Asiatic shrews of the genus Chodsigoa (Soricidae, Eulipotyphla) currently comprise nine species, mostly occurring in southwest China. From May 2017 to August 2020, 11 specimens of Chodsigoa were collected from the Dabie Mountains in Anhui Province, eastern China. Their morphology was compared with other species within the genus and one mitochondrial (cytochrome b) and two nuclear (apolipoprotein B and breast cancer 1) genes were sequenced to estimate the phylogenetic relationships of these specimens. Based on morphological and molecular evidence, these specimens are recognized as a distinct species, Chodsigoa dabieshanensis sp. nov., which is formally described here. Morphologically, the new species is most similar to Chodsigoa hypsibia, but it is distinguishable from all known congeners by the combination of dark brownish pelage, small size, and relatively short tail. Phylogenetic analyses revealed that C. dabieshanensis sp. nov. forms a phylogenetic lineage sister to the clade containing C. parva + C. hypsibia. The-Kimura 2-parameter genetic distances of the cytochrome b (CYT B) gene between the new species and other nominal Chodsigoa species ranged between 8.6 and 17.6%. The new species is distributed at elevations from 750 to 1250 m in the Dabie Mountains and is geographically distant from other species in the genus.

Chodsigoa dabieshanensis, molecular analysis, morphology, new species, taxonomy

Asiatic shrews of the genus Chodsigoa Kastchenko, 1907 are mainly distributed in southwest China, adjacent Myanmar, Vietnam, and Thailand, and have also been recorded in central and eastern China and Taiwan (Hoffmann and Lunde 2008; Wilson and Mittermeier 2018). Animals in this genus are small in size (< 15 g) and mainly occur in mid-to high-montane forests, making them one of the least studied taxa among mammals. The genera Chodsigoa and Episoriculus were regarded as a subgenus of Soriculus (Hoffmann 1985) until recently, when Hutterer (2005) promoted them to full genus status. The most distinctive morphological characters distinguishing Chodsigoa from Soriculus/Episoriculus is the number of upper unicuspids. Chodsigoa has three upper unicuspids while Soriculus/Episoriculus has four. Nine species are currently recognized in Chodsigoa: C. caovansunga Lunde, Musser & Son, 2003, C. furva Anthony, 1941, C. hoffmanni Chen, He, Huang, Wan, Lin, Liu & Jiang, 2017, C. hypsibia (De Winton in De Winton and Styan 1899), C. parca Allen, 1923, C. parva Allen, 1923, C. salenskii (Kastschenko 1907), C. smithii Thomas, 1911 (Thomas 1911a), and C. sodalis Thomas, 1913.

The De Winton’s shrew (C. hypsibia) is endemic to China and is the most widely distributed species (Jiang and Hoffmann 2005). This gray, long-tailed shrew was first described by De Winton (1989) based on specimens from Yangliu-pa (= Yangliu ba), Pingwu, in Sichuan province. It contains two subspecies: C. h. hypsibia, recorded in Qinghai, Sichuan, Shaanxi, Tibet, Yunnan, Anhui, and Henan provinces (Zhang et al. 2018; Zhou et al. 2020) and C. h. larvarum Thomas, 1911 (Thomas 1911b), recorded in Beijing, Hebei, and Shanxi provinces (Liu et al. 2011). Zhang et al. (2018) reported the first record of C. hypsibia in Anhui province based on a specimen collected from Yaoluoping National Nature Reserve, Dabie Mountains. However, the collection site is distant from the known distribution of C. hypsibia, and the genetic distance of the CYT B gene between the specimen and C. hypsibia from Sichuan and Shaanxi (near the type locality in Pingwu, Sichuan) is relatively high (8.4–8.5%), and the two populations form deeply diverged clades in the Bayesian tree (posterior probabilities = 1.00; Zhang et al. 2018). These results suggest that additional studies with more specimens were necessary to confirm the taxonomic status of the population from the Dabie Mountains.

For three years, we conducted extensive field surveys in the Dabie Mountains, during which we collected 11 specimens of Chodsigoa. Based on morphological and molecular phylogenetic analysis, we recognize the population from the Dabie Mountains as distinct from C. hypsibia and other known Chodsigoa species, representing a new species Chodsigoa dabieshanensis sp. nov., which we describe herein.

A total of 11 Chodsigoa specimens were collected from May 2017 to August 2020 from Yaoluoping National Nature Reserve (n = 1), Bancang Natural Reserve (n = 4), and Foziling Natural Reserve (n = 6), all located in the Dabie Mountains, Anhui province, eastern China (Fig. 1). Shrews were sampled using the pitfalls (plastic buckets 15 cm in diameter and 28 cm in depth). Specimens were euthanized and liver or muscle tissues were extracted and preserved in pure ethanol. Skulls were also extracted and cleaned. Specimens and tissues were deposited at the Biological Museum of Anhui University (BMAHU). Animals were handled consistent with the animal care and use guidelines of the American Society of Mammologists (Sikes et al. 2016), and also following the guidelines and regulations approved by the internal review board of Anhui Normal University, and with the permissions of local authorities.

Figure 1. 

Map showing the collection site of Chodsigoa dabieshanensis sp. nov. in the Dabie Mountains, Anhui Province, eastern China.

External measurements including head and body length (HB), tail length (TL), hindfoot length (HF), ear length (EL) were taken in the field with a ruler to the nearest 0.1 mm. The body weight (W) of each specimen was weighed to the nearest 0.01 g using an electronic scale. All craniodental measurements were taken by CZZ using digital calipers graduated to the nearest 0.01 mm following Heaney and Timm (1983), Woodman and Timm (1993), and Chen et al. (2017). The following 19 measurements were taken:

CIL condyloincisive length;

IOB interorbital breadth;

CB cranial breadth;

CH cranial height;

RL rostral length;

PRL postrostral length;

PIL palatoincisive length;

PPL postpalatal length;

UTL upper toothrow length;

M²–M² maximum width across the upper second molars;

P⁴–M³ distance from the upper fourth premolar to the upper third molar;

PPD postpalatal depth;

BMF foramen magnum breadth;

ML mandibular length;

LTR lower toothrow length;

LLI length of lower incisor;

HCP height of coronoid process;

HCV height of coronoid valley;

HAC height of articular condyle.

Comparative morphological data of another 149 Chodsigoa specimens were obtained from our previous study (Chen et al. 2017), including C. caovansunga (3), C. furva (5), C. hoffmanni (14), C. hypsibia (64), C. parca (19), C. parva (31), C. smithii (11), and C. sodalis (2).

To evaluate the morphological variation among populations of Chodsigoa, we performed a principal component analysis (PCA) in SPSS 19.0 (SPSS Inc., USA) using the log₁₀-transformed craniodental measurements. We compared the morphology of the putative new species with other Chodsigoa species stored in Kunming Institute of Zoology (KIZ), the Sichuan Academy of Forestry (SAF), the Museum of Comparative Zoology, Harvard University (MCZ), and the American National Museum of Natural History (AMNH). The terminology for morphological descriptions followed Hoffman (1985), Lunde et al. (2003), and Chen et al. (2017).

Total genomic DNA of 10 C. dabieshanensis specimens were extracted using a DNA extraction kit (Qiagen DNeasy Blood and Tissue Kit, China). The complete CYT B gene and two nuclear gene segments [apolipoprotein B (APOB) and breast cancer 1 (BRCA1)] were amplified using primers and PCR conditions from Chen et al. (2021). The PCR products were purified and sequenced in both directions using the BigDye Terminator Cycle kit v. 3.1 (Invitrogen, USA) on an ABI 3730xl sequencer (Applied Biosystems, USA). Corresponding sequences of other Chodsigoa species were downloaded from GenBank (Table 1) and aligned with our new sequences using MUSCLE (Edgar 2004) and then checked manually by eye. Sequences of Episoriculus caudatus (Horsfield, 1851) and Neomys fodiens (Pennant, 1771) were included in the alignments as outgroup taxa. The Kimura-2-parameter (K2P) distances of the CYT B gene between species were calculated in MEGA 7 (Kumar et al. 2016).

Three datasets were used for the phylogenetic analyses: CYT B gene, concatenated nuclear genes, and concatenated mitochondrial and nuclear genes (Table 1). Maximum likelihood (ML) and Bayesian inference (BI) analyses were performed to reconstruct the phylogenetic relationships in PhyloSuite (Zhang et al. 2020) based on the best-fit partitioning schemes estimated using PartitionFinder v. 2.0 (Lanfear et al. 2012). The ultrafast bootstrap values (UFBoot) ≥ 95 and posterior probabilities (PP) ≥ 0.95 were considered as strong supports (Huelsenbeck and Rannala 2004; Minh et al. 2018).

External and cranial measurements are summarized in Table 2. The PCA based on 128 intact skulls produced two axes with eigenvalues exceeding 1.0, which explained 94.2% of the variation (Table 3). The first axes (PC1) explained 86.2% of the variation and was strongly positively correlated with all variables, indicating it represented the overall skull size (Table 3). The second axis (PC2) explained 8.0% of the variation and was highly positively correlated with CH and BMF (loading > 0.67). A plot of PC1 and PC2 (Fig. 2) showing that C. dabieshanensis are separated well from all named species. This new species occurs in the center of the morphospace, indicating its medium size in the genus. Morphologically, it is most similar to C. hypsibia, with which it occupies the upper left corner morphospace without overlap (Fig. 2), indicating its generally smaller size, larger BMF, and higher CH (Table 2).

Figure 2. 

Results of principal component analysis of Chodsigoa based on the 19 log₁₀-transformed craniodental measurements.

Nine CYT B (1140 bp), nine APOB (513 bp), and nine BRCA1 (768 bp) sequences of C. dabieshanensis were obtained (GenBank accession numbers: OM200108–OM200133; Table 1). The ML and BI trees recovered very similar topologies, and therefore, only the ML gene trees are shown (Fig. 3). The phylogenetic analyses of all three datasets supported Chodsigoa clustered into two major clades (UFboot > 99, PP = 1.00). One clade was composed of C. parva, C. hypsibia, and C. dabieshanensis (Clade I), and the other clade was composed of C. caovansunga, C. furva, C. hoffmanni, C. parca, C. salenskii, C. smithii, and C. sodalis (Clade II). The C. dabieshanensis clade was strongly supported as a monophyletic lineage, sister to the clade containing C. parva and C. hypsibia (UFboot > 98, PP = 1.00). The K2P genetic distances of the CYT B gene between C. dabieshanensis and other nominal Chodsigoa species ranged from 8.6% (with C. hypsibia) to 17.6% (with C. sodalis) (Table 4).

Figure 3. 

Maximum likelihood phylogenetic trees derived from A the CYT B gene B the concatenated nuclear genes C the concatenated mitochondrial-nuclear trees. Branch labels indicate Bayesian posterior probabilities (PP) and ultrafast bootstrap supports (UFBoot). Scale bars represent substitutions per site.

Based on the morphological, morphometric, and molecular evidence and the modern phylogenetic species concept (phylogenetic species concept based on both diagnosability and monophyly as operational criteria) (Mayden 1997; Gutierrez and Garbino 2018), we recognize the population from the Dabie Mountains as a new species of Chodsigoa, which is formally described below.

Prior to this study, nine species were recognized in the genus Chodsigoa (Chen et al. 2017; Wilson and Mittermeier 2018). Our morphological and molecular results support that the specimens from the Dabie Mountains represent a new species of Chodsigoa, C. dabieshanensis sp. nov., based on the diagnosis-and-monophyly-based phylogenetic species concept (Mayden 1997; Gutierrez and Garbino 2018). Chodsigoa dabieshanensis sp. nov. is morphologically closely related to C. hypsibia and was previously considered as a marginal population of that taxon (Zhang et al. 2018). However, it can be distinguished from C. hypsibia by its dark brownish pelage and smaller size (Table 2). The large genetic distance (8.6% by the CYT B gene) and phylogenetic analysis also strongly support they are two distinct species (UFboot > 98, PP = 1.00). As Chodsigoa are mainly distributed in southwest China and adjacent areas (Wilson and Mittermeier 2018), the distribution area of C. dabieshanensis sp. nov. is marginal. It is the only known species of Chodsigoa recorded in Anhui province, separated by at least 500 km from any other member of the genus, i.e., C. hypsibia from Luanxian, Henan Province (Zhou et al. 2020). The new species has no known congeners in Anhui Province; there are only two other soricid taxa recorded, Chimarrogale lender Tomas, 1902 and Crocidura spp. (Wang 1990; Jiang et al. 2015). The former is a large aquatic shrew (W > 20 g), and the latter has white, unpigmented dentition; these taxa are easily distinguishable from the new species.

The new species brings the number of Chodsigoa species to 10, sorted into two major clades; one including C. parva + C. hypsibia + C. dabieshanensis sp. nov. (Clade I), and the other (Clade II) comprised of the remaining species (Fig. 3). These results are also supported by morphology. Compared with the species in Clade II, the cranium of Clade I species is markedly flatter, and the tail of Clade I is relatively shorter (Clade I: TL/HB < 100%; Clade II: TL/HB ≥ 100%). All our gene trees showed C. dabieshanensis sp. nov. forms a subclade inside the main Clade I as the sister group of the subclade C. parva + C. hypsibia (UFboot > 98, PP = 1.00, Fig. 3).

As the most easterly distributed species of Chodsigoa, the discovery of C. dabieshanensis sp. nov. from the Dabie Mountains is important in understanding the macroevolution of the genus. Previous studies suggested that the tribe Nectogalini originated from Europe and migrated eastward to western Siberia and southward along northern China to southwest China (He et al. 2010). While the Hengduan Mountains are considered to serve as an important route for the southward migration (Zhang 2002; He et al. 2010), we have no knowledge of how this group migrated eastward. The oldest fossils of Chodsigoa are from the Early Pliocene in Gansu Provence, northern China (Zhang and Zheng 2001). Fossils of C. cf. hypsibia and C. cf. parva were discovered from the Early Pleistocene in Jianshi, Hubei and Wuhu, Anhui, both in eastern China, and more fossils were found in Wushan, Chongqing, southwest China in the Late Pleistocene (Qiu and Li 2005). These fossil records, together with our finding of C. dabieshanensis sp. nov., diverged earlier than C. hypsibia and C. parva, which suggests that the ancestor of Clade I arrived early in eastern China. Due to the present lack of broad geographic sampling, how the genus migrated to eastern China is still an open question. The Dabie Mountains are an extension of the Qinling fold belt and gradually stabilized by the end of the Tertiary (Feng 1976). Considering that the montane archipelagos always act as refugia and corridors to facilitate the dispersal of terrestrial small mammals (Chen et al. 2015; He and Jiang 2014; He et al. 2019), a parsimonious biogeographic scenario of the migration is via the Qinling and Dabie mountains. The ancestor of new species then became isolated due to climate change and following habitat turnover, resulting in a new species. Finer taxon sampling with additional sequence data is warranted to illustrate the migration patterns of the genus.

The study was supported by the National Natural Science Foundation of China (no. 31900318), National Science & Technology Fundamental Resources Investigation Program of China (grant no. 2019FY101800), the Anhui Provincial Natural Science Foundation (2008085QC106), and the University Synergy Innovation Program of Anhui province (GXXT-2020-075).