Content uploaded by Matthieu Leray
Author content
All content in this area was uploaded by Matthieu Leray on Mar 19, 2015
Content may be subject to copyright.
Mar Biol (2009) 156:1595–1607
DOI 10.1007/s00227-009-1195-0
123
ORIGINAL PAPER
Complex patterns of population structure and recruitment
of Plectropomus leopardus (Pisces: Epinephelidae)
in the Indo-West PaciWc: implications for Wsheries management
Lynne Van Herwerden · J. Howard Choat ·
Stephen J. Newman · Matthieu Leray · Grethe Hillersøy
Received: 10 February 2009 / Accepted: 30 March 2009 / Published online: 16 April 2009
© Springer-Verlag 2009
Abstract Here the population genetic structure of an
ecologically and economically important coral reef Wsh, the
coral trout Plectropomus leopardus, is investigated in the
context of contemporary and historical events. Coral trout
were sampled from four regions (six locations) and partial
mtDNA D-loop sequences identiWed six populations
(Fst = 0.89209, P<0.0001): Scott Reef and the Abrolhos
Islands in west Australia; the Great Barrier Reef (GBR),
represented by northern and southern GBR samples; New
Caledonia and Taiwan, with Taiwan containing two genetic
lineages. Furthermore, this study identiWed source and sink
populations within and among regions. SpeciWcally, the
northern population in west Australia (Scott Reef) was
identiWed, as the source for replenishment of the Abrolhos
population, whilst New Caledonia was a source for recruit-
ment to the GBR. Based on these insights from a single
mtDNA marker, this study will facilitate the development
of rational management plans for the conservation of
P. leopardus populations and therefore mitigate the risk of
population declines from anthropogenic inXuences.
Introduction
Contemporary coral reefs and their associated fauna and
Xora are (and have been) aVected by contemporary
ecological and historical changes including habitat distur-
bance associated with anthropogenic activities and climate
change (Munday 2004; Jones et al. 2004; Smith and
Buddemeier 1992; PandolW et al. 2006; Hubbard et al. 2005).
Such disturbances to the coral reef habitat have modiWed
the genetic structures of populations of many coral reef
inhabitants [examples include invertebrates (Benzie 1999),
sea snakes (Lukoschek et al. 2007) and Wsh (Chenoweth
et al. 1998; McCaVerty et al. 2002)]. The main focus of the
present study is to determine the Indo-West PaciWc popula-
tion genetic structure of the ecologically and economically
important tropical grouper, Plectropomus leopardus a
coral reef inhabitant belonging to the recently resurrected
Epinephelidae (Smith and Craig 2007). This study will
extend previous work on the genetic structure and history
of this species on the Australian plate (van Herwerden
et al. 2002, 2006) to include regions to the north and east
of the Australian plate. As far as can be ascertained, P. leo-
pardus is discontinuously distributed in northern Austra-
lian waters, as it extends southwards down both east and
west Australian coasts, but not between these locations
(Heemstra and Randall 1993). Tropical east and west
Australian coasts have experienced extensive periods of
isolation during the last two interglacial maxima, due to
the presence of the Torres Strait land bridge, which persisted
until 9 kya (Larcombe 2004). Accordingly, P. leopardus
populations to the east and west of the Torres Strait are
genetically partitioned with no evidence of contemporary
gene Xow (van Herwerden et al. 2006). Indeed, using
coalescence analyses, population partitioning by the Torres
Strait Barrier was dated back to the Pleistocene and found
Communicated by T. Reusch.
L. Van Herwerden (&) · J. Howard Choat · G. Hillersøy
School of Marine and Tropical Biology, James Cook University,
Townsville, QLD 4811, Australia
e-mail: lynne.vanherwerden@jcu.edu.au
L. Van Herwerden · M. Leray · G. Hillersøy
Molecular Ecology and Evolution Laboratory,
James Cook University, Townsville, QLD 4811, Australia
S. J. Newman
Department of Fisheries, Western Australian Fisheries an d Marine
Research Laboratories, Government of Western Australia,
North Beach, WA 6920, Australia
1596 Mar Biol (2009) 156:1595–1607
123
to predate the last glacial maximum (LGM) (van Herwerden
et al. 2006).
This species is of particular interest, as it is the most
valuable component of the coral reef Wsh Wshery on the
GBR (Russ et al. 2008), an industry that has expanded on
the GBR during the last decade, particularly as it is a pri-
mary target species for the live Wsh trade in Asia (Welch
et al. 2008). This study examines the population genetic
structure and recent history of P. leopardus from an
extended geographic range, since its distribution in the
Western PaciWc extends from southern Japan in the north to
east and west Australia (Queensland and Western Austra-
lia) in the south, as well as eastward to the Caroline Islands
and Fiji (Heemstra and Randall 1993). The aims of the
present study are fourfold, each evaluating a speciWc pre-
diction that emerged from the earlier P. leopardus study
(van Herwerden et al. 2006):
(1) Given the observed regional genetic diVerentiation
between east and west Australian P. leopardus during
the Pleistocene (van Herwerden et al. 2006), how are
P. leopardus populations structured when incorporating
samples representing regions to the north (Taiwan) and
east (New Caledonia) of Australia? Given the geogra-
phy of these and the Australian localities, we would
expect the Taiwanese and New Caledonian populations
to be diVerentiated as well. Phylogenetic and population
genetic analyses will be employed to evaluate this.
(2) Assuming that multiple regional P. leopardus popula-
tions are identiWed, not only within Australia, but also
beyond, what is the timeframe within which these
populations became partitioned? The prediction from
our earlier work is that population genetic partitioning
in additional regions is also associated with Pleistocene
glacial–interglacial cycles, predating the LGM, as
already shown for a number of tropical east–west
Australian populations, including P. leopardus. This
will be evaluated using coalescence analyses, whereby
the mean and 95% conWdence interval of the age of the
most recent common ancestor (mrca) to each geneti-
cally diVerentiated population will be estimated. Age
estimates of regional populations will then be compared
to determine if divergence(s) occurred in response to a
single or multiple historical events.
(3) To what extent do populations from Taiwan, west
Australia, east Australia and New Caledonia exchange
recruits? The expectation is that a greater exchange of
recruits will occur between regional populations that
are geographically closer to each other, particularly if
dispersal of pelagic larvae is facilitated by currents.
Isolation by distance (Mantel tests) and the relative
number of migrants between regions will be deter-
mined to evaluate this prediction.
(4) A natural extension of ascertaining the presence of
inter- and intra-regional gene Xow is whether the gene
Xow is uni- or bidirectional and whether there is any
evidence of contemporary or historical changes in the
direction(s) of such gene Xow? The prediction is that
the direction of gene Xow is predominantly consistent
with contemporary currents, speciWcally (a) the south-
ward Xowing Indonesian throughXow, which also
extends westwards into the Indian Ocean, connects
Taiwanese and west Australian populations and (b) the
southward Xowing Leeuwin current transports pelagic
larvae from the north of WA to the south (c) the south-
ward Xowing east Australian current (EAC) transports
larvae from the northern to the southern GBR and (d)
the westward Xowing South Equatorial Current con-
nects New Caledonia populations with the GBR. How-
ever, disruptions to these currents, due to the historic
emergence of land bridges and vast areas of exposed
continental shelves as well as temperature change dur-
ing glacial–interglacial cycles, may produce genetic
footprints that are inconsistent with contemporary cur-
rent directions, such as those reported by Benzie
(1999). Additional migration analyses in a Bayesian
framework and both contemporary and historic ocean-
ographic data will be employed to evaluate this.
Materials and methods
Sampling procedures
A total of 211 Plectropomus leopardus individuals were
collected from six locations, representing four regions, as
shown in Fig. 1 Two east coast samples from the Great
Barrier Reef (GBR) were represented by the latitudinal
extremes of the reef system, extending from 10 to 23°S,
Torres Strait in the Far Northern GBR, n=19 and the
Capricorn Bunkers in the Southern GBR, n=20. Two west
Australian coast (WA) samples were collected from locali-
ties at which P. leopardus occurred in suYcient abundance
(Ayling and Choat personal observation), Scott Reef (14°S)
in the north, n=88 and the Abrolhos Islands (centred
around 29°S) in the south, n=42. This sampling strategy
ensured that GBR samples were from locations that are
similar distances apart to those obtained from WA, for
comparative purposes. All GBR and some WA samples had
already been processed and the data incorporated from van
Herwerden et al. (2006), but additional samples from the
Abrolhos Islands and Scott Reef in WA have been obtained
by line Wshing and are included here. The remaining two
regions, from which Australian populations may be
replenished, and where P. leopardus is abundant, were
Mar Biol (2009) 156:1595–1607 1597
123
represented by Taiwan to the north and New Caledonia to
the east of Australia. The Taiwanese samples were collected
in person by JHC, either from the Magong Wsh market in
PengHu county or by spear Wshing from the PengHu archi-
pelago, which lies along the west coast of Taiwan. Market
managers conWrmed that Wsh sold at this market were all
caught locally from reefs of the PengHu archipelago and
were not imported from outside of this region. New Cale-
donia samples were collected exclusively from local Wsh
markets. A total of 20 P. laevis individuals were collected
as an outgroup from Scott Reef (n=2) in WA, the Herald
Cays in the Coral Sea (n=17) and New Caledonia (n=1)
(see Fig. 1; Table 1). In all cases, Wnclips were obtained and
immediately placed into vials containing 80% ethanol.
Laboratory procedures
DNA was extracted from Wnclips of all samples by protein-
ase K digestion and standard salt/chloroform methods
(Sambrook et al. 1989). The Hypervariable region I (HVR
I) of the mtDNA control region of all samples collected was
ampliWed using primers L16007 and H00651 (Kocher et al.
1989). PCR was performed in a 20 l reaction volume con-
taining 2.5 mM Tris pH 8.7, 5 mM KCl, 5 mM (NH4)2SO4,
2.5 mM MgCl2, 200 M each dNTP, 10 M each primer,
10 ng template DNA and 1 unit of Taq polymerase
(Qiagen). After an initial 2 min denaturation at 94°C, each
of 30 cycles consisted of denaturation for 30 s at 94°C,
annealing at 55°C for 30 s and extension at 72°C for 90 s
Fig. 1 Map of the study area
within the Indo-PaciWc,
indicating seven sampling loca-
tions marked by star symbols,
black Wll for P. leopardus and
white Wll for P. laevis
Australia
Capricorn
Bunkers
Torres Strait
New
Caledonia
Scott Reef
Abrolhos
Islands
Taiwan
23°00N
10°00S
121°50E 146°45E
Japan
Philippines
Solomon
Islands
Fiji
30
30
0
Table 1 Population genetic parameters pertaining to the genetic diversity of each sample and for the total sample are shown
Number of individuals sampled per location (n), number of haplotypes detected per sample (nh), haplotype diversity (h), nucleotide diversity ex-
pressed as a percentage () and mean numbers of pairwise diVerences within each sample
Location nn
hh (§SE) % (§SE) Mean pairwise diVerences (§SE)
Scott Reef 88 60 0.97 (0.012) 1.34 (0.704) 6.893 (3.274)
Abrolhos Isl. 42 36 0.99 (0.009) 2.87 (1.454) 14.709 (6.717)
GBR 39 37 0.99 (0.008) 2.19 (1.120) 12.764 (5.879)
New Caledonia 18 15 0.97 (0.029) 2.512 (1.321) 14.471 (6.802)
Taiwan 24 22 0.99 (0.014) 2.899 (1.498) 14.869 (6.894)
Total 211 170 0.99 (0.002) 10.74 (5.146) 63.576 (27.534)
1598 Mar Biol (2009) 156:1595–1607
123
with a Wnal 10 min extension at 72°C. All PCR products
were evaluated on a 2% agarose gel, puriWed by standard
isopropanol precipitation (Sambrook et al. 1989) and
sequenced directly in both directions using the PCR
primers. Automated sequencing was performed using dye
terminator chemistry (ABI) following the manufacturer’s
instructions. Sequences were obtained following PAGE on
ABI 310 capillaries (Macrogen Inc., Korea). Nucleotide
sequences were aligned using the programmes ClustalW
(http://www.ebi.ac.uk/Tools/clustalw2/index.html) and Bio-
Edit (Hall 1999). Sequences generated in this study were
submitted to GenBank (XXX-XXX).
Phylogenetic analyses were performed on all sequences.
This was done for two reasons: (1) to identify the evolu-
tionary links between and within regions, (2) to inform
population genetic analyses, by identifying lineages that are
consistently retrieved and supported. The evolutionary rela-
tionships between samples were inferred using Maximum
Likelihood (ML) methods implemented in PAUP 4.0*
(SwoVord 1999) and (Zwickl 2006). First a likelihood
approach was used in Modeltest version 3.06 (Posada and
Crandall 2001) to Wnd the best substitution model for ML
analyses. ML analyses were performed ten times in Garli
and 100 bootstrap replicates were also performed in Garli
for evaluation of the level of support for the tree topology.
All trees were outgroup rooted using P. laevis as this spe-
cies is closely related to P. leopardus.
Coalescence analyses were implemented in Beast
(Drummond and Rambaut 2007) to estimate genetic bottle-
neck/expansion events of the mtDNA lineages. Knowing
what the timeframe of expansion events was, allows us to
generate hypotheses about the historic events associated
with these expansions from population refuges or newly
colonised populations. These analyses were done using the
best tree obtained from ten independent Garli ML analyses
as a constraint for the Beast analyses. P. leopardus clades
and their most recent common ancestor (mrca) ages were
then calculated based on the branch lengths recovered dur-
ing ten independent Beast analyses, each performing
9 million post-burn-in MCMC iterations (with a burn-in of
1,000,000). The uncertainty of branch lengths was captured
by parameters for all post-burn-in trees from all ten inde-
pendent Beast analyses combined, as all analyses were con-
sistent. The coalescent, Tau (§95% CI) in number of
generations, for the root node was then used in conjunction
with one of the best trees (tree 9,773,000 from the sixth
analysis, LnL score of ¡4460.98647) to compute the time-
frame within which the mrca to the P. leopardus lineage
and each P. leopardus clade emerged (Table 2). This was
converted from time expressed as number of generations to
an age in kya as per van Herwerden et al. (2006), where
mutation rate and generation times of female P. leopardus
were previously calibrated (van Herwerden et al. 2006).
Mrca ages from the root node to the tips were then calcu-
lated based on mean, upper and lower 95% CI branch
lengths leading to nodes, as follows: root age ¡(root
age £timescale to tmrca P. leopardus ingroup node) =
P. leopardus tmrca age (age of most recent common ancestor).
Given the mean age of the root node, based on Tau, the
mean age of the P. leopardus ingroup was calculated as
follows: 554.03 ¡(554.03 £0.837) = 90.31 kya. These cal-
culations were then repeated as before for lower and upper
ages based on 95% CI values obtained for each node age
from upper and lower ages of the preceding node. The
chronogram was also based on the tree from the sixth Beast
run (see Fig. 1) and calculated ages (mean §95% CI) are
presented in Table 2.
Population genetic analyses were performed in order to
measure genetic connectivity within and among regions
sampled. First, population genetic parameters pertaining to
the genetic diversity of each sample and for the total sample
were obtained using Arlequin 2.0.1.1 (Schneider et al.
2000). SpeciWcally, haplotype diversity (h), nucleotide
diversity expressed as a percentage () and mean numbers
of pairwise diVerences within each sample were deter-
mined, since genetic diversity is an indicator of the genetic
resilience of populations to perturbations (van Oppen and
Gates 2006). Second, to identify regional population
genetic partitioning and to identify samples associated with
genetic partitioning, pairwise Fst analyses were performed
as implemented in Arlequin version 2.0.1.1 (Schneider
et al. 2000). Third, an indication of the levels of migration
between and within regions was determined as per Slatkin
(1987), where M=Nm, the number of migrants per genera-
tion for haploid data. The question of directionality of gene
Xow was addressed using Wve independent repeated ML
analyses implemented in MIGRATE to ensure that parame-
ter space was widely sampled, thereby ensuring that restric-
tions to local optima were avoided (Beerli and Felsenstein
2001). Theta and M values were generated from the Fst cal-
culation and mean proWle likelihoods of M from all Wve
runs combined are reported (with 95% conWdence intervals,
Table 2 Population pairwise FST values for Wve of the six samples
studied
Sites are in WA; ScR = Scott Reef and AbI = Abrolhos Islands,
GBR = Combined TS and CB samples (see “Results”), NCa = New
Caledonia and Tai = Taiwan
All P values were highly signiWcant (P<0.0001)
Site ScR AbI GBR NCa Tai
ScR 0.000
AbI 0.361 0.000
GBR 0.936 0.900 0.000
NCa 0.940 0.900 0.310 0.000
Tai 0.720 0.460 0.893 0.885 0.000
Mar Biol (2009) 156:1595–1607 1599
123
CI). All MIGRATE analyses were performed assuming a
constant mutation rate, with ten short and three long chains
of 500 and 5,000 steps, respectively. Ten thousand of
100,000 trees were discarded per chain as burnin. Fourth, to
explicitly show spatial and historical patterns of haplotype
distributions for each clade identiWed by the phylogenetic
analyses, including expansion events within and among
regions, a haplotype diagram based on a single minimum
spanning tree (MST) was constructed, using Arlequin ver-
sion 2.0.1.1 (Schneider et al. 2000) as per Rohlf (1973).
Finally, this partitioning was more formally examined to
quantify genetic diVerentiation between regions, among
populations within regions and between populations, using
hierarchical measures of genetic structure as determined by
AMOVA (as per ExcoYer et al. 1992) in Arlequin 2.0.1.1
(Schneider et al. 2000). The hierarchical structure involved
grouping the samples into three regions, WA, east Australia
(GBR and New Caledonia) and Taiwan, whilst signiWcance
was determined following 1,023 permutations. Finally, an
isolation by distance (IBD) Mantel test was performed to
determine if there were correlations between genetic and
geographical distances. This is important, because conWrma-
tion of IBD has implications for sustainable management: if
population connectivity diminishes as geographic distance
between them increases, populations will not be adequately
replenished by recruitment from distant populations at eco-
logically relevant timescales, and will rely on more localised
recruitment. This has clear implications for sustainable
management of the Wshery. The Mantel test was imple-
mented in the Isolation by distance web service (IBDWS)
(Jensen et al. 2005). IBD was evaluated by reduced major
axis regression, which also estimates the slopes and inter-
cepts of relationships between the two distances. Geographic
distances were measured as the most direct connections
between sampled locations via water and genetic distances
were measured as Fst/1-Fst (Rousset 1997). SigniWcant cor-
relation between genetic and geographic distance matrices
indicates that isolation by distance is operating.
Results
Phylogenetic analyses
Evolutionary relationships between samples were identiWed
by phylogenetic analyses and one of the best phylogenetic
trees obtained is depicted as a chronogram (Fig. 2). This
analysis suggests that regional populations are genetically
partitioned. It identiWes three major lineages, within which
further genetic subdivisions are identiWed. A consensus tree
of ML trees obtained using both PAUP (SwoVord 2000)
and (Zwickl 2006) retains the three major lineages, but fails
to resolve the polytomy of these three major clades and can
therefore not conWrm which lineage diverged from the
P. leopardus mrca Wrst. This will be further explored using
more appropriate coalescence age estimates for each clade
(see later). Regardless, it is apparent that the chronogram of
Australian and other regional P. leopardus populations to
the north and east of Australia, consists of two reciprocally
monophyletic regional lineages, which are sister to each
other: (1) East Australia (GBR-New Caledonia) and
(2) West Australia–Taiwan. Within each of these major
lineages are two geographically deWned regional clades that
are not monophyletic, since there is evidence of predomi-
nantly (but not exclusively) unidirectional gene Xow
between locations in each region. East Australia has two
lineages represented by the Great Barrier Reef and New
Caledonia (clade 3) with evidence that the east Torres Strait
population was the source of the east coast population and
the GBR population has received recruits from New Cale-
donia (clades 3.2 and 3.3), because the New Caledonia
clade also contains a reasonable proportion of residents
from the GBR. However, New Caledonia has not received
recruits from the GBR, since there are no New Caledonia
individuals that bear GBR-derived haplotypes (clade 3.1).
West Australia and Taiwan are sister lineages that are
almost completely monophyletic and mutually exclusive
(clades 2 and 1, respectively), with the exception of a single
Taiwan haplotype in the WA clade, either representing gene
Xow to Taiwan from WA or an ancestral haplotype that has
persisted in both populations. Likewise, the Taiwan lineage
is subdivided into two clades, but the geographic contribu-
tions to these two Taiwanese lineages cannot be identiWed
without samples from other locations in the west PaciWc
(most notably Japan). Populations of P. leopardus in West-
ern Australia appear to be geographically partitioned as
well, with one lineage representing the majority of Abrolhos
Islands individuals (clades 2.3 and 2.4) and the second line-
age representing the majority of Scott Reef individuals
(clades 2.1 and 2.2). There appears to be bidirectional gene
Xow over evolutionary time between Scott Reef and the
Abrolhos, but this appears to be temporally partitioned.
Coalescence analyses
In order to resolve the relative timing of lineage diver-
gences between the four regions sampled and recruitment
events between locations, we estimated the age of the most
recent common ancestor (tmrca) of each lineage (Fig. 2;
Table 3). It appears that all lineages emerged within a
relatively short and recent evolutionary time during the last
77–82 kya (mean coalescent times), with the 95% conWdence
intervals of the coalescent ages spanning 51–113 kya. The
order in which the diVerent regional lineages emerged
cannot be determined with any conWdence. This is
supported by the observation that a consensus tree of
1600 Mar Biol (2009) 156:1595–1607
123
100 ML bootstrap replicates was unable to resolve the
polytomy for the three regional lineages: Taiwan, WA and
EA (including New Caledonia). Clearly, all coalescent
events for each regional P. leopardus population examined
here occurred during the Pleistocene, possibly in response
to the same historic trigger. Population genetic analyses
will explicitly inform whether what has been suggested by
phylogenetic and coalescence analyses are supported.
Population genetic analyses
Population genetic diversity measures indicated that genetic
diversities were high for all samples investigated (Table 1),
with the lowest genetic diversity observed for the Scott Reef
sample, which had about half the nucleotide diversity and
half the mean pairwise diVerences compared to all other
samples. This suggests that Scott Reef is either younger than
all other samples investigated, or that it is much more depen-
dent on self-recruitment than are the other populations, since
both of these circumstances will result in lower genetic
diversity than observed when there are older, more stable
and persistent populations or when there is recruitment from
genetically divergent sources to a population.
Pairwise tests provided insights into where signiWcant
genetic partitioning occurred and whether the partitioning
Fig. 2 Chronogram of
Plectropomus leopardus lineage
diversiWcation, obtained from
coalescence analyses performed
in Beast (Drummond and
Rambaut 2007). Each major
lineage is enclosed in a box and
the composition of each clade
within all major lineages is
depicted by a pie diagram, the
pie slices corresponding to
the number of individuals from
the speciWc location sampled,
as per the embedded key to the
Wgure. The timescale along the
x
-axis is relative to the tree root
height. Numbers on branches
correspond to relative time from
the previous node to the MRCA
of the next divergence and
dotted vertical lines indicate
relative time from ¡1.0 to the
present, 0. Note that the relative
time from ¡1.0 to 0.5 units was
truncated (//) to Wt onto the page
Scott
Abrolhos
Taiwan
Torres S
Capricorn B
N Caledonia
EA clade
WA - Taiwan clade
0.1278
0.837
0.9823
0.0984
0.0245
0.0313
0.0298
0.0164
0.0223
0.0176
-1.0 -0.5 -0.25 0.0
P. leopardus
P. laevis
2
2.4
2.3
3
3.4
3.3
3.2
3.1
1.1
1
1.2
2.1
2.2
Table 3 Population average pairwise diVerences in bp for Plectropomus
leopardus samples in this study
Above the diagonal are the average number of pairwise diVerences
between populations (PiXY), diagonal elements are the average number
of pairwise diVerences within populations (PiX) and below the diagonal
are the corrected average pairwise diVerences (PiXY ¡(PiX+PiY)/2)
for each population pair. Sampling codes are as before (Table 2)
All P values were highly signiWcant (P<0.0001)
Site ScR AbI GBR NCa Tai
ScR 6.893 16.056 137.439 137.111 32.732
AbI 5.254 14.710 134.618 133.005 27.369
GBR 127.611 120.882 12.764 19.540 127.032
NCa 126.429 118.415 5.923 14.471 127.502
Tai 21.851 12.579 113.215 112.832 14.870
Mar Biol (2009) 156:1595–1607 1601
123
was extensive, moderate, minimal or absent. There was no
signiWcant genetic partitioning between sampled locations
at the northern and southern extremes of the GBR (TS and
CB pairwise Fst = 0.01533, P= 0.18215). This conWrms,
with an extended data set, what was previously demon-
strated by van Herwerden et al. (2006) and suggests that the
GBR population may be considered a single population for
the purpose of further analyses in this study. In contrast,
when pairwise Fst values were obtained for each of the
remaining four samples and the combined GBR sample,
highly signiWcant population partitioning was observed at
all levels (Table 2). These ranged from a low of 0.301
(GBR-New Caledonia) to a high of 0.940 (Scott Reef-New
Caledonia). Correspondingly, the corrected population
average pairwise diVerences in base pairs (bp) show signiW-
cant pairwise diVerences between all populations presented,
the lowest being 5.254 (Scott Reef-Abrolhos Islands) and
the highest for Scott Reef-New Caledonia (126.429), as
before. It also clearly identiWes that the lowest within popu-
lation pairwise bp diVerence was at Scott Reef, 6.893
(Table 3), which was at least half that observed within the
remaining populations, again consistent with previous
observations of genetic diversity. These Wndings are further
supported by measures of the levels of migration (M),
where M was by far the greatest between the two GBR sam-
ples (TS and CB), M= 32.12, followed by migration
between New Caledonia and the GBR, which was an order
of magnitude lower (Table 4). Not surprisingly, the lowest
measured migration was between Scott Reef and the east
coast samples, two orders of magnitude lower than the
GBR-New Caledonia migration levels (Table 5). Bayesian
analyses of the estimated migration levels, using
MIGRATE produced results consistent with the above Wnd-
ings and indicate that the direction of greatest gene Xow is
with the direction of contemporary oceanic currents
(Table 6; Fig. 3). The currents are the southward Xowing
east Australian current along the Great Barrier Reef, which
connects Torres Strait to Capricorn Bunkers, the westward
Xowing South Equatorial current, which connects New
Caledonia to the GBR and the south Xowing Leeuwin
current along the WA coast connecting Scott Reef to the
Abrolhos Islands. The only exception to this is the apparent
northwards direction of greater gene Xow from the Abrol-
hos Islands to Taiwan rather than southward gene Xow by
the Indonesian throughXow current, since the contemporary
current direction is southward, not northward (Table 7).
In addition, the MST (Fig. 4) supports the genetic parti-
tioning between all four regions and within the WA region. It
is also consistent with the lack of partitioning within the GBR
and suggests that each of the Wve genetically diVerentiated
populations has undergone an expansion event. Furthermore,
it also depicts the presence of extensive haplotype
exchanges between the two distant GBR samples, reduced
exchanges between the GBR and New Caledonia and
Table 4 AMOVA analysis was performed on Plectropomus leopar-
dus sequences structured by region as per Weir and Cockerham 1984,
ExcoYer et al.1992; Weir 1996
For this purpose, GBR consisted of TS and CB samples (as before),
west Australia consisted of Scott Reef and Abrolhos Islands samples,
whilst Taiwan and New Caledonia samples were each in their own
group. AG represents the genetic variance among groups, APWG rep-
resents the genetic variance among populations within groups and WP
represents the genetic variance within populations
Source of
variation
df Sum of
squares
Variance
components
Percentage
of variation
AG 3 5382.064 43.557 Va 85.29
APWG 2 163.733 2.000 Vb 3.92
WP 205 1129.672 5.511 Vc 10.79
Total 210 6675.469 51.068
Table 5 Matrix of migration, M between P. leopardus from the four
regions sampled, where M=Nm for haploid data as per Slatkin 1987
Sample codes as per Table 2
Site ScR AbI GBR NCa
AbI 0.886
GBR 0.034 0.057
NCa 0.032 0.062 1.117
Tai 0.195 0.587 0.060 0.065
Table 6 Direction of gene Xow between P. leopardus samples
Values reported are for the mean number of migrants, M from Wve
independent Bayesian analyses. The lower and upper M values, respec-
tively are the upper and lower values of the 95% ConWdence Interval.
Numbers in parentheses are standard deviations for each of the Wve
independent Bayesian analyses, as per “Materials and methods”. In the
WA region, SR = Scott Reef and Abr = Abrolhos Islands; north o
f
WA, the Taiwan population = Tai; along the GBR, TS = Torres Strait,
CB = Capricorn Bunkers and eastward, NC = New Caledonia. The
direction of gene Xow for each population pair given, that has the great-
est value based on M, is in bold
Mean M Lower M Upper M
SR to Abr 98.5 (37.1) 53.9 (22.9) 162.8 (56.5)
Abr to SR 7.4 (5.5) 2.8 (2.6) 18.9 (8.5)
TS to CB 450.9 (232.5) 314.5 (182.6) 624.3 (288.3)
CB to TS 48.1 (51.3) 26.8 (32.4) 80.7 (76.8)
NC to TS 38.1 (29.4) 20.2 (18.6) 65.1 (44.4)
TS to NC 10.9 (13.1) 2.2 (4.3) 15.8 (21.7)
NC to CB 132.5 (99.9) 71.4 (64.6) 205.5 (143.4)
CB to NC 23.9 (28.2) 14.2 (11.3) 55.1 (46.0)
Tai to SR 2.9 (3.1) 0.5 (0.7) 9 (8.9)
SR to Tai 4.9 (3.1) 0.5 (0.5) 17.8 (16.4)
Tai to Abr 8 (8.4) 4.7 (6.7) 25.7 (26.0)
Abr to Tai 43.1 (44.6) 20.7 (27.0) 80.4 (66.1)
1602 Mar Biol (2009) 156:1595–1607
123
reduced exchanges between the isolated locations within WA.
All of the above is consistent with the phylogenetic analyses.
Finally, AMOVA analyses, which are hierarchical, also
conWrmed these Wndings. All three Wxation Indices (Fsc, Fst
and Fct) were relatively large and highly signiWcant.
Fsc = 0.26631 (P<0.0001), Fst = 0.89209 (P<0.0001) and
Fct = 0.85293 (P<0.020), indicating strong genetic partition-
ing at all three hierarchical levels. Most of the genetic variance
Fig. 3 Minimum Spanning Tree (MST) of Plectropomus leopardus
haplotypes from the sampled sites. Colour denotes location sampled as
indicated by the embedded key. Roman numerals indicate the number
of substitutions separating haplotypes sampled in this study. Circle
sizes indicate the proportional abundance of haplotpyes
Scott Reef
Abrolhos Islands
II
X
II
II
VI
IV
II I
VI
IV
IV
III III
III
IV
VII
5
Taiwan
2
2
IX
III
III
III
III
IIII
I
II
XV
XI
III
III
II
IV
I
X
II
II
II
III
I
VIII
VI IV
X
XV
I
I
V
Torres Strait
CXI
VI
VI VII
IV
III III IV
VI
II
I
III
III
IV
3
III
I
II III
II
II
V
I
II
II
VV
VI II
III
I
I
II
III
I
II
I
I
I
I
I
I
I
II
III
II
I
II
II
I
I
II
I
III
III
I
II
II
II
III
I
3
III
I
III
VII
I
I
I
II
Capricorn Bunkers
New Caledonia
3
IV
II
II
II
II
X
XI
III
X
XX
XXV
VI
V
IV
II
I
IV
III
IV
II
III
X
IV
III II IX
IV III III
III
I
II II
III
VI
II
II
I
I
V
I
II
III
III
I
I
V
II
Mar Biol (2009) 156:1595–1607 1603
123
(more than 85%) was due to genetic diVerences among groups
(i.e. regions) as expected, given the deep phylogenetic split at
the inter-regional level. Given the strong genetic diVerentiation
observed among these regions, it is important to determine if
there is evidence of isolation by distance between the Wve
regional locations. Isolation by distance analyses using Mantel
tests (Jensen et al. 2005) indicated that genetic and geographic
distance matrices are not signiWcantly correlated for any of the
analyses (neither linear, nor log-transformed data), as P>0.05
in all cases, ranging from 0.07 to 0.09 for all analyses. Further-
more, reduced major axis (RMA) regressions suggested that
geographic distance only accounted for 13.1–23.1% of the
genetic variation, suggesting that factors other than geographic
distance are more important contributors to genetic partition-
ing between populations of P. leopardus sampled here.
Discussion
In this study, we have identiWed that there are distinct popu-
lations of coral trout in each of four Indo-West PaciWc
regions sampled and that there is even population structure
within two of the four regions. This provides a strong basis
for the eVective sustainable management of coral trout Wsh-
eries in the Indo-West PaciWc. Furthermore, this study iden-
tiWed the historic and contemporary mechanisms that have
determined the observed population structure of P. leopar-
dus in the Indo-West PaciWc.
Population structure and history informed
by phylogenetic structure
Phylogenetic and population genetic analyses conWrmed
that there are at least four regional populations of Plectrop-
omus leopardus. These are Taiwan and WA (each of which
is further subdivided into two genetic lineages within the
region), GBR and New Caledonia. When these regional
populations became partitioned is of interest, since it may
allow us to understand how regional coral trout populations
have responded to historical and/or contemporary environ-
mental change experienced by coral reefs.
Timeframes of change associated with observed genetic
structure
There is no contemporary hard barrier to geneXow between
the regions sampled, but historically there was a physical
barrier (the Torres Strait land bridge) to gene Xow between
WA and EA P. leopardus populations. The coalescent age
for the major split into the two (EA-New Caledonia and
WA-Taiwan) clades identiWed here, spans much of the
interglacial period between the two most recent glacial
maxima (150 and 18 kya, respectively, Voris 2000). During
this interglacial, sea levels varied between extreme lows of
about ¡80 m and highs of about 0 m (sensu Benzie 1999).
The genetic diVerentiation between these populations was
maintained, even after the landbridge became submerged,
as it has been for the last 9 kya. The Torres Strait land
bridge has been documented as a barrier associated with
genetic diVerentiation for several other tropical marine
species (e.g. starWsh, Williams and Benzie 1998; Wsh,
Chenoweth et al. 1998; sea snakes, Lukoscheck et al.
2007), but not others (e.g. parrotWsh, Bay et al. 2004; sur-
geonWsh, Klanten et al. 2007; Horne et al. 2008). Besides
the formation of barriers to dispersal, such as the Torres
Strait Barrier, the impact of changing sea levels and water
temperatures on coral reefs is well documented. Pertinent
here is the work by Webster et al. (2008) documenting that
the present GBR shelf edge is where the GBR existed for
85% of time during the last 500 kya, due to climate change
and associated sea level Xuctuations across the shelf edge.
At lower sea levels than present, the GBR clearly had a
complex history of growth and erosion, identiWed by the
study of fossil GBR reefs (Webster et al. 2008).
Notwithstanding the Torres Strait Barrier, the remaining
genetically diVerentiated P. leopardus regional populations
identiWed in this study (Taiwan, New Caledonia) were not
and are not presently isolated from the other populations
by such a Wxed barrier. The remaining populations were
instead, isolated by oceanographic features. Consistent with
this, the coalescent analyses identiWed that the remaining
population pairs have also experienced (albeit incomplete)
isolation more recently than was determined for the com-
pletely isolated WA and EA populations. SpeciWcally, the
WA and Taiwan clades are more evolutionarily related to
each other, more recently, than WA and EA populations
are, with almost complete isolation. Apparently, gene Xow
Table 7 Coalescent ages of major Plectropomus leopardus clades and
lineages sampled in this study, as calculated from analyses using Beast
(Drummond and Rambaut, 2007), assuming a molecular clock o
f
0.45% mutations per site per Mya and a generation time of 5 years (as
per van Herwerden et al. 2006)
Clade Tau Divergence time (kya)
Mean Range (95% CI)
Tree root 1.19 554.03 365.47–767.73
All ingroup clades Calculated* 90.31 59.57–125.14
Tmrca to clade 1–2 Calculated* 81.42 53.71–112.83
Tmrca to clade 2 Calculated* 80.22 52.39–110.07
Tmrca to clade 1 Calculated* 78.87 52.03–109.30
Clade 2 (ex 2.4) Calculated* 78.90 51.53–108.26
Clade 2.4 Calculated* 77.83 50.83–106.79
EA clade 3 Calculated* 78.77 51.96–109.15
GBR clade 3.1 Calculated* 77.38 51.05–107.23
New Cal clade (3.2–3.3) Calculated* 77.01 50.80–106.72
1604 Mar Biol (2009) 156:1595–1607
123
between Taiwanese and Australian P. leopardus, aided by
the Indonesian throughXow in contemporary times, may
have been disrupted in the past, thereby restricting past
recruitment from the north PaciWc to Australia. This has
been documented for other tropical marine species (e.g.
scleractinian corals, Knittweis et al. 2009; gastropods,
Crandall et al. 2008). Such disrupted gene Xow was proba-
bly due to the disruptive impact of the Torres Strait Barrier
on the adjacent Indonesian throughXow. In contrast, the
most closely related GBR and New Caledonia clades were
more recently partitioned, and there is evidence of greater,
possibly contemporary, connectivity between these popula-
tions. This suggests that oceanography and currents (both
historic and contemporary) as well as larval behaviour
(Leis and Carson-Ewart 1999) together may contribute to
reasonably eVective barriers to dispersal by these coral
trout larvae.
Gene Xow, IBD and relative levels of migration
Geographic distance is also considered an important con-
tributor to the extent or lack of gene Xow between popula-
tions. There was no signiWcant correlation between genetic
and geographic distances between P. leopardus popula-
tions, with less than 23% of the genetic distance correlated
with geographic distance. This suggests that the level and
direction of gene Xow between P. leopardus populations is
more strongly inXuenced by biogeographic and oceano-
graphic features than by geographic distance (e.g. the
Torres Strait Barrier and oceanic current direction of Xow).
This was clearly illustrated by the diVerences in estimated
levels of gene Xow, either to or from a particular popula-
tion, when considering pairs of populations in each of the
two discrete P. leopardus lineages (WA-Taiwan and
EA-New Caledonia), within which gene Xow occurred.
Apparently, coral trout larvae rarely traverse open water
distances in the order of thousands of km, unless there are
intermediate, suitable coral reefs present. This is evident
from the extensive contemporary gene Xow between the
Torres Strait and Capricorn Bunkers, which span the entire
GBR, but much less gene Xow between the more isolated
locations in WA or between New Caledonia and the GBR.
Direction of gene Xow and oceanography: contemporary
and historic currents
The major contemporary ocean currents relevant to this
study are the Leeuwin and Indonesian throughXow for the
WA-Taiwan lineage and the East Australian (EAC) and
South Equatorial (SEC) currents for the EA-New Caledo-
nia lineage. Greatest gene Xow always coincided with
downstream currents, except for the anomalous observation
Fig. 4 Map of the sampling
area, indicating the six
P. leopardus sampling locations
and the relative strength and
direction of migration between
sites to the east and west of
Australia. Arrow length indi-
cates relative strength of gene
Xow in direction shown. Sam-
pling sites are identiWed by star
symbols. A Abrolhos, S Scott
Reef, Tw Taiwan, TS Torres
Strait, CB Capricorn Bunkers,
N
C New Caledonia
Australia
Capricorn
Bunkers
T orres
Strait
New
Caledonia
Scott Reef
Abrolhos
Islands
Taiwan
23°00N
10°00S
121°50E 146°45E
Fiji
30
30
0
TS
CB
CB
TS TS
NC
NC
CB
A
A
S
S
Tw
STw
S
A
Tw
Tw
A
TS
NC
Mar Biol (2009) 156:1595–1607 1605
123
that counter-current gene Xow from the Abrolhos Islands
to Taiwan was Wvefold greater than it was in the reverse,
down current direction. This may either be an anomaly,
due to the persistence of an ancestral polymorphism of
Taiwanese ancestry in the Abrolhos Islands, or due to his-
toric changes in the Leeuwin and Indonesian throughXow
currents, which permitted reversed gene Xow. Both con-
temporary and historic changes (e.g. ENSO events and the
closure of the Torres Strait) have had an impact on these
four ocean currents (Brinkman et al. 2002; Meyers 1996).
Furthermore, there are two counter currents in WA, one is
oVshore along the continental shelf edge of the WA coast,
albeit at a depth of 300 m (Godfrey and Ridgway 1985)
and the other is inshore. Ancestral polymorphism or past
changes in contemporary current direction along the WA
coast are also evidenced by the one tenth-fold estimated
gene Xow from the Abrolhos Islands to Scott Reef, which
appears to be older than the predominant southerly direc-
tion of contemporary gene Xow between these WA loca-
tions. This is consistent with the contemporary strong
southward Xowing Leeuwin current that acts as a conveyor
belt and delivers larvae from the Scott Reef population to
the Abrolhos Islands population. It also suggests that there
is no evidence of an ancestral Scott Reef population and
that previous Scott Reef populations have been largely lost
and replaced more recently by a younger lineage, possibly
from Indonesia, and subsequently this recent Scott Reef
population has provided recruits to the Abrolhos Islands
population. Thus “reverse” gene Xow northwards is not
due to contemporary P. leopardus larval transport by an
inshore counter-current. This has immediate relevance to
the management of the west Australian populations, albeit
based on a single molecular marker, since contemporary
population structure indicates that Scott Reef is an impor-
tant source population to the Abrolhos, which is a sink
population. Further implications for Wshery managers are
that each regional population should be treated as a dis-
crete population, since inter-regional levels of geneXow at
ecological time scales are insuYcient to replenish popula-
tions from distant regional sources. If regional populations
of P. leopardus are to be managed sustainably, the reliance
on self-recruitment at the regional scale must be recogni-
sed. Such self-recruitment is likely facilitated by spawning
behaviour rather than pelagic larval duration (PLD, 25–
35 days, Doherty et al. 1994; Williamson and Evans
unpublished) per se, as large primary aggregations are
formed at speciWc locations around the new moon phase
and spawning events coincide with minimal current Xow
(Samoilys 1997). Failure to recognise the reliance of Wsh
populations on self-recruitment may have been a contribut-
ing factor in the decline of many marine Wsheries (e.g.
Roberts 1997; Pauly et al. 2002; Hutchings 2000; Myers
and Worm 2003).
Conclusion
This study has identiWed strong population genetic structure
between and within four regional locations. Even in the
absence of comprehensive sampling throughout the distri-
bution range of this species, we propose, on the basis of our
Wndings, that there are probably at least another six geneti-
cally diVerentiated regional populations, in addition to the
six populations identiWed here. The additional six regional
populations we suggest are to the east of Australia and New
Caledonia at Fiji and the Caroline Islands; to the north of
Australia in Indonesia, the Solomon Islands, the Phillip-
pines; and north of Taiwan, in southern Japan. The exten-
sive population genetic subdivision identiWed here suggests
relatively high levels of self-recruitment at regional scales
are responsible for population replenishment, interspersed
sporadically over evolutionary timeframes by recruitment
from other nearby sources that are generally located “up-
current”. The direction of recruit replenishment was shown
to be along the lines of the major contemporary oceanic
currents, but may change when or if oceanic current direc-
tions change, as they have, albeit rarely, over recent evolu-
tionary timescales, within the last 50–113 kya.
Regional Wsheries for this species should be carefully
managed to prevent populations from becoming extirpated
and Wsheries from collapsing, since the levels of gene Xow
identiWed at the spatial and temporal scales examined here,
suggests that over-exploited populations will suVer from
extensive reductions in recruitment, due to their depen-
dence on self-recruitment from within the region. For
example, if any of the identiWed populations of P. leopar-
dus suVer from localised depletion (removal of spawners
and older age classes) and thus suVer reduced levels of
regional recruitment (no juveniles available in subsequent
years), then the recovery of the population will depend on
the rate of supply of external recruits. Under this scenario,
it will be diYcult for regional populations to recover as
levels of contemporary geneXow are insuYcient to replenish
populations from external regional sources and the recovery
cycle (if any) will be in the order of decades or longer.
As such, Wsheries management agencies should aim to
maintain demographic structures and adequate levels of the
spawning biomass of P. leopardus within each regional
population. In Western Australia, there is a clear need for
prudent and robust management of the Scott Reef area given
its key role in the phylogeography of P. leopardus and its
contribution to the Abrolhos population. Moreover, as Scott
Reef is in the middle of an oil and gas development area
there is an urgent need to develop rational management
plans for the conservation of P. leopardus populations.
Further work is currently underway to expand the
regional sampling presented here and to examine levels of
self-recruitment and population partitioning in groupers,
1606 Mar Biol (2009) 156:1595–1607
123
using combined larval tagging, demographic and additional
genetic approaches.
Acknowledgments This study was funded by grants to JHC and
LVH from James Cook University’s Competitive Research Incentive
Grant Scheme; the Department of Environment, Heritage, Water and
Arts for funding collection and processing of material from the Coral
Sea; and the Department of Fisheries, Government of Western Austra-
lia for logistical support to collect samples from WA. We thank JP
Hobbs, Craig Skepper, Kim Nardi, Chris Dibden, Glenn Almany and
Will Robbins, for assisting in sample collection from WA and the
Coral Sea. Brett Molony and Vanessa Messmer kindly provided sam-
ples from New Caledonia. We thank Emmanuelle Botte for extracting
DNA from new tissue samples from Scott Reef, which were used in
this study. Final thanks to David Williamson and Richard Evans for
unpublished observations pertaining to PLD values of coral trout.
References
Bay LK, Choat JH, van Herwerden L, Robertson DR (2004) High
genetic diversities and complex genetic structure in an Indo-
PaciWc tropical reef Wsh (Chlorurus sordidus): evidence of an
unstable evolutionary past? Mar Biol 144:757–767
Beerli P, Felsenstein J (2001) Maximum likelihood estimation of a
migration matrix and eVective population sizes in n subpopula-
tions by using a coalescent approach. Proc Natl Acad Sci USA
98:4563–4568
Benzie JAH (1999) Genetic structure of coral reef organisms: ghosts of
dispersal past. Am Zool 39:131–145
Brinkman R, Wolanski E, Deleersnijder E, McAllister F, Skirving W
(2002) Oceanic inXow from the Coral Sea into the Great Barrier
Reef. Estuar Coast Shelf Sci 54:655–668
Chenoweth SF, Hughes JM, Keenan CP, Lavery S (1998) When
oceans meet: a teleost shows secondary intergradation at an Indi-
an–PaciWc interface. Proc R Soc Lond B 265:415–420
Crandall ED, Frey MA, Grossberg RK, Barber PH (2008) Contrasting
demographic history and phylogeographical patterns in two Indo-
PaciWc gastropods. Mol Ecol 17:611–626
Doherty PJ, Fowler AJ, Samoilys MA, Harris DA (1994) Monitoring
the replenishment of coral trout (Pisces: Serranidae) populations.
Bull Mar Sci 54(1):343–355
Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary
analysis by sampling trees. BMC Evol Biol 7:214
ExcoYer L, Smouse PE, Quattro JM (1992) Analysis of molecular
variance inferred from metric distances among DNA haplotypes:
application to human mitochondrial DNA restriction data. Genet-
ics 131:479–491
Godfrey JS, Ridgway KR (1985) The large-scale environment of the
poleward-Xowing Leeuwin current, Western Australia: longshore
steric height gradients, wind stresses and geostrophic Xow. J Phys
Oceanogr 15:481–495
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
Heemstra PC, Randall JE (1993) FAO species catalogue, vol 16.
Groupers of the world (family Serranidae, subfamily Epinepheli-
nae). An annotated and illustrated catalogue of the grouper, rock-
cod, hind, coral grouper and lyretail species known to date. FAO
Fish Synop 125(16):382
Horne JB, van Herwerden L, Choat JH, Robertson DR (2008) High
population connectivity across the Indo-PaciWc: congruent lack of
phylogeographic structure in three reef Wsh congeners. Mol
Phylogenet Evol 49:629–638
Hubbard DK, Zankl H, Van Heerden I, Gill IP (2005) Holocene Reef
Development along the Northeastern St. Croix Shelf, Buck
Island, U.S. Virgin Islands. J Sediment Res 75:97–113
Hutchings JA (2000) Collapse and recovery of marine Wshes. Nature
406:534–536
Jensen JL, Bohonak AJ, Kelley ST (2005) Isolation by distance, web
service. BMC Genetics 6: 13. v.3.15. http://ibdws.sdsu.edu/)
Jones GP, McCormick MI, Srinivasan M, Eagle JV (2004) Coral
decline threatens Wsh biodiversity in marine reserves. Proc Natl
Acad Sci USA 101:8251–8253
Klanten O, Choat J, van Herwerden L (2007) Extreme genetic diversity
and temporal rather than spatial partitioning in a widely distrib-
uted coral reef Wsh. Mar Biol 150:659–670
Knittweis L, Kraemer WE, Timm J, Kochzius M (2009) Genetic struc-
ture of Heliofungia actiniformis (Scleractinia: Fungiidae) popula-
tions in the Indo-Malay Archipelago: implications for live coral
trade management eVorts. Conserv Genet 10:241–249
Kocher TD, Thomas WK, Meyer A, Edwards SV, Pääbo S, Villablanca
FX, Wilson AC (1989) Dynamics of mitochondrial DNA evolu-
tion in animals: ampliWcation and sequencing with conserved
primers. Proc Natl Acad Sci USA 86:6196–6200
Larcombe P (2004) Holocene Great barrier Reef: sedimentary controls
and implications for environmental management. In: Gostin VA
(ed) Gondwana to greenhouse: Australian environmental geosci-
ence. Geological Society of Australia Special Publication 21
Leis JM, Carson-Ewart BM (1999) In situ swimming and settlement
behaviour of larvae of an Indo-PaciWc coral-reef Wsh, the coral
trout Plectropomus leopardus (Pisces: Serranidae). Mar Biol
134:51–64
Lukoschek V, Waycott M, Marsh H (2007) Phylogeography of the
olive sea snake, Aipysurus laevis (Hydrophiinae) indicates
Pleistocene range expansion around northern Australia but low
contemporary gene Xow. Mol Ecol 16:3406–3422
McCaVerty S, Bermingham E, Quenouille B, Planes S, Hoelzer G,
Asoh K (2002) Historical biogeography and molecular systemat-
ics of the Indo-PaciWc genus Dascyllus (Teleostei: Pomacentri-
dae). Mol Ecol 11:1377–1392
Meyers G (1996) Variation of Indonesian throughXow and the El
Niño-Southern Oscillation. J Geophys Res 101:12255–12264
Munday PL (2004) Habitat loss, resource specialization, and extinction
on coral reefs. Glob Chang Biol 10:1642–1647
Myers RA, Worm B (2003) Rapid worldwide depletion of predatory
Wsh communities. Nature 423:280–283
PandolW JM, Tudhope AW, Burr G, Chappell J, Edinger E, Frey M,
Steneck R, Sharma C, Yeates A, Jennions M, Lescinsky H,
Newton A (2006) Mass mortality following disturbance in Holo-
cene coral reefs from Papua New Guinea. Geology 34:949–952
Pauly D, Christensen V, Gúenette S, Pitcher T, Sumaila UR, Walters
C, Watson R, Zeller D (2002) Towards sustainability in world
Wsheries. Nature 418:689–695
Posada D, Crandall KA (2001) IntraspeciWc gene genealogies: trees
grafting into networks. Trends Ecol Evol 16:37–45
Roberts CM (1997) Connectivity and management of Caribbean coral
reefs. Science 278:1454–1457
Rohlf (1973) Algorithm 76. Hierarchical clustering using the mini-
mum spanning tree. Comput J 16:93–95
Rousset F (1997) Genetic diVerentiation and estimation of gene Xow
from F-Statistics under isolation by distance. Genetics 145:1219–
1228
Russ GR, Cheal AM, Dolman AM, Emslie MJ, Evans RD, Miller I,
Sweatman H, Williamson DH (2008) Rapid increase in Wsh num-
bers follows creation of world's largest marine reserve network.
Curr Biol 18:514–515
Sambrook J, Fritch EF, Maniatis T (1989) Molecular Cloning: a
Laboratory Manual, 2nd edn. Cold Spring Harbour Laboratory
Press, Cold Spring Harbour
Mar Biol (2009) 156:1595–1607 1607
123
Samoilys MA (1997) Periodicity of spawning aggregations of coral
trout Plectropomus leopardus (Pisces:Serranidae) on the northern
Great Barrier Reef. Mar Ecol Prog Ser 160:149–159
Schneider S, Roessli D, ExcoYer L (2000) A software for population
genetics data analysis (ARLEQUIN)
Slatkin M (1987) GeneXow and the geographic structure of natural
populations. Science 236:787–792
Smith SV, Buddemeier RW (1992) Global change and coral reef eco-
systems. Ann Rev Ecol Syst 23:89–118
Smith WL, Craig MT (2007) Casting the Percomorph net widely: the
importance of broad taxonomic sampling in the search for the
placement of Serranid and Percid Fishes. Copeia 2007(1):35–55
SwoVord DL (1999) PAUP*: phylogenetic analysis using parsimony.
Sinauer Associates, Champaign
SwoVord DL (2000) PAUP*. Phylogenetic analysis using parsimony
(*and other methods). Version 4. 10b. Sinauer Associates,
Sunderland
van Herwerden L, Davies CR, Choat JH (2002) Phylogenetic and evolu-
tionary perspectives of the Indo-PaciWc grouper, Plectropomus spe-
cies on the Great Barrier Reef, Australia. J Fish Biol 60:1591–1596
van Herwerden L, Choat JH, Dudgeon CL, Carlos G, Newman SJ,
Frisch A, van Oppen MJH (2006) Contrasting patterns of genetic
structure in two species of the coral trout Plectropomus (Serrani-
dae) from east and west Australia : Introgressive hybridisation or
ancestral polymorphisms. Mol Phylogenet Evol 41:420–435
Van Oppen MJH, Gates R (2006) Conservation genetics and the resil-
ience of reef-building corals. Mol Ecol 15:3863–3883
Voris HK (2000) Maps of Pleistocene sea levels in Southeast Asia:
shorelines, river systems and time durations. J Biogeogr 27:1153–
1167
Webster JM, Beaman RJ, Bridge T (2008) From Corals to Canyons:
the Great Barrier Reef margin. EOS Trans Am Geophys Union
89:217–218
Weir BS (1996) Genetic data analysis II: methods for discrete genetic
data. Sinauer Assoc., Inc., Sunderland, MA, USA
Weir BS, Cockerham CC (1984) Estimating F statistics for the analysis
of population structure. Evolution 38:1358–1370
Welch DJ, Mapstone BD, Begg GA (2008) Spatial and temporal vari-
ation and eVects of changes in management in discard rates from
the commercial reef line Wshery of the Great Barrier Reef, Austra-
lia. Fish Res 90:247–260
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
Zwickl DJ (2006) Genetic algorithm approaches for the phylogenetic
analysis of large biological sequence datasets under the maximum
likelihood criterion. PhD dissertation, The University of Texas at
Austin
