First record of disk-footed bat Eudiscopus denticulus (Chiroptera, Vespertilionidae) from China and resolution of phylogenetic position of the genus.

The disk-footed bat Eudiscopus denticulus(Osgood, 1932) is a rare species in Southeast Asia. During two chiropteran surveys in the summer of 1981 and 2019, eight and three small Myotis-like bats with distinct disk-like hindfeet were collected from Yunnan Province, China, respectively. External, craniodental, and phylogenetic evidence confirmed these specimens as E. denticulus, representing a new genus in China. The complete mitochondrial genome consistently showed robust support for E. denticulus as a basal lineage within Myotinae. The coding patterns and characteristics of its mitochondrial genome were similar to that of other published genomes from Myotis. The echolocation signals of the newly collected individuals were analyzed. The potential distribution range of Eudiscopus in Southeast Asia inferred using the MaxEnt model indicated its potential occurrence along the southern border region of Yunnan, China.

DEAR EDITOR,

The disk-footed bat Eudiscopus denticulus(Osgood, 1932) is a rare species in Southeast Asia. During two chiropteran surveys in the summer of 1981 and 2019, eight and three small Myotis-like bats with distinct disk-like hindfeet were collected from Yunnan Province, China, respectively. External, craniodental, and phylogenetic evidence confirmed these specimens as E. denticulus, representing a new genus in China. The complete mitochondrial genome consistently showed robust support for E. denticulus as a basal lineage within Myotinae. The coding patterns and characteristics of its mitochondrial genome were similar to that of other published genomes fromMyotis. The echolocation signals of the newly collected individuals were analyzed. The potential distribution range of Eudiscopus in Southeast Asia inferred using the MaxEnt model indicated its potential occurrence along the southern border region of Yunnan, China.

In 1932, Osgood described a new genus and species of vespertilionid bat based on six specimens collected at Phong Saly in northern Laos, externally resembling a small Myotis species but with a striking adhesive disk on the hindfoot and with a flattened skull, and was named Discopus denticulus Osgood, 1932. Conisbee (1953) noted, however, that the name Discopus was preoccupied and proposed Eudiscopus as a replacement name. Till today, only a limited number of specimens have been found in Southeast Asia and the biological information available on the species is limited (Wilson & Mittermeier, 2019). Presently, the species is reported from Laos (Osgood, 1932), Myanmar (Koopman, 1970), Thailand (Kock & Kovac, 2000; Soisook et al., 2016), and Vietnam (Kruskop, 2010, 2013; Zsebők et al., 2014) (Figure 1F). Based on morphological examinations and phylogenetic analyses of 11 specimens collected from Yunnan, we herein report on the first occurrence of this poorly known bat from China.

Figure 1

External (A–C), skull, and dentition (D, E) characteristics of Eudiscopus denticulus from China (GZHU 19159), its potential distribution areas in Southeast Asia predicted by MaxEnt (F), and two maximum-likelihood phylogenetic trees using IQ-tree (G, H)

Live individual (A), ear (B), and hindfoot (C); lateral view of skull and mandible (D); ventral view of skull (E). Scale bar: 5 mm. Photos by Yi Wu (A–C); drawings by Wen-Hua Yu (D, E). F: Black circles mark sampling localities in Xishuangbanna, Yunnan, China; green circles represent historical occurrences from literature and Global Biodiversity Information Facility (GBIF) database (Occurrence dataset https://doi.org/10.15468/igaciv accessed via GBIF.org on 2019-10-14). Red regions are good potential distributions based on localities known so far; orange areas are predictions with inclusion of present records. G: Phylogenetic tree containing 360 presumed vespertilionid species representing large-scale sampling strategy. H: Phylogenetic tree based on 101 complete mitochondrial sequences. Red, green, and blue trapezoids represent Eudiscopus, Submyotodon, and Myotis, respectively.

Due to the limitation of available samples and sequences of E. denticulus, its phylogenetic position remains contradictory. Traditional morphology-based systematics place it within Vespertilioninae (Simmons, 2005; Tate, 1942), or close to Myotis (Borisenko & Kruskop, 2003). Several phylogenies support the inclusion of E. denticulus in Myotinae (Amador et al., 2018; Tsytsulina et al., 2007; Yu et al., 2014); whereas others reveal different positions, including placement outside of Myotinae (Amador et al., 2018; Shi & Rabosky, 2015), or — albeit with low support — as a sister taxon to Hesperoptenus tickelli (Görföl et al., 2019). Clarifying its phylogenetic position would not only improve our understanding of the morphological evolutionary process of this unique species but also help determine the convergent evolution of flattened skulls and the presence of the disk-like hindfoot in Vespertilionidae. Two data matrices representing a large-scale sampling of Vespertilionidae (~5 kb, 360 species) and a compilation of all currently deposited complete mitochondrial genomes (~15 kb, 101 species) were herein generated to determine the phylogenetic position of this species. Acoustic characteristics of the echolocation calls were further analyzed. Through maximum entropy (MaxEnt) modelling, potentially suitable habitats were also inferred.

Eudiscopus denticulusfrom China was sampled from three sites in Xishuangbanna, Yunnan Province (see Supplementary Data). The eight specimens sampled in 1981 were initially misidentified as Tylonycteris sp. but were reclassified as E. denticulus in 2019. All specimens in the study are adults based on the status of the epiphyseal cartilage gap in the metacarpal joint (Kunz & Anthony, 1982). They are presently stored in the collections of the Kunming Institute of Zoology, Chinese Academy of Sciences, Yunnan, China (KIZ 811351–811358) and the School of Life Sciences, Guangzhou University, Guangdong, China (GZHU 19159, 19160, 19164). The specimens are all small-sized individuals with a forearm length of 34.8–38.5 mm (Table 1), and externally resemble a small Myotis species. The fur is dense and soft, reddish brown at the dorsum (Figure 1A). The ears reach the tip of the muzzle when laid forward, and the tragus is straight and distinctly narrowing, ending in a blunt tip (Figure 1B). Noticeable disk-like adhesive pads are present on the feet, similar to that in Tylonycteris and Glischropusbut more pronounced ventrally (Figure 1C). The skull is relatively broad and strikingly flattened; rostrum is elongated and relatively long and upturned anteriorly (Figure 1D, E). Dental formula is I 2/3, C 1/1, P 2/3, M 3/3. Upper canine is Myotis-like without any supplementary cusps; first upper incisor is higher than the second, as in Myotis, p3 is largely reduced, displaced to the lingual side of the tooth row, and compressed between p2 and p4; lower molars are myotodont (Figure 1D, E). Detailed information and measurements of all specimens used in this study are listed in Table 1 and Supplementary Appendix I. To illustrate potential geographic variation, principal component analysis (PCA) with varimax rotation and t-test were performed. According to the integrity of our data, 10 craniodental measurements were selected in the PCA. The first two components (44% and 33% for principal components PC1 and PC2, respectively) accounted for 77% of total variance (Supplementary Figure S1). For PC1, greatest length of skull (GTL), length of maxillary toothrow (UCM3L), length of mandibular toothrow (LCM3L), and greatest length of mandible (MANL) had high positive loadings (Table 1), reflecting an external overall size effect. PC2 was mostly related to braincase height (BCH), mastoid width (MAW), and interorbital width (IOW) (Table 1), generally implying width and height of braincase. These patterns indicated that individuals from China and Laos have a wider and more flattened braincase than those from the middle regions of Vietnam and Myanmar (Table 1 and Supplementary Figure S1). Such differences were also revealed from comparisons between China and Vietnam using t-test, which additionally revealed a longer forearm in the Chinese population (Table 1).

Table 1

Descriptive statistics of external and craniodental measurements, and echolocation parameters of Eudiscopus denticulusfrom China and nearby countries

Index	Yunnan, China	Pu Huong, Vietnam	t-value	Pegu, Myanmar	Laos	PC1	PC2
Abbreviations can be found in text and Supplementary Materials and Methods. Values are given as means±SD (if n>3) and minimum-maximum (min-max).t-value is from Students t-test between China and Vietnam specimens when measurement distribution fits normality, and * represents P<0.05. Using a Pettersson D500X ultrasound detector (Pettersson Elektronik AB), echolocation calls were recorded from three Chinese-sampled bats (collected in 2019) flying in a room (5 m×4 m×3 m).# indicates secondary means and minimum and maximum values based on mean values of three individuals from Zsebők et al. (2014). Note, scores in first parentheses in echolocation measurements indicate number of calls analyzed in this study; those from Zsebők et al. (2014) represent number of individuals in their study. Indices in bold indicate variables used in principal component analysis, PC1 and PC2 scores in bold indicate variables with greatest loadings in respective component. －: Not available.
W (g)	5.2±1.14 (11) (4.0－7.0)	－	－	－	－	－	－
HB (mm)	40.5±2.28 (11) (36.0－45.3)	38.0±2.19 (5) (35.9－40.5)	2.06	－	－	－	－
T (mm)	40.3±2.87 (11) (35.0－45.0)	38.8±2.56 (5) (34.9－41.4)	1.05	－	－	－	－
E (mm)	11.2±1.75 (11) (8.0－13.0)	11.6±0.66 (4) (10.9－12.5)	－0.51	－	－	－	－
HF (mm)	6.0±0.82 (11) (5.4－8.1)	6.7±0.63 (5) (5.8－7.2)	－1.56	－	－	－	－
FA (mm)	36.7±1.10 (11) (34.8－38.5)	34.5±0.88 (5) (33.5－35.6)	3.98*	－	－	－	－
Tib (mm)	17.2±0.54 (11) (16.2－17.9)	16.3±0.64 (5) (15.4－16.9)	2.68*	－	－	－	－
GTL (mm)	14.34±0.34 (8) (13.71－14.79)	14.00±0.22 (10) (13.59－14.32)	2.56*	13.43	14.49	0.81	0.54
CCL (mm)	13.28±0.28 (8) (12.83－13.74)	12.79±0.21 (11) (12.47－13.04)	4.43*	12.15	13.16	0.72	0.62
CBL (mm)	14±0.32 (8) (13.35－14.42)	－	－	－	－	－	－
BCW (mm)	6.84±0.14 (8) (6.63－7.01)	6.74±0.1 (11) (6.53－6.87)	1.76	6.36	6.99	0.46	0.56
BCH (mm)	4.53±0.23 (8) (4.16－4.78)	3.76±0.12 (11) (3.54－3.97)	8.86*	3.07	3.54	0.35	0.72
ZYW (mm)	9.48±0.29 (4) (9.16－9.82)	9.21±0.15 (6) (8.97－9.43)	－	－	－	－	－
MAW (mm)	7.72±0.19 (8) (7.41－7.92)	7.51±0.16 (11) (7.34－7.81)	2.64*	7.20	7.75	0.49	0.75
PL (mm)	6.38±0.11 (8) (6.25－6.6)	－	－	－	－	－	－
IOW (mm)	3.73±0.11 (8) (3.60－3.89)	3.61±0.16 (11) (3.4－3.94)	1.95	3.47	3.71	－0.14	0.87
UIM3L (mm)	6.42±0.18 (8) (6.12－6.68)	－	－	－	－	－	－
UCM3L (mm)	5.44±0.11 (8) (5.31－5.58)	5.36±0.11 (10) (5.21－5.49)	1.61	5.20	5.48	0.88	0.25
UCCW (mm)	3.77±0.06 (8) (3.71－3.91)	3.70±0.09 (9) (3.60－3.91)	1.60NS	3.59	3.80	0.62	0.53
UM3M3W (mm)	5.86±0.12 (8) (5.62－6.01)	5.81±0.09 (10) (5.74－6.05)	1.24	－	5.88	－	－
LIM3L (mm)	6.80±0.15 (8) (6.56－7.06)	－	－	－	－	－	－
LCM3L (mm)	5.74±0.12 (8) (5.54－5.88)	5.70±0.12 (11) (5.52－5.85)	0.77	5.49	5.68	0.89	－0.07
MANL (mm)	10.23±0.18 (8) (9.82－10.42)	10.2±0.15 (11) (9.91－10.46)	0.34	9.74	10.63	0.80	0.30
PCH (mm)	3.29±0.11 (8) (3.13－3.40)	3.16±0.11 (11) (3－3.38)	2.63*	3.00, 3.09	3.25	－	－
HF (kHz)	98.3±3.25 (24) (92.0－104.0)	108.9 (3) (106.5－112.9)#	－	－	－	－	－
LF (kHz)	49.9±2.02 (24) (45.0－54.4)	52.0 (51.1－53.9)#	－	－	－	－	－
FMAX (kHz)	53.3±1.29 (24) (50.1－55.0)	61.7 (60.9－63.0)#	－	－	－	－	－
DUR (ms)	3.1±0.24 (24) (2.8－3.7)	2.03 (1.87－2.11)#	－	－	－	－	－

The complete mitochondrial genome of E. denticuluswas ~16 500 bp in length, containing 13 protein-coding genes, two ribosomal RNA genes, 22 transfer RNA genes, and a control region (GenBank accession No.: MW085031 for GZHU 19159, Supplementary Figure S2). Most genes were encoded on the H-strand, except for eight tRNA and ND6 genes. The coding patterns and characteristics are similar to those in other published mitochondrial genomes of Myotinae (e.g., Chung et al., 2018; Jebb et al., 2017; Jiang et al., 2016; Platt et al., 2017; Yu et al., 2016). Our large-scale vespertilionid maximum-likelihood phylogenies highly supported E. denticulus as a monophyletic clade (bootstrap value: 100; Figure 1G and Supplementary Figures S3A) as well as its inclusion within Myotinae (Figure 1G, H). Furthermore, E. denticulus appeared as a sister taxon to a clade clustering Myotis and Submyotodon, indicating a basal position within Myotinae (Figure 1G). For intraspecific relationships, an ambiguous haplotype network emerged, which implied a close matrilineal relationship between our samples and individuals from Bu Gia Map National Park, BinhPhuc, Vietnam, based on mitochondrial markers (GenBank accession No. of Chinese individuals: MT822523 for GZHU 19159, MT822524 for GZHU 19160; Supplementary Figure S3B).

Eudiscopus denticulus uses a broadband, downward frequency modulated (FM) call when flying in closed environments (see Supplementary Data; Supplementary Figure S4). Both our recordings and those from Zsebők et al. (2014), from specimens flying in a tent or room, showed a broadband FM signal with a narrowband FM end, rather than the broadband FM calls reported by Hughes et al. (2011) in two individuals released in a cluttered environment. The call structure and relatively long and broad wings support the view that E. denticulus is an edge space aerial forager (Schnitzler et al., 2003; Zsebők et al., 2014). From the specimens in China, specifically, the maximum start frequency was found to be around 98 kHz, and the minimum end frequency was around 50 kHz (Table 1). At the start of the echolocation call, the narrowband FM portion was always relatively short (Supplementary Figure S4). The maximum energy was at 53.3 kHz, on average (Supplementary Figure S4), while the mean length of the echolocation calls was around 3.1 ms (Table 1). The geographic differences in morphological measurements were mirrored in the acoustic signals, as our samples emitted lower frequency echolocation calls (e.g., highest frequency (HF), lowest frequency (LF), and frequency with most energy (FMAX)) than those from Vietnam (Zsebők et al., 2014) (Table 1). These differences seem to be in line with the negative correlation between body size and call frequency (Barclay et al., 1999; Robinson, 1996; Schnitzler & Kalko, 2001; Yoshino et al., 2006); however, more elaborate experiments are needed to verify these differences due to the variances in methods used for recording, which may have affected echolocation parameters.

As the present record is the northernmost locality of E. denticulus, we ran two MaxEnt models (with and without the new record) to evaluate its influence on potential distribution of the species. Both model outputs provided satisfactory results, with AUC values of the models with and without our records of 0.89 and 0.75, respectively. Mean temperature of warmest quarter (BIO10) and precipitation of coldest quarter (BIO19) were the most important influencing factors. The model for current potential distribution of E. denticulus in Southeast Asia showed that the most suitable habitat is medium warm and moist forest with bamboo stands. The large potential area with the inclusion of the present Chinese records is approximately twice the size as the prediction based on published occurrences so far (Figure 1F). However, the MaxEnt model almost always provides over-estimates compared to the realized niche of a species, as it only considers niche-based presence data and predicts a species’ fundamental niche rather than realized niche (Phillips & Dudík, 2008; Phillips et al., 2004). Nevertheless, the inference obtained should facilitate subsequent targeted investigation on this rare species.

In the IUCN Red List, E. denticuluswas shifted from Lower Risk/Near Threatened in 1996, to Data Deficient in 2008, and Least Concern at present. This classification is based on its wide distribution across Southeast Asia, lack of major threats, and no sign of decline at a rate that would qualify it in a threatened category (Soisook et al., 2016). Nevertheless, its rarity and our poor knowledge on its biology have been highlighted (Kock & Kovac, 2000; Koopman, 1970; Kruskop, 2010, 2013; Tsytsulina et al., 2007; Wilson & Mittermeier, 2019). Furthermore, our distribution predictions using MaxEnt show that suitable habitat areas are extremely fragmentary (Figure 1F). Of note, the large temporal gap in capture/collection of this species in China (38 years) also indicates its rarity and fragility. As such, particular conservation attention should be reinforced in the future.

Additional contribution herein refers to the scanning of a skull of E. denticulus (KIZ 811355) using a RexcanDS3 Silver 3D Scanner designed for small objects with a maximum resolution of 0.01 mm (Supplementary Figure S5). The scanned PHY files are deposited as supplementary materials (Suppl. PHY files). Although micro-computed tomography (μCT) scanners would capture more accurate details with better resolution, the considerably smaller size of our files than those produced by a μCT scanner (e.g., ~40 MB vs. ~600 MB in the case of a skull scan), and sufficient accuracy suggest that laser 3D scanners can be used as an alternative for shape analyses and morphological studies (Marcy et al., 2018). We believe that 3D digitization and virtual access platforms could facilitate future species determination and boost international academic cooperation.

SCIENTIFIC FIELD SURVEY PERMISSION INFORMATION

Permission for field surveys in Xishuangbanna, Yunnan Province, China, were granted by the Xishuangbanna Nature Reserve Administration.

Supplementary data to this article can be found online.

Click here for additional data file.

COMPETING INTERESTS

The authors declare that they have no competing interests.

AUTHORS' CONTRIBUTIONS

W.H.Y., G.C., Y.W., and S.L. designed the study; Z.L. Y.W., R.C.Q., Q.Y.W., Y.W., and S.L. collected materials for the study; W.H.Y., Y.N.L., G.C., H.Y.S., and Y.W. performed morphometric and phylogenetic analyses; W.H.Y., G.C., Y.W., and S.L. interpreted the results and prepared the manuscript, photographs, and figures for the study. All authors read and approved the final version of the manuscript.

ACKNOWLEDGEMENTS

We are grateful to the anonymous referees for their constructive comments. We also thank Yi-Feng Hu, Yang Yue, and Han-Bo Zhang for their help in field survey and assistance during lab work.