Accepted Manuscript
Molecular phylogenetic reconstructions identify East Asia as the cradle for the
evolution of the cosmopolitan genus Myotis (Mammalia, Chiroptera)
Manuel Ruedi, Benoît Stadelmann, Yann Gager, Emmanuel J.P. Douzery,
Charles M. Francis, Liang-Kong Lin, Antonio Guillén-Servent, Alice Cibois
PII: S1055-7903(13)00333-3
DOI: http://dx.doi.org/10.1016/j.ympev.2013.08.011
Reference: YMPEV 4692
To appear in: Molecular Phylogenetics and Evolution
Please cite this article as: Ruedi, M., Stadelmann, B., Gager, Y., Douzery, E.J.P., Francis, C.M., Lin, L-K.,
Guillén-Servent, A., Cibois, A., Molecular phylogenetic reconstructions identify East Asia as the cradle for the
evolution of the cosmopolitan genus Myotis (Mammalia, Chiroptera), Molecular Phylogenetics and Evolution
(2013), doi: http://dx.doi.org/10.1016/j.ympev.2013.08.011
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Molecular phylogenetic reconstructions identify East Asia as the cradle for the evolution
of the cosmopolitan genus Myotis (Mammalia, Chiroptera)
Manuel Ruedia, Benoît Stadelmanna, b, Yann Gagerc, d, Emmanuel J. P. Douzery c, d, Charles
M. Francise, Liang-Kong Linf, Antonio Guillén-Serventg, and Alice Ciboisa
a Department of Mammalogy and Ornithology, Natural History Museum of Geneva, Route de
Malagnou 1, BP 6434, 1211 Geneva 6, Switzerland
b Department of Zoology and Animal Biology, University of Geneva, 30 quai Ernest-
Ansermet, 1211 Geneva 4, Switzerland
c Université Montpellier II, Place Eugène Bataillon, 34 095 Montpellier Cedex 5, France
d CNRS, Institut des Sciences de l’Evolution (UMR 5554), CC064- Place Eugène Bataillon,
34095 Montpellier Cedex 5, France
e Canadian Wildlife Service, Environment Canada, Ottawa, Ontario, K1A 0H3, Canada
f Laboratory of Wildlife Ecology, Department of Biology, Tunghai University, Taichung,
Taiwan 407, R.O.C.
g Instituto de Ecología A.C., Carretera Antigua a Coatepec 351, El Haya, Xalapa 91070,
Veracruz, México
* Corresponding author. Address: Department of Mammalogy and Ornithology, Natural
History Museum of Geneva, Route de Malagnou 1, BP 6434, 1211 Geneva (6), Switzerland.
Tel.: +41 22 418 6320; Fax: +41 22 418 6301.
E-mail address: [email protected] (Manuel Ruedi)
SUGGESTED RUNNING TITLE: Evolutionary origins of Myotis bats.
2
ABSTRACT
___________________________________________________________________________ 1
Sequences of the mitochondrial cytochrome b (1140 bp) and nuclear Rag 2 (1148 bp) genes 2
were used to assess the evolutionary history of the cosmopolitan bat genus Myotis, based on a 3
worldwide sampling of over 88 named species plus 7 species with uncertain nomenclature. 4
Phylogenetic reconstructions of this comprehensive taxon sampling show that most radiation 5
of species occurred independently within each biogeographic region. Our molecular study 6
supports an early divergence of species from the New World, where all Nearctic and 7
Neotropical species plus a lineage from the Palaearctic constitute a monophyletic clade, sister 8
to the remaining Old World taxa. The major Old World clade includes all remaining Eurasian 9
taxa, most Oriental species, one Oceanian, and all Ethiopian species. Another lineage, 10
including M. latirostris from Taiwan, appears at the base of these two major biogeographic 11
clades and, because it bears nyctalodont molars, could be considered as a distinct genus. 12
However, this molar configuration is also found in crown-group species, indicating that these 13
dental characters are variable in the genus Myotis and may confound interpretation of the 14
fossil record . Molecular datings suggest an origin of all recent Myotis in the early Miocene 15
(about 21 MYA with 95% highest posterior density interval 23-20 MYA). This period was 16
characterized by a global climatic cooling that reduced the availability of tropical habitats and 17
favoured the development of more temperate vegetation. This sharp climatic change might 18
have triggered the evolution of Myotis in the Northern continents, because Myotis ancestors 19
seem to have been well adapted and successful in such temperate habitats. Ancestral area 20
reconstructions based on the molecular phylogeny suggest that the eastern portion of the 21
Asian continent was an important centre of origin for the early diversification of all Myotis 22
lineages, and involved relatively few subsequent transcontinental range expansions. 23
24
Keywords: Myotis, worldwide radiation, Vespertilionidae, Lagrange, cytochrome b, Rag 2, 25
molecular dating, biogeography 26
27
28
1. Introduction 29
The evolutionary history of species is the outcome of a complex balance between 30
processes of extinction and speciation. At any time, some groups may thus be highly 31
diversified and successful while sister clades are depauperate (Purvis et al., 2000). The causes 32
that may promote or inhibit these processes in a particular clade and not in others are still 33
highly debated, but may include intrinsic key innovations; extrinsic processes linked to plate 34
3
tectonics, climatic fluctuations or vegetation changes; differential rates of gene evolution; and 35
niche competition. One difficulty in identifying the common processes underlying the 36
evolution of species diversification and natural history of species is the need for accurate 37
knowledge of the phylogenetic relationships of the group considered (Paradis, 1998; 38
Sanderson and Donoghue, 1996). In particular complete or nearly complete phylogenies 39
including all major branches (i.e. an exhaustive phylogenetic hypothesis for the evolution of a 40
whole group of extant species), allow for robust analysis of the factors that influenced the 41
rates of speciation and extinction (Purvis et al., 2000). Moreover, these complete phylogenies 42
can help to elaborate conservation plans aimed to preserve most of the current evolutionary 43
history of these groups (Bininda-Emonds et al., 2000; Sechrest et al., 2002). 44
Bats from the order Chiroptera are part of the Laurasiatheria clade of mammals, and 45
have diversified in all continents except Antarctica, with over 1100 currently recognized 46
species worldwide (Simmons, 2005), a number that continues to increase with new taxonomic 47
work. Within the superfamily Vespertilionoidea (Teeling et al., 2005), the genus Myotis 48
belongs to the family Vespertilionidae, and is classified in a distinct subfamily, the Myotinae 49
(Hoofer and Van den Bussche, 2003; Simmons, 2005). Because these bats have a rather 50
“generalist morphology” (Horácek et al., 2000) and are very speciose, their taxonomy is 51
difficult. In a phenetic study based on external and cranial morphology, Findley (1972) 52
estimated that the current fauna included about 60 species of Myotis. This number has 53
increased through morphological and genetic studies and recent revisions now exceed 100 54
species (Koopman, 1994; Simmons, 2005). Indeed, new species continue to be described 55
regularly, even in well-known areas such as Europe (Helversen et al., 2001; Ibáñez et al., 56
2006). Thus, Myotis is one of the most diverse genera of mammals, second only to Crocidura 57
shrews, and offers an exceptional substrate for studying speciation and diversification at a 58
worldwide scale. Myotis bats have colonised most terrestrial habitats, except polar regions, 59
and represents the only mammal genus that is naturally distributed on every continent except 60
Antarctica. However, the species diversity is unequal by region, with the maximum diversity 61
found in the northern continents (see inset of Fig. 1). 62
To date, few molecular studies have approached the evolution of the genus Myotis. It 63
has been demonstrated that the prevailing morphology-based subdivision of the genus Myotis 64
into four or more subgenera (Findley, 1972; Koopman, 1993; Tate, 1941) does not reflect 65
phylogenetic history (Hoofer and Van den Bussche, 2003; Kawai et al., 2003; Lack et al., 66
2010; Ruedi and Mayer, 2001). Instead, it appears that adaptive convergences have produced 67
similar ecomorphs independently through some deterministic processes (Fenton and 68
Bogdanowicz, 2002). By contrast, several independent biogeographic radiations emerged 69
4
from phylogenetic reconstructions, including one that unites all Ethiopian taxa (Stadelmann et 70
al., 2004b) and another that comprises all New World species (Lack et al., 2010; Stadelmann 71
et al., 2007). However, Myotis from the Palaearctic and the Oriental regions have been 72
underrepresented in previous studies, impeding phylogenetic inference for the whole group 73
(Kawai et al., 2003; Zhang et al., 2009). 74
By expanding the Old World sampling to include over 80% of all named Myotis 75
species (Simmons, 2005) as well as several unnamed taxa, we produced a comprehensive and 76
robust phylogeny of extant species covering all major branches. With this almost complete 77
picture of the Myotis radiation, we first investigate phylogenetic relationships within Old 78
World taxa. Second, we evaluate the timing and biogeographic evolution of Myotis worldwide 79
by using a Bayesian relaxed molecular clock approach, as well as likelihood reconstructions 80
of ancestral geographic distribution to identify the possible area of origin of this striking 81
evolutionary radiation. 82
83
2. Materials and Methods 84
2.1 Taxon and geographic sampling 85
According to the last available revision of the genus (Simmons, 2005), 103 species of 86
Myotis are recognized worldwide. Unless indicated hereafter, we follow Simmons‟ taxonomic 87
arrangement to refer to the nominal species analysed in this study (see Table S1). Additional 88
taxa recognized here include M. aurascens from Central and East Asia [distinct from M. 89
mystacinus (Tsytsulina et al., 2012)], M. cf. browni from the Philippines and M. latirostris 90
from Taiwan [both distinct from M. muricola (Lack et al., 2010; Stadelmann et al., 2007)], 91
M. escalerai from the Iberian peninsula [distinct from M. nattereri (Ibáñez et al., 2006)], M. 92
formosus flavus [specifically distinct from M. formosus (Jiang et al., 2010)], M. gracilis from 93
the Far East [distinct from M. brandtii though M. sibiricus may be a senior synonym and 94
hence the appropriate name for this species (Horácek et al., 2000; Kruskop et al., 2012)], M. 95
petax from the Eastern Palaearctic [distinct from M. daubentonii and a senior synonym of M. 96
abei (Kruskop et al., 2012; Matveev et al., 2005; Tsytsulina, 2004)], M. phanluongi from 97
Vietnam [distinct from M. siligorensis (Borisenko et al., 2008)], and M. taiwanensis [distinct 98
from M. adversus (Han et al., 2010)]. In turn the European taxon oxygnathus is maintained as 99
a subspecies of M. blythii (Evin et al., 2008), while the Japanese hosonoi, ozensis and 100
yesoensis are now all considered as synonyms of M. ikonnikovi (Abe et al., 2005; Kawai et al., 101
2003). Species for which no tissue or GenBank entries were available include M. adversus 102
(Oriental), M. aelleni (Neotropical), M. australis (Oceanian), M. bucharensis (Eastern 103
Palaearctic), M. cobanensis (Neotropical), M. csorbai (Oriental), M. findleyi (Neotropical), M. 104
5
fortidens (Neotropical), M. gomantongensis (Oriental), M. hajastanicus (Western Palaearctic), 105
M. hermani (Oriental), M. insularum (Oceanian), M. melanorhinus (Nearctic), M. 106
moluccarum (Oceanian), M. morrisi (Ethiopian), M. nipalensis (Eastern Palaearctic), M. 107
oreias (Oriental), M. peninsularis (Nearctic), M. planiceps (Neotropical) and M. stalkeri 108
(Oriental). We also included 7 additional species of Myotis that diverge genetically and 109
morphologically from other currently recognized taxa and await proper taxonomic description 110
(see e.g. Francis et al., 2010; Larsen et al., 2012b; Salicini et al., 2011). Although not 111
complete, we feel that this taxonomic coverage (Table S1) represents most, if not all, major 112
lineages in the Myotis radiation, as most of the missing taxa appear morphologically or 113
genetically to be part of clades that are represented in our sampling, a conclusion confirmed 114
for New World taxa by recently published molecular surveys (Larsen et al., 2012a; Larsen et 115
al., 2012b). 116
In total, 105 taxa representing 88 nominal species, 10 distinct subspecies and 7 117
additional unnamed/unidentified species were analysed (see Table S1). Because of potential 118
concerns about species assignments and nomenclature, especially for the Asian Myotis (Bates 119
et al., 1999), most sequences were derived from vouchered specimens that can be examined in 120
public institutions. In addition to the newly sequenced specimens, some taxa were represented 121
by published sequences (Bickham et al., 2004; Ibáñez et al., 2006; Jones et al., 2006; Kawai et 122
al., 2003; Lack et al., 2010; Larsen et al., 2012b; Ruedi and Mayer, 2001; Stadelmann et al., 123
2004a; Stadelmann et al., 2004b; Stadelmann et al., 2007; Tsytsulina et al., 2012; Zhang et al., 124
2009) available in GenBank (see Table S1). We refrained from considering several other 125
unpublished Myotis sequences available in public databases because they were of uncertain 126
taxonomic origin or had experimental problems. For instance, the sequence of “Myotis 127
hajastanicus” AY665138 was discarded because it proved to be a chimera between partial 128
cytochrome b sequences of M. gracilis and of Bos indicus. 129
The Kerivoulinae and Murininae are the closest relatives of Myotis (e. g. Hoofer and 130
Van den Bussche, 2003; Kawai et al., 2002). Therefore, six species representing the genera 131
Harpiocephalus, Kerivoula and Murina from the Oriental region (Hoofer et al., 2003; Lack et 132
al., 2010; Ruedi et al., 2012; Stadelmann et al., 2004b; Stadelmann et al., 2007) were used as 133
a composite outgroup (Table S1). 134
135
2.2 DNA amplification and sequencing 136
Tissue samples preserved in ethanol or DMSO were soaked for 1 hour in sterile water 137
before extraction. Total genomic DNA was then isolated following a salting out protocol 138
(Miller et al., 1988), as detailed in Castella et al. (2001) and rediluted into 200 µl of low TE 139
6
buffer. The complete mitochondrial cytochrome b gene (Cyt b; 1140 bp) and the partial 140
nuclear Recombination Activating Gene 2 (Rag 2; 1148 bp) were amplified, and sequenced in 141
both directions following the procedures described in previous papers (Stadelmann et al., 142
2004b; Stadelmann et al., 2007). Sequences were checked for possible stop codons or indels, 143
which would indicate the presence of pseudogenes. As none of the coding sequences 144
presented such anomalies, they were assumed to be the functional gene and alignment of 145
orthologs was thus straightforward. Heterozygous individuals at the nuclear Rag 2 gene were 146
identified by double peaks on both sequencing strands. These heterozygous positions were not 147
phased out, but replaced by „N‟. Fragments shorter than the target length were also completed 148
by „N‟s to replace missing data. 149
150
2.3 Phylogenetic analyses 151
Phylogenetic reconstructions were conducted on the Cyt b (141 sequences of 1140 bp; 152
601 variable positions in the ingroup) and Rag 2 datasets (122 sequences, 1148 bp, 220 153
variable positions) analysed separately and in combination (141 lineages of 2288 bp) with 154
strings of Ns replacing the missing sequences. These missing data represented 1.4% and 3.5% 155
of the character matrix for the Cyt b and Rag 2 datasets, respectively. Probabilistic methods 156
were used to reconstruct phylogenetic trees with the Maximum Likelihood (ML) approach 157
implemented in RAxML (Stamatakis, 2006) and Bayesian inferences (BA) in MrBayes v3.2.0 158
(Ronquist and Huelsenbeck, 2003). All analyses were done on a fully partitioned model, 159
where each gene and/or codon partition was allowed to have its own set of model parameters. 160
The most appropriate model of nucleotide substitution for each partition was evaluated using 161
MrModeltest 2.3 (Nylander, 2004) and the Akaike Information Criterion (AIC; data 162
summarized in Table S2). The General Time Reversible (GTR) model with rate variation 163
among sites (Γ) and a proportion of invariable sites (I) represented the best fitting model of 164
nucleotide substitution for the Cyt b and the combined Cyt b + Rag 2 dataset (Table S2). 165
Topological searches were initiated from random trees. 166
Bayesian analyses were performed with a Markov chain Monte-Carlo technique and 167
run for 10x106 generations, with a sampling every 1000 generations. The chains were checked 168
for convergence and appropriate effective sample size (ESS>200) with Tracer v.1.5 (Rambaut 169
and Drummond, 2009). Chains converged and the log-likelihood reached an asymptote after 170
about 1x106 (Cyt b or Rag 2) or 3.5 x106 (combined) generations. These initial trees were 171
discarded as burn-in. Posterior probabilities (PP) were subsequently computed from the 172
consensus of the remaining sampled trees. Reliability of nodes in the ML analyses was 173
assessed by 100 standard bootstraps (BP) with RAxML. 174
7
175
2.3 Molecular dating analyses 176
We estimated the mean divergence times between taxa by a relaxed uncorrelated 177
lognormal molecular clock as implemented in BEAST v1.7.4 (Drummond and Rambaut, 178
2007). Similarly to the MrBayes analyses, we used the combined dataset partitioned into 6 179
categories (3 codon positions each for Cyt b and Rag 2) with the most appropriate model of 180
nucleotide substitution for each codon partition determined with MrModeltest (Table S2). 181
Tree searches started from a random tree and assumed a constant lineage birth rate for each 182
branch (Yule tree prior; Drummond et al., 2006); all other parameters were kept at default 183
values. Chains were sampled every 1000 generations over 10x106 generations, with a burn-in 184
of 10%. 185
Fossil calibrations were used to place temporal constraints on two nodes, as minimum 186
and maximum soft bounds. The choice of these calibrations was difficult as extant Myotis 187
species show a combination of ancestral and derived character states and are difficult to 188
discriminate in the fossil record (Gunnell et al., 2012; Horácek, 2001). This is especially true 189
for the different dental remains that constitute the fossils of ancient Myotis-like taxa (Horácek, 190
2001). The first calibration constraint was the split of M. daubentonii and M. bechsteinii, 191
estimated to have diverged between 5 and 11.6 MYA (Topál, 1983). We used an exponential 192
prior distribution (offset 5.0, mean 2.5) to encapsulate this calibration in the 95% CI. The 193
second calibration involved the most recent common ancestor of Myotis estimated to have 194
diverged in the Late Oligocene - Early Miocene (Horácek, 2001; Gunnell et al., 2012), some 195
20-31 MYA. If Khonsunycteris is also part of the early Myotinae radiation, as hypothesized 196
by Gunnell et al. (2012), this lower boundary would be slightly older (34 MYA), but still 197
included in the lognormal prior distribution entered in our BEAST analysis (offset 20.0, S.D. 198
1.6). 199
200
2.4 Biogeographic reconstructions 201
Species distributions were obtained from the geographic information available in 202
Simmons (2005) and were coded in seven biogeographic regions (bioregions): Nearctic, 203
Neotropical, Western and Eastern Palaearctic, Oriental (= Indomalayan, Corbet and Hill, 204
1992), Ethiopian and Oceanian regions (see inset of Fig. 1 and Table S1 for details of 205
geographic assignations). The dispersal-extinction-cladogenesis (DEC) likelihood model 206
implemented in LAGRANGE (Ree and Smith, 2008) was used to investigate the 207
biogeographic evolution of the genus Myotis. Following a probabilistic approach, this 208
program traces from the tips to the root of a phylogenetic tree changes in distribution and 209
8
reconstructs the ancestral state by giving probabilities of each state at each node. We used the 210
ultrametric tree issued from the BEAST analysis as a template topology for all biogeographic 211
reconstructions. Although Myotis bats can be strong flyers, areas separated by large stretches 212
of open sea represent important, if not impassable barriers to dispersal (Castella et al., 2000; 213
García-Mudarra et al., 2009; Larsen et al., 2012b), while adjacent land masses can be reached 214
more easily. Hence, we performed biogeographic reconstructions assuming two dispersal 215
models, one without any constraint on dispersal across the planet (all regions can be reached 216
from any area; H0), and one where possible dispersals were restricted to adjacent land masses 217
(H1). In this constrained model, an adjacency matrix of connections between regions was 218
implemented as following: any dispersal was allowed within the Old World, or within the 219
New World, but between those continents, only dispersals from/to the Eastern Palaearctic and 220
the Nearctic were allowed (i.e. across the Beringian Strait). Similarly, connections with the 221
Oceanian region was only allowed to/from the Oriental region (i.e. across Wallacea), as 222
suggested by previous genetic studies of Myotis species living in this region (Kitchener et al., 223
1995). Because of the global scale of these analyses, we assumed that ancestral areas could 224
not cover more than two bioregions, and set the analyses accordingly. The best fitting results 225
were chosen for interpretation by comparing their log-likelihoods (see Ree and Smith, 2008). 226
The influence of the biogeographic origin of the outgroup species (here they were all sampled 227
within the Oriental region, see Table S1) may introduce a bias in the reconstruction of the 228
deepest nodes of the ingroup (personal observation). To avoid this potential problem, we 229
artificially assigned both the Ethiopian and Oriental regions to the geographic range of 230
Kerivoula, and the Oriental and Eastern Palearctic regions to the distribution of Murina, in 231
order to reflect the global distribution of these outgroup genera (Simmons, 2005). 232
233
3. Results 234
3.1 mtDNA and nuclear diversity 235
Thirty-nine sequences of the Cyt b gene were newly obtained, which represent 32 236
Oriental, 6 Palaearctic and one Ethiopian lineages of Myotis (Table S1). These sequences 237
were deposited in GenBank under accession numbers KF312497-KF312535. The final dataset 238
for the Cyt b gene thus consisted of 141 sequences of 1140 bp (135 ingroup and 6 outgroup 239
taxa). Whenever possible, we ensured that the same individual was sequenced for the Cyt b 240
and Rag 2 genes, but this was not possible for all species. Furthermore, because tissue sample 241
was not available for all species used in the Cyt b dataset, the taxon sampling of the Rag 2 242
dataset was reduced to 122 sequences (including 6 outgroups). The 48 new partial Rag 2 243
sequences obtained in this study were deposited in GenBank under accession numbers 244
9
KF312536-KF312583. A complete list of all specimens and sequences is provided in Table 245
S1. 246
247
3.2 Phylogenetic reconstructions 248
Phylogenetic trees based on ML or BA reconstructions were very similar, although 249
nodes tended to be less well supported by bootstraps (BP) when compared to posterior 250
probabilities (PP, Table 1), a well-known bias inherent to the different methods used to 251
estimate nodal support (Alfaro et al., 2003; Douady et al., 2003). The overall levels of support 252
were also higher towards the terminal clades than for deeper nodes, indicating that additional 253
mitochondrial and nuclear markers are required to resolve the entire Myotis radiation with 254
high statistic support. Deep nodes that were well supported in most analyses include the 255
monophyly of all current Myotis species and of the New World clade (Table 1). As expected 256
at low taxonomic levels, trees reconstructed with the more variable Cyt b gene (Fig. S1) were 257
much more resolved when compared with those generated with the Rag 2 gene alone (Fig. 258
S2). Phylogenetic incongruences between gene partitions include the position of latirostris or 259
the clade mystacinus-ikonnikovi, both of which appear within the New World clade (albeit 260
with low bootstrap) in the nuclear tree (Fig. S2), whereas it is basal in reconstructions 261
including the Cyt b data set (Figs. 1 and S1). Another study including many more nuclear 262
characters (2352 variable positions; Lack et al., 2010) also places with strong support 263
latirostris as the sister-group relative to the remaining Myotis, not within the New World 264
clade, suggesting that inference from the nuclear gene tree of Fig. S2 might require many 265
more characters or a larger sampling of nuclear genes to be more reliable. The combination of 266
both gene datasets resulted in significant improvement of most nodal supports (Table 1). We 267
therefore report only phylogenetic reconstructions resulting from this combined Cyt b + Rag 2 268
dataset and representing 141 species or lineages (Fig. 1). Phylogenetic tree can be accessed in 269
TreeBase (http://purl.org/phylo/treebase/phylows/study/TB2:S14471). 270
Three major monophyletic groupings are statistically supported (PP and BP 72-100 %) 271
in the early radiation of Myotis: the latirostris lineage, the New World clade, and the Old 272
World clade (Fig. 1). The latirostris lineage comprises M. latirostris from Taiwan and an 273
unidentified taxon (M. sp 4) from South China. This lineage is clearly basal to the Myotis 274
radiation. The monophyletic and strongly supported New World clade includes all sequenced 275
New World species of Myotis plus the Palaearctic brandtii lineage. The latter is represented 276
by the Western Palaearctic M. brandtii and the Eastern Palaearctic M. gracilis. The New 277
World clade is comprised of two further, well-supported subclades, one containing 278
10
exclusively Nearctic species, while the other is composed of both Nearctic and Neotropical 279
species, as detailed in Stadelmann et al. (2007). 280
The Old World clade encompasses the greatest species diversity with over 60 distinct 281
taxa, and is structured in two major groups: the Ethiopian clade, and the Eurasian assemblage, 282
the latter with low nodal support (Fig. 1, Table 1). Such low support can be due to a 283
succession of short internodes or to the uncertain position of two species, M. alcathoe and M. 284
dasycneme (called here “floating” species) that branch close to the base of this Eurasian 285
radiation, but with little statistical support. The Ethiopian clade is highly supported with all 286
methods of reconstruction, and has been described in Stadelmann et al. (2004b). It unites all 287
species from sub-Saharan Africa and islands of the western Indian Ocean (Weyeneth et al., 288
2011), plus the circum-Mediterranean M. emarginatus, and two Asian species, M. cf. 289
formosus and M. formosus flavus. Finally, the Eurasian assemblage includes all remaining 290
species from the Palaearctic, Oriental and Oceanian regions (Fig. 1). This assemblage can be 291
divided in several clades already described in previous studies (Clade II, III and IV, Ruedi and 292
Mayer 2001) and in five new ones, as indicated by continued Roman numerals (Fig. 1 and 293
Table 1). Some of these clades are composed exclusively of lineages from the Oriental region 294
(e.g. Clade VI or IX), while others, such as Clade II, are geographically more heterogeneous. 295
The European M. capaccinii and the South-East Asian M. annectans are monotypic lineages 296
with phylogenetic positions unsettled within this Eurasian assemblage. 297
Interestingly, species endemic to the Western Palaearctic region, such as M. 298
bechsteinii, M. escalerai, M. capaccinii, M. alcathoe or those from the Far East (e.g. M. 299
pruinosus, M. macrodactylus, M. yanbarensis, M. ikonnikovi) are largely unrelated and 300
dispersed across the entire phylogenetic tree (Fig. 1), while those from Africa or from the 301
Americas are much more closely related to their respective geographic congeners. Species 302
from the Oriental region are either closely related within an endemic clade (e.g. Clade VII or 303
IX) or part of geographically more widespread assemblages (e.g. Clade II or V). The most 304
dramatic endemic radiation involves species from the Nearctic and Neotropical subclades 305
(Stadelmann et al., 2007) that largely speciated within their biogeographic regions. 306
307
3.3 Divergence times 308
Divergence time estimates obtained with a relaxed molecular clock on the combined 309
dataset are reported on Fig. 2 and suggest that all Myotis diverged from their closest sister 310
group (represented here by several Murininae and Kerivoulinae species) about 26 MYA (HPD 311
31-21 MYA). The latirostris lineage diverged about 21 MYA from the lineage leading to all 312
other modern Myotis. The early divergence between the New World clade and the Old World 313
11
clade represents the second major split within the Myotis radiation and occurred about 19 314
MYA. Then, the Nearctic subclade started diverging about 12 MYA, while the brandtii 315
lineage split from the Neotropical subclade 11 MYA. The most recent common ancestor 316
shared by the Ethiopian clade and the Eurasian assemblage is about 17 MYA. The lineages 317
that have further diversified in the seven corresponding clades of the Eurasian assemblage 318
originated within a narrow time frame in the middle Miocene, between 16 and 11 MYA (Fig. 319
2). 320
Compared to our previous study based on the same markers but on a reduced taxon 321
sampling of Myotis species and on a different method to calibrate nodes (see Stadelmann et 322
al., 2007), the dates inferred here are largely concordant for the recent nodes (< 8 MYA), but 323
for the deeper ones, the dates inferred here are much older. For instance, the basal node for all 324
modern Myotis is estimated here at about 20.9 MYA (with 95% highest posterior density 325
interval, HPD = 23-20 MYA; Table 1), while it was at 13.0 ± 2.2 MYA in Stadelmann et al. 326
(2007). Another example is the origin of the New World clade, which was estimated at 9.9 ± 327
1.7 MYA by Stadelmann et al. (2007), while it appears about two million years older in the 328
present estimates (12.3 MYA, HPD 15-10 MYA). Other dated molecular trees aimed to 329
resolve the Vespertilionidae radiation, based on different genes, fossil calibration and taxon 330
sampling (Lack et al., 2010), are also more concordant with the present estimates. We 331
therefore feel that divergence times presented here are consistent with the best available 332
information, even if largely imprecise due to the usual uncertainty associated with molecular 333
tree calibrations. This calibrated phylogeny served as the basis for the following Lagrange 334
biogeographic reconstructions. 335
336
3.4 Biogeographic inference 337
The likelihood value of biogeographic reconstructions of the unconstrained-dispersal 338
model (H0 -lnL = 195.3) was significantly worse than the constrained one (H1 -lnL = 180.3; 339
delta lnL = 15; p < 0.001), and hence only this second model with dispersals limited to 340
adjacent bioregions will be discussed further. Overall, 111 out of the 134 ingroup nodes had 341
robust ancestral area reconstructions (AAR) with over 70% relative likelihood. For the 342
remaining nodes some uncertainties existed due to alternative, albeit less likely, AAR. For 343
instance, the AAR for the deepest node of all modern Myotis was not clearly identified, with 344
the co-occurrence of the Oriental and Eastern Palaearctic regions being more likely (35% 345
relative probability, as shown in Fig. 3), but with either of those two bioregions alone only 346
slightly less likely (23 and 21 %, respectively). Likewise, there is considerable uncertainty 347
about the most likely scenario for the area occupied by the ancestor leading to the Ethiopian 348
12
clade. This AAR was either identified as being Oriental (46% relative probability) or Western 349
Palaearctic (45%). To ease interpretation, we recorded only the ancestral area with the highest 350
relative likelihood on Fig. 3, but acknowledge that some of those AAR are uncertain. 351
Globally, the biogeographic scenario for the Myotis evolution inferred under the DEC 352
model (Fig. 3) suggests relatively few intercontinental colonisation events (noted as range 353
expansions on the tree). These are 21 in total, 13 of which stem from the Oriental, four from 354
the Western Palaearctic, two from the Eastern Palearctic, and one each from the Ethiopian and 355
Nearctic regions (see inset of Fig. 3). With only eight occurrences recorded on the tree, 356
vicariant events leading to daughter lineages were much rarer than range expansions (Fig. 3). 357
One such event occurred in the early ancestor of most modern Myotis that occupied the 358
Eastern Palaearctic plus Nearctic regions, and which subsequently split into an American 359
branch (leading eventually to all extant New World species (Clade I) and an Oriental branch 360
(leading to the Eurasian assemblage, clades II to X). Another major vicariant event was 361
identified in the Neotropical subclade (Clade I, Fig. 3) where the common ancestor of the 362
clade comprising M. grisescens, M. austroriparius, M. velifer and M. yumanensis (all are 363
Nearctic species) was split by vicariance from the remaining Neotropical taxa. According to 364
the dated phylogeny of Fig. 3, these inferred colonisation and vicariance events apparently 365
occurred sporadically during the entire evolution of the Myotis, with no temporal aggregation. 366
367
4. Discussion 368
369
4.1 Morphological convergences and taxonomy 370
371
The first major dichotomy in molecular reconstructions is the basal split of the 372
latirostris lineage that represent the sister group of all other modern Myotis species (Fig. 1). 373
This phylogenetic position of the latirostris lineage also appeared in analyses involving a 374
much larger character sampling (over 6000 aligned nucleotide positions) and wider taxonomic 375
coverage (Lack et al., 2010), confirming that this outlier position is not an artefact of our 376
sampling. Species representing this lineage (M. latirostris from Taiwan and an unnamed 377
species from China, M. sp4; table S1) have a lower molar configuration (nyctalodoncy, Fig. 378
S3) that would exclude them from members of the Myotinae, as this subfamily is supposed to 379
contain species with exclusively myotodont molars (Gunnell et al., 2012; Menu and Sigé, 380
1971). Based on those genetic and morphologic lines of evidences, the latirostris lineage 381
could be treated as a distinct genus (Lack et al., 2010; Stadelmann et al., 2007). However the 382
presence of other fully nyctalodont species among crown-group Myotis (e.g. M. siligorensis 383
13
alticraniatus; Fig. S3) suggest that this character state was acquired several times within the 384
group and that other truly synapomorphic characters are needed to diagnose the unique 385
latirostris lineage. Remarkably, the overall conserved morphology and very close 386
resemblance to other small-sized, Asian Myotis led most authors to include latirostris in the 387
synonymy of M. muricola (Corbet and Hill, 1992; Simmons, 2005), which is clearly 388
inappropriate owing to their paraphyletic position (Fig. 1). 389
The second dichotomy within the Myotis radiation separates all New World (Clade I) 390
from most Old World species (clades II to X), the only exceptions being two Palaearctic taxa 391
embedded within Nearctic species (Fig. 1). These two boreal species, M. brandtii and M. 392
gracilis [for which M. sibiricus may be the appropriate name according to Kruskop et al. 393
(2012)] are morphologically almost indistinguishable from each other and were long 394
considered as synonyms of M. mystacinus. They differ from mystacinus by subtle penial and 395
dental characteristics, yet appear as distantly related on the Myotis tree (Fig. 1). This further 396
illustrates that close morphological resemblance in this difficult genus does not necessarily 397
imply close phylogenetic relationships (Ruedi and Mayer, 2001). This observation also holds 398
for other chiropteran taxa, for example the Phyllostomidae genera (Davalos et al., 2012). 399
The evolution of the New World clade was already described in a previous paper 400
(Stadelmann et al., 2007). Since the current sampling only adds a few more species (M. 401
nesopolus and M. nyctor) which do not alter the main phylogenetic relationships within this 402
clade, we will not discuss it in more detail here. We do note that the multiple new lineages or 403
cryptic species recently evidenced in South America (Larsen et al., 2012) are all nested within 404
the Nearctic or the Neotropical subclades, as predicted by their geographic origins. This 405
supports our supposition that our sampling of Myotis species, although not complete, is 406
representative of the major lineages existing in this continent. 407
The Old World clade is fairly well supported in our reconstructions (Table 1) and 408
includes all species from the remaining continents in the Western Hemisphere (clades II to X, 409
Fig. 1). This large assemblage unites morphologically disparate species, including the 410
smallest (M. siligorensis, less than 4 grams) and the largest species (M. chinensis, over 30 411
grams), orange (M. anjouanensis) and solid black species (M. macropus), short-eared (M. 412
daubentonii) and long-eared species (M. bechsteinii) or short-footed (M. frater) and long-413
footed species (M. pilosus). Conversely, examples of morphologically very similar species 414
that appear distantly related within this large clade are also numerous. For instance the 415
Eastern Palaearctic M. petax was long considered a single species with the Western 416
Palaearctic M. daubentonii, but multiple lines of evidence (Datzmann et al., 2012; Kruskop et 417
al., 2012; Matveev et al., 2005) clearly show that they are not even sister-species (they appear 418
14
in clades X and III, respectively, on Fig. 1). Even species sharing a remarkable black-and-419
orange pigmentation (e.g. M. welwitschii, M. cf. formosus and M. formosus flavus) are not 420
each other‟s closest relatives, although they do all belong to the Ethiopian clade (Clade V, 421
Fig. 1). Phylogenetic relationships rather indicate that the striking black-and-orange 422
patterning of these Myotis is either a plesiomorphic feature among members of the Ethiopian 423
clade, or is a convergent character. These examples confirm that convergence in several 424
morphological characters can mislead the systematics of Myotis bats. 425
Strong, misleading resemblances of unrelated taxa may also explain the difficulties in 426
classifying properly many species. In some cases, species that clearly fit within the phylogeny 427
of Myotis were originally classified in different genera [e.g. M. annectans and M. ridleyi were 428
originally classified as Pipistrellus and M. rosseti as Glischropus because they often lack the 429
middle upper and lower premolars, although other morphological and genetic characters 430
clearly indicate they belong in Myotis (Hill and Topál, 1973; Topál, 1970)]. In other cases, 431
multiple unrelated lineages with similar morphology have been classified as the same species 432
(e.g. M. nigricans Larsen et al., 2012a). Hence several of the specimens included in our 433
molecular sampling, although keying out to known species, bear “cf.” in their name to warn 434
that they are not assigned confidently to a particular taxonomic name. This is particularly 435
evident in the M. horsfieldii, M. montivagus and M. muricola species groups that are certainly 436
composed of several cryptic taxa (Francis et al., 2010; Görföl et al., 2013; Volleth and Heller, 437
2012). Other unnamed taxa are more divergent, both morphologically and genetically, and 438
await formal description (e.g. unknown species 1 to 7). Such descriptions will require 439
integrated approaches combining molecular and morphological characters to ensure that they 440
are accurately described in relation to currently recognized taxa. 441
The next major phylogenetic subdivision of the Myotis tree separates the Ethiopian 442
Clade V from the Eurasian clade (clades II to IV and VI to X, Fig. 1). The former is well 443
supported in all analyses and was described by Stadelmann et al. (2004b). We add to this 444
clade M. anjouanensis from the Comoros and both Oriental species of the formosus species 445
complex (Jiang et al., 2010). Interestingly, the latter are not each other closest relatives as M. 446
formosus flavus is sister to the circum-Mediterranean M. emarginatus, while M. cf. formosus 447
is more closely related to the African M. welwitschii. The Ethiopian clade is thus composed of 448
all sampled sub-Saharan species (M. morrisi remains unsampled) plus at least three Eurasian 449
and Oriental species. 450
The Eurasian assemblage is the most speciose and includes at least 57 recognized 451
species and a number of undescribed taxa (Fig. 1). The monophyly of this Eurasian 452
assemblage is not supported statistically (Table 1) probably due to the four “floating” species 453
15
that are variably placed near the base or in a more internal position within this assemblage. 454
These “floating” species include M. alcathoe, M. dasycneme, and M. capaccinii, all 455
distributed mainly in Europe, and the Asian species M. annectans (Fig. 1). It is unclear if the 456
phylogenetic uncertainty reflects hard polytomies (i.e. episodes of rapid radiation), or merely 457
soft polytomies that could be better resolved with increased character sampling (Lack and 458
Van Den Bussche, 2010). Published DNA barcode sequences for M. gomantongensis suggest 459
that it may be allied to M. annectans (Francis et al., 2010) and hence to this Eurasian 460
assemblage, but we were not able to include it in our analysis as we lacked the matching 461
genes. 462
Our molecular results further demonstrate the existence of at least eight strongly 463
supported clades within this Eurasian assemblage. Clade IX is composed exclusively of 464
species endemic to the Oriental region. Clade II is geographically more heterogeneous as it 465
includes species endemic to Europe, to Central or to Eastern Asia, all within Eurasia. Clade 466
III is a composite of geographically and ecomorphologically very distinct Myotis. It includes 467
the surface gleaner M. bechsteinii, the aerial feeder M. frater, the trawling bat M. daubentonii, 468
as well as species living in mountain forests of the Himalayas (M. sicarius) or found in 469
lowland bamboo forests in Taiwan (M. sp3). Similarly, the Oriental Clade IV includes several 470
trawling species with large feet (e.g., macrotarsus, horsfieldii, and hasseltii) as well as some 471
unusual small species with small feet and only two premolars (M. ridleyi and M. rossetti). 472
This diversity illustrates the rapid morpho-anatomical and/or ecological changes that can 473
occur within these clades. 474
475
4.2 Biogeographic evolution of the genus Myotis 476
477
Among placental mammals, molecular characters have shown that there is a 478
widespread pattern of phylogenetic relationships matching with the geographical origin of the 479
members of the corresponding major clades (Murphy et al., 2001). At lower taxonomic scales 480
within laurasiatherians, we here recover a similar pattern for Myotis bats as continental 481
assemblages of species, e.g. those found in Africa, in North and South America, or in Eurasia 482
tend to appear in monophyletic groups (Fig. 1). This suggests that most speciation events and 483
diversification occurred within each continent independently. The strong influence of the 484
biogeographic origin of the Myotis species in shaping global phylogenetic relationships 485
corroborate one earlier study based on less complete taxonomic sampling (Stadelmann et al., 486
2007). Indeed, the vast majority of species segregate in clades with a strong geographic 487
component, especially if we consider that the Eurasian continent is a huge landmass dissected 488
16
by rather loose biogeographic boundaries. For instance, the northeastern portion of the 489
Oriental region has no major topological discontinuity that would separate its faunal elements 490
from those in the Eastern Palaearctic region. This is reflected by a broad transition zone 491
comprising faunal elements typical of both regions (Corbet and Hill, 1992), and in the 492
phylogenetic relationships of Myotis by a mixture of endemics from both regions in the same 493
clades (Fig. 1). The same is true between the Western and the Eastern Palaearctic regions 494
where the limits are fuzzy and not bounded to major geographic discontinuities. The evolution 495
of ancestral areas reconstructed onto the phylogenetic relationships of Myotis (Fig. 3) reflect 496
these fuzzy biogeographic boundaries within the Eurasian continent, as the vast majority (16 497
out of 21) of ancestral range expansions recorded worldwide occurred within this huge land 498
mass (see inset of Fig. 3). 499
Barriers between the other biogeographic regions, such as mountain ranges, large 500
deserts or sea channels, appear to be much more effective at reducing dispersal of Myotis 501
species (Castella et al., 2000; García-Mudarra et al., 2009). For instance, none of the species 502
found on either side of the Sahara seem to have a common ancestor that lived across this 503
desert. Likewise no current Myotis species have a Holarctic distribution, and AAR suggest 504
that a single expansion across the Beringian Strait in the Middle Miocene could explain the 505
occurrence of Myotis species in the New World (Fig. 3). This scenario is strongly supported 506
by the Lagrange analysis (relative probability >98%) and contradicts the alternative scenario 507
of a back colonization of a putative common ancestor of brandtii/gracilis lineage coming 508
from the Nearctic region. 509
If we assume that all current species living on the Australian continent form a 510
monophyletic assemblage (Cooper et al., 2001), including M. macropus represented here in 511
Clade IV, the colonisation of this southern landmass can also be traced back to a single event 512
that occurred at the end of the Miocene epoch (Fig. 3). Interestingly, Myotis species from the 513
Japanese Archipelago (Kawai et al., 2003) appear in various independent clades (e.g. M. 514
macrodactylus, M. yanbarensis or M. ikonnikovi), suggesting that these islands were 515
colonized independently by several lineages before evolving eventually into endemic taxa. 516
The high proportion of expansion/colonisation events originating from the Oriental 517
region clearly highlights the pivotal importance of this region in the diversification of the 518
Myotis species worldwide. AAR (Fig. 3) also suggests that only eight ancestral ranges were 519
split by vicariance into descendent lineages, whereas most other diversification events took 520
place within the bioregions. However, as the geographic scale of each bioregion is very large 521
(i.e. continent-level) and probably many more vicariant events may have occurred within each 522
of these large and heterogeneous regions, our data are not sufficient to infer more precisely 523
17
which mode of speciation (sympatric versus allopatric) prevailed during the evolution of 524
Myotis. 525
The global number of species contained in each biogeographic region is very unequal, 526
with the northern continents much more diversified than the southern ones. For instance, the 527
number of species currently known to inhabit the Ethiopian region (up to 6 species, excluding 528
Madagascar and the Comoros; Happold and Happold, 2013) is in sharp contrast with the 529
species diversity found on other continents (except Australia; see inset of Fig. 1). This 530
suggests that Myotis species from the African continent did not radiate into many endemic 531
species despite the fact that genuine Myotis apparently entered into this region relatively early 532
(Middle Miocene; Fig. 3) during the evolution of modern species. Given the poor record of 533
Tertiary fossils of vespertilionids in Africa (Gunnell et al., 2012), it is not possible to know if 534
this paucity of species is due to an increased rate of extinction or a decreased rate of 535
diversification. By contrast, the Oceanian region was likely the last part of the world that was 536
reached by Myotis species, about 5.0 MYA (Fig. 3), suggesting that a late arrival might have 537
contributed to low current diversity of Myotis species found in Australia. 538
539
4.3 Fossils and molecules in the Myotinae 540
541
The Myotis species are rather abundant and already diverse in the European fossil 542
record of the Miocene through the Pleistocene, and recent studies (Gunnell et al., 2012; 543
Horácek, 2001) suggest that their ancestry might date back to the Late Oligocene. The 544
difficulty in assigning the most ancient fossils, often represented by isolated tooth (Menu et 545
al., 2002) or dentary fragments, is that the genus Myotis itself is defined by a combination of 546
plesiomorphic characters (e.g. three upper and lower premolars), as well as more derived ones 547
shared with other genera (e.g. single-rooted third premolars). The most prominent feature 548
believed to be characteristic of the genus Myotis is the myotodont configuration of lower 549
molars (Menu and Sigé, 1971). According to these dental characters, several of the more 550
ancient fossils initially attributed to Myotis (e.g. the Early Oligocene “Myotis” misonnei) are 551
now assigned to other genera (e.g. Quietia or Stehlinia) (Horácek, 2001) that are not related to 552
the Myotinae. By relying mainly on these dental characters, Gunnel et al. (2012) recently 553
revised the fossil origins of the Myotinae and stressed that there is a large gap between the 554
divergence times estimated from the molecules and the first appearance of fossil Myotis on 555
the different continents. The dated molecular phylogeny of Fig. 2 and divergence dates 556
independently obtained by Lack et al. (2010) now suggest that the most recent common 557
ancestor of all modern Myotis (including latirostris) is older than previously estimated, with a 558
18
divergence of 21 MYA (95% HPD 23-20 MYA, Table 1) or 18 MYA (HPD 23-13 MYA, 559
Lack et al., 2010). As a lower bound these molecular reconstructions also suggest that the 560
closer outgroups to the Myotinae (the Kerivoulinae-Murininae subfamilies) diverged in the 561
Late Oligocene (about 26 MYA, HPD 31-21 MYA; Fig. 3 and Lack et al., 2010). These 562
molecular estimates are still over 10 million years younger than the presumed oldest Myotinae 563
fossil (Khonsunycteris, 34 MYA, Gunnel et al., 2012). A plausible explanation for these large 564
temporal discrepancies between molecular and fossil interpretation is that the dental 565
characteristics used to assign fossil fragments to the Myotinae are convergent characters that 566
do not identify unambiguously members of this subfamily. For instance, it is now clear that 567
species in the siligorensis group are nyctalodont or semi-nyctalodont (Borisenko et al., 2008; 568
Tiunov et al., 2011), including all specimens of M. s. alticraniatus analyzed here (Fig. S3 and 569
Table S1). These specimens are nested within the Asian clade IX (Fig. 1), but on a 570
paleontological point of view, their cranial characteristics would have excluded them from the 571
Myotinae. Myotis latirostris and the unknown species 4 both have entirely nyctalodont molars 572
(Fig. S3) and are basal to, but clearly part of the early Myotinae radiation (Fig. 1), yet have 573
diverged only about 21 MYA from other recent species of Myotis (Table 1). These modern 574
taxa share striking dental and cranial similarities with a fossil genus, Submyotodon that was 575
described recently from the Upper part of the Middle Miocene deposits in Europe, but 576
classified away from Myotis lineages owing to its dental features (Ziegler, 2003). Finally, 577
members of the genus Cistugo have all the dental characteristics of Myotinae, including 578
myotodont molars (Fig. S3), yet they differ by the presence of unique wing glands and a 579
distinctive karyotype (Rautenbach et al., 1993), and are also genetically so divergent from any 580
other Vespertinionidae that they are now classified in their own family (Lack et al., 2010). 581
Again, paleontologically, Cistugo species would have been classified within the genus Myotis 582
as all those distinguishing features would have been invisible on fossils. Given these 583
morphological ambiguities, we therefore postulate that the most ancient, Oligocene fossils of 584
Myotis-like bats (including Khonsunycteris or Leuconoe = Myotis) are stem taxa not directly 585
related to extant Myotis, nor even part of the subfamily Myotinae. Their exact phylogenetic 586
affinities should therefore be reevaluated. This problematic paleontological assignation of 587
dental remains might also explain the great apparent gap between the Early Miocene (about 588
20 MYA) presence of Myotis–like bats in Australia, as suggested by Menu et al. (2002), and 589
our molecular reconstructions that date their earliest arrival on this continent in the Pliocene 590
(Fig. 3), about 5 MYA. Such a late, Pliocene, arrival of Myotinae in Australia is also 591
compatible with a scenario based on chromosomal evolution as shown for Australian 592
Vespertilionini (Volleth and Tidemann, 1991). 593
19
Molecular divergence times of younger nodes are more concordant with 594
paleontological data. Indeed, fossil deposits show a burst of diversification during the Middle 595
Miocene, in particular in Europe (Ziegler, 2003), which corroborates the emergence of most 596
basal lineages within the Eurasian assemblage (Fig. 2). 597
598
4.4 Paleoenvironmental changes and Myotis evolution 599
600
The Middle Miocene is characterized by a major climatic transition where global 601
cooling provoked a drop in sea level, and the development of more extensive temperate 602
habitats. These major environmental changes correspond to faunal turnovers (Janis, 1993) that 603
coincide with the basal radiation of crown-group Myotis, including the divergence of the Old 604
World versus New World clades (Fig. 2) that presumably occurred after a range expansion of 605
an ancestral Eastern Palaearctic taxon into North America , some 12.3 MYA (HPD 15-10 606
MYA, Table 1). 607
The northward drift of the African-Arabian plate resulted about 18-19 MYA in the 608
gradual closure of the Tethyan Seaway which had previously separated Africa from Eurasia. 609
Some mammals experienced faunal interchanges at that early time, but Myotis (and other 610
groups such as horses) apparently dispersed much later between these two continents, as 611
indicated by the colonization date of ancestors of the Ethiopian clade (12.3 MYA, HPD 15-10 612
MYA, Table 1, which corresponds to the Astracian age). It seems that besides the creation of 613
the land bridge, a change in habitat, probably driven by climatic factors, was necessary to the 614
establishment of dispersal routes suitable for Myotis. Our biogeographic reconstructions 615
suggest that the Ethiopian clade is most probably linked to an Oriental stock (Fig. 3). This 616
supposed Oriental origin together with the fact that the Sahara and its adjacent regions have 617
gradually shifted from a tropical to an arid environment since the Neogene warmth climax, 618
suggest that Myotis took advantage of these paleoenvironmental changes to reach Africa from 619
Asia (Fig. 3). The Asian M. cf. formosus represents the sister lineage of the African M. 620
welwitschii and colonized secondarily the Oriental region out of Africa (Fig. 3). Likewise the 621
biogeographic interpretation of the AAR of Fig. 3 suggests that M. emarginatus, now 622
confined to the Mediterranean region, descended from widespread ancestors that ranged 623
across Asia and Africa, then became isolated in the Oriental region, and finally reached 624
Europe through a range expansion (Fig. 3). 625
The Late Miocene epoch coincides with further cooling of the climate and 626
aridification, resulting in a continuing reduction of both tropical forests, and expansion of 627
more open and temperate habitats. This might have triggered the diversification of the Old 628
20
World Myotis lineages, especially the species-rich assemblage of Eurasian taxa. Indeed, 629
temperate forests, whether present in high latitude regions or in mountain ranges of 630
intertropical regions, characterize the habitats occupied today by most species of Myotis (La 631
Val, 1973). 632
633
4.5 Geographic origins of the Myotinae 634
635
The fossil record of Tertiary vespertilionids in tropical Asia (Mein and Ginsburg, 636
1997) or Africa (Butler, 1984) is poorly documented compared to that found in Europe (e.g. 637
Horácek, 2001; Ziegler, 2003) or North America (Gunnell and Simmons, 2005). Hence any 638
firm biogeographic inference based solely on the relative abundance or on the first occurrence 639
of Myotis fossils in sediments is difficult. The presumed African origin for the subfamily 640
Myotinae proposed by Gunnell and colleagues (Gunnell et al., 2012) strongly relies on the 641
dentary fragments of Khonsunycteris (a Late Eocene fossil found in Egypt) being part of that 642
subfamily. We however showed that this interpretation is questionable, as taxa unrelated to 643
the Myotinae, such as the endemic African Cistugo, can share all dental characteristics 644
believed to be unique to the genus Myotis. Contrary to the current interpretation of the fossil 645
record, our molecular phylogeny (Fig. 1) and likelihood biogeographic reconstructions (Fig. 646
3) identify eastern Asia as the cradle of Myotis evolution. There is some uncertainty whether 647
the Oriental region alone or in combination with the Eastern Palaearctic (as shown on Fig. 3) 648
is the origin of early Myotinae, but alternative bioregions are clearly less likely. In addition, 649
13 of the 21 intercontinental range expansions originated from the Oriental region (inset of 650
Fig. 3) suggesting that this area is a center of origin for the radiation of Myotis species. The 651
deepest lineage identified in molecular reconstructions, the latirostris lineage, that is sister to 652
all remaining modern Myotis (Fig. 1 and Lack et al., 2010), is also endemic to this region, 653
further supporting an ancient, East Asian origin for this radiation. Finally, the Myotis species 654
diversity in the Oriental region (inset of Fig. 1) is already the highest in the world based on 655
currently recognized taxa, and is likely strongly underestimated (Francis et al., 2010), again 656
pointing to this area as a cradle of evolution for the genus. 657
658
4.6 Conclusions 659
660
Reconstruction of the biogeographic origin of Myotis (Fig. 3) suggests that the East 661
Asian region is the center of origin for the radiation of modern Myotis and is thus probably 662
also the cradle for the entire Myotinae. The Indo-Malayan, Indo-Chinese, Sino-Himalayan, 663
21
and East Asiatic floras and faunas (Corbet and Hill, 1992) meet in this region and engendered 664
the highest biodiversity hotspots on earth (Myers et al., 2000). This geographical area, which 665
has been subjected to gradual uplifting of the Tibetan plateau (the trans-Himalayan range) 666
during the last several million years, includes several high mountain ranges exhibiting sharp 667
topographical complexity, and shows climates ranging from tropical to arctic. Such high 668
environmental and topographic complexity could have promoted speciation. 669
Considering the potential vagility of bats, and their current worldwide distribution, the 670
overall number of transcontinental migrations in the Myotis radiation is relatively low (Fig. 3). 671
This illustrates the relative inability of Myotis species to cross some physical barriers. In turn, 672
Myotis species can occupy a variety of ecological niches, and it is not exceptional to find up 673
to 12 species coexisting in sympatry. This and the high ecological and morphological 674
diversity found within each Myotis assemblage at a worldwide scale support the central 675
impact of biogeography on the Myotis evolutionary history. Climatic and vegetation changes 676
since the Middle Miocene might also have favored speciation during the evolution of Myotis. 677
Recent molecular surveys of Myotis taxa from different parts of the world identified 678
that several species currently diagnosed by morphological characters certainly contain species 679
complexes: M. muricola, M. siligorensis, M. horsfieldii (Francis et al., 2010), M. montivagus 680
(Görföl et al., 2013; Volleth and Heller, 2012), M. nigricans (Larsen et al., 2012a), and even 681
European species such as M. nattereri (Salicini et al., 2011). Many more cryptic species are 682
present in the Oriental region, as suggested by the presence of numerous unnamed taxa in our 683
reconstructions (Table S1). Here, phylogenetic tools based on the comparison of polymorphic 684
mitochondrial and nuclear molecular markers will be of central interest to identify cryptic 685
species, refine the Myotis systematics, and help choosing among priorities for biological 686
conservation purposes. Unfortunately, this region of high biodiversity is also facing high rates 687
of habitat destruction, and is in urgent need of protection to maintain this evolutionary 688
diversity (Sodhi et al., 2004). 689
690
Acknowledgements 691
We are deeply indebted to those individuals and institutions who generously donated 692
tissues for genetic analyses. These samples provide the foundation of such broad taxonomic 693
surveys and would not be possible without their help. These individuals are: P. Benda 694
(Natural History Museum, Prague), L. R. Heaney and W. Stanley (Field Museum of Natural 695
History, Chicago, USA), G. Jones (University of Bristol, UK), K. Kawai (Institute of Low 696
Temperature Science, Hokkaido University, Sapporo, Japan), T. Kingston (Boston 697
University, Boston, USA), S. Kruskop and A. V. Borisenko (Zoological Museum of Moscow 698
22
State University), T. H. Kunz (Boston University, Boston, USA), J. Ma (Beijing, China), V. 699
Matveev (Moscow Lomonosov State University, Russia), late O. von Helversen, and F. 700
Mayer (University of Erlangen, Germany). This research was supported by grants from the 701
Swiss National Funds for Scientific research to M. Ruedi and A. Cibois (3100A0-105588), 702
and from Bat Conservation International to B. Stadelmann. This publication is contribution No 703
2013-XXX of the Institut des Sciences de l‟Evolution de Montpellier (UMR 5554 – CNRS - 704
IRD). 705
706
References 707
Abe, H., Ishii, N., Itoo, T., Kaneko, Y., Maeda, K., Miura, S., Yoneda, M., 2005. A guide to 708 the mammals of Japan. Tokai University Press, Kanagawa, Japan. 709 Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A simulation study comparing 710 the performance of Bayesian Markov Chain Monte Carlo sampling and bootstrapping in 711 assessing phylogenetic confidence. Mol. Biol. Evol. 20, 255-266. 712 Bates, P.J.J., Hendrichsen, D.K., Walston, J.L., Hayes, B., 1999. A review of the mouse-eared 713 bats (Chiroptera: Vespertilionidae: Myotis) from Vietnam with significant new records. Acta 714 Chiropterol. 1, 47-74. 715 Bickham, J.W., Patton, J.C., Schlitter, D.A., Rautenbach, I.L., Honeycutt, R.L., 2004. 716 Molecular phylogenetics, karyotypic diversity, and partition of the genus Myotis (Chiroptera: 717 Vespertilionidae). Mol. Phylogenet. Evol. 33, 333-338. 718 Bininda-Emonds, O.R.P., Vazquez, D.P., Manne, L.L., 2000. The calculus of biodiversity: 719 integrating phylogeny and conservation. Trends Ecol. Evol. 15, 92-94. 720 Borisenko, A.V., Kruskop, S.V., Ivanova, N.V., 2008. A new mouse-eared bat (Mammalia: 721 Chiroptera: Vespertilionidae) from Vietnam. Russian J. Theriol. 7, 57-69. 722 Butler, P.M., 1984. Macrocelidea, Insectivora and Chiroptera from the Miocene of East 723 Africa. Paleovertebrata 14, 117-200. 724 Castella, V., Ruedi, M., Excoffier, L., 2001. Contrasted patterns of mitochondrial and nuclear 725 structure among nursery colonies of the bat Myotis myotis. J. Evolution. Biol. 14, 708-720. 726 Castella, V., Ruedi, M., Excoffier, L., Ibáñez, C., Arlettaz, R., Hausser, J., 2000. Is the 727 Gibraltar Strait a barrier to gene flow for the bat Myotis myotis (Chiroptera: 728 Vespertilionidae)? Mol. Ecol. 9, 1761-1772. 729 Cooper, S.J.B., Day, P.R., Reardon, T.B., Schulz, M., 2001. Assessment of species 730 boundaries in Australian Myotis (Chiroptera : Vespertilionidae) using mitochondrial DNA. J. 731 Mammal. 82, 328-338. 732 Corbet, G.B., Hill, J.E., 1992. The mammals of the Indomalayan region: a systematic review. 733 Oxford University Press, Oxford, UK. 734 Datzmann, T., Dolch, D., Batsaikhan, N., Kiefer, A., Helbig-Bonitz, M., phel, U., Stubbe, M., 735 Mayer, F., 2012. Cryptic diversity in Mongolian vespertilionid bats (Vespertilionidae, 736 Chiroptera, Mammalia). Results of the Mongolian-German biological expeditions since 1962, 737 No. 299. Acta Chiropterol. 14, 243-264. 738 Davalos, L.M., Cirranello, A.L., Geisler, J.H., Simmons, N.B., 2012. Understanding 739 phylogenetic incongruence: lessons from phyllostomid bats. Biol. Rev. 87, 991-1024. 740 Douady, C., Delsuc, F., Boucher, Y., Doolittle, W., Douzery, E., 2003. Comparison of 741 Bayesian and maximum likelihood bootstrap measures of phylogenetic reliability. Mol. Biol. 742 Evol. 20, 248-254. 743 Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics and 744 dating with confidence. Plos Biol. 4, 699-710. 745
23
Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling 746 trees. Bmc Evolutionary Biology 7, 214. 747 Evin, A., Baylac, M., Ruedi, M., Mucedda, M., Pons, J.M., 2008. Taxonomy, skull diversity 748 and evolution in a species complex of Myotis (Chirotpera: Vespertilionidae): a geometric 749 morphometric appraisal. Biol. J. Linn. Soc. 95, 529-538. 750 Fenton, M.B., Bogdanowicz, W., 2002. Relationships between external morphology and 751 foraging behaviour: bats in the genus Myotis. Can J. Zool. 80, 1004-1013. 752 Findley, J.S., 1972. Phenetic relationships among bats of the genus Myotis. Syt. Zool. 21, 31-753 52. 754 Francis, C., Borisenko, A., Ivanova, N., Eger, J., Lim, B., Guillen-Servent, A., Kruskop, S., 755 Mackie, I., Hebert, P., 2010. The role of DNA barcodes in understanding and conservation of 756 mammal diversity in Southeast Asia. Plos One 5, e12575. 757 García-Mudarra, J.L., Ibañez, C., Juste, J., 2009. The Straits of Gibraltar: barrier or bridge to 758 Ibero-Moroccan bat diversity? Biol. J. Linn. Soc. 96, 434-450. 759 Görföl, T., Estók, P., Csorba, G., 2013. The subspecies of Myotis montivagus - taxonomic 760 revision and species limits (Mammalia: Chiroptera: Vespertilionidae). Acta Zool. Hung. 59, 761 41-59. 762 Gunnell, G.F., Eiting, T.P., Simons, E.L., 2012. African Vespertilionoidea (Chiroptera) and 763 the antiquity of Myotinae. In: Gunnell, G.F., Simmons, N.B. (Eds.), Evolutionary history of 764 bats. Fossils, molecules and morphology. Cambridge University Press, New York, pp. 252-765 266. 766 Gunnell, G.F., Simmons, N.B., 2005. Fossil evidence and the origin of bats. J. Mamm. Evol. 767 12, 209-246. 768 Han, N.-J., Zhang, J.-S., Reardon, T., Lin, L.-K., Zhang, J.-P., Zhang, S.-Y., 2010. 769 Revalidation of Myotis taiwanensis Ärnbäck-Christie-Linde 1908 and its molecular 770 relationship with M. adversus (Horsfield 1824) (Vespertilionidae, Chiroptera). Acta 771 Chiropterol. 12, 449-456. 772 Happold, M., Happold, D.C.D., 2013. The Mammals of Africa: Hedgehogs, Shrews and Bats. 773 Bloomsbury Publishing, London, UK. 774 Helversen, v.O., Heller, K.-G., Mayer, F., Nemeth, A., Volleth, M., Gombkötö, P., 2001. 775 Cryptic mammalian species: a new species of whiskered bat (Myotis alcathoe n. sp.) in 776 Europe. Naturwissenschaften 88, 217-223. 777 Hill, J.E., Topál, G., 1973. The affinities of Pipistrellus ridleyi Thomas 1898 and Glischropus 778 rosseti Oey, 1951 (Chiroptera: Vespertilionidae). Bull.br.Mus.nat.Hist.Zool. 24, 447-454. 779 Hoofer, S.R., Reeder, S.A., Hansen, E.W., Van den Bussche, R.A., 2003. Molecular 780 phylogenetics and taxonomic review of noctilionoid and vespertilionoid bats (Chiroptera: 781 Yangochiroptera). J. Mammal. 84, 809-821. 782 Hoofer, S.R., Van den Bussche, R.A., 2003. Molecular phylogenetics of the chiropteran 783 family Vespertilionidae. Acta Chiropterol. 5 (supplement), 1-63. 784 Horácek, I., 2001. On the early history of vespertilionid bats in Europe: the Lower Miocene 785 record from the Bohemian Massif. Lynx 32, 123-154. 786 Horácek, I., Hanák, V., Gaisler, J. (Eds.), 2000. Bats of the Palearctic region: a taxonomic 787 and biogeographic review. Platan publishing house, Warsaw, Poland. 788 Ibáñez, C., García-Mudarra, J.L., Ruedi, M., Stadelmann, B., Juste, J., 2006. The Iberian 789 contribution to cryptic diversity in European bats. Acta Chiropterol. 8, 277-297. 790 Janis, C.M., 1993. Tertiary mammal evolution in the context of changing climates, vegetation, 791 and tectonic events. Annu. Rev. Ecol. Syst. 24, 467-500. 792 Jiang, T.L., Sun, K.P., Chou, C.H., Zhang, Z.Z., Feng, J., 2010. First record of Myotis flavus 793 (Chiroptera: Vespertilionidae) from mainland China and a reassessment of its taxonomic 794 status. Zootaxa 2414, 41-51. 795 Jones, G., Parsons, S., Zhang, S.Y., Stadelmann, B., Benda, P., Ruedi, M., 2006. Echolocation 796 calls, wing shape, diet and phylogenetic diagnosis of the endemic Chinese bat Myotis 797 pequinius. Acta Chiropterol. 8, 451-463. 798
24
Kawai, K., Nikaido, M., Harada, M., Matsumura, S., Lin, L.K., Wu, Y., Hasegawa, M., 799 Okada, N., 2002. Intra- and interfamily relationships of Vespertilionidae inferred by various 800 molecular markers including SINE insertion data. J. Mol. Evol. 55, 284-301. 801 Kawai, K., Nikaido, M., Harada, M., Matsumura, S., Lin, L.K., Wu, Y., Hasegawa, M., 802 Okada, N., 2003. The status of the Japanese and East Asian bats of the genus Myotis 803 (Vespertilionidae) based on mitochondrial sequences. Mol. Phylogenet. Evol. 28, 297-307. 804 Kitchener, D.J., Cooper, N., Maryanto, I., 1995. The Myotis adversus (Chiroptera: 805 Vespertilionidae) species complex in Eastern Indonesia, Australia, Papua New Guinea and the 806 Solomon Islands. Rec. West. Aust. Mus. 17, 191-212. 807 Koopman, K.F., 1993. Order Chiroptera. In: Wilson, D.E., Reeder, D.M. (Eds.), Mammal 808 species of the World. A taxonomic and geographic reference. Smithsonian Institution Press, 809 Washington, pp. 137-242. 810 Koopman, K.F., 1994. Chiroptera: Systematics. In: Niethammer, J., Schliemann, H., Starck, 811 D. (Eds.), Handbuch der Zoologie. de Gruyter, Berlin, Germany, pp. 100-109. 812 Kruskop, S.V., Borisenko, A.V., Ivanova, N.V., Lim, B.K., Eger, J.L., 2012. Genetic diversity 813 of northeastern Palaearctic bats as revealed by DNA barcodes. Acta Chiropterol. 14, 1-14. 814 La Val, R.K., 1973. A revision of the Neotropical bats of the genus Myotis. Natural History 815 Museum of Los Angeles County, Sciences Bulletin 15, 1-54. 816 Lack, J.B., Roehrs, Z.P., Stanley Jr, C.E., Ruedi, M., Van den Bussche, R.A., 2010. Molecular 817 phylogenetics of Myotis indicate familial-level divergence for the genus Cistugo (Chiroptera). 818 J. Mammal. 91, 976-992. 819 Lack, J.B., Van Den Bussche, R.A., 2010. Identifying the confounding factors in resolving 820 phylogenetic relationships in Vespertilionidae. J. Mammal. 91, 1435-1448. 821 Larsen, R.J., Knapp, M.C., Genoways, H.H., Khan, F.A.A., Larsen, P.A., Wilson, D.E., 822 Baker, R.J., 2012a. Genetic diversity of Neotropical Myotis (Chiroptera: Vespertilionidae) 823 with an emphasis on South American species. Plos One 7. 824 Larsen, R.J., Larsen, P.A., Genoways, H.H., Catzeflis, F.M., Geluso, K., Kwiecinski, G.G., 825 Pedersen, S.C., Simal, F., Baker, R.J., 2012b. Evolutionary history of Caribbean species of 826 Myotis, with evidence of a third Lesser Antillean endemic. Mamm. Biol. 77, 124-134. 827 Matveev, V.A., Kruskop, S.V., Kramerov, D.A., 2005. Revalidation of Myotis petax Hollister, 828 1912 and its new status in connection with M. daubentonii (Kuhl, 1817) (Vespertilionidae, 829 Chiroptera). Acta Chiropterol. 7, 23-37. 830 Mein, P., Ginsburg, L., 1997. Les mammifères du gisement miocène inférieur de Li Mae 831 Long, Thaïlande: systématique, biostratigraphie et paléoenvironnement. Geodiversitas 19, 832 783-844. 833 Menu, H., Hand, S., Sigé, B., 2002. Oldest Australian vespertilionid (Microchiroptera) from 834 the early Miocene of Riversleigh, Queensland. Archeringa 26, 319-331. 835 Menu, H., Sigé, B., 1971. Nyctalodontie et myotodontie, importants caractères de grades 836 évolutifs chez les chiroptères entomophages. Comptes Rendus de l'Académie des Sciences de 837 Paris 272, 1735-1738. 838 Miller, S.A., Dykes, D.D., Polesky, H.F., 1988. A simple salting procedure for extracting 839 DNA from human nucleated cells. Nucleid Acids Res. 16, 215. 840 Murphy, W.J., Eizirik, E., O'Brien, S.J., Madsen, O., Scally, M., Douady, C.J., Teeling, E., 841 Ryder, O.A., Stanhope, M.J., de Jong, W.W., Springer, M.S., 2001. Resolution of the early 842 placental mammal radiation using Bayesian phylogenetics. Science 294, 2348-2351. 843 Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. 844 Biodiversity hotspots for conservation priorities. Nature 403, 853-858. 845 Nylander, J.A.A., 2004. MrModeltest v. 2.3. Program distributed by the author. . Uppsala 846 University, Sweden 847 Paradis, E., 1998. Detecting shifts in diversification rates without fossils. Am. Nat. 15, 176-848 187. 849 Purvis, A., Agapow, P.-M., Gittleman, J.L., Mace, G.M., 2000. Nonrandom extinction and the 850 loss of evolutionary history. Science 288, 328-330. 851
25
Rambaut, A., Drummond, A.J., 2009. Tracer v1.5. Available from 852 http://beast.bio.ed.ac.uk/Tracer 853 Rautenbach, I.L., Bronner, G.N., Schlitter, D.A., 1993. Karyotypic data and attendant 854 systematic implications for the bats of southern Africa. Koedoe 36, 87-104. 855 Ree, R.H., Smith, S.A., 2008. Maximum likelihood inference of geographic range evolution 856 by dispersal, local extinction, and cladogenesis. Syst. Biol. 57, 4-14. 857 Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference under 858 mixed models. Bioinformatics 19, 1572-1574. 859 Ruedi, M., Biswas, J., Csorba, G., 2012. Bats from the wet: two new species of Tube-nosed 860 bats (Chiroptera: Vespertilionidae) from Meghalaya, India. Rev. suisse Zool. 119, 111-135. 861 Ruedi, M., Mayer, F., 2001. Molecular systematics of bats of the genus Myotis 862 (Vespertilionidae) suggests deterministic ecomorphological convergences. Mol. Phylogenet. 863 Evol. 21, 436-448. 864 Salicini, I., Ibáñez, C., Juste, J., 2011. Multilocus phylogeny and species delimitation within 865 the Natterer‟s bat species complex in the Western Palearctic. Mol. Phylogenet. Evol. 61, 888-866 898. 867 Sanderson, M.J., Donoghue, M.J., 1996. Reconstructing shifts in diversification rates on 868 phylogenetic trees. Trends Ecol. Evol. 11, 15-20. 869 Sechrest, W., Brooks, T.M., da Fonseca, G.A., Konstant, W.R., Mittermeier, R.A., Purvis, A., 870 Rylands, A.B., Gittleman, J.L., 2002. Hotspots and the conservation of evolutionary history. 871 Proc. Natl. Acad. Sci. USA 99, 2067-2071. 872 Simmons, N.B., 2005. Order Chiroptera. In: Wilson, D.E., Reeder, D.M. (Eds.), Mammal 873 species of the World. A taxonomic and geographic reference. Johns Hopkins University Press, 874 Washington, pp. 312-529. 875 Sodhi, N.S., Koh, L.P., Brook, B.W., Ng, P.K.L., 2004. Southeast Asian biodiversity: an 876 impending disaster. Trends Ecol. Evol. 19, 654-660. 877 Stadelmann, B., Herrera, G., Arroyo-Cabrales, J., Ruedi, M., 2004a. Molecular systematics of 878 the piscivorous bat Myotis (Pizonyx) vivesi. J. Mammal. 85, 133-139. 879 Stadelmann, B., Jacobs, D., Schoeman, C., Ruedi, M., 2004b. Phylogeny of African Myotis 880 bats (Chiroptera, Vespertilionidae) inferred from cytochrome b sequences. Acta Chiropterol. 881 6, 177-192. 882 Stadelmann, B., Kunz, T.H., Lin, L.K., Ruedi, M., 2007. Molecular phylogeny of New World 883 Myotis (Chiroptera, Vespertilionidae) inferred from mitochondrial and nuclear DNA genes. 884 Mol. Phylogenet. Evol. 43, 32-48. 885 Stamatakis, A., 2006. RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses 886 with thousands of taxa and mixed models. Bioinformatics 22, 2688-2690. 887 Tate, G.H., 1941. A review of the genus Myotis (Chiroptera) of Eurasia, with special 888 reference to species occurring in the East Indies. B. Am. Mus. Nat. Hist. 78, 537-565. 889 Teeling, E.C., Springer, M.S., Madsen, O., Bates, P., O'Brien, S.J., Murphy, W.J., 2005. A 890 molecular phylogeny for bats illuminates biogeography and the fossil record. Science 307, 891 580-584. 892 Tiunov, M.P., Kruskop, S.V., Feng, J., 2011. A new mouse-eared bat (Mammalia: Chiroptera, 893 Vespertilionidae) from South China. Acta Chiropterol. 13, 271-278. 894 Topál, G., 1970. On the systematic status of Pipistrellus annectans Dobson, 1871 and Myotis 895 primula Thomas, 1920 (Mammalia). Annals hist.-nat. Mus. nat. hung. 62, 373-379. 896 Topál, G., 1983. New and rare fossil mouse-eared bats from the Middle Pliocene of Hungary 897 (Mammalia, Chiroptera). Fragm. Min. Palaeont. 11, 43-54. 898 Tsytsulina, K., 2004. On the taxonomical status of Myotis abei Yoshikura, 1944 (Chiroptera, 899 Vespertilionidae). Zool. Sci. 21, 963-966. 900 Tsytsulina, K., Dick, M.H., Maeda, K., Masuda, R., 2012. Systematics and phylogeography of 901 the steppe whiskered bat Myotis aurascens Kuzyakin, 1935 (Chiroptera, Vespertilionidae). 902 Russian J. Theriol. 11, 1-20. 903
26
Volleth, M., Heller, K.G., 2012. Varations on a theme: Karyotype comparison in Eurasian 904 Myotis species and implications for phylogeny. Vespertilio 16, 329-350. 905 Volleth, M., Tidemann, C.R., 1991. The origin of the Australian Vespertilioninae bats, as 906 indicated by chromosomal studies. Z. Saugetierkd. 56, 321-330. 907 Weyeneth, N., Goodman, S.M., Ruedi, M., 2011. Do diversification models of Madagascar‟s 908 biota explain the population structure of the endemic bat Myotis goudoti (Chiroptera: 909 Vespertilionidae)? J. Biogeogr. 38, 44-54. 910 Zhang, Z.Z., Tan, X.Y., Sun, K.P., Liu, S., Xu, L.J., Feng, J., 2009. Molecular systematics of 911 the Chinese Myotis (Chiroptera, Vespertilionidae) inferred from cytochrome-b sequences. 912 Mammalia 73, 323-330. 913 Ziegler, R., 2003. Bats (Chiroptera, Mammalia) from Middle Miocene karstic fissure fillings 914 of Petersbuch near Eichstätt, Southern Franconian Alb (Bavaria). Geobios 36, 447-490. 915 916 917
918
27
Table 1 Clade name, level of support (over 50%) and inferred age estimated from the 919
different molecular analyses. Supports were recovered from a maximum likelihood (ML) or a 920
Bayesian approach (BA), and data sets consisted of either Cyt b or Rag 2 sequences analyzed 921
separately or in combination. Values are expressed as percent bootstraps (for ML) or posterior 922
probabilities (for BA). Node age is expressed in million years ago (MYA) followed by the 923
95% highest posterior density interval (HPD). 924
925
Table S1 Origin and biogeographic assignation of the specimens analyzed for the Cyt b and 926
Rag 2 genes. Species names not listed in Simmons (2005) are marked with a star (*) and their 927
taxonomy discussed in the material and methods section. The symbol ø denotes information 928
unavailable. Vouchers (if any) are deposited in the following institutions or belong to personal 929
collections: Instituto Politécnico Nacional in Mexico (CDR), Estacion Biologica de Doñana 930
(EBD), Frieder Mayer (FM), Field Museum of Natural History in Chicago (FMNH), Gareth 931
Jones (GJ), Jean-François Maillard, DIREN, Martinique (JFM), Kuniko Kawai (KK), Kishi 932
Maeda (KM), Natural History Museum of Geneva (MHNG), National Natural History 933
Museum of Paris (MNHN), Manuel Ruedi (MR), Museum of Vertebrate Zoology at Berkeley 934
(MVZ), National Museum Prague (NMP), Osaka City University Graduate School of 935
Medicine (OCUMS), Royal Ontario Museum (ROM), Sumiko Matsumura (SM), 936
Senckenberg Museum of Frankfurt (SMF), T.H. Kunz (THK), Tigga Kingston (TiK), 937
Museum of Texas Tech University (TK), Transvaal Museum, South Africa (TM), University 938
of Alaska Museum (UAM), UKM (Universiti Kebangsaan Malaysia), Zoological Museum of 939
Moscow State University (ZMMU), Z. Zhang (ZZ). 940
941
Figure 1 Maximum likelihood tree for 135 sequences of Myotis and six outgroup species 942
based on analyses of the combined mitochondrial and nuclear data set (1140 bp of Cyt b and 943
1148 bp of Rag 2). Nodal support is represented as standard bootstrap value (BP) obtained 944
with RAxML or posterior probabilities (PP) obtained with MrBayes. Empty circles represent 945
28
support values comprised between 50 and 70 %, grey circles values between 71 and 90% and 946
filled circles values above 90% in both ML and BA reconstructions. Species with uncertain 947
phylogenetic position (“floating species”) are indicated with a filled diamond. The various 948
clades discussed in the text are highlighted by boxes or broken lines. The inset represents the 949
geographic distribution of the 103 Myotis species listed in Simmons (2005) (right value) 950
versus the numbers sampled for the molecular reconstructions (left values) in the different 951
biogeographic regions of the world. 952
953
Figure 2 Chronogram of Myotis taxa based on Bayesian dating analysis using BEAST and the 954
combined mitochondrial and nuclear DNA gene dataset. Mean divergence values (expressed 955
as million year ago, MYA) are given at each node and horizontal bars represent the 95% 956
highest posterior density ranges. Clade names correspond to those given in Fig. 1. 957
958
Figure 3 Biogeographic evolution of ancestral areas of Myotis species reconstructed in the 959
chronogram of Fig. 2. The most likely ancestral areas determined with Lagrange under a 960
model with dispersals limited to adjacent bioregions are given in boxes at each node. The 961
inset map illustrates the seven biogeographic regions included for this study – A: the Western 962
Palaearctic, B: the Eastern Palaearctic, C: the Oriental, D: the Oceanian, E: the Ethiopian, F: 963
the Nearctic and G: the Neotropical region. Inferred range expansion events are highlighted 964
by black boxes (on the chronogram) or by arrows (on the world map), and vicariance events 965
by filled diamonds. The bottom scale represents time (in million years before present, MYA) 966
and the approximate subdivision of paleontological epochs. 967
968
Figure S1 Majority rule (50%) consensus Bayesian tree with posterior probability values 969
obtained with the analysis of 135 mitochondrial Cyt b sequences of Myotis taxa. Six outgroup 970
taxa from the Murininae and Kerivoulinae were used to root the tree. 971
972
29
Figure S2 Majority rule (50%) consensus Bayesian tree with posterior probability values 973
obtained with the analysis of 122 nuclear Rag 2 sequences of Myotis taxa. Six outgroup taxa 974
from the Murininae and Kerivoulinae were used to root the tree. 975
976
Figure S3 Occlusal views of right dentaries showing the nyctalodont lower molars (where the 977
postcristid links the hypoconulid) of Myotis latirostris (A.) and M. siligorensis alticraniatus 978
(B.), and the myotodont molars (where the postcristid links the entoconid) of Cistugo seabrae 979
(C.). See Menu and Sigé (1971) for a definition of lower molar configurations. 980
981
Table 1 : Clade name, level of support (over 50%) and inferred age estimated from the different molecular analyses. Supports were recovered from a maximum likelihood (ML) or a Bayesian approach (BA), and data sets consisted of either Cyt b or Rag 2 sequences analyzed separately or in combination. Values are expressed as percent bootstraps (for ML) or posterior probabilities (for BA). Nodes age are expressed in million years ago (MYA) followed by the 95% highest posterior density interval (HPD).
Clade name Cyt b Rag 2 combined Age (HPD)
ML BA ML BA ML BA MYA Myotis (incl. latirostris) 96 100 75 100 99 100 20.94 (23-20)
Myotis (excl. latirostris) 76 100 - - 80 100 18.66 (21-16) I New World clade 93 100 - - 97 100 12.34 (15-10) II-X Old World clade 62 100 - - 72 100 16.96 (20-14) II-III + VI-X Eurasian clade - 51 - - - 65 16.22 (19-14) II Large Myotis clade 92 100 - - 98 100 9.93 (12- 8) III Myotis clade 100 100 - - 100 100 9.33 (12-7) IV Oriental Myotis clade 92 100 - - 90 100 11.92 (14-10) V Ethiopian clade 94 100 - - 98 100 12.33 (15-10) VI Whiskered Myotis clade 85 100 - - 88 100 9.06 (11-7) VII muricola clade 72 100 - - 76 100 10.24 (13-8) VIII Eastern Myotis clade 100 100 - - 97 100 9.55 (12-7) IX Asian Myotis clade 87 100 52 96 87 100 8.77 (11-6) X Trawling Myotis clade 95 100 - - 99 100 7.30 (9-5)
50 < BP & PP < 70
70 < BP & PP < 90
90 < BP & PP
"floating" species
blythii omari
annamiticus
taiwanensis
cf. montivagus 4
cf. montivagus 2cf. montivagus 1
cf. montivagus 3
h. horsfieldii 2horsfieldii deignani
h. horsfieldii 1
cf. muricola 4
Murina annamiticus
latirostris lineage
Ne
w W
orld
Eth
iop
ian
wh
iskere
dm
uri
co
laO
rie
nta
lA
sia
nla
rge
My
oti
s
I
V
VI
VII
VIII
IV
IX
III
X
II
Eura
sia
n
Mu. pluvialis
daubentonii nathalinae
Nearc
tic
subcla
de
Neotr
op
ical
subcla
de
brandtii lineage
Old
World
pilosus
tra
wli
ng
cf. browni
cf. formosus 2cf. formosus 1
formosus flavus
WesternPalaearctic
EasternPalaearctic
Ethiopian
Oriental
13/14
1/4
6/7 21/26
8/14 17/19
13/19
Nearctic
0.0 MYA5.010.015.020.025.0
nesopolus
tricolor 2tricolor 1
brandtii
23.1
26.2
20.9
18.7
12.3
7.8
16.4
12.3
12.3
17.0
12.3
16.2
15.5
15.2
12.9
10.2
9.1
12.1
14.8 13.4
12.68.8
4.5
3.1
14.2
12.5
13.1
7.811.9
9.8
9.9
8.1
9.4
0.6
0.3
1.6
1.3
3.2
4.44.9
4.2
1.4
2.8
2.9
4.0
5.64.1
2.34.4
0.58
0.35
0.55
1.1
1.4
1.4
0.8
0.6
0.6
0.5
1.2
0.7
pilosus
b. oxygnathusb. omari
b. blythiib. ancilla
b. amurensis
n. tschuliensis
d. nathalinae
taiwanensis
s. alticraniatus 3s. alticraniatus 2
s. alticraniatus 1s. alticraniatus 4
lucifugus
I
V
VI
VII
VIII
IX
IV
X
III
II
cf. montivagus 4
formosus flavus
cf. formosus 1cf. formosus 2
6.6
6.0
4.46.01.0
2.7
1.1
8.5
10.9
9.4
6.43.8
8.5
6.8
6.0
5.58.1
9.3
8.2
11.3
11.6 9.53.1
1.7
7.8
5.9
2.5
2.6
2.9
2.4
9.6
8.8
5.5
4.9 2.2
3.4
3.0
2.2
8.6 2.0
6.34.9
7.3
6.0
5.11.7
9.38.4
6.8
7.1
1.0
8.98.2
6.42.7
1.34.4
5.94.3
3.0
1.8
1.9
1.0
0.9
1.1
1.5
1.9
1.71.1
1.6
0.9
0.5
1.3
0.8
brandtii
pilosus
Miocene Pliocene Plei
s. alticraniatus
FF
cf.muricola1-3cf.muricola 4
Harpiocephalus
l. carissima
n. larensis
cf. browni
taiwanensis
b. ancillab. blythii
n. tschuliensis
Oligocen.
I
V
VI
VII
VIII
IX
IV
X
III
II
f. flavus
cf. formosus
C
C
C
CCC
C
C
C
C
C
C
C
C
CC
CC
C
C
C C
C
CC
CC
C
C
C C
C
C
CC
C
C
C
F
F
F
F
F
F
F
FF
F
F
F
FF
F
F
GG
G
G
G
G
G
G
G
G
G
GG
G
GG
G
B
B
B
B
BB
BB
B
E
E
E
E
E
BF
A
FG
FG
FG
CE
CA
E
E
A
AC
A
AC
AC
A A
C
BC
A
C
AC
A
C
AC
A
B
AB
A
ABA
F
A
AA A
CB
D
A
C
vicariance (8x)
expansions (21x)
0.0 MYA5.010.015.020.025.0
AB
E ECC CE
C AC
A ABC BC
A AB
C BC
C BC
C CB
C CD
B BF
B BA
F FG
C CA
C AC
C AC
C AC
A AC
A AB
C BC
C BC
Ethiopian
clade
latirostris lineage
New World Myotis
Nearctic
subclade
brandtii lineage
Neotropical
subclade
Eurasian
clade
25 15 10 5 MYA 20 0
Old World Myotis
BC
C
B => BF
C => CE
C => CA
B
C
Nearly complete phylogeny of Myotis distributed worldwide
New World versus Old World clade
Likelihood models of range evolution identify Eastern Asia for their origins
Ancestral area reconstructions suggest 21 intercontinental dispersals during 21 MYA of
evolution
Discrepancies in molecular dating and interpretation of fossil record due to homoplasies
in dental evolution