ALLOPATRIC DIVERGENCE AND SPECIATION
IN CORAL REEF FISH: THE THREE-SPOT
DASCYLLUS, DASCYLLUS TRIMACULATUS,
Matthieu Leray,1,2Ricardo Beldade,1Sally J. Holbrook,3Russell J. Schmitt,3Serge Planes,2
and Giacomo Bernardi1,4
1Department of Ecology and Evolutionary Biology, University of California Santa Cruz, Santa Cruz, California 95060
2UMR 5244, CNRS-EPHE-UPVD, Centre de Biologie et d’Ecologie Tropicale et M´ editerran´ eenne, Universit´ e de Perpignan,
66860 Perpignan Cedex, France
3Coastal Research Center, Marine Science Institute, and Department of Ecology, Evolution, and Marine Biology, University
of California Santa Barbara, Santa Barbara, California 93106
Received July 30, 2009
Accepted November 5, 2009
Long pelagic larval phases and the absence of physical barriers impede rapid speciation and contrast the high diversity observed in
marine ecosystems such as coral reefs. In this study, we used the three-spot dascyllus (Dascyllus trimaculatus) species complex to
evaluate speciation modes at the spatial scale of the Indo-Pacific. The complex includes four recognized species and four main color
morphs that differ in distribution. Previous studies of the group using mitochondrial DNA revealed a noncongruence between
color morphs and genetic groupings; with two of the color morphs grouped together and one color morph separated into three
clades. Using extensive geographic sampling of 563 individuals and a combination of mitochondrial DNA sequences and 13 nuclear
microsatellites, we defined population/species boundaries and inferred different speciation modes. The complex is composed of
seven genetically distinct entities, some of which are distinct morphologically. Despite extensive dispersal abilities and an apparent
lack of barriers, observed genetic partitions are consistent with allopatric speciation. However, ecological pressure, assortative
mating, and sexual selection, were likely important during periods of geographical isolation. This study therefore suggests that
primarily historical factors later followed by ecological factors caused divergence and speciation in this group of coral reef fish.
KEY WORDS: Allopatry, coral reef fish diversity, microsatellites, phylogeny, speciation.
Species level phylogenies provide indirect records of the evo-
lutionary history of speciation. A link between phylogeny and
distribution range can then be used to infer the most parsimo-
nious geographical mode of speciation (Barraclough and Vogler
2000; Barraclough and Nee 2001) under two major assumptions:
(1) the current distribution range of species reflects their mode of
diversification and (2) the gene tree is congruent with the species
tree as substantiated by multiple independent molecular markers.
In addition, environmental characteristics, life-history traits, and
biological interactions gained from field observations and exper-
iments can inform the processes that drive diversification.
In the context of understanding speciation mechanisms,
as described above, marine systems challenge conventional al-
lopatric speciation. The paucity of physical barriers combined
with the potential for long-range dispersal of early life-history
C ?2010 The Author(s). Journal compilation C ?2010 The Society for the Study of Evolution.
Evolution 64-5: 1218–1230
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
extreme diversity seems to coincide with highly connected envi-
ronments. Additional factors also play important roles: (1) Dis-
Most coral reef fish have a bipartite life cycle (Leis 1991) with
sedentary adults and pelagic larvae that spend from a few days
to several months in the water column (Wellington and Victor
1989). Although biotic (Planes 1993; Riginos and Victor 2001;
Ovenden et al. 2004) and abiotic (Frith et al. 1986; Shulman and
between patchily distributed reefs, evidence of current and histor-
ical long distance dispersion and colonization have already been
documented for coral reef fish (Lessios et al. 1998). A number
of taxa in various families have widespread ranges of distribution
1999; Planes and Fauvelot 2002; Horne et al. 2008). (2) Besides
the permanent hard continental barriers that have led to separa-
tion between ocean basins (e.g., the rise of the Isthmus of Panama
between the Atlantic and Pacific Oceans, Lessios 2008), only a
few barriers to coral reef fish dispersal such as oceanic currents
(Lessios et al. 1999; Barber et al. 2000), open-ocean distances
(Vermeij 1987; Lessios and Robertson 2006), and freshwater out-
flows (Rocha 2004; Floeter et al. 2008; Beldade et al. 2009) have
been identified as important factors in marine speciation (but see
Connolly et al. 2003 for critical assessment). (3) Barriers that
are currently recognized have changed over time. Environmental
variations such as sea level fluctuations related to climatic cycles
are also likely to have impacted the effectiveness of these barriers
(Fauvelot et al. 2003; Rocha et al. 2005a; Bowen et al. 2006;
Robertson et al. 2006; Renema et al. 2008).
Due to all the factors mentioned above, dispersal abilities
probably play only a modest role in allopatric speciation in the
marine realm (Mayr 1954; Hellberg 1998). As a result, the con-
cept of ecological speciation, already recognized in the terrestrial
environment, was suggested as a potential explanation for coral
reef fish diversity (Streelman et al. 2002; Rocha et al. 2005b;
Selective pressure on color polymorphisms, so striking in
coral reef fish, may result from ecological factors such as com-
munication, competition, camouflage, habitat differences, and
mating behavior (DeMartini and Donaldson 1996; Crook 1997;
Although the role of color variation in speciation processes has
lation). By contrast, Rocha et al. (2005b) found that ecological
speciation was mediated by differential temperature between ad-
jacent habitats and Munday et al. (2004) demonstrated the key
role of competition for limited resources in host shift for sym-
patric species of Gobiodon. Taken together, these studies suggest
a strong influence of ecological pressure in speciation processes
in coral reef fish.
In this study, we evaluated the geographic mechanisms of
speciation in a group of coral reef fish, the three-spot dascyl-
lus (Dascyllus trimaculatus) species complex where geographic
distance, allopatry, coloration patterns, and ecological special-
ization are all encountered in differing degrees of intensity. To
that end, we combined phylogenetic data (mitochondrial control
region sequences and nuclear microsatellite genetic partitioning
investigation), and range distributions to infer species boundaries
and modes of speciation, as well as life-history traits and environ-
mental characteristics to evaluate the role of ecological pressure
Our results underscore the high resolution of nuclear mi-
crosatellites for investigations of macroevolutionary phenomena
underlying initial speciation.
Materials and Methods
THE D. TRIMACULATUS COMPLEX
The complex comprises four species, D. trimaculatus (R¨ uppell
1828–30), D. albisella (Gill 1862), D. strasburgi (Klausewitz
1960), and D. auripinnis (Randall and Randall 2001). Dascyllus
trimaculatus, a black fish with three small white spots, has the
widest range of distribution, from the East African Coast to the
Central Pacific, whereas the three remaining species have much
white flanks is endemic to Hawaii, D. strasburgi is a grayish fish
that is restricted to the Marquesas Islands and the recently de-
scribed D. auripinnis has yellow fins, a yellow lower body and is
only found in the Line and Phoenix Islands (Fig. 1). All species
innis that shows a parapatric distribution to D. trimaculatus with
a small sympatric area in the Northern Cook Islands (Randall and
Randall 2001; H. Debelius, pers. comm.). Some small local vari-
ation in morphology have been observed. For example, in Fiji,
fish that are described as D. trimaculatus have some yellow on
their body, potentially due to the presence of turbid waters (Allen
to be brownish (G. Bernardi, pers. obs.). In Johnston Atoll, a
few individuals, recorded as D. albisella, have yellow fin tips
(E. DeMartini, pers. comm.). Molecular studies revealed that the
four species of the D. trimaculatus complex, which display only
small morphometric and meristic differences, have only recently
Pleistocene) (Bernardi and Crane 1999; McCafferty et al. 2002).
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
Figure 1. Distribution range of the four species in the Dascyllus trimaculatus species complex. Each species is labeled with a color code
and symbol, D. trimaculatus (circle), D. albisella (triangle), D. auripinnis (square), and D. strasburgi (star). Sampling localities are shown
with their acronyms as described in Table 1. NC denotes the location of the northern Cook Islands.
recognition using sound (Lobel and Mann 1995) followed by ex-
and Holzberg 1974; Thresher 1984), larvae hatch and disperse in
the water column for 22–26 days (Wellington and Victor 1989).
where they remain protected for a few months whereas D. al-
Schmitt 1997; Ramon et al. 2008). On very rare occasions, how-
ever, D. trimaculatus have been observed to settle on corals, both
in the Indian Ocean (Oman) and the Pacific (Moorea) (pers. obs.).
Recent divergence, unbalanced distribution range, extensive
dispersal capabilities, small morphological differences, strong
habitat association, and specific biological interaction, that char-
acterize the members of the D. trimaculatus complex, make its
coral reef fish. Several phylogenetic studies have already focused
on the D. trimaculatus complex describing relationships among
the species (Bernardi and Crane 1999; Bernardi et al. 2001, 2002,
2003; McCafferty et al. 2002). A study based on mitochondrial
control region sequences found, as expected, that individuals of
D. albisella and D. strasburgi were reciprocally monophyletic
(Bernardi et al. 2002). In contrast, despite their morphological
similarity throughout the Indo-Pacific, D. trimaculatus individu-
als clustered into three mitochondrial clades that were separated
geographically. The Indian Ocean clade was considered as the
ancestral population within the complex, the two other clades
corresponded to the Tuamotu/Society Islands of French Poly-
nesia and the “West-Central Pacific.” Mitochondrial control re-
gions also clustered D. auripinnis individuals together with D.
trimaculatus West-Central Pacific individuals, raising doubts re-
garding the validity of D. auripinnis as a species. In addition,
Bernardi et al. (2003) found that 4% of D. trimaculatus individu-
als from Polynesia (termed OC3 haplotypes) also clustered with
West-Central Pacific individuals (Fig. 2). Thus, although much
information has already been gathered on the three-spot dascyllus
Figure 2. Summarized phylogenetic relationship between 563 in-
dividuals sampled in the D. trimaculatus species complex based
on mitochondrial control regions using the Maximum Likelihood
nodes and branches supported with bootstrap values superior to
50 are presented. Size of pie charts is proportional to the number
1220 EVOLUTION MAY 2010
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
species complex, many questions regarding speciation processes
and species relationships remained unanswered.
SAMPLE COLLECTION AND DNA EXTRACTIONS
the complete distribution range of the species (Table 1, Fig. 1).
Individuals were collected using hand spears while free or scuba
Table 1. Number of sampled individuals per location used for mi-
tochondrial control region (Dloop) and microsatellites (SSR) phy-
logenetic analysis. Numbers in parentheses are the new samples
that were sequenced. The sample grouping was based on (1)
the mitochondrial DNA genetic partitioning described in Bernardi
acronyms are used on the map (Fig. 1).
SpeciesGroup locality Acronym
D. trimaculatus Indian Ocean42
West Central Pacific
Great Barrier Reef GBR
Hawai + Johnston
Kure, PH, Midway KUR
diving. Immediately after collection, fin clips were placed in 95%
ethanol and stored at ambient temperature in the field, and then
at 4◦C in the laboratory. Total genomic DNA was prepared from
20 mg of fin tissue by proteinase K digestion in lysis buffer
(10 mM Tris, 400 mM NaCl, 2 mM EDTA, 1% SDS) overnight at
55◦C. This was followed by purification using phenol/chloroform
extractions and alcohol precipitation (Sambrook et al. 1989).
DNA SEQUENCES AND MICROSATELLITE DATA
Previously published sequences from the mitochondrial control
region (Dloop) were included in this study (Bernardi and Crane
1999; Bernardi et al. 2001, 2002, 2003; Ramon et al. 2008),
together with an additional 126 new samples (Table 1). These
samples represent the entire distribution of the species complex
except for the region comprised between western Australia and
Cocos-Keeling. Using microsatellites markers allowed to com-
pare nuclear genetic partitioning within the D. trimaculatus com-
plex with the genetic partitioning observed using mitochondrial
DNA sequences and reveal patterns and processes such as incom-
plete lineage sorting, hybridization, and introgression. Therefore,
the selection of samples used for microsatellite data analysis was
based on described morphological species and on the mitochon-
drial clades. At least 48 individuals per clade were randomly
selected for screening and all the D. auripinnis available (49 in-
dividuals) were also genotyped. Within the West-Central Pacific
and Hawaiian archipelago, where particular processes of interest
seem to take place based on previous mitochondrial DNA anal-
ysis, the selection of samples included a few individuals from
different locations chosen to integrate all the genetic variability.
Finally, all Polynesian samples that cluster with the West-Central
of Bernardi et al. 2003) were genotyped.
MITOCHONDRIAL GENE AMPLIFICATION AND
Sequence data of 366 bp for the mitochondrial control region
were amplified for 126 new samples using previously detailed
conditions (Bernardi et al. 2001) and primers (Lee et al. 1995).
Mitochondrial sequences were aligned using Bioedit (Hall 1999).
For microsatellite analyses, we tested 52 loci isolated from a
D. trimaculatus individual sampled in Moorea (French Polyne-
sia). Of these, 13 successfully amplified all samples and were
found to be highly polymorphic. Leray et al. (2009) provide a full
description of amplification procedures and microsatellite char-
acteristics. GENEMAPPER3.7 Applied Biosystems (Foster City,
CA) was used for scoring microsatellite genotypes.
PHYLOGENETIC RELATIONSHIP AND GENETIC
To evaluate phylogenetic relationships based on mitochondrial
control regions, 10 independent maximum likelihood (ML)
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
replicates were computed using the program GARLI (Zwickl
2006) that uses a genetic algorithm approach to simultaneously
find the topology, branch lengths and model parameters that max-
imize the likelihood of the tree. Out of the 10 best trees, the one
with the highest likelihood overall was selected. Branch support
on the resulting best tree was assessed via 100 bootstraps imple-
mented by GARLI. In addition to using a ML approach, pairwise
genetic distances were estimated using Arlequin 3.0 (Excoffier
et al. 2005) to quantify the divergence between previously iden-
tified groups. To infer which model of nucleotide substitution
best fit the data, Modeltest (Posada and Crandall 1998) was used.
To account for differences between gene divergence and popula-
tion divergence due to the presence of ancestral polymorphisms
(Edwards and Beerli 2000), average pairwise distances between
populations were the mean number of pairwise differences be-
tween two populations minus the average distance between in-
dividual within those populations. Significance was tested with
10,000 replicates and a sequential Bonferroni correction for the
level of significance was used.
Samples were divided into groups previously identified based on
the ML phylogenetic tree. Haplotype diversity, nucleotide diver-
sity, and mean number of pairwise differences were calculated
with DNAsp (Rozas et al. 2003).
ANALYSIS OF THE NUCLEAR MICROSATELLITES
Based on predefined groups (mitochondrial clades and morpho-
logical species), the number of alleles, observed and expected
heterozygosities were computed per locus using Fstat (Goudet
1995). Linkage disequilibrium between loci and deviations from
Hardy–Weinberg equilibrium (HWE) were also tested for each
locus and group. Significance was tested with 10,000 replicates.
Levels of significance for multiple comparisons of loci across
samples were adjusted using a standard Bonferroni correction
(Rice 1989). Departures from HWE can be caused by inbreeding,
Wahlund effect, or technical causes such as null alleles, misscor-
ing due to stuttering, and allelic dropout. The proportion of null
alleles that could influence the genetic signal was evaluated using
the algorithm implemented in FREENA (Dempster et al. 1977;
Chapuis and Estoup 2007).
To explore and decompose the genetic variability into gene pools
without providing prior information on the geographical origin
of the samples, a Bayesian clustering approach implemented in
Structure 2.2 was used (Pritchard et al. 2000). The program si-
multaneously defines clusters and assigns individual multilocus
genotypes to the defined clusters. Allele frequencies were pre-
sumed uncorrelated and null alleles were coded as recessive to
et al. 2007). The most likely number of clusters in the dataset was
identified based on posterior probabilities and 10 runs were im-
plemented in Structure 2.2 to test for robustness of the results
following Pritchard et al. (2000).
COALESCENCE ESTIMATES AND MIGRATION RATES
Historical demography parameters were examined for groups
based on genetic partitioning and morphology to evaluate pos-
sible events of population fluctuation based on a coalescent ap-
proach. Estimates of ? (=2Nμ, where μ is the mutation rate)
were made for each genetic group as well as the whole species
complex. The parameter ? was estimated under two conditions:
an unconstrained exponential growth parameter and an assump-
tion of constant N (g = 0). We used FLUCTUATE (Kuhner et al.
1998) to estimate the ML of the parameters ? and g (the ex-
ponential growth parameter in units μ−1). Seeds for all analyses
were generated randomly. Analyses were repeated 10 times per
region to ensure stability of parameters estimates. Final analyses
each and five long chains of length 20,000, with a sample incre-
ment of 20. Exchanges and range expansions (immigration) be-
tween groups were investigated with MIGRATE 2.0 (Beerli and
Felsenstein 2001; Beerli 2003) to identify reproductively isolated
between genetic partitions and morphological species. Multiple
length of the analysis until reaching consistent results. The final
parameters used 10 short Monte Carlo chains with 5000 recorded
genealogies and six long chains with 50,000 recorded genealo-
between chains running(in parallel) at different temperatures was
implemented, the “colder” chain exploring less genealogy space
than the “hotter” one. Again 10 replicates were realized to ensure
the reproducibility and stability of the estimation.
The time of coalescence was estimated by assuming that
coalescence was reached when the population size was reduced
to 1% of its present day value, following Wares and Cunningham
(2001). To estimate coalescence time, we used the mutation rate
(μ = substitutions per site per generation) that was determined
for the mitochondrial control region for other damselfish species,
the Panama Trans-Isthmian geminates Chromis atrilobata and C.
multilineata (Domingues et al. 2005).
MITOCHONDRIAL DNA ANALYSIS
Although the dataset used here was more extensive, results were
1222 EVOLUTION MAY 2010
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
consistent with previously published data (Bernardi et al. 2002,
2003). The Indo-Pacific separated into five distinct geograph-
ical areas: the Indian Ocean (clade 1), the West-Central Pacific
the Marquesas Islands, northern French Polynesia (clade 4), and
the Hawaiian archipelago (clade 5) including Johnston atoll.
Dascyllus strasburgi and D. albisella each segregated into
monophyletic clades, corresponding to Marquesas (clade 4) and
Hawaii (clade 5) samples, respectively. The West-Central Pacific
group (clade 2) included D. trimaculatus from West-Central Pa-
cific locales, a small number of D. trimaculatus from Society and
Tuamotu islands (OC3 samples), and all D. auripinnis individ-
uals. The genetic partitioning across the rest of the Indo-Pacific
did not match with taxonomically described species. Indeed, the
wide-ranging D. trimaculatus separated into three deeply diver-
gent groups (clade 1, 2, and 3, Fig. 2). Moreover, within the
West-Central Pacific, where the two species D. auripinnis and
D. trimaculatus occur, all sampled individuals grouped in a sin-
gle well-supported clade (clade 2). Within clade 2, two subclades
(weakly supported and not shown) were found, suggesting poten-
tial population structure within this clade. D. auripinnis individ-
uals were found in both subclades.
Genetic diversity and historical demography in the
D. trimaculatus complex
Genetic diversity reflected the historical demography of the dif-
ferent genetic entities of the D. trimaculatus species complex.
Table 2. Genetic characteristics of Dascyllus trimaculatus clades based on the mitochondrial control region. Genetic diversity and
coalescence data are presented, with numbers in parentheses being standard deviations. Estimated coalescence times are given in
millions of years (My).
Number of sequences
Number of segregating sites
Mean number of nucleotide
Coalescence time (My)
Genetic diversity was highest in the Indian Ocean (Table 2). For
example, the mean number of nucleotide differences between
Indian Ocean haplotypes was twice that in all the other clades
(Table 2). Conversely, the genetic diversity of the West-Central
Pacific was the lowest despite the broad geographic origin of
the individuals: all West-Central Pacific samples (two described
species D. trimaculatus and D. auripinnis) and 25 samples from
Society and Tuamotu islands (OC3 samples).
Overall coalescence for the D. trimaculatus complex oc-
curred approximately between 560,000 and 630,000 years ago
(ya). Coalescence estimates are more recent for the Indian Ocean
(190,000–210,000 ya) than for the Pacific Ocean (470,000–
530,000 ya). In addition, coalescence estimates for each clade,
which ranged from 150,000 to 400,000 ya, were lower for the
West-Central Pacific clade than for Society + Tuamotu, Marque-
sas, and even Hawaii clades (Table 2).
NUCLEAR DNA ANALYSIS
Genetic partitioning based on nuclear microsatellites
Specific characteristics of the microsatellite data used here are
provided in Table S1. A partition of our samples in seven clusters
was found to best fit the observed microsatellite variability using
the clustering method (Fig. 3, and Table S2). All individuals from
high levels of reassignment to their own cluster (average: 0.995).
assigned to their own cluster (reassignment = 0.996), including
the OC3 samples that in the mitochondrial dataset clustered with
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
Figure 3. Bayesian population assignment test based on 13 microsatellites loci. The seven clusters that partition the data are displayed
with different colors and mapped on the lower panel. Each vertical line represent one individual and its assignment likelihood to belong
to one of the cluster (Y scale) is shown by the color. Black vertical lines represent the limit between predefined groups. These are the five
mitochondrial clades, Indian Ocean (dark blue), Western Central Pacific (blue), Tuamotu + Society Islands (pale blue), Hawaii (green), and
Marquesas Islands (red). Within the Hawaiian Archipelago, Johnston Atoll + some French Frigate shoals individuals (dark green) separate
from other Hawaiian individuals (green). Within the Western Central Pacific, we have separated D. trimaculatus (blue) and D. auripinnis
(yellow). The arrow represents the increasing physical distance to D. auripinnis distribution range for D. trimaculatus individuals sampled
in the West-Central Pacific.
the West-Central Pacific haplotypes (Fig. 3). Within the Hawaii
cluster, samples were divided in two different subclusters. One
subcluster comprised all 20 sampled individuals from Johnston
Atoll (JOHN) and five samples from the northwestern Hawai-
ian island of French Frigate Shoals (FFS), and one subcluster
comprised all other remaining Hawaiian samples (this cluster in-
cluded samples both from the main island group and from the
Northwestern Hawaiian island of Kure, Table 1). West-Central
Pacific individuals presented a mosaic picture. Unlike the mito-
chondrial results, genotype assignment divided D. auripinnis and
D. trimaculatus in two clusters (Fig. 3). All D. auripinnis were
assigned to a single cluster, except one individual (collected in
Palmyra) that clustered with D. trimaculatus.
We specifically chose a suite of D. trimaculatus individuals
from the West-Central Pacific that would span a region with
varying distances to the contact zone between D. trimaculatus
and D. auripinnis. Of these 48 samples, 19 clustered with D.
auripinnis, and the presence of D. auripinnis genotypes occurred
in a decreasing proportion with increasing distance from the Line
Islands. Indeed, five of eight fish in Wallis and nine Fijian indi-
viduals of 15 displayed a D. auripinnis genotype. Although no
coloration pattern was available for the Wallis samples, samples
from Fiji displayed yellow anal and pelvic fins with streaks of or-
ange in the caudal fin and the spinous part of the dorsal fin (Allen
1991, pers. obs.). This color variant, intermediate between the
bright yellow D. auripinnis and the pure black D. trimaculatus, is
consistent with a mixed D. auripinnis–D. trimaculatus genotype.
Mitochondrial molecular markers have extensively been used to
infer phylogenetic relationships and species boundaries. In the
consistent with previously published results solely based on mi-
tochondrial markers. However, the use of highly variable mi-
crosatellites allowed for additional information that could not be
obtained with mitochondrial markers alone.
Nuclear and mitochondrial molecular markers were consis-
tent in genetically identifying the two species D. strasburgi and
1224 EVOLUTION MAY 2010
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
D. albisella endemic to the Marquesas and Hawaii, respectively.
Within D. albisella, mitochondrial sequences put Johnston Atoll
individuals in a single cluster within the remaining Hawaiian
samples (Ramon et al. 2008), whereas nuclear microsatellites,
with more segregating power, grouped Johnston Atoll individ-
uals, with five samples collected at FFS (Northwestern Hawai-
ian islands) and separated them from the remaining Hawaiian
individuals. Based on the mitochondrial control region, strong
population structure also exists (not presented) within the D.
albisella distribution range and warrants further investigation
(Ramon et al. 2008). Thus, overall, for both D. albisella and
D. strasburgi coloration patterns coincide with genetic partitions
even though a cryptic divergence appears to be occurring within
genetic partitioning was found in D. trimaculatus (Bernardi et al.
2002). Although intraspecific morphological, behavioral, or eco-
1966; Fricke 1973), both mitochondrial and nuclear markers di-
vided the widespread D. trimaculatus into three deeply divergent
genetic partitions that separate the Indo-Pacific biogeographical
and southern French Polynesia.
Although Indian Ocean and southern French Polynesian re-
gions seem well defined genetically, complex processes seem to
take place within the West-Central Pacific. Dascyllus auripinnis
had been considered a phenotypic variant of D. trimaculatus, and
only recently was described as a new species based on striking
different color patterns but “modest” morphometric differences
measured on a small number of samples (Randall and Randall
2001). These include a larger size, one fewer gill raker, and a
shorter average length of the paired fins. Although Bernardi et al.
(2002) could not determine whether D. auripinnis was a valid
species, an emerging species, or an isolated population of D. tri-
provided new insights in the geographic dynamic of introgression
black phenotypes. These results are consistent with “directional”
mating that, very likely, is driving introgression from D. aurip-
innis into D. trimaculatus through repeated backcrossing. In the
contact zone, where both species live in sympatry, at the north-
ern Cook Islands of Penhryn (Tongareva) and Suwarrow Islands
(Randall and Randall 2001), the yellow form is much more abun-
dant than the black form (H. Debelius, pers. comm.). Therefore,
if hybridization occurs at this location, rare chance events of in-
trogression of the yellow on the black fish are more likely than in
the opposite direction. Directional hybridization in species with
different abundances has also been observed in other coral reef
species (Yaakub et al. 2006). Although assortative mating based
on coloration pattern is likely to be present, as seen in other dam-
selfish (Planes et al. 2001; van Herwerden and Doherty 2006), a
direct examination of mating behavior in this area (sneak mating
was observed in Yaakub et al. 2006), coupled with paternity anal-
ysis, would allow a test of this hypothesis. An alternative to this
ecological explanation, that would require further testing, is that
a selective sweep may be responsible for the clinal distribution of
this haplotype (e.g., Galtier et al. 2000).
GEOGRAPHY OF SPECIATION
The use of multiple markers to infer the phylogeny of a very
recently diverged group such as the D. trimaculatus species com-
plex allows inferences about geographical modes of speciation
using comparison of present day distribution range between sister
species (Barraclough and Vogler 2000; McCafferty et al. 2002).
Recently allopatrically diverged taxa will likely have a nonover-
lapping distribution. Both D. albisella and D. strasburgi group
in monophyletic clades and cluster based on mitochondrial and
nuclear DNA. Bernardi et al. (2002) suggested that these species
emerged as the result of rare chance long-distance dispersal of
propagules, which once they settled in remote locations diverged
in allopatry as the result of low level of gene flow. The congru-
ent biogeographical and genetic patterns shared by several taxa
and independent genetic markers support the allopatric speciation
scenario in the Marquesas and Hawaiian archipelagos (Planes
and Fauvelot 2002). Although the Hawaii separation is mainly
explained by its geographical isolation, the separation between
the Marquesas and Society + Tuamotu islands is likely related
to the barrier of larval exchange created by the South Equatorial
direction to the South Equatorial Current (Vermeij 1987; Planes
1993). Importantly, although such an oceanographic barrier may
potentially prevent any larval crossing, an interchange of a few
propagules may still be possible. In this case, however, the rela-
tive success of the recruits is probably the cause of the observed
At a smaller scale, emerging divergence was detected within
drial DNA. Because all samples from the most remote location,
Johnston Atoll, group together, the phylogenetic structure is very
likely the outcome of reduced gene flow between islands within
the Hawaiian archipelago. Larval transport simulations described
in Ramon et al. (2008) also support this hypothesis. Indeed, sev-
eral dispersion corridors were identified between Johnston Atoll
and the Hawaiian archipelago but only for organisms with a
pelagic larval duration (PLD) longer than 40 days (Kobayashi
2006). However, although PLD is short in D. albisella (approxi-
mately 25 days), the presence of five samples from the Hawaiian
gests long distance, but rare, larval exchange.
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
Allopatric speciation, as observed within D. trimaculatus,
match numerous previous genetic comparisons between West-
ern Pacific and Indian Ocean populations of marine invertebrates
and fish (e.g., Lacson and Clark 1995; McMillan and Palumbi
1995; Chenoweth et al. 1998; Duda and Palumbi 1999; Lessios
et al. 1999; Barber et al. 2000; Hobbs et al. 2009). The gener-
ally accepted explanation for this observed pattern in many taxa
is that sea level fluctuations culminating during the last Pleis-
tocene glaciation, exposed the Sunda shelf, between Malaysia
and northern Australia, thus limiting water exchange between the
two oceanic provinces. In the case of D. trimaculatus, the Pacific
and Indian Oceans cryptic species are genetically partitioned for
both mitochondrial DNA and nuclear markers, indicating a cur-
rent absence of gene flow, despite their adjacent distribution. The
contact zone, represented in our samples by the Cocos-Keeling
population, does not seem to indicate the presence of gene flow
in the region of secondary contact either. However, defining the
precise boundary between the Indian Ocean and West-Central
Pacific clades will require additional sampling in the Western
Australia–Cocos-Keeling area. Indeed, as for other genetic stud-
ies, it is likely that Western Australia samples may group with
West-Central Pacific samples rather than Indian Ocean ones,
and additional samples from Cocos-Keeling/Christmas Island
may reveal some level of gene exchange in the region (Hobbs
et al. 2009).
The second genetic boundary within the D. trimaculatus
species, between the Tuamotu\Society Islands and the West-
Central Pacific results from a different process. Mitochondrial
individuals (OC3) group with the West-Central Pacific clade,
whereas microsatellite analysis clustered those same individu-
als exclusively with the Polynesian group. Bernardi et al. (2002)
suggested that the presence of two distinct clades in Polynesia
was the result of larval transport from the West-Central Pacific
to South Polynesia. The present study, with new information pro-
vided by highly polymorphic microsatellites, is consistent with
such an ecological explanation combined with introgression. Al-
with a lack of resolution for the mitochondrial DNA. Using the
mitochondrial control region, genetic structure between South
Polynesia and West-Central Pacific has previously been observed
data), two species that have a comparable PLD to D. trimacula-
tus (Juncker et al. 2006). In comparison Acanthurus triostegus,
a species with a PLD twice as long, 60–70 days (Randall 1961;
Juncker et al. 2006), does not show significant genetic structure
(S. Planes, unpubl. data). Therefore, these patterns suggest that
highly restricted gene flow (which may be linked to the PLD or
to differential survival of the propagules) between the two geo-
graphical areas is consistent among different species, resulting in
a strong genetic break between South Polynesia and the West-
ulatus have very different sizes of geographic range. The range
of D. auripinnis is less than 5% of the range of D. trimaculatus.
In itself, this may suggest peripatric speciation, as divergence of
a peripheral population in allopatry (Losos and Glor 2003). The
small overlapping range observed in the Northern Cook Islands
would also be an indication of a range shift. The hypothesis of
peripatric speciation assumes the temporary emergence of a bar-
rier that isolated a peripheral population in the Line and Phoenix
Islands. It is possible, however, that a parapatric ecological mode
of speciation (Wu and Ting 2004), or rapid sexual selection, may
have played a role in the divergence of this species.
Taken together, the geography of speciation in the D. trimac-
ulatus species complex points toward allopatric and parapatric
modes of speciation that led to the diversification of the com-
plex with no evidence for sympatric speciation given the overall
lack of overlap of the distribution range of the genetic partitions
(Fig. 4). Our data cannot distinguish whether the divergence was
due to natural selection, drift, or both. Besides simple geographic
partitioning, it is possible that additional ecological pressure on
speciation may also have played an important role in the diversi-
fication process (e.g., Hemmer-Hansen et al. 2007).
Rocha et al. (2005b) argued that the paradox of reef fish diversity
and their high dispersal capabilities may be solved by ecological
speciation. They demonstrated that differential selection would
operate in adjacent habitats with contrasting environments and
ability, predation, and competition are also thought to promote
divergent or disruptive selection leading to speciation via charac-
tat, or behavioral changes with or without geographic isolation
(Schluter 2000; Rundle and Nosil 2005). Munday et al. (2004)
presented a case of speciation via host shift in highly specialized
species of coral dwelling fish (genus Gobiodon), and proposed a
key role for competition in the divergence. In general, D. trimac-
ulatus remain protected from predators within anemones during
their juvenile phase (12 months). Yet, D. albisella, a species en-
demic to the Hawaiian Archipelago, where suitable anemones are
virtually absent, recruit on branching corals. Although species in
the D. trimaculatus complex most likely arose in allopatry, early
host shift may therefore have played an important role in their
divergence. Occasionally, D. trimaculatus settles on branching
corals or on Diadema urchins (G. Bernardi, pers. obs.), indicating
that ecological divergence in the D. trimaculatus species complex
is observed. An additional ecological signal may be uncovered
1226EVOLUTION MAY 2010
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
Figure 4. Schematic representation of potential modes of speciation in the Dascyllus trimaculatus species complex. The different sizes
of the arrow represent unbalanced levels of gene flow.
in the D. auripinnis–D. trimaculatus hybrid zone. Indeed, the
presence of sound production during mating, the variation in col-
oration patterns, and the imbalance between D. auripinnis versus
D. trimaculatus haplotypes and genotypes suggest that female
choice, assortative mating, and sexual selection may play an im-
portant, yet unexplored role in the genetic partitioning of individ-
CONCLUSIONS AND PERSPECTIVES
The combined analysis of mitochondrial and nuclear markers
presented here provides interesting insights on early processes
of speciation in the D. trimaculatus species complex. Mitochon-
drial markers allowed us to broadly define five genetic clades that
served as the basis of understanding of the complex relationship
between established taxonomy, distribution, and coloration pat-
terns of the species. Highly variable nuclear markers refined this
view by settling alternative scenarios (the presence of unsorted
OC3 haplotypes in south Polynesian samples), and uncovering
subtle differences (Johnston Atoll individuals, D. trimaculatus–
D. auripinnis hybrid zone). The study of speciation processes in
an opportunity to shed light on such intricate systems.
We would like to thank R. Galzin and C. Fauvelot for comments and
discussion, and Editor M. Hellberg and three anonymous reviewers for
comments. We would like to very much thank G. Lecaillon for providing
so many samples over the years from Mayotte, Philippines, and Vietnam.
We would also like to thank P. Nelson, D. R. Robertson, J. H. Choat,
and J. McIlwain for providing samples. NSF MCR LTER Award OCE
for this research.
Allen, G. R. 1991. Damselfishes of the world. Mergus, Germany
Balloux, F., and N. Lugon-Moulin. 2002. The estimation of population differ-
entiation with microsatellite markers. Mol. Ecol. 11:155–165.
Balloux, F., H. Brunner, N. Lugon-Moulin, J. Hausser, and J. Goudet. 2000.
Microsatellites can be misleading: an empirical and simulation study.
Barber, P. H., S. R. Palumbi, M. V. Erdmann, and M. K. Moosa. 2000. Bio-
geography – A marine Wallace’s line? Nature 406:692–693.
Barraclough, T. G., and S. Nee. 2001. Phylogenetics and speciation. Trends
Ecol. Evol. 16:391–399.
Barraclough, T. G., and A. P. Vogler. 2000. Detecting the geographical pattern
of speciation from species-level phylogenies. Am. Nat. 155:419–434.
Beerli, P. 2003. Migrate—a maximum likelihood program to estimate
gene flow using the coalescent, Tallahassee/Seattle. http://evolution.
Beerli, P., and J. Felsenstein. 2001. Maximum likelihood estimation of a
migration matrix and effective population sizes in n subpopulations by
using a coalescent approach. Proc. Natl. Acad. Sci. USA 98:4563–4568.
Beldade, R., J. B. Heiser, D. R. Robertson, J. L. Gasparini, S. Floeter, and
G. Bernardi. 2009. Historical biogeography and speciation in the Creole
wrasses (Labridae, Clepticus). Mar. Biol. 156:679–687.
Bernardi, G., and N. L. Crane. 1999. Molecular phylogeny of the humbug
damselfishes inferred from mtDNA sequences. J. Fish Biol. 54:1210–
Bernardi, G., S. J. Holbrook, and R. J. Schmitt. 2001. Gene flow at three
spatial scales in a coral reef fish, the three-spot dascyllus, Dascyllus
trimaculatus. Mar. Biol. 138:457–465.
Bernardi, G., S. J. Holbrook, R. J. Schmitt, N. L. Crane, and E. DeMartini.
three-spot damselfish (Dascyllus trimaculatus) species complex. Proc.
R. Soc. Lond. B 269:599–605.
Bernardi, G., S. J. Holbrook, R. J. Schmitt, and N. L. Crane. 2003. Genetic
evidence for two distinct clades in a French Polynesian population of
the coral reef three-spot damselfish Dascyllus trimaculatus. Mar. Biol.
Bowen, B. W., A. Muss, L. A. Rocha, and W. S. Grant. 2006. Shallow mtDNA
a recent invasion from the Indian Ocean. J. Hered. 97:1–12.
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
of population differentiation. Mol. Biol. Evol. 24:621–631.
Chenoweth, S. F., J. M. Hughes, C. P. Keenan, and S. Lavery. 1998. When
oceans meet: a teleost shows secondary intergradation at an Indian-
Pacific interface. Proc. R. Soc. Lond. B 265:415–420.
the argument. J. Biogeogr. 33:967–968.
Connolly, S. R., D. R. Bellwood, and T. P. Hughes. 2003. Indo-Pacific bio-
diversity of coral reefs: deviations from a mid-domain model. Ecology
Crook, A. C. 1997. Colour patterns in a coral reef fish—Is background com-
plexity important? J. Exp. Mar. Biol. Ecol. 217:237–252.
DeMartini, E. E., and T. J. Donaldson. 1996. Color morph-habitat relations in
the arc-eye hawkfish Paracirrhites arcatus (Pisces: Cirrhitidae). Copeia
Dempster, A. P., N. M. Laird, and D. B. Rubin. 1977. Maximum likeli-
hood from incomplete data via the EM algorithm. J. R. Stat. Soc.
chromis. Mol. Ecol. 14:4051–4063.
Duda, T. F., and S. R. Palumbi. 1999. Population structure of the black tiger
prawn, Penaeus monodon, among western Indian Ocean and western
Pacific populations. Mar. Biol. 134:705–710.
Edwards, S. V., and P. Beerli. 2000. Perspective: gene divergence, population
divergence, and the variance in coalescence time in phylogeographic
studies. Evolution 54:1839–1854.
software package for population genetics data analysis. Evolutionary
Bioinformatics Online 1:47–50.
Falush, D., M. Stephens, and J. K. Pritchard. 2007. Inference on population
structure using multilocus genotype date: dominant markers and null
alleles. Mol. Ecol. Notes 7:574–578.
Fautin, D. G., and G. R. Allen. 1992. Field guide to the anemone fishes and
their host anemones. Western Australian Museum, Perth.
DNA diversity of coral reef fish provide evidence of population bottle-
necks resulting from Holocene sea-level change. Evolution 57:1571–
Fishelson, L. 1966. Observations on the biology and behaviour of Red Sea
coral fishes. Contribution to the knowledge of the Red Sea, No. 30.
Deep-Sea Res. 13:322.
Floeter, S. R., L. A. Rocha, D. R. Robertson, J. C. Joyeux, W. Smith-Vaniz, P.
Wirtz, A. J. Edwards, J. P. Barreiros, C. E. L. Ferreira, J. L. Gasparini,
et al. 2008. Atlantic reef fish biogeography and evolution. J. Biogeogr.
trimaculatus Pisces Pomacentridae. Z. Tierpsychol. 32:225–256.
Fricke, H. W., and S. Holzberg. 1974. Social units and hermaphroditism in a
pomacentrid fish. Naturwissenschaften 61:367–368.
Frith, C. A., J. M. Leis, and B. Goldman. 1986. Current in the Lizard Island
region of the Great Barrier Reef Lagoon and their relevance to potential
movements of larvae. Coral Reefs 5:81–92.
Galtier, N., F. Depaulis, and N. H. Barton. 2000. Detecting bottlenecks and
Gill, T. N. 1862. Catalogue of the fishes of Lower California in the Smith-
sonian Institution, collected by Mr. J. Xantus. Proc. Acad. Natl. Sci.
J. Hered. 86:485–486.
Hall, T. 1999. BioEdit: a user-friendly biological sequence alignment editor
and analysis program for Windows 95/98/NT. Nucleic Acids Symp. Ser.
of speciation in the marine gastropod Tegula. Evolution 52:1311–1324.
Hemmer-Hansen, J., E. E. Nielsen, J. Frydenberg, and V. Loeschcke. 2007.
Adaptive divergence in a high gene flow environment: Hsc70 varia-
tion in the European flounder (Platichthys flesus L.). Heredity 99:592–
Hobbs, J.-P. A., A. J. Frisch, G. R. Allen, and L. Van Herwerden. 2009.
Marine hybrid hotspot at Indo-Pacific biogeographic border. Biol. Lett.
Holbrook, S. J., and R. J. Schmitt. 1997. In situ nocturnal observations of reef
fishes using infrared video. Pp. 805–812 in B. S. J. Y. Sire, ed. Proceed-
ings 5th Indo-Pacific Fish Conference, Noum´ ea, New Caledonia.
Horne, J. B., L. van Herwerden, J. H. Choat, and D. R. Robertson. 2008.
High population connectivity across the Indo-Pacific: congruent lack of
Juncker, M., L. Wantiez, and D. Ponton. 2006. Flexibility in size and age at
settlement of coral reef fish: spatial and temporal variations in Wallis
Islands (South Central Pacific). Aquat. Living Resour. 19:339–348.
Klausewitz, W. 1960. Dascyllus strasburgi, ein neuer Fisch aus dem Pazifik.
(Pisces, Perciformes, Pomacentridae). Bull. Aquat. Biol. 2:45–49.
Kobayashi, D. R. 2006. Colonization of the Hawaiian Archipelago via
Johnston Atoll: a characterization of oceanographic transport corri-
dors for pelagic larvae using computer simulation. Coral Reefs 25:407–
Kuhner, M. K., J. Yamato, and J. Felsenstein. 1998. Maximum likelihood
estimation of population growth rates based on the coalescent. Genetics
Lacson, J. M., and S. Clark. 1995. Genetic divergence of Maldivian and
biocellata, C. glauca and C. leucopoma (Pomacentridae). Mar. Biol.
Lee, W. J., J. Conroy, W. H. Howell, and T. D. Kocher. 1995. Structure and
evolution of Teleost mitochondrial control regions. J. Mol. Evol. 41:54–
fishes. In P. F. Sale, ed. The ecology of fishes on coral reefs. Academic
Press, San Diego, CA.
Leray, M., R. Beldade, S. J. Holbrook, R. J. Schmitt, S. Planes, and G.
Bernardi. 2009. Isolation and characterization of 13 polymorphic nu-
clear microsatellite primers for the widespread Indo-Pacific three-spot
damselfish, Dascyllus trimaculatus, and closely related D. auripinnis.
Mol. Ecol. Res. 9:213–215.
Lessios, H. A. 2008. The Great American Schism: divergence of marine
organisms after the rise of the Central American Isthmus. Annu. Rev.
Ecol. Evol. Syst. 39:63–91.
Lessios, H. A., and D. R. Robertson. 2006. Crossing the impassable: genetic
connections in 20 reef fishes across the eastern Pacific barrier. Proc.
Roy. S. Lond. B 273:2201–2208.
Lessios, H. A., B. D. Kessing, and D. R. Robertson. 1998. Massive gene flow
across the world’s most potent marine biogeographic barrier. Proc. R.
Soc. Lond. B 265:583–588.
Lessios, H. A., B. D. Kessing, D. R. Robertson, and G. Paulay. 1999. Phy-
logeography of the pantropical sea urchin Eucidaris in relation to land
barriers and ocean currents. Evolution 53:806–817.
Lobel, P. S., and D. A. Mann. 1995. Spawning sounds of the damselfish, Das-
cyllus albisella (Pomacentridae), and relationship to male size. Bioa-
1228 EVOLUTION MAY 2010
ALLOPATRIC DIVERGENCE AND SPECIATION IN CORAL REEF FISH
Losos, J. B., and R. E. Glor. 2003. Phylogenetic comparative methods and the
geography of speciation. Trends Ecol. Ecol. 18:220–227.
Marshall, N. J., K. Jennings, W. N. McFarland, E. R. Loew, and G. S. Losey.
2003. Visual biology of Hawaiian coral reef fishes. III. Environmental
McCafferty, S., E. Bermingham, B. Quenouille, S. Planes, G. Hoelzer, and K.
Asoh. 2002. Historical biogeography and molecular systematics of the
Indo-Pacific genus Dascyllus (Teleostei : Pomacentridae). Mol. Ecol.
McMillan, W. O., and S. R. Palumbi. 1995. Concordant evolutionary pat-
terns among Indo-West Pacific Butterflyfishes. Proc. R. Soc. Lond. B
McMillan, W. O., L. A. Weigt, and S. R. Palumbi. 1999. Color pattern evolu-
tion, assortative mating, and genetic differentiation in brightly colored
Butterflyfishes (Chaetodontidae). Evolution 53:247–260.
Moland, E., and G. P. Jones. 2004. Experimental confirmation of aggressive
mimicry by a coral reef fish. Oecologia 140:676–683.
Munday, P. L., P. J. Eyre, and G. P. Jones. 2003. Ecological mechanisms for
evaluation. Oecologia 137:519–526.
Munday, P. L., L. van Herwerden, and C. Dudgeon. 2004. Evidence for sym-
patric speciation by host shift in the sea. Curr. Biol. 14:1498–1504.
population structure in a potentially vagile fish species (Pristipomoides
multidens, Teleostei: Perciformes : Lutjanidae) from the East Indies
triangle. Mol. Ecol. 13:1991–1999.
of the convict surgeonfish Acanthurus triostegus in French Polynesia.
Mar. Ecol. Prog. Ser. 98:237–246.
Planes, S., and P. J. Doherty. 1997. Genetic relationships of the colour morphs
of Acanthochromispolyacanthus (Pomacentridae) onthenorthern Great
Barrier Reef. Mar. Biol. 130:109–117.
Planes, S., and C. Fauvelot. 2002. Isolation by distance and vicariance drive
genetic structure of a coral reef fish in the Pacific Ocean. Evolution
Planes, S., P. J. Doherty, and G. Bernardi. 2001. Unusual case of extreme ge-
netic divergence in a marine fish, Acanthochromis polyacanthus, within
the Great Barrier Reef and the Coral Sea. Evolution 55:2263–2273.
Posada, D., and K. A. Crandall. 1998. MODELTEST: testing the model of
DNA substitution. Bioinformatics 14:817–818.
Pritchard, J. K., M. Stephens, and P. Donnelly. 2000. Inference of population
structure using multilocus genotype data. Genetics 155:945–959.
Puebla, O., E. Bermingham, F. Guichard, and E. Whiteman. 2007. Colour
pattern as a single trait driving speciation in Hypoplectrus coral reef
fishes? Proc. R. Soc. Lond. B 274:1265–1271.
Phylogeography, historical demography, and the role of post-settlement
Randall, J. E. 1961. A contribution to the biology of convict surgeonfish
of the Hawaiian islands, Acanthurus triostegus sandwicensis. Pac. Sci.
———. 1998. Zoogeography of shore fishes of the Indo-Pacific region. Zool.
Randall, J. E., and H. E. Randall. 2001. Dascyllus auripinnis, a new pomacen-
trid fish from atolls of the central Pacific Ocean. Zool. Stud. 40:61–67.
Renema, W., D. R. Bellwood, J. C. Braga, K. Bromfield, R. Hall, K. G.
Johnson, P. Lunt, C. P. Meyer, L. B. McMonagle, R. J. Morley, et al.
2008. Hopping hotspots: global shifts in marine biodiversity. Science
Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223–225.
Riginos, C., and B. C. Victor. 2001. Larval spatial distributions and other
Pacific blennioid fishes. Proc. R. Soc. Lond. B 268:1931–1936.
Robertson, D. R., F. Karg, R. L. de Moura, B. C. Victor, and G. Bernardi.
2006. Mechanisms of speciation and faunal enrichment in Atlantic par-
rotfishes. Mol. Phylogenet. Evol. 40:795–807.
Rocha, L. A. 2004. Mitochondrial DNA and color pattern variation in three
western Atlantic Halichoeres (labridae), with the revalidation of two
species. Copeia 2004:770–782.
Rocha, L. A., D. R. Robertson, C. R. Rocha, J. L. Van Tassell, M. T. Craig,
and B. W. Bowen. 2005a. Recent invasion of the tropical Atlantic by an
Indo-Pacific coral reef fish. Mol. Ecol. 14:3921–3928.
speciation in tropical reef fishes. Proc. R. Soc. Lond. B 272:573–579.
DNA polymorphism analyses by the coalescent and other methods.
Rundle, H. D., and P. Nosil. 2005. Ecological speciation. Ecol. Lett. 8:336–
R¨ uppell, W. P. E. S. 1828–30. Atlas zu der Reise im n¨ ordlichen Africa. Fische
des Rothen Meeres. Fische Rothen Meeres 1:141–143.
Sambrook, J., E. F. Fritschi, and T. Maniatis. 1989. Molecular cloning: a
laboratory manual. Cold Spring Harbor Laboratory Press, New York.
Schluter, D. 2000. The ecology of adaptive radiation. Oxford Univ. Press,
and the population genetics of Caribbean reef fishes. Evolution 49:897–
Streelman, J. T., M. Alfaro, M. W. Westneat, D. R. Bellwood, and S. A. Karl.
2002. Evolutionary history of the parrotfishes: biogeography, ecomor-
phology, and comparative diversity. Evolution 56:961–971.
Thresher, R. E. 1984. Reproduction in reef fishes. TFH Publishing, NJ.
Van Herwerden, L., and P. J. Doherty. 2006. Contrasting genetic structures
across two hybrid zones of a tropical reef fish, Acanthochromis polya-
canthus (Bleeker 1855). J. Evol. Biol. 19:239–252.
for molluscan speciation and extinction. Evolution 41:1046–1058.
Wares, J. P., and C. W. Cunningham. 2001. Phylogeography and his-
torical ecology of the north Atlantic intertidal. Evolution 55:2455–
Wellington, G. M., and B. C. Victor. 1989. Planktonic larval duration of 100
Wu, C. I., and C. T. Ting. 2004. Genes and speciation. Nat. Rev. Genet.
Yaakub, S. M., D. R. Bellwood, L. van Herwerden, and F. A. Walsh. 2006.
Hybridization in coral reef fishes: introgression and bi-directional gene
exchange in Thalassoma (family Labridae). Mol. Phylogenet. Evol.
of large biological sequence datasets under the maximum likelihood
criterion. Ph.D. dissertation, The Univ. of Texas at Austin, TX.
Associate Editor: M. Hellberg
EVOLUTION MAY 2010
MATTHIEU LERAY ET AL.
The following supporting information is available for this article:
Figure S1. Phylogenetic relationship between 563 individuals sampled in the D. trimaculatus species complex based on mitochon-
drial control regions using the Maximum Likelihood method. A simplified tree is provided in the inset with size of pie charts being
proportional to the number of individuals.
Table S1. Number of alleles (Na), expected heterozygosity (He), and observed heterozygosity (Ho) across 13 loci and six groups.
Supporting Information may be found in the online version of this article.
Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting information supplied by the
authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
1230 EVOLUTION MAY 2010