Multi-locus phylogeny using topotype specimens sheds light on the systematics of Niviventer (Rodentia, Muridae) in China
- Bin Zhang†1, 2,
- Kai He†1, 3, 9Email author,
- Tao Wan1,
- Peng Chen4,
- Guozheng Sun5,
- Shaoying Liu6,
- Truong Son Nguyen7,
- Liangkong Lin8 and
- Xuelong Jiang1Email author
© The Author(s). 2016
Received: 8 June 2016
Accepted: 17 November 2016
Published: 1 December 2016
Niviventer is a genus of white-bellied rats that are among the most common rodents in the Indo-Sundaic region. The taxonomy of the genus has undergone extensive revisions and remains controversial. The current phylogeny is unresolved and was developed primarily on the basis of mitochondrial genes. Identification is extremely difficult, and a large number of GenBank sequences seem to be problematic. We extensively sampled specimens of Niviventer in China and neighboring northern Vietnam, including topotypes of the most reported species (n = 6), subspecies (n = 8), and synonyms (n = 4). We estimated phylogenetic relationships on the basis of one mitochondrial and three nuclear genes, using concatenation and coalescent-based approaches. We also employed molecular species delimitation approaches to test the existence of cryptic and putative new species.
Our phylogeny was finely resolved, especially for the N. confucianus-like species. Our data provided the first support for N. brahma and N. eha as sister species, an assignment that is congruent with their morphological similarities. Species delimitation analyses provided new insight into species diversity and systematics. Three geographic populations of N. confucianus and one of N. fulvescens were supported as genetically distinct in our species delimitation analyses, while three recognized species (N. coninga, N. huang, and N. lotipes) were not strongly supported as distinct.
Our results suggested that several genetically distinct species may be contained within the species currently known as N. confucianus and N. fulvescens. In addition, the results of Bayesian Phylogenetics and Phylogeography (BPP) for N. coninga, N. huang, and N. lotipes indicated that either inter-specific gene flow had occurred or imperfect taxonomy was present. Morphological examinations and morphometric analyses are warranted to examine the molecular results.
Taxonomy and systematics are crucial for understanding biological diversity and the conservation of species. Nonetheless, a complete understanding of species diversity remains a distant goal for most groups of organisms. A recent review of newly discovered mammal species showed that saturation of species discovery in Mammalia has not been reached. The newly added species include not only new described species, but also revisionary new species, which are synonyms or subspecies that have been recognized as valid species . The latest version of Mammal Species of the World (2005) recognized 5,339 valid species, as well as 6,351 subspecies and 15,881 species-level synonyms . A large proportion of synonyms and subspecies are assigned those ranks because of the lack of specimens, comprehensive analyses, and critical evaluation. This is especially the case for taxa that are widely distributed and morphologically unremarkable, such as the white-bellied rats of the genus Niviventer. These murine rodents are small to medium-sized and are mainly distributed in China, southern Himalaya, and Southeast Asia in various habitats over a wide range of elevations [3, 4]. While most species of Niviventer have very limited distributions, a few species, such as N. cremoriventer, N. confucianus, and N. fulvescens, are widely spread across large geographic areas and various landscapes [5–7].
The taxonomy of Niviventer has a long and complicated history. Representatives of Niviventer were originally included in the genus Mus (1836–1911), after which they were placed in Epimys (1911–1916), and Rattus (1916–1981; Rattus = Epimys, Hollister ), successively. The genus was established by Musser , who included 15 species and divided the genus into the N. andersoni-division and N. niviventer-division. According to the taxonomy presented by Musser and Carleton in 2005 , 17 taxa had full species status, while another 65 were recognized as subspecies or synonyms. The widespread N. cremoriventer, N. confucianus, and N. fulvescens had 10, 14, and 26 subspecies and synonyms, respectively. Recent studies integrating karyotypic information, DNA sequences, and morphological characters recognized N. lotipes (a synonym of N. tenaster) , N. huang, and N. bukit (synonyms of N. fulvescens) as full species  but placed N. langbianis into Chiromyscus . Molecular phylogenetic studies further found that the widely distributed species were paraphyletic or even polyphyletic (e.g., N. andersoni, N. confucianus, N. excelsior, and N. fulvescens) . Despite the possibility of misidentification, the number of putative species always exceeded that of recognized species [13, 14], implying that species diversity was underestimated. Thus, it was reasonable to re-evaluate the large body of remaining subspecies/synonyms within this genus.
The complex distribution pattern, large number of subspecies/synonyms, and morphological similarities make taxonomic and systematic revision for the genus Niviventer extremely difficult. Many species are morphologically very similar to one another . For example, geometric morphometric analyses in a previous study showed that two distinctive genetic lineages largely overlapped (i.e., N. confucianus and N. fulvescens; Fig. 6 in Lu et al. ). Examination of type specimens is always crucial for taxonomic revision and essential for re-evaluating the synonyms. However, the type series for most species distributed in Asia are deposited in natural history museums in Europe and the United States. Thus, comprehensive genetic analyses using topotypes may uncover cryptic species diversity and help to determine key taxa that need to be diagnosed before the type series is examined. In addition to the existing taxonomic problems, the phylogeny of Niviventer remained largely obscure. Although several major clades and species groups have been recognized, the relationships among them were unresolved, and the phylogenetic positions of N. coninga and N. culturatus were contradictory [11, 15]. More than 540 Niviventer specimens have been sequenced, but most of the previous studies were conducted using only mitochondrial cytochrome b (CYT B) and cytochrome c oxidase (COI) genes [11, 13, 15, 16]. In another previous study using both mitochondrial (CYT B, COI, and D-loop) and nuclear (IRBP) loci, the gene trees were conflicting, suggesting that inclusion of multiple nuclear genes is warranted .
According to the taxonomy of Musser and Carleton  and recent studies [10–12], there are 10 recognized Niviventer species distributed in China, accounting for 55% of Niviventer species, including 5 endemic to China (N. andersoni, N. coninga, N. culturatus, N. excelsior, and N. lotipes) and 5 both inside and outside China (N. brahma, N. eha, N. confucianus, N. fulvescens, and N. huang). Nine Niviventer species (N. cremoriventer, N. cameroni, N. rapit, N. fraternus, N. lepturus, N. hinpoon, N. tenaster, N. bukit, and N. niviventer) are distributed in Southeast Asia outside China. Another 24 taxa are recognized as synonyms or subspecies of N. andersoni (n = 3), N. confucianus (n = 14), N. eha (n = 1), N. excelsior (n = 1), N. fulvescens (n = 1), and N. huang (n = 4). Moreover, according to Lu et al. , undescribed species may exist in the mountains of southwestern China, an area characterized by extremely complex topography .
In the present study, we extensively sampled white-bellied rats throughout China and neighboring northern Vietnam, capturing topotypes of 18 species, subspecies, and synonyms. We sequenced four unlinked loci, reconstructed phylogenetic relationships, and delimited species boundaries. The goals were i) to clarify the status of species, subspecies, and synonyms, and ii) to uncover underestimated species diversity.
Topotypes (sample locality number)
N. andersoni (Thomas, 1911)
Mt. Emei, Sichuan, China
N. andersoni pianmaensis Li & Yang, 2009
Pianma, Yunnan, China
N. andersoni ailaoshansis Li & Yang, 2009
Mt. Ailao, Yunnan, China
Y (13, 18, 21)
N. andersoni lushuiensis Wu & Wang, 2002
N. brahma (Thomas, 1914)
Mishmi Hills, Arunachal Pradesh
N. confucianus (Milne-Edwards, 1871)
Baoxing, Sichuan, China
Rattus confucianus littoreus Cabrera, 1922
Fuzhou, Fujian, China
Epimys zappeyi Allen, 1912
Mt. Wa, Sichuan, China
Rattus confucianus yaoshanensis Shih, 1930
Luoxiang, Guangxi, China
Rattus confucianus sinianus Shih, 1931
Lechang, Guangdong, China
Rattus elegans Shih, 1931
Lechang, Guangdong, China
N. confucianus sacer (Thomas, 1908)
Yantai, Shandong, China
Mus confucianus luticolor Thomas,1908
Yan’an, Shaanxi, China
Epimys confucianus canorus Thomas, 1922
Wenxian, Gansu, China
N. confucianus chihliensis (Thomas, 1917)
Eastern Tombs, Hebei, China
N. confucianus mentosus (Thomas, 1916)
N. confucianus yushuensis (Wang & Zheng, 1981)
Yushu, Qinghai, China
N. confucianus naoniuensis (Zhang & Zhao, 1984)
Mt. Naoniu, Baicheng, Jilin, China
N. confucianus yajiangensis, Deng & Wang, 2000
Bajiaolou, Yajiang, Sichuan, China
N. confucianus deqinensis, Deng & Wang, 2000
Adong, Deqin, Yunnan, China
N. coninga (Swinhoe, 1864)
Y (11, 19)
N. culturatus (Thomas, 1917)
Y (15, 19)
N. eha (Wroughton, 1916)
Lachen, Sikkim, India
N. eha ninus (Thomas, 1922)
Salween-Mekong divide, Yunnan, China
N. excelsior (Thomas, 1911)
Kangding, Sichuan, China
N. excelsior tengchongensis, Deng & Wang, 2002
N. fulvescens (Gray, 1847)
Rattus huang vulpicolor Allen, 1926
Nating River, Yunnan, China
N. huang (Bonhote, 1905)
Guadun, Fujian, China
Mus ling Bonhote, 1905
Chungfengling, Fujian, China
Rattus flavipilis Shih, 1930
Luoxiang, Guangxi, China
Rattus flavipilis minor Shih, 1930
Luoxiang, Guangxi, China
Rattus wongi Shih, 1931
Lechang, Guangdong, China
N. lotipes (Allen, 1926)
Nada, Hainan, China
Mitochondrial tree and species recovery
Sample sizes and species used for each analysis
Sample size (from China)
Species sampling (from China)
Data source(s) (current study + GenBank)
Collected for this study
Download from GenBank
ML and ABGD
CYT B data set
ML and ABGD
157 + 517
Concatenated data set
ML and MrBayes
157 + 47
Full data set
BEAST, *BEAST, and BPP
147 + 0
Results of Automatic Barcode Gap Discovery (ABGD) analyses with K80 distance model
X (Gap width)
p = 0.0056
p = 0.0121
Multilocus gene trees
Species trees and species delimitation
Results of Bayesian phylogenetics and phylogeography (BPP) analyses using coalescent species trees as guard trees
8, 5 + 9
29, 23 + 30
22, 23+ 29 + 30
18 + 19, 22+ 29 + 23 + 30
21, 18 + 19 + 22 + 29 + 23 + 30
MtDNA and nuclear Algorithm 0
MtDNA and nuclearAlgorithm 1
nuclear Algorithm 0
nuclear Algorithm 1
When nuclear genes alone were used, BPP supported N. coninga (group 21), N. lotipes (group 23) and several populations of N. confucianus as one species (Fig. 5; Table 4; Additional file 5: Table S5). All analyses supported group 24 from western Yunnan and southeastern Tibet (identified as N. confucianus mentosus); group 26 from Gongbu, Tibet (identified as N. confucianus; putative sp. 4); and group 27 from southern China and northern Vietnam (identified as N. confucianus; putative sp. 5) as three putative species (Fig. 5; Table 4; Additional file 5: Table S5).
Correct naming of species
According to the famous Chinese teacher Confucius, “If names be not correct, language is not in accordance with the truth of things. If language be not in accordance with the truth of things, affairs cannot be carried on to success” . The species N. confucianus was named after Confucius , and his quote regarding names is indeed relevant to the taxonomy of the genus Niviventer. Species diversity within this genus has been underestimated [13, 14], and our understanding of its taxonomy was incomplete . These two issues have prevented a full understanding of the systematics and evolution of the genus. Multi-locus sequencing for topotype, conducted in this study, indicates the need for several taxonomic and systematic revisions.
Implications of the molecular species delimitation analyses
Based on the barcoding gap found from the pairwise distances of CYT B sequences, the single gene species delimitation method, ABGD, supported the existence of 22 groups/putative species in China (represented by 589 sequences from China), in agreement with previous reports by He and Jiang  and Lu et al. . Although far fewer samples (n = 147) were available for multilocus species delimitation, BPP analyses still recognized 15 (using nuclear plus mitochondrial genes) or 11 (using three nuclear genes alone) putative species.
BPP molecular species delimitation approaches have been developed and widely used in mammals [24, 25], amphibians [26, 27], and reptiles [28, 29]. It should be noted that the biological species concept associated with BPP may not be true in practice , and that simulations showed that BPP is robust in the case of incomplete lineage sorting but sensitive to hybridization. Strongly supported putative species could be either valid species or long-term isolated conspecific populations. On the other hand, when named species were not strongly supported (PP < 0.95) as putative species by BPP, the results either suggested genetic homogenization (i.e., conspecific relationships) or gene flow between valid species. In both cases, the taxonomic status must still be evaluated carefully in a taxonomic context and must not violate the International Code of Zoological Nomenclature . The BPP analysis that used nuclear genes alone did not strongly support N. huang, N. lotipes, or N. coninga as putative species. On the other hand, four putative new species (N. confucianus mentosus and Niviventer sp. 3–5) need further attention. Comprehensive morphological and morphometric examinations are warranted to test species status for all these taxa.
Value of topotypical samples
Species in the genus Niviventer are morphologically very similar. The current taxonomy roughly relies on the size of skull, the length of ear and tail, and the color of skin and feet . Morphological identification is not only time-consuming but also challenging. All but one of the topotypes of Niviventer sequenced in the current study were presented for the first time. We carefully assigned each individual to recognized species on the basis of morphology (Suppl. File 2) and the molecular results. These voucher specimens and sequences are valuable for barcoding-based species identification and taxonomic revision. As shown in our CYT B gene tree, GenBank sequences under the same species names were found in different groups or major clades (Fig. 1), indicating misidentification.
Implications for taxonomy
The analyses of topotype sequences of named synonyms and subspecies revealed multiple putative species within the currently recognized species of N. confucianus, N. fulvescens, N. andersoni, and N. eha. Within N. confucianus, the subspecies N. confucianus mentosus should be examined more closely, as it was supported as a distinct group/putative species in all analyses. This subspecies was named and described as a species by Thomas  on the basis of the holotype from Hkampti, Myanmar, after which it was assigned as a subspecies of N. niviventer  or a synonym of N. confucianus . Animals from Chayu (Tibet Autonomous Region), and Nujiang (Yunnan Province) were assigned to N. c. mentosus on the basis of their longer hind feet (≥30 mm) and white-tipped tail . Niviventer sp. 4 collected from Guangxi and Yunnan were marked by a white tail tip and conspicuous brush, a typical pattern of N. confucianus that distinguishes it from N. c. yaoshanensis, N. niviventer, or N. fulvescens. Niviventer sp. 3 is similar to N. fulvescens in body size and color pattern but could be distinguished from N. fulvescens and N. cremoriventer by its fuzzy bicolor tail.
Our analyses also suggest that diversity in some groups may be overestimated. Niviventer lotipes (group 23), recognized by Li et al.  for its distinct karyotype, was not supported as a putative species by any of our BPP analyses. It was originally described as a subspecies of N. confucianus  on the basis of the holotype from Nada, Hainan, then treated as a subspecies of N. niviventer [33, 36] or a synonym of N. tenaster . We did not included any topotype of N. lotipes (Danzhou, Hainan), but all specimens we included from Hainan, Fujian and Guangxi could be affiliated with the holotype (AMNH M-59303) of N. lotipes based on the sharp bicolor appearance of the tail, and are different from specimens used in Li et al. , showing a sharp bicolor appearance. However, individuals from other localities had a white-tipped tail. Another recognized species, N. coninga (group 21), was not supported by the BPP analyses using nuclear genes. Furthermore, most recognized subspecies of N. confucianus were clustered into two lineages or groups (subspecies N. c. yushuensis, N. c. yajiangensis, and N. c. deqinensis in group 29, and the nominated subspecies N. c. confucianus, N. c. chihliensis, and N. c. naoniuensis in group 30). They were not supported as putative species by BPP analyses. Therefore, the taxonomic status of all these clades must be re-examined.
In this study, we collected extensive samples of Niviventer in China, including most topotypes of species, subspecies, and synonyms. We developed a robust phylogeny on the basis of one mitochondrial and three nuclear genes. Four major clades were recovered by both concatenation and coalescent-based approaches. The sister relationship between N. brahma and N. eha was well supported, and relationships among N. confucianus-like species were mostly well resolved. Our coalescent-based species delimitation analyses indicated that species diversity of Niviventer may be overestimated by single locus in previous studies, and three species should be re-evaluated using morphological or morphometric approaches. However, our analyses also indicated that cryptic or putative new species may exist in the mountains of southwest China, an area characterized by extremely complex topography, climate conditions, and geographic history.
Collection of samples and sequences
We collected 157 individuals of Niviventer, as well as five Chiromyscus langbianis (previously included in Niviventer), from China and neighboring northern Vietnam (Additional file 1: Table S1). These samples were collected at 67 locations (Fig. 1), including the type localities of 6 species, 8 subspecies, and 4 synonyms (Table 1). Voucher specimens from mainland China were deposited in the Kunming Institute of Zoology, Chinese Academy of Sciences, and Sichuan Academy of Forestry, China, and voucher specimens from Taiwan were deposited in Tunghai University. Specimens were carefully identified on the basis of external and skull characters following the original descriptions (Additional file 6: Table S6, [3, 5, 6]) and in consideration of the results of molecular phylogenetic analyses (see Results section). We followed the taxonomy of Musser and Carleton  but recognized N. huang, N. bukit, and N. lotipes as valid species [10, 11], and treated N. langbianis as a member of Chiromyscus following Balakirev et al. . After completing our mitochondrial phylogeny, we carefully examined and described the morphological characters of each clade (Additional file 7: Table S7).
Primers and polymerase chain reaction (PCR) cycling conditions used in this study
Nucleotide sequence 5′ to 3′
Annealing Temperature (°C)
Fragment Length (bp)
He et al. 2010 
Stanhope et al. 1992 
Steppan et al. 2004 
Three data sets were used for phylogenetic analyses: 1) the CYT B data set, which included 674 in-group sequences (157 sequences from the current study and 517 sequences downloaded from GenBank; Table 2), representing 16 of 19 recognized Niviventer species and two potential new species identified in a previous study (namely, Niviventer sp. 1 and sp. 2) , with Rattus andamanensis and Chiromyscus langbianis as the outgroup. The first 60 bp and the last 63 bp of CYT B were deleted from the data set to avoid sequencing errors . 2) A four-gene concatenated data set, including sequences of 197 individuals of Niviventer plus outgroups (157 sequences from the current study and 40 sequences downloaded from NCBI GenBank; Tables 2). One to four sequences downloaded from GenBank were selected as representative of each group based on the CYT B gene tree. 3) The full data set, including 147 sequences from the current study plus 8 individuals representing six outgroup species (C. langbianis, L. edwardsi, L. neilli, R. andamanensis, R. norvegicus, and M. musculus), all of which have four gene segments (CYT B, GHR, IRBP, and RAG1).
PartitionFinder v. 1.0.0  was used to estimate the best-fit partitioning schemes and substitution evolutionary models under the Bayesian Information Criterion [42, 43]. For concatenated gene tree/species tree analyses, we concatenated gene alignments and defined by genes and codon positions. We did not consider the parameter for the proportion of invariant sites suggested by the author of RAxML . Thus, we considered 12 models for each MrBayes and BEAST. For RAxML analyses, only GTR + gamma models were considered because RAxML did not allow any other models. For coalescent species tree estimation (see below), each of the three nuclear genes was treated as one partition and the CYT B was partitioned by codon positions. jModelTest 2 was used to estimate the best-fit evolutionary model for each nuclear gene .
The phylogenetic relationships of Niviventer were reconstructed using Bayesian inference and ML approaches. ML analyses and Bayesian analyses were conducted for the CYT B data set and the full data set, respectively, and both ML and Bayesian analyses were used for the concatenated data set. The ML analyses were performed using RAxML v7.2.8 on the CIPRES Science Gateway v 3.3  (http://www.phylo.org). The BS were obtained using a rapid bootstrapping algorithm with 500 bootstrap replicates . Bayesian analyses of the concatenated data set were implemented in MrBayes v. 3.2 and repeated twice . Each analysis consisted of two independent runs, using four chains, sampled every 1,000 generations. We used Tracer v. 1.6  to access the convergence of the two independent runs and stationary state of each parameter, and ESSs higher than 200 were considered adequate. Finally, each analysis was run for 2 million generations, and the first 25% of the generations were discarded as burn-in. Bayesian analyses of the full data set were implemented in BEAST v. 1.8.2 . Each BEAST analysis used a lognormal relaxed clock model, a birth-death tree prior  and the other default parameters. The analyses were repeated twice; each was run for 200 million generations, and we sampled every 20,000 interactions. We considered BS ≥ 70 and PP ≥ 0.95 as strong support [50, 51].
Species delimitation and species tree estimation
We used ABGD  to recover candidate species, and species tree and Bayesian Phylogenetics and Phylogeography (BPP v. 3.1; ) for validation. The ABGD approach assumed that the largest intraspecific divergence is still less than the smallest interspecies divergence. The gap between intra- and interspecific distance observed from the distribution of sequence distances is defined as a barcoding gap. ABGD automatically clusters the sequences into groups or hypothetical species based on the barcoding gap. Thus, detection of pairwise differences is the premise of this method. The only input parameter, although imprecisely, was the maximum prior intraspecific distance, p. We chose ABGD instead of bPTP  or GMYC  because this method was independent of phylogenetic trees and particularly sensitive to recent speciation events . This analysis was performed on the ABGD website (http://wwwabi.snv.jussieu.fr/public/abgd/abgdweb.html) and used the 674 CYT B sequences as input data (Table 2). We selected the K80 model and set the transition/transversion ratio at 4.73, as estimated using MEGA 7 . We set both the number of steps and bins to 25 and ran the program with different gap widths as 0.5, 1.0, 1.5, 2.0 and 2.5. Genetic distances of CYT B between candidate species were calculated using MEGA 7  under the K2P model.
We inferred the species tree using the *BEAST model  implemented in BEAST v. 1.8.2. Species with at least one sequence per gene were included (as required by the model). We assigned the full data set of 147 individuals to 18 candidate species groups according to the results of ABGD species delimitation analyses (Additional file 1: Table S1). We run the analyses in BEAST v. 1.8.2  for 400 million generations with the same priors as in the BEAST analyses.
To validate our ABGD proposal, a multilocus coalescent species delimitation method, BPP, was performed. Based on the phylogeny, ancestral population size, and root age, BPP employed the Reversible-Jump Markov Chain Monte Carlo (rjMCMC) method to generate the posterior probability that the two groups form a single species. Two data sets were used in these analyses: i) mitochondrial and nuclear genes and ii) three nuclear genes alone. Individuals represented only by mitochondrial sequences (sequences downloaded from GenBank) were not included in our BPP analyses. Using the results of AGBD, we assigned 147 individuals to 18 candidate species (Additional file 1: Table S1). We ran BPP for each of the four major clades recovered in our phylogenetic analyses separately (see Results), and used the *BEAST coalescent species tree as the guide tree. We performed the analyses using both rjMCMC algorithms 0 and 1. Analyses were run for one million generations with a pre-burn-in of 10,000 generations and a sampling interval of 100. We repeated BPP analyses 12 times with different combinations of priors and parameters following Wan et al. .
We thank Mr. Changzhe Pu for collecting specimens.
This research was funded by the National Natural Science Foundation of China (No. 31272276) and the Fund of State Key Laboratory of Genetic Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences (GREKF13-07).
Availability of data and material
All newly conducted DNA sequences were submitted to GenBank with accession numbers KY068361–KY069019. Accession numbers of each gene locus for all specimens are also provided in Additional file 1: Table S1 and Additional file 2: Table S2.
BZ, KH, and TW contributed to the laboratory work, data analyses, and manuscript writing. PC, GZS, SYL, TSN, and LKL conducted field work and collected specimens. XLJ conceived the study and commented on the manuscript. All authors have read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
The specimen collections were approved by the internal review board of Kunming Institute of Zoology, Chinese Academy of Sciences (approval ID: SYDW-20160606001).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Reeder DAM, Helgen KM, Wilson DE. Global trends and biases in new mammal species discoveries. Occasional papers The Museum of Texas Tech University. 2007;269:1–35.Google Scholar
- Wilson DE, Reeder DM. Mammal species of the world. A taxonomic and geographic reference (3rd ed). Baltimore: Johns Hopkins University Press; 2005.Google Scholar
- Musser GG. Notes on systematics of Indo-Malayan murid rodents, and descriptions of new genera and species from Ceylon, Sulawesi, and the Philippines. Bull Am Mus Nat Hist. 1981;168:229–334.Google Scholar
- Richharia M, Westbrook LD. Walker's mammals of the world. Baltimore: Johns Hopkins University Press; 1983.Google Scholar
- Allen GM. The mammals of China and Mongolia, vol. 6. New York: The American Museum of Natural History; 1940.Google Scholar
- Marshall JT. Family Muridae Rats and Mice. In: Lekagul B, McNeely BA, editors. Mammals of Thailand. Thailand: Association for the Conservation of Wildlife; 1977. p. 397–485.Google Scholar
- Musser GG. Species-limits of Rattus cremoriventer and Rattus langbianis, Murid Rodents of Southeast Asia and the Greater Sunda Islands. Am Mus Novit. 1973;2525:1–65.Google Scholar
- Hollister N. The generic names Epimys and Rattus. Proc Biol Soc Washington. 1916;29:126.Google Scholar
- Musser GG, Carleton MD. Superfamily Muroidea. In: Wilson DE, Reeder DM, editors. Mammal species of the world: a taxonomic and geographic reference (3rd ed), vol. 2. Baltimore: Johns Hopkins University Press; 2005. p. 894–1531.Google Scholar
- Li Y, Wu Y, Harada M, Lin LK, Motokawa M. Karyotypes of three rat species (Mammalia: Rodentia: Muridae) from Hainan Island, China, and the valid specific status of Niviventer lotipes. Zoolog Sci. 2008;25:686–92.View ArticlePubMedGoogle Scholar
- Balakirev AE, Abramov AV, Rozhnov VV. Taxonomic revision of Niviventer (Rodentia, Muridae) from Vietnam: A morphological and molecular approach. Russian J Theriol. 2011;10:1–26.Google Scholar
- Balakirev AE, Abramov AV, Rozhnov VV. Phylogenetic relationships in the Niviventer-Chiromyscus complex (Rodentia, Muridae) inferred from molecular data, with description of a new species. Zookeys. 2014;451:109–36.View ArticleGoogle Scholar
- He K, Jiang XL. Mitochondrial phylogeny reveals cryptic genetic diversity in the genus Niviventer (Rodentia, Muroidea). Mitochondrial DNA. 2015;26(1):48–55.Google Scholar
- Lu L, Ge D, Chesters D, Ho SYW, Ma Y, Li G, Wen Z, Wu Y, Wang J, Xia L, et al. Molecular phylogeny and the underestimated species diversity of the endemic white-bellied rat (Rodentia: Muridae:Niviventer) in Southeast Asia and China. Zool Scr. 2015;44:475–94.View ArticleGoogle Scholar
- Jing M, Yu H-T, Wu S-H, Wang W, Zheng X. Phylogenetic relationships in genus Niviventer (Rodentia: Muridae) in China inferred from complete mitochondrial cytochrome b gene. Mol Phylogenet Evol. 2007;44:521–9.View ArticlePubMedGoogle Scholar
- Balakirev AE, Rozhnov VV. Phylogenic relationships and species composition in the genus Niviventer (Rodentia, Muridae) based on studies of the cytochrome b gene of mtDNA. Mosc Univ Biol Sci Bull. 2010;65:170–3.View ArticleGoogle Scholar
- He K, Jiang XL. Sky islands of southwest China. I: an overview of phylogeographic patterns. Chin Sci Bull. 2014;59:585–97.View ArticleGoogle Scholar
- Pagés M, Chaval Y, Herbreteau V, Waengsothorn S, Cosson J-F, Hugot J-P, Morand S, Michaux J. Revisiting the taxonomy of the Rattini tribe: A phylogeny-based delimitation of species boundaries. BMC Evol Biol. 2010. doi:10.1186/1471-2148-10-184.
- Lanfear R, Calcott B, Ho SYW, Guindon S. PartitionFinder: combined selection of partitioning schemes and substitution models for phylogenetic analyses. Mol Biol Evol. 2012;29:1695–701.View ArticlePubMedGoogle Scholar
- Kumar S, Stecher G, Tamura K. MEGA7: Molecular evolutionary genetics analysis version 7.0 for bigger datasets. Mol Biol Evol. 2016;33:1870–4.View ArticlePubMedGoogle Scholar
- Puillandre N, Lambert A, Brouillet S, Achaz G. ABGD, Automatic Barcode Gap Discovery for primary species delimitation. Mol Ecol. 2012;21:1864–77.View ArticlePubMedGoogle Scholar
- Confucius, Legge, J. Confucian analects, the Great learning, the Doctrine of the mean, and the Works of Mencius with English translation and notes. Montana: Kessinger Publishing; 2006.Google Scholar
- Milne-Edwards H. Nouvelles Archives du Museum D'Histoire Naturelle De Paris. 1871.Google Scholar
- He K, Shinohara A, Jiang X-L, Campbell KL. Multilocus phylogeny of talpine moles (Talpini, Talpidae, Eulipotyphla) and its implications for systematics. Mol Phylogenet Evol. 2014;70:513–21.View ArticlePubMedGoogle Scholar
- Wan T, He K, Jiang X-L. Multilocus phylogeny and cryptic diversity in Asian shrew-like moles (Uropsilus, Talpidae): implications for taxonomy and conservation. BMC Evol Biol. 2013;13:232.Google Scholar
- Salerno PE, Senaris JC, Rojas-Runjaic FJM, Cannatella DC. Recent evolutionary history of Lost World endemics: population genetics, species delimitation, and phylogeography of sky-island treefrogs. Mol Phylogenet Evol. 2015;82:314–23.View ArticlePubMedGoogle Scholar
- Zhou W-W, Wen Y, Fu J, Xu Y-B, Jin J-Q, Ding L, Min M-S, Che J, Zhang Y-P. Speciation in the Rana chensinensis species complex and its relationship to the uplift of the Qinghai-Tibetan Plateau. Mol Ecol. 2012;21:960–73.Google Scholar
- Leache AD, Fujita MK. Bayesian species delimitation in West African forest geckos (Hemidactylus fasciatus). Proc Biol Sci. 2010;277:3071–7.Google Scholar
- Brown RM, Siler CD, Grismer LL, Das I, McGuire JA. Phylogeny and cryptic diversification in Southeast Asian flying geckos. Mol Phylogenet Evol. 2012;65:351–61.View ArticlePubMedGoogle Scholar
- Rannala B. The art and science of species delimitation. Curr Zool. 2015;61:846–53.View ArticleGoogle Scholar
- ICZN. International Code of Zoological Nomenclature (4th ed). London: International Trust for Zoological Nomenclature; 1999.Google Scholar
- Thomas O. On the Rat known as Epimys jerdoni from Upper Burma. J Bomb Nat Hist Soc. 1916;24:643–4.Google Scholar
- Ellerman JR, Morrison-Scott TCS. Checklist of Palaearctic and Indian mammals, 1758 to 1946. London: British Museum (Natural History); 1951.Google Scholar
- Feng ZJ, Cai GQ, Zheng CL. The mammals of Xizang. Beijing: Science Press; 1986.Google Scholar
- Allen GM. Rats (genus Rattus) from the Asiatic Expeditions. Am Mus Novit. 1926;217:1–16.Google Scholar
- Ellerman JR, Hayman RW, Holt GWC, British M. The families and genera of living rodents, vol. 3. London: British museum (Natural History); 1949.Google Scholar
- Sambrook J, Russell D. Molecular cloning: A laboratory manual. New York: Cold Spring Harbor Laboratory Press; 2001.Google Scholar
- He K, Li Y, Brandley M, Lin L, Wang Y, Zhang Y, Jiang X. A multi-locus phylogeny of Nectogalini shrews and influences of the paleoclimate on speciation and evolution. Mol Phylogenet Evol. 2010;56:734–46.View ArticlePubMedGoogle Scholar
- Stanhope MJ, Czelusniak J, Si J-S, Nickerson J, Goodman M. A molecular perspective on mammalian evolution from the gene encoding interphotoreceptor retinoid binding protein, with convincing evidence for bat monophyly. Mol Phylogenet Evol. 1992;1:148–60.View ArticlePubMedGoogle Scholar
- Steppan SJ, Storz BL, Hoffmann RS. Nuclear DNA phylogeny of the squirrels (Mammalia: Rodentia) and the evolution of arboreality from c-myc and RAG1. Mol Phylogenet Evol. 2004;30:703–19.View ArticlePubMedGoogle Scholar
- Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004;32:1792–7.View ArticlePubMedPubMed CentralGoogle Scholar
- Luo A, Qiao H, Zhang Y, Shi W, Ho SY, Xu W, Zhang A, Zhu C. Performance of criteria for selecting evolutionary models in phylogenetics: A comprehensive study based on simulated datasets. BMC Evol Biol. 2010. doi:10.1186/1471-2148-10-242.
- Fang X, Nevo E, Han L, Levanon EY, Zhao J, Avivi A, Larkin D, Jiang X, Feranchuk S, Zhu Y, et al. Genome-wide adaptive complexes to underground stresses in blind mole rats Spalax. Nat Commun. 2014. doi:10.1038/ncomms4966.
- Stamatakis A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics. 2014;30:1312–3.View ArticlePubMedPubMed CentralGoogle Scholar
- Darriba D, Taboada GL, Doallo R, Posada D. jModelTest 2: more models, new heuristics and parallel computing. Nat Methods. 2012;9:772.View ArticlePubMedPubMed CentralGoogle Scholar
- Miller MA, Pfeiffer W, Schwartz T. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. In: Gateway Computing Environments Workshop (GCE). IEEE. 2010;14:1–8.Google Scholar
- Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, Höhna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP. MrBayes 3.2: efficient bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012;61:539–42.View ArticlePubMedPubMed CentralGoogle Scholar
- Drummond AJ, Suchard MA, Xie D, Rambaut A. Bayesian phylogenetics with BEAUti and the BEAST 1.7. Mol Biol Evol. 2012;29:1969–73.View ArticlePubMedPubMed CentralGoogle Scholar
- Gernhard T. The conditioned reconstructed process. J Theor Biol. 2008;253:769–78.View ArticlePubMedGoogle Scholar
- Hillis DM, Bull JJ. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst Biol. 1993;42:182–92.View ArticleGoogle Scholar
- Huelsenbeck JP, Rannala B. Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Syst Biol. 2004;53:904–13.View ArticlePubMedGoogle Scholar
- Rannala B, Yang ZH. Bayes estimation of species divergence times and ancestral population sizes using DNA sequences from multiple loci. Genetics. 2003;164:1645–56.PubMedPubMed CentralGoogle Scholar
- Zhang J, Kapli P, Pavlidis P, Stamatakis A. A general species delimitation method with applications to phylogenetic placements. Bioinformatics. 2013;29:2869–76.View ArticlePubMedPubMed CentralGoogle Scholar
- Fujisawa T, Barraclough TG. Delimiting species using single-locus data and the generalized mixed yule coalescent approach: a revised method and evaluation on simulated data sets. Syst Biol. 2013;62:707–24.View ArticlePubMedPubMed CentralGoogle Scholar
- Schwarzfeld MD, Sperling FA. Comparison of five methods for delimitating species in Ophion Fabricius, a diverse genus of parasitoid wasps (Hymenoptera, Ichneumonidae). Mol Phylogenet Evol. 2015;93:234–48.View ArticlePubMedGoogle Scholar
- Heled J, Drummond AJ. Bayesian inference of species trees from multilocus data. Mol Biol Evol. 2010;27:570–80.View ArticlePubMedGoogle Scholar