Content uploaded by Jean-Paul Adrian Hobbs
Author content
All content in this area was uploaded by Jean-Paul Adrian Hobbs on Feb 23, 2015
Content may be subject to copyright.
Coral Reefs (2007) 26:841–850
DOI 10.1007/s00338-007-0273-3
123
REPORT
Hybridization of reef Wshes at the Indo-PaciWc biogeographic
barrier: a case study
A. D. Marie · L. van Herwerden · J. H. Choat ·
J-P. A. Hobbs
Received: 1 October 2006 / Accepted: 21 June 2007 / Published online: 17 July 2007
© Springer-Verlag 2007
Abstract Hybridization is recognized as an important
source of genetic variation. In some reef Wshes, including
the Acanthuridae, hybridization has been detected due to
intermediate colouration. This study used a molecular
genetic approach to investigate hybridization in two
Acanthurid species: Acanthurus leucosternon and Acanthu-
rus nigricans, which have Indian and PaciWc Ocean distri-
butions respectively and are sympatric in the eastern Indian
Ocean. In this area a putatitve hybrid, Acanthurus cf. leuco-
sternon has been recognized based on intermediate colour-
ation and restriction to the sympatric region of otherwise
allopatric putative parental species. This study aimed to test
this hypothesis using genetic tools. The three species were
sampled from Cocos (Keeling) and Christmas Islands, the
biogeographic boundary where many Indian and PaciWc
Ocean biota meet. Representatives from allopatric popula-
tions of both parental species and outgroups were also
sampled. Mitochondrial COI and intron 1 of the nuclear
ribosomal protein S7 were sequenced from 13 and 30 spec-
imens respectively. Although sample sizes in this study are
relatively small and more genetic data, including an
extended phylogeographic sampling, is required to further
evaluate these Wndings, the COI results support hybrid ori-
gins of Acanthurus cf. leucosternon, but S7 data are incon-
clusive due to the possibility of incomplete lineage sorting.
The fourfold more abundant Acanthurus nigricans is most
often the maternal parent. Inter-fertile hybrids apparently
backcross with rare Acanthurus leucosternon males, trans-
ferring Acanthurus nigricans mitochondria to this species.
These results suggest that Acanthurus leucosternon may
eventually be lost from these islands, due to their relative
rarity and introgressive hybridization.
Keywords Acanthuridae · Hybridization · Introgression ·
COI and S7 intron 1 · Biogeography · Mating behaviour
Introduction
Coral reefs support a great diversity of Wshes, which are
equally diverse in colouration. The responsible mechanisms
for this diversity are poorly understood, particularly as
there are few barriers to dispersal in the marine environ-
ment (McMillan et al. 1999; Bernardi et al. 2002). The
identiWcation of many closely related coral reef Wshes relies
on colour diVerences (Lieske and Meyers 1999) as they are
often morphologically and meristically indistinguishable
(Jansson et al. 1991; Wilkins et al. 1994). However, many
species have multiple colour morphs (Taylor and Hellberg
2003; Messmer et al. 2005). The relationship between
colour pattern and genetic divergence has been identiWed
in several studies (for example McMillan et al. 1999;
Communicated by Biology Editor M. van Oppen.
A. D. Marie · L. van Herwerden · J. H. Choat · J-P. A. Hobbs
Molecular Ecology and Evolution Laboratory,
James Cook University, Townsville, QLD 4811, Australia
L. van Herwerden (&)
Discipline Marine Biology and Aquaculture,
School of Marine and Tropical Biology,
James Cook University, Townsville, QLD 4811, Australia
e-mail: lynne.vanherwerden@jcu.edu.au
J-P. A. Hobbs
ARC Centre of Excellence for Coral Reef Studies,
James Cook University, Townsville, QLD 4811, Australia
A. D. Marie
Department of Biology, University of Sherbrooke,
Sherbrooke, QC, Canada J1K 2W5
842 Coral Reefs (2007) 26:841–850
123
Messmer et al. 2005). Moreover, there are cryptic species,
where no colour or meristic diVerences are apparent, but
populations are genetically diVerentiated (e.g. Dascyllus;
Bernardi et al. 2002). This is further complicated by the
fact that genetic diVerentiation has been identiWed between
some colour morphs, but not others, both within and
between populations. For example, the spiny damselWsh,
Acanthochromis polyacanthus, has diVerent colour morphs
on the same reef but these cannot be distinguished geneti-
cally (Planes and Doherty 1997; van Herwerden and Doh-
erty 2006). The contrary was found for the same species on
other reefs (van Herwerden and Doherty 2006). In addition,
it appears that same colour morphs exhibit genetic diVer-
ences between isolated locations (van Herwerden and Doh-
erty 2006). Contrary to the spiny damselWsh, at a range of
locations in the Caribbean, the hamlets of the genus Hypo-
plectrus consists of many colour morphs, but these are
genetically indistinguishable (Ramon et al. 2003; Garcia-
Machado et al. 2004).
It is therefore important to use genetic approaches to
identify whether there is gene Xow between populations
and/or colour morphs (Knowlton 2000). The genetic identi-
Wcation of species can however, be confounded by gene
Xow between them due to hybridization or incomplete line-
age sorting. Hybrids are individuals that are the oVspring of
dissimilar parents, which are distinguishable by one or
more heritable characters (Harrison 1993). Hybrids gener-
ally occur in hybrid zones which are geographical regions
in which previously isolated populations/species have come
into contact and inter-bred (Harrison 1993). Such inter-
breeding is favoured by factors such as abundance diVer-
ences between the participant species (Arnold 1997). There
are many examples in the terrestrial environment of
diverged species hybridizing when they come into second-
ary contact at biogeographic borders (see review by Mallet
2005). Biogeographic borders also exist in the marine envi-
ronment (Williams and Benzie 1998; Lessios et al. 1999;
Bernardi et al. 2001), however, due to a lack of studies it is
unclear whether secondary contact at biogeographic bound-
aries is a typical cause of hybridization in coral reef Wshes.
The use of molecular genetic analyses is particularly
helpful in the study of hybridization and has been applied to
a number of marine species (e.g. Brown 1995; Gardner
1996), especially corals (see references in van Oppen and
Gates 2006). The Atlantic cod, Gadus morhua, which has
been the subject of population genetic studies for several
decades (Ruzzante et al. 2000; Hutchinson et al. 2001) pro-
vides evidence of mixing between diVerentiated popula-
tions of Atlantic cod in a temperate marine hybrid zone
(e.g. Ruzzante et al. 2000). An increasing number of studies
on coral reef Wshes have further improved our understand-
ing of hybridization in the marine environment, for exam-
ple in Siganus (Lacson and Nelson 1993), several
damselWsh genera (Lacson 1994; Lacson and Clark 1995,
van Herwerden and Doherty 2006), Chaetodon (McMillan
et al. 1999), serranids (van Herwerden et al. 2006) and
wrasses (Yaakub et al. 2006, 2007), but none of these stud-
ies have examined hybridization at a signiWcant marine
biogeographic boundary.
In this study, three morphologically similar Wsh “species”
that have distinct colours were characterised genetically:
Acanthurus nigricans (Linnaeus 1758), the whitecheek
surgeonWsh, Acanthurus leucosternon (Bennett 1833), the
powderblue surgeonWsh and Acanthurus cf. leucosternon, a
putative hybrid of the afore-mentioned species. These spe-
cies are Perciformes in the class Actinopterygii. Acanthurus
nigricans and Acanthurus leucosternon are widespread in
the PaciWc and Indian Oceans respectively (Fig. 1). These
allopatric species are separated by the east Indian Ocean
biogeographic barrier and their ranges overlap at Christmas
and Cocos (Keeling) Islands (Fig. 1) where they are sympat-
ric. These sympatric populations are in contact, presumably
due to range expansions by each species from their respec-
tive locations. Thus, gene Xow between these species is the-
oretically possible and hybridization may therefore occur.
In order to test the hypothesis that hybridization has pro-
duced the
Acanthurus cf leucosternon type, genetic analy-
ses were performed using both mitochondrial and nuclear
DNA markers. The mtDNA will permit identiWcation of the
maternal contributor and the nuclear marker will identify
both parental contributions in Wrst generation (F1) hybrids,
if complete lineage sorting has been achieved. Thus, the
aims of this study were (1) to determine the status of the
hybrid; (2) to identify if there is a single or multiple mater-
nal parent(s) and (3) to examine the ecological factors that
Fig. 1 Christmas (indicated by a closed square) and Cocos (Keeling)
(indicated by a closed circle) Islands in the east Indian Ocean are situ-
ated 900 km apart. The ocean basin within which each species occurs
is shown on the plate by dashed, dotted and solid lines, respectively.
The zone of sympatry sampled in this study is indicated by an opaque
rectangle
Coral Reefs (2007) 26:841–850 843
123
may promote hybridization between these two species that
come into contact in the vicinity of a known marine biogeo-
graphic boundary.
Materials and methods
Sampling procedures
Cocos (Keeling) and Christmas Island are approximately
900 km apart, situated at 12°10⬘S, 96°52⬘E and 10°30⬘S,
105°40⬘E respectively (Fig. 1). These islands are situated at
the biogeographic break between Indian and PaciWc Ocean
species (Williams and Benzie 1998). Individuals whose
colour patterns were intermediate to the putative parent
species were labelled as “putative hybrids” (also known as
Acanthurus cf. leucosternon) (see Table 1). There was an
abundance of additional possible putative hybrids that
appeared similar in colour to one parent, but contained
some minor colouration from the other putative parent.
Numbers of specimens of the three species sampled at
Christmas and Cocos (Keeling) Islands are given in
Table 2. Additional specimens from allopatric populations
of the putative parental species were obtained, three from
west Australian Scott Reef (Acanthurus nigricans) in the
east Indian Ocean and one from the Seychelles in the west
Indian Ocean (Acanthurus leucosternon). Two specimens
each of Naso lituratus (from the Marquesas, PaciWc Ocean
and from west Australian Scott Reef) and N. elegans (Sey-
chelles) were also obtained for outgroup rooting purposes
(see Klanten et al. 2004), as per Table 2. Fish were col-
lected using spearWshing equipment in May 2004 and June
2006. Tissues (Wn clips) were preserved in 80% ethanol.
Laboratory procedures
DNA extraction procedures followed Sambrook and Russell
(2001). A PCR cocktail was made to a Wnal concentration
Table 1 Relative abundance, ecological and morphological aspects of Acanthurus nigricans, the putative hybrid and Acanthurus leucosternon as
per Kuiter and Debelius (2001)
a
Putative hybrids sampled were all F1s, based on intermediate colouration, however a range of putative post-F1 hybrids, not sampled here, were
also observed at much greater abundance than the F1s and A. leucosternon (J. P. Hobbs, personal observation)
Relative abundance A. nigricans Putative hybrid A. leucosternon
16-fold Onefold F1, eightfold
(F1 and post-F1)
a
Twofold
Ecological strategies
Habitat Hard substrate areas and seaward
reefs from the lower surge zone
Around rocky reefs subjected
to strong currents
Reef Xats and along upper
seaward slopes
Mating strategy Monogamous No data Monogamous
Clustering strategy Solitary or in group Solitary or in small group Solitary or in a large feeding
aggregation
Diet Filamentous algae No data Benthic algae
Morphological aspects
Maximal size 21 cm 20 cm 23 cm
Number of dorsal spines 9–9 No data 9–9
Number of dorsal soft rays 28–31 No data 28–30
Number of anal spines 3–3 No data 3–3
Number of anal soft rays 26–28 No data 23–26
Number of gill rakers on
the anterior row
17–19 No data 17–19
Number of gill rakers on
the posterior row
18–20 No data 18–20
Colouration
Below the eye A horizontally elongated white spot A white spot No distinct white spot
Pectoral Wn Black and blue Blue with yellow margins Yellow
Dorsal Wn Black with blue margins and yellow base Yellow with white margins
and black submarginal
Yellow with white margins
and black submarginal
Ventral and anal Wn Blue with yellow base and white margins Blue with yellow base
and white margins
Blue with white margins
Caudal peduncle Yellow surrounded by blue Yellow surrounded by blue Yellow surrounded by yellow
Caudal Wn Sub-marginal yellow band Sub-marginal yellow band Sub-marginal black band
844 Coral Reefs (2007) 26:841–850
123
of 2.5 mM MgCl
2
, 1 £ BuVer (10mM Tris–HCL, 5mM
KCL, pH 8.3), 0.2 mM deoxynucleotide triphosphates,
10 M each primer and 0.75 units Taq polymerase. Two
markers, both of which diVerentiate between congeneric
species (e.g. Read et al. 2006; Bernardi et al. 2004) and
both of which have been previously used to identify hybrid-
ization in coral reef Wsh (Yaakub et al. 2006) were used
in this study. The cytochrome oxidase I (COI) region of
the mitochondrial genome was ampliWed using universal
primers LCO 1490, 5⬘-GGTCAACAAATCATAAAGAT
ATTGG-3⬘ and HCO 2198, 5⬘-TAAACTTCAGGGTGA
CCAAAAAATCA-3⬘ (Folmer et al. 1994). The Wrst intron
of the nuclear ribosomal S7 protein was also ampliWed
using universal primers S7RPEX1F, 5⬘-TGGCCTCTTCC
TTGGCCGTC-3⬘ and S7RPEX2R, 5⬘-AACTCGTCTGG
CTTTTCGCC-3⬘(Chow and Hazama 1998). For both
markers, PCR conditions varied among samples, as diVer-
ent annealing temperatures and MgCl
2
concentrations were
required. COI ampliWcations were performed at an anneal-
ing temperature of 48°C for 35 cycles or at 50°C for 5
cycles, followed by 30 cycles at 48°C. S7 ampliWcations
were either performed at an annealing temperature of 50°C
or using the same touchdown procedure described for the
COI ampliWcation. All PCR conditions included an initial
denaturation step at 94°C for 2 min, followed by 35 cycles
of PCR at 94°C for 30 s, annealing at the speciWed tempera-
ture for 30 s and an extension at 72°C for 1 min and 30 s.
A Wnal extension was done for 10 min to complete the PCR
procedure. PCR products were puriWed using either isopro-
panol precipitation (as per Sambrook and Russell 2001) or
using a QIAquick gel extraction procedure in the event of
multiple PCR products being produced, by following man-
ufacturer’s instructions (Qiagen).
The samples were sent to Macrogen (Seoul, Republic of
Korea) for sequencing on an ABI 377 sequencer in both
directions. DNA sequences were inspected individually
for quality and then spliced together using the computer
program BioEdit Sequence Alignment Editor (Hall 1999).
Sequences were aligned using ClustalW from within Bio-
Edit and alignments were checked and modiWed by hand
where necessary. Aligned data was exported as a NEXUS
Wle, and used for phylogenetic analyses to determine the
relationships between individuals.
Analytical procedures
Sequences have been submitted to Genbank with accession
numbers EF648221 to EF648275. To identify the evolu-
tionary relationships between samples, neighbour joining,
(NJ) and maximum parsimony (MP) methods were used.
Phylogenetic analyses were conducted using MEGA
version 3.1 (Molecular Evolutionary Genetics Analysis)
(Kumar et al. 2004) and PAUP* version 4 (SwoVord 1999).
A likelihood approach was used in Modeltest version 3.06
(Posada and Crandall 1998) to Wnd the best substitution
model for NJ analyses. One thousand bootstrap replicates
were performed during the NJ analysis for evaluation of the
level of support for the tree topology. Both complete and
pairwise deletions of gaps and missing data were used for
S7 sequences. Only lineages supported by bootstrap values
greater than 50% were retained when majority rule boot-
strap consensus trees were constructed.
The haplotype diversity, h, was calculated (as per Nei
1987) for COI sequences and nucleotide diversities, , were
calculated for both S7 and COI data as implemented in the
software program Arlequin, version 3.1 (ExcoYer et al.
2005). A minimum spanning tree was constructed for COI
haplotypes using Arlequin version 3.1 (ExcoYer et al.
2005) in order to identify the relationships between haplo-
types.
Results
COI
Sequence data was obtained from Wve Acanthurus leuco-
sternon, Wve Acanthurus nigricans and three putative
hybrids. The nucleotide frequency was: A = 26.22%,
C = 25.17%, G = 17.07% and T = 31.54%. There were 14
Table 2 Material examined with collection locations, number of sam-
ples (n) from each location and marker used for each species from each
location
Genus Species Location n Primer
used
A
canthurus leucosternon Christmas Island 4 COI
A
canthurus leucosternon Cocos Island 1 COI
A
canthurus cf. leucosternon Christmas Island 3 COI
A
canthurus nigricans Christmas Island 4 COI
A
canthurus nigricans Cocos Island 1 COI
N
aso elegans Seychelles 2 COI
N
aso lituratus Marquesas 2 COI
N
aso lituratus West Australia 2 COI
A
canthurus leucosternon Christmas Island 7 S7
A
canthurus leucosternon Cocos Island 4 S7
A
canthurus leucosternon Seychelles 1 S7
A
canthurus cf. leucosternon Christmas Island 7 S7
A
canthurus nigricans Christmas Island 7 S7
A
canthurus nigricans Cocos Island 1 S7
A
canthurus nigricans West Australia 3 S7
N
aso elegans Seychelles 2 S7
N
aso lituratus Marquesas 2 S7
N
aso lituratus West Australia 2 S7
Coral Reefs (2007) 26:841–850 845
123
transversions and 2 transitions. There were no gaps in the
partial COI sequence and when sequences were translated,
functional partial COI protein was obtained, as veriWed by a
Blast search of the GenBank sequence database. Of 686
bases, only 16 were variable, eight of which (1.17%) were
parsimony informative. The remaining eight variable sites
were singletons.
The best substitution model identiWed by the Akaike
Information Criterion was HKY + I + G, with I = 0.5418
and G = 0.3050. The MP and NJ trees had very similar
topologies, therefore MP bootstrap support values are indi-
cated on the NJ tree (Fig. 2). The only diVerence between
the two is in the placement of L4CI, which is part of the
Acanthurus nigricans clade in the MP, but distinct in the NJ
analysis. There were 190 equally likely MP trees, each with
a length of 16 steps. The consistency index (CI) and reten-
tion index (RI) were both one (Fig. 2). Samples separated
into two strongly supported clades, one for each species
(Figs. 2, 4). All eight parsimony informative sites contrib-
uted to the split between these two species clades. The
Acanthurus nigricans clade contained all Acanthurus nigri-
cans samples for which data was obtained as well as one or
two (analysis dependent) of the Wve Acanthurus leucoster-
non and two of the three putative hybrids for which COI
sequences were obtained. There were six substitutions in
this clade, all of which were transitions. Only one of the six
haplotypes in this clade was shared (Fig. 3). Four individu-
als, two of which were Acanthurus nigricans (N1 and N4),
one Acanthurus leucosternon (L1) and one hybrid (H2)
shared this haplotype, suggesting that these four individuals
are all descendants of an Acanthurus nigricans mother.
This clade also includes Wve other specimens, three of
Fig. 2 The outgroup rooted COI phylogram generated by NJ analysis.
B
ranch lengths show the number of substitutions between specimens.
Majority rule support values (>50%) obtained from 1,000 bootstrap
replicates are indicated above branches for NJ analysis and below
branches are support values for the 50% majority rule MP consensus
tree. Two clades are identiWed, an Acanthurus nigricans and an
A
canthurus leucosternon clade. Individuals are identiWed by an alpha-
numeric code, the numeric part of the code uniquely identiWes a spe-
ciWc Wsh while species and location are identiWed by the alphabetic part
of the code as speciWed in the key to the Wgure
Fig. 3 Relationships between hybrids (Acanthurus cf leucosternon),
A
canthurus leucosternon and Acanthurus nigricans haplotypes repre-
sented in an minimum spanning tree. Hybrid (H), A. leucosternon (L)
and A. nigricans (N) specimens from Christmas and Cocos (Keeling)
Islands are identiWed by diVerent Wlls, as indicated in the Wgure key.
The sizes of circles indicate the number of individuals sharing that par-
ticular haplotype. Crossbars on the line connecting haplotypes repre-
sent the number of substitutions separating them. The two clades are
encircled
A. nigricans from Christmas Island
A. leucosternon from Christmas Island
A. leucosternon from Cocos Island
A. nigricans from Cocos Island
A. cf. leucosternon from Christmas Island
Clades
One change
846 Coral Reefs (2007) 26:841–850
123
which are Acanthurus nigricans (N2, N3 and N5), one of
which is a hybrid (H3) and one Acanthurus leucosternon
(L4). The three Acanthurus nigricans specimens may be
purebred, since, like N1 and N4, they have Acanthurus nig-
ricans colouration. L4 is a descendant of either Acanthurus
leucosternon (e.g. L1) or a hybrid (e.g. H2), as it has
Acanthurus leucosternon colouration, despite having
Acanthurus nigricans mtDNA from its mother. Both puta-
tive hybrids, H2 and H3, have intermediate colouration,
but have Acanthurus nigricans mtDNA.
The Acanthurus leucosternon clade was strongly sup-
ported and contained the remaining three Acanthurus leu-
costernon individuals and one of the three putative hybrids.
No Acanthurus nigricans occurred in the Acanthurus leuco-
sternon clade. There were only two substitutions, one tran-
sition and one transversion. There was one shared
haplotype (shared by two Acanthurus leucosternon individ-
uals) and a total of three unique haplotypes in this clade.
The N. lituratus and N. elegans outgroups were also par-
titioned by species, as the N. elegans (NeSY) were strongly
supported as a distinct clade from all of the N. lituratus,
irrespective of location (NlMQ and NlWA).
The haplotype diversity, h, within clades was relatively
high (h =0.83 for both Acanthurus nigricans and Acanthu-
rus leucosternon clades), with low nucleotide diversities
( =0.19§ 0.15% and =0.14§0.14%) for each clade
respectively (Fig. 3
). Both clades contained individuals
from both locations (Christmas and Cocos Islands).
S7 intron 1
S7 intron 1 sequences were highly variable, with a rela-
tively conserved block of 352 bp at the 5⬘ end and 293 bp at
the 3⬘ end. Seventy-four base pairs were eliminated from
the sequence centre, as they were too variable and polymor-
phic to align among individuals with conWdence. Two hun-
dred and twenty eight sites were variable, of which 122
were parsimony informative. Nucleotide frequencies were:
A = 23.56%, C =19.30%, G = 25.05% and T =32.09%.
The optimal substitution model was GTR + G, G = 0.4982
and produced an NJ phylogram of broadly similar topology
to that of the MP tree. Therefore, an outgroup rooted NJ
phylogram is presented with support from 1,000 bootstrap
replicates and MP majority rule consensus support values
indicated (Fig. 4). There were 22 gaps, which were treated
as a complete deletion. Despite the reasonable number of
parsimony informative sites, there is little supported struc-
ture in this nuclear data (Fig. 4). Two major lineages were
resolved, one of which contains only one individual: The
only representative of the allopatric Acanthurus leucosternon
population (L3298SY) from the Seychelles, in the west
Indian Ocean, for which data was obtained. This is pro-
posed to be a west Indian Ocean Acanthurus leucosternon
lineage. The second clade contains all remaining samples of
both species from the east Indian Ocean, including the west
Australian Acanthurus nigricans (N26WA, N27WA and
N28WA). Two subclades were identiWed within this line-
age, neither of which was species-speciWc either. Putative
hybrids were dispersed throughout this mixed species
lineage. The MP majority rule consensus tree of 370
equally likely trees was 21 steps in length, with CI, RI
and rescaled consistency indices of 0.904762, 0.857143
and 0.77551 respectively. The same two lineages were
identiWed as before, but only one of the subclades was
resolved by the MP analysis (Fig. 4).
Like the ingroup, the N. lituratus and N. elegans out-
group sequences also failed to form species-speciWc lin-
eages, unlike the case for the mtDNA. The N. elegans
(NeSY) and N. lituratus from west Australia (NlWA)
formed a strongly supported clade, which excluded the con-
speciWc N. lituratus samples from the Marquesas in the
PaciWc Ocean (NlMQ).
Fig. 4 An outgroup rooted S7 intron 1 phylogram generated by NJ
analysis. Branch lengths show the number of substitutions between
specimens. NJ bootstrap support values (>50%) of 1,000 bootstrap rep-
licates are indicated above branches for NJ analysis and below branch-
es are support values for the 50% majority rule MP consensus tree.
Individuals are identiWed by an alphanumeric code, the numeric part o
f
the code uniquely identiWes a speciWc Wsh while species and location
are identiWed by the alphabetic part of the code as speciWed in the key
to the Wgure
Coral Reefs (2007) 26:841–850 847
123
Discussion
Relationship between colour pattern and genetic divergence
This study tested the hypothesis that Acanthurus cf. leuco-
sternon is a hybrid of Acanthurus leucosternon and
Acanthurus nigricans as suggested by meristics, colour pat-
tern and also by the fact that Acanthurus cf. leucosternon is
restricted to the area where its putative parents are sympat-
ric and the relative abundances of all three at these sites
(Table 1). The results of this study support hybridization,
since mtDNA of the parental species (and sister species of
the outgroup) are genetically diVerentiated, although there
is evidence of some mtDNA introgression from the most
abundant, Acanthurus nigricans into the much rarer
Acanthurus leucosternon. Furthermore, mtDNA from both
parental species is present in the putative hybrids, Acanthu-
rus cf leucosternon. However, due to incomplete lineage
sorting of the nuclear marker in both the study species (and
the sister species of the outgroups), conclusions from this
work require conWrmation from additional nuclear markers
which do not suVer from incomplete lineage sorting.
In another study, the presence of hybridization between
meristically indistinguishable Wsh species with diVerences
in colouration was tested using genetics (Yaakub et al.
2006) and conWrmed that the wrasses Thalassoma quinque-
vittatum and T. jansenii hybridize to produce the intermedi-
ate colour morphs. These wrasses are sympatric through
much of their distribution ranges, but hybrids have only
been found at an isolated coral reef in the Coral Sea, where
the species abundances are very diVerent (Yaakub et al.
2006). DiVerent relative abundances of hybridizing species
are considered one of the hallmarks of hybridization
(Arnold 1997). In contrast, seven colour morphs of the
genus Hypoplectrus (hamlet Wsh) that co-occur on the same
Caribbean reefs do not appear to be genetically distinguish-
able (e.g. Ramon et al. 2003), despite strong colour pattern-
based assortative mating. However, closer examination by
McCartney et al. (2003) identiWed a complex mosaic of
genetic diVerentiation among Hamlet colour morphospe-
cies, providing evidence for reproductive isolation at some,
but not other study sites. In the Hamlets, which are a very
young group of morphospecies, haplotype and microsatel-
lite genotype frequency diVerences among the morphospe-
cies were highly signiWcant at speciWc locations, despite a
lack of concordance between haplo-, geno- and morpho-
types (McCartney et al. 2003
). Another example is pre-
sented in the study of Pseudochromis fuscus, which has
more than six colour morphs at various locations. Genetic
analyses variably identiWed diVerences between morphs
(Messmer et al. 2005) suggesting that colour alone is not a
reliable indicator of distinct species. All these examples of
coral reef Wsh colour variants/species indicate that descriptions
based on meristics and colour patterns alone, may not be
suYcient to resolve the question of whether species are dis-
tinct, whether they hybridize or are simply colour variants.
Moreover, such studies do not provide information on the
genetic contribution of each putative parent, which is an
essential factor to understand what ecological and behavio-
ural conditions permit hybridization amongst coral reef
Wshes.
Putative parental contributors
MtDNA
This study identiWed two mtDNA clades, one containing all
Acanthurus nigricans samples and a few Acanthurus leuco-
sternon individuals, the other containing only Acanthurus
leucosternon and no Acanthurus nigricans individuals.
Putative hybrids were present in both clades.
This data suggests that Acanthurus nigricans is the
mother in hybrid matings with Acanthurus leucosternon
more often than not, but that Acanthurus leucosternon
females can also mate with Acanthurus nigricans males or
hybrids. This may be due to Acanthurus nigricans being at
least eightfold more abundant than Acanthurus leucoster-
non and twofold more abundant than putative hybrids at the
study site (Table 1). In several other studies of reef Wsh
hybridization, only one of the parents has acted exclusively
as the mother during hybridization (e.g. coral trout, van
Herwerden et al. 2006; wrasses, Yaakub et al. 2006; the
damselWsh Acanthochromis polyacanthus from the “south-
ern hybrid zone”, van Herwerden and Doherty 2006). How-
ever, like the Acanthurid species presently studied, some
species of butterX
yWsh (McMillan et al. 1999) and some
morphs of Acanthochromis polyacanthus from the “north-
ern hybrid zone” (van Herwerden and Doherty 2006) can
hybridize using either species or colour morph as the
hybridizing mother.
Incomplete lineage sorting is an alternative interpreta-
tion for the lack of absolute partitioning between species at
the mtDNA marker (e.g. see van Herwerden et al. 2006).
However, this is much less likely, given: (1) the relative
abundances in the contact zone, (2) meristics, (3) colour-
ation, (4) biogeography (see below), (5) ecology of these
species at the study sites (see below), (6) resolution
obtained with this marker for the outgroup sister species
and (7) that COI has been successfully and extensively used
to diVerentiate congeneric species of tropical coral reef Wsh
in phylogenetic studies (e.g. Bernardi et al. 2004; Read
et al. 2006, Yaakub et al. 2007). Additional more compre-
hensive phylogeographic data based on larger sample sizes
is required to discriminate between these alternative
hypotheses: hybridization vs. incomplete lineage sorting.
Partitioning of the lineages from allopatric populations of
848 Coral Reefs (2007) 26:841–850
123
these two species into the two clades identiWed in this study
would support hybridization, whilst a similarly inter-mixed
genetic signal from distant allopatric populations to that
observed here in the zone of sympatry, may favour incom-
plete lineage sorting (see McMillan et al. 1999; van Her-
werden et al. 2006).
Nuclear DNA
The nuclear data suggests either incomplete lineage sorting
or introgressive hybridization, since there is no partitioning
of this marker in the parental species sampled within the
east Indian Ocean. The allopatric west Indian Ocean
Acanthurus leucosternon does however appear to be genet-
ically distinct, although this is only based on a single indi-
vidual and requires additional data for conWrmation.
Support for incomplete lineage sorting (rather than hybrid-
ization) at this marker is provided by the genetic structure
within the outgroup, since well-established sister-species
(Klanten et al. 2004) are not diVerentiated by this marker,
but share a well-supported lineage. This is in marked con-
trast to the clear genetic structure revealed for these out-
group species using the mtDNA marker.
Hybridization, introgression, ecology and evolution
Hybridization can be facilitated when parental species,
which have similar ecological and morphological traits,
come into contact, especially if there is a numerical imbal-
ance and one of the species is rare relative to the other, as is
the case for Acanthurus nigricans and Acanthurus leuco-
sternon at Christmas and Cocos (Keeling) Islands.
Acanthurus nigricans is generally found on hard substrates
and seaward reefs from the lower surge zone, whilst
Acanthurus leucosternon is generally found on reef Xats
and along upper seaward slopes (Kuiter and Debelius
2001). In the contact zone studied here, these species (and
the putative hybrids) were all found in shallow water (1–
4 m), feeding in the same area (J. P. Hobbs, personal obser-
vation). Furthermore, Acanthurus leucosternon was rare
relative to the putative hybrids (F1 and post-F1) and
Acanthurus nigricans, thereby providing an incentive for
inter-breeding by Acanthurus leucosternon. Although, mat-
ing behaviour of these species has not been documented to
date, it is noted that most of the putative hybrids are post-
F1, which is eightfold more abundant than F1 hybrids, sug-
gesting that de novo F1 hybrids are generated or survive to
adulthood less frequently than do subsequent generations of
hybrids.
It is possible that hybridization will re-orient evolution
(at this location at least), because eventually the already rel-
atively rare Acanthurus leucosternon may disappear from
this location (as per Rhymer and SimberloV 1996; Grant
et al. 2005). Furthermore, hybrid oVspring may disperse
from the hybrid zone leading to more widespread introgres-
sion as recorded for hybridizing butterXyWshes (McMillan
et al. 1999). An alternative, but less likely consequence of
hybridization may be the merging of both parental species
due to hybridization (“reverse speciation”) as has recently
been documented for a pair of lake inhabiting sticklebacks
following the introduction of an exotic crayWsh (Taylor
et al.
2006). However, additional data is required from
additional nuclear markers that do not suVer from incom-
plete lineage sorting, to further evaluate this.
Little is known of the Wtness of reef Wsh hybrids com-
pared to their parents, but the hybrids are often conWned to
the hybrid zone, suggesting that they are less Wt than both
parent species outside of the hybrid zone (e.g. wrasses in
Yaakub et al. 2006; damselWsh in van Herwerden and Doh-
erty 2006; Acanthurids in this study), but see McMillan
et al. (1999) for a case in contrast, where butterXyWsh
mtDNA introgression extends across nearly 12,000 km.
Some authors aYrm that hybrids are as Wt or Wtter than their
parents in the hybrid zone at least, which is often intermedi-
ate to the respective parental habitats (e.g. Arnold and Hod-
ges 1995). This suggests that hybrids can contribute to
adaptation in spite of the genetic barriers (Barton 2001).
The present study suggests that hybrids in the hybrid zone
are at least as Wt as the fourfold rarer parent, Acanthurus
leucosternon. Other authors suggest that the Wtness of
hybrids relative to parents can be reduced (e.g. Barton and
Hewitt 1985). Hybrid abundance is diYcult to evaluate in
the natural environment, because unlike F1 hybrids, post-
F1 hybrids may have colouration that is diYcult to distin-
guish from the parental colouration. This study, like the
Yaakub et al. (2006) and van Herwerden et al. (2006) stud-
ies, has shown that reef Wsh hybrids can be much more
abundant than is apparent from colouration alone. This may
eventually lead to the loss of the rarer parental species from
the hybrid zone. Only through genetic analyses can the
presence of such introgressive hybridization be determined.
Finally, the observed introgressive hybridization sug-
gests that there is no assortative mating behaviour, at least
not in the rare Acanthurus leucosternon or the F1 hybrids,
nor prezygotic isolation or habitat partitioning between spe-
cies in the hybrid zone, since these species co-occur and
probably spawn concurrently. Hybridization in these spe-
cies appears to be due to secondary contact at a biogeo-
graphic border and the absence of prezygotic isolating
mechanisms (behavioural and gamete compatibility). A
behavioural study is necessary to determine if hybridization
between the parental species is due to active behavioural
responses of mating individuals of both species, sneak mat-
ing by the rarer Acanthurus leucosternon males or acciden-
tal fertilization by compatible gametes. Data from other
hybridizing coral reef Wsh species suggest that sneak mating
Coral Reefs (2007) 26:841–850 849
123
by males of the numerically rare species is likely to be more
common (Frisch and van Herwerden 2006; Yaakub et al.
2006). Regardless, this study identiWes a deWnite risk that
Acanthurus leucosternon may go locally extinct as a
consequence of its relative rarity and possibly extensive
introgressive hybridization at the Cocos (Keeling) and
Christmas Islands (as per Rhymer and SimberloV 1996).
This may ensure that these species are maintained in allopa-
try, however, if hybrid larvae successfully disperse from
the hybrid zone, then introgression may be more wide-
spread and eventually result in the merging of two species
into one, as has been documented for some sticklebacks
(Taylor et al. 2006). Given the relative isolation of these
islands, it is unlikely that hybrids will populate and “pol-
lute” distant purebred stocks of the parental species, e.g.
Acanthurus leucosternon at the Seychelles, rather, they are
likely to remain conWned to the hybrid zone and adjacent
reefs, e.g. Indonesia and Reefs oV the northwest Australian
coast. Further genetic analyses, involving widespread phy-
logeographic sampling is necessary to determine the geo-
graphic extent of introgression in these species.
Acknowledgements Our thanks to: Justin Gilligan, Jay Hender and
Dr Robertson, who provided Weld assistance and helped in collecting
the specimens. Thanks also to Parks Australia Christmas Island, John
Clunies-Ross and Geof Christie for providing valuable logistical sup-
port. We also thank Selma Klanten for guidance in the laboratory and
COI sequences for the outgroups. Thanks are further due to Claire
Farnsworth and Kate Winters for assistance with labwork. This work
was funded by James Cook University funds awarded to JH Choat, GP
Jones, P Munday and D Jerry.
References
Arnold ML (1997) Natural hybridization and evolution. Oxford Uni-
versity Press, New York
Arnold ML, Hodges SA (1995) Are natural hybrids Wt or unWt relative
to their parents? Trends Ecol Evol 10:67–71
Barton NH (2001) The role of hybridization in evolution. Mol Ecol
10:551–568
Barton NH, Hewitt GM (1985) Analysis of hybrid zones. Annu Rev
Ecol Syst 16:113–148
Bernardi G, Holbrook SJ, Schmitt RJ (2001) Gene Xow at three spatial
scales in a coral reef Wsh, the three-spot dascyllus, Dascyllus
trimaculatus. Mar Biol 138:457–465
Bernardi G, Holbrook SJ, Schmitt RJ, Crane NL, DeMartini E (2002)
Species boundaries, populations and colour morphs in the coral
reef three-spot damselWsh (Dascyllus trimaculatus) species com-
plex. Proc R Soc B Biol Sci 269:599–605
Bernardi G, Bucciarelli G, Costagliola D, Robertson DR, Heiser JB
(2004) Evolution of coral reef Wsh Thalassoma spp. (Labridae). 1.
Molecular phylogeny and biogeography. Mar Biol 144:369–375
Brown LD (1995) Genetic evidence for hybridization between Haliotis
rubra and H. laevigata. Mar Biol 123:89–93
Chow S, Hazama K (1998) Universal PCR primers for S7 ribosomal
protein gene introns in Wsh. Mol Ecol 7:1255–1256
ExcoYer L, Laval G, Schneider S (2005) Arlequin ver. 3.0: an inte-
grated software package for population genetics data analysis.
Evol Bioinformatics Online 1:47–50
Folmer O, Black M, Hoen W, Lutz R, Vrijenhoek R (1994) DNA prim-
ers for ampliWcation of mitochondrial cytochrome c oxidase
subunit I from diverse metazoan invertebrates. Mol Mar Biol
Biotechnol 3:294–299
Frisch A, van Herwerden L (2006) Field and experimental studies of
hybridization between coral trouts, Plectropomus leopardus and
Plectropomus maculatus (Serranidae), on the Great Barrier Reef,
Australia. J Fish Biol 68:1013–1025
Garcia-Machado E, Chevalier Monteagudo P, Solignac M (2004) Lack
of mtDNA diVerentiation among hamlets (Hypoplectrus, Serrani-
dae). Mar Biol 144:147–152
Gardner JPA (1996) The Mytilus edulis species complex in southwest
England: eVects of hybridization and introgression upon interlo-
cus associations and morphometric variation. Mar Biol 125:385–
399
Grant PR, Grant BR, Petren K (2005) Hybridization in the recent past.
Am Nat 166:56–67
Hall TA (1999) Bioedit: a user-friendly biological sequence alignment
editor and analysis program for Windows 95/98/NT. Nucleic
Acids Symp Ser 41:95–98
Harrison RG (1993) Hybrid zones and the evolutionary process.
Oxford University Press, Oxford
Hutchinson WF, Carvalho GR, Rogers SI (2001) Marked genetic
structuring in localised spawning populations of cod (Gadus
morhua) within the North Sea and adjoining waters as revealed by
microsatellites. Mar Ecol Prog Ser 223:251–260
Jansson H, Holmgren I, Wedin K, Andersson T (1991) High frequency
of natural hybrids between Atlantic salmon, Salmo salar L., and
brown trout, S. truta L., in a Swedish river. J Fish Biol 39:343–
348
Klanten S, van Herwerden L, Choat JH, Blair D (2004) Patterns of line-
age diversi
Wcation in the genus Naso (Acanthuridae). Mol Phylo-
genet Evol 32:221–235
Knowlton N (2000) Molecular genetic analyses of species boundaries
in the sea. Hydrobiologia 420:73–90
Kuiter RH, Debelius H (2001) SurgeonWshes, rabbitWshes and their rel-
atives. A comprehensive guide to Acanthuroidei (marine Wsh
families). TMC Publishing, Chorleywood
Kumar S, Tamura K, Nei M (2004) MEGA3: integrated software for
molecular evolutionary genetics analyses and sequence align-
ment. BrieWngs in Bioinformatics 5:150–163
Lacson JM (1994) Fixed allele frequency diVerences among Palauan
and Okinawan populations of the damselWshes Chrysiptera cya-
nea and Pomacentrus coelestis. Mar Biol 118:359–365
Lacson JM Clark S (1995) Genetic divergence of Maldivian and
Micronesian demes of the damselWshes Stegastes nigricans,
Crysiptera biocellata, C. glauca and C. leucopoma (Pomacentridae).
Mar Biol 121:585–590
Lacson JM, Nelson SG (1993) Genetic distances among Wshes of the
genus Siganus (Siganidae) from the western PaciWc Ocean. Mar
Biol 116:187–192
Lessios HA, Kessing BD, Robertson DR, Paulay G (1999) Phylogeog-
raphy of the pantropical sea urchin Eucidaris in relation to land
barriers and ocean currents. Evolution 53:806–817
Lieske E, Meyers R (1999) Coral reef Wshes. Princeton University
Press, Princeton
Mallet J (2005) Hybridization as an invasion of the genome. Trends
Ecol Evol 20:229–237
McCartney MA, Acevedo J, Heredia C, Rico C, Quenoville B, Ber-
mingham E, McMillan O (2003) Genetic mosaic in a marine
species Xock. Mol Ecol 12:2963–2973
McMillan WO, Weight LA, Palumbi SR (1999) Color pattern evolu-
tion, assortative mating, and genetic diVerentiation in brightly
colored butterXyWshes (Chaetodontidae). Evolution 53:247–260
Messmer V, van Herwerden L, Munday P, Jones GP (2005) Phyloge-
ography of colour polymorphism in the coral reef Wsh Pseudochromis
850 Coral Reefs (2007) 26:841–850
123
fuscus, from Papua New Guinea and the Great Barrier Reef. Coral
Reefs 24:392–402
Nei M (1987) Molecular evolutionary genetics. Columbia University
Press, New York
Planes S, Doherty P (1997) Genetic and color interactions at a contact
zone of Acanthochromis polyacanthus—a marine Wsh lacking
pelagic larvae. Evolution 51:1232–1243
Posada D, Crandall KA (1998) Modeltest: testing the model of DNA
substitution. Bioinformatics 14:817–818
Ramon ML, Lobel S, Sorenson MD (2003) Lack of mitochondrial
genetic structure in hamlets (Hypoplectrus spp.); recent speciation
or ongoing hybridization. Mol Ecol 12:2975–2980
Read C, Bellwood D, van Herwerden L (2006) Ancient origins of Indo-
PaciWc coral reef Wsh biodiversity: a case study of the leopard
wrasses (Labridae: Macropharyngodon). Mol Phylogenet Evol
38:808–819
Rhymer JM, SimberloV D (1996) Extinction by hybridization and
introgression. Annu Rev Ecol Syst 27:83–109
Ruzzante DE, Taggart CT, Lang S, Cook D (2000) Mixed-stock anal-
ysis of Atlantic cod near the Gulf of St Lawrence based on micro-
satellite DNA. Ecol Appl 10:1090–1109
Sambrook J, Russell DW (2001) Molecular cloning: a laboratory man-
ual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor
SwoVord DL (1999) PAUP*: phylogenetic analysis using parsimony.
Sinauer Associates, Champaign
Taylor EB, Boughman JW, Groenenboom M, Sniatynski M, Schluter
D, Gow JL (2006) Speciation in reverse: morphological and
genetic evidence of the collapse of a three-spined stickleback
(Gasterosteus aculeatus) species pair. Mol Ecol 15:343–356
Taylor MS, Hellberg ME (2003) Genetic evidence for local retention
of pelagic larvae in a Carribbean reef Wsh. Science 299:107–109
van Herwerden L, Doherty P (2006) Contrasting genetic structures
across two hybrid zones of a tropical reef Wsh, I (Bleeker 1855).
J Evol Biol 19:239–252
van Herwerden L, Choat JH, Dudgeon CL, Carlos G, Newman SJ,
Frisch A, van Oppen M (2006) Contrasting patterns of genetic
structure in two species of the coral trout Plectropomus (Serrani-
dae) from east and west Australia: introgressive hybridization or
ancestral polymorphisms. Mol Phylogenet Evol 41:420–435
van Oppen MJH, Gates RD (2006) Conservation genetics and the resil-
ience of reef-building corals. Mol Ecol 15:3863–3883
Wilkins NP, Courtney HP, Gosling E, Linnane A, Jordan C, Curatolo
A (1994) Morphometric and meristic characters in salmon, Salmo
salar L., trout, Salmo truta L., and their hybrids. Aquaculture and
Fisheries Management 25:505–518
Williams` ST, Benzie JAH (1998) Evidence of a biogeographic break
between populations of a high dispersal starWsh: congruent
regions within the Indo-West PaciWc deWned by color morphs,
mtDNA, and allozyme data. Evolution 52:87–99
Yaakub SM, Bellwood DR, van Herwerden L, Walsh FM (2006)
Hybridization in coral reef Wshes: introgression and bi-directional
gene exchange in Thalassoma (family Labridae). Mol Phylogenet
Evol 40:84–100
Yaakub SM, Bellwood DR, van Herwerden L (2007) A rare hybridiza-
tion event in two common Caribbean wrasses (genus Halichoeres;
family Labridae). Coral Reefs. doi: 10.1007/s00338-007-0240-z
