Is the Gibraltar Strait a barrier to gene flow for the bat
Myotis myotis (Chiroptera: Vespertilionidae)?
V. C A S T E L L A , * M . R U E D I , * † L . E X C O F F I E R , ‡ C . I B Á Ñ E Z , § R . A R L E T T A Z * and J . H A U S S E R *
*Laboratoire de Zoologie, Institut d’Ecologie, Bâtiment de Biologie, Université de Lausanne, CH-1015 Lausanne, Switzerland,
‡Laboratoire de Génétique et Biométrie, Département d’Anthropologie, Université de Genève, 1211 Genève 24, Switzerland,
§Estación Biológica de Doñana, CSIC, Apartado 1056, 41080 Sevilla, Spain
Abstract
Because of their role in limiting gene flow, geographical barriers like mountains or seas
often coincide with intraspecific genetic discontinuities. Although the Strait of Gibraltar
represents such a potential barrier for both plants and animals, few studies have been
conducted on its impact on gene flow. Here we test this effect on a bat species ( Myotis myotis)
which is apparently distributed on both sides of the strait. Six colonies of 20 Myotis myotis
each were sampled in southern Spain and northern Morocco along a linear transect of
1350 km. Results based on six nuclear microsatellite loci reveal no significant population
structure within regions, but a complete isolation between bats sampled on each side of the
strait. Variability at 600 bp of a mitochondrial gene (cytochrome b) confirms the existence of
two genetically distinct and perfectly segregating clades, which diverged several million years
ago. Despite the narrowness of the Gibraltar Strait (14 km), these molecular data suggest that
neither males, nor females from either region have ever reproduced on the opposite side
of the strait. Comparisons of molecular divergence with bats from a closely related species
(M. blythii) suggest that the North African clade is possibly a distinct taxon warranting
full species rank. We provisionally refer to it as Myotis cf punicus Felten 1977, but a definitive
systematic understanding of the whole Mouse-eared bat species complex awaits further
genetic sampling, especially in the Eastern Mediterranean areas.
Keywords: Chiroptera, cytochrome b, Gibraltar, microsatellite, Myotis, population structure
Received 27 February 2000; revision received 5 June 2000; accepted 5 June 2000
Introduction
Recent reviews of intraspecific genetic variation (Taberlet
et al. 1998; Avise & Walker 1999; Hewitt 1999) have shown
that many animal or plant species present important genetic
discontinuities when they are sampled over wide ranges.
These discontinuities often result from recent admixture
of populations, which diverged in allopatry (e.g. in glacial
refugia) or coincide with current or historical barriers to
gene flow (e.g. mountain ranges or ancient glaciers). The
Gibraltar Strait, which separates the Iberian Peninsula
from the Maghreb (Fig. 1) by a minimum gap of 14 km of
open sea, could thus represent a barrier to gene flow. Indeed,
Correspondence: M. Ruedi: †Present address: Muséum d’histoire
naturelle, CP 6434, CH-1211 Genève 6, Switzwerland. Fax: + 41
22 4186301; E-mail: manuel.ruedi@mhn.ville-ge.ch
© 2000 Blackwell Science Ltd
since the last Messinian crisis (about 5.5 Ma) when most
of the present-day Mediterranean Sea dried-up, no land
bridge existed between these two land masses (see, e.g.
Steininger et al. 1985). Thus, dispersal of both plants and
animals across this Strait must have been severely limited
since the Pliocene. Restricted over-water dispersal is
suggested by Valdés (1991) who estimated that as much
as 25% of the 3500 species of plants distributed in southern
Spain and northern Morocco only occur in one side of the
Strait. For mammals, even highly vagile species like the
wolf (Canis lupus) or the jackal (Canis aureus) are currently
found in Iberia and the Maghreb, respectively, but neither
apparently ever crossed the Gibraltar Strait successfully.
Genetic studies in Wild mice (Mus spretus) and Wood mice
(Apodemus sylvaticus) have also suggested that the Gibraltar
Strait is an effective barrier to gene flow (Boursot et al.
1985; Filippucci 1992). More generally, Dobson (1998)
Fig. 1 Distribution of the Greater Mouseeared bat (Myotis myotis) in the western
Mediterranean area (in black) according
to Arlettaz et al. (1997b). The broken line
delineates the limits of the North African
form, referred here as M. cf punicus Felten
1977. The inset provides an enlargement of
the area around the Gibraltar Strait. Numbers
indicate the position of the following colonies
(locality, province, country): Boumahden,
Agadir, Morocco (1); Azrou, Meknes, Morocco
(2); Mina del Agua, Ceuta, Spain (3); Cueva
del Agua, Cadiz, Spain (4); Canillas de
Aceituno, Malaga, Spain (5); Denia, Alicante,
Spain (6); Inca, Mallorca, Spain (7); Oleta,
Corsica, France (8); Ulassai, Sardinia, Italy
(9). See also the Appendix for locations of
other examined specimens and for those
of M. blythii.
considers that only four of the 17 nonflying mammal
species occurring wild in both sides of the Strait were not
transported by man. In contrast, H. Dobson stresses that
most bat species found in north-west Africa (26 species)
and in Iberia (25 species) are naturally occurring in both
regions, and are thus much better over-water dispersers.
In fact, several other taxa are distributed on both sides of
the Strait, but its impact as a geographical barrier to
restrict gene flow between populations has rarely been
addressed specifically. For instance, a genetic survey of
rare plants distributed in southern Spain and Morocco
(genus Androcymbium, Colchicaceae) revealed that population differentiation was similar when comparisons were
made within countries or across Gibraltar, suggesting that
this strait was not a strong barrier for these species (PedrolaMonfort & Caujapé-Castells (1994). More recently, Burban
et al. (1999) found that both the maritime pine and its
homopteran parasite Matsococcus feytaudi shared common
haplotypes in southern Spain and extreme northern
Morocco, while populations from central Morocco were
completely distinct.
The Greater Mouse-eared bat (Myotis myotis) is a
widespread species, which is apparently distributed from
Morocco to Europe and the Middle east, including several
Mediterranean islands (reviewed in Benda & Horacek
1995; Arlettaz et al. 1997b). It also probably occurs in the
remote Azore Islands, more than 1400 km west from the
Iberian Peninsula (Palmeirim 1979). Ringing studies have
shown that this strong flier is able to cover annually several
hundreds of kilometers between summer and winter
roosts (Horacek 1985; Paz et al. 1986). Moreover, females
may commute daily up to 25 km between their nursery
colonies and the feeding territories during lactation
(Arlettaz 1996, 1999). Owing to such behavioural characteristics, we can anticipate that populations of the Greater
Mouse-eared bat found on both sides of the Gibraltar
Strait may exchange migrants across the channel. We,
therefore, assessed the impact of the strait as a barrier by
comparing levels of gene flow within southern Spain and
within North Africa, and across Gibraltar, using a combination of nuclear and mitochondrial DNA (mtDNA)
markers. These two classes of genetic markers enabled
us to estimate in which proportion males and females
contributed to the migrant pool, and provided a phylogenetic perspective of their evolution.
Materials and methods
Taxonomic and geographical sampling
As European populations of Mouse-eared bats are
protected in most countries, animals were caught and
sampled noninvasively under appropriate license (see
Acknowledgements). Upon capture, individuals were
identified according to their lengths of forearm, ear and
upper toothrow, (Arlettaz et al. 1997b) and sexed. A
total of 119 bats from six colonies were sampled with a
medical biopsy punch on the membrane of each wing
(Worthington Wilmer & Barratt 1996). Animals were
released back to their colony within one hour of capture.
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Table 1 Observed (HO) and expected heterozygosity (HE), number of alleles (A) and inbreeding coefficient (FIS) of Myotis myotis
populations calculated for each locus and over all loci. The mean values and standard deviations (SD) were calculated among three
nursery colonies within each country. Twenty bats were sampled in each population, except in Ceuta (Moroccan side) where only 19
individuals were captured
Spanish populations
B11
C113
D9
E24
G25
H29
Overall
Moroccan populations
HO
HE
A
FIS
HO
HE
A
FIS
0.87 ± 0.06
0.12 ± 0.03
0.92 ± 0.06
0.83 ± 0.08
0.55 ± 0.09
0.90 ± 0.05
0.67 ± 0.03
0.83 ± 0.02
0.11 ± 0.03
0.90 ± 0.03
0.85 ± 0.08
0.62 ± 0.01
0.91 ± 0.01
0.71 ± 0.02
8.3 ± 0.6
2.0 ± 0.0
12.7 ± 0.6
13.7 ± 1.2
3.7 ± 1.2
11.7 ± 1.5
8.7 ± 0.3
– 0.038
– 0.039
– 0.017
0.021
0.114
0.013
0.012
0.54 ± 0.14
0.03 ± 0.03
0.80 ± 0.08
0.98 ± 0.03
0.03 ± 0.06
0.83 ± 0.08
0.49 ± 0.04
0.56 ± 0.10
0.03 ± 0.03
0.88 ± 0.06
0.92 ± 0.01
0.07 ± 0.06
0.88 ± 0.04
0.56 ± 0.02
5.0 ± 1.0
1.7 ± 0.6
11.0 ± 2.0
14.3 ± 1.2
1.7 ± 0.6
10.3 ± 0.6
7.3 ± 0.6
0.033
0.000
0.092
– 0.069
0.493
0.054
0.035
Three colonies of Mouse-eared bats were sampled on
each side of the Strait (Fig. 1). A distance of 270 – 770 km
separated the North African colonies (nos 1– 3; geographical
distances were measured on a map and rounded to the
nearest 10 km), while those from southern Spain (nos 4 – 6)
were sampled at intervals of 130 – 580 km. These six
colonies represent animals living along a nearly linear
transect of 1350 km. The colony of Ceuta (referred here
collectively with the Moroccan colonies, though is politically
a province of Spain) is located just in front of the Iberian
colony of Cadiz. They are separated by only 60 km of
land and 14 km of sea (see inset of Fig. 1). Because there
is a long lasting controversy about the taxonomic position of North African Mouse-eared bats (Felten et al. 1977;
Arlettaz et al. 1997b), 18 Lesser Mouse-eared bats (Myotis
blythii) coming from various places in Europe and Asia
were analysed for comparative purposes. We also included
several Mouse-eared bats from Mallorca, Corsica, and
Sardinia (see Fig. 1 and Appendix). Myotis nattereri, a
smaller species closely related to the large Myotis (Ruedi
& Mayer submitted), was used as an outgroup to root the
phylogenetic trees.
DNA amplification
DNA was extracted from half wing punches using a salt
protocol (Miller et al. 1988). For the purpose of measuring
the genetic structure of bat populations at the nuclear
level, six microsatellite loci developed specifically for
the Greater Mouse-eared bat were analysed (Table 1). The
GenBank accession nos and polymerase chain reaction
(PCR) conditions used to amplify these loci are detailed
in Castella & Ruedi (2000). The female contribution to
gene flow was evaluated by sequencing the first 600 bp of
a mtDNA marker, the cytochrome b gene. This gene was
amplified with primers L14724 (Kocher et al. 1989) and
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
MVZ16 (Smith & Patton 1993), and sequenced with an
automated DNA sequencer (Applied Biosystems model
373XL) following standard methods. This mitochondrial
gene was sequenced in the following subsample of
animals from the transect: 14 M. myotis from Spain (10
from Cadiz, two each from Malaga and Alicante), and 14
from Morocco (10 from Ceuta, and two each from Azrou
and Agadir). The same gene was also sequenced in the
other Myotis sampled elsewhere in Europe and Asia (see
Appendix). Sequences were aligned and edited visually
using Sequencher 3.0 (Gene Codes Corp.). All different
haplotypes are deposited in GenBank under accession
nos AF246241–AF246261. Abbreviation of haplotypes or
populations refer to Morocco (MA), Spain (ES), Sardinia
(SA), Corsica (CO), Switzerland (CH), Greece (GR), the
Czech Republic (TS), and Kirghizstan (KZ).
Microsatellite analysis
The genetic variability found within the six populations
was quantified by the number of alleles and by the
observed (HO) and expected (HE, Nei 1987) heterozygosity.
Means and standard deviations of these variables were
calculated separately for both the three pooled Spanish
and the three pooled Moroccan colonies. To determine
whether the six surveyed loci represent independent
samples of the Mouse-eared bats’ genome, genotypic
linkage disequilibrium was investigated for each pair
of loci using the program genepop 3.1c (Raymond &
Rousset 1995). Genotypic proportions were tested for
departure from Hardy–Weinberg (H–W) expectations
using randomization tests available in the fstat 2.9
package (Goudet 1999a). For H–W expectations within
populations, alleles were permuted 1000 times among
individuals, while alleles were permuted among populations for H–W expectations within regions. The same
program was used to test whether inbreeding coefficients
(FIS) deviated significantly from zero. In order to get
mean FIS values for each region, computations were run
independently for the two sets of three colonies. In both
cases, P-values were obtained following 1800 permutations of alleles within populations with the specific
alternative hypothesis of heterozygote deficiency. Sequential
Bonferroni corrections were used to compute the critical
significance levels for all simultaneous statistical tests
(Rice 1989).
Nuclear population structure was quantified by
estimating the differentiation between colonies with Fstatistics (Wright 1978). Pairwise FST were computed
according to Weir & Cockerham (1984) and population
differentiation was tested by randomizing (1500 permutations) multilocus genotypes between each pair of
samples with fstat 2.9. A hierarchical design (Schneider
et al. 1996), including the colony, the regional and the
whole transect levels, was also used to test for the
effect of the Gibraltar Strait on the apportionment of
molecular variance. To fully exploit genetic information
at the individual level, the multilocus genotype of each
bat was introduced in a principal component analysis
using the program pca-gen 1.2 (Goudet 1999b). As with
assignment tests (Paetkau et al. 1995, 1997), this method
allows the identification of alien genotypes within local
population.
Haplotype analysis and phylogenetic reconstructions
The small number of bats sequenced in most populations
precluded any test of a hierarchical design of population
differentiation. Contrary to the microsatellites, we only
compared haplotypic variability within regions (i.e. Spain
or Morocco) and among regions. Thus, the mitochondrial
population structure was analysed by estimating the
correlation of haplotypes drawn from the same side of
the Gibraltar Strait to haplotypes drawn randomly from
the total sample in terms of φST (Excoffier et al. 1992).
Patterns of DNA evolution among haplotypes and all
phylogenetic reconstructions were done with paup 4.0b4
(Swofford 1996). Phylogenetic relationships were recovered
with a uniform parsimony analysis (branch-and-bound
search), and reliability of nodes ascertained with 500
bootstrap resampling. Other methods of phylogenetic
inference (i.e neighbour-joining or maximum likelihood
methods) gave similar results. Allozyme data previously
used to differentiate M. myotis and M. blythii (Arlettaz
et al. 1997b) were reanalysed here in a phylogenetic context
based on a complete data set of 35 loci (see Ruedi et al.
1990). Populations analysed here included M. myotis from
Spain (n = 12), and Switzerland (n = 5), M. blythii from
Spain (n = 9), Switzerland (n = 5) and Kighizstan (n = 7),
Moroccan bats (n = 7) and a composite population of
three bats from Sardinia and two from Corsica. Allozyme
allele frequencies were transformed into Nei’s genetic
distance (Nei 1987) and phylogenetic relationships were
reconstructed with the neighbour subroutine in the
program phylip 3.5c (Felsenstein 1993). The resulting tree
was also bootstrapped 500 times with phylip.
Results
Microsatellite variability
Within all colonies, bats showed considerable genetic
variability at the six microsatellite loci with total number
of alleles ranging from 41 to 54 and gene diversity from
0.54 to 0.72 (Table 1; individual genotypes are available
at http://www.unil.ch/izea/research.html#mmyotis). The
loci C113 and G25 were the least variable as they were
nearly fixed in one or both regions (Table 1 and Fig. 2).
One test of linkage disequilibrium out of the 15 possible
was highly significant in the colony of Ceuta (P < 0.001)
and involved loci D9 and H29. This association remained
significant even after a sequential Bonferroni procedure
(Rice 1989). In all other populations, no evidence of association was detected for any locus combination. Because one
significant test out of 15 would be expected by chance
alone at the population level (P > 0.1; binomial test with
α = 0.05), there was no evidence for strong dependence
between loci. All colonies, either considered independently
or grouped by region, were in Hardy–Weinberg equilibrium
for each locus and over all loci. The only population which
had an inbreeding coefficient that differed significantly
from zero (FIS = 0.105, Azrou) was no more deviating from
random mating when sequential Bonferroni correction
was applied. The high (but not significant) FIS value for
the locus G25 in Moroccan populations (Table 1) was due
to the occurrence of one particular individual in the
colony of Ceuta. This bat was homozygous for a rare
Moroccan allele (13 repeats; Fig. 2), while all remaining
bats were either homozygous for the most common allele
(56 bats) or heterozygous for both alleles (2 bats).
Patterns of genetic variation within regions
A total of 69 and 56 alleles were scored in the three
Spanish and three Moroccan colonies, respectively. Most
of these alleles were widespread among colonies within
regions, except 19 and 10 of them which were limited to
one Iberian or one Moroccan colony, respectively. The
majority of these private alleles only occurred at very low
frequencies (mostly one single copy). The rather uniform
number of alleles and heterozygosity calculated among
colonies (low standard deviations in Table 1) also suggests
that the microsatellite variability is high but evenly distributed among populations from the same region. This is
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Fig. 2 Distribution of allele frequencies at
the six microsatellite loci for 60 Spanish
(black bars) and 59 Moroccan Mouseeared bats (white bars). Allele sizes are
indicated in repeat numbers on the x-axis.
Some loci clearly differentiate bats from
both regions.
Table 2 Microsatellite estimates of population differentiation
within and between regions (pairwise FST) are indicated above
the diagonal. Asterisks below the diagonal indicate significant
genetic structuring among Spanish and Moroccan samples
obtained with the exact G-test (uncorrected P-values). Sample
size is the same as in Table 1
Spain
Cadiz
Malaga
Alicante
Agadir
Azrou
Ceuta
Morocco
Cadiz
Malaga
Alicante
Agadir
Azrou
Ceuta
0
NS
NS
***
***
***
– 0.002
0
NS
***
***
***
0.009
0.004
0
***
***
***
0.344
0.340
0.357
0
NS
*
0.336
0.333
0.353
– 0.002
0
NS
0.320
0.314
0.337
0.017
0.019
0
NS, non significant; *P < 0.5, ***P < 0.001.
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
confirmed by exact tests of population differentiation,
which indicate that colonies from the same side of the
strait are not significantly differentiated from each other
(Table 2). The only apparent exception concerned the
colonies of Agadir and Ceuta, but the differentiation
was no more significant after a Bonferroni correction.
Assignment tests based on the multilocus genotype of
individuals (Paetkau et al. 1995) sampled within a given
region classify only 40% of bats in their correct colony of
origin (results not shown). Although the colonies were
sampled up to 770 km from each other in Spain or in
Morocco, the microsatellite data suggest high ongoing
nuclear gene flow and weak genetic subdivision among
colonies within regions.
Because microsatellites are nuclear markers inherited
biparentally (Tautz & Renz 1984), they reflect the movements of both males and females. Thus, the dispersal
Microsatellites
Among regions
Among populations within region
Within populations
mtDNA
%
Φ
%
Φ
33.3
0.5
66.2
0.333
0.007
0.338
99.1
—*
0.9
0.991
—*
—
*hierarchical level not included for mtDNA.
of one sex, e.g. the males only, would be sufficient to
homogenize populations, even in the absence of female
dispersal. We, therefore, measured patterns of gene
flow through mtDNA variation which is transmitted by
females only (Avise 1994). Twenty-eight individuals were
sequenced for 600 bp of the cytochrome b gene (see
Appendix). The 14 Greater Mouse-eared bats from Spain
differed at only two positions (one transition and one
transversion substitution), determining three different
haplotypes. The most common haplotype (ES1) was found
in nine bats from Cadiz and two from Malaga. Similarly,
only three positions were variable (all transitions) in the
cytochrome b of the 14 Moroccan bats, determining four
closely related haplotypes. The most common Moroccan
haplotype (MA1) was found in all individuals from
Ceuta, and one bat from Azrou. Some haplotypes were
found in one colony alone (e.g. MA3 from Agadir, or ES3
from Alicante) suggesting that females may be more
philopatric than males to their natal colonies. The low
number of sequenced individuals in most colonies precludes any formal testing of this hypothesis, but a similar
bias of sex dispersal was already suggested in a population
of M. myotis from Germany (Petri et al. 1997).
Patterns of genetic variation across the Gibraltar Strait
In sharp contrast to the genetic homogeneity and
high levels of nuclear gene flow inferred within regions,
comparisons of bats sampled across the Gibraltar
Strait yielded strong levels of genetic differentiation. For
the microsatellite data set, three loci showed contrasting
patterns of variation. In the B11 locus, none of the 17
different alleles was distributed on both sides of the Strait.
All alleles comprising 4 –11 repeats were found exclusively
in Morocco, while all those comprising 15 – 25 repeats
were typical of Spanish bats (Fig. 2). In the G25 locus, a
single copy (out of 120 sampled) of the 14-repeat allele
was found in the Spanish population of Cadiz, while
this allele was nearly fixed (mean frequency 0.97) in all
Moroccan colonies (Fig. 2). Finally, the locus C113 was
nearly fixed for the 7-repeat allele in Spain (average
frequency = 0.94) and for the 6-repeat allele in Morocco
Table 3 Analysis of microsatellite and
mtDNA variance across the Gibraltar
Strait. The percentage of variance (%) and
the Φ-statistic are given at each hierarchical
level (see Excoffier et al. 1992). Microsatellite
results are based on the same sample size
as in Table 1, but for mtDNA only 14 bats
were analysed for Spanish and for Moroccan
populations (see text)
(mean frequency 0.98; Fig. 2). For both the G25 and C113
loci, results from a broader population survey involving
420 M. myotis from western Europe (V. Castella & M. Ruedi,
unpublished data) showed that rare ‘Moroccan’ alleles
can be found sporadically in several European populations
sampled as far away as Switzerland or Poland. The three
remaining loci showed more overlapping allele distribution
across the strait (Fig. 2).
Contrary to the situation observed within regions,
all pairs of populations were significantly differentiated
when comparisons were made across the Gibraltar Strait,
even after Bonferroni adjustment (Table 2). Moreover,
classical FST values were one or two orders of magnitude
larger when pairs of populations were compared between
regions rather than within regions (Table 2). As a consequence, assignment tests among regions classified all
bats in their correct country of origin (results not shown). A
hierarchical analysis of molecular variance showed that
most of the microsatellite variability (66%) was distributed
within colonies. However, along the whole transect of
1350 km, only a negligible part of the genetic variance
(0.5%) was due to differences among colonies, while 33%
was explained by the effect of the Gibraltar Strait (Table 3).
In order to make comparisons with a closely related
species, we analysed the same six microsatellite loci in 18
M. blythii sampled in various locations in Europe and
Asia (see Appendix). This heterogeneous sample included,
in particular, two individuals from a nursery colony in
Tarifa (Spain) located just on the Iberian border of the
Gibraltar Strait (Fig. 1). Among the 52 alleles that were
scored in this new sample, 20 were exclusively found in
M. blythii. In particular, most alleles (15 out of 17) of the
dinucleotide locus E24 showed single nucleotide shifts
when compared to Iberian and Moroccan alleles. This
suggests that an insertion or deletion unique to M. blythii
has occurred within this locus. In the diallelic C113 locus,
both the 6-repeat (typical of Moroccan bats) and the 7repeat allele (typical of Iberian bats) were present in this
species. The 6-repeat allele was the most frequent
(frequency of 0.72) while the other allele was found in 10
heterozygotes. All M. blythii tested were fixed for the
17-repeat allele at the G25 locus; this allele is otherwise
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Fig. 3 Factor map of the two main axis of the Principal Component Analysis (PCA) carried out with individual microsatellite
genotypes. The x-axis represents 30% inertia (FST = 0.19), while the y-axis has 7% inertia (FST = 0.04). All Spanish bats (represented by
black triangles) display positive values on the first axis, while negative values characterize Moroccan bats (represented by white
squares). No individual of these two groups lay in an intermediate position. European Myotis blythii, represented by grey circles,
appear more closely related to Iberian rather than to Moroccan bats.
typical of M. myotis from the Iberian region. As shown by
a principal component analysis, multilocus genotypes of
these 18 M. blythii appear to be more closely related to the
Iberian M. myotis samples, than to the Moroccan bats
(Fig. 3).
To come back to our main species, patterns of mtDNA
variation across the Gibraltar Strait supports the same
conclusion as the nuclear markers. Nearly all the molecular variance was observed for comparisons made across
the Strait (99.1%; Table 3). Thus, all haplotypes could be
unambiguously assigned to their correct region of origin.
The two groups of haplotypes differed by a mean (± SD)
genetic distance of 10.6 ± 0.2% (Kimura 2-parameter correction). This represents 54 – 59 observed substitutions
over the 600 bp sequenced. To expand the geographical
coverage of Mouse-eared bats from the Mediterranean
region, we also sequenced animals from other parts of
Europe and from the islands of Mallorca (Spain), Corsica
(France), and Sardinia (Italy), and also compared sequences
from M. blythii (see Appendix). M. myotis from Spain
(including bats from Mallorca) and Switzerland were all
very similar to each other (within 1% of sequence divergence). Surprisingly, four Spanish M. blythii from Tarifa,
Cadiz and Alicante had the same haplotype (ES1b) that is
found in most Iberian M. myotis. M. blythii from the Czech
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Republic and Greece were slightly more divergent (up to
2.8% sequence divergence), while the two central Asian
Lesser Mouse-eared bats differed by a mean of 6.2 ± 0.4%
from European samples. By contrast, sequences from
Corsica and Sardinia were all closely related to the Moroccan
haplotypes. Thus, European M. myotis and M. blythii are
much more closely related to each other than to the
Moroccan sample. Accordingly, a phylogenetic analysis of
all different haplotypes places the Moroccan, Corsican,
and Sardinian samples in a single, well-supported clade
(Fig. 4A). All other haplotypes form a distinct clade, also
highly supported by bootstrap resampling. This clade
include two Kirghiz M. blythii which emerge first as an
early offshoot, while sequences of European M. blythii
and M. myotis constitute a closely related crown-group
which did not attain reciprocal monophyly.
Phylogenetic analysis of Mouse-eared bat populations
using 35 allozyme loci and M. nattereri as an outgroup
confirms the close relationships of Moroccan bats with
those of Corsica and Sardinia (93% bootstrap support;
Fig. 4B). However, contrary to results of microsatellites
(Fig. 3) and DNA sequencing (Fig. 4A), this allozyme tree
suggests a closer phylogenetic relationship of this North
African taxon with M. myotis, but this node is not firmly
established (63% bootstrap support).
Fig. 4 Phylogenetic relationships of various Mouse-eared bats
deduced from mtDNA haplotypes (gene tree A), or from allozyme
data (population tree B). The haplotype tree is a consensus of all
equally parsimonious trees (score 154 steps) obtained from a branchand-bound search using Myotis nattereri as an outgroup. Notice
that haplotypes of M. myotis and M. blythii are not reciprocally
monophyletic. The allozyme tree is a neighbour-joining tree based
on Nei’s (Nei 1987) genetic distances calculated among populations. Nodes with bootstrap proportions above 50% are indicated.
Abbreviation of haplotypes or populations refer to samples
from Morocco (MA), Spain (ES), Sardinia (SA), Corsica (CO),
Switzerland (CH), Greece (GR), the Czech Republic (TS), and
Kirghizstan (KZ).
Discussion
Greater Mouse-eared bats are strong fliers able to cover
several hundreds of kilometers annually (Horacek 1985;
Paz et al. 1986). Consistent with this high dispersal ability,
microsatellite data show that colonies within Spain, or
within Morocco are only weakly differentiated (Table 2).
This suggests that high levels of gene flow prevail among
continental colonies over large geographical distances (at
least up to 770 km). Accordingly, all mtDNA haplotypes
sequenced within Spain or within Morocco were identical
or very similar to each other. These patterns of microsatellite
and mtDNA variation within each region are consistent
with recent genetic data obtained for another Palaearctic
species of bat, Nyctalus noctula. This true migratory species
covers several hundreds of kilometers every year between
breeding and wintering grounds, and exhibits no significant
nuclear genetic differentiation within a core area spawning
more than 3000 km (Petit & Mayer 1999).
In sharp contrast to the weak differentiation of populations observed within Spain or within Morocco, the same
molecular markers revealed a clear dichotomy when
comparisons were made across the Gibraltar Strait. For
instance, the Moroccan and Iberian populations share no
mtDNA haplotypes in common, and at the microsatellite
locus B11, all short alleles are exclusively found in North
Africa, while longer repeat alleles are all Iberian (Fig. 2).
Thus, haplotypes and multilocus genotypes of all sampled
individuals can be easily assigned to their correct region
of origin (Figs 3 and 4A). The 119 bats analysed from only
six populations may represent a small sample to detect
direct migrants. This is certainly true for first generation
migrants which could occur in the area without reproducing, and which could have remained unsampled.
However, unlike direct methods (e.g. ringing studies),
which provide punctual information on the movements
of individuals, genetic data are representative of the
movement of genes accumulated during many generations (Slatkin 1994). In a principal component analysis
based on the multilocus genotype of each individual,
recent hybrid genotypes are expected to occur in an intermediate position between pure-bred genotypes (but see
Lugon-Moulin et al. 1999). As shown in Fig. 3, all Iberian
and Moroccan bats appear to be pure-bred, with no
apparent sign of recent admixture of both groups. Because
mtDNA markers are inherited clonally through the mother,
they could introgress into a new population through a
single reproducing, alien female. Yet haplotypes segregate in two clades endemic to each region. Thus, the
evidence supports that no Spanish migrant contributed to
any significant gene flow into the Moroccan populations
or vice-versa. Moreover, the amount of genetic divergence
among cytochrome b haplotypes measured between
European Myotis myotis and African bats (10.6 ± 0.2% for
Kimura 2-parameter corrected distance) suggests that
they became separated long ago. Comparable amounts
of genetic divergence were measured between the two
cryptic species of pipistrelle bats (11% Barratt et al. 1997)
or among congeneric mammals in general (reviewed in
Johns & Avise 1998). To maintain such high levels of genetic
difference between conspecific populations, a strong,
persistent and ancient barrier preventing any significant
gene flow would be needed. Fossils of typical M. myotis
are known at least since the Pleistocene in the Iberian
Peninsula (Sevilla 1989). Thus, owing to their vagility, these
bats should have had ample time to exchange migrants
with North Africa. Moreover, Mouse-eared bats have
successfully colonized all major islands of the Mediterranean
Sea, suggesting that they can be very effective colonizers.
For instance, bats on Mallorca, which is about 200 km away
from mainland populations (Fig. 1), have haplotypes identical or very similar to Spanish ones (Fig. 4A). Both these
temporal and physical arguments suggest that the 14 km
of open sea, which separates Spain from Morocco (Fig. 1),
is certainly not sufficient per se to prevent gene flow.
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
We excluded the possibility that the North African bats
were conspecific with another known species, namely M.
blythii, as formerly hypothesized (Bogan et al. 1978).
According to the microsatellite data (Fig. 3) and to the
molecular tree of Fig. 4(A), M. blythii and M. myotis
appear to be closely related to each other (mean haplotypic divergence 2.1 ± 0.3%), but both differ considerably
from the North African form (mean 11.8% and 10.9%,
respectively). In a previous review of ecological, morphological and allozymic variation among M. myotis and
M. blythii, Arlettaz et al. (1997b) analysed basically the
same populations as those presented in this paper. These
authors showed that the morphologically intermediate
bats from North Africa, Corsica and Sardinia represented
a homogeneous, monotypic assemblage (Figs 1 and 2 in
Arlettaz et al. 1997b) and have the same allozyme alleles
as M. myotis at two discriminant loci (Arlettaz 1995; Arlettaz
et al. 1997b). Accordingly, a reanalysis of this data set using
35 allozyme loci and M. nattereri as an outgroup, suggests a
closer phylogenetic relationship between the African
taxon and M. myotis rather than with M. blythii (Fig. 4B).
However, these results depend on few variable allozyme
loci (11 out of 35 essayed), most of which are di- or triallelic
within Mouse-eared bats. For instance, the ADA and
GOT-1 loci are fixed for alternative alleles in continental
M. myotis and M. blythii (Arlettaz 1995; Arlettaz et al. 1997b),
and were used to identify bats from Morocco, Sardinia
and Corsica. Yet, the same alleles are found in other
species (e.g. in M. nattereri or M. daubentonii), suggesting
that retention of plesiomorphic alleles, homoplasy, parallelism or undetected mutations could be common in these
allozyme loci. According to the highest bootstrap values
obtained for the mtDNA tree (Fig. 4A) and to the larger
number of individuals and informative loci used for the
microsatellite data set, we are more confident in the
phylogenetic relationships obtained in the present
molecular data.
Thus, our most likely interpretation of the available
data is that the Moroccan and Iberian bats collected along
the transect represent distinct biological species, which
do not interbreed because occasional migrants across
the strait, if any, do not seem to transmit their genes to the
resident populations. Unlike M. myotis and M. blythii, which
are found in strict sympatry over most of their European
range, the North African form does not apparently coexist
with either species, at least in Morocco, Sardinia and
Corsica. Interestingly, the niche of the latter is very similar
to that of European continental M. myotis as they both
feed essentially on ground-dwelling arthropods (carabid
beetles, ground crickets, scorpions, etc.) (Arlettaz et al. 1997a;
Arlettaz 1999). Competitive exclusion would, therefore,
be more likely between M. myotis and bats of the North
African clade. In contrast, M. blythii exploits a quite
distinct niche throughout its range, it’s diet consisting
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
mostly of grass-dwelling prey, namely bush crickets.
Whether the current distribution of these three sibling
species is due entirely to historical processes of colonization, to competitive exclusion, or a combination of both
remains open to debate. If one applies the ‘standard’ rate
of 2% per million years for cytochrome b divergence in
mammals ( Johns & Avise 1998), the difference observed
between both European Mouse-eared bats and the North
African taxon (about 11%) would support a divergence
dating back to the Pliocene epoch. This very ancient
date does not mean that these taxa face each other at
Gibraltar since several million years. Rather, they probably diverged from a common ancestor somewhere else,
while the descendents colonized the present distribution
area more recently.
As the North African taxon is morphologically intermediate in size between M. myotis and M. blythii (Felten
et al. 1977; Benda & Horacek 1995), there has been considerable debate concerning its systematic position: it was either
treated as a smaller form of M. myotis (e.g. Ellerman &
Morrison-Scott 1966; Benda & Horacek 1995), or as a
larger form of M. blythii (Felten et al. 1977; Bogan et al.
1978; Corbet 1978). Based on samples from the Maghreb,
Sardinia, Corsica and Malta, Felten (in Felten et al. 1977)
described them under the new name M. blythii punicus
Felten 1977. Thus, we provisionally refer to the North
African, Corsican, and Sardinian ‘intermediate’ Mouseeared bats as Myotis cf punicus Felten 1977. Of course, the
definitive systematic position and distribution of this
interesting complex of species awaits further morphological and molecular comparisons especially with bats
from the eastern Mediterranean area and from elsewhere
in the Middle east.
In conclusion, even if the narrow Gibraltar Strait doesn’t
seem to represent an insuperable obstacle for Mouse-eared
bats (they can cross far broader stretches of open sea), it
may still prevent gene flow. For these bats, it materializes
the borders of distribution of sibling species, with two of
them (M. myotis and M. blythii) apparently not coexisting
with the third one (M. cf punicus). It remains to be shown
whether other species of bats actually exchange migrants
across Gibraltar, or whether they colonized southern Spain
and northern Morocco via other routes like the majority
of other wild mammals (Dobson 1998).
Acknowledgements
We thank Javier Juste, Juan Quetglas, Elena Migens, Association
Sportive de Speleologie de Agadir, Asociación Roncadell, Juan R.
Boyero, Pepe Ganfornina and many other bat workers for assistance in the field. Authorizations to take biopsy samples in Spain
were provided by the Consejeria de Medio Ambiente de la Junta
de Andalucia and Consellería de Agricultura y Medio Ambiente
de la Generalitat Valenciana. Nelly diMarco provided help with
DNA extraction. We thank also Eric Petit and François Balloux
for helpful comments on early versions of the manuscript. Financial support was provided by the following institutions: Swiss
National Funds for Scientific Research (grant # 3100 – 04 9245.96
to MR), DGICYT (project PB 90 – 0143 to CI), Fondation du 450e
anniversaire (University of Lausanne), Société Académique Suisse,
Société Suisse de Zoologie, Natural History Museum of Geneva,
and Institut Menorquí d’Estudis.
References
Arlettaz R (1995) Ecology of the Sibling Mouse-Eared Bats (Myotis
Myotis and Myotis Blythii): Zoogeography, Niche, Competition, and
Foraging. Horus Publishers, Martigny, Switzerland.
Arlettaz R (1996) Feeding behaviour and foraging strategy of
free-living Mouse-eared bats, Myotis myotis and Myotis blythii.
Animal Behaviour, 51, 1–11.
Arlettaz R (1999) Habitat selection as a major resource partitioning mechanism between the two sympatric sibling bat species
Myotis myotis and Myotis blythii. Journal of Animal Ecology, 68,
460 – 471.
Arlettaz R, Perrin N, Hausser J (1997a) Trophic resource partitioning and competition between the two sibling bat species
Myotis myotis and Myotis blythii. Journal of Animal Ecology, 66,
897– 911.
Arlettaz R, Ruedi M, Ibañes C, Palmeirim J, Hausser J (1997b) A
new perspective on the zoogeography of the sibling mouseeared bat species Myotis myotis and M. blythii: morphological,
genetical and ecological evidence. Journal of Zoology (London),
242, 45 – 62.
Avise J, Walker D (1999) Species realities and numbers in sexual
vertebrates: Perspectives from an asexually transmitted genome.
Proceedings of the National Academy of Sciences of the USA, 96,
992 – 995.
Avise JC (1994) Molecular markers, natural history and evolution.
Chapman & Hall, New York.
Barratt EM, Deaville R, Burland TM et al. (1997) DNA answers
the call of pipistrelle bat species. Nature, 387, 138 –139.
Benda P, Horacek I (1995) Biometrics of Myotis myotis and Myotis
blythi. Myotis, 32-33, 45 – 55.
Bogan MA, Setzer HW, Findley JS, Wilson DE (1978) Phenetics of
Myotis blythi in Morocco. In: Proceedings of the Fourth International
Bat Research Conference, Nairobi, pp. 217 – 230.
Boursot P, Jacquart T, Bonhomme F, Britton-Davidian J, Thaler L
(1985) Différenciation géographique du génôme mitochondrial
chez Mus spretus Lataste. Comptes Rendus de l’Académie des
Sciences de Paris, 301, 161–166.
Burban C, Petit RJ, Carcreff E, Jactel H (1999) Rangewide variation of the maritime pine bast scale Matsucoccus feytaudi Duc.
(Homoptera: Matsucoccidae) in relation to the genetic structure
of its horst. Molecular Ecology, 8, 1593 –1602.
Castella V, Ruedi M (2000) Characterisation of highly variable
microsatellite loci in the bat Myotis myotis (Chiroptera: Vespertilionidae). Molecular Ecology, 9, 1000 –1002.
Corbet GB (1978) The Mammals of the Palaearctic Region: a taxonomic review. Cornell University Press, London.
Dobson M (1998) Mammal distributions in the western Mediterranean: the role of human intervention. Mammal Review, 28,
77– 88.
Ellerman JR, Morrison-Scott TCS (1966) Checklist of Palaearctic and
Indian mammals, 1758 –1946. Alden Press, Oxford.
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular
variance inferred from metric distances among DNA haplotypes:
application to human mitochondrial DNA restriction data.
Genetics, 131, 479 – 491.
Felsenstein J (1993) PHYLIP (Phylogeny Inference Package). University of Washington, Seattle, WA.
Felten H, Spitzenberger F, Storch G (1977) Zur Kleinsäugerfauna
West-Anatoliens. Teil IIIa. Senckenbergiana Biologica, 58, 1– 44.
Filippucci MG (1992) Allozyme variation and divergence among
European, Middle Eastern, and North African species of the
genus Apodemus (Rodentia, Muridae). Israel Journal of Zoology,
38, 193 – 218.
Goudet J (1999a) FSTAT 2.9, a program to estimate and test gene
diversities and fixation indices (updated from Goudet 1995).
Lausanne, Switzerland. http://www.unil.ch/izea/softwares/
fstat.html.
Goudet J (1999b) PCA-GEN, Version 1.2. Lausanne, Switzerland.
http://www.unil.ch/izea/softwares/pcagen.html.
Hewitt GM (1999) Post-glacial re-colonization of European biota.
Biological Journal of the Linnean Society, 68, 87–112.
Horacek I (1985) Population ecology of Myotis myotis in Central
Bohemia (Mammalia: Chiroptera). Acta Universitas Carolinae —
Biologica, 8, 161– 267.
Johns GC, Avise JC (1998) A comparative summary of genetic
distances in the vertebrates from the mitochondrial Cytochrome
b gene. Molecular Biology and Evolution, 15, 1481–1490.
Kocher TD, Thomas WK, Meyer A et al. (1989) Dynamics of
mitochondrial DNA evolution in animals: amplification and
sequencing with conserved primers. Proceedings of the National
Academy of Sciences of the USA, 86, 6196 – 6200.
Lugon-Moulin N, Brünner H, Wyttenbach A, Hausser J, Goudet J
(1999) Hierarchical analyses of genetic differentiation in a hybrid
zone of Sorex araneus (Insectivora: Soricidae). Molecular Ecology,
8, 419 – 431.
Miller SA, Dykes DD, Polesky HF (1988) A simple salting procedure for extracting DNA from human nucleated cells. Nucleic
Acids Research, 16, 215.
Nei M (1987) Molecular Evolutionary Genetics. Columbia University
Press, New York, USA.
Paetkau D, Calvert W, Stirling I, Strobeck C (1995) Microsatellite
analysis of population structure in Canadian polar bears.
Molecular Ecology, 4, 347– 354.
Paetkau D, Waits L, Clarkson P, Craighead L, Strobeck C (1997)
An empirical evaluation of genetic distance statistics using
microsatellite data from bear (Ursidae) populations. Genetics,
147, 1943 –1957.
Palmeirim J (1979) First record of Myotis myotis on the Azores
Islands (Chiroptera: Vespertilionidae). Arquivos do Museu Bocage,
notas e suplementos, 46, 1– 2.
Paz Od, Fernandez R, Benzal J (1986) El anillamiento de Quiropteros en el centro de la peninsula iberica durante el periodo
1977-86. Boletin de la Estacion Central de Ecologia, 30, 113 –138.
Pedrola-Monfort J, Caujapé-Castells J (1994) Allozymic and
morphological relationships among Androcymbium gramineum,
A. europaeum, and A. psammophilum (Colchicaceae). Plant
Systematics and Evolution, 191, 111–126.
Petit E, Mayer F (1999) Male dispersal in the noctule bat (Nyctalus
noctula): where are the limits? Proceedings of the Royal Society of
London, Series B, 266, 1717–1722.
Petri B, Pääbo S, Von Haeseler A, Tautz D (1997) Paternity assessment and population subdivision in a natural population of
the Larger Mouse-eared bat Myotis myotis. Molecular Ecology, 6,
235 – 242.
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Raymond M, Rousset F (1995) An exact test for population
differentiation. Evolution, 49, 1280 –1283.
Rice WR (1989) Analysis tables of statistical tests. Evolution, 43,
223 – 225.
Ruedi M, Arlettaz R, Maddalena T (1990) Distinction morphologique
et biochimique de deux espèces jumelles de chauves-souris:
Myotis myotis (Bork.) et Myotis blythi (Tomes) (Mammalia;
Vespertilionidae). Mammalia, 54, 415 – 429.
Schneider S, Kueffer J-M, Roessli D, Excoffier L (1996) Arlequin: a
Software Package for Population Genetics. Genetics and Biometry
Lab, Dept. of Anthropology, University of Geneva, Geneva.
Sevilla P (1989) Quaternary fauna of bats in Spain: Paleoecologic
and biogeographic interest. In: European Bat Research 1987 (eds
Hanak V, Horacek I, Gaisler J), pp. 349 – 355. Charles University
Press, Praha, Tchechia.
Slatkin M (1994) Gene flow and population structure. In: Ecological Genetics (ed. Real LA), pp. 3 –17. Princeton Universty Press,
Princeton.
Smith MF, Patton JL (1993) The diversification of South American
murid rodents: evidence from mitochondrial DNA sequence
data for the akodontine tribe. Biological Journal of the Linnean
Society, 50, 149 –177.
Steininger FF, Rabeder G, Rögl F (1985) Land mammal distribution
in the Mediterranean Neogene: A consequence of geokinematic
and climatic events. In: Geological Evolution of the Mediterranean
Basin (eds Stanley DJ, Wezel F-C), pp. 559 – 571. Springer-Verlag,
New York.
Swofford DL (1996) PAUP*: Phylogenetic analysis using parsimony
(* And Other Methods), version 4.0. Sinauer Associates, Sunderland,
Masachusetts.
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772
Taberlet P, Fumagalli LW, ¸st-Saucy A-G, Cosson J-F (1998) Comparative phylogeography and postglacial colonization routes
in Europe. Molecular Ecology, 7, 453 – 464.
Tautz D, Renz M (1984) Simple sequences are ubiquitous repetitive components of eukaryotic genomes. Nucleic Acids Research,
12, 4127– 4138.
Valdés B (1991) Andalucia and the Rif. Floristic links and a common flora. Botanika Chronika, 10, 117–124.
Weir BS, Cockerham CC (1984) Estimating F-statistics for the
analysis of population structure. Evolution, 38, 1358 –1370.
Worthington Wilmer J, Barratt E (1996) A non-lethal method of
tissue sampling for genetic studies of Chiropterans. Bat Research
News, 37, 1– 3.
Wright S (1978) Variability within and among natural populations.
Evolution and the genetics of populations. University of Chicago
Press, Chicago.
VC’s PhD project focuses on phylogeography and population
genetics of the greater Mouse-eared bat. MR is conducting various
studies on the molecular systematics and phylogeny of bats. LE
is a human population geneticist with interest in inferring population history and demography from molecular markers in various
species. CI is interested in the phylogeography of bats from Iberia,
Maghreb and Macaronesia. RA works on community and population ecology, behavioural ecology, evolutionary ecology, and
conservation biology of bats and birds. JH’s prime research
interests are population genetics and the evolutionary history of
soricine shrews.
Appendix
List of material and locations used in the genetic analyses. The number of females (F) and males (M) genotyped at all microsatellite loci
and the number of the sequenced haplotypes (see Fig. 4A) appear in parentheses
Myotis myotis — SPAIN: Cueva del Agua, Cadiz (15 F and 5 m; 9 × ES1, 1 × ES2); Canillas de Aceituno, Malaga (19 F and 1 m; 2 × ES1);
Denia, Alicante (12 F and 8 m; 2 × ES4); Inca, Mallorca (2 F and 1 m; 1 × ES1, 2 × ES5). SWITZERLAND: Perreux, Neuchâtel (1 F and 1 m;
2 × CH1).
Myotis blythii — SPAIN: Tarifa, Cadiz (2 F; 2 × ES1b); Cueva del Agua, Cadiz (3 F; 1 × ES1b); Canillas, Malaga (1 F and 1 m); Denia,
Alicante (1 m; 1 × ES1b); Castellon (1 m and 1 F); Malgrat, Barcelona (1 F). PORTUGAL: Loucal, Querença (1 indet.). FRANCE:
Montredon, Narbonne (1 F and 1 m); Nice (2 F and 2 m). CZECH REPUBLIC: Bohemia (1 indet.; 1 × TS1). GREECE: Macedonia (1 indet.;
1 × GR1). KIRGHIZSTAN: Adzidar cave, Os (1 F and 1 m; 1 × KZ1, 1 × KZ2). Note that the latter four bats were not included in the
microsatellite analyses.
Myotis cf punicus — MOROCCO: Azrou, Meknes (16 F and 4 m; 1 × MA1, 1 × MA2); Boumahden, Agadir (5 F and 15 m; 1 × MA3,
1 × MA4). SPAIN: Mina del Agua, Ceuta (6 F and 13 m; 10 × MA1). FRANCE: Oleta, Corsica (2 F; 1 × CO1, 1 × CO2). ITALY: Mt Maiore,
Ulassai, Sardinia (1 F and 2 m; 2 × SA1).
Myotis nattereri — GREECE: Macedonia (not sexed).
© 2000 Blackwell Science Ltd, Molecular Ecology, 9, 1761–1772