Zootaxa, a New Species of Newt of the Genus Paramesotriton
Total Page：16
File Type：pdf, Size：1020Kb
Zootaxa 2494: 45–58 (2010) ISSN 1175-5326 (print edition) www.mapress.com/zootaxa/ Article ZOOTAXA Copyright © 2010 · Magnolia Press ISSN 1175-5334 (online edition) A new species of newt of the genus Paramesotriton (Salamandridae) from southwestern Guangdong, China, with a new northern record of P. longliensis from western Hubei YUNKE WU1, 3, KE JIANG2 & JAMES HANKEN1 1Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts, 02138, USA 2College of Life Sciences, China West Normal University, Nanchong, Sichuan, 637002, China 3Corresponding author. E-mail: [email protected] Abstract We report two previously unknown populations of Asian warty newts (Salamandridae: Paramesotriton) in China. The first population, from southwestern Guangdong, is described as a new species, which is closely related to P. guangxiensis based on morphological and molecular data. The second new population, from western Hubei, is assigned to P. longliensis, which extends the known range of this species 400 km northwards. Limited genetic differentiation between P. longliensis and P. zhijinensis suggests that these two names may refer to the same (single) species. Key words: Amphibia; salamander; Mitochondrial DNA; phylogenetics; taxonomy Introduction The salamandrid genus Paramesotriton is popular in the international amphibian pet trade. Their peculiar warty skin and variable color pattern make these salamanders appealing to most herpetological hobbyists. Illegal field collections, however, may seriously threaten natural populations of some species of Paramesotriton, especially those with restricted ranges. Indeed, some scientists and conservationists advocate withholding the locality data for newly described species that are potentially valuable in commercial markets (Stuart et al. 2006). Ten species of Paramesotriton are currently recognized from southern China, northern Laos and northern Vietnam: P. caudopunctatus Liu and Hu; P. chinensis Gray; P. deloustali Bourret; P. ermizhaoi Wu, Rovito, Papenfuss and Hanken; P. fuzhongensis Wen; P. guangxiensis Huang, Tang and Tang; P. hongkongensis Myers and Leviton; P. laoensis Stuart and Papenfuss; P. longliensis Li, Tian, Gu and Xiong; and P. zhijinensis Li, Tian and Gu. Recent studies suggest that P. laoensis constitutes a different lineage from Paramesotriton (Weisrock et al. 2006; Zhang et al. 2008), and a new monotypic genus, Laotriton, is proposed for this species (Dubois & Raffaëlli 2009). On the basis of external morphology and osteological characters, Freytag (1983) considered P. caudopunctatus to be distinct from congeners and proposed a new genus, Allomesotriton, to contain only this species. Freytag’s proposed taxonomic change has not received wide acceptance from later workers, who instead recognize Allomesotriton as a subgenus or species group within Paramesotriton (Pang et al. 1992; Fei et al. 2006; Dubois & Raffaëlli 2009). The first species of Paramesotriton was described 150 years ago (Gray 1859), and an additional five species were described over the next 140 years. Four new species, however, have been reported in just the last ten years (Stuart & Papenfuss 2002; Li et al. 2008a, b; Wu et al. 2009). This recent surge suggests that species-level diversity within Paramesotriton may remain underestimated, especially in poorly surveyed areas of Southeast Asia. Here we present a molecular and morphological analysis of the taxonomic status of two previously unreported populations of Paramesotriton from China. We describe one population as a new species and identify the other population as a new record of the recently described P. longliensis. Accepted by M. Vences: 20 Apr. 2010; published: 4 Jun. 2010 45 Material and methods Nine specimens from the first new population were collected from streams on Yunwu Mountain in southwestern Guangdong. Seven specimens from the second population were caught by seine net from a river in Xianfeng County, western Hubei. Euthanized newts were fixed and preserved in 75% ethanol. Freshly cut liver samples from each animal were preserved in 100% ethanol for molecular analyses. We took the following linear measurements from preserved specimens with a digital caliper: TTL, total length; SVL, snout-vent length, measured from the tip of the snout to the posterior edge of the vent; TAL, tail length, measured from the posterior edge of the vent to the tail tip; TAD, maximum tail depth; HL, head length, measured from the tip of the snout to the posterior edge of the parotoid gland; HW, maximum head width; IO, interocular distance, measured from the anterior corner of each eye; EN, distance from the anterior corner of the right eye to the right nostril; IN, internostril distance; AG, distance between the axilla and the groin along the left side of the body; AL, average length of both forelimbs; PL, average length of both hind limbs. Osteological features were assessed by dissection and digital radiography at the Museum of Comparative Zoology, Harvard University (MCZ). Material for morphological comparisons was examined from the Chengdu Institute of Biology, Chinese Academy of Sciences (CIB), the Museum of Vertebrate Zoology, University of California at Berkeley (MVZ), and the Kunming Institute of Zoology, Chinese Academy of Sciences (KIZ). We also incorporated data from the original descriptions of previously named species. Phylogenetic relationships of the two new populations of Paramesotriton to congeneric species were derived from molecular genealogical reconstruction based on the mitochondrial ND2 fragment and flanking tRNAs. This work extended the analysis of Wu et al. (2009). Sequences were generated or retrieved from GenBank for all known species (Table 1) except P. laoensis, which exhibits closer affinity to the sister genus Pachytriton—our phylogenetic outgroup—than to other Paramesotriton on the basis of a mitogenomic analysis (Zhang et al. 2008). All taxa were sampled from or near their respective type localities (Fig. 1, Table 1). Genomic DNA extraction, PCR and sequencing methods were performed as in Wu et al. (2009); PCR primers were available from Wu et al. (2010). Mitochondrial sequences were aligned manually in Se-Al 2.0 (Rambaut 1995). No premature stop codons or indels were found in the protein-coding region. We inferred the matriarchal genealogical relationships within Paramesotriton using both maximum likelihood (ML) and Bayesian inference (BI) methods. The optimal evolutionary model was determined by the Akaike Information Criterion (AIC) implemented in MODELTEST 3.7 (Posada & Crandall 1998); the ML analysis was conducted in Garli v.0.951 (Zwickl 2006). The search for the ML tree was continued until the ML had not been improved by 0.01 for 500,000 generations. Bootstrap values were calculated for 100 replicates with the termination threshold reduced to 100,000 generations. Bayesian inference was executed in MrBayes 3.1.2 (Huelsenbeck & Ronquist 2001). Sequence data were partitioned according to tRNA and codon positions because a mitogenomic study on plethodontid salamanders suggests significant improvement of model likelihoods by defining more partitions (Muller et al. 2004). Independent GTR+I+G models were assigned to each data partition. Two independent runs with three heated chains and one cold chain were carried out for four million generations. Parameter convergence was assessed by the average standard deviation of split frequencies. The first two million generations were discarded as burn-in. To assess the effect of data partitioning on nodal support, we also conducted an unpartitioned analysis, which, with other conditions unchanged, applied one GTR+I+G model to the entire mtDNA sequence. Harmonic means of model likelihood were compared between the partitioned and unpartitioned strategy using Bayes Factor (BF). A BF > 10 indicates strong evidence favoring the better model (Kass & Raftery 1995). Additionally, we performed a maximum parsimony (MP) analysis in PAUP* 4.0b10 (Swofford 2002). A heuristic search was carried out for 1000 random-sequence-addition replicates; the MP bootstrap support was calculated by 1000 nonparametric pseudo-replicates. Interspecific divergences were assessed through uncorrected pairwise genetic distances (p-distance) in PAUP*; sequences retrieved from Lu et al. (2004) were excluded due to the large amount of missing data. 46 · Zootaxa 2494 © 2010 Magnolia Press WU ET AL. FIGURE 1. Sample localities. 1: Paramesotriton sp. (western Hubei); 2: P. caudopunctatus; 3: P. zhijinensis; 4: P. longliensis; 5: P. chinensis; 6: P. hongkongensis; 7: P. e r m iz ha oi; 8: P. fuzhongensis; 9: P. guangxiensis; 10: P. deloustali. Asterisk (11) denotes the type locality of P. yunwuensis sp. nov. The Yunwu Mountains are shaded orange. Inset: red square shows the focal area in Asia. Results and discussion Mitochondrial ND2 and flanking tRNA sequences yield an alignment for 33 specimens that contains 1216 nucleotide sites; 432 of these are variable and 356 are potentially parsimony-informative. Two single-site indels at the tRNA-Trp region are present in the sequences from Pachytriton (outgroup), and a single-site deletion is fixed at the tRNA-Ala region in the new population of Paramesotriton from southwestern Guangdong. TrN+G is the optimal evolutionary model for the ML analysis (alpha = 0.2876). Branch lengths from BI are unrealistically long (nearly ten times longer than those from ML), which results from the dependence of branch length estimation on branch length prior (Marshall 2010). Deleting sequences