogenetic analyses are needed to confirm the species validity of H. tenuis across the entire distribution range of H. hidamontanus.

The distribution area of H. utsunomiyaorum has been changed following the present description, i.e., it is not distributed in Hyogo Prefecture (Table 1; Figure 1). Although the distribution range of H. utsunomiyaorum decreased, H. utsunomiyaorum still has a wider distribution range than those of the other four new species (Figure 1). The estimated distribution ranges of H. pseudoutsunomiyaorum sp. nov., H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov., and H. ushiromontanus sp. nov. are approximately less than 120, 10, 30, and 20 km², respectively. In Japanese Hynobius, the species with the narrowest distribution range is H. tosashimizuensis, i.e., around 0.35 km^{2 4}. The scale of the distribution ranges of H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov., and H. ushiromontanus sp. nov. are larger than that of H. tosashimizuensis, but the narrow distribution ranges may be comparable with that of Hynobius amakusaensis ⁵. The distribution range of H. pseudoutsunomiyaorum sp. nov. is wider than that of the other three new species, but it is narrower than those of H. hidamontanus and Hynobius abei ³⁸, which are well-protected by local governments in law (e.g., Internationally Rare Species of Wild Fauna/Flora). The rank of H. utsunomiyaorum in the Japanese Red Data List is “vulnerable” (VU) ³⁹. However, the conservation status of the four new species should be reappraised. To conserve these species, it is necessary to conduct ecological studies on them (e.g., migratory distance, life-span, or survival rate in habitats) or the environment of their habitats (e.g., soil temperature, water temperature, or dissolved oxygen concentration of their breeding ponds). Therefore, ecological studies of these species should be conducted at the earliest opportunity to avoid their extinction.

6. Conclusions

The clade of Hiba salamander, H. utsunomiyaorum (sensu lato), including H. hidamontanus was divided into six species, and the four new species that were described based on the three species concepts used in this study are as follows: H. pseudoutsunomiyaorum sp. nov., H. mengamemontanus sp. nov., H. sakuhokumontanus sp. nov., and H. ushiromontanus sp. nov.. The distribution areas between H. pseudoutsunomiyaorum sp. nov. vs. H. mengamemontanus sp. nov., H. pseudoutsunomiyaorum sp. nov. vs. H. utsunomiyaorum, and H. sakuhokumontanus sp. nov. vs. H. utsunomiyaorum are peripatric or parapatric based on our field surveys; therefore, introgression between them may have occurred currently or in the recent past. Moreover, there is a possibility of genetic pollution due to human activities. Therefore, studies are required on conservation genetics using microsatellite or single nucleotide polymorphism markers to elucidate the detailed genetic structure of the population.

ACKNOWLEDGEMENTS

We are grateful to Toshiaki Iwata and Kazue Iwata for help with field surveys and sampling; Shoko Kumasaki, Nobue Akai, Masayuki Kitagawa, and Masashi Yamamoto for providing habitat information of the species used in this study; and Shingo Nakamura and Masaki Shindoh of Shobara Municipal Hiwa Museum for Natural History, Kei Ichisawa of Tottori Prefectural Museum, and Kiyoshi Hagiwara of Yokosuka City Museum for supporting the registration of the type specimens used in this study; the governments of Nagano Prefecture (permission number: 3 HOKUACHISO DAI 14 GO) and Hakuba Village (permission number: 2 HAKUKYOSYO DAI 55 GO) for giving the collecting permission of H. hidamontanus. We would like to thank Enago for the English review of this manuscript. This study was funded by JSPS KAKENHI, grant number JP18K02583.

References