Population genetics and management units of invasive common carp Cyprinus carpio in the...
Transcript of Population genetics and management units of invasive common carp Cyprinus carpio in the...
Journal of Fish Biology (2009) 75, 295–320
doi:10.1111/j.1095-8649.2009.02276.x, available online at www.interscience.wiley.com
Population genetics and management units of invasivecommon carp Cyprinus carpio in the Murray–Darling
Basin, Australia
G. D. Haynes*†, D. M. Gilligan‡, P. Grewe§ and F. W. Nicholas*
*Faculty of Veterinary Science, University of Sydney, NSW 2006, Australia, ‡NSW Department ofPrimary Industries (Fisheries), Batemans Bay, NSW 2536, Australia and §CSIRO Division of
Marine Research, Castray Esplanade, Hobart, TAS 7000, Australia
(Received 14 November 2008, Accepted 05 March 2009)
Common carp Cyprinus carpio were introduced into Australia on several occasions and are nowthe dominant fish in the Murray-Darling Basin (MDB), the continent’s largest river system. In thisstudy, variability at 14 microsatellite loci was examined in C. carpio (n = 1037) from 34 sitesthroughout the major rivers in the MDB, from 3 cultured populations, from Prospect Reservoir inthe Sydney Basin and from Lake Sorrell in Tasmania. Consistent with previous studies, assignmenttesting indicated that the Boolara, Yanco and koi strains of C. carpio are present in the MDB.Unique to this study, however, the Prospect strain was widely distributed throughout the MDB.Significant genetic structuring of populations (Fisher’s exact test, AMOVA and distribution ofthe different strains) amongst the MDB sub-drainages was detected, and was strongly associatedwith contemporary barriers to dispersal and population history. The distributions of the strainswere used to infer the history of introduction and spread of C. carpio in the MDB. Fifteenmanagement units are proposed for control programmes that have high levels of genetic diversity,contain multiple interbreeding strains and show no evidence of founder effects or recent populationbottlenecks. © 2009 The Authors
Journal compilation © 2009 The Fisheries Society of the British Isles
Key words: Boolara; freshwater fish; invasive species; koi carp; Prospect Reservoir; Yanco.
INTRODUCTION
Common carp Cyprinus carpio are a highly invasive species of freshwater fish.Native to Eurasia, they have been successfully introduced to parts of the Americas,Oceania, Africa, Asia, Europe and Australia (Koehn, 2004). Cyprinus carpio havebeen introduced into Australian rivers several times since the late 19th century(Anderson, 1920; Clements, 1988; Koehn et al., 2000) and have spread fromintroduction sites through natural range expansions and through intentional andaccidental releases (Koehn et al., 2000). They have been in the Murray–DarlingBasin (MDB), Australia’s largest river system, since at least 1917 (Anderson,
†Author to whom correspondence should be addressed. Tel.: +61 2 9351 4789; fax: +61 2 9351 3957;email: [email protected]
295© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles
296 G . D . H AY N E S E T A L .
1920; Clements, 1988). After extensive flooding in 1974–1975, C. carpio numbersincreased sharply, and C. carpio became the dominant species in the MDB (Harris& Gehrke, 1997; Koehn et al., 2000). There is much interest in C. carpio control,because C. carpio have a detrimental effect on the aquatic environment and areconsidered a pest in most Australian states (Koehn et al., 2000). The extent ofpopulation sub-structure (i.e., single panmitric unit or a single independent sub-population is present) can be a useful guide for implementing pest-managementstrategies, and can be assessed through the statistical analysis of data from multiplegenetic loci.
Previous genetic studies indicated the presence of at least four C. carpio strainsin Australia: Prospect, Yanco, Boolara and koi (Shearer & Mulley, 1978; Daviset al., 1999). The Prospect strain was founded in Sydney from 14 fingerlingsof unknown origin in 1907–1908 (Stead, 1929) and was used to seed severalwaterways in the Sydney Basin (Clements, 1988). The Yanco strain was introducedinto the MDB between 1910–1950 (Brown, 1996). Individuals of this strain wereoriginally a distinctive orange colour (Shearer & Mulley, 1978), a trait which isnow rarely observed in the MDB carp (K. Bell, pers. comm.). Interbreeding withother strains, and possibly natural selection, has presumably led to the replacementof this colouration with the wild-type phenotype in contemporary populations. TheBoolara strain was probably illegally imported from Germany in the late 1950s,was deliberately spread throughout Victoria and invaded the Murray River in 1968(Clements, 1988; Koehn et al., 2000). Koi are an ornamental strain of C. carpiofrom Japan (Balon, 1995), sometimes illegally released into waterways (Koehnet al., 2000; Graham et al., 2005). Previous studies detected Yanco C. carpio attwo sites and koi at one site in the MDB, the Boolara strain throughout the MDBand the Prospect strain only in the Sydney Basin (Shearer & Mulley, 1978; Daviset al., 1999). The introduction history of these strains may provide insights into thecontemporary genetic structuring of C. carpio in the MDB.
In the present study, repeat-length variability in 14 microsatellite loci was surveyedto determine the distributions of the various strains, to estimate the extent of geneticstructuring between sub-drainages and to assess levels of genetic diversity withinthe MDB. The distributions of the various strains are interpreted in conjunctionwith historical and demographic data to infer the history of colonization andexpansion of carp in the MDB since their introduction. In addition, the microsatellitevariability between sub-drainages is used to identify barriers to migration which,when considered with the geography of the region, is used to define managementunits that can inform strategies for control programs.
MATERIAL AND METHODS
S A M P L E C O L L E C T I O N
Common carp were collected by electrofishing from March 2004 to October 2006. A finclip was taken from each individual and immediately placed in 70% ethanol. Effort wasmade to collect at least 30 fish from each major river catchment in the MDB. Sampleswere collected upstream and downstream of major dams to assess the effect of the dams onmigration. Cyprinus carpio were also sampled from Lake Sorell, Tasmania, where they werefirst reported in 1995 (Koehn et al., 2000). Prospect strain C. carpio were collected from
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 297
Prospect Reservoir in the Sydney catchment, and koi were obtained from two fish breeders,one in Germany and one in Sydney. Mirror-scale domestic C. carpio were obtained from afish farm in Jaenschwalde, Germany, to represent ‘pure’ European carp that have not interbredwith non-European strains. Sample site names and coordinates, and sample sizes, are givenin Table I. Sample site locations are given in Fig. 1.
P C R A N D G E N OT Y P I N GDNA was extracted according to Wasko et al. (2003), and samples were genotyped for 14
di-, tri- or tetranucleotide microsatellite loci, including Cca02, Cca07, Cca09, Cca17, Cca19,Cca65, Cca67, Cca72, GF1, Koi5-6, Koi29-30, Koi41-42, MFW6 and MFW26 (Zheng et al.1995; Crooijmans et al., 1997; David et al., 2001; Yue et al., 2004). Microsatellite DNAwas amplified using the polymerase chain reaction (PCR) in eight single-locus and threemultiplex reactions. Primers for Cca65, Cca09, Cca07, Cca17, Koi5-6, Koi29-30 and Koi41-42 were redesigned to anneal at higher temperatures and to change the size of the PCRproducts to facilitate multiplexing. Non-template sequence was included in the 5′ end ofsome primers to facilitate multiflex PCR Shuber et al., 1996 and accurate allele size scoriusBrownstein et al., 1995 Primer sequences appear in Appendix I. Optimal conditions for eachPCR consisted of 1 μl (10–100 ng) total genomic DNA, 1x PCR buffer (Fisher Biotech;www.fisherbiotech.com), 200 μM each dNTP, 1 unit Taq DNA polymerase, primer and MgCl2concentrations (Appendix I) and sterile water to 15 μl total volume. PCR amplificationswere made with touch-down protocols (Appendices I and II). PCR products were pooledinto two groups and genotyped using an ABI 3730 DNA Analyzer (Applied Biosystems;www.appliedbiosystems.com). Genotypes were scored with GeneMapper 3.1 and checked byeye by at least two individuals.
S TAT I S T I C A L A NA LY S I S
Allelic diversityAllelic size ranges and numbers at each locus were summarized using GenAlEx 6.0
(Peakall & Smouse, 2006). Departures of genotype frequencies from Hardy–Weinberg (HW)proportions were tested in GENEPOP 4.0 (Raymond & Rousset, 1995). As a large numberof sites were tested, the HW P -values were adjusted for multiple tests using the Benjamini-Hochberg (BH) method (Benjamini & Hochberg, 1995), which has been demonstrated to berobust and effective at minimizing type 1 errors (Reiner et al., 2003).
Assignment testsAssignment tests were made with a Bayesian algorithm in Structure 2.1 (Pritchard et al.,
2000; Falush et al., 2003), which uses HW expectations and linkage disequilibrium to assignindividuals to population groups. Analyses were run for K = 1–10 potential populationgroups with 500 000 burn-in steps and 1 000 000 Markov-chain Monte-Carlo steps. The ‘allelefrequencies correlated’ and ‘use prior population information to assist clustering’ models wereused, as preliminary analyses indicated that these two models were best able to differentiatebetween the populations analysed here. Three replicates were made for each value of K .The �K statistic (Evanno et al., 2005) was used to estimate the actual number of populationgroups present (i.e. the true value of K). This statistic is the change in the log probabilityvalues [lnP (D)] between successive values of K , and when plotted against K produces a sharppeak at the most likely value of K (Evanno et al., 2005). The Prospect, koi and Jaenschwaldestrains were included in the analysis to test how effectively Structure differentiated amongisolated populations and to estimate the extent to which these strains were introduced intothe MDB. The USEPOPINFO parameter was set to 1 for these samples to indicate they werelearning samples and set to 0 for the remaining samples. Koi from Sydney and Germany werepooled in this analysis.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
298 G . D . H AY N E S E T A L .
TA
BL
EI.
Cyp
rinu
sca
rpio
site
nam
esan
dlo
catio
ns,s
ampl
esi
ze(N
)an
dpr
obab
ility
(P)
for
test
sof
fitto
over
allH
ardy
–W
einb
erg
(HW
)pr
opor
tions
adju
sted
for
mul
tiple
test
ing
usin
gth
eB
Hm
etho
d.Si
gnifi
cant
(P<
0·05)
depa
rtur
esfr
omH
Wpr
opor
tion
sar
esh
own
inbo
ld
Loc
atio
n
Sam
ple
site
Riv
erco
llect
edSt
ate
Iden
tifier
Lat
itude
(S)
Lon
gitu
de(E
)N
HW
P
Wal
gett
Bar
won
Riv
erN
SWW
G30
·017◦
148·1
00◦
300·7
38N
ynga
nB
ogan
Riv
erN
SWN
G31
·555◦
147·1
69◦
300·8
00C
oona
mbl
eC
astle
reag
hR
iver
NSW
CN
30·96
7◦14
8·381
◦30
0·229
Mud
gee
Cud
gego
ngR
iver
NSW
MG
32·34
0◦14
9·350
◦30
0·043
Bou
rke
Dar
ling
Riv
erN
SWB
K30
·089◦
145·9
38◦
90·8
41W
ilcan
nia
Dar
ling
Riv
erN
SWW
C31
·550◦
143·3
67◦
280·4
35D
enili
quin
Edw
ards
Riv
erN
SWD
Q35
·516◦
144·9
59◦
310·9
19M
oree
Gw
ydir
Riv
erN
SWM
R29
·465◦
149·8
44◦
330·1
40L
ake
Car
gelli
goL
achl
anR
iver
NSW
CW
33·26
7◦14
6·400
◦30
0·468
Wya
ngal
aD
amL
achl
anR
iver
NSW
WY
33·95
0◦14
8·933
◦30
0·020
Bur
rend
ong
Dam
Mac
quar
ieR
iver
NSW
BD
32·65
0◦14
9·100
◦30
0·897
Dub
boM
acqu
arie
Riv
erN
SWD
B32
·233◦
148·6
00◦
250·9
21W
ellin
gton
Mac
quar
ieR
iver
NSW
WN
32·53
3◦14
8·933
◦30
0·201
Ech
uca
Mur
ray
Riv
erN
SWE
C36
·116◦
144·8
05◦
300·9
00L
ake
Hum
eM
urra
yR
iver
NSW
LH
35·96
7◦14
7·067
◦31
0·000
Wen
twor
thM
urra
yR
iver
NSW
WT
34·10
5◦14
1·912
◦30
0·242
Bur
rinj
uck
Dam
Mur
rum
bidg
eeN
SWB
J34
·983◦
148·5
83◦
300·0
26C
oom
aM
urru
mbi
dgee
NSW
CM
36·23
7◦14
9·125
◦30
0·212
Nar
rand
era
Mur
rum
bidg
eeN
SWN
D34
·733◦
146·5
50◦
300·0
00L
ake
Kee
pit
Nam
oiR
iver
NSW
KP
30·85
0◦15
0·500
◦30
0·006
Nar
rabr
iN
amoi
Riv
erN
SWN
B30
·360◦
149·8
31◦
310·0
67C
onda
min
eC
onda
min
eQ
LD
CD
M27
·901◦
148·6
37◦
300·7
51Pa
roo
Riv
erPa
roo
Riv
erQ
LD
PR28
·056◦
145·3
68◦
300·9
14C
harl
evill
eW
arre
goR
iver
QL
DC
V26
·402◦
146·9
38◦
300·4
36
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 299
TA
BL
EI.
Con
tinue
d
Loc
atio
n
Sam
ple
site
Riv
erco
llect
edSt
ate
Iden
tifier
Lat
itude
(S)
Lon
gitu
de(E
)N
HW
P
Low
erL
akes
Mur
ray
Riv
erSA
LL
35·50
7◦13
8·956
◦30
0·387
Avo
caA
voca
Riv
erV
ICA
V37
·087◦
143·4
63◦
250·3
93B
roke
nR
iver
Bro
ken
Riv
erV
ICB
R36
·551◦
145·9
66◦
300·9
48C
ampa
spe
Cam
pasp
eR
iver
VIC
CS
36·48
1◦14
4·613
◦30
0·213
Gou
lbur
nG
oulb
urn
Riv
erV
ICG
B36
·718◦
145·1
65◦
300·7
11K
iew
aK
iew
aR
iver
VIC
KIW
36·11
9◦14
6·940
◦30
0·681
Lak
eE
ildon
Lak
eE
ildon
VIC
EI
37·20
8◦14
6·035
◦30
0·909
Lod
don
Lod
don
Riv
erV
ICL
D37
·082◦
144·0
13◦
300·8
90O
vens
Ove
nsR
iver
VIC
OV
36·05
6◦14
6·187
◦25
0·777
Hor
sham
Wim
mer
aR
iver
VIC
WM
36·71
8◦14
2·184
◦25
0·940
Tasm
ania
∗L
ake
Sore
llTA
STA
S42
·100◦
147·1
67◦
240·5
17Pr
ospe
ct∗
Pros
pect
Res
ervo
irN
SWP
33·81
5◦15
0·901
◦24
0·001
Jaen
schw
alde
∗na
naJ
Na
na30
0·868
Koi
(Syd
ney)
∗na
naK
bN
ana
460·0
00K
oi(G
erm
any)
∗na
naK
Na
na30
0·001
∗ Not
part
ofth
eM
DB
.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
300 G . D . H AY N E S E T A L .
Queensland
CV
CDM
MR
WG
NBKP
CN
DBWN
BDMG
WY
Burrendong Dam
Keepit Dam
Wyangala Dam
Sydney
BJND
LHKIW
BR
EI
OVEC
GB
DQ
CS
LDAVWM
WT
LL
200 km1000
Adelaide
Hume DamEildon Dam
MelbourneVictoria
Canberra
Burrinjuck Dam
CM
Brisbane
PR
BK
NG
CW
WC Darling riv
er
TAS
New South WalesSouth
Australia
Murray river
Macquarie
river
Lachlan river
Murrumbidgee river
FIG. 1. Collection sites for Cyprinus carpio. Murray–Darling Basin is indicated in white, the rest of Australiain gray. Sample-site coordinates and full names are given in Table I.
Genetic structureThe F -statistics of Weir & Cockerham (1984) were estimated with Genepop 4.0 (Raymond
& Rousset, 1995), and analysis of molecular variance (AMOVA) (Excoffier et al., 1992)performed with GenAlEx 6.0 (Peakall & Smouse, 2006). The significances of the AMOVAresults were tested against an empirical null distribution derived from 9999 permutations.As the large dams at river headwaters likely limit C. carpio dispersal, the AMOVA andF -statistic analyses were conducted on three groups: (1) all MDB samples, (2) below-damMDB samples and (3) above-dam MDB samples. In addition, departure of allele frequenciesfrom the null hypothesis of panmixia was tested for each pair of above-dam v. below-damsites (KP, BD, WY, BJ, LH and EI sites against the NB, WN, CW, ND, KIW and GB sites,respectively) using Fisher’s exact test in Genepop 4.0.
To test for isolation-by-distance population structure, geographic distances between MDBsample sites were measured in Google Earth, both ‘as the crow flies’ and following the shortestpath along river channels. Two measures of the fixation index between sub-populations, FST,were calculated between all pairs of sample sites in Arlequin 3.1 (Excoffier et al., 2005).These were Slatkin’s estimate of FST (Slatkin, 1991) and Reynolds’ estimate of FST, derivedfrom the coancestry-based genetic distance of Reynolds et al. (1983). Correlations betweengeographical and genetic distances were estimated for each combination of geographic (along-river and crow-flies) and genetic (Reynolds’ and Slatkin’s estimates of FST) distance. Twelve
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 301
combinations of sample sites were tested to account for barriers to dispersal imposed byimpoundments and limited flows (Appendix III). The Bourke (BK) site was excluded, becauseits small sample size (n = 9) could skew results. The significance of each correlation wasdetermined using a Mantel permutation test in GenAlEx 6.0, with 9999 permutations. Astests were not independent (i.e. same sample sites used in multiple tests), P -values wereadjusted for multiple tests using the Benjamini and Yekutieli (BY) procedure (Benjamini &Yekutieli, 2001) in the R-Multitest package (Pollard et al., 2008), rather than the BH procedure(Benjamini & Hochberg, 1995) used previously, as this false discovery rate correction takesinto account that P -values may not be independent.
Barriers to dispersalBarriers to dispersal were identified with Barrier 2.2 (Manni et al., 2004), which uses
geographic and genetic distances to identify genetic discontinuities between regions. Thepotential number of barriers (predefined by user) can range from 1 to the number of samplesites. Non-MDB sites were excluded from these analyses. The BK site was also excluded,because of its small sample size. Barrier was run for each of the two measures of geneticdistance (Reynolds’ and Slatkin’s estimates of FST). Bootstrap values for each barrier weregenerated by sub-sampling with replacement from each sample to generate 100 randomlyre-sampled datasets, by computing a genetic distance matrix for each of the 100 re-sampleddatasets and by analysing these matrices with Barrier. Bootstrap values for each apparentbarrier can range from 1 (barrier detected in one of the re-sampled matrices) to 100% (barrierdetected from each of the 100 re-sampled matrices). Bootstrap values were arbitrarily classedas strong (>80%), weak (40–79%) or not significant (<40%).
Defining management unitsManagement units can be defined as populations ‘connected by such low levels of gene
flow that they are functionally independent’, at least on the time scale relevant to short-termmanagement, and identified by the presence of divergent allele frequencies between regions(Moritz, 1994). Management units were proposed in this study based on genetic differentiationbetween regions implicit in the assignment tests (i.e. different population groups present indifferent regions), on genetic discontinuities being consistently detected by Barrier for thetwo genetic distance measures and on the known physical barriers to dispersal (primarilycatchment boundaries within the MDB). As the dams at river headwaters almost certainly limitC. carpio dispersal, the level of bootstrap support for the barriers detected by Barrier betweenthe above-dam and below-dam sites was used as a guide to the minimal level of bootstrapsupport necessary to delimit a management unit from the Barrier results. Consistency betweenassignment tests and Barrier analysis was desirable, but not strictly necessary to delimit amanagement unit.
Genetic diversity and population bottlenecksGenetic diversity was estimated as allele richness (Ar), mean number of alleles per
locus (A) and observed (HO) and expected (HE) heterozygosity. These measures werecalculated for the MDB as a whole, for the proposed management units (see Discussion)and for the Tasmania, Prospect, koi and Jaenschwalde samples. Ar, A and HE were estimatedwith HP-Rare (Kalinowski, 2005) and HO with Genepop 4.0 (Raymond & Rousset, 1995).For Ar, rarefaction was used to adjust for different sample sizes. As the smallest groupanalysed consisted of 24 individuals, the number of genes per locus was set to 48 for thiscalculation.
Departures from mutation-drift equilibrium indicative of a recent population bottleneck(inflated heterozygosity relative to heterozygosity expected at mutation-drift equilibrium)were tested using Bottleneck 1.2.02 (Cornuet & Luikart, 1996; Pirey et al., 1999). A two-phase model (TPM) of mutation was used, and significance was assessed with a two-tailedWilcoxon sign rank test, which provides relatively large power with as few as four loci.Departures from expected values under mutation-drift equilibrium were tested for the MDB
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
302 G . D . H AY N E S E T A L .
as a whole, separately for each proposed management unit and for Tasmanian, Jaenschwalde,koi and Prospect C. carpio. Koi from Sydney and Germany were analysed separately. P -values were adjusted for multiple testing using the BH method (Benjamini & Hochberg,1995), and adjusted values of <0·05 were considered significant.
RESULTS
A L L E L E D I V E R S I T Y
The number of detected alleles ranged from 4 (GF1 and Cca07) to 17 (MFW26),with allele size ranges being consistent with size ranges reported in the literature(Zheng et al. 1995; Crooijmans et al. 1997; David et al., 2001; Yue et al., 2004;Hanfling et al., 2005), expect in some instances where primers were re-designedto anneal in different regions (Table II). Nine of the 39 sample sites showed asignificant (P<0·05) overall departure from HW after adjustment for multiple testing(Table I).
A S S I G N M E N T T E S T S
The graph of �K against K produced a single, distinctive peak at K = 5 (datanot shown), indicating the presence of five population groups in the analysis. Thedistribution of these population groups is illustrated in Fig. 2. Jaenschwalde and koiC. carpio corresponded closely to population groups 2 and 3, respectively. ProspectC. carpio correspond most strongly with group 1, although about 30% of theiroverall genetic variation was assigned to group 2. Population group 4 is distributedubiquitously throughout the MDB and is the dominant group in Victoria. Populationgroup 5 is also widely distributed and is dominant in the Murrumbidgee catchment(ND site). A high percentage (59%) of individuals from the MDB and Tasmania
TABLE II. Cyprinus carpio microsatellite alleles and allelic size ranges detected
Size range
Locus Reported∗ Detected Number of alleles
Cca02 173–194 159–205 12Cca09a 303–387 332–380 11Cca65a 184–194 150–160 5Cca72 244–299 237–304 12GF1 337–353 335–376 4Koi41-42a 228 285–316 6MFW26 122–150 125–170 17MFW6 144–152 116–168 15Cca07a 216–245 224–236 4Cca17a 322–367 371–389 5Cca19 262–370 291–299 5Cca67 228–254 231–267 11Koi29-30a 247 334–344 5Koi5-6a 189 234–255 6
∗References for allelic sizes are listed in Table III. Reported ranges differ greatly from detected rangesin some cases because primers were re-designed to anneal in different regions.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 303
Population groupsPopulation group 1 (Prospect strain)Population group 2
(Jaenschwalde–Prospect strain)
Population group 3 (koi)Population group 4 (Boolara strain)
Population group 5 (Yanco strain)
New South Wales
Non-MDB samples
Prospect Reservoir
Jaenschwalde
Koi
Tasmania
Adelaide
Victoria Melbourne
Sydney
Brisbane
0 100 200 km
FIG. 2. Assignment results from Structure for K = 5 population groups. Pie diagrams indicate the overallproportions of each population group (1–5) to the genetic diversity of each sample site.
were allocated to more than one population group. The distribution of the populationgroups suggested approximately 11 genetically different regions in the MDB.
G E N E T I C S T RU C T U R E
Significant allele frequency differences were detected among sample sites. TheAMOVA showed significant variation among sites (Table III), with 11% of variationamong sites and 89% within sites in the MDB overall. As expected, the percentage ofamong-site variation was smaller (7%) among below-dam samples and larger (20%)among above-dam samples. F -statistics also indicated that population structuring wasgreatest among the above-dam samples (FST = 0·1724), lowest among the below-dam samples (FST = 0·0384) and intermediate among all samples (FST = 0·0720).All exact test-comparisons between the above- and below-dam samples were highlysignificant (P<0·001).
In the plots of genetic distance against geographic distance that were generated totest for isolation by distance, the data points showed little scatter about the y-axis(genetic distance) (data not shown). None of the 48 correlations between geographicand genetic distance was significant after BY adjustment (Benjamini & Yekutieli,2001) for multiple testing.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
304 G . D . H AY N E S E T A L .
TABLE III. F -statistics (Weir & Cockerham, 1984) and AMOVA results. Statistics werecalculated across all 14 microsatellite loci
F -statistics AMOVA
Variation VariationComparison FST FIS FIT within sites (%) among sites (%) P
All MDB sites 0·0720 0·0237 0·0940 89 11 0·010Below-dam
MDB sites0·0384 0·0043 0·0426 93 7 0·010
Above-damMDB sites
0·1724 0·0990 0·2543 79 21 0·001
Barriers to dispersalBarriers to dispersal identified by Barrier were similar for Slatkin’s and Reynolds’
FST, differing more in bootstrap values than in location (Fig. 3). Since the population-assignment results from Structure indicated 11 genetically differentiated regions inthe MDB (Fig. 2), the results for the detection of 12 barriers were used, as theseallowed the detection of these discontinuities along with an additional boundary notidentified by Structure. Strong barriers (>80% bootstrap support) were consistentlydetected around the Broken, Campaspe and Goulburn rivers in Victoria (sites BR,CS and GB, respectively), the Murrumbidgee catchment (ND), the Paroo andWarrego Rivers (PR and CV) and Lake Eildon (EI) and Wyangala (WY) dams.Combinations of weak (40–79% support) and strong barriers were detected aroundthe Macquarie River sites (DB and WN), between the Avoca (AV) and Loddon(LD) Rivers and the rest of the MDB, and Burrinjuck (BJ, CM), Burrendong (BD,MG) and Lake Keepit (KP) dams. Both FST measures indicated weak barriersaround Lake Hume (LH), Burrendong Dam (BD) and the Condamine River (CDM),and between the upper (OV, KIW) and mid-Murray (EC, DQ). Slatkin’s FST alsodetected a strong barrier between the Wimmera catchment (WM) and the rest ofthe MDB. Minimal bootstrap support for a barrier to delimit a management unitwas set at 41%, as this was the lowest bootstrap value for a barrier detectedbetween above and below-dam sites (Slatkin’s FST, between the LH and KIWsites).
G E N E T I C D I V E R S I T Y A N D P O P U L AT I O N B OT T L E N E C K S
No significant departures from mutation-drift equilibrium (P<0·05) were detectedfor any management unit by BOTTLENECK after adjustment for multiple test-ing (data not shown). For management units, Ar ranged from 2·1 to 4·0, A
from 2 to 5, HO from 0·179 to 0·467 and HE from 0·182 to 0·498 (Fig. 4).Genetic diversity was highest in the Murrumbidgee catchment, and lowest inthe Wimmera catchment (A and Ar), Lake Keepit (A and Ar) and Burren-dong Dam (all measures) management units. When the MDB is considered over-all, Ar and A are much higher than in the individual management units, bothbeing 8·3.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 305
(a)
Queensland
New South WalesSouth
Australia
New South WalesSouth
Australia
Adelaide
Adelaide
43
5652 97
9480
MelbourneVictoria
MelbourneVictoria
100
91
83 4743
78 64 46Canberra
Brisbane
83
6988
94
94
67
64
94
884378
95
82
8088
Sydney
50
86
100
0 100 200 km
0 100 200 km
89
9493100
10098
100
100
100
41
42
8383
98
71
93
80
80
43
647762
5770
99
99100
9854
99100
(b)
Queensland
Reynolds' FST
Slatkin's FST
Canberra
Sydney
Brisbane
FIG. 3. Putative barriers to dispersal of Cyprinus carpio calculated from (a) Reynolds’ estimate of FST and(b) Slatkin’s estimate of FST. Polygons around each sample site represent the Voronı tessellations drawnaround each sample site by Barrier. Bold lines represent putative barriers to dispersal. The level ofbootstrap support for each barrier is indicated by the number associated with the barrier and by thethickness of the barrier. Bootstrap values <40 are not shown.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
306 G . D . H AY N E S E T A L .
Ar
A
HE
HO
Management unit or population
Management unit or population
Gen
etic
div
ersi
tyG
enet
ic d
iver
sity
(a) 16·0
14·0
12·0
10·0
8·0
6·0
4·0
2·0
0·0
1·000(b)
0·800
0·600
0·400
0·200
0·000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
FIG. 4. Genetic diversity in Cyprinus carpio in the MDB. (a) Allele richness (Ar) and the mean number ofalleles per locus (A). (b) Observed and expected heterozygosity (HO and HE, respectively). Geneticdiversity indices from this study are shown in comparison with published data from C. carpio in theirnative range, other invasive species of freshwater fish and freshwater fish in general. Allele richnesswas not reported in the previous studies, and HO was not reported for freshwater fish in general. Datafrom C. carpio in Australia are given for the MDB as a whole and for each individual managementunit (Fig. 5). (1) all MDB, (2) Paroo-Warrego catchments, (3) Condamine catchment, (4) Macquariecatchment, (5) Main MDB, (6) Wimmera catchment, (7) Avoca–Loddon catchments, (8) Murrumbidgeecatchment, (9) central Victoria, (10) upper Murray, (11) Burrendong Dam, (12) Lake Keepit, (13)Wyangala Dam, (14) Lake Eildon, (15) Lake Hume, (16) Burrinjuck Dam, (17) Tasmania, (18) ProspectReservoir, (19) Jaenschwalde, (20) koi (Sydney fish farm), (21) koi (German fish farm), (22) C. carpio,European, wild*, (23) C. carpio, European, domestic*, (24) C. carpio, Central Asian, wild*, (25) C.carpio, east Asian, wild*, (26) Petromyzon marinus**, (27)Poecilia reticulata†, (28) Freshwater fishoverall††. *Kohlmann et al. (2005), **Bryan et al. (2005), †Lindholm et al. (2005) and ††DeWoody &Avise (2000).
DISCUSSION
This research is the most comprehensive population genetic study of C. car-pio in a single river basin to date. Consistent with the findings of previousAustralian studies, this study confirmed that the Boolara, Yanco and koi strainsare present in the MDB (Shearer & Mulley, 1978; Davis et al., 1999). Theresults of this study also showed that the Prospect strain is widely distributedthroughout the MDB. Significant genetic structuring appears across the MDBand is strongly associated with contemporary barriers to dispersal. Levels ofgenetic variation in the MDB were similar to those in domestic populations(koi and Jaenschwalde), indicating that C. carpio are not genetically depauper-ate in Australia. A history of introduction and spread of the various C. car-pio strains in Australia is proposed below, based on the current distribution
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 307
of the strains. The MDB is divided into 15 management units for controlprogrammes, each corresponding to natural or man-made barriers to C. carpio dis-persal.
S T R A I N S O F CYPRINUS CARPIO I N T H E M U R R AY – DA R L I N GBA S I N
Five population groups of carp were identified with Structure (Fig. 2). Groups 1,2 and 3 probably represent the Prospect, Jaenschwalde and koi carp, respectively,as these strains correspond most closely with these groups. The imperfect separationbetween groups 1 and 2 in the Prospect strain is likely to be a result of their smallsample size (24) of Prospect individuals, the limited number of microsatellite loci (14)and genetic similarity between Prospect and Jaenschwalde C. carpio. Group 4 likelyrepresents the Boolara strain, as it is ubiquitously distributed throughout the MDBand is the dominant group in Victoria (Davis et al., 1999). Group 5 likely representsthe Yanco strain, as it is the dominant group at Narrandera in the Murrumbidgeecatchment (ND site), close to where Shearer & Mulley (1978) caught the Yancostrain individuals in their study. The ability of Structure to detect these strains inthe MDB, despite several generations of potential interbreeding, may stem from thelongevity of C. carpio. Older individuals of ‘pure’ strain ancestry may have beencaught alongside younger, intercrossed progeny, as carp over 50 years in age havebeen caught in the wild (P. Sorenson, pers. comm.).
P O P U L AT I O N G E N E T I C S T RU C T U R E
Significant variation among sites (AMOVA) and the heterogeneous distributionof the strains indicated that C. carpio in the MDB exhibit considerable populationgenetic structure. Dams play a role in limiting gene flow, as among-site variationmeasured by AMOVA was greater when samples from above dams were includedthan when only below-dam samples were analysed. All pairwise comparisons ofallele frequencies between above and below-dams sites showed highly significantdepartures from panmixia. This genetic structuring is not associated with isolationby distance. The lack of scatter around the y-axis (genetic distance) in the plots ofgenetic distance against geographic distance is similar to a relationship theoreticallyand empirically demonstrated by Hutchinson & Templeton (1999), in which thereis a lack of regional equilibrium, and migration and gene flow play a largerrole in shaping genetic structure than does genetic drift. The pattern of geneticstructure can therefore be attributed to contemporary barriers to dispersal that limitmigration and gene flow, as well as historical patterns of introduction and rangeexpansion.
G E N E T I C D I V E R S I T Y
Although many invasive species show decreased levels of genetic diversity intheir introduced range relative to their native range (e.g. Hamner et al., 2007),some invasives have comparable or greater levels of genetic diversity, because theyoriginated from multiple source populations, or rapid population growth followedestablishment so that the loss of genetic diversity through drift was minimized
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
308 G . D . H AY N E S E T A L .
(Zenger et al., 2003; Frankham, 2005; Hanfling, 2007). Cyprinus carpio in the MDBgenerally have high levels of genetic diversity, with multiple strains detected inall regions, a high percentage (59%) of individuals showing mixed-strain ancestryand no evidence for a recent population bottleneck. Only 3 of the 15 managementunits (Burrendong Dam, Lake Keepit and the Wimmera catchment) showed greatlyreduced A, Ar, HE or HO relative to the domestic populations (koi and JaenschwaldeC. carpio) analysed here. The high level of genetic diversity in the Murrumbidgeecatchment management unit is consistent with the presence of a self-sustainingpopulation of Yanco strain C. carpio before the introductions of the Boolara andProspect strains. Overall values of A and Ar in MDB populations are greaterthan in domestic populations in Europe (Kohlmann et al., 2005), invasive lampreysPetromyzon marinus L. in the Great Lakes of North America (Bryan et al., 2005) andinvasive guppies Poecilia reticulata Peters in Queensland, Australia (Lindholm et al.,2005). Genetic diversity, however, is less than estimates for indigenous populationsof wild C. carpio reported by Kohlmann et al. (2005), although this may be dueto the use of a different set of microsatellite loci by Kohlmann et al. (2005). HEand HO for the management units and the MDB as a whole are also lower thanprevious estimates for wild and domestic C. carpio in their native range (Kohlmannet al., 2005), freshwater fish overall (DeWoody & Avise, 2000), and invasive P.marinus and P. reticulata, and may have resulted from the inclusion of differentstrains in the samples (Wahlund’s effect). The high level of genetic diversity ofC. carpio in the MDB may have facilitated invasiveness and adaptation to newenvironments.
H I S TO RY O F I N T RO D U C T I O N A N D R A N G E E X PA N S I O N O FCYPRINUS CARPIO
The following possibilities for the introductions and spread of C. carpio in theMDB are proposed. (1) As the Prospect strain was detected throughout the MDB, itwas probably introduced early, and perhaps expanded its range during the extensive1950s floods. (2) The widespread distribution of the Boolara strain is consistent witha range expansion during large-scale floods in 1974–1975 (Reid et al., 1997; Koehnet al., 2000), perhaps aided by heterosis (hybrid vigour) resulting from mating withthe already present Prospect strain. (3) The scarcity of the Yanco strain in someregions indicates a range expansion after the expansion of the Prospect and Boolarastrains. Prospect and Boolara C. carpio and their intercrossed progeny may not haveentered the Murrumbidgee catchment in significant numbers until the 1974–1975floods. Prospect and Boolara carps may have bred with the resident Yanco carps,resulting in further heterosis and providing the genetic diversity necessary for thedescendents of introduced Yanco carp to lose their conspicuous orange colourationand expand their range. Descendents of Yanco carp are now scarce in some of therivers in the Darling River catchment, because these rivers have remained partiallyisolated from the rest of the MDB since the 1974–1975 floods. The Yanco strainwas also possibly prevented from penetrating far into the Victorian rivers and theupper reaches of the Murray River by weirs and by the abundance of adult Boolaraand Prospect strain C. carpio already present in these regions. (4) Koi C. carpiohave been released in low numbers throughout the MDB, but have contributedlittle to the overall population. Thirty-seven C. carpio with 5–50% koi ancestry
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 309
were detected above Burrinjuck Dam (BJ and CM sites) and 7 in the samplefrom Tasmania, consistent with the detection by Davis et al. (1999) of putative koihaplotypes in Lake Burley Griffin (which is also located above Burrinjuck Dam) andin Tasmania.
The establishment of carp above six of the large dams in the MDB indicatesthat carp were either present before the dams were constructed or were introducedby humans, as these dams are too large to be submerged by flooding. Dispersalof sticky carp eggs on the feet or plumage of waterfowl has been postulated asa mechanism of dispersal (Gilligan & Rayner, 2007), although to date there is noempirical evidence to support this. The following is proposed for these populations.(1) The carp above the Eildon (EI) and Hume (LH) dams were probably introducedfrom adjacent waterways, possibly those immediately downstream, as they have asimilar strain composition to these adjacent rivers. (2) The Keepit (KP), Wyangala(WY) and Burrinjuck (BJ, CM) dam populations were probably introduced beforethe expansion of the Yanco strain, as these populations include the Prospect andBoolara strains but not the Yanco strain. (3) The reduced levels of genetic diversityand prevalence of the Prospect strain above Burrendong Dam (BD, MG) are con-sistent with a founding by a small number of Prospect strain C. carpio, that mayhave been introduced from the Sydney Basin. This strain was unlikely to have beenpresent before the construction of Burrendong Dam in 1967, as ageing data fromotoliths indicate that the oldest of 300 C. carpio caught in Burrendong Dam wasspawned in 1989 (D. M. Gilligan, unpublished data). As C. carpio can live over50 years in the wild (P. W. Sorenson, pers. comm.), the rivers above BurrendongDam were probably not populated with C. carpio prior to the dam’s construction.Whether these introductions are the results of accidental releases, through use ofC. carpio as live bait or contamination of stocked native fish with C. carpio fry(Koehn et al., 2000), or of deliberate introductions is unknown.
BA R R I E R S TO D I S P E R S A L A N D M A NAG E M E N T U N I T S
The presence of 15 discrete genetic entities that could be classified as individualmanagement units were identified by the assignment tests and Barrier analyses, inconjunction with known dispersal barriers in the MDB. These management unitsare illustrated in Fig. 5, and supporting information appears in Appendix IV. Eachmanagement unit corresponds with the presence of impoundments, naturally limitedflows and catchment boundaries. These units should be interpreted with some caution,however, for two reasons. First, the ongoing construction of fishways (Stuart et al.,2008) and improved flow management may increase connectivity between popula-tions in various regions and may render some units obsolete, although this could beminimized by the inclusion of William’s carp-separation cages to reduce the move-ment of C. carpio (Stuart et al., 2006). Second, these units are defined over a broadarea, including the whole river catchment within the MDB. As additional barriers todispersal may be present within each unit, the fine details of the hydrology of eachriver system should also be considered when implementing control programmes. Theproposed units, however, indicate which catchments can be managed independentlyand which should be managed in conjunction with each other units for the effectivelong-term control of invasive C. carpio.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
310 G . D . H AY N E S E T A L .
Queensland
New South Wales
Main MDBMurrumbidgee
catchment
Wimmera catchment
Avoca-Loddon catchments
Central Victoria
Lake Eildon
Upper Murray
Lake Hume
Burrinjuck Dam
Wyangala Dam
Sydney
Brisbane
Canberra
Burrendong Dam
Macquarie catchment
Lake Keepit
Condamine catchment
Paroo-Warrego catchments
Adelaide
0 100 200 km
South Australia
FIG. 5. Proposed management units for Cyprinus carpio in the MDB. Units are based on genetic discontinuitiesand geographic barriers to dispersal (see Appendix IV for supporting details).
We thank L. Miles and J. Gongora for assistance with calling genotypes, C. Moran and L. A.Rollins for assistance with manuscript preparation and Z. Doan for technical support. We areindebted to K. Kohlmann for supplying samples from Germany and to L. Faulks, V. Carracher,P. Boyd, B. Smith, M. Hutchinson, S. Backhouse, P. Brown, D. Hartwell, C. McGregor andJ. Patil for collecting samples from Australia. Funding support was provided by the FisheriesR&D Corporation, the Murray–Darling Basin Commission, the Invasive Animals CooperativeResearch Centre, the NSW Department of Primary Industries and the University of Sydney.
ReferencesAnderson, H. K. (1920). Rescue operation on the Murrumbidgee River. The Australian
Zoologist 1, 157–160.Balon, E. K. (1995). Origin and domestication of the wild carp, Cyprinus carpio: from
Roman Gourmets to the swimming flower. Aquaculture 129, 3–48. doi: 10.1016/0044-8486(94)00227-F
Benjamini, Y. & Hochberg, Y. (1995). Controlling the false discovery rate: a practical andpowerful approach to multiple testing. Journal of the Royal Statistical Society B 57,289–300.
Benjamini, Y. & Yekutieli, D. (2001). The control of the false discovery rate in multiplehypothesis testing under dependency. Annals of Statistics 29, 1165–1188.
Brown, P. (1996). Fish Facts 4 Carp in Australia . Narrandera: NSW Fisheries.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 311
Brownstein, M. J., Carpten, J. D. & Smith, J. R. (1995). Modulation of non-templatednucleotide addition by Taq DNA polymerase: primer modifications the facilitategenotyping. BioTechniques 20, 1004–1010.
Bryan, M. B., Zalinski, D., Filcek, K. B., Libants, S., LI, W. & Scribner, K. T. (2005).Patterns of invasion and colonization of the sea lamprey (Petromyzon marinus) in NorthAmerica as revealed by microsatellite genotypes. Molecular Ecology 14, 3757–3773.doi: 10.1111/j.1365-294X.2005.02716.x
Clements, J. (1988). Salmon at the Antipodes a History and Review of Trout, Salmon and Charand Introduced Coarse Fish in Australasia Ballarat, Victoria, Australia: Author.
Cornuet, J. M. & Luikart, G. (1996). Description and power analysis of two tests fordetecting recent population bottlenecks from allele frequency data. Genetics 144,2001–2014.
Crooijmans, R. P. M. A., Bierbooms, V. A. F., Komen, J., Van der Poel, J. J. & Groe-nen, M. A. M. (1997). Microsatellite markers in common carp (Cyprinus carpio L.).Animal Genetics 28, 129–134. doi: 10.1111/j.1365–2052.1997.00097.x
David, L., Rajasekaran, P., Fang, J., Hillel, J. & Lavi, U. (2001). Polymorphisms in ornamen-tal and common carp strains (Cyprinus carpio L.) as revealed by AFLP analysis anda new set of microsatellite markers. Molecular Genetics and Genomics 266, 353–362.doi: 10.1007/s004380100569
Davis, K. M., Dixon, P. I. & Harris, J. H. (1999). Allozyme and mitochondrial DNA analysisof carp, Cyprinus carpio L., from south-eastern Australia. Marine and FreshwaterResources 50, 253–260. doi: 10.1071/MF97256
DeWoody, J. A. & Avise, J. C. (2000). Microsatellite variation in marine, freshwater andanadromous fishes compared with other animals. Journal of Fish Biology 56, 461–473.doi: 10.1111/j.1095–8649.2000.tb00748.x
Evanno, G., Regnaut, S. & Goudet, J. (2005). Detecting the number of clusters of individ-uals using the software STRUCTURE: a simulation study. Molecular Ecology 14,2611–2620. doi: 10.1111/j.1365-294X.2005.02553.x
Excoffier, L., Laval, G. & Schneider, S. (2005). Arlequin ver. 3.0: an integrated softwarepackage for population genetics data analysis. Evolutionary Bioinformatics Online 1,47–50.
Excoffier, L., Smouse, P. E. & Quattro, J. M. (1992). Analysis of molecular variance inferredfrom metric distances among DNA haplotypes: application to human mitochondrialDNA restriction data. Genetics 131, 479–491.
Falush, D., Stephens, M. & Pritchard, J. K. (2003). Inference of population structure usingmultilocus genotype data: linked loci and correlated allele frequencies Genetics 164,1567–1587.
Frankham, R. (2005). Resolving the genetic paradox in invasive species. Heredity 94, 385.doi: 10.1038/sj.hdy.6800634
Gilligan, D. & Rayner, T. (2007). The distribution, spread, ecological impacts, and poten-tial control of carp in the upper Murray River. NSW Department of Primary Indus-tries–Fisheries Research Report Series 14.
Graham, K. J., Lowry, M. B. & Walford, T. R. (2005). Carp in NSW: assessment of distri-bution, fishery and fishing methods. NSW Department of Primary Industries–FisheriesFinal Report Series 72.
Hamner, R. M., Freshwater, D. W. & Whitfield, P. E. (2007). Mitochondrial cytochrome$b$ analysis reveals two invasive lionfish species with strong founder effectsin the western Atlantic. Journal of Fish Biology 71(Suppl. B), 214–222. doi:10.1111/j.1095–8649.2007.01575.x
Hanfling, B. (2007). Understanding the establishment success of non-indigenous fishes:lessons from population genetics. Journal of Fish Biology 71(Suppl. D), 115–135.doi: 10.1111/j.1095–8649.2007.01685.x
Hanfling, B., Boton, P., Harley, M. & Carvalho, G. R. (2005). A molecular approach todetect hybridisation between crucian carp (Carassis carassius) and non-indigenouscarp species (Carassisus cpp. and Cyprinus carpio). Freshwater Biology 50, 403–417.doi:10.1111/j.1365–2427.2004.01330.x
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
312 G . D . H AY N E S E T A L .
Harris, J. H. & Gehrke, P. C. (1997). Fish and Rivers in Stress - the NSW Rivers Survey .Cronulla: NSW Fisheries Office of Conservation & the Cooperative Research Centrefor Freshwater Ecology.
Hutchinson, D. W. & Templeton, A. R. (1999). Correlations of pairwise genetic distance andgeographic distance measures: inferring the relative influences of gene flow and drifton the distribution of genetic variability. Evolution 53, 1898–1914.
Kalinowski, S. T. (2005). HP-Rare 1.0: a computer program for performing rarefac-tion on measures of allelic richness. Molecular Ecology Notes 5, 187–189. doi:10.1111/j.1471–8286.2004.00845.x
Koehn, J. (2004). Carp (Cyprinus carpio) as a powerful invader in Australian waterways.Freshwater Biology 49, 882–894. doi: 10.1111/j.1365–2427.2004.01232.x
Koehn, J., Brumley, B. & Gehrke, P. (2000). Managing the Impacts of Carp Canberra: Bureauof Rural Sciences, Department of Agriculture, Fisheries and Forestry - Australia.
Kohlmann, K., Kersten, P. & Flajshans, M. (2005). Microsatellite-based genetic variabilityand differentiation of domesticated, wild and feral common carp (Cyprinus carpio L.)populations. Aquaculture 247, 253–256. doi: 10.1016/j.aquaculture.2005.02.024
Lindholm, A. K., Brenden, F., Alexander, H. J., Chan, W.-K., Thakurta, S. G. & Brooks, R.(2005). Invasion success and genetic diversity of introduced populations of gup-pies Poecilia reticulata in Australia. Molecular Ecology 14, 3671–3682. doi:10.1111/j.1365-294X.2005.02697.x
Manni, F., Guerard, E. & Heyer, E. (2004). Geographic patterns of (genetic, morphologic,linguistic) variation: how barriers can be detected by “Monmonier’s algorithm”. HumanBiology 76, 173–190.
Moritz, C. (1994). Defining evolutionary significant units for conservation. Trends in Ecologyand Evolution 9, 373–375. doi: 10.1016/0169-5347(94)90057-4
Peakall, R. & Smouse, P. E. (2006). GENALEX 6: genetic analysis in EXCEL. Populationgenetic software for teaching and research. Molecular Ecology Notes 6, 288–295. doi:10.1111/j.1471–8286.2005.01155.x
Pirey, S., Luikart, G. & Cornuet, J. M. (1999). BOTTLENECK: a computer program fordetecting recent reductions in the effective population size using allele frequency data.Journal of Heredity 90, 502–503.
Pritchard, J. K., Stefens, M. & Donelly, P. (2000). Inference of population structure usingmultilocus genotype data. Genetics 155, 945–959.
Raymond, M. & Rousset, F. (1995). GENEPOP (version 1.2): population genetics softwarefor exact tests and ecumenicism. Journal of Heredity 86, 248–249.
Reid, D. D., Harris, J. H. & Chapman, D. J. (1997). NSW Inland Commercial Fishery DataAnalysis FRDC Project No. 94/027. Sydney: Fisheries Research & DevelopmentCorporation, NSW Fisheries, Cooperative Research Centre for Freshwater Ecology.
Reiner, A., Yekutieli, D. & Benjamini, Y. (2003). Identifying differentially expressed genesusing false discovery rate controlling procedures. Bioinformatics 19, 368–375.
Reynolds, J., Weir, B. S. & Cockerham, C. C. (1983). Estimation of the coancestry coeffi-cient: basis for a short-term genetic distance. Genetics 105, 767–779.
Shearer, K. D. & Mulley, J. C. (1978). The introduction and distribution of the carp, Cyprinuscarpio Linnaeus, in Australia. Australian Journal of Marine and Freshwater Research29, 661–563.doi:10.1071/MF9780551
Shuber, A. P., Grondin, V. J. & Klinger, K. W. (1995). A simplified procedure for developingmultiplex PCRs. Genome Research 5, 488–493.
Slatkin, M. (1991). Inbreeding coefficients and coalescence times. Genetical Research 58,167–175.
Stead, D. G. (1929). Introduction of the great carp Cyprinus carpio into waters of New SouthWales. Australian Zoologist 6, 100–102.
Stuart, I. G., Williams, A., McKenzie, J. & Holt, T. (2006). Managing a migratory pestspecies: a selective trap for common carp. North American Journal of FisheriesManagement 26, 888–893.
Stuart, I. G., Zampatti, B. P. & Baumgartner, L. J. (2008). Can a low-gradient vertical-slotfishway provide passage for a lowland river fish community? Marine and FreshwaterResearch 59, 332–346. doi: 10.1071/MF07141
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 313
Wasko, A. P., Martins, C., Oliveira, C. & Foresti, F. (2003). Non-destructive genetic sam-pling of fish. An improved method for DNA extraction from fish fins and scales.Hereditas 138, 161–165. doi:10.1034/j.1601–5223.2003.01503.x
Weir, B. S. & Cockerham, C. C. (1984). Estimating F -statistics for the analysis of populationstructure. Evolution 38, 1358–1370.
Yue, G. H., Ho, M. Y., Orban, L. & Komen, J. (2004). Microsatellites within genes and ESTsof common carp and their applicability in silver crucian carp. Aquaculture 234, 85–98.doi: 10.1016/j.aquaculture.2003.12.021
Zenger, K. R., Richardson, B. J. & Vachot-Griffin, A.-M. (2003). A rapid population expan-sion retains genetic diversity within European rabbits in Australia. Molecular Ecology12, 789–794. doi: 10.1046/j.1365-294X.2003.01759.x
Zheng, W., Slacey, N. E., Coffin, J. & Strobeck, C. (1995). Isolation and characterization ofmicrosatellite loci in the gold fish Carassius accretus . Molecular Ecology 4, 791–792.
Electronic Reference
Pollard, K. S., Dudoit, S. & van der Laan, L. J. (2008). Multiple Testing Pro-cedures: R Multtest Package and Applications to Genomics . U.C. Berke-ley Division of Biostatistics Working Paper Series 164. Available at:http://www.bepress.com/ucbbiostat/paper164
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
314 G . D . H AY N E S E T A L .
APP
EN
DIX
I.PC
Rco
nditi
ons
and
prim
erse
quen
ces
for
Cyp
rinu
sca
rpio
mic
rosa
telli
telo
ci.
PCR
sA
,B
and
Cco
mpr
ise
mul
tiple
xes
oftw
olo
ci;
all
othe
rPC
Rs
ampl
ify
asi
ngle
locu
s.N
on-t
empl
ate
com
pone
nts
ofth
epr
imer
sequ
ence
s(S
hube
ret
al.,
1995
;B
row
nste
inet
al.,
1996
)in
italic
s.Pr
imer
nam
esw
ithan
‘a’
suffi
xha
vebe
enre
-des
igne
d.PC
Rcy
clin
gpr
otoc
ols
appe
arin
App
endi
xII
and
PCR
prod
uct
size
rang
esin
Tabl
eII
.The
5′-G
TT
TC
TT
-3′ n
on-t
empl
ate
sequ
ence
,w
hich
isad
ded
toth
ere
vers
epr
imer
sto
faci
litat
eac
cura
teal
lele
size
scor
ius
(Bro
wns
tein
etal
.,19
95)
issh
own
inbo
ld
Prim
erPr
imer
MgC
l 2PC
RPC
RL
ocus
Prim
ers
sequ
ence
conc
entr
atio
nco
ncen
trat
ion
prot
ocol
∗
AC
ca72
∗F-
NE
DC
AG
GC
CA
GA
TC
TAT
CA
TC
AT
CA
A0·2
μM2·5
mM
TD
5060
RG
TT
TC
TT
CT
GC
TG
TT
GG
ATA
TG
CA
CTA
CA
TC
0·2μM
Cca
02∗
F-V
ICA
TG
CA
GG
GC
TC
AT
GT
TG
CT
CA
TAG
0·2μM
RG
TT
TC
TT
GC
AG
AC
AG
AC
AC
GT
TG
CT
CT
CG
0·2μM
BM
FW6∗∗
F-N
ED
AC
CT
GA
TC
AA
TC
CC
TG
GC
TC
0·2μM
2m
MT
D68
50R
GT
TT
CT
TT
TG
GG
AC
TT
TTA
AA
TC
AC
GT
TG
0·2μM
MFW
26∗∗
F-V
ICC
CC
TG
AG
ATA
GA
AA
CC
AC
TG
0·2μM
RG
TT
TC
TT
CA
CC
AT
GC
TT
GG
AT
GC
AA
AA
G0·2
μM
CK
oi41
−42
†Fa
-VIC
GC
GG
TC
CC
AA
AA
GG
GT
CA
GT
AT
CT
CT
GA
AA
AG
CC
CA
ATA
TG
TC
AA
0·17μ
M1·5
mM
TD
6452
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
CT
GTA
AA
TC
TT
CA
TG
GT
GT
GT
GT
CC
0·17μ
M
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 315
APP
EN
DIX
I.C
ontin
ued
Prim
erPr
imer
MgC
l 2PC
RPC
RL
ocus
Prim
ers
sequ
ence
conc
entr
atio
nco
ncen
trat
ion
prot
ocol
∗
Cca
09∗
F-6F
AM
GC
GG
TC
CC
AA
AA
GG
GT
CA
GT
AA
TG
CC
TAT
TC
AC
AT
TAT
GA
AA
AT
0·2μM
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
AA
TC
AG
GTA
TAG
TG
GT
TATA
TG
AG
TT
0·2μM
DG
F1††
F-N
ED
GC
GG
TC
CC
AA
AA
GG
GT
CA
GT
AT
GA
AG
GG
TAG
GA
AA
AG
TG
TG
A0·2
μM2
mM
TD
6452
RG
TT
TC
TT
CA
AA
AG
GG
TC
AG
TC
AG
GT
TAG
GG
AG
AA
GA
AG
GA
AT
0·2μM
Da
Cca
65∗
Fa-6
FAM
AA
GT
GA
GC
GG
GA
GA
CA
GA
GA
0·17μ
M1·5
mM
TD
6452
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
CA
GA
CA
AG
TG
TG
CA
TG
AG
TG
G0·1
7μM
FC
ca19
∗F-
HE
XG
CG
GT
CC
CA
AA
AG
GG
TC
AG
TC
CT
GA
CC
CT
GA
AG
AG
AA
CA
AC
TAC
0·2μM
2m
MT
D64
52
RG
TT
TC
TT
CA
AA
AG
GG
TC
AG
TT
GG
CC
TC
AT
CA
AA
GA
CA
TC
AA
G0·2
μM
GC
ca67
∗F-
VIC
GTA
GC
CC
CA
AA
AG
AT
GTA
GC
A0·2
μM1·5
mM
TD
6850
RG
TT
TC
TT
TG
GT
CA
AG
TT
CA
GA
GG
CT
GTA
T0·2
μM
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
316 G . D . H AY N E S E T A L .
APP
EN
DIX
I.C
ontin
ued
Prim
erPr
imer
MgC
l 2PC
RPC
RL
ocus
Prim
ers
sequ
ence
conc
entr
atio
nco
ncen
trat
ion
prot
ocol
∗
HK
oi5-
6†Fa
-NE
DG
CG
GT
CC
CA
AA
AG
GG
TC
AG
TT
TT
GT
GT
TT
TC
TG
TT
GTA
GG
CT
CT
G0·2
μM1·5
mM
TD
6452
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
TT
TTA
CT
TC
AT
CT
CT
CG
CA
CT
CA
TC
T0·2
μM
IK
oi29
-30†
Fa-N
ED
GC
GG
TC
CC
AA
AA
GG
GT
CA
GT
CC
CT
GA
CC
CT
GA
AG
AG
AA
CA
AC
TAC
0·2μM
1·5m
MT
D64
52
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
GC
CT
CA
TC
AA
AG
AC
AT
CA
AG
0·2μM
JC
ca07
∗Fa
-6FA
MG
CG
GT
CC
CA
AA
AG
GG
TC
AG
TC
AT
TG
CG
CT
GTA
ATA
TG
AG
GT
TT
CT
0·2μM
1·5m
MT
D64
52
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
CT
CG
TT
CC
TT
TT
CT
GA
CG
CT
TT
T0·2
μM
KC
ca17
∗Fa
-6FA
MG
CG
GT
CC
CA
AA
AG
GG
TC
AG
TC
AG
GT
CT
TG
AT
TTA
CT
GC
TG
TC
TT
T0·2
μM1·5
mM
TD
6452
Ra
GT
TT
CT
TC
AA
AA
GG
GT
CA
GT
GA
TAA
CT
GC
GT
GTA
GG
CT
CT
GTA
TT
0·2μM
∗ Yue
etal
.(2
004)
.∗∗
Cro
oijm
ans
etal
.(1
997)
.†D
avid
etal
.(2
001)
.††
Zhe
nget
al.
(199
5).
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 317
APP
EN
DIX
II.
PCR
cycl
ing
prot
ocol
s
PCR
prot
ocol
Den
atur
ing
step
Touc
h-do
wn
cycl
eSt
anda
rdcy
cle
Fina
lex
tens
ion
step
TD
6850
95◦
C10
min
Den
atur
ing
95◦
Cfo
r45
sD
enat
urin
g95
◦C
for
45s
72◦
C30
min
Ann
ealin
g68
◦C
for
90s∗
Ann
ealin
g50
◦C
for
60s
Ext
ensi
on72
◦C
for
60s
Ext
ensi
on72
◦C
for
60s
Tota
lcy
cles
9To
tal
cycl
es30
TD
6050
95◦
C10
min
Den
atur
ing
95◦
Cfo
r30
sD
enat
urin
g95
◦C
for
30s
72◦
C30
min
Ann
ealin
g60
◦C
for
30s∗
∗A
nnea
ling
50◦
Cfo
r30
sE
xten
sion
72◦
Cfo
r30
sE
xten
sion
72◦
Cfo
r30
sTo
tal
cycl
es10
Tota
lcy
cles
30T
D64
5295
◦C
10m
inD
enat
urin
g95
◦C
for
30s
Den
atur
ing
95◦
Cfo
r30
s72
◦C
30m
inA
nnea
ling
64◦
Cfo
r60
s∗∗A
nnea
ling
52◦
Cfo
r30
sE
xten
sion
72◦
Cfo
r60
sE
xten
sion
72◦
Cfo
r30
sTo
tal
cycl
es12
Tota
lcy
cles
30
∗ Dec
reas
eby
2◦C
each
cycl
e.∗∗
Dec
reas
eby
1◦C
each
cycl
e.
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
318 G . D . H AY N E S E T A L .
APP
EN
DIX
III.
Cyp
rinu
sca
rpio
sam
ples
used
inis
olat
ion-
by-d
ista
nce
anal
yses
Nam
eof
anal
ysis
Sam
ples
site
s
All
site
sC
DM
,PR
,C
V,
DB
,W
N,
WG
,N
G,
CN
,W
C,
MR
,N
B,
DQ
,E
C,
ND
,C
W,
LL
,W
T,W
M,A
V,B
R,C
S,B
G,K
IW,L
D,O
V,B
D,M
G,K
P,W
Y,E
I,L
H,B
J,C
MB
elow
dam
sC
DM
,PR
,C
V,
DB
,W
N,
WG
,N
G,
CN
,W
C,
MR
,N
B,
DQ
,E
C,
ND
,C
W,
LL
,W
T,
WM
,A
V,
BR
,C
S,B
G,
KIW
,L
D,
OV
Mai
nM
DB
man
agem
ent
unit
WG
,N
G,
CN
,W
C,
MR
,N
B,
DQ
,E
C,
CW
,L
L,
WT
Mur
ray
Bas
inD
Q,
EC
,N
D,
CW
,L
L,
WT
,W
M,
AV
,B
R,
CS,
GB
,K
IW,
LD
,O
VM
urra
yR
iver
(LH
incl
uded
)D
Q,
EC
,L
L,
WT
,K
IW,
OV
,L
HM
urra
yR
iver
(LH
excl
uded
)D
Q,
EC
,L
L,
WT
,K
IW,
OV
Dar
ling
Bas
in–
1C
MD
,PR
,C
V,
DB
,W
T,
WG
,N
G,
CN
,W
C,
MR
,N
B,
LL
,W
TD
arlin
gB
asin
–2
WG
,N
G,
CN
,W
C,
MR
,N
B,
LL
,W
TD
arlin
gB
asin
–3
CM
D,
PR,
CV
,D
B,
WT
,W
G,
NG
,C
N,
WC
,M
R,
NB
Dar
ling
Bas
in–
4W
G,
NG
,C
N,
WC
,M
R,
NB
Dar
ling
Riv
erW
G,
WC
,N
B,
LL
,W
NM
urra
yR
iver
+D
arlin
gR
iver
LL
,W
T,
EC
,D
Q,
OV
,K
IW,
WC
,W
G,
MR
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
C Y P R I N U S C A R P I O I N T H E M U R R AY – DA R L I N G BA S I N 319
APP
EN
DIX
IV.
Man
agem
ent
units
for
Cyp
rinu
sca
rpio
inth
eM
DB
(map
inFi
g.4)
Uni
tSa
mpl
esi
tes
Rea
son
for
delim
iting
asa
man
agem
ent
unit
Mai
nM
DB
LL
,W
T,
EC
,D
Q,
CW
,B
K,W
C,W
G,M
R,N
G,
CN
,N
B,
WM
Mul
tiple
know
nba
rrie
rsto
disp
ersa
l,m
ultip
lege
netic
disc
ontin
uitie
sde
tect
edby
Stru
ctur
e(p
redo
min
antly
Pros
pect
,Y
anco
and
Boo
lara
stra
ins
pres
ent)
and
Bar
rier
.A
lthou
ghth
eY
anco
stra
inis
mor
epr
eval
ent
inth
eD
arlin
gca
tchm
ent
than
inth
eM
urra
yca
tchm
ent,
site
sfr
ombo
thca
tchm
ents
are
incl
uded
inth
em
anag
emen
tun
itsas
age
netic
disc
ontin
uity
was
not
dete
cted
byB
arri
erbe
twee
nth
etw
oca
tchm
ents
.Pa
roo
–W
arre
goca
tch-
men
tsPR
,C
VG
enet
icdi
scon
tinui
tyde
tect
edby
Stru
ctur
e(p
redo
min
antly
Pros
pect
and
Boo
lara
stra
in)
and
Bar
rier
;Pa
roo
and
War
rego
Riv
ers
linke
dby
irri
gatio
nch
anne
ls.
Con
dam
ine
catc
hmen
tC
DM
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyB
oola
rast
rain
)an
dB
arri
erM
acqu
arie
catc
hmen
tW
N,
DB
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyPr
ospe
ctan
dB
oola
rast
rain
)an
dB
arri
er.B
oth
site
sin
the
Mac
quar
ieR
iver
(WN
and
DB
)ar
epr
opos
edto
bepa
rtof
the
sam
em
anag
emen
tun
it,de
spite
disc
ontin
uitie
sbe
ing
cons
iste
ntly
dete
cted
betw
een
them
byB
arri
er,b
ecau
seth
ere
are
nom
ajor
barr
iers
todi
sper
sal
betw
een
the
two
site
s.T
hedi
scon
tinui
tyis
prob
ably
anar
tefa
ctof
the
pred
omin
antly
Pros
pect
stra
inC
ypri
nus
carp
ioin
Bur
rend
ong
Dam
disp
ersi
ngdo
wns
trea
man
dhe
nce
bein
gm
ore
prev
alen
tat
the
WN
site
imm
edia
tely
belo
wth
eda
mou
tlet
than
atth
em
ore
dist
ant
DB
site
.M
urru
mbi
dgee
catc
h-m
ent
ND
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyY
anco
stra
in)
and
Bar
rier
Wim
mer
aca
tchm
ent
WM
Stro
ngly
isol
ated
from
othe
rpa
rts
ofth
eM
DB
,ge
netic
disc
ontin
uity
dete
cted
for
Slat
kin’
sF
ST
byB
arri
erA
voca
–L
oddo
nca
tch-
men
tsA
V,
LD
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyPr
ospe
ctan
dB
oola
rast
rain
)an
dB
arri
erC
entr
alV
icto
ria
BR
,G
B,
CS
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyPr
ospe
ctan
dB
oola
rast
rain
)an
dB
arri
er
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320
320 G . D . H AY N E S E T A L .
APP
EN
DIX
IV.
Con
tinue
d
Uni
tSa
mpl
esi
tes
Rea
son
for
delim
iting
asa
man
agem
ent
unit
Upp
erM
urra
yO
V,
KIW
Gen
etic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(pre
dom
inan
tlyPr
ospe
ctan
dB
oola
rast
rain
),w
eak
gene
ticdi
scon
tinui
tyde
tect
edfo
rSl
atki
n’s
FS
Tby
Bar
rier
Lak
eK
eepi
tK
PL
arge
dam
atri
ver
head
wat
ers
limits
C.c
arpi
odi
sper
sal,
gene
ticdi
scon
tinui
tyde
tect
edby
Stru
ctur
e(p
redo
min
antly
Pros
pect
and
Boo
lara
stra
in)
and
Bar
rier
Bur
rend
ong
Dam
BD
,M
GL
arge
dam
atri
ver
head
wat
ers
limits
C.c
arpi
odi
sper
sal,
gene
ticdi
scon
tinui
tyde
tect
edby
Stru
ctur
e(p
redo
min
antly
Pros
pect
stra
in)
and
Bar
rier
Wya
ngal
aD
amW
YL
arge
dam
atri
ver
head
wat
ers
limits
C.c
arpi
odi
sper
sal,
gene
ticdi
scon
tinui
tyde
tect
edby
Stru
ctur
e(p
redo
min
antly
Pros
pect
and
Boo
lara
stra
in)
and
Bar
rier
Bur
rinj
uck
Dam
BJ,
CM
Lar
geda
mat
rive
rhe
adw
ater
slim
itsC
.car
pio
disp
ersa
l,ge
netic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(gre
ater
cont
ribu
tion
from
koiC
.car
pio
and
muc
hle
sser
cont
ribu
tion
from
Yan
cost
rain
than
dow
nstr
eam
site
s)an
dB
arri
erL
ake
Hum
eL
HL
arge
dam
atri
ver
head
wat
ers
limits
C.c
arpi
odi
sper
sal,
gene
ticdi
scon
tinui
tyde
tect
edby
Bar
rier
Lak
eE
ildon
EI
Lar
geda
mat
rive
rhe
adw
ater
slim
itsC
.car
pio
disp
ersa
l,ge
netic
disc
ontin
uity
dete
cted
bySt
ruct
ure
(Pro
spec
tst
rain
mor
epr
eval
ent
than
atdo
wns
trea
msi
tes)
and
Bar
rier
© 2009 The AuthorsJournal compilation © 2009 The Fisheries Society of the British Isles, Journal of Fish Biology 2009, 75, 295–320