Chrysopteron Jentink, 1910 is 1 of the 7 subgenera of Myotis Kaup, 1829 recognized by Tate that traditionally comprises Asian and African species characterized by conspicuously parti-colored wing membranes. Definition of Myotis subgenera has long challenged taxonomists and prior to the present study the systematic status of numerous forms within Chrysopteron remained unclear. Following examination of material (including available type specimens) in 21 European, North American, and Asian collections, and using morphological (external, cranial, and dental characters) and genetic data, we evaluate the validity of the Chrysopteron subgenus, revise the taxonomy of the named Asian forms, and review their distinguishing characters, distribution, and taxonomic history. We argue that Chrysopteron is an available name for a monophyletic “Ethiopian clade” recovered with high support in our analyses, which comprises species characterized by striking reddish or yellowish dorsal fur that strongly differentiates them from congeners. We also determine that M. formosus sensu lato contains several morphologically distinct forms, some of which occur in sympatry and some in widely separated localities. A morphological key is provided for all Asian species of Chrysopteron revealed by our study: M. bartelsi Jentink, 1910 (Java and Bali), M. formosus (Hodgson, 1835) (Afghanistan, India, Nepal, China, Taiwan, and Vietnam), M. hermani Thomas, 1923 (Sumatra, Thailand, and Malaysia), M. rufoniger (Tomes, 1858) (Korea, Japan, China, Taiwan, Laos, and Vietnam), M. rufopictus (Waterhouse, 1845) (Philippines), and M. weberi (Jentink, 1890) (Sulawesi).
Chrysopteron Jentink, 1910 is 1 of the 7 subgenera of Myotis Kaup, 1829 recognized by Tate (1941; alongside Selysius Bonaparte, 1841, Isotus Kolenati, 1856, Paramyotis Bianchi, 1916, Myotis, Leuconoe Boie, 1830, and Rickettia Bianchi, 1916) and traditionally comprises Asian and African species characterized by conspicuously parti-colored wing membranes. Originally established by Jentink (1910) for C. bartelsi from Java, several Asian forms with similar dichromatic wing patterns (Vespertilio formosa Hodgson, 1835; V. rufopictus Waterhouse, 1845; V. rufoniger Tomes, 1858; Kerivoula pallida Blyth, 1863; V. auratus Dobson, 1871; V. dobsoni Anderson, 1881; K. weberi Jentink, 1890; V. andersoni Trouessart, 1897; M. tsuensis Kuroda, 1922; M. hermani Thomas, 1923; M. watasei Kishida, 1924; M. chofukusei Mori, 1928; and M. flavus Shamel, 1944) were described and subsequently included in major treatments as separate species, subspecies, or synonyms within Chrysopteron (Tate 1941; Honacki et al. 1982; Koopman 1989; Corbet and Hill 1992; Simmons 2005). The African species M. welwitschii (Gray, 1866) and its synonym M. venustus Matschie, 1899 also have been included in the subgenus (Tate 1941; Meester et al. 1986).
Definition of Myotis subgenera has long challenged taxonomists, and their compositions have consequently varied from author to author. Chrysopteron is no exception. Tate (1941:539) defined Chrysopteron as “near Myotis [the subgenus], but distinguished by peculiar dichromatic wing-pattern, somewhat like that of K. picta, and by the presence of four well-developed lobes on i1 and i2. Braincase rather higher and rostrum lower than in Myotis.” He admitted, however, that these craniodental characters showed no clear segregation, for example, “in Chrysopteron the braincase is slightly fuller and rostrum a little more depressed,” and “the 4-cusped condition of i1 and i2 … present also in true Myotis” (Tate 1941:541). Tate's (1941) remark that M3 shows reduction in Myotis, but not in Chrysopteron, also does not appear to be true.
When Findley (1972) included Chrysopteron in the subgenus Myotis in his phenetic treatment, he found evidence that M. flavus, M. formosus, M. hermani, M. rufoniger, M. rufopictus, and M. welwitschii form a separate clade that he named the formosus-group, and linked these species to his newly established emarginatus-group (including M. tricolor, M. emarginatus, and M. goudoti). While Findley (1972) also found M. bocagii was placed close to M. goudoti, he regarded this position as “misplaced” (Findley 1972:38). Although the above Asian taxa were listed as operational taxonomic units and occupied positions similarly separated from each other as from other species in his depictions of phenetic space, all of these taxa were subsumed under the species M. formosus within the formosus-group without explanation in his classification (Findley 1972:42). This view was followed by Honacki et al. (1982), and subsequently, these sometimes conspicuously different forms were generally regarded as conspecifics. For instance, Corbet and Hill (1992) retained Chrysopteron as a valid subgenus (with just 2 species, M. formosus and M. hermani) and separated it from the subgenus Myotis by its wing coloration and differences in the dorsal profiles of the skull. The latter view is, however, hardly justifiable, because all Indomalayan Chrysopteron have gradually sloping braincases similar to those of several species assigned by the same authors to the subgenus Myotis. Koopman (1994) also adopted Findley's (1972) treatment, including the segregation of 2 externally similar species into 2 different subgenera, namely: M. scotti to Selysius and M. bocagii (regarded by Tate  as belonging to Selysius) to the large-footed Leuconoe, but in the latter case also remarking, “foot relatively small” (Koopman 1994:106).
In the era of molecular systematics, growing evidence has revealed the paraphyletic nature of Myotis subgenera. Characters once thought to be diagnostic are often now regarded as convergent morphological traits related more to modes of food procurement than phylogeny (Stadelmann et al. 2004b). Some analyses also suggest that the biogeographical origins of species are a better predictor of phylogenetic relationships than morphology (Stadelmann et al. 2007). Nonetheless, species once treated as belonging or relating to Chrysopteron by Findley (1972) have been shown to have close relationships irrespective of their geographic origin. For instance, Ruedi and Mayer (2001) found M. welwitschii (from Uganda and South Africa) and M. emarginatus (from Greece) monophyletic in all analyses. In incorporating M. formosus from South Korea in their phylogenetic study, Kawai et al. (2003) placed the taxon in a group that included M. welwitschii, M. emarginatus, and M. dasycneme. Although subsequent molecular investigations have not placed M. dasycneme among these species (Ruedi et al. 2013), they have proven the validity of grouping M. formosus with the other species.
The work of Stadelmann et al. (2004b) was the 1st study of Myotis to include all of the traditional subgenera and biogeographic regions. This also included additional African species (M. bocagii, M. tricolor, M. goudoti, and M. scotti) in a phylogenetic analysis inferred from cytochrome-b (Cytb) sequences that found that (together with M. emarginatus and M. formosus) these form a well-supported lineage, which they named the “Ethiopian clade.” Sampling completeness was further improved by the phylogenetic study of Jiang et al. (2010), which included M. watasei and M. flavus and concluded that the Indomalayan taxa investigated represented more than 1 species, and that these also nested within the “Ethiopian clade” of Stadelmann et al. (2004b, 2007). Unfortunately, however, the nomenclature used by Jiang et al. (2010) seriously misconceived the taxa involved (see details under the headings “Status of Chrysopteron” and “Taxonomic remarks” of M. formosus) and consequently lent further confusion to taxonomic and nomenclatural questions surrounding Asian Chrysopteron. Finally, Ruedi et al. (2013), using an even wider sample and a combination of mitochondrial and nuclear sequences, confirmed the monophyly of the “Ethiopian clade” with the additional inclusion of M. anjouanensis (from Anjouan, Comoro Islands).
It has long been supposed that the pelage colors of Chrysopteron species are species-specific (hence the many species descriptions emphasizing differences between pale yellowish and deep reddish taxa), yet, with the exception of Jiang et al. (2010), systematic and taxonomic studies of Asian Myotis that recognize the importance of color differences are thus far confined to Taiwanese field guides (Fang 2007; Cheng et al. 2010) and 1 university thesis (Chou 2004). Major contemporary treatments admit only 2 species (M. formosus and M. hermani) distinguished by size (Simmons 2005; Francis 2008; Smith and Xie 2008), and critical investigation of these has been hampered by their apparent rarity. As noted by Corbet and Hill (1992:121) “known specimens are quite inadequate for any study of infraspecific variation.” In light of this, and following examination of material (including available type specimens) in 21 European, North American, and Asian collections, we discuss the validity of the subgenus Chrysopteron, revise the taxonomy of the named Asian forms, and review their distinguishing characters, distributions, and taxonomic histories. A key to the Asian species of Chrysopteron also is provided.
Materials and Methods
All specimens examined are listed under the corresponding taxa in Appendix I. Abbreviations referring to collections include AMNH: American Museum of Natural History, New York, United States; BMNH: The Natural History Museum (formerly British Museum [Natural History]), London, United Kingdom; CPC: Cuc Phuong National Park Reference Collection, Cuc Phuong, Vietnam; DWNP: Department of Wildlife and National Parks, Kuala Lumpur, Malaysia; FMNH: Field Museum of Natural History, Chicago, United States; HNHM: Hungarian Natural History Museum, Budapest, Hungary; IEBR: Institute of Ecology and Biological Resources, Hanoi, Vietnam; MHC: Harada Collection, Osaka City University, Osaka, Japan; MHNG: Muséum d'histoire naturelle de Genève, Geneva, Switzerland; MNHN: Museum national d'histoire naturelle, Paris, France; MZB: Museum Zoologicum Bogoriense, Bogor, Indonesia; NF: Kim Hy Nature Reserve Collection, Hanoi, Vietnam; NSMT: National Museum of Nature and Science (formerly National Science Museum), Tokyo, Japan; NMNS: National Museum of Natural Science, Taichung, Taiwan; PSUZC: Prince of Songkla University Zoological Collection, Hat Yai, Thailand; RMNH: Naturalis Biodiversity Center (formerly Rijksmuseum van Natuurlijke Historie), Leiden, Netherlands; ROM: Royal Ontario Museum, Toronto, Canada; SMC: Sumiko Matsumura Collection, Yamaguchi, Japan; SMF: Forschungsinstitut und Natur-Museum Senckenberg, Frankfurt a. M., Germany; THU: Tunghai University, Taichung, Taiwan; USNM: National Museum of Natural History (formerly United States National Museum), Washington, D.C, United States; ZMB: Museum für Naturkunde (formerly Zoological Museum), Berlin, Germany.
Forearm length (FA) data were compiled from the literature or measured by the authors from museum specimens to the nearest 0.1 mm. Craniodental measurements were taken to the nearest 0.01 mm using digital calipers and a stereomicroscope. Measurements include only those taken from nonjuveniles, as indicated by the presence of fully ossified metacarpal–phalangeal joints. Abbreviations and definitions for craniodental measurements follow Bates and Harrison (1997) and are: GTL—greatest length of skull, from the anterior of the 1st upper incisor to the most posteriorly projecting point of the occipital region; CCL—condylo–canine length, from the exoccipital condyle to the most anterior part of the canine; CCW—greatest width across the upper canines from their buccal borders; M3M3W—greatest width across the crowns of the last upper molars from their buccal borders; IOW—least width of the interorbital constriction; ZYW—greatest width of the skull across the zygomatic arches; MAW—greatest distance across the mastoid region; BCW—greatest width of the braincase; BCH—braincase height, from the basisphenoid at the level of the hamular processes to the most dorsal part of the skull, including the sagittal crest (if present); CM3L—maxillary toothrow length, from the anterior of the upper canine to the posterior of the crown of the 3rd molar; CP4L—distance from the anterior of the upper canine to the posterior of the crown of the last premolar; ML—mandible length, from the anterior rim of the alveolus of the 1st lower incisor to the most posterior part of the condyle; cm3L—mandibular toothrow length, from the anterior of the lower canine to the posterior of the crown of the 3rd lower molar; and CPH—height of the coronoid process, from its dorsal tip to the apex of the indentation on the ventral surface of the ramus adjacent to the angular process.
All statistical analyses were carried out with R 2.13.2 (R Development Core Team 2012). Measurements were compared using Welch 2-sample t-tests. A 1-way analysis of variance (ANOVA) model with Tukey's pairwise tests was used for comparisons of selected external and craniodental measurements. All variables showed normal distributions in quantile-comparison plots and all tests were 2-tailed. To eliminate false discoveries in multiple tests, the adjustment method of Benjamini and Hochberg (1995) was applied.
Thirty-six Cytb sequences (1,140 base pairs) were downloaded from GenBank to build a phylogenetic tree of available species belonging to Chrysopteron and a number of other representative Myotis taxa (Table 1). In addition, homologous sequences of 3 outgroup taxa (Murina cyclotis, Mu. pluvialis, and Kerivoula papillosa) were used to root the trees. Sequences were aligned in MEGA5 (Tamura et al. 2011), and the analyses were completed in MrBayes version 3.2.1 (Ronquist and Huelsenbeck 2003) using the Bayesian interference method and in RAxML (Stamatakis 2006) to obtain a maximum-likelihood tree. Reliability of nodes in maximum-likelihood analyses was assessed by 1,000 standard bootstraps with RAxML. All analyses were done using a partitioned scheme where each codon position was allowed to have specific model parameters. The general time reversible model with “gamma” and “invariant sites” was used in each partition, as suggested by the results of jModelTest version2.1.3 (Darriba et al. 2012) applied to the alignment.
Origin and GenBank numbers with corresponding references of the 39 cytochrome-b sequences from Myotis analyzed. Both species of Murina and Kerivoula were used as outgroups for the phylogenetic reconstructions.
To obtain Bayesian interference trees, MrBayes was run for 10 × 106 generations and sampled every 1,000 generations. The first 1 × 106 generations were discarded as burn-in. Posterior probabilities were subsequently computed from the consensus of the remaining sampled trees. Two independent replicate analyses were performed on the same data set, and the results were then combined. Effective sample sizes for the estimated parameters and posterior probability were calculated with Tracer version 1.5 (Rambaut and Drummond 2007) and were all higher than 200.
Status of Chrysopteron
The monophyletic lineage known as the “Ethiopian clade” within the phylogenetic tree of Myotis (Stadelmann et al. 2004b, 2007; Jiang et al. 2010; Ruedi et al. 2013) was recovered in our analyses with high support (> 95% posterior probability and boostrap value) and includes 6 Ethiopian, 1 Palearctic, and 2 Indomalayan species (Fig. 1). The majority of the taxa grouped in this clade had previously been regarded as closely related based on phenetic analyses (Findley 1972). The species referred to as “M. flavus” (e.g., Jiang et al. 2010) is identified by us as M. formosus, and it does not form a sister group with M. rufoniger, the other species of Indomalayan parti-colored bat (variously designated as “M. watasei” or “M. formosus” by different authors [e.g., Jiang et al. 2010]). Rather, M. formosus (as understood here) is associated with the European M. emarginatus and the African M. tricolor (low nodal support among these 3 species impedes more precise understanding of their phylogenetic relationships), whereas M. rufoniger is sister to the African M. welwitschii (with high support). The Ethiopian clade is itself part of the Old World species assemblage as detailed in a more comprehensive analysis of the genus (Ruedi et al. 2013).
Although our investigations did not reveal any craniodental characters distinguishing these taxa from all other species of Myotis, 1 phenetic character was shared by all of the studied taxa. This is the characteristic reddish or yellowish dorsal fur that strongly differentiates these bats from congeners. An additional feature that may be a synapomorphic character of Chrysopteron is the peculiar texture of the hairs, variously described as “thick and woolly” (Dobson [1871:187] for V. auratus and Yoshiyuki [1989:110] for M. tsuensis), “cottony” (Tomes [1858:84] for V. rufoniger), “woolly” (Blyth [1863:34] for K. pallida, Hill and Morris [1971:43] for M. morrisi, and Dietz et al. [2009:242] for M. emarginatus), “thick, soft and cottony” (Tomes [1858:81] for V. emarginatus), or “shaggy” (Rosevear [1965:303] for M. bocagii). This feature also is reflected in the vernacular name, “hairy bats,” frequently used for Ethiopian species (Taylor 2000), and, with the exception of 2 sister species, M. goudoti and M. anjouanensis from islands in the Indian Ocean, is typical for all Chrysopteron species investigated by us.
Chrysopteron is the earliest available name for the “Ethiopian clade.” Its type species, bartelsi Jentink, 1910 from Java, is currently considered a subspecies of M. formosus (Simmons 2005), which, in turn, is part of the “Ethiopian clade.” This renders Chrysopteron the only traditional subgenus of Myotis currently validated by molecular systematics.
With the separation of Cistugo from Myotis (Stadelmann et al. [2004b]; subsequently confirmed by Horácek et al.  and Lack et al. ), all sub-Saharan species of Myotis were regarded (or anticipated in the case of M. morrisi) as part of the “Ethiopian clade” by Stadelmann et al. (2004b). Our morphological examination of the holotype of M. morrisi indicated that the species has the fur characters regarded here as diagnostic of Chrysopteron, thereby justifying its inclusion in the “Ethiopian clade.” The 2 Indian Ocean insular species, M. goudoti and M. anjouanensis, which are clearly part of the Ethiopian clade in molecular reconstructions (Ruedi et al. 2013; Fig. 1), also have very strong overall reddish to orange dorsal fur. In contrast, the recently recognized species M. dieteri Happold, 2005 may be distinct from the “Ethiopian clade.” Although M. dieteri resembles M. bocagii in size, Happold (2005) noted its dorsal fur coloration as dark brown with a pale auburn tip (among other distinguishing characters), instead of reddish brown with a rufous tip, as in M. bocagii. Based on this external characteristic M. dieteri would be the only sub-Saharan Myotis that does not belong to Chrysopteron, but molecular data are needed to substantiate this interpretation.
Although Stadelmann et al. (2004b) supposed the dichromatic wing coloration evolved only once within the “Ethiopian clade,” the presence of the conspicuous black-and-orange pattern in 2 independent lineages of Chrysopteron and in several genetically unsampled species suggests that the developmental pathway may be conserved in many (if not all) species of the subgenus. It is, however, worth noting that within the subgenus this wing coloration is always accompanied by large body size and is observed only in M. welwitschii, the largest Ethiopian species (Taylor 2000; Ratcliffe 2002), and much less conspicuously so in some specimens M. tricolor from Malawi (M. Happold, pers. comm.) and in similar-sized Asian taxa. Interestingly, the distantly related K. picta (Vespertilionidae: Kerivoulinae) also shows the very same dichromatic pattern, suggesting that this peculiar coloration has appeared more than once in the evolution of bats.
Biogeographic inferences based on a likelihood model of ancestral area reconstruction (DEC model—Ree and Smith 2008) and a nearly complete, worldwide taxonomic sampling of Myotis species (Ruedi et al. 2013) indicate that species of Chrysopteron evolved after an initial range expansion of an Asian ancestor. This widespread ancestor radiated in sub-Saharan Africa to give rise to several taxa. The 2 lineages that led to the current, nonsister species M. rufoniger and M. formosus apparently resulted from 2 independent recolonizations of the Asian continent by these African forms.
Revision of Asian species of Chrysopteron
The 1st publication that grouped all of the Asian dichromatic taxa into a single species was that of Findley (1972), who provided limited justification for this action. However, his opinion was adopted by Honacki et al. (1982), Koopman (1989, 1994), and, with the exception of recognizing M. hermani as a distinct species, by Corbet and Hill (1992) and Simmons (2005). The presence of more than 1 species within M. formosus sensu lato is suggested by more recent publications employing external traits (Cheng et al. 2010), craniodental features (Heaney et al. 1998; Chou 2004), and molecular genetic studies (Jiang et al. 2010; Ruedi et al. 2013).
Our phylogenetic reconstruction (Fig. 1) confirms that specimens formerly included in the M. formosus complex are not monophyletic, but form 2 strongly supported, nonsister clades. The nomenclature used in the tree is based on our identification of voucher specimens or is inferred from distributional information and genetic distances.
One of these clades includes all representatives of M. formosus (= M. flavus sensu Jiang et al. 2010), whereas the remaining sequences, including those representing M. rufoniger (= M. watasei sensu Jiang et al. 2010), group in a distinct clade. The latter clade is clearly sister to M. welwitschii from Africa in all reconstructions (Fig. 1), but the phylogenetic position of the M. formosus clade is uncertain within a group that also contains M. emarginatus and M. tricolor with no bootstrap or posterior probability supporting either sister-group relationship. The mean Kimura 2-parameter genetic distance is about 2 orders of magnitude larger between (18.9%) than within (0.2%) these 2 clades and corresponds well to a pattern of interspecific and intraspecific comparisons, respectively (Bradley and Baker 2001).
Myotis formosus sensu lato, in fact, contains several forms that are morphologically distinct and can be found either in sympatry (e.g., in Taiwan) or in widely separated geographic localities. Having investigated specimens from the entire distribution, we recognize 6 species (M. bartelsi, M. formosus, M. hermani, M. rufoniger, M. rufopictus, and M. weberi), revise their distributions, and summarize their distinguishing characters. In view of the history of taxonomic confusion, a brief review of previous taxonomic opinions also is included.
Subgenus Chrysopteron Jentink, 1910
Chrysopteron Jentink, 1910:74. Original designation, type species Chrysopteron bartelsi.
Dichromyotis Bianchi, 1916:78. Original designation, type species Myotis formosus.
Myotis bartelsi Jentink, 1910
Chrysopteron Bartelsii Jentink, 1910:74. Type locality Mt. Pangrango, Java, Indonesia.
Myotis bartelsi: Tate 1941:542. First use of current name combination.
Myotis formosus: Honacki et al. 1982:187 (part). Name combination.
Myotis formosus bartelsi: Koopman 1994:101. Name combination.
This is a large species of Chrysopteron (Table 2). Individual dorsal hairs are black basally, pale yellow distally, then darken to deep red before terminating in a black tip. The general impression of the dorsal fur is thus red tipped with black. Individual ventral hairs possess a black base, followed by either a pale yellow section that progressively darkens distally to deep red, or are otherwise entirely deep red. The ear is conspicuously edged with black and the thumb and underside of hind foot are entirely black. This combination of fur and flesh coloration is henceforth referred to as the “rufoniger-type,” this being the 1st taxon described to have these characters (Figs. 2 and 3a).
Selected external and craniodental measurements (in mm) of Asian Chrysopteron (Myotis) species. Values are given as mean ± SD (n ≥ 5), minimum–maximum (n). Acronyms and definitions for measurements are given in the text.
The skull has a definite, albeit shallow, rostral depression, and an elongated supraorbital region. The sagittal crest is prominent, and the lambdoid crests are strong. The dentition is robust. The upper canine has a wide base, and attains a height twice that of P4. P3 is fully out of line of the rest of the toothrow, has less than half the basal area of P2, and is not visible in lateral view of the skull. The mandible of the holotype is missing (Fig. 4).
In describing C. bartelsi (separated from C. weberi from Sulawesi based on its larger size and small differences in color patterns), Jentink (1910) differentiated the cranial (flatter head) and dental (smaller middle upper premolar and number of cusps on lower incisors) features of these 2 species from another, unrelated genus containing parti-colored bats, Kerivoula, and placed both in his newly established genus Chrysopteron. He also noted that the species of Chrysopteron “have some characters in common with … Myotis” (Jentink 1910:73) and tentatively included M. formosus in the new genus.
This taxon is quite similar to M. hermani in craniodental traits. Although M. bartelsi is smaller in several mensural characters (e.g., FA, M3M3W, IOW, and CM3L; and is therefore tentatively regarded here as a separate species), in other measurements (e.g., GTL, CCW, and ZYW) it falls within the known variation of M. hermani. When more specimens and genetic information for both taxa become available, they may prove to be conspecific.
Myotis formosus (Hodgson, 1835)
Vespertilio formosa Hodgson, 1835:700. Type locality Kathmandu Valley, Nepal.
Kerivoula pallida Blyth, 1863:34. Type locality Chaibassa, Orissa, India.
Vespertilio auratus Dobson, 1871:186. Type locality Darjeeling, West Bengal, India.
Myotis formosus: Tate 1941:541. First use of current name combination.
Myotis flavus Shamel, 1944:191 Type locality Yuanli, Miaoli, Taiwan.
Myotis formosus formosus: Koopman 1994:101. Name combination.
A medium-sized species of Chrysopteron (Table 2). Individual dorsal hairs are mid-brown at base (approximately 10% of hair length), pale yellow distally for 80–100% of hair length, or instead terminate in a mid-brown tip slightly darker than the base. Banding is sometimes not evident as the color changes gradually. The general aspect of the dorsal fur is light yellow-brown. Ventral hairs either possess a mid-brown base followed distally by pale yellow, or are entirely light yellow. The darker hair bases are not apparent from a superficial view. The ear is only faintly edged with black, and the thumb and hind foot are brown, not black. This combination of fur and flesh coloration is henceforth referred to as the “formosus-type,” this being the 1st taxon having these characters to be described (Figs. 3b and 6).
The skull has a distinctly elevated frontal region, with a globose braincase. The sagittal crest is missing or very weak and the lambdoid crests are weakly developed. The upper canine is moderately robust. P3 occupies at most half the basal area of P2, is fully or mostly out of the toothrow (sometimes missing), and is not visible in lateral view of the skull. The lower middle premolar (p3) occupies approximately half the basal area of p2 and is usually situated within the toothrow (Fig. 7).
The very vague description of K. pallida given by Blyth (1863:34) contains a short comparison with Kerivoula picta and is limited to some external measurements.
Dobson's (1871:187) description of V. auratus “… hairs tipped with light golden brown; beneath light fawn color” defines it as having the same general color as M. formosus, with which he later (together with K. pallida) synonymized the taxon (Dobson 1878). Without referring to M. formosus, Shamel (1944:191) described M. flavus from Taiwan and characterized it as a “pale yellow bat.” He pointed out that this taxon is larger than M. rufoniger and M. watasei and that its P3 is much more reduced. He also described the characteristically differing coloration of M. watasei from a Taiwanese specimen (Shamel 1944:192). Lacking exact information on the characters of M. formosus sensu stricto, Jiang et al. (2010) mistakenly thought that M. flavus and M. formosus differ in coloration and craniodental traits (they likely regarded M. formosus as the same taxon as M. rufoniger). Although they correctly recognized that 2 separate species occur in Taiwan and in mainland Asia, these populations were subsequently associated with the wrong names. As a consequence, they named the larger, yellow species “flavus” and the smaller, reddish one “formosus,” and M. watasei was regarded as a junior synonym of M. formosus. The specific identity of the Jianxi (China) specimen (Jiang et al. 2010) is confirmed by the accompanying photographs (Jiang et al. 2010:46–47) and its Cytb sequence (Fig. 1).
Although the limited geographic coverage does not suggest major genetic subdivision within M. formosus (n = 19), based on Welch 2-sample t-tests, the Taiwanese population (M. flavus, n = 12) is significantly larger than the mainland population in the following measurements: M3M3W (t10.34 = 3.42, P = 0.006), CM3L (t9.26 = 3.89, P = 0.003), CP4L (t10.89 = 3.36, P = 0.006), cm3L (t7.62 = 3.75, P = 0.006), ML (t7.39 = 2.92, P= 0.021) and CCW (t8.54 = 2.41, P = 0.041). After controlling for false discoveries in multiple tests, differences in M3M3W, CM3L, CP4L, and cm3L remain significant (P = 0.023 in each case). We regard M. formosus flavus herein as a valid subspecies.
Afghanistan, India (Jammu and Kashmir, Himachal Pradesh, Punjab, Maharashtra, Uttar Pradesh, Bihar, West Bengal, Sikkim, Assam, and Meghalaya—for details see Bates and Harrison  and Mandal et al. ), Nepal, China (Tibet and Jianxi), Taiwan, and Vietnam (Fig. 5). The IEBR XL-15B specimen from Thanh Hoa Province, Vietnam (an adult female, collected on 18 April 2012), represents the 1st and only record of M. formosus sensu stricto from the country.
The skull is very robust with a shallow but distinct frontal depression, posteriorly elongated supraoccipitale, and exceptionally developed sagittal and lambdoid crests. The basal dimensions of C1 exceed those of P4, whereas P3 is minute, fully displaced lingually, and obscured in lateral view of the skull. The lower middle premolar (p3) is half the size of p2 and partly out of the toothrow (Fig. 8).
Thomas (1923) allied M. hermani with M. weberi and M. bartelsi and noted that M. hermani is much larger size and has a well-marked sagittal crest. He also briefly discussed the earlier generic placements of M. weberi and M. bartelsi (in Kerivoula and Chrysopteron, respectively) and was the first to recognize that all 3 species belong to Myotis.
Although larger, M. hermani is evidently closely related to, and might be conspecific with M. bartelsi, in which case M. bartelsi would have priority (see also remarks under that species).
Myotis rufoniger (Tomes, 1858)
Vespertilio rufo-niger Tomes, 1858:82. Type locality Shanghai, China.
Myotis tsuensis Kuroda, 1922:43. Type locality Tsushima Island, Japan.
Myotis Watasei Kishida, 1924:36. Type locality Manjhou, Pingtung, Taiwan.
Myotis chofukusei Mori, 1928:389. Type locality “Kaishu,” Hwanghae-Namdo, North Korea.
Myotis formosus chofukusei: Kuroda 1938:97. Name combination.
Myotis formosus tsuensis: Kuroda 1938:97. Name combination.
Myotis formosus watasei: Kuroda 1938:97. Name combination.
Myotis rufoniger: Tate 1941:541. First use of current name combination.
Myotis sicarius tsuensis?: Tate 1941:548. Name combination.
Myotis formosus: Findley 1972:42 (part). Name combination.
Myotis formosus rufoniger: Koopman 1994:101. Name combination.
Myotis formosus tsuensis: Koopman 1994:101. Name combination.
Myotis formosus watasei: Koopman 1994:101. Name combination.
The skull has a slightly, but distinctly, elevated frontal part, and moderately strong sagittal and lambdoid crests. The dentition (including the canines) is moderately robust. The basal area of P3 is approximately two-thirds that of P2, and is usually in line and visible from outside, but rarely displaced inward. The lower middle premolar (p3) is well developed and at least two-thirds of the size of p2 basally, but often closely approaches its basal dimensions (Fig. 9).
Tomes (1858) provided a diagnosis for his V. rufoniger based on the color differences (ears edged with black, and dorsal and ventral hairs tipped with bright rufous) but hesitated to recognize it as a distinct species or a “variety” of V. formosus.
When Kuroda (1922) described M. tsuensis as having reddish brown fur dorsally and ventrally and compared it only with M. macrodactylus and M. nattereri bombinus, he unsurprisingly found it specifically distinct. Details provided by Kishida (1924:40) for M. watasei unambiguously define the species from Taiwan as belonging to the “rufoniger-type”: “ear red brown at base, edged with black … upper [dorsal] and under [ventral] body fur basal two-thirds brown-yellow, terminal one-third brown … feet black” (translated from Japanese).
Mori (1928:390) compared M. chofukusei, characterized by its “capucine orange” dorsal fur and ears emarginated with dark margins, with M. bechsteinii. Under the heading of formosus, Howell (1929:15) stated “the mainland bat of this rufous and black type has been described under the name rufo-niger Tomes, but I believe that the validity of the latter has not been satisfactorily established.” The geographical basis of this remark likely stemmed from his erroneous idea that M. formosus was described from Formosa (= Taiwan).
Imaizumi (1970:223), however, regarded M. tsuensis as “indistinguishable from M. chofukusei described from Korea” and gave its range as Tsushima and the Korean Peninsula. Yoshiyuki (1989) and Yoon (2010) also regarded the Korean population as belonging to the subspecies M. formosus tsuensis.
Kim et al. (2011) published the complete mitochondrial genome of a specimen of “M. formosus” from South Korea (GenBank accession number is HQ184048, not HQ184084 as published in Kim et al. 2011); genetically, this corresponds fully (100% match at Cytb, see Fig. 1) to “M. formosus” in Kawai et al. (2003) and “M. watasei” in Jiang et al. (2010) and thus represents M. rufoniger, not M. formosus sensu stricto.
Although earlier studies (Imaizumi and Yoshiyuki 1969) found no differences between specimens from Tsushima Island (M. tsuensis) and the geographically closer Korean Peninsula (M. chofukusei), in 1-way ANOVA models the Tsushima Island population (n = 4) proved to be significantly smaller than either M. chofukusei (n = 10) or the Taiwan population (M. watasei, n = 8), respectively, in the following measurements GTL (P = 0.020 and 0.006, F = 4.82), MAW (P = 0.005 and 0.010, F = 5.36), CM3L (P = 0.037 and 0.001, F = 5.39), CP4L (P = 0.005 and 0.001, F = 11.11), and CPH (P = 0.021 and 0.001, F = 9.36). After controlling for false discoveries in multiple tests, differences between M. tsuensis and M. chofukusei in MAW and CP4L remained significant (P = 0.038 in both cases). For M. tsuensis versus M. watasei, all of the above differences remained significant (P = 0.023, 0.03, 0.005, 0.005, and 0.005, respectively). Abe et al. (2008) asserted that the Tsushima population was vagrant from the Korean Peninsula and did not breed on the island although a female with young had been reported on Tsushima Island in August by Imaizumi (1970). Recent observations indicate these bats are present year-round on the island (S. Matsumura, pers. comm.). Specimens from Tsushima have not yet been included in phylogenetic analysis, but all other samples studied show almost identical Cytb sequences, indicating no major genetic subdivision within the species.
The only known Lao record of M. rufoniger is included in several phylogenetic reconstructions (Stadelmann et al. 2004b; Jiang et al. 2010; Ruedi et al. 2013; present paper). Its specific identity is supported by the color plate provided by Francis et al. (1999) and by its genetic data (e.g., approximately 0.8% sequence divergence from South Korean or Taiwanese M. rufoniger).
North Korea, South Korea, Japan (Tsushima), China (Fujian, Jiangxi, Jilin, Shanghai, and Sichuan), Taiwan, Laos, and Vietnam (Fig. 5). The Vietnamese specimens from the provinces of Hai Duong, Bac Kan, and Ninh Binh represent the 1st published records of the species from the country.
The skull profile ascends almost evenly with no frontal depression. The sagittal and lambdoid crests are only moderately developed, whereas the skull is globose posteriorly. The canines are moderately strong, and the upper molars are relatively robust with developed talons. The P3s are missing from the holotype. The p3s are very small and intruded lingually half-way out of the line of the toothrow (Fig. 10).
Waterhouse (1845) gave a detailed description of the fur of his new species, which was compared by him only with K. picta and separated from that species by its much larger size and differently shaped ear.
Tomes (1858) directly compared the type specimens of M. rufopictus and M. formosus and noted the larger size (despite immaturity) and missing middle premolars of the former.
Large differences in skull measurements within the Philippines (CCL 16.1–17.9, and CM3L 7.1–8.5, n = 4) are reported by Ingle and Heaney (1992) and records of a reddish form in the islands (L. Heaney, Field Museum of Natural History, pers. comm., 2013) imply the presence of a 2nd species in addition to M. rufopictus.
Philippines (Fig. 5). An overview of the distribution and habitat of the species is given by Ong et al. (2008).
Myotis weberi (Jentink, 1890)
Kerivoula weberi Jentink, 1890:129. Type locality Loka, Bantaeng, Sulawesi, Indonesia.
Myotis weberi: Tate 1941:542. First use of current name combination.
Myotis formosus: Honacki et al. 1982:187 (part). Name combination.
Myotis formosus weberi: Koopman 1994:101. Name combination.
The skull has a distinct frontal depression, moderately developed sagittal and lambdoid crests, and posteriorly elongated supraoccipital region. The upper canine has a wide base but is not especially high, whereas P2 is much reduced in size, fully or partly displaced lingually, and is accordingly obscured or visible in the lateral view. The p3 occupies half the basal area of p2 at most and is situated within the toothrow (Fig. 11).
In placing M. weberi in the genus Kerivoula, Jentink (1890) was misled by the apparent similarity of their wing patterns, and within that genus he distinguished his new species by its larger size; comparisons were inappropriately confined to K. picta.
Myotis weberi is erroneously given by Yoshiyuki (1989) as the type species of Chrysopteron.
Indonesia (Sulawesi [Fig. 5]).
Key to the Asian Species of Chrysopteron
2a. M3M3W < 7.6 mm; basal area of P3 more than half of P2 (“mainland red” species; Fig. 9)M. rufoniger
2b. M3M3W > 7.6 mm; basal area of P3 less than half of P2 (“Sunda red” species)3
3a. Sagittal crest moderately developed; CCL < 18.0 mm; M3M3W < 8.5 mm (Sulawesi; Fig. 11)M. weberi
3b. Sagittal crest very strong; CCL > 18.0 mm; M3M3W > 8.5 mm4
4a. FA > 56 mm; M3M3W > 8.8 mm (Sumatra and Malay Peninsula; Fig. 8)M. hermani
4b. FA 53.4 mm; M3M3W 8.63 mm (Java and Bali; Fig. 4) M. bartelsi
5a. Skull with distinct frontal depression; very weak sagittal crest; CCW near or over 5 mm (“mainland yellow” species; Fig. 7)M. formosus
5b. Cranial profile straight; moderately developed sagittal crest; CCW 4.75–4.76 mm (“Philippine yellow” species; Fig. 10)M. rufopictus
We thank N. Simmons and E. Westwig (AMNH); P. Jenkins and R. Portela Miquez (BMNH); Le Trong Dat (CPC); L. Heaney (FMNH); M. Harada (MHC); J. Cuisin and Jean-Marc Pons (MNHN); Shin-Ichiro Kawada and S. Shimoinaba (NSMT); Yen-Jean Chen (NMNS); C. Satasook and P. Soisook (PSUZC); C. Smeenk, S. van der Mije, and W. van Bohemen (RMNH); B. K. Lim and J. L. Eger (ROM); S. Matsumura (SMC); K. Krohmann (SMF); K. Helgen and D. Lunde (USNM); and F. Mayer and N. Lange (ZMB) for kindly providing essential information and access to specimens in their care. We are grateful to P. Racey (University of Exeter) and P. Bates (Harrison Institute) for their longstanding support of bat research in Southeast Asia, to Z. Vas for his help with the statistical comparisons, and to A. Honfi for the final elaboration of drawings. This research received support from the SYNTHESYS Project, which is financed by European Community Research Infrastructure Action under the FP7 “Capacities” Program, the NHMG, the FMNH, and the Smithsonian Institution. Field studies in Vietnam were funded by the National Foundation for Science and Technology Development of Vietnam (project 106.11–2012.02) and helped by the Directorate and Scientific Council of IEBR, and Le Dinh Thuy.
Kerivoula sp.—JAVA: RMNH 35401 (registered as M. formosus). Myotis anjouanensis—COMOROS: MNHN 1886-266, 1886-1265, 1886-1267, 1886-1269, 1886-1536. M. bartelsi—JAVA: MZB 10573 (holotype). M. bocagii—CAMEROON: BMNH 184.108.40.206, 89.723; DR CONGO: BMNH 59.508, 59.509; GABON: MNHN 1985-1928; KENYA: BMNH 220.127.116.11 (holotype of M. hildegardeae); MALAWI: BMNH 87.1080, 87.1081. M. emarginatus—HUNGARY: HNHM 71.7.1; PAKISTAN: BMNH 1909.1.4.33 (holotype of M. desertorum). M. formosus—AFGHANISTAN: SMF 38752; NEPAL: BMNH 18.104.22.168 (holotype), HNHM 98.8.22; NORTH INDIA: BMNH 22.214.171.124, 126.96.36.199, 188.8.131.52, FMNH 85057; TAIWAN: USNM 239908 (holotype of M. flavus), BMNH 184.108.40.206, 220.127.116.11, HNHM B000054, MHNG B000065, B000100, NMNS t-4607, t-4614, t-4546, THU B000060, B030004, ZMB 54193; TIBET: BMNH 18.104.22.168; VIETNAM: IEBR XL-15B; UNKNOWN: BMNH 22.214.171.124. M. goudoti—MADAGASCAR: BMNH 126.96.36.1993 (holotype of M. madagascariensis), 87.146, 87.147, 99.9, MNHN 1949-310, 1981-869. M. hermani—SUMATRA: BMNH 188.8.131.52 (holotype); THAILAND: PSUZC M05.1. M. morrisi—ETHIOPIA: BMNH 70.488 (holotype); NIGERIA: BMNH 84.840. M. scotti—ETHIOPIA: BMNH 184.108.40.206–220.127.116.11 (paratypes). M. rufoniger—CHINA: BMNH 18.104.22.168 (holotype), 22.214.171.1242, 126.96.36.199, AMNH 84843, USNM 241369, ZMB 4139; JAPAN (TSUSHIMA ISLAND): NSMT 11886, 21191, without number; SOUTH KOREA: FMNH 48375, HNHM 2003.37.8–2003.37.10, 2003.37.24, 2003.37.45, MHC 5289, 5296, NSMT 5732, 5888, 11671, 27178; TAIWAN: MHC 7222, 7223, MHNG B030022, B000098, NMNS t-4613, t-4611, THU B000048, B000053, B030063, B030046, USNM 239909, ZMB 88447, 88448; VIETNAM: IEBR T.080511.1, NF 170906.7, CPC DB0295; UNKNOWN: NSMT 34440. M. rufopictus—PHILIPPINES: BMNH 188.8.131.523 (holotype), FMNH 1114. M. tricolor—KENYA: BMNH 76.29.30, 76.29.31; SOUTH AFRICA: BMNH 1881.17.1. M. weberi—SULAWESI: AMNH 224579, RMNH 35827 (holotype), ZMB 5416, 88450.