The specimens collected in this study comprise 3 adult males and 2 adult females, according to the development degree of molars and the degree of ossification of forelimb joints. No pregnant females were found. Voucher specimens are deposited in the Kunming Natural History Museum of Zoology, Kunming Institute of Zoology, the Chinese Academy of Sciences (KIZ, CAS), Kunming, China and registered under the numbers KIZ 20220058, 20220089, 20230357, 20230425 and 20230463.

Following the manufacturer’s protocol, total genomic DNA was extracted from muscle samples using the Ezup Column Animal Genomic DNA Purification Kit (Sangon Biotech, China). The complete sequence of the mitochondrial Cyt b gene was amplified and sequenced with the primer pair LGL765: GAAAAACCAYCGTTGTWATTCAACT and LGL766: GTTTAATAAGAATYTYAGCTTTGGG (Bickham et al. 1995). Polymerase chain reactions (PCR) were carried out in a total volume of 25 μl containing 1 μl of template DNA, 1 μl of each primer at 10 μM, 1 μl of dNTPmix at 10 mM, 0.2 μl of Taq Plus DNA polymerase (Sangon Biotech, China) at 5 U/μl, 2.5 μl of 10x Taq Buffer and added H₂O to 25 μl. The PCR thermal profile was: 1, 95°C 5 min for initial denaturation; 2, 94°C 30 sec for denaturation; 3, 63°C (decreases by 0.5°C per cycle) 30 sec for annealing; 4, 72°C 30 sec for extension; 5, 10 cycles of steps 2 to 4; 6, 95°C 30 sec for denaturation; 7, 58°C 30 sec for annealing; 8, 72°C 30 sec for extension; 9, 30 cycles of steps 6 to 8; 10, 72°C 10 min final extension; 11, 4°C for renaturation. PCR products were detected by agarose gel electrophoresis and purified using the SanPrep Column DNA Gel Extraction Kit (Sangon Biotech, China). Finally, purified samples were sequenced by the ABI 3730XL DNA Analyzer (USA) at Sangon Biotech (Shanghai, China). Sequences were edited and assembled using SeqMan in Lasergene 7.1 (DNASTAR Inc., Madison, WI, USA).

The full-length Cyt b sequences (1,140 bp) were compared with those from the National Center for Biotechnology Information (NCBI). The sequences were aligned using the default parameters of the ClustalW algorithm in the software MEGA11 (Tamura et al. 2021) and the uncorrected P-distances were calculated between pairwise sequences. The pairwise deletion option was used to remove ambiguous positions when calculating genetic distances. The phylogeny of the subfamily Murininae was reconstructed by MEGA11 using the Maximum Likelihood method under a GTR+G+I nucleotide substitution model and the branch support was evaluated by 1,000 bootstrap replicates. ModelFinder (Kalyaanamoorthy et al. 2017) on PhyloSuite v.1.2.2 (Zhang et al. 2020) was used to select the best-fit model (GTR+G+I nucleotide substitution model), based on the Bayesian Information Criterion (BIC) .

The morphological characters of the five Yunnan specimens were compared with those described for Harpiola grisea and H. isodon (Dobson 1878, Thomas 1915, Bhattacharyya 2002, Kuo et al. 2006, Kruskop et al. 2006). We also took external, cranial and dental measurements from our specimens, as described below. Head and body length, tail length, foot length, ear length, forearm length, thumb length, metacarpal lengths and tibia length were measured in the field. The cranial and dental measurements were taken according to Kuo et al. (2006) as follows: total length of skull — from the anterior rim of alveolus of the first upper incisor to the most projecting point of the occipital region; condylobasal length — from the exoccipital condyle to the posterior rim of the alveolus of the first upper incisor; upper canine width— taken across the outer borders of upper canines; upper molar width — taken across the outer crowns of the last upper molars; zygomatic width — the greatest width of the skull across the zygomatic arches; mastoid width — the greatest distance across the mastoid region; postorbital width — the least width of the postorbital constriction; maxillary toothrow length — from the front of the upper canine to the back of the crown of the third molar; upper canine–premolar length — the largest distance from the front of the upper canine to the back of the crown of the posterior premolar; length of mandible — from the anterior rim of the alveolus of the first lower incisor to the most posterior part of the condyle; mandibular toothrow length — from the front of the lower canine to the back of the crown of the third lower molar; lower canine–premolar length — the greatest distance from the front of the lower canine to the back of the crown of the posterior premolar; height of the coronoid process — taken perpendicularly from the extremity of the coronoid process to the ramus mandibulae. Each craniodental measurement was taken three times by Xin Mou with a caliper accurate to 0.01 mm and the average value was reported.

Following the Chinese laws and regulations on the protection of wild terrestrial animals (State Council Decree 1992), the field investigations of small mammals in Guanyin Mountains Provincial Nature Reserve and the collection of specimens were approved by the Conservation Area Management Committee of Guanyin Mountains Provincial Nature Reserve and the Ethics Committee of KIZ, CAS.

The two species of Harpiola, H. isodon and H. grisea, differ from each other in the following characters: the PM⁴ is wider than long in H. isodon (versus as wide as long in H. grisea); both M¹ and M² have the mesostyles present in H. isodon (versus M² lacking mesostyle in H. grisea); M¹ of H. isodon has a post-cingular platform (versus no post-cingular platform in H. grisea) (Kuo et al. 2006). The five Chinese mainland Harpiola specimens showed these dental characters in line with H. isodon. We noted that the Chinese mainland taxon had a range of maxillary toothrow length surpassing that of the Taiwanese H. isodon provided in Kuo et al. (2006) (Table 1). Nevertheless, we should warn that the measurements were taken by different people across the two studies and those taken by the same investigator are warranted in the future for a valid evaluation of the morphometric differences between the two taxa.

We compared the Cyt b sequences of our five specimens with those downloaded from NCBI for 19 Vespertilionids, including Harpiola isodon from Taiwan, Harpiocephalus harpia, 15 Murina, one Kerivoula and one Myotis species (Table 2). The novel sequences are deposited in the NCBI GenBank database under accession numbers PP476123 (KIZ20220058), PP476124 (KIZ20220089), PP476125 (KIZ20230357), PP476126 (KIZ20230425) and PP476127 (KIZ20230463).

Based on the reconstructed phylogenetic tree, the Chinese mainland Harpiola formed a monophyletic group with H. isodon (Fig. 3), which indicates a close relationship between them. However, the two taxa showed genetic distances of 4.8–5.1% between each other, which exceeded the difference between Murina recondita and Murina gracilis (3.7%) and was not much lower than those between Murina suilla and Murina florium (7.7%), Murina gracilis and Murina eleryi (7.7%) and Murina puta and Murina huttoni rubella (7.2%) (Suppl. material 1). Thus, there are substantial sequence differences in the studied uniparentally inherited gene between Taiwanese and mainland Harpiola bats, but the taxonomic interpretation of these differences would be premature and additional material and nuclear genes should be included in subsequent analyses.

Figure 3.  

Maximum Likelihood phylogenetic reconstruction of 25 Vespertilionidae samples using an 1140 bp alignment of the mitochondrial cyt b gene. Bootstrap (BS) values are indicated adjacent to nodes (nodes with BS < 0.50 are not labelled).

With a synthetic consideration of morphological and molecular evidence, we refer to the specimens from the Guanyin Mountains, Yuanyang, Yunnan as the first records of H. isodon on the Chinese mainland.

The specimens were captured at two mountain sites (23.03N, 102.95E and 22.99N, 102.99E) with mist nets. Their locations and the distribution map of Harpiola isodon are shown in Fig. 4. The habitat is a mid-mountain evergreen broad-leaved forest at elevations of 2,381 m and 2,463 m, respectively. The canopy is well closed and the forest is rich in shrubs and has small streams. There are no caves in this area, but there are many large trees with tree holes. We speculate that this insect-eating bat mainly roosts in tree cavities or under the dense canopy during the daytime.

Figure 4.  

Distribution map of Harpiola isodon. Coordinate information is derived from Kuo et al. (2006), Kruskop et al. (2006) and the Global Biodiversity Information Facility (Anonymous 2024).