The genus Myotis is a diverse group of vespertilionid bats found on nearly every continent. One clade in this group, the subgenus Chrysopteron, is characterized by reddish to yellowish fur and, in some cases, visually striking dichromatic wing pigmentation. Here, we describe a new dichromatic species of Myotis (Chrysopteron) from the Nimba Mountains in Guinea. The new species is superficially similar to Myotis welwitschii, but phylogenetic analyses based on cytochrome b data indicated that it is actually more closely related to M. tricolor. Discovery of this new taxon increases the number of Myotis species known from mainland Africa to 11 species, although patterns of molecular divergence suggest that cryptic species in the Chrysopteron clade remain to be described. This discovery also highlights the critical importance of the Nimba Mountains as a center of bat diversity and endemism in sub-Saharan Africa.
INTRODUCTION
The genus Myotis is the most speciose genus of bats with over 120 extant species and a range that covers most of the world (Simmons, 2005; Simmons and Cirranello, 2020). Myotis bats are primarily insectivorous and range from tiny species that weigh only a few grams (e.g., Myotis elegans, 3–5 g; Reid, 2009) to quite large species (e.g., Myotis chinensis, 30–40 g; Francis, 2001), and have color patterns that range from brown or black to brightly colored yellow or orange, including some dichromatic taxa with particolored orange and black wings (e.g., Myotis welwitschii; Monadjem et al., 2010). Members of the genus are recognized by a combination of traits including ears that are longer than they are wide, a lanceolate tragus, a simple narial region with nostrils that open anterolaterally and not dorsally, a dental formula of I2/3, C1/1, P3/3, M3/3 = 38 (although some species lack p3 and P3), a single-rooted p3 that is somewhat smaller than p2, myotodont lower molars (although a few species are seminyctalodont), lower molars that have a relatively deep talonid basin, well-developed entocristid, and lack a lingual cingulid, a relatively high-crowned lower canine with well-developed mesial and distolingual shelves, a tall, daggerlike upper canine with a distinct lingual ridge, anterior upper premolars that are reduced in size compared to the P4, which is simple with rounded labial shelf, and an M1 and M2 with a distally open protofossa and lacking a paraconule and metaloph, and having a hypocone that is either small or absent (Findley, 1972; Koopman, 1994; Ruedi et al., 2015; Gunnell et al., 2017). Many species are cryptic and difficult to distinguish from similar congeners, and more than 20 new species have been described or elevated from synonymy in the last decade (Simmons, 2005; Simmons and Cirranello, 2020).
Only 11 species of Myotis are currently known from continental Africa, and only eight of these occur in tropical sub-Saharan regions, suggesting that new and/or cryptic species of Myotis likely remain to be discovered given the vast size and diverse habitats of the African continent (Patterson et al., 2019; Simmons and Cirranello, 2020). A recent study of genetic variation among and within populations of African Myotis by Patterson et al. (2019) found evidence of strong geographic structure within two species (M. tricolor and M. welwitschii) suggesting possible cryptic species in these complexes. As with many African bat taxa, sampling effort has been very uneven, especially in tropical areas, and large sampling gaps remain for all Myotis species (Patterson et al., 2019).
The Nimba Mountains are an isolated mountain range in West Africa that straddles the borders of Guinea, Liberia, and Côte d'Ivoire (fig. 1). With peaks rising between 1600–1750 m above sea level, the mountain range is surrounded by lowland habitats and occupies a transition zone between the tropical forest zone and moist woodlands. The Nimba Mountains support exceptional biodiversity including a diverse bat fauna that includes several threatened and endangered species (Brosett, 2003; Denys et al., 2013; Monadjem et al., 2016). In January and February 2018, and in March 2019, we conducted field surveys to assess the species diversity and distribution patterns of subterranean roosting bats in the northern (Guinean) portion of the Nimba Mountains. Surveys were carried out in the Mining Concession and Perimeter, managed by Société des Mines de Fer de Guinée S.A (SMFG), as well as in the adjacent World Heritage Site and surrounding lowlands, as part of SMFG's environmental and social impact assessment for the Nimba Iron Ore Project. As part of this work, we captured two individuals of a spectacular dichromatic orange and black species clearly belonging to Myotis but that did not fit the diagnosis of any previously described species in the genus. We analyzed morphological, morphometric, echolocation, and molecular data, and herein describe these individuals as representing a new species of Myotis (subgenus Chrysopteron).
FIGURE 1.
Relief map with satellite imagery overlaid showing the capture locations of the Myotis nimbaensis holotype (red triangle: captured 26 January 2018; red circle: collected 2 February 2018) and other sites where this species was likely detected with acoustic monitoring at entrances of underground sites (yellow squares). The mining concession footprint is shown as the unshaded area within the Nimba Mountains Strict Nature Reserve and Mount Nimba World Heritage Site (identical boundaries) that are overlaid by light green shading. Dominant habitat types are visible as different textures from the satellite imagery showing how gallery forests occur along steep canyons surrounded by savanna at higher elevations. Roads and trails in the mining concession are also visible as light brown features. Location of the Nimba Mountains in relation to Guinea, Liberia and Côte d'Ivoire is shown in the inset.
Genetic diversity and species limits of bats in the monophyletic subgenus Chrysopteron (= “Ethiopian Clade” of Stadelmann et al., 2004, and “Clade V” of Ruedi et al., 2013, 2015) were recently reviewed by Csorba et al. (2014) and Patterson et al. (2019). Chrysopteron as defined by Csorba et al. (2014) is the only traditional subgenus of Myotis currently validated by molecular systematics (Csorba et al., 2014; Patterson et al., 2019; Morales et al., 2019). This taxon, which is diagnosed morphologically by reddish or yellowish dorsal fur with a cottony or wooly texture, is distributed from Africa and Madagascar to the Mediterranean region, across India and southern Asia to China, Taiwan, Korea, the Philippines, Malaysia, and Indonesia (Csorba et al., 2014). Myotis (Chrysopteron) currently contains 14 species: Myotis anjouanensis, M. bartelsi, M. bocagii, M. emarginatus, M. formosus, M. goudoti, M. hermani, M. morrisi, M. rufoniger, M. rufopictus, M. scotti, M. tricolor, M. weberi, and M. welwitschii (Csorba et al., 2014; Patterson et al., 2019).
FIGURE 2.
Photographs of roosting and surrounding habitats at the type locality in the Guinean Nimba Mountains. A, Entrance of Kaiser Adit 1. B, Entrance of Kaiser Adit 3 with harp trap placed for bat capture. C, Ecotone of savanna and gallery forest habitats at the headwaters of the Zié river viewable from where bats were captured at adit entrances.
MATERIALS AND METHODS
We surveyed subterranean habitats (abandoned mine adits and natural caves) by placing harp traps (a two-bank 4.2 m2 harp trap [manufactured by Austbat] or a “cave catcher” 0.9 m2 harp trap [Bat Conservation and Management]) at entrances at least 30 minutes before sunset (fig. 2). Harp traps were checked a maximum of every 10–15 minutes for the presence of bats for two to three hours after sunset and removed after bat emergence activity had subsided. Captured bats were placed in individual cloth bags and processed to identify species, sex, age, and reproductive status prior to release at place of capture following standard methods as described in Kunz and Parsons (2009). We collected standard field measurements (forearm length, tibia length, hind-foot length, tail length, ear length, tragus length, body length, and mass) and consulted Mammals of Africa, volume 4 (Hedgehogs, Shrews and Bats; Happold and Happold, 2013) to identify species. We collected tissue samples for DNA analysis using a 3 mm biopsy punch from the wing membrane and stored in desiccant. Echolocation calls were recorded upon release for echolocating bat species using a Pettersson M500 full-spectrum bat detector (Pettersson Elektronik) at a sampling rate of 500 kHz. Files were saved as uncompressed “.wav” format and analyzed using BatSound v.4.1 (Pettersson Elektronik: FFT size 1024, Hanning window) to determine the following parameters for each pulse: duration (D), maximum frequency (FMAX), minimum frequency (FMIN), peak frequency (PF), and interpulse interval (IPI). D, FMAX, FMIN, and IPI were measured from spectrograms, and PF from power spectrum.
A male bat captured and released on 26 January 2018 at Kaiser Adit 3 could not be identified using keys for the bats of Africa. Due to the likelihood that this individual represented an unknown species of Myotis, we resurveyed two adjacent adits (Kaiser Adit 1 and Kaiser Adit 3) on 2 February 2018 and collected two individuals (one male, one female) emerging from Kaiser Adit 1 to confirm and describe the species. The collected specimens included the male bat originally captured and released on 26 January 2018, identifiable by the 3 mm hole in its wing membrane from tissue sampling.
We employed analyses of mitochondrial cytochrome b gene sequences as well as standard morphological comparisons in our evaluation of the new taxon. The specimens examined and tissues used for this study (appendices 1 and 2) belong to the following collections:
AMNH
American Museum of Natural History, New York
BMNH
Natural History Museum, London
FMNH
Field Museum of Natural History, Chicago
USNM
National Museum of Natural History, Smithsonian Institution, Washington, DC
Molecular Analyses
Our analysis was designed to cover all the major lineages previously recognized in the subgenus Chrysopteron to allow placement of our new species in phylogenetic context. We downloaded GenBank sequences of Myotis compiled or originally published by Csorba et al. (2014) and were given early access to sequence data subsequently published in Patterson et al. (2019); see appendix 2. Data for an outgroup, Submyotodon latirostris, were also obtained from GenBank (appendix 2). Genomic DNA from the new species of Myotis was extracted at CIBIO-InBIO, University of Porto, Portugal, using Qiagen DNeasy kits (Qiagen, Crawley, UK) and stored at -20° C. Mitochondrial cytochrome b (cyt b) gene was amplified by polymerase chain reaction (PCR) using the primers MOLCIT-F (5′-AATGACAT-GAAAAATCACCGTTGT-3′; Ibáñez et al., 2006) and MVZ16-R (5′-AAATAGGAARTATCAYTCTGGTTTRAT-3′; Smith and Patton, 1993). PCR's were performed in a 10 µL volume, which included 1 µL of DNA extract, 0.4 µL of each primer (10 µM), 5 µL of Qiagen Master Mix, and double-distilled water was added until final volume was reached. Reactions were performed under the following conditions: 95° C for 15 min; 40 cycles of 95° C for 30 s, 50° C for 45 s, 72° C for 1 min; 60° C for 10 min, and DNA sequencing performed on an ABI3700 DNA sequencer (Applied Biosystems).
Sequences were aligned using Geneious Prime v.2020.0.2 and edges of incomplete sequences were trimmed to reduce missing data. The final alignment included 102 individuals with 634 BP and 277 variables sites. Models of sequence evolution were explored in jModel test v.2.1.10 (Darriba et al., 2012) and likelihood scores calculated using PhyML v.3.0 (Guindon et al., 2010). The model that best fits the data, TrN+I+G (Tamura and Nei, 1993) was identified using the Akaike Information Criterion (wAIC = 0.9915). Phylogenetic inferences were performed using two approaches. First, we used a Bayesian approach as implemented in BEAST v.2.6.0 (Bouckaert et al., 2014). A Yule prior was used with a uniform distribution. Two independent chains were run, each for 50 million steps sampled every 5000, and a burn-in equivalent to 25%. Convergence was assessed using effective sampling size in Tracer v.1.7.1 (Rambaut et al., 2018). Trees from the posterior distribution were summarized using a maximum clade credibility tree using mean node heights and a burning equivalent to 25% as implemented in Tree Annotator v.2.6.0 (Drummond and Rambaut, 2007). Second, we use a maximum likelihood (ML) approach as implemented in RAxML v.8 (Stamatakis, 2014). The best scoring ML tree was estimated using 20 replicates. Nodal support was calculated by 100 bootstrap replicates. Finally, the percentage of genetic divergence among species of the subgenus Chrysopteron was calculated using Nei's dA index (Nei and Kumar, 2000) as implemented in strataG v.2.4.905 (Archer et al., 2017). Only taxa with more than two samples were included in the analysis.
Morphological Analyses
We examined 34 specimens of adult Myotis (15 males, 16 females, 3 of unknown sex; appendix 1) as well as literature descriptions of all Chrysopteron species, and evaluated external and osteological characters including but not restricted to those discussed by Hill and Morris (1971), Smithers (1983), Koopman (1994), Taylor (2000), Monadjem et al. (2010), and Csorba et al. (2014). Dental homology nomenclature for premolars follows that of Csorba et al. (2014): 1st upper premolar (P2), 2nd or middle upper premolar (P3); 3rd upper premolar (P4), 1st lower premolar (p2), 2nd or middle lower premolar (p3), and 3rd lower premolar (p4). All measurements reported herein are from adult individuals with closed epiphyses. The first five measurements listed below were taken from skin labels or other records made by the original collector unless otherwise noted; all other measurements were taken by us using digital calipers and were recorded to the nearest 0.01 mm. Linear measurements are given in millimeters (mm), and weights are reported in grams (g). Descriptive statistics (mean and observed range) were calculated for all samples. Measurements are defined as follows:
Mass (M): Body weight in grams.
Total length (TL): Distance from the tip of the snout to the tip of the last caudal vertebra.
Length of tail (LT): Measured from the point of dorsal flexure of the tail with the sacrum to the tip of the last caudal vertebra.
Hind-foot length (HF): Measured from the anterior edge of the base of the calcar to the tip of the claw of the longest toe.
Ear length (Ear): Measured from the ear notch to the fleshy tip of the pinna.
Forearm length (FA): Distance from the elbow (tip of the olecranon process) to the wrist (including the carpals). This measurement is made with the wing at least partially folded.
Greatest length of skull (GLS): Distance from the posteriormost point on the occiput to the anteriormost point on the premaxilla (excluding the incisors).
Condyloincisive length (CIL): Distance between a line connecting the posteriormost margins of the occipital condyles and the anteriormost point on the upper incisors.
Condylobasal length (CBL): Distance between a line connecting the posteriormost margins of the occipital condyles and the anteriormost point on the skull not including the upper incisors.
Postorbital breadth (POB): Breadth across the narrowest portion of the postorbital region.
Braincase breadth (BB): Greatest breadth of the globular part of the braincase, excluding mastoid and paraoccipital processes.
Mastoid breadth (MB): Greatest breadth across the mastoid processes.
Zygomatic hreadth (ZB): Greatest breadth across the zygomatic processes.
Maxillary toothrow length (MTRL): Distance from the anteriormost surface of the upper canine to the posteriormost surface of the crown of M3.
Breadth across molars (BM): Greatest breadth across the outer margins of the upper molar teeth.
Breadth across canines (BC): Greatest breadth across the outer margins of the upper canines.
SYSTEMATICS
Family Vespertilionidae Gray, 1821
Subfamily Myotinae Tate 1942
Genus Myotis Kaup, 1829
Subgenus Chrysopteron Jentink, 1910
Myotis nimbaensis, new species
Nimba Myotis
Figures 3–8
Holotype: AMNH 279589, an adult male captured and released on 26 January 2018 at Kaiser Adit 3 and then collected on 2 February 2018 by Jon Flanders and Eric Moïse Bakwo Fils at Kaiser Adit 1 (fig. 2), Nimba Mountains, Guinea (N07.66499, W008.37223), field number BCIGU133. The specimen was preserved whole in formalin and stored in ethanol, and the skull was subsequently extracted and cleaned. A 3 mm wing biopsy tissue sample was collected and stored in desiccant before the DNA was extracted at CIBIO-InBIO, University of Porto, Portugal.
Paratype: AMNH 279590, an adult female collected at the same time and place as the holotype, field number BCIGU132. Like the holotype, the paratype was preserved in formalin and stored in ethanol after a 3 mm wing biopsy tissue sample was collected and stored in desiccant before the DNA was extracted. The skull was subsequently extracted and cleaned.
Other records: In addition to the captures of the holotype and paratype, we recorded echolocation calls from the male holotype on 26 January 2018. Once its distinctive call parameters were identified, we searched for its echolocation call signature in sound files recorded at the entrances of 10 mine adits between 2018–2019, including the type locality, within the Nimba Mountains' Mining Concession in Guinea (fig. 1), using Song Meter SM4BAT acoustic detectors (Wildlife Acoustics). Echolocation calls closely matching those recorded from the release calls of the holotype were detected at the entrances of a total of five mine adits (fig. 1) with seasonal variation between sites and the highest activity rates recorded at the type locality. While the Nimba Mountains support an exceptionally diverse bat fauna, this is the first Myotis species observed using the adits and only the second Myotis species to be recorded across the whole mountain range (Monadjem et al., 2016). Details of call frequency and structure, as well as comparisons with calls of congeneric species, are provided below in the section entitled Echolocation Calls.
Distribution: Known only from the type locality and vicinity in the Guinean Nimba Mountains.
Etymology: Myotis nimbaensis (“from Nimba”) is named in recognition of the mountain range in which it was discovered. As an epithet referring to a place, nimbaensis is spelled the same way whether applied in combination with either a masculine or a feminine genus name. Woodman (1993) argued that Myotis should be considered feminine in gender, but Pritchard (1994) disagreed. Both of these authors overlooked a 1958 ruling by the International Commission on Zoological Nomenclature that fixed the gender of Myotis as masculine and placed the name as such on the Official List of Generic Names in Zoology (International Commission on Zoological Nomenclature, 1958). So in this case, nimbaensis is masculine.
Diagnosis and Description: Myotis nimbaensis is diagnosed by a combination of the following characteristics: large size (FA 52.4–55.2 mm, mass 15.5–17.0 g; table 1) with females apparently somewhat larger than males; dorsal fur bright orange with strongly tricolored hairs (basal 1/3 of hair shaft black, middle 1/3 of shaft creamy white, distal 1/3 and tip bright orange to coppery red); ventral fur paler than dorsal fur, tricolored on belly (black base, buff-colored shaft that grades to orange at tip) grading to bicolored (lacking black base) on the sides near the wing membranes; ruff of brighter fur present around neck; pointed face with pale skin covered with orange fur, skin clearly visible through the fur around eye, mouth, and on rostrum; no black spots present on face; pinna with a rounded tip, relatively long and reaching beyond tip of nose when laid forward; pinna with strong distal emargination; skin of pinna pale orange brown becoming slightly darker near the tip, no rim of black around ear; tragus lanceolate, slightly less than half the length of the pinna; thumb brown; wing membranes strongly dichromatic black and orange, with plagiopatagium and dactylopatagium mostly black with narrow bands of orange along the metacarpals, phalanges, and across the membrane behind the forearm and upper arm; black pigmentation extends medially nearly to body and posteromedially to tibia; membrane between metacarpals I and II pale orange; propatagium pale orange; uropatagium orangish brown with orange fur on the proximal 1/3 of the dorsal surface and pale cream-colored fur on proximal 1/5 of ventral surface; no black spots on patagia; no conspicuous gland present in plagiopatagium behind humerus; pale ventral fur extends onto proximal plagiopatagium in narrow strip along body, does not extend past knee; relatively small foot with length approximately 2/5 of tibia length; foot and toes brown, with sparse long brown hairs on the dorsal surface of each toe; wing membrane essentially naked except close to the body, where it has a sparse covering of cream colored hairs on the ventral surface; wing membrane attaches to foot at base of first toe; calcar long, more than twice the length of the hind foot, runs approximately 2/3 of the length of the uropatagium border; skull large for Myotis (GLS = 19.48–19.73 mm; CIL = 18.92–18.93 mm; CBL = 18.31–18.81 mm) with globular braincase, well-defined sloping forehead, and elongated supraorbital region all contributing to the appearance of a clearly defined and elongate rostrum; rostrum with shallow medial depression in nasal region; sagittal crest moderately well developed in both sexes; lambdoid crest weakly developed in both sexes; dental formula I2/3, C1/1, P 3/3, M3/3 = 38; I1 and I2 each with a well-developed posterior cusp; upper canine robust, crown height approximately 2× that of P4, basal area (as seen in occlusal view) roughly equal to that of P4; anterior two premolars (P2 and P3) much smaller than last premolar (P4); P3 with less than half the basal area of P2 and shifted lingually to be partly excluded from the toothrow although still visible in lateral view; P4 large, sharply pointed, taller than M1; lower incisors each with 4 main cusps; i3 with distal accessory cuspules; p3 relatively large, approximately three quarters the basal area of p2, and fully in line in toothrow, not shifted lingually; lower molars myotodont.
Comparisons: External and craniodental measurements for Myotis nimbaensis and other congeneric African species with which it might be confused are provided in table 1. In addition to M. nimbaensis, three other large Myotis (subgenus Chrysopteron) species with orange dorsal fur occur in Africa: M. tricolor, M. welwitschii, and M. morrisi. Descriptions and measurements of these species can be found in Hill and Morris (1971), Hill et al. (1988), Taylor (2000), and Monadjem et al. (2010); the latter work also includes color photographs and echolocation call spectrograms.
Myotis nimbaensis can be easily distinguished from M. morrisi on the basis of size, with M. morrisi being smaller in all external and craniodental dimensions (table 1). The ventral pelage of M. morrisi is also different—rather than being tricolored as in M. nimbaensis, M. morrisi has unicolored dull cream white ventral fur that is tinged a faint brown on the flanks and chin. The skull and dentition of M. morrisi is very similar to that of M. nimbaensis although the braincase of M. morrisi is somewhat more globular and lacks the sagittal crest seen in M. nimbaensis.
Myotis nimbaensis is similar in size and general coloration to the widespread species M. welwitschii, but these taxa can be easily distinguished because M. nimbaensis completely lacks the prominent black spots seen on the face and patagia of M. welwitschii (figs. 3–6). Color of the pinnae is also different—bright orange to coppery red with a black edge in M. welwitschii compared to pale orange brown with no black edge in M. nimbaensis. The thumb in M. welwitschii is bright orange whereas it is brown in M. nimbaensis. Both species are strongly dichromatic with black wing membranes and orange along the digits and forearm, but pigmentation of the plagiopatagium near the body and hind legs is different in the two species. Myotis welwitschii has a broad band of orange along the side of the body that extends anteriorly past the elbow to run along the forearm and posteriorly to the ankle, so black pigmentation is limited to the more distal and posterior portions of the plagiopatagium. In contrast, the orange is much less extensive and the black is more extensive in M. nimbaensis; the black pigmentation approaches the body wall, extending anteriorly into the plagiopatagium behind the humerus and posteriorly to the tibia in M. nimbaensis, so there is no broad band of orange along the body in that species. The uropatagium in M. welwitschii is also bright orange whereas it is darker and more brownish in M. nimbaensis. There are additional differences in features of the dorsal and ventral pelage—the dorsal fur in M. welwitschii is bicolored (black at base and otherwise coppery red) rather than being tricolored as in M. nimbaensis, and M. welwitschii has unicolored ventral pelage (whitish tinged with coppery red) rather than the distinctive tricolored ventral fur in seen in M. nimbaensis.
Craniodental measurements (table 1) suggest overall similarity of M. nimbaensis with M. welwitschii; small differences in a few dimensions are not interpretable given the small sample sizes. Both species have lambdoidal crests that are weakly developed but clearly visible, and moderately well-developed sagittal crests in both sexes. The dentition of M. welwitschii includes two very small anterior upper premolars that are both included in the toothrow; P3 is smaller than P2, but P3 is not displaced lingually as it is in M. nimbaensis.
Myotis tricolor is a widespread species that shows considerable morphometric variation across its range (Koopman, 1989, 1994; Taylor, 2000; Monadjem et al., 2010). Myotis nimbaensis is roughly equivalent in overall size (e.g., forearm length and body mass) to the largest individuals of M. tricolor reported in the literature, although none of the specimens of M. tricolor that we measured were as large as M. nimbaensis. The only known specimen of M. tricolor from West Africa—a male from Liberia referred to that species by Koopman (1989)—is much smaller than M. nimbaensis, having a forearm length of 46.0 mm compared to 52.4 mm for the male paratype of M. nimbaensis. The pelage of M. tricolor is similar to that of M. nimbaensis, and M. tricolor similarly lacks black spots on the face or tail membrane. However, these species can be distinguished easily based on coloration of the wing membranes: M. tricolor has wings that are dark brown rather than the distinctive dichromatic orange and black wings seen in M. nimbaensis. The pinnae of M. tricolor are dark brown compared with the pale orange brown seen in M. nimbaensis, and the emargination is more proximally located (nearly at the midpoint of the lateral edge of the pinna) in M. tricolor compared to the clearly distal location of the emargination in M. nimbaensis. The uropatagium of M. tricolor is also darker brown (rather than orange brown as in M. nimbaensis) and has a dense covering of coppery red fur proximately (more sparse in M. nimbaensis). In terms of craniodental features, M. nimbaensis has a greater condylobasal length (18.9 mm for both sexes) than seen in M. tricolor (15.9–18.5 mm including data from table 1 and Monadjem et al., 2010). Although our comparative sample of M. tricolor was not large, we found that most other craniodental measurements similarly appear to distinguish M. nimbaensis and M. tricolor, with M. nimbaensis the larger of the two species.
Sagittal crest development in Myotis tricolor is variable, with females typically lacking any sagittal crest development. In contrast, both male and female M. nimbaensis have well-developed sagittal crests (though again, our sample size is small). Upper premolars in M. tricolor are somewhat variable, but P3 is always less than half the basal area of P2 and is always shifted lingually as in M. nimbaensis. However, in many individuals the reduction in P3 is greater than in M. nimbaensis, and this tooth is completely excluded from the toothrow (only partly excluded in M. nimbaensis). The p3 of M. tricolor is one half to two thirds the basal area of p2, and p3 is sometimes displaced lingually from the toothrow, compared to a slightly larger p3 that is at least three quarters the size of p2 and which is not displaced in M. nimbaensis.
Another orange Myotis species from Africa is M. bocagii, which occurs across much of tropical Africa and is broadly sympatric with M. tricolor and M. welwitschii (Monadjem and Jacobs, 2017a; Patterson, 2019). Myotis bocagii can be easily distinguished from M. nimbaensis based on its dark brown wings and considerably smaller size in all dimensions (e.g., FA = 36.0–40.5 mm; mass = 6.0–9.0 g.; CBL = 13.0–15.0 mm; CIL = 14.0–15.2 mm; Koopman, 1994; Monadjem et al., 2010).
Myotis (subgenus Chrysopteron) species with yellow or orange dorsal fur and dichromatic orange and black wings are not limited to Africa, but also occur in Asia. At least six Asian species (M. bartelsi, M. formosus, M. hermani, M. rufoniger, M. rufopictus, and M. weberi) exhibit color patterns with dichromatic wing pigmentation superficially similar to that of M. nimbaensis (Csorba et al., 2014). Csorba et al. (2014) gave diagnoses and measurements of these species that provide a basis for distinguishing them from M. nimbaensis. These authors recognized two pelage and skin color patterns among the Asian species: a “rufoniger type” and a “formosus type.” Species exhibiting “rufoniger type” morphology include M. bartelsi, M. hermani, M. rufoniger, and M. weberi. These species can be distinguished from M. nimbaensis based on darker dorsal pelage that has four bands of color (individual dorsal hairs black basally, pale yellow distally, then darkening to deep red before terminating in a black tip; Csorba et al., 2014) rather than the bright orange tricolored dorsal fur of M. nimbaensis. Species with “rufoniger type” coloration are also considerably darker ventrally than M. nimbaensis, having hairs with a black base, followed by either a pale yellowish section that progressively darkens distally to deep red, or are otherwise entirely deep red (Csorba et al., 2014). The ear is also edged with black in rufoniger-type species (Csorba et al., 2014), a trait lacking in M. nimbaensis (or any other African species of Chrysopteron). The thumb and underside of hind foot are entirely black in species with “rufoniger type” coloration (Csorba et al., 2014) but not in M. nimbaensis, which has an orange-brown thumb and a brown foot. Myotis rufoniger has flight membranes with a broad band of red that extends from the ankle to the forearm along the side of the body, and black pigmentation is limited to the more distal and posterior portions of the plagiopatagium (Bhak et al., 2017). In contrast, the black pigment is more extensive in M. nimbaensis, extending much closer to the body including reaching the tibia and into the membrane behind the humerus, and the pale portions of the wings are more orange than red.
Members of the Myotis rufoniger group vary somewhat in size, but they are all large bats roughly comparable in size to M. nimbaensis. Measured forearm lengths of M. rufoniger (FA = 45.0–56.0; Csorba et al., 2014) overlap with those of M. nimbaensis, but cranial dimensions of M. rufoniger (e.g., GLS = 16.98–19.24 mm; ZB = 10.04–12.24 mm) do not overlap, with M. nimbaensis being the slightly larger species. Dental morphology is also somewhat different; M. rufoniger has a P3 that is roughly two thirds the size of P2 and which is usually in line with the other check teeth (rarely displaced lingually; Csorba et al., 2014), compared with M. nimbaensis, in which P3 is less than half the size of P2 and is lingually displaced. Considerable overlap in most dimensions exists between M. nimbaensis and M. weberi (FA = 49.7–53.5 mm; GLS = 19.15–19.72 mm; ZB = 12.37–12.67 mm; Csorba et al., 2014). However, M. weberi as described by Csorba et al. (2014) can be distinguished by having a relatively short upper canine that is only about 1.5× the height of P4 (C height is ∼2× height of P4 in M. nimbaensis) and a p3 that is no more than half the basal area of p2 (p3 is at least three quarters the basal area of p2 in M. nimbaensis).
In contrast to Myotis rufoniger and M. weberi, both M. bartelsi (FA = 53.4 mm; GLS = 20.42 mm; ZB = 13.41 mm) and M. hermani (FA = 56.0–60.0 mm; GLS = 20.10–21.77 mm; ZB = 13.4–14.10 mm) are somewhat larger than M. nimbaensis in most dimensions (Csorba et al., 2014). As described by Csorba et al. (2014), M. bartelsi can be distinguished by having strong lambdoid crests (weakly developed in M. nimbaensis) and a P3 that is fully out of line with the rest of the toothrow and not visible in lateral view (partially out of line and visible in lateral view in M. nimbaensis). The forehead is also less clearly developed and strongly sloped in M. bartelsi compared to M. nimbaensis, and the rostrum appears somewhat shorter in M. bartelsi in lateral view. Myotis hermani as described by Csorba et al. (2014) can be distinguished by its overall very robust skull and exceptionally well-developed sagittal and lambdoid crests (more weakly developed in M. nimbaensis), very large upper canine with a basal area exceeding that of P4 (basal areas of C and P4 subequal in M. nimbaensis), minute P3 less than one quarter the size of P2 (P3 only slightly less than one half the size of P2 in M. nimbaensis), P3 fully out of line with the rest of the toothrow and not visible in lateral view (only partially out of line and visible in lateral view in M. nimbaensis), p3 with half the basal area of p2 (p3 is at least three quarters the basal area of p2 in M. nimbaensis), and p3 partly lingually displaced out of the toothrow (p3 fully in toothrow in M. nimbaensis).
Asian Myotis (subgenus Chrysopteron) species exhibiting “formosus type” external morphology are M. formosus and M. rufopictus (Csorba et al., 2014). These species can be distinguished from M. nimbaensis based on pelage color and banding. The dorsal fur of formosus-type species has individual hairs that have a very narrow brown base and are pale yellow distally for 80–100% of their length, or grade into a brown tip, with banding sometimes not evident as the color changes gradually (Csorba et al., 2014). Overall, the general aspect of the dorsal fur in M. formosus and M. rufopictus is pale yellow-brown (Csorba et al., 2014). This pattern is very different from the distinctly banded tricolored orange fur of M. nimbaensis. Ventral fur of M. formosus and M. rufopictus is either bicolored light yellow with a narrow brown base, or unicolored light yellow (Csorba et al., 2014), again quite different from tricolored buff to orange ventral fur of M. nimbaensis. The ear is faintly edged with black in the formosus-type species (Csorba et al., 2014) but not in M. nimbaensis. As in M. rufoniger, the dark pigment in the wing membranes of M. formosus is limited to more distal and posterior portions of the plagiopatagium and does not extend to the tibia and membranes next to the body as it does in M. nimbaensis.
Measurements and craniodental features also distinguish Myotis nimbaensis from M. formosus and M. rufopictus. In terms of measurements, M. nimbaensis is only trivially different or falls within the range of variation in measurements reported for M. formosus (FA = 45.5–53.0 mm; GLS = 17.91–19.45 mm; ZB = 11.76–12.94 mm; Csorba et al., 2014). However, M. formosus as described by Csorba et al. (2014) can be distinguished based on having a sagittal crest that is missing or only very weakly developed (moderately well developed in M. nimbaensis), P3 that is not visible in lateral view of the skull (visible in M. nimbaensis), and a p3 with half the basal area of p2 (p3 is at least three quarters the basal area of p2 in M. nimbaensis). In addition, the skull of M. formosus lacks an expanded supraorbital region and appears to have a shorter rostrum than M. nimbaensis when seen in lateral view. Myotis rufopictus (FA = 51.0–52.5 mm; GLS = 18.2 mm; ZB = 11.17; Csorba et al., 2014) appears slightly smaller than M. nimbaensis, but both species are known from very few specimens, so the value of this observation is questionable. As described by Csorba et al. (2014), the skull profile of M. rufopictus ascends almost evenly with no frontal depression (i.e., no clearly defined break between rostrum and sloping forehead as seen in M. nimbaensis). Like Myotis formosus, M. rufopictus additionally lacks an expanded supraorbital region and it appears to have a shorter rostrum than M. nimbaensis when seen in lateral view. Morphology of P3 in M. rufopictus is unknown, but p3 is a very small tooth less than one quarter the basal area of p2 and displaced lingually halfway out of the line of toothrow (p3 at least three quarters the area p2 and not displaced in M. nimbaensis).
Echolocation calls: Based on the echolocation calls recorded on the release of the holotype (n = 22), Myotis nimbaensis emits a broad bandwidth (84.1 ± 13.1 kHz; FMAX: 104 kHz, FMIN: 20 kHz) steep frequency-modulated pulse with a peak frequency of 42.5 kHz (± 4.2 kHz), an average call duration of 3.5 ms (± 0.3 ms), and an interpulse interval of 113 ms (± 5.3 ms; fig. 9). In comparison, echolocation calls of Myotis welwitschii have a lower peak frequency (34 kHz), shorter bandwidth (28.3 kHz) and shorter duration (2.4 ms; Schoeman and Jacobs, 2008: Monadjem et al., 2010: Collen, 2012). While there may be geographic variation in call parameters between East African and Southern African Myotis tricolor populations, calls recorded from Southern Africa have a slightly higher peak frequency (47.8±3.1 kHz) but are shorter in bandwidth (46±23.9 kHz) and duration (3.3±0.6 ms; Schoeman and Jacobs, 2008: Monadjem et al., 2010). Echolocation calls of Myotis morrisi are unknown. Myotis bocagii, the other known Myotis species found on the Nimba Mountains, have calls with a similar peak frequency (41–43 kHz) but are shorter in bandwidth (36 kHz) and duration (2–2.8 kHz) (Schoeman and Waddington, 2011; Collen, 2012; Monadjem et al., 2017a).
Molecular Analyses: Molecular analyses of mitochondrial cytochrome b sequences from selected specimens (appendix 2) support recognition of Myotis nimbaensis as a distinct species (table 2, figs. 10 and 11). Our analyses recovered well-supported clades consistent with results of Csorba et al. (2014) and Patterson et al. (2019), and in this context Myotis nimbaensis was found to represent a distinct lineage that nests within the subgenus Chrysopteron as sister to the M. tricolor complex (the clade consisting of M. tricolor 1, 2, and 3 of Patterson et al., 2019). Monophyly of Myotis nimbaensis, the M. tricolor complex, and a clade consisting of these two lineages was highly supported (>95% bootstrap). Myotis nimbaensis is 5.18% different than M. tricolor 1, 5.43% from M. tricolor 2, and 5.08% than M. tricolor 3 (table 2).
Natural History: The type locality habitat of Myotis nimbaensis occurs at the ecotone of high-altitude grassland, montane savanna, and gallery forest habitats of the Nimba Mountains around 1400 m elevation (fig. 1). As far as is known, the preferred (or exclusive) roosting habitat for this species is subterranean features. The two individuals that we captured were emerging from day roosting in an abandoned mine adit (fig. 1), and echolocation calls similar to the release calls of Myotis nimbaensis recorded at other nearby adits, apparently confirming association of M. nimbaensis with these underground structures. Myotis nimbaensis coroosts with other subterranean roosting bats and was detected at mine adits that house colonies of several hundred Rhinolophus guineensis and Hipposideros lamottei, the latter a bat species endemic to the Nimba Mountains (Monadjem et al., 2013).
Foraging habitat for Myotis nimbaensis is unknown, but the six mine adits where M. nimbaensis echolocation calls were recorded occur in similar ecotone habitats at the interface of high-elevation grassland, montane savanna, and gallery forests along headwaters of the Zié, Gouan, and Zougué rivers in the northern Guinean Nimba Mountains.
FIGURE 3.
Portrait of Myotis nimbaensis (AMNH 279589, holotype). A, View of right side of upper body and head showing the pale ventral fur and bright orange fur on the head and the ruff around the neck; also note the brown color of the thumb. B, Anterior view of the left ear showing the pale orange-brown color of the pinna, strong distal emargination, and rounded pinna tip. C, Anterior view of right ear showing the ridges in the pinna and the relative length of the lanceolate tragus, which is slightly less than half the length of the pinna. D, Close-up view of the right side of the head showing the pale skin visible through the fur around eye, mouth, and on the rostrum; also note the strongly tricolored fur on the top of the head.
TABLE 1.
Comparative measurements of Myotis nimbaensis, n. sp., and related African species. Mass (M) is in grams, all other measurements are in mm; see Materials and Methods for descriptions of measurements. Summary statistics (mean, observed range in parentheses, and sample size) are given for samples that include multiple individuals.
FIGURE 4.
Fur color and banding in Myotis nimbaensis (AMNH 279589, holotype). A, View of the dorsal fur showing the overall bright orange coloration; creamy white bands on the hairs proximal to the bright orange fur tips are visible on the lower right near the leg where the fur has been somewhat disturbed. B, Close-up of dorsal fur over the lower back showing tricolored fur in a spot on the center left where the fur has been slightly teased apart; where the fur is undisturbed, the banding of individual hairs is not visible. C, dorsal fur over lower back clearly showing the tricolored banding in an area where the fur has been separated by blowing. D, Ventral fur over torso showing overall paler coloration than dorsal fur, and tricolored banding of the hairs with less orange at the tips. E, Ventral fur over left side of the thorax showing bicolored banding (lack of a black basal band) near the wing membrane.
FIGURE 5.
Coloration of the wing membranes in Myotis nimbaensis (AMNH 279589, holotype). A, Dorsal surface of the wing showing the dichromatic black and orange skin coloration. The plagiopatagium and dactylopagial membranes are mostly black with thin orange bands along the metacarpals, phalanges, and forearm; the black pigmentation also extends nearly to the body wall in the area between the forearm and the hind leg. The propatagium is pale orange. B, Dorsal surface of the anterior and proximal portion of the wing showing the patterning of the black pigmentation near the body. C, Dorsal surface of the distal wing showing the brown thumb and orange (not black) pigmentation of the membrane between digits II and III.
FIGURE 6.
Uropatagium, foot, calcar, and flight membrane attachments in Myotis nimbaensis (AMNH 279589, holotype). A, Dorsal side of the uropatagium showing orangish brown skin with orange fur on the proximal 1/3 of the surface of the membrane. B, Ventral side of uropatagium showing pale cream-colored fur on proximal 1/5 of ventral membrane surface; this fur continues across the femur onto the proximal plagiopatagium in a narrow strip that does not extend past the knee. C, Dorsal view of distal leg, foot, calcar, and associated flight membranes. Note that the wing membrane (on right of the foot) attaches to the foot at the base of the first toe, and the calcar (on the left of the foot) is more than twice the length of the hind foot. D, Close-up of the dorsal surface of the foot showing sparse, long brown hairs on each toe. E, Close-up of the ventral side of foot and toes showing dark brown coloration. F, Ventral view of the hind leg and foot showing the relatively small foot size (foot length = 2/5 of tibia length) and the patterning of black membrane pigmentation near the leg and foot.
FIGURE 7.
Skull and jaws of Myotis nimbaensis (AMNH 279589, holotype): A, dorsal view of skull; B, ventral view of skull; C, lateral view of skull and lower jaws. (Drawings by Patricia J. Wynne.)
DISCUSSION
Discovery of a new large Myotis in western Africa is perhaps not surprising given the complexity of habitat types and relative lack of exploratory studies across much of the region. Moreover, considering the large gaps in the ranges of many known species, it seems likely that many populations remain unsampled, and other new species will be discovered in future years. Little is known about the habits of Myotis nimbaensis other than characteristics of the type locality and roosting habits at the type locality. However, this information provides a basis for comparisons with other species in the subgenus Chrysopteron.
Habitat use and Roosting habits in African Chrysopteron: Comparisons of the natural history of M. nimbaensis with that of similar African congeners is hindered by lack of information about these unusual bats. The best known of these is Myotis tricolor, a taxon known from Ethiopia, Liberia, and the Democratic Republic of Congo south to South Africa (Simmons, 2005; Monadjem and Jacobs, 2017b; Patterson et al., 2019). Myotis tricolor as currently recognized apparently represents a species complex (see Patterson et al., 2019; figs. 10, 11), and most of what is known about these bats is based on observations of South African populations. The M. tricolor complex occurs primarily in savanna woodland in mountainous areas, penetrating into drier, more open terrain in the southern part of their range (Smithers, 1983; Taylor, 2000; Monadjem et al., 2010). These bats apparently roost exclusively underground in caves and mines, where they are gregarious and roost in tight clusters (Taylor, 1998, 2000; Monadjem et al., 2010). Average group size in South Africa is apparently a few dozen individuals, but M. tricolor may congregate in groups as large as 1400 to 2000 individuals (McDonald et al., 1990; Taylor, 1998, 2000; Monadjem et al., 2010). Myotis tricolor may share roosts with Miniopterus species (Taylor, 1998, 2000). Myotis tricolor is known to be migratory in parts of South Africa, flying hundreds of kilometers between summer maternity caves and winter hibernation caves (Taylor, 1998, 2000). Smithers (1983) and Monadjem et al. (2010) speculated that the occurrence of M. tricolor is probably governed more by the availability of caves and mine adits than by the vegetational associations or terrain in which they are found. In this respect M. nimbaensis may be similar. Despite numerous bat surveys in the Nimba region (summarized by Monadjem et al., 2016), the only place this species has been found is in association with mine adits in the Mining Concession area in the Guinean Nimba Mountains. We recorded echolocation activity at sunset and sunrise at mine adit entrances, indicating M. nimbaensis regularly roosts in abandoned adits. Colony size of M. nimbaensis may be small, as small as a few to single individuals, based on typically very low levels of echolocation activity at mine entrances where the species was detected. With known occurrence at only a single locality at 1400 m of elevation, M. nimbaensis is likely much more of a habitat specialist than are members of the M. tricolor complex, which appear to have a broader geographic range.
TABLE 2.
Percentage of sequence divergence among species of the subgenus Chrysopteron based on Nei's dA index and 634 bp of mitochondrial cytochrome b. Values of divergence between Myotis nimbaensis and closely related species are shown in bold. Myotis tricolor and M. welwitchii groups are clades identified by Patterson et al. (2019) as follows: Myotis tricolor 1 includes populations from Kenya and Uganda; M. tricolor 2 and M. tricolor 3 are both from South Africa; M. welwitschii 1 is from Kenya and Uganda; and M. welwitschii 2 includes populations from Tanzania, Malawi, and South Africa.
Myotis welwitschii (known from Ethiopia and the Democratic Republic of Congo south to South Africa; Simmons, 2005; Monadjem et al., 2017b; Patterson et al., 2019) is thought to be more of a habitat specialist tied to paramontane areas covered by woodland or woodland-forest mosaic vegetation; it has been captured also in riverine coastal forest adjacent to sugar cane fields and open thornveld (Smithers, 1983; Taylor, 1991, 1998; 2000; Fahr and Ebigbo, 2003). Myotis welwitschii is thought to be a solitary species (Taylor, 1991, 2000; Ratcliffe, 2002). Roosting habits of M. welwitschii are poorly known, but this species has been found roosting in vegetation including low bushes and shrubs, trees, furled banana leaves, and a single individual has also been captured in a cave (Smithers and Wilson, 1979; Rautenbach, 1982; Smithers, 1983; Skinner and Smithers, 1990; Taylor, 1998, 2000; Monadjem et al., 2010). At least two specimens of M. welwitschii were collected hanging in the open during the day, and in at least one case the bat was mistaken for a dead leaf before being revealed as a roosting bat (Taylor, 2000). It is possible that the unusual fur color and dichromatic wings of this species help to provide camouflage under such conditions.
Myotis morrisi is a rare species known only from Ethiopia and Nigeria (Hill and Morris, 1971; Hill et al., 1988; Simmons, 2005), a range that suggests either that many populations remain unsampled or that it represents a species complex. The holotype of M. morrisi was collected in 1968 on the Great Abbai Expedition in Ethiopia, where it was captured over the Blue Nile about 1 m above the water (Hill and Morris, 1971). The surrounding habitat was mostly maize fields and thick brush (Hill and Morris, 1971). The Nigerian specimen of M. morrisi was captured by mistnet in the Sudan savannah in a region of short grass and acacia bush, with some silk cotton trees, near the River Benue (Hill et al., 1988). Nothing is known about the roosting habits of this species.
Foraging habits: Three ecomorphs of Myotis are generally recognized: aerial netters (also called aerial hawkers), gleaners, and trawlers (Findley, 1972; Koopman, 1994; Ruedi and Mayer, 2001; Morales et al., 2019). Each of these is characterized by a suite of phenotypic traits involving ear length, skull shape, wing length and width, leg and foot length, and attachment site of the posterior plagiopatagium (Morales et al., 2019). Morales et al. (2019) used these traits to determine the ecomorph class for over 90 Myotis species, and they classified the majority of Chrysopteron species as gleaners including Myotis welwitschii, M. tricolor, M. rufoniger, M. formosus, M. emarginatus, and M. goudoti. Myotis scotti was classified as an aerial netter, and M. bocagii as a trawler (Morales et al., 2019). Given the morphological similarity of Myotis nimbaensis to M. welwitschii, M. tricolor, M. rufoniger, and M. formosus, we might expect that M. nimbaensis is also a gleaner. However, the usefulness of these categorizations for Chrysopteron species remains to be determined. Stoffberg and Jacobs (2004) tested the abilities of captive Myotis tricolor to catch prey presented in a variety of ways, and they found that while this species easily captured aerial prey, it did not glean. Those authors noted that M. tricolor has more pointed wings than many bats known to glean, and also varies from known gleaners in the genus Myotis (e.g., M. lucifugus and M. septentrionalis) in using short bandwidth echolocation calls that typically lack harmonics. In contrast, M. septentrionalis consistently produces two harmonics in addition to the dominant one, and M. lucifugus produces one harmonic in addition to the dominant harmonic (Stoffberg and Jacobs, 2004) effectively making the bandwidth of their calls even broader. While we did not detect any harmonics in the release calls of M. nimbaensis, we found that it produced broader bandwidth calls than those of Myotis tricolor. Although its relatively long call duration may suggest M. nimbaensis is not specialized for gleaning, it is similar to that of Myotis lucifugus, which is both an aerial hunter and gleaner (Ratcliffe and Dawson 2003).
Details of foraging habits and diet of African Chrysopteron are poorly known. Nothing is known of the diet of M. morrisi, and a description of M. welwitschii feeding on small beetles is based on fecal pellets from a single individual (Moratelli, 2019). Captive individuals apparently prefer soft-shelled insects (Happold and Happold, 2013). Myotis welwitschii has been observed entering houses at night while foraging (Skinner and Smithers, 1990; Taylor, 1998) but has been also been observed flying low over streams and a farm pool, suggesting that it may also forage by trawling over open water (Ratcliffe, 2002; Moratelli, 2019). Myotis tricolor in South Africa feeds primarily on hard-bodied prey with coleopterans making up the greatest proportion of the diet, followed by hemipterans (Stoffberg and Jacobs, 2004). There is no evidence that they feed on lepidopterans, and no insects or other arthropods that are potential substrate prey have been found in dietary analyses based on fecal samples (Stoffberg and Jacobs, 2004). The diet of M. nimbaensis remains unknown.
FIGURE 10.
Maximum likelihood phylogenetic reconstruction of subgenus Chrysopteron using an alignment of 634 base pairs of mitochondrial gene cytochrome b. Colored circles at nodes represent support values as bootstrap percentage from maximum likelihood analyses. Support values lower than 50% at shallow nodes are not shown. Tip labels indicate GenBank accession number and locality. Myotis tricolor 1, 2, and 3 and M. welwitschii 1 and 2 are labeled following Patterson et al. (2019).
Phylogenetic relationships, biogeography, and evolution: Molecular phylogenies indicate that Chrysopteron split from other Old World Myotis early in the radiation of the genus (Stadelmann et al., 2004; Morales et al., 2019). Stadelmann et al. (2004) dated this split to the middle Miocene (∼10–13 Mya), but a slightly older Miocene date (∼16–18 Mya) was recently proposed using revised calibration points and a more extensive data set (Morales et al., 2019). Not surprisingly given the overlap of our data sets, our phylogenetic results echoed those of Csorba et al. (2014) and Patterson et al. (2019) in confirming monophyly of Chrysopteron with moderate support. Within this clade, each of the currently recognized species (sensu Csorba et al., 2014) was found to be distinct and, if represented by more than one sequence, monophyletic (figs. 10, 11). Although support values for some basal relationships with Chrysopteron were not high, some biogeographically and morphologically interesting clades were strongly supported in our analyses. Two dichromatic species from different continents, Myotis rufoniger (Asia) and the M. welwitschii complex (Africa) were found to be sister taxa, suggesting that they shared a dichromatic common ancestor. Rather than grouping with this clade, the two other dichromatic species, M. nimbaensis and M. formosus, group in different parts of the tree. Myotis nimbaensis is strongly supported as the sister taxon of the M. tricolor species complex, which does not have dichromatic coloring although both are from mainland Africa; Myotis formosus from Asia forms a weakly supported clade with nondichromatic M. emarginatus from the Mediterranean region. Another strongly supported clade within Chrysopteron is comprised of monochromatic species known only from Africa and nearby Indian Ocean islands (Myotis bocagii, M. goudoti, and M. anjouanensis). Although low support values limit the value of further interpretations, it seems clear that dichromatic coloration evolved at least twice and possibly three or four times in Chrysopteron, and that intercontinental dispersal events occurred at least twice in this subgenus although directionality of these events cannot be determined with any certainty given the data at hand. Additional data from genes capable of resolving more basal relationships within Chrysopteron will be necessary to fully resolve these questions.
The Nimba Mountains as a Center of Bat Diversity and Endemism: The Nimba Mountains are well known as a center of biodiversity and endemism (Brosset, 2003; Lamotte and Roy, 2003; Wieringa and Poorter, 2004; Sandberger et al., 2010; Denys and Aniskine, 2012; Denys et al., 2013; Marshall and Hawthorne, 2013; Monadjem et al., 2013, 2016, 2019). As summarized by Monadjem et al. (2016), many surveys of the bats of the Nimba region have been conducted over the past several decades, with the majority of work done on the Liberian side of the mountain range. The discovery of Myotis nimbaensis brings the known number of bat species in the Nimba Mountains to 62 species (Monadjem et al., 2016, 2019). Five new species of bats have been described from the region in the last decade: M. nimbaensis (this study), Neoromicia roseveari (Monadjem et al., 2013), N. isabella (Decher et al., 2015), Parahypsugo happoldorum (Hutterer et al., 2019), and Miniopteris nimbae (Monadjem et al., 2019). Two species are endemic to the mountain range (Myotis nimbaensis and Hipposideros lamottei) and two more are highly restricted regional highland endemics (Hipposideros marisae and Rhinolophus ziama; Monadjem et al., 2016). Hipposideros marisae is known only from highland regions of Guinea, Liberia, and Côte d'Ivoire (Fahr, 2013), and was observed by the authors in 2018 roosting in a small natural cave at 491 m in the buffer zone of the World Heritage Site in the Guinean Nimba Mountains (unpublished data). Similarly, Rhinolophus ziama is restricted to Guinea and Liberia and likely a specialist restricted to mountain habitats (Fahr et al., 2002; Simmons and Cirranello, 2020). Given this pattern, if Myotis nimbaensis has a range extending beyond the Nimba Mountains, we would expect it to similarly occur in highland regions elsewhere in Guinea, Liberia, and Côte d'Ivoire. Acoustic surveys of these areas would be a good way to determine if M. nimbaensis has a broader ranger or is an endemic restricted in range to only to the Nimba Mountains.
FIGURE 11.
Bayesian phylogenetic reconstruction of subgenus Chrysopteron using an alignment of 634 base pairs of mitochondrial gene cytochrome b. Colored circles at nodes represent support values as posterior probability from Bayesian analyses. Support values lower than 50% at shallow nodes are not shown. Tip labels indicate GenBank accession number and locality. Myotis tricolor 1, 2, and 3 and M. welwitschii 1 and 2 are labeled following Patterson et al. (2019).
Two species coroosting in mine adits on the Nimba Mountains with M. nimbaensis are considered threatened by IUCN: Rhinolophus guineensis (IUCN status: Endangered) and Hipposideros lamottei (IUCN status: Critically Endangered) (Shapiro and Cooper-Bohannon, 2020; Mickleburgh et al., 2008). In addition, two other species occurring in the Nimba range that have been assessed by IUCN are considered threatened—Hipposideros marisae (IUCN status: Vulnerable) and Rhinolophus ziama (IUCN status: Endangered) (Monadjem et al. 2016). Including Myotis nimbaensis, 10 taxa—one sixth of the total bat species diversity of the Nimba Mountains—have yet to have their conservation status fully assessed by the IUCN. Although some of these lineages may be fairly widespread in West Africa, others will likely turn out to mirror the endemic Hipposideros and Rhinolophus of the area that are strictly endemic to a one or a few mountainous areas. Myotis nimbaensis, which may be limited by both underground roosting habits and affinity for habitat and vegetation combinations found only in the highlands, may well be one of these taxa. The striking color pattern of M. nimbaensis and its lack of black facial spots (spots highly characteristic of M. welwitschii, the only bat in West Africa that it superficially resembles) suggests that M. nimbaensis is unlikely to have been overlooked or misidentified in previous surveys in the region. It therefore seems very likely that it is an uncommon to rare endemic with a very small geographic range. In this context, we expect that M. nimbaensis will be classified as a critically endangered species when it is assessed by the IUCN based on having a known extent of occurrence less than 100 km2 (IUCN 2012).
ACKNOWLEDGMENTS
The authors would like to thank Patrick Jules Atagana and Marcel Leno for their assistance during the 2018 field expedition to Nimba. We would also like to thank Darrin Lunde (USNM) and Roberto Portela Miguez (BMNH) for access to specimens needed for this project, and for kindly hosting N.B.S. during visits to their respective institutions. We also thank Bruce Patterson and Terry Demos (FMNH) for prepublication access to their data, and Bruce Patterson and Gabor Csorba for thoughtful reviews of the manuscript. A.E.M. was supported by a Gerstner Postdoctoral Fellowship from the American Museum of Natural History.
REFERENCES
Appendices
APPENDIX 1
Specimens Examined
The following list includes all specimens used in the morphological components of this study with data on their respective localities. See Materials and Methods for institutional abbreviations.
Myotis morrisi (N = 1). ETHIOPIA— Blue Nile Gorge, Mouth of Didessa River, Forward Base Three (BMNH 70.488 [holotype]).
Myotis nimbaensis (N = 2). GUINEA—Nimba Mountains, Kaiser Adit 2, N07.66499, W008.37223 (AMNH 279589 [holotype], 279590 [paratype]).
Myotis tricolor (N = 28). ETHIOPIA— Amhara Region, Agew Awi Zone, Dangila (BMNH 37.2.24.13); KENYA—Rift Valley Province, 20 mi. SW of Kitale, R. Barberton Farm (USNM 351060, 351061); Rift Valley Province, Crater of Mt. Menengai, 7400 ft. (USNM 317927, 317928); Rift Valley Province, 3mi NW Nakuru, Menengai Crater, 6500 ft. (BMNH 75.2549—75.2554); Rift Valley Province, West Pokot County, Sigor, Wei-Wei River (BMNH 75.2555); LIBERIA—Grand Cape Mt., Bomi wood concession (AMNH 257053); MALAWI—Zomba District, Zomba (BMNH 87.1082); SOUTH AFRICA—Eastern Cape Province, King William's Town (AMNH 146789); KwaZulu-Natal, Estcourt, Will Brook (BMNH 14.5.4.2); KwaZulu-Natal, Otto's Bluff (AMNH 232029, 232030); KwaNatal, Pietermaritzburg, Town Bush (USNM 292066); Mpumalanga Provence, Barberton, Louws Creek [= Low's Creek] (USNM 238099); Transvaal, 32 mi W of Pretoria, Uitkomst Farm (USNM 342643—342645); Transvaal, Kruger National Park, ca. 4 km east of Skukuza (AMNH 257358); UGANDA—Sebai District, Near Sipi, Cave at Kyema (BMNH 64.172); Mt. Elgon, Kapretwa (BMNH 40.740, 40.741); ZAMBIA—Cave near Mujimbe Hill (AMNH 89776).
Myotis welwitschii (N = 3). MALAWI—Southern Region, Cholo [= Thyolo], Ruo River (BMNH 22.12.17.75); Southern Region, Near Chiromo (BMNH 22.12.17.76); SOUTH AFRICA—Transvaal, 50 miles ENE of Lydenburg (BMNH 0.11.6.1).
