Molecular Ecology (2006) 15, 2935–2947
doi: 10.1111/j.1365-294X.2006.02993.x
Apparent ‘sympatric’ speciation in ecologically similar
herbivorous beetles facilitated by multiple colonizations of
an island
Blackwell Publishing Ltd
B J A R T E H . J O R D A L , B R E N T C . E M E R S O N and G O D F R E Y M . H E W I T T
Centre for Ecology, Evolution and Conservation, School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK
Abstract
Coexistence of recently diverged and ecologically similar sister species in complete sympatry
represents a particularly compelling case for sympatric speciation. This study investigates
the possible sympatric origin of two coexisting bark beetle taxa that utilize the same host
plant on the island of La Palma in the Canary Islands. Aphanarthrum subglabrum and
Aphanarthrum glabrum ssp. nudum breed inside dead twigs of Euphorbia lamarckii plants
and are closely related to the allopatric A. glabrum ssp. glabrum in Tenerife, El Hierro and
La Gomera. We tested the various speciation hypotheses in a genealogical context, using
mitochondrial gene fragments from Cytochrome Oxidase I and 16S, and nuclear gene
α and Histone H3. Phylogenetic analyses of the
fragments from Enolase, Elongation Factor 1α
combined nuclear DNA data strongly supported a sister relationship between two sympatric
and reproductively isolated taxa in La Palma. However, network analyses of subdivided
nonrecombinant segments of the Enolase locus indicated a closer relationship between the
two allopatric A. glabrum subspecies, suggesting multiple colonizations of this island. A
bimodal distribution of mtDNA haplotypes in La Palma further documented the independent colonization of this island, with asymmetric introgression of mtDNA between two
lineages. Consequently, the sympatric origin of the La Palma species is concluded to have
involved allopatric phases before the parallel colonization of this island and subsequent
introgression at some loci. The clear genetic and morphological evidence for reproductive
isolation between these species suggests that the sympatric completion of divergence was
either due to initial genetic incompatibility, morphological character displacement in male
genitalia, or a combination of these factors.
Keywords: Aphanarthrum, Canary Islands, Euphorbia, mitochondrial introgression, Scolytinae,
sympatric speciation
Received 10 November 2005; revision received 3 April 2006; accepted 12 April 2006
Introduction
Speciation by means of allopatric divergence, either due to
genetic drift or selection in different environments, is the
most common mode of speciation in nearly all organisms
(Turelli et al. 2001). Divergence due to geographical
isolation is well documented (Coyne & Orr 2004) and
comparative analyses of a wide range of organisms have
shown that the most recently diverged sister species have
allopatric ranges (Barraclough et al. 1998; Barraclough &
Correspondence: Bjarte H. Jordal, Present address: Department of
Biology, University of Bergen, Allégt. 41, N-5007 Bergen, Norway.
Fax: +47 55589667; E-mail: [email protected]
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
Nee 2001). The few well-documented cases of sympatric
speciation all involve some sort of resource division (e.g.
Barluenga et al. 2006), in particular host switching
(Berlocher & Feder 2002), or divergent phenologies (e.g.
Savolainen et al. 2006) or habitat-associated mating
behaviour (Via 2001; Coyne & Orr 2004). These traits are
necessary to reduce gene flow between diverging
populations, with the ultimate elimination of intermediate
‘generalist’ genotypes (Fry 2003). Sympatric divergence
solely through assortative mating may be possible in theory
(Kondrashov et al. 1998; Kondrashov & Kondrashov 1999),
but genotype-dependent ‘homogameous’ mating has not
yet been demonstrated empirically (Via 2001). Hence,
divergence of sympatric sister lineages that exploit the
2936 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
same fundamental niche seems an unrealistic scenario for
successful speciation (Arnegard & Kondrashov 2004).
Observations of closely related sympatric taxa in the
same habitat nevertheless do occur from time to time in nature
(e.g. Schluter 1994; Via 2001: Verheyen et al. 2003), although
these may not necessarily represent sister species
(Machado et al. 2002; Coyne & Orr 2004). If sympatric
species share the same ancestor, their common origin is
either explained by a historical allopatric phase (e.g.
Bernatchez & Dodson 1990) or cryptic ecological differences during divergence (e.g. Thompson et al. 1997;
Barluenga & Meyer 2004; Rolan-Alvarez et al . 2004;
Barluenga et al. 2006). To be able to test between these
hypotheses with genealogical data, several criteria must be
satisfied to allow any robust inference on speciation mode
(Berlocher 1998; Coyne & Orr 2004). First, the phylogeny of
the group must be complete, including all closely related
taxa that could potentially bridge artificially demarcated
populations (e.g. Funk 1999; Funk & Omland 2003).
Sampling outside the area of sympatry would therefore
always be critical to obtain confidence in sister lineage
assessment. Second, the geographical range of all putatively
close relatives must be well known and sampling of these
areas must be representative. Undersampling could potentially lead to erroneously inferred sympatric distributions
in species that are actually more widely distributed and
sympatric only for a small part of their respective distributions. Lastly, the geographical variation in morphological
and ecological traits must be well characterized and based
on repeated sampling from different kinds of related
habitats. An allopatric origin for two species then may be
rejected if they are fully sympatric and reproductively
isolated sister species, with historical allopatric phases
unlikely in a biogeographical context (Coyne & Orr 2004).
What is ‘unlikely’ in this context, however, is not readily agreed
on, but small oceanic islands are among the most promising
settings when microallopatry within the island is similarly
unlikely (e.g. Barluenga et al. 2006; Savolainen et al. 2006).
Here we investigate the possible sympatric origin of two
bark beetle taxa associated with the same host plant on the
island of La Palma in the Canary archipelago (Fig. 1).
Aphanarthrum subglabrum and Aphanarthrum glabrum
nudum are endemic to La Palma and are closely related to
the allopatric A. glabrum glabrum , which has a wider
distribution on Tenerife, La Gomera and El Hierro
(Israelson 1972; Jordal & Hewitt 2004). The three taxa are
distinguished mainly by differences in male genitalia, but
most specimens of A. subglabrum are further recognized by
a slightly darker colour and somewhat longer pubescence
on the elytra (Fig. 2). The A. glabrum species complex is part
of a much larger radiation of crypturgine bark beetles,
where each species of Aphanarthrum and Coleobothrus is
restricted to a single host plant of Euphorbia spurges (Jordal
2006). All three members of the A. glabrum complex feed
and reproduce exclusively in dead wood of Euphorbia
lamarckii, where each individual completes their entire life
cycle (Jordal 2006). Aphanarthrum subglabrum and A. g. nudum
have completely overlapping distributions in La Palma,
and both species are often found together in intermingled
egg tunnel systems in the same plant. These taxa, and the
other Aphanarthrum species outside the complex, have been
collected repeatedly since Wollaston’s (1862) pioneer work
on the group (Schedl et al. 1959; Israelson 1972; Jordal
2006), which gives high confidence in distributional and
host-plant data.
The A. glabrum species complex has several additional
features that make it a very promising candidate for testing
hypotheses on speciation in a genealogical context. The
phylogeny of Macaronesian Aphanarthrum is complete and
is based on dense sampling from all islands and the
African mainland (Jordal & Hewitt 2004). Furthermore, the
A. glabrum complex is distinct from all other Aphanarthrum
lineages, and thus, excludes potential close relations to
other taxa outside the species complex. At the same time,
the evolution within the complex is recent, based on the
observed low level of genetic variation within and between
each morphologically defined taxon (Jordal & Hewitt
2004). The recent divergence within this group thus reduces
the likelihood of range changes through a brief span of
evolutionary history and therefore increases the information
potential from genealogical data.
Within the framework of ecologically, biogeographically,
and phylogenetically well-characterized populations, we
can ask several questions regarding the distinctness of
island populations. Of primary importance to sympatric
speciation is the question of whether the two morphologically defined taxa in La Palma are reproductively isolated,
thus conforming to the biological species concept. A second
question is whether the morphologically defined subspecies
A. glabrum nudum in La Palma is sister taxon or paraphyletic
with the allopatric A. glabrum glabrum in El Hierro, La
Gomera and Tenerife (Fig. 3a vs. 3b), both of which imply
parallel colonization of La Palma. The alternative implies
monophyly of the two endemic taxa in La Palma (Fig. 3c),
a relationship not at odds with a history of sympatric
speciation. However, particular care must be invested to
examine the possibility for historical introgression between
sympatric taxa that may have led to their current sister
relationship (Coyne & Orr 2004).
Genealogical data were assembled from partial DNA
sequences of the mitochondrial genes cytochrome oxidase
I (COI) and the large ribosomal subunit (16S), and from
three nuclear encoding genes: Elongation Factor 1-alpha
(EF-1α), Enolase and Histone H3. Studies on phylogeography
and speciation in animals have so far largely relied on
mitochondrial DNA, partly because of the higher rate of
nucleotide substitution than in nuclear DNA, but also
because of the smaller genetic population size of the
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2937
Fig. 1 Collecting sites for the Aphanarthrum glabrum complex. Black dots signify sample sites of A. glabrum in La Palma (ssp. nudum),
Tenerife, La Gomera and El Hierro (ssp. glabrum), and the white dot signifies A. subglabrum. The black and white graded dots indicate
sample sites where both taxa were taken from the same plant in La Palma. Numbers associated with each sample site refer to the individuals
analysed.
maternally inherited mitochondrial genome. Together with
the lack of recombination, the smaller genetic population
size results in shorter coalescence times in mitochondrial
loci, which more readily reveal weak phylogeographical
signals (Avise 2000). On the downside, mitochondrial
DNA is susceptible to selective sweeps and asymmetrical
introgression that can severely mislead the interpretation
of the evolutionary history for a given focal group (e.g.
Ballard 2000; Shaw 2002; Machado & Hey 2003; Chan &
Levin 2005). Analyses of multiple loci from both the
nuclear and mitochondrial genomes are therefore required
to confidently demarcate species boundaries and identify
deeper phylogeographical break points.
Materials and methods
Beetles of the Aphanarthrum glabrum complex were sampled
from dead branches of Euphorbia lamarckii shrubs throughout
their known distributional range (Fig. 1). Genitalia of the
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
males were examined on microscope slides to assess
consistency in genitalic differences between the three
described taxa (see Israelson 1972). When available, at least
two males of each taxon were analysed from each sampling
locality.
PCR amplification, sequencing, and haplotype assessment
Extraction of genomic DNA, polymerase chain reaction
(PCR) amplification and sequencing followed the protocols
described in Jordal & Hewitt (2004). The Histone H3 gene
fragment was amplified with primers HexAF and HexAR
(Svenson & Whiting 2004) using the mtDNA PCR conditions.
Direct sequencing of Enolase and EF-1α resulted in several
putative heterozygous sites revealed by two overlaying
peaks in the sequencing chromatograms (see, e.g. Brumfield
et al . 2003). Heterozygosity was low in EF-1 α and
haplotypes were assigned by direct comparison with
homozygous sequences. Some heterozygous sequences of
2938 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
sites in a complete alignment of all Enolase sequences.
Based on known haplotypes from homozygous PCR
products, and from the cloned sequences, we statistically
assigned the unknown haplotypes in a Bayesian framework,
using the computer program phase version 2.0, which takes
recombination into account (Stephens et al. 2001). Only five
haplotypes could not be assigned to a known haplotype
in the population, but only one of these was more than
one substitution different from a known haplotype and had
no effect on the reconstruction of genealogies, and negligible
effect on the estimation of population parameters. Nuclear
sequences were also analysed phylogenetically in their
polymorphic states for each individual (direct sequences),
to compare with analyses of the probabilistically determined
haplotypes. New sequences added to those used in
Jordal & Hewitt (2004) can be downloaded from GenBank
under the following accession numbers: DQ460224–DQ460428.
Data analysis
Fig. 2 Dorsal view of adults and male genitalia of Aphanarthrum
subglabrum (a, b) and Aphanarthrum glabrum (c–e). Line drawings
illustrate each of the A. glabrum subspecies (d, e) as defined by
Israelson (1972).
Enolase could not have their gametic phases identified
directly and two different approaches were used to identify
the two alleles. First, one heterozygous Enolase PCR product
from each of the three taxa was cloned in bacterial plasmids,
with five clones sequenced for each PCR product. Cloned
sequences were thereafter compared to the directly
sequenced PCR products, to identify true heterozygous
Alignments were easily obtained by eye for all data partitions,
with two unambiguous indels necessary only for the 16S
data. The dnasp software (Rozas & Rozas 1999) was used
to estimate the number of segregating sites and nucleotide
diversity (k, the average number of pairwise substitutions;
π, the average number of pairwise substitutions per site).
The nuclear gene fragments were also examined for the
minimum number of recombination events and had their
putatively recombined sites localized. The minimum
number of recombination events, Rm, was based on the fourgamete test by Hudson & Kaplan (1985) as implemented in
the dnasp software.
Recombination within the nuclear genes can potentially
obscure the relationships among interbreeding populations
because different portions of the haplotype may have
different genealogies (Hare 2001). However, for the purpose
of distinguishing species and deeper phylogeographical
splits between reproductively isolated populations, gene
trees from recombining nuclear loci are still very useful.
The rationale is that fixed nucleotide substitutions are
allowed to accumulate in recombining genes only when
gene flow (and thus, recombination) is reduced between
populations. Recombining loci can also provide further
Fig. 3 Different hypothetical relationships between the three closely related taxa of the Aphanarthrum glabrum complex. (a) The two
subspecies of A. glabrum are sister lineages, which implies two colonizations of La Palma, unless the entire complex originated on that island;
(b) same as (a), but with current gene flow between La Palma and some or all of the other Canary Islands; (c) sister relationship between
the two endemic taxa in La Palma, which may imply a sympatric origin of the two taxa.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2939
clues about more recent genealogies when these loci are
divided into segments of putatively nonrecombining DNA
before analyses (Hare 2001). The recombinant Enolase was
therefore divided into five unambiguously nonrecombined segments for further network analyses.
Phylogenies were reconstructed in a Bayesian framework, using mrbayes version 3.1.1 (Huelsenbeck &
Ronquist 2001). Evolutionary models were selected with
the software modeltest version 3.7 (Posada & Crandall
1998) in conjunction with paup* (Swofford 2002). All best
models were variants of the GTR + I + Γ model and hence
implemented in mrbayes. The two physically linked
mitochondrial loci were combined in all phylogenetic
analyses. Trees were rooted with two species (Aphanarthrum wollastoni and A. aeoni) from the most likely sister
lineage of the Aphanarthrum glabrum complex (see fig. 6 in
Jordal & Hewitt 2004). Statistical parsimony networks
(Templeton et al. 1992) were reconstructed with the software package tcs (Clement et al. 2000) to assess the level of
multiple haplotype connections that reflect uncertainties
generated by recombination or parallel mutations (Posada
& Crandall 2001).
Results
DNA polymorphism
Sequence variation was lower within all morphologically
defined taxa than across taxon boundaries, with the highest
maximum divergence across all three taxa measured in
COI (4.6%). The maximum sequence divergence for the
combined mtDNA loci was 3.2%. Overall nucleotide
diversity (π) was low for all five loci, and ranged between
0.48% in Histone H3, to 1.48% in mtDNA (Table 1). The
greatest nucleotide variation, and the highest number of
haplotypes, was found in the most widespread taxon,
Aphanarthrum glabrum glabrum, for each locus except
Histone H3. The two sympatric taxa in La Palma,
Aphanarthrum subglabrum and A. glabrum nudum, had quite
similar levels of variation, except for mtDNA where
the first taxon had a distinct bimodal distribution of
haplotypes (see Fig. 4a).
Recombination in nuclear loci
Recombination was not detected for Histone H3, and was
only detected within one population in La Gomera for
EF-1α. Hence, the latter gene fragment was also treated as
nonrecombining in the genealogical analyses. A minimum
of seven recombination events was estimated from the
complete Enolase data set, six of these occurred in A. g. glabrum,
six occurred in A. g. nudum, and three events occurred in
A. subglabrum (Table 1). Putative recombinant sites in the
689-bp-long fragment were found between sites 9 and 18,
150 and 265, 282 and 300, 300 and 315, 327 and 384, 531 and
558, and 558 and 570. Five nonrecombining segments were
defined based on the localization of recombinant sites,
excluding very short segments of less than 40 bp. The five
segments included positions 10–150, 151–299, 301–383,
384–557 and 559–689 of the Enolase gene fragment.
Table 1 Nucleotide polymorphism for mitochondrial DNA (COI + 16S) and three nuclear (EF-1α, Enolase, Histone H3) gene fragments in
the Aphanarthrum glabrum complex
Locus
Taxon
No.
of individuals
Total sites
N haplotypes
Segregating
sites
k
π
Rm
mtDNA
All populations
glabrum
nudum
subglabrum
All populations
glabrum
nudum
subglabrum
All populations
glabrum
nudum
subglabrum
All populations
glabrum
nudum
subglabrum
56
27
15
14
56
27
15
14
56
27
15
14
17
5
5
7
1117
1117
1117
1117
689
689
689
689
904
904
904
904
328
328
328
328
38
20
7
12
58
29
18
11
17
13
2
2
3
1
2
1
60
44
8
37
35
17
9
8
23
15
1
1
3
0
1
0
15.97
15.03
2.31
14.84
8.01
3.29
2.96
2.77
4.59
1.95
0.11
0.40
1.56
0
0.40
0
0.0148
0.0135
0.0021
0.0137
0.0116
0.0048
0.0043
0.0040
0.0051
0.0022
0.0001
0.0004
0.0048
0
0.0012
0
—
—
—
—
7
6
6
3
1
1
0
0
0
0
0
0
Enolase
EF-1α
Histone H3
Abbreviations: k, average number of pairwise substitutions; π, average number of pairwise substitutions per site; Rm, minimum number of
recombination events per locus.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
2940 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
Fig. 4 Phylogenetic trees reconstructed
from Bayesian analyses of (a) mtDNA data
(COI + 16S) and (b) combined Enolase, EF1α and Histone H3 data (direct sequences).
Posterior probabilities above 95% are shown
on internodes. The trees were rooted with
Aphanarthrum wollastoni and A. aeoni.
Numbers refer to individual beetles and
their sample location (see Fig. 1). Grey
boxes highlight A. subglabrum and stars
indicate subsampled individuals for
Histone H3 sequences (cf. Fig. 5, inset).
Genealogical patterns
The mtDNA data were highly structured geographically,
with each island genetically cohesive except for four
paraphyletic A. subglabrum haplotypes (Fig. 4a). These four
haplotypes grouped with strong support to the El Hierro
and La Gomera haplotypes of A. g. glabrum. Haplotypes
of A. g. glabrum from Tenerife grouped closer to the
remaining haplotypes of A. subglabrum and those of A. g.
nudum. The haplotypes of A. g. nudum were furthermore
nested within the more diverse A. subglabrum, and one
haplotype was shared between five individuals of A. g.
nudum and one A. subglabrum (all from different localities).
Several other haplotypes of the two taxa were 1 or 2 bp
different from each other.
Statistical parsimony analysis (95% connection limit) of
the mitochondrial sequence data resulted in two unconnected networks (not shown) divided at the major split
between the Tenerife–La Palma clade, and the La Gomera–
El Hierro (–La Palma) clade in the Bayesian tree. The four
deviating A. subglabrum haplotypes were connected via
three mutational steps to a haplotype from La Gomera, and
the El Hierro haplotypes were connected by four mutational
steps to a haplotype from La Gomera. The grouping of
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2941
Tenerife haplotypes was connected to a haplotype of A.
subglabrum by five mutational steps.
All three nuclear loci revealed a very dissimilar pattern
to the one observed for the mitochondrial loci (Fig. 5). Very
little or no geographic structure occurred among the three
island populations of A. g. glabrum, with six haplotypes of
Enolase (Fig. 5a), two haplotypes of EF-1α (Fig. 5b), and all
Histone H3 haplotypes (Fig. 5, inset) shared between at
least two of the islands of Tenerife, La Gomera or El Hierro.
Contrary to the inter-island panmictic structure in A. g. glabrum,
each of the two taxa in La Palma were strongly supported
as monophyletic groups by the Enolase data (Fig. 5a),
distinguished by six fixed nucleotide substitutions (Table 2).
They were further distinguished from each other by one
fixed substitution in the EF-1α data (Fig. 5b), while all
Histone H3 haplotypes from A. subglabrum were shared
with A. g. nudum (Fig. 5, inset). Two of the nuclear loci
strongly supported the sister relationship between the two
La Palma taxa, by six synapomorphic character changes in
the EF-1α data and two in the Histone H3 data. Enolase
resulted in a less-resolved topology, with monophyly for
each of the La Palma taxa, but ambiguity surrounding the
relationship between these two clades.
Statistical parsimony analyses of each of the nuclear data
sets revealed networks mostly congruent with the main
features of the Bayesian analyses. However, the recombined
Enolase haplotypes produced an irresolvable network of
multiple connections within each taxon. Relationships between
taxa were less ambiguous though, with A. g. nudum more
closely connected to A. g. glabrum than to A. subglabrum.
The first two taxa were connected by three mutational
steps, while A. subglabrum was connected to A. g. glabrum
by seven mutational steps. A subsequent analysis of
Enolase was undertaken that divided the locus into five
putatively nonrecombined segments for network analyses,
with lengths ranging from 83 to 173 bp (Fig. 6). Only one
of the segments (Fig. 6b) revealed shared haplotypes
between A. subglabrum and A. g. nudum. In three other
segments (Fig. 6a, d, e), a large proportion of haplotype
segments in A. g. nudum were shared with the allopatric
A. g. glabrum. In the remaining segment (Fig. 6c), A. g. nudum
was connected to A. g. glabrum via a single mutational step,
while its sympatric congener A. subglabrum was connected
via three mutational steps through A. g. glabrum. Based on
the analyses of nonrecombinant Enolase segments, it seems
likely that A. subglabrum has a longer history of isolation
from the other two taxa for this locus.
Morphological differences
All males of A. subglabrum (n = 10) and A. g. glabrum (n = 14)
had spines on the end plates (a pair of apical sclerites, see
Fig. 2) of the male genitalia, but specimens of the first taxon
had consistently fewer spines away from the genital apex
(as depicted in Fig. 2b). There were no apparent differences
in the genitalia between A. g. glabrum from Tenerife, La
Gomera and El Hierro. All males of A. g. nudum examined
(n = 10) were, as the species epithet implies, free of spines
on the end plates. The ventral sclerite of this taxon had in
addition a pair of antero-lateral spines not seen in the other
two taxa (Fig. 2d).
Discussion
Table 2 Polymorphism and divergence between taxa of the
Aphanarthrum glabrum complex
Locus
mtDNA
Taxon
k
comparison k (all) (between taxa) SX1 SX2 SS SF
sub-nud
sub-glab
nud-glab
EF-1α
sub-nud
sub-glab
nud-glab
Enolase
sub-nud
sub-glab
nud-glab
Histone H3 sub-nud
sub-glab
nud-glab
9.95
16.42
16.21
0.80
4.19
4.75
8.20
6.62
6.13
0.20
1.60
1.53
11.66
18.10
20.83
1.31
7.21
8.27
13.39
10.81
9.43
0.20
3.00
2.80
33
14
4
1
1
1
8
6
7
0
0
1
4
21
40
1
15
15
9
15
15
1
0
0
4
23
4
0
0
0
0
2
2
0
0
0
Abbreviations: k, average number of pairwise susbstitutions;
SX1, polymorphism exclusive to the first mentioned species;
SX2, polymorphism exclusive to the second mentioned species;
SS, polymorphism common to both species;
SF, fixed differences between species.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
0
0
4
1
6
7
6
5
2
0
3
2
Sympatric sister species
The majority of data collected in this study supported
several of the criteria necessary to infer sympatric speciation.
First of all, each of the two taxa in La Palma was clearly
reproductively isolated from each other and from A. g.
glabrum on the neighbouring islands. Genetic differences
between the two La Palma species were found at both
the EF-1α and the Enolase loci, with as many as six fixed
substitutions at the recombinant Enolase locus. Because
reshuffling of recombined alleles would immediately
eliminate such fixed differences during periods of gene
flow between two populations (Ortiz-Barrientos et al. 2002),
it would be impossible to accumulate this many differences
unless there has been significant barriers to gene flow over
a major part of the genome. The observation of completely
sorted lineages at two of the three nuclear loci examined is
by itself a convincing argument for complete speciation
insofar as recent speciation events rarely result in high
proportions of sorted loci (Edwards & Beerli 2000; Wu
2001; Machado & Hey 2003). Reproductive isolation was
2942 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2943
Fig. 6 Networks resulting from the statistical parsimony analyses of nonrecombined segments of the Enolase locus. Each
diagram represents a network from the
analysis of sites 10–150 (a), 151–299 (b),
301–383 (c), 384–557 (d), and 559–689 (e).
furthermore reflected in the clear differences observed in
male genitalia (presence vs. absence of spines). Differences
in genitalia are usually much more subtle between closely
related scolytine beetles and this kind of exaggerated change
in reproductive characters is only known from scolytine
species with parapatric distributions (Jordal 1998). Thus it
is tempting to infer some kind of character displacement
process responsible for the abnormal character differences
in male genitalia.
Second, the two taxa in La Palma appeared to be sister
groups in the separate and combined analyses of several
nuclear DNA loci. If they truly share a common ancestor,
the implication could be a real case for sympatric speciation,
especially if allopatric divergence within the island
would be similarly unlikely (Berlocher 1998; Coyne & Orr
2004; Barluenga et al. 2006). However, the sympatric divergence hypothesis seems very unlikely in the context of
there being no recognizable ecological differences
between A. subglabrum and A. g. nudum — a necessary
condition for divergence in sympatry (e.g. Fry 2003). The
close relationship between these species therefore calls for
a more detailed examination of the genetic data in
particular, but also the extent of geographical overlap at
the time of divergence.
Contemporary sympatry of sister species rarely provides
conclusive evidence for sympatric speciation because
historical periods of range expansion and contraction that
may have included allopatric phases are difficult to
disprove (Losos & Glor 2003; but see Barluenga et al. 2006
and Savolainen et al. 2006 for the opposite view). Range
constriction has recently been demonstrated for flightless
weevils on La Palma, revealing genetic signatures of
geographic isolation for species now continuously distributed throughout the island (Emerson et al. 2006). Flightcapable beetles associated with Euphorbia seem much more
mobile however, lacking any evidence for a positive
correlation between distance and genetic divergence on La
Palma (Mantel test, e.g. mtDNA: A. g. nudum, r = −0.17 NS;
A. subglabrum, r = 0.10 NS). An important factor in reducing
local divergence in these beetles is the high abundance and
ubiquitous presence of their common host plant, Euphorbia
lamarckii (see, e.g. Jordal 2006). This plant is part of the
early successional plant community on volcanic soil and is
abundant all over the Canary Islands below 1000 m altitude.
Fig. 5 Phylogenetic trees reconstructed from Bayesian analyses of (a) Enolase and (b) EF-1 haplotype data. Posterior probabilities above 95%
are shown on the internodes. One unique amino acid substitution was observed in the Enolase data (Proline to Alanine). The trees were
rooted with Aphanarthrum wollastoni and A. aeoni. Inset: statistical parsimony network of subsampled Histone H3 sequences. Colour codes
signify A. subglabrum (n = 5, black), A. g. nudum (n = 5, grey) and A. g. glabrum (n = 7, white). Squared box indicate the ancestral haplotype
inferred by the tcs software, and the size of circles and boxes indicate the relative number of individuals with identical haplotypes. See
Fig. 4a for sample identities.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
2944 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
Consequently, it seems unlikely that there would ever be
sufficient opportunities for host-plant patchiness within
islands, and hence populations of flight-capable beetles may
move across an island within only a few generations.
Intra-island allopatric divergence would therefore be quite
unlikely in Aphanarthrum beetles, meaning that an eventual
sister relationship between sympatric species would be
consistent with sympatric speciation.
Multiple colonizations of La Palma?
A more likely resolution of the allopatry–sympatry controversy
can be obtained by a more fine-scaled analysis of the nuclear
DNA data. These data may at first look convincingly
congruent in their support for a monophyletic sister
relationship between the two species in La Palma.
However, the mtDNA data did not support their monophyly
and the recombinant Enolase data demonstrated ambiguity
around this relationship. It has been argued that analyses
of recombined DNA sequences could obscure the correct
genealogical history of individuals that carry gene
fragments of different genealogical descent (Hare 2001).
Thus, when we analysed the putatively nonrecombined
stretches of DNA from the Enolase locus, we did indeed
find a closer relationship between the two allopatric
subspecies of A. glabrum. A more recent history of
recombination between these lineages thus implies that La
Palma was colonized twice and not once, as the sympatric
monophyly would imply.
The multiple-colonization scenario was also supported
by the mtDNA data where A. subglabrum showed a
markedly bimodal distribution of haplotypes. One set of
haplotypes was closely related to A. g. nudum and the
Tenerife population of A. g. glabrum, and the other set of
haplotypes was closely related to A. g. glabrum in El Hierro
and La Gomera (cf. Fig. 4a). Consequently, the mitochondrial
gene pool for A. subglabrum is composed of two geographical
sources that may have contributed to some of the interactions
between genomes that are necessary to keep marginally
different incipient species apart (e.g. Fry 2003; but see also
Bernatchez & Dodson 1990 and Feder et al. 2003). This is
particularly important when there are no evident ecological
differences as is the case between the two lineages in La
Palma. It must be assumed, however, that the final divergence that led to the currently recognizable species on this
island occurred in sympatry. This is evident from the low
genetic divergence between the two main mitochondrial
lineages (current average of 2.17% COI divergence), which
suggests that the origin of these species is very recent.
Furthermore, the longer history of recombination between
the two allopatric A. glabrum taxa at the Enolase locus, and
the bimodal distribution of mitochondrial haplotypes in La
Palma, may suggest that sufficient genetic differences had
evolved at the time of secondary contact.
Asymmetric introgression and the mosaic genomes of
recently diverged species
Genealogical incongruence among different loci is a familiar
problem in studies of recent speciation events (e.g. Funk &
Omland 2003). Reconstructing the correct genealogical
pathway is particularly challenging during and just after the
incipient stage of speciation, and low levels of gene flow
between recently diverged species can further contribute
to discordant gene genealogies long after their origin.
Consequently are diverging genomes expected to be
mosaics with respect to genealogies from different gene
regions for quite some time (Wu 2001). Taken together with
the expected natural variation in coalescence times between
neutrally evolving loci of similar genetic population size
(Edwards & Beerli 2000), the modest genealogical incongruence observed here for recently diverged Aphanarthrum
should not at all be unexpected.
We do nevertheless anticipate that loci with the smallest
genetic population size, such as those in the mitochondrial
genome, should be among the first to sort into monophyletic
groups. Compared to an average nuclear allele, coalescence
time for mitochondrial alleles is much shorter, on average
only one-fourth the time needed to sort a nuclear allele
(Moore 1995; Avise 2004). Hence, the significant geographical
structure between all four islands for the mitochondrial
haplotypes fits well with coalescent theory. Nuclear
haplotypes, on the other hand, exhibited no geographical
structure across the three islands of El Hierro, Tenerife and
La Gomera in A. g. glabrum. Although the lack of structure
in nuclear DNA could potentially be explained by the
larger expected genetic population size for these loci, the
presence of shared nuclear haplotypes across islands in
A. g. glabrum cannot be reconciled by arguments of incomplete lineage sorting alone. This is particularly so for the
most variable locus, Enolase, where the shared haplotypes
between islands is indicative of considerable nuclear DNA
gene flow among these island populations. The observation of nuclear DNA gene flow in the absence of maternally contributed mitochondrial gene flow suggests that
dispersal between these islands is strongly male-biased
(Avise 2004). Similar genealogical patterns of strongly
structured mitochondrial alleles in an otherwise panmictic
nuclear gene pool are known from animals with observed
male-biased dispersal (reviewed in Prugnolle & de Meeus
2002). Most examples include studies on birds and mammals,
but see Doums et al. (2002) for an illustrative example
on male dispersed, queenless ants.
In the context of geographically well-structured
mitochondrial genealogies for Tenerife, El Hierro and La
Gomera, it seems difficult to reconcile the extensive mitochondrial paraphyly in La Palma with a small mtDNA
population size. Moreover, loci with larger expected genetic
population sizes, such as Enolase and EF-1α, were nicely
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2945
sorted into reciprocally monophyletic groups in La Palma,
consistent with morphological differences in the male
genitalia. Thus, the paraphyletic pattern of mtDNA in La
Palma is most likely not due to random lineage sorting of
alleles. One rather unlikely possibility for obtaining
misleading mitochondrial data could be unwanted amplification of nuclear insertions of mitochondrial pseudogenes (Bensasson et al. 2001). However, none of the
sequence traces showed clear background peaks, which
could be indicative of multiple copies. Furthermore, there
is no good reason for why pseudogenes should exhibit
strong geographical structure as observed in the mtDNA
data. Also, the fully congruent results from separate
analyses of the 16S and COI data imply, if they are pseudogenes, that the entire stretch of mitochondrial DNA
between the 16S and COI primer pairs (> 5332 bp) must
have been inserted into the nuclear genome in one batch,
which is not very likely.
Having dismissed each of the pseudogene and ancestral
polymorphism hypotheses in explaining the mitochondrial
paraphyly in La Palma, a third hypothesis of introgression
provides a more compelling mechanism. Given that both
A. subglabrum and A. g. nudum utilize the same host plant
and are found frequently together in the same tunnel
systems ( Jordal 2006), there are basically no ecological
barriers to hybridization between the two species. Thus,
the observation of one shared haplotype between one
subglabrum and five nudum individuals, and that several
more are only 1 or 2 bp different, suggests that mitochondrial
introgression has occurred between the two taxa in La
Palma. In cases where taxa are genetically well separated
by the majority of loci — as in this case for some of the
nuclear gene fragments — shared haplotypes are strongly
indicative of recent introgression through hybridization
(e.g. Alves et al. 2003; Besansky et al. 2003). Introgression is
further supported by the observed nested structure of A. g.
nudum haplotypes within the more variable A. subglabrum,
which is a typical signature for recently introgressed taxa
(Funk & Omland 2003). That only one haplotype was shared,
and many more were a few base pairs different, suggests
that introgression has occurred more frequently in the past
and is most likely not an ongoing process.
Cytoplasmic capture of organelle DNA is a well-known
phenomenon from plants and animals alike, and several
possible processes have been used to explain this kind of
asymmetric gene flow (Avise 2004). One explanation for
mitochondrial introgression suggests that frequent hybridization provides a channel for mtDNA flow across species
boundaries, provided that strong selection occurs against
hybrids to maintain the parental genotype (Shaw 2002).
However, we may expect a significant proportion of
morphological intermediates under this scenario (Morrow
et al. 2000), but these have not been observed. Rare events
of hybridization in the past thus seems a more likely
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
explanation, under which mitochondrial haplotypes of one
species may have been displaced by the introgressing
species’ haplotypes, due either to selection for new optimal
combinations of mitochondrial DNA in a new genetic
background, or by drift. That mitochondrial DNA is rather
frequently introgressed between closely related animal
species (e.g. Wang et al. 1997; Kliman et al. 2000; Morrow
et al. 2000; Ting et al. 2000; Machado et al. 2002; Shaw 2002;
Alves et al. 2003; Besansky et al. 2003; Mallet 2005) is
perhaps not very surprising given the presumed inactive
role of mitochondrial genes in reproductive isolation
(Barton & Jones 1983; Machado & Hey 2003; Chan & Levin
2005).
Acknowledging that hybridization has taken place between
the two species in La Palma, we may further speculate as
to whether introgression has also occurred at some of the
nuclear loci. All except one haplotype of the Histone H3
locus were shared between the two species, and only one
fixed substitution was found between the two at the EF-1α
locus. Each of these loci has very low substitution rates
throughout the eukaryotes and it may well be that the little
divergence between A. subglabrum and A. g. nudum reflects
the slow evolutionary rates in these genes. On the other
hand, these taxa were clearly distinct from the allopatric
A. g. glabrum at these same loci, which could instead imply
that the similarity between the two taxa in La Palma is due
to historical introgression at these loci. Under this scenario,
introgressed alleles must have completely replaced the
other incipient species’ alleles shortly after the introgression event. Partial or complete reproductive isolation
nevertheless developed soon after the initial stage of
hybridization, allowing EF-1α and Enolase to accumulate
fixed differences, and polymorphism in the Histone H3
locus.
Given the intricate nature of the genetic data that we
have assembled here, the most crucial question related to
the monophyly of sympatric species remains unanswered.
We do know, however, that each species in the A. glabrum
species complex is reproductively isolated despite a
history involving hybridization. On the other hand, we do
not know how influential these processes of hybridization
have been on genealogies from various sites in the genome.
If the extent of introgression has indeed been over a large
part of the genome, it will certainly reduce our ability to
resolve the evolutionary origin of these species. Only a
greater amount of additional genomic data can bring us
any closer to deciding whether the genetic similarity
between sympatric species is due to introgression or to
evolutionary descent.
Conclusion
Island ecosystems have been a recent focal point for
understanding the processes important for the generation
2946 B . H . J O R D A L , B . C . E M E R S O N and G . M . H E W I T T
of species diversity, with species diversity itself being
recognized as an important driver of diversification
(Cadena et al. 2005; Emerson & Kolm 2005a, b). Here we
have used multiple genetic markers to resolve origins and
understand the paradox of two ecologically indistinguishable
and sympatric bark beetle species on the Canary Island
of La Palma. Our results support an origin via two
colonizations of La Palma with a subsequent history of
low-level introgression, followed by character divergence
and reproductive isolation. While we have dismissed a scenario
of truly sympatric speciation, clearly our results implicate
the interaction between forms for the origin of the two endemic
taxa on La Palma, a phenomenon consistent with the
hypothesis of Emerson and Kolm (2005a, b). Additionally,
our results highlight the importance of multiple genetic
markers for the robust inference of evolutionary history
within a flight-capable group of invertebrates with the
potential for frequent dispersal events between islands.
Acknowledgements
We would like to thank P. Oromi for help with logistics and
permits, and each of the ‘Cabildos’ in La Palma, El Hierro, La
Gomera and Tenerife for issuing collecting permits. B.H.J. was
supported by a Marie Curie fellowship HPMF-CT2001-01323 from
the European Union.
References
Alves PC, Ferrand N, Suchentrunk F, Harris DJ (2003) Ancient
introgression of Lepustimidus mtDNA into L. granatensis and L.
europeus in the Iberian Peninsula. Molecular Phylogenetics and
Evolution, 27, 70– 80.
Arnegard ME, Kondrashov AS (2004) Sympatric speciation by
sexual selection alone is unlikely. Evolution, 58, 222–237.
Avise JC (2000) Phylogeography: The History and Formation of Species.
Harvard University Press, Cambridge, Massachusetts.
Avise JC (2004) Molecular Markers, Natural History, and Evolution.
Sinauer Associates, Sunderland, Massachusetts.
Ballard JWO (2000) When one is not enough: introgression of
mitochondrial DNA in Drosophila. Molecular Biology and
Evolution, 17, 1126–1130.
Barluenga M, Meyer A (2004) The Midas cichlid species complex:
incipient sympatric speciation in Nicaraguan cichlid fishes.
Molecular Ecology, 13, 2061–2076.
Barluenga M, Stölting KN, Salzburger W, Muschick M, Meyer A
(2006) Sympatric speciation in Nicaraguan crater lake cichlid
fish. Nature, 439, 719 –723.
Barraclough TG, Nee S (2001) Phylogenetics and speciation.
Trends in Ecology & Evolution, 16, 391– 399.
Barraclough TG, Vogler AP, Harvey PH (1998) Revealing the
factors that promote speciation. Philosophical Transactions of the
Royal Society of London. Series B, Biological Sciences, 353, 241–249.
Barton NH, Jones JS (1983) Mitochondrial DNA: new clues about
evolution. Nature, 306, 317–318.
Bensasson D, Zhang D, Hartl DL, Hewitt GM (2001) Mitochondrial pseudogenes: evolution’s misplaced witnesses. Trends in
Ecology & Evolution, 16, 314 –321.
Berlocher SH (1998) Can sympatric speciation via host or habitat
shift be proven from phylogenetic and biogeographic evidence.
In: Endless Forms: Species and Speciation (eds Howard DJ, Berlocher
SH. ), pp. 99–113. Oxford University Press, Oxford.
Berlocher SH, Feder JL (2002) Sympatric speciation in phytophagous insects: moving beyond controversy. Annual Review
of Entomology, 47, 773–815.
Bernatchez L, Dodson JJ (1990) Allopatric origin of sympatric
populations of lake whitefish (Coregonus clupeaformis) as revealed
by mitochondrial-DNA restriction analysis. Evolution, 44, 1263 –
1271.
Besansky NJ, Krzywinski J, Lehmann T et al. (2003) Semipermeable species boundaries between Anopheles gambiae and
Anopheles arabiensis: evidence from multilocus DNA sequence
variation. Proceedings of the National Academy of Science, USA,
100, 10818–10823.
Brumfield RT, Beerli P, Nickerson DA, Edwards SV (2003) The
utility of single nucleotide polymorphisms in inferences of
population history. Trends in Ecology & Evolution, 18, 249–
256.
Cadena CD, Ricklefs RE, Jimenéz I, Bermingham E (2005) Is
speciation driven by species diversity? Nature, 438, E1–E2.
Chan KMA, Levin SA (2005) Leaky prezygotic isolation and
porous genomes: rapid introgression of maternally inherited
DNA. Evolution, 59, 720–729.
Clement M, Posada D, Templeton AR (2000) tcs: a computer
program to estimate gene genealogies. Molecular Ecology, 9,
1657–1659.
Coyne JA, Orr HA (2004) Speciation. Sinauer Associates, Sunderland, Massachusetts.
Doums C, Cabrera H, Peeters C (2002) Population genetic structure and male-biased dispersal in the queenless ant Diacamma
cyaneiventre. Molecular Ecology, 11, 2251–2264.
Edwards SV, Beerli P (2000) Gene divergence, population divergence, and the variance in coalescence time in phylogeographic
studies. Evolution, 54, 1839–1854.
Emerson BC, Forgie S, Goodacre S, Oromi P (2006) Testing phylogeographic predictions on an active volcanic island: Brachyderes
rugatus (Coleoptera: Curculionidae) on La Palma (Canary
Islands). Molecular Ecology, 15, 449–458.
Emerson BC, Kolm N (2005a) Species diversity can drive speciation. Nature, 434, 1015–1017.
Emerson BC, Kolm N (2005b) Is speciation driven by species
diversity? Nature, 438, E2.
Feder JL, Berlocher SH, Roethele JB et al. (2003) Allopatric genetic
origins for sympatric host–plant shifts and race formation in
Rhagoletis. Proceedings of the National Academy of Science, USA,
100, 10314–10319.
Fry JD (2003) Multilocus models of sympatric speciation: Bush
versus Rice versus Felsenstein. Evolution, 57, 1735–1746.
Funk DJ (1999) Molecular systematics of cytochrome oxidase I and
16S from Neochlamisus leaf beetles and the importance of
sampling. Molecular Biology and Evolution, 16, 67–82.
Funk DJ, Omland KE (2003) Species-level paraphyly and polyphyly:
frequency, causes and consequences, with insights from animal
mitochondrial DNA. Annual Review of Ecology and Systematics,
34, 397–423.
Hare MP (2001) Prospects for nuclear gene phylogeography.
Trends in Ecology & Evolution, 16, 700–706.
Hudson RR, Kaplan NL (1985) Statistical properties of the number
of recombination events in the history of a sample of DNA
sequences. Genetics, 111, 147–164.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
S Y M P A T R I C S P E C I A T I O N I N A P H A N A R T H R U M 2947
Huelsenbeck JP, Ronquist F (2001) mrbayes: Bayesian inference of
phylogeny. Bioinformatics, 17, 754 –755.
Israelson G (1972) Male copulatory organs of Macaronesian
species of Aphanarthrum Wollaston. With designations of lectotypes and descriptions of new taxa (Col. Scolytidae). Entomologica
Scandinavica, 3, 249–257.
Jordal BH (1998) A review of Scolytodes ferrari (Coleoptera:
Scolytidae) associated with Cecropia (Cecropiaceae) in the
northern Neotropics. Journal of Natural History, 32, 31–84.
Jordal BH (2006) Community structure and reproductive biology
in bark beetles (Coleoptera: Scolytinae) associated with
Macaronesian Euphorbia shrubs. European Journal of Entomology,
103, 71–80.
Jordal BH, Hewitt GM (2004) The origin and radiation of Macaronesian beetles breeding in Euphorbia: the relative importance of
multiple data partitions and population sampling. Systematic
Biology, 53, 711–734.
Kliman RM, Andolfatto P, Coyne JA et al. (2000) The population
genetics of the origin and divergence of the Drosophila simulans
complex species. Genetics, 156, 1913 –1931.
Kondrashov AS, Kondrashov FA (1999) Interactions among quantitative traits in the course of sympatric speciation. Nature, 400,
351–354.
Kondrashov AS, Yampolsky LY, Shabalina SA (1998) On the
sympatric origin of species by means of natural selection. In:
Endless Forms: Species and Speciation (eds Howard DJ, Berlocher
SH), pp. 90–98. Oxford University Press, Oxford.
Losos JB, Glor RE (2003) Phylogenetic comparative methods and
the geography of speciation. Trends in Ecology & Evolution, 18,
2003.
Machado CA, Hey J (2003) The causes of phylogenetic conflict in
a classic Drosophila species group. Proceedings of the Royal Society
of London. Series B, Biological Sciences, 270, 1193 –1202.
Machado CA, Kliman RM, Markert JA, Hey J (2002) Inferring the
history of speciation from multilocus DNA sequence data: the
case of Drosophila pseudoobscura and close relatives. Molecular
Biology and Evolution, 19, 472– 488.
Mallet J (2005) Hybridization as an invasion of the genome. Trends
in Ecology & Evolution, 20, 229–237.
Moore WS (1995) Inferring phylogenies from mtDNA variation:
mitochondrial-gene trees versus nuclear-gene trees. Evolution,
49, 718–726.
Morrow J, Scott L, Congdon B et al. (2000) Close genetic similarity
between two sympatric species of tephritid fruit fly reproductively isolated by mating time. Evolution, 54, 899 –910.
Ortiz-Barrientos D, Reiland J, Hey J, Noor MAF (2002) Recombination and the divergence of hybridizing species. Genetica,
116, 167–178.
Posada D, Crandall KA (1998) modeltest: testing the model of
DNA substitution. Bioinformatics, 14 (9), 817– 818.
Posada D, Crandall KA (2001) Intraspecific gene genealogies: trees
grafting into networks. Trends in Ecology & Evolution, 16, 37–45.
Prugnolle F, de Meeus T (2002) Inferring sex-biased dispersal from
population genetic tools: a review. Heredity, 88, 161–165.
Rolan-Alvarez E, Caballo M, Galindo J et al. (2004) Nonallopatric
and parallel origin of local reproductive barriers between two
snail ecotypes. Molecular Ecology, 13, 3415 –3424.
Rozas J, Rozas R (1999) DnaSP version 3: an integrated program
for molecular population genetics and molecular evolution
analyses. Bioinformatics, 15, 174 –175.
© 2006 The Authors
Journal compilation © 2006 Blackwell Publishing Ltd
Savolainen V, Anstett M, Lexer C et al. (2006) Sympatric speciation
in palms on an oceanic island. Nature, 441, 210–213.
Schedl KE, Lindberg H, Lindberg H (1959) Coleoptera insularum
canariensium. II. Scolytidae. Societas Scientiarum Fennica, Commentationes Biologicae, 20, 1–78.
Schluter D (1994) Experimental evidence that competition promotes divergence in adaptive radiation. Science, 266, 798– 801.
Shaw KL (2002) Conflict between nuclear and mitochondrial DNA
phylogenies of a recent species radiation: what mtDNA reveals
and conceals about modes of speciation in Hawaiian crickets.
Proceedings of the National Academy of Science, USA, 99, 16122–
16127.
Stephens M, Smith NJ, Donelly P (2001) A new statistical method
for haplotype reconstruction from population data. American
Journal of Human Genetics, 68, 978–989.
Svenson GJ, Whiting MF (2004) Phylogeny of Mantodea based on
molecular data: evolution of a charismatic predator. Systematic
Entomology, 29, 359–370.
Swofford D (2002) PAUP*. Phylogenetic Analysis Using Parsimony
(*and Other Methods), Version 4. Sinauer Associates, Sunderland,
Massachusetts.
Templeton AR, Crandall KA, Sing CF (1992) A cladistic analysis
of phenotypic associations with haplotypes inferred from
restriction endonuclease mapping and DNA sequence data. III.
Cladogram estimation. Genetics, 132, 619–633.
Thompson CE, Taylor EB, McPhail JD (1997) Parallel evolution of
lake-stream pairs of threespine stickleback (Gastersteus
aculeatus) inferred from mtDNA variation. Evolution, 51, 1955–
1965.
Ting C, Tsaur S, Wu C (2000) The phylogeny of related species as
revealed by the genealogy of a speciation gene, Odysseus.
Proceedings of the National Academy of Sciences, USA, 97, 5313–
5316.
Turelli M, Barton NH, Coyne JA (2001) Theory and speciation.
Trends in Ecology & Evolution, 16, 330–343.
Verheyen E, Salzburger W, Snoeks J, Meyer A (2003) The origin of
the superflock of cichlid fishes from Lake Victoria, East Africa.
Science, 300, 325–329.
Via S (2001) Sympatric speciation in animals: the ugly duckling
grows up. Trends in Ecology & Evolution, 16, 381–390.
Wang RL, Wakeley J, Hey J (1997) Gene flow and natural selection
in the origin of Drosophila pseudoobscura and close relatives.
Genetics, 147, 1091–1106.
Wu C-I (2001) The genic view of the process of speciation. Journal
of Evolutionary Biology, 14, 851–865.
Bjarte Jordal is a research fellow in biosystematics at the
Department of Biology, University of Bergen. His research focuses
on the taxonomy and molecular systematics of wood boring
weevils, including studies on speciation processes and adaptive
radiation. Brent Emerson is a Senior Lecturer in Evolutionary
Biology at the University of East Anglia, UK, with interests in the
application of molecular data to interpret phylogenetic history
and population dynamics, particularly within island ecosystems.
Godfrey Hewitt is an Emeritus Professor in Biology at the
University of East Anglia and is particularly interested in the use
of genetic data to understand postglacial colonization and the
structure of biodiversity.