Zootaxa 3852 (2): 203–226 www.mapress.com /zootaxa / Copyright © 2014 Magnolia Press

Article

ISSN 1175-5326 (print edition)

ZOOTAXA

ISSN 1175-5334 (online edition)

http://dx.doi.org/10.11646/zootaxa.3852.2.3 http://zoobank.org/urn:lsid:zoobank.org:pub:1A49E97D-CD5D-4666-B738-062109FAB0A7

Three new species of the salamander genus Hynobius (Amphibia, Urodela, Hynobiidae) from Kyushu, Japan KANTO NISHIKAWA1 & MASAFUMI MATSUI Graduate School of Human and Environmental Studies, Kyoto University, Yoshida Nihonmatsu-cho, Sakyo-ku, Kyoto 606-8501, Japan 1 Corresponding author. E-mail: [email protected]

Abstract Three new species of lotic breeding Hynobius, formerly assigned to H. boulengeri, are described from the Kyushu region, southwestern Japan. They differ from all the known congeners by a unique combination of body size, character ratios, coloration, mtDNA, and allozymic characteristics. Together with H. stejnegeri they form a clade, which is not a sister group of H. boulengeri, and their speciation in Kyushu is surmised to have occurred at the end of Miocene, accompanied by differentiations in larval period and metamorphosing size. Measures of conservation of these new species are discussed briefly. Key words: Urodela, Hynobius, new species, Kyushu, Japan, taxonomy

Introduction Hynobius boulengeri (Thompson 1912) is a small-sized Japanese salamander of pure lotic breeding behavior, like H. hirosei, H. katoi, H. kimurae, H. naevius, H. stejnegeri, and H. yatsui (after Nishikawa et al. 2007). Although H. tsuensis and H. okiensis also breed in lotic waters, they are thought to have acquired this habit secondarily (Matsui et al. 2007; Kim et al. 2007). Hynobius boulengeri was originally described under a distinct genus Pachypalaminus (Thompson 1912), and this name was long used (e.g. Sato 1943). After the species was moved to Hynobius (Nishio et al. 1987), intraspecific variation was little studied. In 2001, we first examined allozymic variation in populations of H. boulengeri, and found that they were separated into three groups, i.e., Kii Peninsula on Honshu mainland, Shikoku Island, and Kyushu Island groups, that differ from each other at distinct specific levels (Nishikawa et al. 2001). Later, we found that H. stejnegeri, endemic to the Kyushu Central Mountains, Kyushu Island, was genetically nested into the Kyushu populations of H. boulengeri (Nishikawa et al. 2005). On the basis of morphological and allozymic studies on H. boulengeri sensu lato and H. stejnegeri (hereafter, collectively called the H. boulengeri complex), we proposed to treat the population from the Kii Peninsula as H. boulengeri sensu stricto, that from Shikoku Island as H. hirosei Lantz 1931, and the population from the Kyushu Central Mountains as H. stejnegeri Dunn 1923. However, the name H. boulengeri sensu lato was retained for the remaining populations from Kyushu because of the paucity of comparable specimens (Nishikawa et al. 2007). In order to determine the taxonomic status of the Kyushu populations of H. boulengeri sensu lato, we accumulated additional samples and compared them by use of mitochondrial DNA sequencing and external morphology. As a result, we confirmed independent specific status for Kyushu populations from the SoboKatamuki Mountains, Amakusa Islands, and Osumi Peninsula. Thus, in the present paper, we describe each of them as a new species.

Material and methods Salamanders collected in the field were fully anesthetized by saturated acetone-chloroform solution and tissues

Accepted by M. Vences: 21 Jul. 2014; published: 14 Aug. 2014

203

were removed for genetic study. They were then fixed in 10% formalin before final preservation in 70% ethanol. Sex and maturity were determined by direct gonad observation. These specimens were stored at the Graduate School of Human and Environmental Studies, Kyoto University (KUHE). We also examined preserved specimens in the collections of the National Science Museum of Tokyo (NSMT), the Osaka Museum of Natural History (OMNH), and Institute for Amphibian Biology, Graduate School of Science, Hiroshima University (IABHU), and private collections of Dr. M. Sakamoto (SK), Mr. S. Sato (ST), and Mr. S. Tanabe (T). Details of samples used for comparisons are shown in Appendix 1, and distribution of the species of the H. boulengeri complex in Kyushu is shown in Fig. 1.

FIGURE. 1. Map of Kyushu Island showing ranges of species of the Hynobius boulengeri complex from Kyushu.

For genetic comparisons, we obtained the complete sequence data of the gene encoding cytochrome b (cyt b) of mitochondrial DNA (mtDNA) from muscle or liver tissue samples preserved in 99% ethanol. In order to assign the Kyushu populations of H. boulengeri sensu lato, we sequenced one sample each from Mt. Sobo (KUHE 22889), Osumi Peninsula (KUHE 24797), and Amakusa Island (KUHE 30338), and some species of Hynobius (H. stejnegeri: KUHE 22817 from Miyazaki; H. hirosei: T 3009 from Kochi; H. yatsui: KUHE 28011 from Kumamoto; H. nebulosus: KUHE 24693 from Nagasaki). Methods for DNA extraction, and amplification and sequencing of the mtDNA fragments are the same as those reported by Aoki et al. (2013). We deposited the resultant sequence in GenBank (Accession numbers AB921162–921168). For comparisons, we added GenBank data of H. boulengeri from Nara (KUHE 25653: AB266675), H. katoi from Shizuoka (KUHE 37128: AB266673), H. kimurae from Shiga (KUHE16689: AB266674), H. retardatus from Hokkaido (KUHE 13034: AB363609) and

204 · Zootaxa 3852 (2) © 2014 Magnolia Press

NISHIKAWA & MATSUI

Salamandrella keyserlingii from Hokkaido (KUHE 13057: AB363573). For tree construction and calculation of genetic distances (uncorrected p-distance), we followed Aoki et al. (2013). We estimated the time of divergence using the evolutionary rate for cyt b of 1.28% per My (Weisrock et al. [2001]: see Yoshikawa et al. [2008] for discussion on this rate in hynobiids). For morphological comparisons of adults, the following 18 measurements and five counts were taken for specimens in preserved state (for the definition of characters, refer to Nishikawa et al. [2007]): 1) SVL (snout-vent length); 2) HL (head length); 3) HW (head width); 4) SL (snout length); 5) IND (internarial distance); 6) IOD (interorbital distance); 7) UEL (upper eyelid length); 8) AGD (axilla-groin distance); 9) TAL (tail length); 10) BTAW (basal tail width); 11) MTAW (medial tail width); 12) BTAH (basal tail height); 13) MTAH (medial tail height); 14) FLL (forelimb length); 15) HLL (hindlimb length); 16) 3TL (third toe length); 17) VTW (vomerine tooth series width); 18) VTL (vomerine tooth series length); 19) UJTN (number of upper jaw teeth); 20) LJTN (number of lower jaw teeth); 21) VTN (number of vomerine teeth); 22) CGN (number of costal grooves following Misawa [1989]); 23) LON (number of costal folds between adpressed limbs). For the holotypes, we also measured MXHW (maximum head width); LJL (lower jaw length); UEW (upper eyelid width); TRL (trunk length); MXTAH (maximum tail height); 2FL (second finger length); 3FL (third finger length); and 5TL (fifth toe length). For examining overall morphological variation among three target populations from Kyushu Island and H. stejnegeri, we conducted Canonical Discriminant Analysis (CDA) using loge-transformed metric values of a total of the 18 measurements. The number of presacral vertebrae including atlas (PSVN) was counted from a soft X-ray photograph using Fuji Medical X-Ray Film (RX-U). For larvae, we determined the developmental stage following the table proposed for congeneric species H. nigrescens (Iwasawa & Yamashita 1991). In this table, stages (St.) 50 to 62 (from early stage of digital differentiation until full-grown), 63 (full-grown), and 64 to 68 (from onset to its completion of metamorphosis), correspond to young, full-grown, and metamorphosing larvae, respectively. Student’s t-test and Tukey-Kramer test were used for morphometric comparisons, while Mann-Whitney’s Utest and Kruskal-Wallis test (or Dunn’s multiple comparison test) were performed for ratio values and detection of the presence or absence of differences in the frequency distributions. In comparing PSVN, we combined all sex and age groups. Significance level of 95% was used in all statistical tests. All statistical analyses were performed by SAS (1990). In the following description, average values are shown with 1 standard deviation (±SD). We used skeletochronology in order to determine ages of some adult salamanders. The methods were basically same as those used in Nishikawa & Matsui (2008).

Results As a result of molecular phylogenetic analyses, we obtained complete sequence of cyt b gene of mtDNA for all samples. Of 1141 nucleotide sites, 414 were variable, and 278 were parsimoniously informative. The best substitution models chosen for ML by Treefinder (Jobb 2008) were J2 (Jobb 2008)+G (1st and 3rd codons) and TN93 (Tamura & Nei 1993)+G (2nd), and those for BI by MrBayes v3.1.2 (Huelsenbeck & Ronquist 2001) were SYM (Zharkikh 1994)+G (1st) and HKY (Hasegawa et al. 1985)+G (2nd and 3rd). The likelihood values (-lnLs) of the ML and BI trees were 5027.198 and 5054.320, respectively. Phylogenetic analyses employing two different optimality criteria yielded nearly identical relationships, differing only in the position of H. nebulosus, and nearly matched a previous report (Weisrock et al. 2013, where positions of H. boulengeri and H. hirosei are mistakenly shown). As shown in the BI tree in Fig. 2, three populations of the Kyushu populations of H. boulengeri sensu lato, together with H. stejnegeri, proved to form a well-supported clade (MLBS=98, Bayesian posterior probability [BPP]=1.00), whereas H. boulengeri sensu stricto formed a clade with H. kimurae, and H. hirosei with H. katoi. In the Kyushu clade, H. stejnegeri was the sister group to the Amakusa population with a high support value in ML, although not in BI (MLBS=91, BPP70 % and Bayesian posterior probability (bpp)>95 %.

In the allozymic analyses by Nishikawa et al. (2005), the four populations also formed one group, which greatly differed from H. boulengeri sensu stricto. The group included three subgroups, 1) H. boulengeri from the Sobo-Katamuki Mountains, 2) that from Amakusa Islands and Osumi Peninsula, and 3) H. stejnegeri. Although their relationship was not resolved, they differed from each other at different species levels (Nishikawa et al. 2005, 2007). The second subgroup was further separated into two local populations, Amakusa Islands and Osumi Peninsula, by a nearly fixed allelic difference in PGM-C locus, although the genetic distance between them was small (Nei’s [1978] D=0.03: Nishikawa et al. 2005). The relationships among the four populations were different between mtDNA and allozyme analyses. The allozymic subgroup of Amakusa and Osumi was not recognized in mtDNA tree, in which the Amakusa population was first grouped with H. stejnegeri, then with the Osumi population, although the nodes were not supported (Fig. 2). These lines of information indicate that a total of four populations, the three Kyushu populations of H. boulengeri and H. stejnegeri, are genetically differentiated from each other.

206 · Zootaxa 3852 (2) © 2014 Magnolia Press

NISHIKAWA & MATSUI

In addition, no positive data, suggesting an admixture among the four populations, have been obtained. Hynobius stejnegeri occupies the widest range among them (ca. 70 km along longitude and 110 km along latitude), yet samples covering this wide range were genetically fairly uniform and well grouped by allozyme similarity (Nishikawa et al. 2005). The range of H. stejnegeri is surrounded by the three remaining, geographically isolated populations (Fig. 1) but is clearly separated by small to moderate geographic gaps (min. 10 km from SoboKatamuki, 30 km from Amakusa, and 90 km from Osumi). Furthermore, in congruence with this genetic separation, the male specimens of the four populations are also separated from each other in morphometric and meristic characters. The Amakusa population significantly differs from all samples of H. stejnegeri by RSL, RIND, RTAL, and PSVN, and from Gokase samples of that species by SVL, RAGD, RBTAW, RMTAW, RBTAH, RMTAH, RFLL, RHLL, R3TL, and LON, from the Osumi population by SVL, UJTN, LJTN, and VTN, from the Sobo-Katamuki population by SVL, RHL, RSL, RIND, RIOD, RTAL, RMTAW, RBTAH, RFLL, RHLL, R3TL, RVTW, VTW/VTL, CGN, LON, VTN, and PSVN. The Osumi population significantly differs from all samples of H. stejnegeri by SVL, RSL, RIND, and PSVN, and from Gokase samples of that species by RHW, RAGD, R3TL, and LJTN, and from the Sobo-Katamuki population by SVL, RSL, RIND, RTAL, RFLL, R3TL, RVTW, VTW/VTL, LON, UJTN, LJTN, and PSVN. The Sobo-Katamuki population significantly differs from Gokase samples of H. stejnegeri by RHL, RVTW, VTW/VTL, UJTN, and VTN. Although smaller in number than males, the female samples show tendencies similar to those found in males (Tables 2–5). In body coloration, the four populations are distinguishable from each other as well. Especially, H. stejnegeri is unique in having a dorsum with amber-like color, which is never seen in the three other populations. This characteristic color pattern is constantly found in its wide range of distribution. The multivariate analysis (CDA) showed a tendency of all the four populations to be separated in overall morphology (details in the systematic accounts). We thus retain current taxonomic status of H. stejnegeri, and describe each of the remaining three Kyushu populations formerly assigned to H. boulengeri as new taxon:

Systematics Hynobius shinichisatoi sp. nov. (Japanese name: Sobo-sanshouo) (Figs. 3, 4, and 5) Pachypalaminus boulengeri: Sato (1954): 195; Hynobius (Pachypalaminus) boulengeri: Nakamura and Uéno (1963): 13 (part); Hynobius boulengeri: Nishikawa et al. (2001): 281 (part); Sato (2001): 8; Nishikawa et al. (2003): 2 (part); Nishikawa et al. (2005): 54 (part); Nishikawa et al. (2007): 752 (part); Nishikawa and Matsui (2008): 10 (part); Nishikawa et al. (2008): 29.

Holotype: KUHE 24878, an adult male from Saeki-shi, Oita Prefecture, Kyushu, Japan (alt. 900 m a.s.l.), collected by K. Nishikawa on 18 March 1998. Paratypes: A total of 32 specimens all from Kyushu, Japan: KUHE 18920 and 18922 (one male and one female), 22889 (one male), 24967 (one male), 32394 (one male), and 32395 and 32396 (two males) from the type locality, collected on unknown date February 1996 by S. Sato, on 21 April 1995 by K. Nishikawa, on 9 April 1995 by S. Sato, on 29 April 1979 by S. Sato, and on 10 June 1979 by S. Sato, respectively; KUHE 26141 and 26142 (one male and one female), 27183–27185 (two males and one female), 28748–28757 (nine males and one female), and 35423 (one female) from Bungo-ono-shi, Oita Pref. (alt. 900 m a.s.l.), collected on 12 April 1999 by S. Sato, on 6 May 2000 by K. Nishikawa, on 5 May 2001 by K. Nishikawa and S. Sato, and on 1 April 2004 by K. Nishikawa, respectively; KUHE 21619 (one male), 22813 (one male), and 25096 (one male) from Takachiho-cho, Miyazaki Pref. (alt. 550 m a.s.l.), collected on 6 April 1996, 31 March 1997, and 23 April 1998 by K. Nishikawa, respectively; KUHE 27125 and 27126 (one male and one female) from Nobeoka-shi, Oita Pref. (alt. 800 m a.s.l.), collected on 3 May 2000 by K. Nishikawa; KUHE 27159 (one juvenile), 28200 (one juvenile), and 28484 (one juvenile) from the type locality, collected on 3 May 2000 by K. Nishikawa, 3 April 2000 by S. Sato, and 19 March 2001 by K. Nishikawa, respectively; IABHU J518 (one female) from Bungo-ono-shi, Oita Pref. (alt. 600 m a.s.l.), collected on 22 May 1953 by S. Sato.

THREE NEW SALAMANDERS FROM JAPAN

Zootaxa 3852 (2) © 2014 Magnolia Press ·

207

FIGURE. 3. Alive male holotype (KUHE 24878) of Hynobius shinichisatoi sp. nov.

Referred specimens: Three adult specimens: ST 1, 2, and 5 (two males and one female) from Bungo-ono-shi, Oita Pref. (alt. 600 m a.s.l.); 14 juvenile specimens: KUHE 24507, 26745–26752 from Takachiho-cho, Miyazaki Pref. (alt. 550 m a.s.l.), ST 7–10 from the type locality, and ST 6 from Bungo-ono-shi, Oita Pref. (alt. 600 m a.s.l.); 66 larval specimens: KUHE unnumbered from Takachiho-cho, Miyazaki Pref. (alt. 550 m a.s.l.). Etymology: The specific name is dedicated to Mr. Shin’ichi Sato of Oita Prefecture, who first found this species in the Sobo-Katamuki Mountains (Sato, 1954) and has studied its natural history for 60 years. Diagnosis: A comparatively large-sized species (adult SVL 73.6–92.4 mm in males and 70.8–95.8 mm in females) within the lotic breding Hynobius, breeding in montane streams; dorsum immaculate in adults; limbs and tail long; tips of fore- and hindlimbs adpressed on body scarcely meeting (overlap of -1.5 to 0.5 costal folds in males and -2.5 to 0 in females); fifth toe well developed; ova large, pigmentless; egg sacs long and string-like, without distinct whiptail structure on free end; most similar to H. hirosei, but with smaller and thinner body, and deeper vomerine tooth series; distinctly different from sympatric H. yatsui in having much larger and immaculate body. Description of holotype: Head-body large and slender; head oval and moderately depressed, distinctly longer than wide; snout rounded, slightly projecting beyond lower jaw; nostril close to snout tip; labial fold absent; eye large, prominently protruded, slightly inset from edge of head in dorsal view; upper eyelid well developed, shorter than snout; gular fold distinct, curving slightly anteriorly; parotoid gland evident, extending from angle of jaw to gular fold; postorbital grooves distinct, branching posterior to angle of jaw, one short and running down to lower jaw, the other long and posteriorly to parotoid gland; vomerine tooth series wider than long, V-shaped, anterior margin distal to choana; tongue broad, both sides free from mouth floor; fore- and hindlimbs long and thin; number of costal grooves between axilla and groin 13; depressed limbs separated by one costal fold; relative length of fingers I

Three new species of the salamander genus Hynobius (Amphibia, Urodela, Hynobiidae) from Kyushu, Japan.

Three new species of lotic breeding Hynobius, formerly assigned to H. boulengeri, are described from the Kyushu region, southwestern Japan. They diffe...
4MB Sizes 3 Downloads 5 Views