﻿First record of the flat-skulled woolly bat Kerivouladepressa and the Indochinese woolly bat K.dongduongana (Chiroptera, Vespertilionidae) in China

﻿Abstract Recent studies have revealed that the Kerivouladepressa complex should be divided into two species, K.depressa distributed mainly in Myanmar, Vietnam, Laos and Cambodia, and K.dongduongana found only in the Annamite Mountains of Vietnam, Laos and Cambodia. In November 2018 and April 2019, 24 woolly bats were collected by two-band harp traps in Xishuangbanna, Yunnan, China. Based on morphological, morphometric, and phylogenetic (COI, Cytb, and RAG2 gene sequences) analyses, these bats were identified as K.depressa and K.dongduongana, representing two new species records for the country. Including the new records, six Kerivoula species have been recorded in China, namely K.depressa, K.dongduongana, K.furva, K.kachinensis, K.picta and K.titania. To facilitate their identification and biological research in the future, we have provided an up-to-date key to all Kerivoula species occurring in China.

In China, four species of woolly bats from the genus Kerivoula have been recorded, including K. picta (Pallas, 1767), K. furva, K. kachinensis, and K. titania (Bates et al., 2014) (Wu et al. 2012;Kuo et al. 2017;Tu et al. 2018;Yu et al. 2018;Wilson and Mittermeier 2019;Yu et al. 2022). In November 2018 and April 2019, a series of chiropteran surveys were conducted in the southwestern region of Yunnan Province, and 24 Kerivoula individuals were sampled. Based on morphology, morphometric analyses, and phylogenetic inferences using COI, Cytb, and RAG2 sequences, they were identified as K. depressa and K. dongduongana, which represent two new records of Kerivoula in China. In this paper, we provided details about these findings, new distribution information and an up-to-date key to identify all Kerivoula species occurring in China.

Specimen sampling and morphological measurements and analyses
In November 2018 and April 2019, 24 Kerivoula bats were collected using two-band harp traps during field surveys in Xishuangbanna Tropical Botanical Garden, Yunnan,China (21°57'17"N,101°15'26"E and 21°30'58"N,101°30'38"E). All field survey and sample collection protocols complied with the current laws of Yunnan Province, China. We followed the guidelines of the American Society of Mammalogists (Sikes 2016) for the care and use of animals. All voucher specimens were determined to be adults based on the degree of epiphyseal-diaphyseal fusion (Brunet-Rossinni and Wilkinson 2009). The specimens were preserved in 75% ethanol and deposited at the School of Life Sciences, Guangzhou University, China.
External and skull measurements were taken with a digital caliper to the nearest 0.01 mm following Bates and Harrison (1997) and Bates et al. (2004). Body mass was measured with an electronic scale. Twenty-four adult specimens were examined using six external and eight craniodental measurements following Tu et al. (2018) and Yu et al. (2018), and further morphometric analyses were performed using 20 specimens (Table 1). We conducted a principal component analysis (PCA) and discriminant analysis of principal components (DAPC) of craniodental measurements using the R Core Team (2013) and R add-in packages: psych (Revelle 2013), ade4 (Dray et al. 2007), adegenet (Jombart 2008), FactoMineR (Le et al. 2008), and gg-plot2 (Wickham 2016).

Morphological examination
Kerivoula depressa is a moderate-sized species with a forearm (FA) length of 30.75 ± 1.08 mm. Ears are small and rounded, and the posterior margin of the pinnae has a deep, smoothly concave emargination just below the apex. Overall pelage color is buff brown to dark brown. The lower part of ventral hair is dark brown, whereas its tip is light brownish yellow. Dorsal fur is of black base but with dark brown tip ( Fig. 1A-C). The domed skull is small, with the greatest length of 13.75 ± 0.17 mm. The mid-portion of the braincase exceeds the frontal region in height. Its lateral profile is flattened from the rostrum to the forehead. A sagittal crest is not evident, and the lambdoid crests are relatively weak. The dental formula is I 2/3, C 1/1, P 3/3, M 3/3. The second upper incisor (I 3 ) is about half of the first upper incisor (I 2 ) in height, and the latter is one half the height of the upper canine. The third upper premolar (P 4 ) is distinctly higher than the anterior two. The third upper molar (M 3 ) is degenerated. The crown area of the first and second lower molars is approximately equal and slightly larger than the last molar (M 3 ). Kerivoula dongduongana, with FA of 34.02 ± 0.94 mm, is slightly larger than K. depressa. Its ears are rounded with a tiny, smooth depression near the tip. Kerivoula dongduongana is obviously yellower than K. depressa in pelage coloration (Fig. 1D-F).   The ventral pelage is golden brown, and the base and middle portions are medium brown with golden-brown tips. The dorsal pelage is dark brown with golden-brown tips. Craniodental features are similar to K. depressa, but K. dongduongana is characterized by a flatter and longer skull (Fig. 2C, G; Table 1).

Multivariate comparison analysis
PCA based on eight craniodental measurements revealed 96.2% of the total variance from the first two principal components (PCs) (86.9% and 9.3% for PC1 and PC2, respectively) in the scatter plot of the six morphological groups (Fig. 3A). For PC1, all measurements had positive loadings (Table 1), reflecting the skull size. Larger bats were characterized by higher PC1 scores; thus, specimens of K. kachinensis clustered to the right compared with those of other taxa (Fig. 3A). For PC2, all measurements had low loadings except for the braincase height (BH) (Table 1). Therefore, based on PC2, K. picta and K. titania, which had a larger BH, were clustered to the top of the plot, whereas K. dongduongana and K. kachinensis were assigned to the bottom (Fig. 3). For DAPC, we entered the first two PCs from the PCA results and obtained two discriminant functions (DFs) to distinguish among studied Kerivoula specimens. PCA and DAPC scatter plots showed that K. depressa and K. dongduongana specimens formed distinct and separated clusters (pale pink and green triangles in Fig. 3), although some scatter values overlapped with those from K. furva. Meanwhile, K. picta, K. titania, and K. kachinensis clustered into three distinguishable groups (Fig. 3).

Discussion
The major interspecific phylogenetic relationships of our analyses are comparable with those reported by Kuo et al. (2017) and Tu et al. (2018). Our studies similarly confirmed the monophyly of K. depressa and K. dongduongana. However, the topology of the phylogenetic tree based on the RAG2 gene remains unresolved and needs further study. Finally, combining the results of external and craniodental examination and multivariate analyses, 24 specimens were determined as K. depressa and K. dongduongana (Table 1).
Our discovery of K. depressa and K. dongduongana in China indicates that six species of Kerivoula live in China. According to morphological analyses, K. picta is easily distinguished by its unique pelage color pattern and skull shape (Wilson and Mittermeier 2019), whereas K. kachinensis is the largest species with a distinctly flattened skull (Bates et al. 2004;Tu et al. 2018;Yu et al. 2022). As for the remaining four similar-sized species, K. titania has a distinctly longer tibia and higher braincase than the others (Kuo et al. 2017;Tu et al. 2018). In pelage coloration, K. furva has the darkest fur color, varying from black brown to black gray, whereas K. depressa and K. dongduongana are pale brown.
Among the four species, K. dongduongana has the shortest BH (Kuo et al. 2017;Tu et al. 2018). A key to the Kerivoula species occurring in China is provided in Appendix 2.
Until recently, Kerivoula species were considered forest-dependent (Wilson and Mittermeier 2019). They are known in the south of China across Yunnan to Taiwan, and from Hainan to Chongqing. It is worth noting that five of the "Chinese" Kerivoula species are found in the southwest region of Yunnan Province, which is often treated as a biodiversity hotspot near Myanmar, Laos, and Vietnam. Its unique terrain, vegetation, and environmental conditions, including a low latitude, warm and tropical forest, and humid micro-climate, appear suitable for inhabitation and colonization (Kruskop 2013;Wilson and Mittermeier 2019;Qian et al. 2020). The high diversity of woolly bats in tropical forest areas may indicate the origin in their diversification progress.
Based on a comparison of the recorded Kerivoula diversity from the bordering countries of Myanmar (five species), Laos (seven species), and Vietnam (eight species) (Wilson and Mittermeier 2019), we suggest that there is still a risk of underestimating the diversity of Kerivoula in China. More surveys should therefore be conducted, especially on the border/unexplored region and using effective sampling tools such as multi-bank harp traps.
Appendix 1 Table A1. Sample used in molecular analyses, with GenBank accession numbers for COI, Cytb and RAG2 genes provided. Newly generated sequences in this study are shown in bold.