Kerivoula furva Kuo & Soisook & Ho & Csorba & Wang & Rossiter 2017, sp. n.

Kerivoula furva sp. n. (Figs. 1–6, Tables 1–2)

Holotype

NMNS 17595, adult ♀, dry skin, skull and postcranial skeleton, collected by Yi-Wen Chen, 29 Mar 2009.

Type locality

TAIWAN, Yilan County, Yuanshan Township, 3 km East of Shuanglianpi, 24°45.21’N, 121°39.63’E, 180 m a.s.l.

Paratypes

NMNS 17657, adult ♀, dry skin, skull and postcranial skeleton and NMNS 19136, adult ♀, dry skin, skull and postcranial skeleton, both from the type locality; THU 07.X.25.9, adult ♀, dry skin and skull and THU 07.X.25.11, adult ♂, dry skin and skull, both from TAIWAN, Yilan County, Yuanshan Township, Shuanglianpi, 24°45.04’N, 121°38.05’E, around 500 m a.s.l.; HNHM 2016.8.1, adult ♂, in alcohol with skull extracted, collected at TAIWAN, Taoyuan City, Fuxing District, 2.5 km SE of Xiao Wulai, 24°47.01’N, 121°24.07’E, 650 m a.s.l.; TESRI B1398, adult ♀, dry skin, skull and postcranial skeleton, TESRI B1399, adult ♀, dry skin, skull and postcranial skeleton, TESRI B1400, adult ♀, dry skin, skull and postcranial skeleton and TESRI B1401, adult ♀, dry skin, skull and postcranial skeleton, the above four from TAIWAN, Taichung City, Heping District, Wushikeng Low Altitude Experimental Station of TESRI, 24°16.42’N, 120°57.059’E., 1056 m a.s.l.; HNHM 2005.66.1, adult ♂, in alcohol with skull extracted, collected at TAIWAN, Tainan City, Dongshan District, Gaoyuan Village; NMNS 13946, adult ♂, dry skin, skull in THU (B040017), NMNS 13947, adult ♂, dry skin, skull in THU (B040018), NMNS 13948, adult ♀, dry skin, skull in THU (B030041) and NMNS 13952, juvenile ♂, dry skin, skull in THU (B030028), the above four from TAIWAN, Kaohsiung City, Taoyuan District, Meilan Logging Road, 23°17.24’N, 120°49.46’E, 920 m a.s.l.; NMNS 10806, adult ♂, dry skin, skull in THU (B040007) and NMNS 14039, adult ♀, dry skin, skull in THU (B030025), both from TAIWAN, Kaohsiung City, Taoyuan District, Meilong Logging Road, 23°5.70’N, 120°44.28’E, around 800 m a.s.l.; Field number PS160206.3 (to be subsequently deposited in the Zoological Collection of the University of Mandalay), adult ♀, in alcohol with skull extracted, collected from MYANMAR, Kachin, Putao Township, Wang Hlaing Dam Village, 27.479°N, 97.169°E, 863 m a.s.l.

Referred specimens

BMNH 21.1.6.32, adult ♀, INDIA, Meghalaya, Jaintia Hills; BMNH 50.432, adult ♂, and BMNH 50.438, adult ♀, both from MYANMAR, Kachin, Nam Tamai; Field number PS160206.1 (to be subsequently deposited in the Zoological Collection of the University of Mandalay), adult ♂, from MYAN- MAR, Kachin, Putao Township, Wang Hlaing Dam Village. The two specimens from CHINA, Hainan (IBHG 08279 and IBHG 08280, both ♀) which are consistently supported by morphological, karyotypic and molecular evidence to have a close relationship with K. furva sp. n. in Taiwan (Wu et al., 2012) are considered to belong to this new species. In addition, the following specimens from CHINA which formed the COI lineage K. hardwickii D along with K. furva sp. n. in Taiwan (Fig. 1a) are also assigned to this new species: ROM MAM 116079, ♀, Guangxi, Jing Xin County Provincial Nature Reserve; ROM MAM 114918, ♂, ROM MAM 114944, ♀, both from Hunan, Shuhuangshan Nature Reserve; ROM MAM 114961, ♂, ROM MAM 114962, ♀, ROM MAM 114963, ♀, ROM MAM 114964, ♀, ROM MAM 115009, ♀, ROM MAM 115041, ♀, ROM MAM 115042, ♀, ROM MAM 115047,

♂, from Hunan, Daweishan National Park. The assignments of the ROM specimens to K. furva sp. n. are solely informed from a mitochondrial gene, and thus required further verification based on morphological or nuclear gene data (see Discussion).

Etymology

The proposed English name is ‘Dark woolly bat’. The name refers to the very dark pelage of the new species.

Diagnosis

Kerivoula furva sp. n. has a darkest dorsal pelage among the Asian Kerivoula species, with dorsal hairs that are broadly uniformly coloured throughout their length. Its body size is comparable to that of K. titania and K. hardwickii s. str., however, compared to the former, K. furva sp. n. has a proportionally shorter tibia and a proportionally wider postorbital constriction, whereas compar- ed to K. hardwickii s. str., it has a proportionally lower braincase height. Kerivoula furva sp. n. and K. depressa have skulls that are similar in shape while the former taxon has a slightly larger skull than the latter. Kerivoula furva sp. n. has a generally larger and more flattened skull compared to K. krauensis.

Description

A medium-sized Kerivoula, with a body mass of 4–7 g and a forearm length of 30.9–37.5 mm (Tables 1 and 2). The ear is 14.3–15.2 mm in length (n = 4); it is funnel shaped and has a fold extending from the base of the pinna to the half of the posterior border. The edge of the pinna is darkly pigmented, contrasting strongly with the paler central part (Fig. 4a). The tragus is long and pointed, and curves slightly outwards at the tip. In some individual bats examined, the tip of the tragus is dark while in others the tragus is pale in colour throughout its length. The dorsal pelage varies from black brown to black grey. Individual dorsal hairs are dark brown and are broadly uniform in colour except for the tips, which are even darker (Fig. 4b). The ventral pelage is greyish brown. Individual ventral hairs have distinct dark brown bases; those on the chest and along the middle line of the abdomen have medium brown tips while those along the flanks of the abdomen have pale grey tips (Fig. 4c). Some individual bats have grey-tipped hairs over the entire ventral surface. The base of the thumb swells to form an oval, fleshy pad. The insertion point of the wing membrane is at the base of the first toe.

The skull has a GTL of 14.07–15.45 mm (Table 2). The narial emargination is two-fold longer than its width, while the anterior palatal emargination is approximately the same long as its width (Fig. 5a–b). The sagittal crest is absent. The braincase is broadened and flattened. The basioccipital pits are welldeveloped, and widen anteriorly and taper posteriorly beyond the front one-third of the cochleae (Fig. 5b). The C–M 3 is 5.52–6.11 mm (Table 2; see Fig. 5c–d for qualitative characters of the upper dentition). The first upper incisor (I 2) has a height approximately two-thirds that of the upper canine. The second upper incisor (I 3) has a height half that of I 2, while the crown areas of these two are about the same. The upper canine has a height surpassing that of the other teeth in the upper toothrow, and has a length that is approximately equal to its width. The first upper premolar (P 2) and the second upper premolar (P 3) are both slightly shorter than their respective widths; the latter tooth has a height and a crown area approximately 80% and 90% the respective dimensions of the former tooth. The third upper premolar (P 4) has a length along its labial bor- der longer than that along its lingual border, making this tooth somewhat triangular in shape. P 4 has its height and crown area 30–40% and 70% larger than respective dimensions of P 2; P 4 has its height twothirds that of the upper canine while having its crown area 10% larger than that of the latter tooth. The first and second upper molars (M 1 and M 2, respectively) both have typical W-shaped cusp structure, with para-, meso- and metastyle all well developed. The third upper molar (M 3) has reduced mesostyle and metacone and has the metastyle absent.

The mandible has a MDL of 9.59–10.86 mm and has a C–M 3 of 5.86–6.44 mm (Table 2; see Fig. 5e–f for qualitative characters of the lower dentition). The first and second lower incisors (I 1 and I 2, respectively) are tricuspid. The third lower incisor (I 3) is unicuspid, whilst it has a large cingular cusp on the lingual posterior border in addition to another smaller one on the labial posterior border. The lower canine has a large cingular cusp on its lingual anterior border, which is higher than I 3. The first, second and third lower premolars (P 2, P 3 and P 4, respectively) are similar to each other in both height and crown area except for P 3, which has a comparatively smaller (90%) crown area. The three lower premolars have heights of ca. 75% of the lower canine. The first and second lower molar (M 1 and M 2, respectively) both have well developed talonids which have widths exceeding those of the trigonids of corresponding teeth. The third lower molar (M 3) has a reduced talonid whose width is smaller than that of the trigonid of the same tooth.

Comparisons

While Wu et al. (2012) referred to K. furva sp. n. as K. titania, the former species is distinct in having a proportionally shorter tibia. In K. furva sp. n. the ratio of the tibia to the forearm (TIB /FA) ranged 0.48–0.51 (n =13) for museum specimens and 0.48–0.55 (n =159) for live individuals caught in the field. In comparison, the reported TIB /FA for K. titania ranges 0.55–0.57 (n =6) for museum specimens. In the cranium, K. furva sp. n. has a proportionally larger POC than K. titania (Fig. 2 and Fig. 6 bottom row).

External measurements include forearm length (FA) and length of tibia (TIB); craniodental measurements include greatest length of skull (GTL), condylobasal length (CBL), condylocanine length (CCL), zygomatic breadth (ZB), greatest width of the braincase (GBB), braincase height (BH), postorbital constriction (POC), maxillary toothrow length (C–M 3), posterior palatal width (M 3 –M 3), mandible length (MDL) and mandibular toothrow length (C–M); a — Data from Miller (1906 a); b — with an ambiguous reclassification in DAPC (Fig. 3); c — data from Tate (1941); d — data from Francis et 3 al. (2007); e — data from Douangboubpha et al. (2016); f — data from Struebig et al. (2016)

Fig. 5. The skull of K. furva sp. n. (paratype, ESRI B1398), showing (a–b) dorsal and basal views of the cranium, (c) occlusal view of the left upper toothrow, (d–e) lateral views of the cranium and the mandible, and (f) occlusal view of the left lower toothrow.

Scale bars = 5 mm

Among phylogenetically close relatives (i.e., K. hardwickii complex), K. kachinensis is a diagnostically larger-sized species compared to K. furva sp. n. (Fig. 2 and Table 2). Kerivoula hardwickii s. str. and K. furva sp. n. broadly overlap in both body and skull size (Fig. 2 and Table 2), while the former species has a proportionally larger BH than the latter one (Figs. 2–3 and Fig. 6 middle row). Kerivoula depressa has a smaller skull size compared to K. furva sp. n. (Fig. 2 and Table 2). Kerivoula krauensis is smaller than K. furva sp. n. in general, with more extreme differences between the two on TIB and on lengths of the skull (Table 2). In addition, the braincase of K. krauensis is domed like that of K. hardwickii s. str. compared to the flat skull of K. furva sp. n. The ratio of BH to C–M 3 in K. furva sp. n. is smaller than that of K. hardwickii s. str. (Fig. 6 middle right panel) and, based on a measurement reported by Douangboubpha et al. (2016), is much smaller than that of K. krauensis (1.10).

The dorsal pelage of K. furva sp. n. is clearly distinguishable from that of all congeneric species described above. Kerivoula hardwickii s. str. was described to have “dark bases of the fur on the upperparts …with the majority of the hair pale grey or brown” (Francis et al., 2007: 5–6), while K. depressa was described as having “upperparts between buff and cream-buff” (Miller, 1906 a: 65). The dorsal hairs of K. kachinensis have darker grey bases and paler grey-brown upper segments (Bates et al., 2004; Soisook et al., 2007), those of K. titania have light grey middle segments above black bases (Bates et al., 2007), and those of K. krauensis are dark brown with shiny golden-brown tips (Francis et al., 2007; Struebig et al., In press). In comparison, K. furva sp. n. has a diagnostically darker dorsal pelage that shows virtually no colour distinction between the root and the middle section of individual hairs.

Echolocation

Kerivoula furva sp. n. produced echolocation calls in groups consisting of 4.4–9.8 pulses and with repetition rates among bouts of 5.0–9.3 Hz 0.66

0.62

0.58

0.54

fur T fur C dep har tita

Myanmar boundary, dep for K. depressa, har for K. hardwickii and tita for K. titania

(Table 3). Within bouts, individual pulses had peak frequencies at 141.8–163.3 kHz, sweeping down from 182.4–222.1 (maximum frequency) to 105.7– 118.3 (minimum frequency) kHz, and were characterized by durations of 1.4–2.0 ms and repetition rates of 50–83.8 Hz. Further details of parameters for echolocation calls are presented in Table 3. In general, K. furva sp. n. emitted individual calls that were typical to its genus (see Kingston et al., 1999; Schmieder et al., 2012), characterized by very broad bandwidths (69.4–103.9 kHz), extremely high frequencies and very high repetition rates, all of which are considered to be specializations for foraging in cluttered environments such as dense vegetation (also see Siemers and Schnitzler, 2000, 2004; Schnitzler and Kalko, 2001).

Distribution and ecological notes

Currently known from Taiwan, Meghalaya in India, Kachin in Myanmar (this study), Hainan of SE China (Wu et al., 2012) and probably also from Hunan and Guangxi of SE China (Fig. 1). Kerivoula furva sp. n. has a wide range in Taiwan where it occurs in montane and hilly areas with elevations of <1800 m a.s.l. (Kuo et al., 2014). In such areas, K. furva sp. n. has been found in broadleaf evergreen forests with low to medium-level disturbance, including forests mixed with betel palm plantations, banana orchards or bamboo stands. In Myanmar, two individuals of K. furva sp. n. were captured in a lowland degraded forest at a foothill. This forest mainly comprises bamboo and banana with some secondary growth trees, and is located between a village and a primary evergreen forest.

By analysing arthropod fragments in feces, Chiang (2006) reported the diet of K. furva sp. n. to be composed principally of spiders (order Araneae), followed by cockroaches (Blattodea). Chiang (2006) suggested K. furva sp. n. as a gleaner given that its diet comprised arthropods with no or poor flying ability. Liao (2013) demonstrated in a behavioural experiment that K. furva sp. n. can perform aerial hawking in addition to glean prey from the substrate with the interfemoral membrane. Liao’s (2013) second experiment indicated that K. furva sp. n. can hunt using prey-generated sound. It was reported that K. furva sp. n. roosted in bamboo internodes or in furled young leaves of banana plants (Chang et al., 2010; Cheng et al., 2010), with individuals roosting either solitarily or in small colonies composed of two to ten bats. Little is known of their reproductive phenology, however, females were recorded as pregnant from mid-April to mid-May (n = 8) and lactating from mid-May to early July (n = 27) (H. C. Kuo, personal observation).

Genetics

Conventional, G-banded and C-banded karyotypes were reported in Wu et al. (2012) for Hainanese and Taiwanese specimens, showing diploid chromosome number (2n) = 32 and fundamental number (FN) = 51–52. The diploid chromosome number differed from those of K. hardwickii, K. intermedia, K. lanosa, K. lenis, K. minuta and K. papillosa (reviewed in Wu et al., 2012). Phylogenetic analyses conducted in this study based on partial sequences of the mitochondrial COI gene and of the nuclear RAG2 gene consistently recovered close relationships among K. furva sp. n., K. hardwickii s. lato and K. kachinensis, to the exclusion of other congeneric species (Fig. 1). Kerivoula krauensis may also belong to this complex based on the COI -based phylogeny. Recently, Kuo et al. (2014) investigated the phylogeography of K. furva sp. n. across its range in Taiwan. Based on genotyping individuals at nine autosomal microsatellite loci, the authors found strong genetic differentiation among populations across the island, with evidence of restricted East-West gene flow, indicating that the Central Mountain Range acts as a barrier to movement. No such p hylogeographic structure was observed when studying haplotypes of the COI gene, although this appears to reflect this species’ recent colonization history on Taiwan (Kuo et al., 2014).