Variation in Aselliscus stoliczkanus based on morphology and molecular sequence data, with a new record of the genus Aselliscus in China

Transcription

1 Journal of Mammalogy, 97(6): , 2016 DOI: /jmammal/gyw138 Published online September 21, 2016 Variation in Aselliscus stoliczkanus based on morphology and molecular sequence data, with a new record of the genus Aselliscus in China Zong-Xiao Zhang, # Yan-Mei Wang, # Liu-Meng Zheng, Jie Wu, Yi-Fu Zhao, Yan-Zhen Bu, and Hong-Xing Niu* College of Life Science, Henan Normal University, No. 46, Jianshe Road, Muye District, Xinxiang, China (Z-XZ, Y-MW, L-MZ, JW, Y-FZ, Y-ZB, H-XN) * Correspondent: hongxingniu@htu.cn # These authors contributed equally to this work. The taxonomic status of the subspecies of Aselliscus stoliczkanus (Mammalia: Chiroptera: Hipposideridae) in China has yet to reach a consensus. To explore the variation and differentiation among populations of A. stoliczkanus from different regions of China, we conducted a series of bat surveys from 2012 to We performed multivariate morphological analyses using 16 external and 18 skull measurements and compared the sequence divergences of 2 mitochondrial genes (cytochrome b [Cytb] and cytochrome oxidase I [COI]). There were significant differences in external and skull measurements, particularly in the upper and lower canines, in the specimens from Guangxi compared with those from southwestern Yunnan and Guizhou. Sequence variation of the Cytb and COI genes was % and %, respectively, which represents differentiation at the species level. The samples from Guangxi were clustered with A. dongbacana from Vietnam in the phylogenetic trees based on COI sequences, and their sequence divergence was less than 3.2%. Based on the morphological and molecular results, we inferred that the samples from Guangxi were A. dongbacana, which represents a new record in China. Skull measurements of A. stoliczkanus from southwestern Yunnan exhibited significant morphological differences compared with those from Anlong County of Guizhou, and the sequence divergences of the Cytb and COI genes were % and %, respectively. Therefore, A. stoliczkanus from southwestern Yunnan and Anlong County of Guizhou might represent different subspecies. Finally, the phylogenetic analyses and minimal Cytb divergences of % indicated that specimens from southwestern and central Yunnan should be considered the same subspecies. Samples from Anlong County of Guizhou, Sichuan, and South Yunnan, with low Cytb divergences of %, represent another subspecies. A greater COI gene divergence was found between Libo and Anlong County of Guizhou; thus, the samples from Libo might represent yet another subspecies. Key words: China, COI, Cytb, morphology, Aselliscus, subspecies, new record Stoliczka s trident bat, Aselliscus stoliczkanus (Dobson 1871), primarily occurs in Southeast Asia including southeastern China, Myanmar, Thailand, Laos, Vietnam, and the Malay Peninsula (Corbet and Hill 1992; Bates et al. 2000). Originally, A. stoliczkanus was included in the genus Asellia (Dobson 1871). Later, Tate (1941) placed A. stoliczkanus in the genus Aselliscus, and this classification was followed by subsequent reviewers (Sanborn 1952). Based on molecular and morphological data, Tu et al. (2015) regarded A. stoliczkanus of northeastern Vietnam as a new species, Aselliscus dongbacana. A. dongbacana is currently known only from the karst areas of North Vietnam including Khau Ca, Cao Bang, Ba Be, Na Hang, and Hun Lien provinces, and it can be distinguished by its long, robust upper and lower canines (Tu et al. 2015). In China, A. stoliczkanus is primarily distributed in Yunnan, Guizhou, Guangxi, Hunan, Southwest Jiangxi, and Sichuan provinces (Wang 2003; Pan et al. 2007; Smith and Xie 2009). To date, no national comparisons of the morphological and molecular data for A. stoliczkanus are available. The taxonomic status of the subspecies of A. stoliczkanus is still controversial and uncertain. Luo (1993) and Smith and Xie (2009) suggested that A. stoliczkanus was a monotypic species, and no subspecific differentiation occurs in China. Li et al. (2007) compared the Cytb and NDI sequences, diet, and echolocation frequencies of 2016 American Society of Mammalogists,

2 ZHANG ET AL. VARIATION IN ASELLISCUS STOLICZKANUS FROM CHINA Materials and Methods Sample collection. From July 2012 to August 2015, we conducted a series of bat surveys in 9 provinces of China, and A. stoliczkanus was only found in Yunnan, Guizhou, and Guangxi provinces (Fig. 1). A total of 168 specimens of A. stoliczkanus was captured. All of the specimens were adult bats. Bats were captured either using hand nets in roosts or, on occasion, using harp traps or mist nets at the entrances of caves. Each captured bat was carefully removed from the trap or net and placed individually in a cotton bag. Three-millimeter wing membrane biopsy punches were taken and preserved in a 95% alcohol solution for subsequent extraction of genomic DNA (Worthington and Barratt 1996). Of the 168 bats captured, 42 were euthanized by cervical dislocation and brought back to the laboratory for cranial samples, and the rest of the captured bats were released immediately after taking biopsies and measurements. The specimens were subsequently preserved in 95% ethanol and deposited in Henan Normal University, Henan Province, China. All fieldwork was conducted in accordance with the Law of the People s Republic of China on the Protection of Wildlife. All experiments on living bats followed the American Society of Mammalogists guidelines for the use of wild mammals in research (Sikes et al. 2011). Collection sites for A. stoliczkanus are presented in Table 1. External and skull measurements. Digital calipers (accuracy ± 0.01 mm) were used to measure 16 external and 18 skull variables (Bates and Harrison 1997; Pan et al. 2007; Douangboubpha et al. 2010). The following external measurements were included: HB: head and body length from the tip of the snout to the anus, ventrally; TL: tail length from the tip of the tail to the base adjacent to the anus; EL: ear length from the lower border of the external auditory meatus to the tip of the pinna; FA: forearm length from the extremity of the elbow to the extremity of the carpus with the wings folded; TIB: tibia length from the knee joint to the ankle; HF: hindfoot length from the extremity of the heel behind the os calcis to the extremity of the longest digit, not including the hairs or claws; CL: claw length from the extremity of the longest digit to the tip of the claw; 3MT, 4MT, and 5MT: length of the 3rd, 4th, and 5th metacarpals, respectively, from the extremity of the carpus to the distal extremities of the metacarpals; and 3D1P and 3D2P, 4D1P and 4D2P, and 5D1P and 5D2P: length of the 1st and 2nd phalanges of the 3rd, 4th, and 5th metacarpals, respectively (from the proximal to the distal extremities of the phalanges). Fig. 1. Sample locations used in this study. Numbers in parentheses refer to the number of sampled individuals. 1. Baoshan, Yunnan; 2. Lincang, Yunnan; 3. Dali, Yunnan; 4. Kunming, Yunnan (Li et al. 2007); 5. Sichuan (Li et al. 2007); 6. South Yunnan (Sun et al. 2009); 7. Anlong, Guizhou (Li et al. 2007); 8. Anlong, Guizhou; 9. Libo, Guizhou (Tu et al. 2015); 10. Chongzuo, Guangxi; 11. Tuyen Quang, Vietnam (Tu et al. 2015); 12. Lang Son, Vietnam (Tu et al. 2015). GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. A. stoliczkanus specimens collected in Yunnan, Guizhou, and Sichuan provinces, and suggested that these samples might represent different geographic populations rather than distinct subspecies. This view was followed by Sun et al. (2009). However, there is a lack of supporting morphological data. Therefore, further studies regarding the classification of A. stoliczkanus are needed. To describe the variation in A. stoliczkanus from different areas, we conducted a series of bat surveys in 9 provinces of China from 2012 to A. stoliczkanus were captured from 5 locations in 3 provinces. In this study, the taxonomic status of A. stoliczkanus from different regions of China was examined in detail using a combined analysis of external and cranial characteristics and molecular sequences. 1719

3 1720 JOURNAL OF MAMMALOGY Table 1. Collection localities of bats in China, , with their corresponding GenBank accession numbers. GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. Locality Geographical coordinates The following cranial measurements were included: GSL: greatest skull length the greatest anteroposterior diameter of the skull between the most projected points at each extremity regardless of the structure that formed these points; CBL: condylo-basal length from the exoccipital condyle to the alveolus of the anterior incisor; CH: cranial height the distance between the superior border of the braincase and the inferior orbital rim of the auditory bulla; RW: rostral chamber width the greatest width across the rostral chambers; MW: mastoid width the greatest distance across the mastoid region; ZW: zygomatic width the greatest width of the skull across the zygomata; ABG: greatest width of auditory bullas the maximal distance across the auditory bullas; CCL: condylo-canine length the distance from the exoccipital condyle to the anterior alveolus of the canine; C 1 : upper toothrow length the distance from the front of the upper canine to the posterior of the upper 3rd molar; M 3 : upper M 3 (outer) the distance from the upper 3rd molar to a 3rd molar from outer basal face; C 1 : lower toothrow length the distance from the front of the lower canine to the posterior of the lower 3rd molar; ML: mandible length the distance from the most posterior part of the condyle to the most anterior part of the mandible, not including the lower incisors; OCG: greatest width of the occipital condyles the greatest distance across the occipital condyles; ONL: occipitonasal length the distance from the nasal anterior to the posterior of the exoccipital condyle; PL: palatal length the distance from the alveolus of the anterior incisor to the posterior of the palate bone; MAW: width of mandible arthrosis the distance from the tip of the coronoid process to the gonion laterale of the mandible; LUC: length of the upper canine the distance from the root of the upper canine to the tip of the upper canine; and LLC: length of the lower canine the distance from the root of the lower canine to the tip of the lower canine. DNA extraction and sequencing. Total genomic DNA from each individual was extracted using a UNIQ-10 Column Animal Genomic DNA Isolation Kit (Sangon, China). The cytochrome b (Cytb) genes of 19 samples were amplified using the primers L14724 (5 -CGAAGCTTGATATGAAAAACCATCGTTG-3 ) and H15915 (5 -AACTGCAGTCATCTCCGGTTTACAAGAC- 3 Irwin et al. 1991). The cytochrome oxidase I (COI) Cytb (1,106 bp) Haplotype no. (number of the sequenced samples) Accession no. COI (567 bp) Haplotype no. (number of the sequenced samples) Accession no. GX Chongzuo City N, E Hap1 (5) Ku Hap1 (6) Ku Hap2 (1) Ku SWYN Baoshan City N, E Hap3 (2) Ku Hap2 (2) Ku Dali City N, E Hap4 (3) Ku Hap3 (3) Ku Lincang City N, E Hap5 (1) Ku Hap4 (2) Ku Hap6 (2) Ku Hap5 (2) Ku Hap7 (1) Ku GZA Anlong County N, E Hap8 (1) KX Hap6 (2) KX Hap9 (1) KX Hap7 (2) KX Hap10 (2) KX genes of 19 samples were amplified using the forward primer (5 -TTCTCAACCAACCACAAAGACATTGG-3 ) and the reverse primer (5 -TAGACTTCTGGGTGGCCAAAGAATCA-3 ) according to Francis et al. (2010). PCR reactions were performed in 25-µl volumes. The cycling parameters for the Cytb gene were as follows: 94 C for 5 min; 94 C for 45 s, 47 C for 45 s and 72 C for 90 s for 35 cycles; and a final extension at 72 C for 10 min. Similar cycling parameters were used to amplify the COI gene. All PCR products were purified using an EZ-10 Spin Column DNA Gel Extraction Kit and were sequenced by Life Technologies, China. Statistical analysis. Statistical analyses were run in SPSS 20.0; the 16 external and 18 cranial variables of A. stoliczkanus from the 3 provinces are represented as the mean ± SD in Tables 2 and 3. One-way analysis of variance (ANOVA) was used to analyze the variation between males and females. Discriminant function analysis was used to assess the morphometric differences among specimens of A. stoliczkanus from the 3 provinces. The stepwise discriminant function procedure was applied to all variables using the Wilks lambda minimization procedure to determine which variable provided the greatest discrimination. In this study, Wilks lambda was performed in a multivariate setting with a combination of dependent variables. The criterion chosen for this method was the F-value, with an F-toenter = 3.84 and an F-to-remove = Eigenvalues, canonical correlations, Wilks lambda, and unstandardized coefficients were computed. The group to which a specimen belonged was determined based on the Mahalanobis distance. A specimen was included in the group whose centroid was the closest. The discriminant accuracy was then calculated. Original and cross-validation were used to test the stability of the discriminant functions. Principal component analysis (PCA) was used to create new variables which explain as much of the data as possible. PCA was performed using a correlation matrix; all measurements were standardized prior to the analysis. Sequence and phylogenetic analyses. The sequences were aligned using ClustalX v1.8 (Thompson et al. 1997) and then carefully examined and manually edited. The numbers of haplotypes and polymorphic sites were estimated using DnaSP v5 (Librado and Rozas 2009). These sequences were used in this study together with the COI sequences of A. dongbacana

4 ZHANG ET AL. VARIATION IN ASELLISCUS STOLICZKANUS FROM CHINA 1721 Table 2. External measurements of Aselliscus (mean ± SD, length in mm) collected in China, D1P = length of 1st phalanx of 3rd digit; 3D2P = length of 2nd phalanx of 3rd digit; 4D1P = length of 1st phalanx of 4th digit; 4D2P = length of 2nd phalanx of 4th digit; 5D1P = length of 1st phalanx of 5th digit; 5D2P = length of 2nd phalanx of 5th digit; 3MT = length of 3rd metacarpal; 4MT = length of 4th metacarpal; 5MT = length of 5th metacarpal; CL = claw length; EL = ear length; FA = forearm length; HB = head and body length; HF = length of hind-foot; TIB = tibia length; TL = tail length. GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. Character A. stoliczkanus A. dongbacana (holotype) GX (23) SWYN (113) GZA (32) Bac Kan Province (11) (12) (57) (56) (14) (18) (1) HB ± ± ± ± ± ± TL ± ± ± ± ± ± EL ± ± ± ± ± ± FA ± ± ± ± ± ± TIB ± ± ± ± ± ± 0.58 HF 5.80 ± ± ± ± ± ± 0.34 CL 2.24 ± ± ± ± ± ± MT ± ± ± ± ± ± D1P ± ± ± ± ± ± D2P ± ± ± ± ± ± MT ± ± ± ± ± ± D1P ± ± ± ± ± ± D2P 9.99 ± ± ± ± ± ± MT ± ± ± ± ± ± D1P ± ± ± ± ± ± D2P 9.87 ± ± ± ± ± ± 0.35 Table 3. Cranial measurements of Aselliscus (mean ± SD, length in mm) collected in China, ABG = greatest width of auditory bullae; C 1 = lower toothrow length; C 1 = upper toothrow length; CBL = condylo-basal length; CCL = condylo-canine length; CH = cranial height; GSL = greatest skull length; LLC = length of lower canine; LUC = length of upper canine; M 3 = Upper M 3 (outer); MAW = width of mandible arthrosis; ML = mandible length; MW = mastoid width; OCG = greatest width of the occipital condyles; ONL = occipitonasal length; PL = palatal length; RW = rostral camber width; ZW = zygomatic width. GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. Character A. stoliczkanus A. dongbacana (holotype) GX (10) SWYN (21) GZA (11) Bac Kan Province (6) (4) (10) (11) (5) (6) (1) GSL ± ± ± ± ± ± CBL ± ± ± ± ± ± CH 5.92 ± ± ± ± ± ± 0.17 RW 4.64 ± ± ± ± ± ± 0.16 MW 7.25 ± ± ± ± ± ± ZW 7.68 ± ± ± ± ± ± ABG 5.96 ± ± ± ± ± ± 0.05 CCL ± ± ± ± ± ± C ± ± ± ± ± ± M ± ± ± ± ± ± C ± ± ± ± ± ± ML 9.60 ± ± ± ± ± ± OCG 4.49 ± ± ± ± ± ± 0.10 ONL ± ± ± ± ± ± 0.21 PL 4.92 ± ± ± ± ± ± 0.12 MAW 3.10 ± ± ± ± ± ± 0.10 LUC 1.93 ± ± ± ± ± ± LLC 1.49 ± ± ± ± ± ± from Vietnam (HM540152; HM540158) and Libo County of Guizhou Province (JF443870; JQ600013) and the Cytb sequences of A. stoliczkanus from Kunming City of Yunnan Province (DQ888668), South Yunnan Province (EU434954), Anlong County of Guizhou Province (DQ888676), and Sichuan Province (DQ888673). Four species (Hipposideros pratti, Hipposideros armiger, Hipposideros pomona, and Coelops frithii) from GenBank were selected as outgroups for the Cytb (DQ297584, DQ297585, DQ888671, and DQ888674, respectively) and the COI gene (HM540611, HM540326, JF443930,

5 1722 JOURNAL OF MAMMALOGY and HQ918409, respectively). Sequence divergences were calculated with the MEGA 5.0 program (Tamura et al. 2011) using the Kimura 2-parameter model. The program PAUP4.0*b10 (Swofford 2002) was used to construct a phylogenetic tree using the maximum parsimony method. Modeltest version 3.7 (Posada and Crandall 1998) was used to estimate the best-fit evolutionary model of the nucleotide substitutions for the 2 mitochondrial genes. The TVM + I models were selected by Akaike information criterion (AIC) in Modeltest 3.7 for Cytb to fit the concatenated data sets of the Cytb gene sequences (base frequencies: A = , C = , G = , and T = ; rate matrix: A C = , A G = , A T = , C G = , C T = , and G T = ), and the models TrN + I were selected by AIC in Modeltest 3.7 for COI to fit the concatenated data sets of the COI gene sequences (base frequencies: A = , C = , G = , and T = ; rate matrix: A C = , A G = , A T = , C G =1.0000, C T = , and G T=1.0000). The tree topology was evaluated using 1,000 replicates for the MP analysis. Results External measurements. No external or cranial measurements showed significant differences between the sexes (1-way ANOVA, all P > 0.05). Therefore, we did not evaluate sexual dimorphism and pooled data for males and females. Discriminant analysis of the external variables showed that significant differences existed among the samples from Guangxi Province (GX), southwestern Yunnan Province (SWYN), and Anlong County of Guizhou Province (GZA). In the stepwise analysis of all measurements, 10 measurements (HB, EL, FA, CL, 3D2P, 4MT, 4D1P, 5MP, 5D1P, and 5D2P) contributed to the greatest discrimination between the examined groups (Table 4). Individuals were separated into 3 groups (Fig. 2a). Statistically significant differences between the designated species group centroids were evaluated using squared Mahalanobis distances (D 2 ). Of the 113 samples, 93 (82.3%) of the samples from Dali, Baoshan, and Lincang in southwestern Yunnan Province were correctly assigned to the SWYN group, 93.8% (30 of 32) of samples from Anlong County of Guizhou Province were correctly assigned to the GZA group, and 95.7% (22 of 23) of samples from Guangxi Province were correctly assigned to the GX group. The cross-validation was 78.8%, 90.6%, and 91.3%, respectively. The first 3 principal components (PC1, PC2, and PC3) based on 16 external measurements explained 51.28% of the total variation (Table 5). The first component (PC1: 29.67%) included 9 of the 16 variables with high positive values ( ; TL, 3MT, 3D2P, 4MT, 4D1P, 4D2P, 5MT, 5D1P, and 5D2P). The second component (PC2: 13.81%) included a set of 4 variables with positive values (> 0.440; HB, FA, TIB, and HF). The third component (PC3: 7.79%) revealed 1 variable with a positive value (> 0.435; EL). Samples from GX could be separated somewhat from those from SWYN and GZA; however, specimens from SWYN and GZA cannot be effectively separated on the basis of external measurements (Fig. 3a). Cranial measurements. Five measurements (RW, ABG, C 1, ONL, PL, and LLC) were selected in the stepwise Table 4. Results of the stepwise discriminant analysis for samples of Aselliscus collected in China, D2P = length of 2nd phalanx of 3rd digit; 4D1P = length of 1st phalanx of 4th digit; 5D1P = length of 1st phalanx of 5th digit; 5D2P = length of 2nd phalanx of 5th digit; 4MT = length of 4th metacarpal; 5MT = length of 5th metacarpal; ABG = greatest width of auditory bullae; C 1 = lower toothrow length; CL = claw length; EL = ear length; FA = forearm length; HB = head and body length; LLC = length of lower canine; ONL = occipitonasal length; PL = palatal length; RW = rostral camber width. External measurements Canonical discriminant function coefficients Cranial measurements Canonical discriminant function coefficients Character F-to-remove F1 F2 Character F-to-remove F1 F2 HB RW EL ABG FA C CL ONL D2P PL MT LLC D1P MP D1P D2P Constant Eigenvalue Cumulative % Canonical correlation Wilks λ χ d.f P

6 ZHANG ET AL. VARIATION IN ASELLISCUS STOLICZKANUS FROM CHINA 1723 Fig. 2. Scatter plots of canonical DFs for a) external and b) cranial measurements of Aselliscus stoliczkanus sampled in China, DF = discriminant function; GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. Table 5. Variable loadings for the 3 principal components (PC1, PC2, and PC3) from an analysis of external and cranial characters of Aselliscus stoliczkanus collected in China, Values in bold indicate high loading on that particular principal component. 3D1P = length of 1st phalanx of 3rd digit; 3D2P = length of 2nd phalanx of 3rd digit; 4D1P = length of 1st phalanx of 4th digit; 4D2P = length of 2nd phalanx of 4th digit; 5D1P = length of 1st phalanx of 5th digit; 5D2P = length of 2nd phalanx of 5th digit; 3MT = length of 3rd metacarpal; 4MT = length of 4th metacarpal; 5MT = length of 5th metacarpal; ABG = greatest width of auditory bullas; C 1 = lower toothrow length; C 1 = upper toothrow length; CBL = condylo-basal length; CCL = condylo-canine length; CH = cranial height; CL = claw length; EL = ear length; FA = forearm length; GSL = greatest skull length; HB = head and body length; HF = length of hind-foot; LLC = length of the lower canine; LUC = length of the upper canine; M 3 = Upper M 3 (outer); MAW = width of mandible arthrosis; ML = mandible length; MW = mastoid width; OCG = greatest width of the occipital condyles; ONL = occipitonasal length; PL = palatal length; RW = rostral chamber width; TIB = tibia length; TL = tail length; ZW = zygomatic width. External measurements Loadings Cranial measurements Loadings Character PC1 PC2 PC3 Character PC1 PC2 PC3 HB GSL TL CBL EL CH FA RW TIB MW HF ZW CL ABG MT CCL D1P C D2P M MT C D1P ML D2P OCG MT ONL D1P PL D2P MAW LUC LLC Eigenvalues Eigenvalues % total variance % total variance % cumulative total variance % cumulative total variance analysis of all measurements (Table 4). Samples from GX, GZA, and SWYN showed significant differences and were separated into 3 groups (Fig. 2b). All (n = 10) of the samples from Guangxi Province were correctly assigned to the GX group. Of 21 samples from southwestern Yunnan Province, 20 (92.5%) were correctly assigned to the SWYN group; 100% (n = 11) of the samples from Anlong County of Guizhou Province were correctly assigned to the GZA group. Similarly, the cross-validation was 100%. The first 3 PCs explained 57.31% of the total variation, explaining 36.47%, 12.27%, and 8.57% of the total variation, respectively (Table 5). For PC1, 11 variables showed a positive loading ( ; GSL, CBL, MW, ABG, CCL, C 1, ML, OCG, MAW, LUC, and LLC); ZW, ONL, and PL showed a positive loading (> 0.462) for PC2; and CCL and MAW showed a positive loading (> 0.478) for PC3. PC1 and PC2 clearly distinguished the 3 collection sites (Fig. 3b).

7 1724 JOURNAL OF MAMMALOGY Molecular data analysis. The Cytb sequences had a length of 1,106 bp. Ten different haplotypes (Hap1 Hap10) were observed in the 19 sequenced samples collected from the 3 provinces. Six sequenced samples from GX shared 2 haplotypes (Hap1 Hap2), 9 sequenced samples from Baoshan, Dali, and Lincang in SWYN shared 5 haplotypes (Hap3 Hap7), and 4 sequenced samples from GZA shared 3 haplotypes (Hap8 Hap10; Table 1). The sequence divergences of A. stoliczkanus among the 3 sampled sites in SWYN ranged from 0.2% to 0.9%, the divergences among the 3 haplotypes from GZA ranged from 0.1% to 0.7%, and the divergence of the 2 haplotypes from GX was 0.6%. The divergences between the haplotypes from SWYN and GZA were %, the divergences between the haplotypes from GZA and GX were %, and the divergences between the haplotypes from GX and SWYN were % (Table 6). The COI sequences were 567 bp long. Four (Hap2 Hap5) and 2 (Hap6 Hap7) different haplotypes were observed among SWYN and GZA samples, respectively, and the samples from GX shared 1 haplotype (Hap1). The sequence divergences among the A. stoliczkanus specimens from the 3 sampled sites in SWYN were %, and the divergences of the 2 haplotypes from GZA was 0.3%. The divergences between SWYN and GZA were %, the divergences between SWYN and GX were %, and the divergence between GZA and GX was 14.0% (Table 7). Phylogenetic reconstruction. The MP trees constructed based on the Cytb and COI genes were similar in topology. Two highly divergent clades, named A and B, can be distinguished Fig. 3. Scatter plots of principal component analysis based on PC1 and PC2 for a) external and b) cranial measurements of Aselliscus stoliczkanus sampled in China, GX = Guangxi Province; GZA = Anlong County of Guizhou Province; PC = principal component; SWYN = southwestern Yunnan Province. Table 6. Sequence divergence matrix based on complete mitochondrial Cytb (1106 bp) gene sequences of Aselliscus stoliczkanus in China and outgroups (Coelops frithii, Hipposideros armiger, H. pratti, and H. pomona). MEGA 5 was used to calculate the genetic distances based on the Kimura 2-parameter model. Asterisk indicated the sequences obtained from GenBank. 1 2 = A. stoliczkanus, Chongzuo, Guangxi; 3 = A. stoliczkanus, Dali, Yunnan; 4 = A. stoliczkanus, Baoshan, Yunnan; 5 7 = A. stoliczkanus, Lincang, Yunnan; 8 = A. stoliczkanus, Kunming, Yunnan; 9 12 = A. stoliczkanus, Anlong, Guizhou; 13 = A. stoliczkanus, Sichuan; 14 = A. stoliczkanus, South Yunnan; 15 = C. frithii; 16 = H. armiger; 17 = H. pratti; 18 = H. pomona * * 13* 14* 15* 16* 17* 18* * * * * * * * *

8 ZHANG ET AL. VARIATION IN ASELLISCUS STOLICZKANUS FROM CHINA 1725 Table 7. Sequence divergence matrix based on the complete mitochondrial COI (567 bp) gene sequences of Aselliscus in China and outgroups (Coelops frithii, Hipposideros armiger, H. pratti, and H. pomona). MEGA 5 was used to calculate the genetic distances based on the Kimura 2-parameter model. Asterisk indicated the sequences obtained from GenBank. 1 2 = A. dongbacana, Vietnam; 3 = A. stoliczkanus, Chongzuo, Guangxi; 4 = A. stoliczkanus, Dali, Yunnan; 5 6 = A. stoliczkanus, Lincang, Yunnan; 7 = A. stoliczkanus, Baoshan, Yunnan; 8 9 = A. stoliczkanus, Anlong, Guizhou; = A. stoliczkanus, Libo, Guizhou; 12 = C. frithii; 13 = H. armiger; 14 = H. pratti; 15 = H. Pomona. 1* 2* * 11* 12* 13* 14* 15* 1* 2* * * * * * * Fig. 4. Maximum parsimony tree constructed with PAUP4.0*b10 software based on Cytb gene sequences. Hap1 Hap10 are 10 different haplotypes observed in the 3 geographical regions (GX, SWYN, and GZA). Hipposideros pratti, H. armiger, H. pomona, and Coelops frithii were used as the outgroups. The numbers at the nodes are the MP tree posterior probabilities (asterisk indicates a posterior probabilities less than 50). GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. in both the Cytb and COI trees (Figs. 4 and 5). In the tree topologies based on the Cytb genetic data, clade A was formed by 2 haplotypes from GX. In the COI tree, clade A comprised samples of GX and A. dongbacana from northern Vietnam. Based on the levels of genetic divergence in the mtdna sequences, clade B can be further divided into 2 subclades, B1 and B2, on the Cytb and COI trees. Clade B1 comprised samples from Kunming, Dali, Baoshan, and Lincang City in Yunnan Province, and clade B2 comprised samples from Anlong and Libo County in Guizhou Province, Sichuan Province, and southern Yunnan Province. Discussion Animal mtdna is maternally inherited and generally evolves more rapidly than nuclear DNA (Brown et al. 1982; Avise and Ellis 1986). Cytb is the best-known mitochondrial gene with respect to its structure and function. Therefore, sequences of this gene are widely used in phylogenetic studies (Degli Esposti et al. 1993). Bradley and Baker (2001) calculated the Cytb genetic distance of samples from 7 genera of bats and 4 genera of rodents including recognized sister species. The study showed that average genetic distances within subspecies taxa, a single species, and sister taxa were %, %, and 4 11%, respectively. By pairwise comparisons for the sequences of species within a genus, the average genetic distance values ranged from 2.23% to 21.97% with an average of 11.46%. Therefore, Bradley and Baker (2001) suggested that subspecies were recognized with values > 2.34%; values exceeding 11% were thought to indicate species-level divergence. In the present study, the uncorrected divergences between the GX population and the other populations were % and thus reached the species level. For the COI gene, the levels of intraspecific variation in 98% of the animals ranged from 0% to 2%; however, the average value for specieslevel divergence was 11.3% (Hebert et al. 2003). The COI gene

9 1726 JOURNAL OF MAMMALOGY Fig. 5. Maximum parsimony tree constructed with PAUP4.0*b10 software based on COI gene sequences. Hap1 Hap7 are 7 different haplotypes observed in the 3 geographical regions (GX, SWYN, and GZA). Hipposideros pratti, H. armiger, H. pomona, and Coelops frithii were used as the outgroups. The numbers at the nodes are the MP tree posterior probabilities. GX = Guangxi Province; GZA = Anlong County of Guizhou Province; SWYN = southwestern Yunnan Province. divergences between the GX population and the other populations were % and thus also reached the species level. In the present study, stepwise discriminant analysis showed that samples from SWYN, GX, and GZA were separated into 3 groups (Figs. 2a and 2b). PCA indicated that most morphological variables in the GX population were significantly different from the other populations (Figs. 3a and 3b). Therefore, based on these distinct differences in morphological characteristics and molecular data, we propose that the GX population constitutes a distinct species. The external and skull measurements of the GX population were similar to A. dongbacana. Additionally, the GX population and A. dongbacana were clustered in a clade in the MP tree based on the COI gene (Fig. 5). The COI gene divergences between the GX population and A. dongbacana were 3.2%. Therefore, we propose that the samples from Guangxi represent A. dongbacana, and a new record for China. The stepwise discriminant analysis showed that the samples from SWYN and GZA were separated (Figs. 2a and 2b). Although the samples from SWYN and GZA cannot be effectively separated on the basis of external measurements in the scatter plots of the PCA (Fig. 3a), most of the cranial measurements were significantly different (Fig. 3b). The morphological measurements of A. stoliczkanus from GZA were greater than those from SWYN (Tables 2 and 3). The Cytb and COI sequence divergences between the samples from SWYN and GZA ranged from 5.7% to 6.3% and 6.4% to 6.7%, respectively, which reaches the level required for subspecies designation. Based on the morphological and molecular results, we suggest that specimens from GZA and SWYN represent distinct subspecies within A. stoliczkanus in China. This result is consistent with Li et al. (2007) and Sun et al. (2009). In the tree topologies based on the Cytb genetic data, samples from Dali, Baoshan, and Lincang City in southwestern Yunnan Province and Kunming City in central Yunnan Province were clustered in a clade, and their divergences in Cytb ranged from 0.2% to 0.9%. In the tree topologies based on the COI genetic data, samples from Dali, Baoshan, and Lincang City in SWYN were clustered in a clade, and the divergences in COI ranged from 0.5% to 1.4%. The southwestern and central Yunnan populations should belong to the same subspecies. Samples from GZA, Sichuan Province, and southern Yunnan Province were clustered in a clade, and the divergences of Cytb ranged from 0.1% to 0.7%. These samples should also belong to the same subspecies. The COI gene divergences between samples from Anlong and Libo County in Guizhou Province were 2.4% and have reached the subspecies level. Based on these preliminary findings, we propose that samples from Anlong and Libo in Guizhou might represent different geographic races, which are similar to those reported by Tu et al. (2015). Acknowledgments This project was supported by the National Natural Science Foundation of China (NSFC, No ). We thank all those who helped in the field, especially S. Zhao, W. Yang, Y. Zhu, W. Liu, X. He, C. Zhang, H. Zhou, Y. Wang, and X. Sun. We also thank L. Zhao and H. Zhou for tissue sampling. We are especially grateful to Dr. C. Zhou for help with the phylogenetic tree reconstruction. Literature Cited Avise, J. C., and D. Ellis Mitochondrial DNA and the evolutionary genetics of higher animals. Philosophical Transactions of the Royal Society of London, B: Biological Sciences 312: Bates, P. J. J., and D. L. Harrison Bats of the Indian subcontinent. Harrison Zoological Museum, Kent, United Kingdom. Bates, P. J. J., T. Nwe, M. J. Pearch, K. M. Swe, S. S. H. Bu, and T. Tun A review of bat research in Myanmar (Burma) and results of a recent survey. Acta Chiropterologica 2: Bradley, R. D., and R. J. Baker A test of the genetic species concept: cytochrome-b sequences and mammals. Journal of Mammalogy 82: Brown, W. M., M. Pragere, A. Wang, and A. C. Wilson Mitochondrial DNA sequences of primates: tempo and mode of evolution. Journal of Molecular Evolution 18: Corbet, G. B., and J. E. Hill The mammals of the Indomalayan region: a systematic review. Natural History Museum Publication, Oxford, United Kingdom.

10 ZHANG ET AL. VARIATION IN ASELLISCUS STOLICZKANUS FROM CHINA 1727 Degli Esposti, M., D. S. Vries, M. Crimi, A. Ghelli, T. Patarnello, and A. Meter Mitochondrial cytochrome b: evolution and structure of the protein. Biochimica et Biophysica Acta (BBA)- Bioenergetics 1143: Dobson, G. E Description of four new species of Malayan bats, from the collection of Dr. Stoliczka. Journal of the Asiatic Society of Bengal 40: Douangboubpha, B., S. Bumrungsri, P. Soisook, C. Satasook, N. M. Thomas, and P. J. Bates A taxonomic review of the Hipposideros bicolor species complex and H. Pomona (Chiroptera: Hipposideridae) in Thailand. Acta Chiropterologica 12: Francis, C. M., et al The role of DNA barcodes in understanding and conservation of mammal diversity in Southeast Asia. PLoS One 5:1 12. Hebert, P. D. N., S. Ratnasingham, and J. R. Dewaard Barcoding animal life: cytochrome coxidase subunit 1 divergences among closely related species. Proceedings of the Royal Society B Biological Sciences 270:S96 S99. Irwin, D. M., T. D. Kocher, and A. C. Wilson Evolution of the cytochrome b gene of mammals. Journal of molecular evolution 32: Li, G., et al Echolocation calls, diet, and phylogenetic relationships of Stoliczka s trident bat, Aselliscus stoliczkanus (Hipposideridae). Journal of Mammalogy 88: Librado, P., and J. Rozas DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25: Luo, R The mammalian fauna of Guizhou. Guizhou Science and Technology Publishing House, Guizhou, China. Pan, Q. H., Y. X. Wang, and K. Yan A field guide to the mammal of China. China Forestry Publishing House, Beijing, China. Posada, D., and K. A. Crandall Modeltest: testing the model of DNA substitution. Bioinformatics 14: Sanborn, C. C The mammals of the Rush Watkins Zoological Expedition to Siam. Natural History Bulletin of the Siam Society 15:1 20. Sikes, R. S., W. L. Gannon, and The Animal Care and Use Committee of the American Society of Mammalogists Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92: Smith, A. T., and Y. Xie A guide to the mammals of China [in Chinese]. Hunan Education Press, Changsha, China Sun, K., J. Feng, Z. Zhang, L. Xu, and Y. Liu Cryptic diversity in Chinese rhinolophids and hipposiderids (Chiroptera: Rhinolophidae and Hipposideridae). Mammalia 73: Swofford, D. L PAUP* beta version. Phylogenetic analysis using parsimony (* and other methods). Sinauer Associates, Inc., Publishers, Sunderland, Massachusetts. Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and S. Kumar MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28: Tate, G. H. H Results of the Archbold Expedition. No. 36. Remarks on some Old World leaf-nosed bats. American Museum Novitates 1140:1 11. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins The CLUSTAL_X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24: Tu, V. T., et al Description of a new species of the genus Aselliscus (Chiroptera, Hipposideridae) from Vietnam. Acta Chiropterologica 17: Wang, Y. X A complete checklist of mammal species and subspecies in China: a taxonomic and geographic reference. Forestry Publishing House, Beijing, China. Worthington, W. J., and E. Barratt A non-lethal method of tissue sampling for genetic studies of chiropterans. Bat Research News 37:1 3. Submitted 15 May Accepted 2 August Associate Editor was Jorge Ortega.