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.


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



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

.


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



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


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.



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.


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.



(a)

Queensland


Australia


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.


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



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


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



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.


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.



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


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



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


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



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


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).



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.


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



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


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


Population genetics and management units of invasive common carp Cyprinus carpio in the...

Documents

Transcript of Population genetics and management units of invasive common carp Cyprinus carpio in the...