BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofitpublishers, academic institutions, research libraries, and research funders in the common goal of maximizing access tocritical research.

Assessing Moth Migration and Population Structuring inHelicoverpa armigera (Lepidoptera: Noctuidae) at the RegionalScale: Example from the Darling Downs, AustraliaAuthor(s): Kirsten D. Scott , Nicole Lawrence , Corinna L. Lange , Leon J. Scott ,Kendle S. Wilkinson , Melissa A. Merritt , Melina Miles , David Murray , andGlenn C. GrahamSource: Journal of Economic Entomology, 98(6):2210-2219. 2005.Published By: Entomological Society of AmericaDOI: http://dx.doi.org/10.1603/0022-0493-98.6.2210URL: http://www.bioone.org/doi/full/10.1603/0022-0493-98.6.2210

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in thebiological, ecological, and environmental sciences. BioOne provides a sustainable onlineplatform for over 170 journals and books published by nonprofit societies, associations,museums, institutions, and presses.

Your use of this PDF, the BioOne Web site, and all posted and associated contentindicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use.

Usage of BioOne content is strictly limited to personal, educational, and non-commercialuse. Commercial inquiries or rights and permissions requests should be directed to theindividual publisher as copyright holder.

INSECTICIDE RESISTANCE AND RESISTANCE MANAGEMENT

Assessing Moth Migration and Population Structuring inHelicoverpa armigera (Lepidoptera: Noctuidae) at the Regional Scale:

Example from the Darling Downs, Australia

KIRSTEN D. SCOTT, NICOLE LAWRENCE, CORINNA L. LANGE, LEON J. SCOTT,KENDLE S. WILKINSON, MELISSA A. MERRITT, MELINA MILES,1 DAVID MURRAY,1

AND GLENN C. GRAHAM

School of Integrative Biology, The University of Queensland, Brisbane, Queensland 4072, Australia

J. Econ. Entomol. 98(6): 2210Ð2219 (2005)

ABSTRACT Analysis of gene ßow and migration of Helicoverpa armigera (Hubner) in a majorcropping region of Australia identiÞed substantial genetic structuring, migration events, and signiÞcantpopulation genotype changes over the 38-mo sample period from November 1999 to January 2003. Fivehighly variable microsatellite markers were used to analyze 916 individuals from 77 collections across10 localities in the Darling Downs. The molecular data indicate that in some years (e.g., April2002ÐMarch 2003), low levels of H. armigeramigration and high differentiation between populationsoccurred, whereas in other years (e.g., April 2001ÐMarch 2002), there were higher levels of adult mothmovement resulting in little local structuring of populations. Analysis of populations in other Aus-tralian cropping regions provided insight into the quantity and direction of immigration ofH. armigeraadults into the Darling Downs growing region of Australia. These data provide evidence adult mothmovement differs from season to season, highlighting the importance of studies in groups such as theLepidoptera extending over consecutive years, because short-term sampling may be misleading whenpopulation dynamics and migration change so signiÞcantly. This research demonstrates the impor-tance of maintaining a coordinated insecticide resistance management strategy, because in some yearsH. armigera populations may be independent within a region and thus signiÞcantly inßuenced by localmanagement practices; however, periods with high migration will occur and resistance may rapidlyspread.

KEY WORDS microsatellite, molecular biology, resistance management, pest management

Helicoverpa armigera (Hubner) (Lepidoptera: Noc-tuidae) is an insect that has developed resistance toagricultural insecticides (Zhou et al. 2000) and iswidely distributed (Africa, Middle East, India, Aus-tralia, and Asia). It is polyphagous and is a signiÞcantpest on cotton, grains, and horticultural crops in Aus-tralia.H. armigera adults are highly mobile, and adult

movements occur on several spatial scales; betweenÞelds, between areas within a region, and betweenregions (Farrow and Daly 1987). The effective man-agement of Helicoverpa in crops is complicated byvariability in infestation levels between regions andbetween seasons (Rochester et al. 1996). The capacityH. armigera has for extensive movement also compli-cates the understanding of the population dynamics ofthis pest, because local populations may consist ofelements of diverse origins at any one time (Fitt et al.1995). To improve the effectiveness of managing this

signiÞcant pest, it is essential that a greater under-standing of the genetic structure and migratory be-havior of H. armigera be achieved.

Efforts to improve our understanding of the migra-tion events, genetic structuring, and gene ßow of H.armigera on both local and regional scales will beneÞtagricultural industries by enabling a reduction of in-secticide application and resistance buildup (Stokes etal. 1997). Microsatellite technology is well suited todescribing population structure and gene ßow withina species, because it is amenable for use with largesample numbers, and because microsatellites arecodominantly inherited, and highly informative be-cause of high rates of mutation. Microsatellite tech-nology had not been applied to population geneticstudies in H. armigera previously Scott et al. (2003),because they have only been developed in this speciesrelatively recently (Tan et al. 2001, Ji et al. 2003, Scottet al. 2004).

Measures of gene ßow are an important componentin understanding the genetic structuring within a spe-cies and can be obtained directly by measuring dis-

1 Queensland Department of Primary Industries, P.O. Box 102,Toowoomba, Queensland 4350, Australia.

0022-0493/05/2210Ð2219$04.00/0 � 2005 Entomological Society of America

persal and breeding contribution, or they can be in-ferred by measuring allele frequencies (Slatkin 1987,Zhou et al. 2000). Molecular approaches that havebeen used to measure gene ßow in H. armigera inAustralia include isozyme studies (Daly and Gregg1985), mitochondrial DNA variation (McKechnie etal. 1993), and a sodium channel gene (Stokes et al.1997). The early work in Australia by Daly and Gregg(1985) suggested that signiÞcant long-distance dis-persal byH.armigerawas likely because they observedlittle genetic differentiation among populations; how-ever, later work by Stokes et al. (1997) indicated thatthere was clear restriction to free and fast gene ßowbetween populations in the Lower Namoi Valley andSt. George growing regions of Australia. Data fromtraditional trapping methods also vary in their descrip-tion of H. armigera dispersal frequency and distancesin Australia. Recapture experiments by Fitt and Dillon(1993) found that many released H. armigera colo-nized crops within 10 km of their emergence site,whereas earlier work (Fitt and Daly 1990) suggestedthat both local and immigrant H. armigera might berepresented in pheromone trap catches. A recent mi-crosatellite survey of gene ßow between regional oc-currences of H. armigera across Australia has demon-strated that high, moderate, and low gene-ßow yearsoccur, with some seasons having substantial gene ßowbetween H. armigera from distant growing regions,whereas other seasons showed signiÞcant genetic dif-ferentiation between H. armigera in different geo-graphic regions (Scott et al. 2005).

Genetic analysis ofH.armigeraoutside Australia hasshown the occurrence of high migration rates andsmall genetic distances between populations, for ex-ample, in the Mediterranean, by using random ampli-Þed polymorphic DNA (Zhou et al. 2000), and inAfrica and Europe, by using isozymes (Nibouche et al.1998). Isozyme studies on Heliothis virescens (F.) inthe United States (Korman et al. 1993) described somegenetic structuring and high migration rates and onthis basis made recommendations on management ofH. virescens in the United States.

We used variability in microsatellites to quantify atthe regional level the genetic structuring and migra-tion of H. armigera on the Darling Downs (Australia)and to monitor the changing genetic proÞle over the3 yr from November 1999 to January 2003. These datacontribute to our knowledge of the genetics of H.armigera and will assist with deÞning the scale atwhich management programs should operate for thispest.

Materials and Methods

Sample Source and DNA Extraction. H. armigeralarvae and adults were collected from 10 localities onthe Darling Downs over a 38-mo period (Figs. 1 and2; Table 1). The Darling Downs is an agricultural areawith highly fertile soils and grows primarily cotton andgrains with some horticultural crops. Although a strat-iÞed sampling strategy was proposed, variation in pestpressure and seasonal ßux in pest availability (i.e., H.armigera usually only present in high numbers in the

Darling Downs over the summer) has resulted in dif-ferences in the number of individual samples obtainedat various sites and at various times. In addition tothese factors, H. armigera often coexists with anotherspecies,Helicoverpa punctigera (Wallengren), furtherreducing the number ofH. armigera that may be geno-typed for each collection. Samples were obtained fromavarietyofcrops includinggrains, legumesandcotton.

DNA for microsatellite analysis was extracted as forScott et al. (2003). A diagnostic polymerase chainreaction (PCR) (developed at the School of Integra-tive Biology, unpublished data; for details please con-tact us) was used to differentiate betweenH. armigeraand H. punctigera because morphological determina-tion of species after storage in ethanol was problem-atic. Nine hundred and sixteenH. armigera individualswere analyzed from 77 collections consisting of 10geographic locations and multiple collection dates(Table 1). The number of individuals analyzed percollection varied as a result of differing proportions ofH. armigera and H. punctigera present in the samplescollected at each location. Collections with �5 H.armigera individuals were not included.

To determine the extent of insect movement intoand out of the Darling Downs region, additional insectmaterials were used for an inter-regional comparison.These non-Darling Downs collections included an ad-ditional 2,226 individuals from seven other growingregions. The full details of these collections are pre-sented in Scott et al. (2005).Microsatellite Analysis. Five highly variable micro-

satellite loci HaB60 (25 alleles, He � 0.45, Ho � 0.04),HaD25 (58 alleles, He � 0.73, Ho � 0.36), HaD47 (85alleles, He � 0.75, Ho � 0.24), HaC87 (33 alleles, He �0.50, Ho � 0.17), and HaC14 (32 alleles, He � 0.65,Ho � 0.48)] (Scott et al. 2004) were used to analyze916 H. armigera individuals. Microsatellites were Hexlabeled with PCR ampliÞcation conditions and gelseparation as published in Scott et al. (2003).Statistical Analysis. Microsatellite alleles were

scored using ONE-Dscan (version 1.33, ScanalyticsInc., Billerica, MA). Allele sizes were analyzed usingGenAlEx (Peakall and Smouse 2001). Nei distancebetween collections was calculated using the methoddescribed by Weir (1990), and pairwise genetic dis-tances were calculated as in Peakall et al. (1995).Allele frequencies and heterozygosity calculations fol-lowed the formulae of Hartl and Clark (1997). Nm(where N is local population size andm is average rateof migration) was estimated using the private allelemethod of Slatkin (1985) and Slatkin and Barton(1989). Analysis of molecular variance (AMOVA)analysis was as in ExcofÞer et al. (1992), Peakall et al.(1995), and Michalakis and ExcofÞer (1996). Principlecoordinate analysis (PCA) used the algorithm pub-lished by Orloci (1978). Assignment tests were per-formed in GeneClass2.3 (Piry et al. 2004). The criteriafor computation was the Bayesian method of Rannalaand Mountain (1997), the thresholds for declaring aÞrst generation migrant was calculated using the L �L_home/L_max model with 1000 Monte-Carlo resa-mpling of gametes to preserve linkage disequilibrium

December 2005 SCOTT ET AL.: REGIONAL GENE FLOW OF H. armigera in AUSTRALIA 2211

from recent immigrations (Paetkau et al. 2004). Thisassignment test enables the identiÞcation of immi-grant individuals in the current generation. This dif-fers from methods such as Wilson and Rannala (2003)that estimate migration over several generations.Other assignment test methods such as Rannala andMountain (1997) and Cornuet et al. (1999) resamplealleles rather than gametes, randomly distributing mi-grant alleles across the population (Paetkau et al.2004).This is anunrealistic approach fora species suchas H. armigera that has continuous migration andwould led to an excess of “locals” being wrongly iden-tiÞed as immigrant (type 1 error). The approach ofPaetkau et al. (2004) seemed then to be the mostappropriate methodology for H. armigera.We do ac-knowledge that we breach the assumption of HardyÐWeinberg equilibrium by using the Paetkau et al.(2004) approach; however, it has proven to be con-sistent with all our other measures of migration andseems to be informative. Assignment criteria for pop-ulations of �40 individuals used a 1% error rate, andfor populations with �40 individuals an error rate of

5% was applied. This was to account for the increasein type 1 errors when using smaller sample sizes.

Results and Discussion

Microsatellite analysis ofH. armigera showed a highdegree of differentiation between collections over the3 yr of study. Fluctuations in migration quantity anddirectionweredemonstrated throughassignment test-ing.H. armigerapopulations were not in HardyÐWein-berg equilibrium with a deÞcit of heterozygotes. Pre-liminary data from experimental crosses indicate thatthis deviation from HardyÐWeinberg equilibrium isnot a result of excessive null alleles. If null alleles werepresent, you also could expect that some loci wouldremain in HardyÐWeinberg equilibrium, with otherloci having a heterozygote deÞciency, but this is notthe case in H. armigera because all loci are deÞcient.The assumptions of HardyÐWeinberg equilibrium areperhaps then inappropriate for H. armigera, particu-larly the assumptions of no immigration (Scott et al.2005) and no selection, because every generation is

Fig. 1. Darling Downs regional map showing the sample locations. Site numbers 1Ð10 relate to site details as shown inTable 1. Darkly shaded areas mark the cotton growing regions.

2212 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 98, no. 6

under extreme selection pressure because of the ap-plication of insecticides. AMOVA methodology (Ex-cofÞer et al. 1992, Peakall et al. 1995, Michalakis andExcofÞer 1996) and Nm estimates from private alleles(Slatkin 1985, Slatkin and Barton 1989) were used inthese analyses, because these approaches do not re-quire HardyÐWeinberg equilibrium.

PCA of Nei genetic distance between collectionsfrom the Darling Downs from 1999 to 2003 (Fig. 3)illustrates that H. armigera genotypes changed everyfew months. However, it is important to note that thePCA analysis for the collections that fall into the April2002 to March 2003 period also reßects a componentof geographic differentiation within its clustering, be-cause in this period, the geographic differentiationbetween populations on the Darling Downs was sta-tistically signiÞcant(Table2).Thechange ingenotypefrequencies seen in the remaining seasons is not areßection of the geographic origin of the collections,because the AMOVA results presented in Table 2demonstrate no statistically signiÞcant geographicstructuring in the region at these times. This observedchange in genotypes is in concordance with observa-tions made ofH. armigera populations in the DawsonÐCallide valleys (Scott et al. 2003), and Þndings in

monarch butterßies, Danaus plexippus L. (Eanes andKoehn 1978). This may reßect a “founder-like effect”because of nonuniform distribution of the pest inspace, time, or both, and as a result of ßuctuatingpopulation sizes (Eanes and Koehn 1978). Moth im-migration events, insecticide use, agricultural prac-tices, and seasonal variation causing bottleneckswithin the populations are other likely contributors tothe changes in H. armigera genotypes observed.

The use of assignment testing on individual H. ar-migera over several consecutive years has providedinsight into the direction of immigration events intothe Darling Downs and into the quantity of immigra-tion during these seasons (Fig. 4). The level of mi-gration determined by assignment testing correlatedwell with the differentiation between regions deter-mined by AMOVA analysis. In the year with highAMOVA inter-regional differentiation (2002Ð2003),the proportion of H. armigera immigrating into theDarling Downs was only 5.7Ð10.7%. However, in theyear with low AMOVA inter-regional differentiation(2001Ð2002), the proportion of H. armigera immigrat-ing into the Darling Downs was much higher at 22.2Ð29.4%. The consistency between the AMOVA results,Nm from private alleles (both which do not assume

Fig. 2. Map of the east coast of Australia showing the major crop growing localities and regions from which samples wereobtained.

December 2005 SCOTT ET AL.: REGIONAL GENE FLOW OF H. armigera in AUSTRALIA 2213

HardyÐWeinberg) and the assignment test results,support our use of the Paetkau et al. 2004 approach,although we did not fulÞll the HardyÐWeinberg as-sumption.

AMOVA analysis was considered yearly, with eachyear beginning at the end of the summer growingseason (April) through to March of the following year.The gene ßow and genetic structuring varied fromyear to year (Tables 2Ð5).

Growing Season (November 1999–March 2000). Inthe Þrst year of study, November 1999ÐMarch 2000,AMOVA results comparing the Jimbour to the Brook-stead area of the Darling Downs showed no signiÞcantdifferentiation of genotypes, suggesting that in thisinitial study season (Table 2) Jimbour and Brooksteadwere effectively a single population and thus perhapsa single crop management unit for H. armigera. Amigration rate Nm of 1.42 (Table 3) was recorded in

Table 1. H. armigera sample locations within the Darling Downs, including map references, geographical positioning system (GPS)coordinates, host crop, collection date, and sample numbers

Location Map ref. GPS Crop Date collected No. individuals

1999Ð2000Brookstead 1 S27 42 52, E151 19 04 Chickpea 4 Nov. 1999 15Brookstead 1 S27 35 16, E151 13 06 Chickpea 4 Nov. 1999 15Brookstead 1 S27 20 29, E151 16 42 Sorghum 14 Feb. 2000 15Brookstead 1 S27 20 29, E151 16 42 Sorghum 14 Feb. 2000 15Brookstead 1 S27 20 29, E151 16 42 Sorghum 14 Feb. 2000 15Brookstead 1 S27 20 29, E151 16 42 Sorghum 17 Feb. 2000 15Brookstead 1 S27 43 10, E151 17 59 Soyabean 24 Feb. 2000 15Brookstead 1 S27 36 35, E151 14 06 Soyabeans 3 Mar. 2000 15Brookstead 1 S27 32 25, E151 14 28 Sorghum 15 Mar. 2000 15Brookstead 1 S27 41 51, E151 19 00 Mungbean 16 Mar. 2000 15Jimbour 2 S27 00 33, E151 04 18 Chickpea 2 Nov. 1999 15Jimbour 2 S26 59 09, E151 08 51 Chickpea 2 Nov. 1999 15Jimbour 2 S26 59 27, E151 04 30 Maize 18 Feb. 2000 15Jimbour 2 S26 58 48, E151 10 38 Maize 8 Mar. 2000 15Jimbour 2 S27 00 38, E151 06 52 Maize 21 Mar. 2000 15Jimbour 2 S26 59 34, E151 03 54 Mungbean 22 Mar. 2000 15

2000Ð2001Brookstead 1 S27 44 30, E151 19 49 Chickpea 2 Oct. 2000 7Brookstead 1 S27 42 10, E151 20 21 Chickpea 2 Oct. 2000 11Brookstead 1 S27 41 43, E151 21 10 Chickpea 2 Oct. 2000 15Brookstead 1 S27 44 30, E151 19 49 Chickpea 11 Oct. 2000 15Cecil Plains 3 S27 35 10, E151 12 57 Soyabean 2 Mar. 2001 15Jimbour 2 S27 01 13, E151 09 35 Chickpea 18 Oct. 2000 15Jimbour 2 S26 56 40, E151 05 22 Sorghum 8 Mar. 2001 15Kingsthorpe 4 S27 29 57, E151 45 54 Sorghum 8 Mar. 2001 15

2001Ð2002Brookstead 1 S27 41 43, E151 21 10 Pheromone trap 26 Sept.-3 Oct. 2001 15Brookstead 1 S27 43 10, E151 17 59 Chickpea 9 Nov. 2001 9Brookstead 1 S27 41 43, E151 21 10 Pheromone trap 13 Nov.Ð3 Dec. 2001 15Brookstead 1 S27 41 32, E151 16 10 Maize 18 Jan. 2002 15Gatton 5 S37 32 28, E152 19 47 Chickpea 14 Oct. 2001 13Jimbour 2 S26 57 06, E151 10 51 Pheromone trap 17Ð20 Sept. 2001 15Jimbour 2 S26 57 06, E151 10 51 Pheromone trap 12Ð27 Nov. 2001 10Jimbour 2 S26 56 32, E151 07 56 Pigeon pea 20 Feb. 2002 10Jimbour 2 S26 56 32, E151 07 56 Cotton 21 Feb. 2002 13

2002Ð2003Bowenville 10 S27 29 02, E151 23 83 Pheromone trap 20Ð25 Nov. 2002 5Brookstead 1 S27 44 06, E151 24 58 Pheromone trap 20Ð25 Nov. 2002 5Dalby 6 S27 18 26, E151 15 54 Chickpea 3 Oct. 2002 7Dalby 6 S27 18 26, E151 15 54 Cotton 24 Jan. 2003 7Dalby 6 S27 18 26, E151 15 54 Cotton 24 Jan. 2003 10Kingaroy 7 S26 24 56, E151 51 11 Barley 17 Oct. 2002 15Kingaroy 7 S26 21 02, E151 47 33 Wheat 18 Oct. 2002 10Kingaroy 7 S26 43 05, E151 51 10 Wheat 23 Oct. 2002 7Kingaroy 7 S26 43 05, E151 51 10 Barley 25 Oct. 2002 14Kingaroy 7 S26 37 10, E151 25 45 Wheat 25 Oct. 2002 6Murgon 8 S26 16 15, E152 01 01 Wheat 23 Oct. 2002 6Murgon 8 S26 11 26, E151 51 03 Maize 23 Oct. 2002 10Murgon 8 S26 08 93, E151 47 64 Wheat 23 Oct. 2002 14Murgon 8 S26 11 17, E151 43 02 Cotton 10 Jan. 2003 13Murgon 8 S26 11 17, E151 43 02 Cotton 16 Jan. 2003 15Murgon 8 S26 11 26, E151 51 03 Cotton 16 Jan. 2003 15Murgon 8 S26 10 20, E151 49 15 Cotton 16 Jan. 2003 15Murgon 8 S26 10 20, E151 49 15 Cotton 16 Jan. 2003 15Toowoomba 9 S27 55 57, E151 53 04 Barley 25 Oct. 2002 15Toowoomba 9 S27 14 48, E151 09 97 Pheromone trap 20Ð25 Nov. 2002 5Toowoomba 9 S27 22 92, E151 16 23 Pheromone trap 20Ð25 Nov. 2002 6Wooroolin 7 S26 25 11. E151 50 25 Peanut 29 Jan. 2003 5

2214 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 98, no. 6

February 2000 for Jimbour and Brookstead, and an Nmof 2.76 (Table 3) in March 2000 was recorded for theBrookstead and Jimbour localities. The migration ratesin this year were relatively high in comparison withrates in subsequent years on the Darling Downs; how-ever, assignment test still found that 96.1% of individ-uals were of local origin with 3.9% ofH. armigerabeingpossible immigrants from the Central Highlands (Ta-ble 4). Migrations rates of Nm � 1 are considered tobe sufÞcient to allow gene ßow to overcome geneticdrift (Wright 1931). The migration rates in this study

are smaller than those described in H. armigera byKorman et al. (1993), Nibouche et al. (1998), andZhou et al. (2000). This may be partly because achange toward microsatellite technologies, a reßec-tion of the geographic location and/or scale of thestudies, or a consequence of the years in which thesedata were collected, because migration rates in H.armigera in Australia are shown to vary signiÞcantlybetween years (Scott et al. 2005).Growing Season (April 2000–March 2001). In April

2000ÐMarch 2001, migration rates in the DarlingDowns were slightly lower than in the previous year

Table 3. Estimate of migrants (Nm) into localities in the DarlingDowns region by using the private alleles method (Slatkin 1985,Slatkin and Barton 1989)

Location Date Nm

Brookstead and Jimbour Feb. 2000 1.42Brookstead and Jimbour Mar. 2000 2.76Brookstead and Jimbour Oct. 2000 0.76Cecil Plains, Kingsthorpe and Jimboura Mar. 2001 1.14Brookstead and Gattona Oct. 2001 2.43Kingaroy Oct. 2002 0.55Murgon Oct. 2002 0.97Toowoomba Nov. 2002 0.25Murgon Jan. 2003 1.31

Migration rates were calculated for each month, where three ormore collections were available for a locality (the localities deÞnedwhere possible by the AMOVA analysis, Table 2).a AMOVA unavailable.

Fig. 3. Principle coordinate analysis of NeiÕs genetic distance between collections. Collections were made from November1999 to January 2003 in the Darling Downs. The cumulative percentage of variation explained by the two axes of the PCAis 51.86%.

Table 2. AMOVA between localities (refer to Fig. 1) within theDarling Downs from November 1999 to March 2003

Localities compared% betweenlocalities(P value)

% betweencollections(P value)

% withincollections(P value)

1999Ð2000Jimbour to Brookstead 0 (0.378) 12 (0.001) 88 (0.001)

2000Ð2001Jimbour to Brookstead 1 (0.158) 8 (0.001) 91 (0.001)

2001Ð2002Jimbour to Brookstead 0 (0.350) 10 (0.001) 90 (0.001)

2002Ð2003Toowoomba to Kingaroy 1 (0.084) 17 (0.001) 82 (0.001)Tooowoomba to Murgon 4 (0.002) 24 (0.001) 72 (0.001)Kingaroy to Murgon 2 (0.004) 22 (0.001) 76 (0.001)Toowoomba to Dalby 3 (0.039) 25 (0.001) 72 (0.001)Kingaroy to Dalby 4 (0.006) 20 (0.001) 76 (0.001)Murgon to Dalby 2 (0.047) 27 (0.001) 71 (0.001)

P value � 0.05 are signiÞcant.

December 2005 SCOTT ET AL.: REGIONAL GENE FLOW OF H. armigera in AUSTRALIA 2215

with Nm� 0.76 (Table 3) for October 2000 in Brook-stead and Jimbour, and an Nm � 1.14 (Table 3) forMarch 2001 for Cecil Plains, Kingsthorpe, and Jim-bour. AMOVA results comparing the Jimbour with theBrookstead area of the Darling Downs also showed nosigniÞcant differentiation. If more intensive samplinghad been possible in this year the “effective popula-tions” or “management units” at the local level mayhave been evident. During this year, microsatelliteanalysis also was completed for H. armigera from theDawsonÐCallide valleys and Namoi Valley crop grow-ing regions in Australia (Table 5). The AMOVA anal-

ysis of this data showed that the Darling Downs wasgenetically distinct from both the DawsonÐCallidevalleys growing regions (4%, P � 0.01; AMOVA) andthe Lower Namoi Valley (2%, P � 0.01; AMOVA)(Table 5) growing regions, with 87 to 87.3% of indi-viduals in this season being of local origin in the Dar-ling Downs. Although each of the growing regionswere distinct during this period, there is evidencefrom the assignment tests that migration into the Dar-ling Downs still occurred from both of these regions,i.e., immigrants into the Darling Downs sourced fromboth the DawsonÐCallide valleys (1.6%, OctoberÐDe-

Fig. 4. Schematic representation of local recruitment and immigration of H. armigera into the Darling Downs, based onassignment testing. The proportion of local recruitment (in bold) and immigration in the Darling Downs is indicated as apercentage. Each circled region was signiÞcantly differentiated in the AMOVA analysis except in April 2001ÐMarch 2002,where nondifferentiated populations are indicated by grey zones. Locality names are listed in Figs. 1 and 2, results of AMOVAanalyses are summarized in Tables 2 and 5, and assignment data is shown in Table 4.

2216 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 98, no. 6

cember 2000; 4.3%, JanuaryÐMarch 2001) and from theNamoi Valley (11.1%, OctoberÐDecember 2000; 8.7%,JanuaryÐMarch 2001) (Table 4).Growing Season (April 2001–March 2002). During

April 2001ÐMarch 2002, there was again no signiÞcantdifference between the collections of H. armigerafrom Jimbour to those from Brookstead (Table 2).This was accompanied by a high migration rate inOctober for Brookstead and Gatton Nm� 2.43 (Table3). The Darling Downs during this year had AMOVAdifferentiation from the Macintyre Valley (1%, P �0.01; AMOVA) and from the Murrumbidgee (2%, P�0.01; AMOVA) (Table 5). For this period, 22.2Ð29.4%of individuals were immigrants to the Darling Downs,with immigrations sourced most frequently from theDawsonÐCallide valleys (11.1%, JulyÐSeptember 2001;17.6%, OctoberÐDecember 2001; 9.1%, JanuaryÐMarch2002) and from the Macintyre Valley (11.1%, JulyÐSeptember 2001; 5.9%, OctoberÐDecember 2001) withoccasional contributions from the Murrumbidgee Val-ley (18.2%, JanuaryÐMarch 2002). There was no sta-tistical support for differentiation of the DarlingDowns collections from those of the DawsonÐCallidevalleys or Lower Namoi Valley at this time (Table 5).The absence of differentiation between the DawsonÐ

Callide valleys and the Darling Downs is easily ex-plained by the signiÞcant and extended immigrationfrom the DawsonÐCallide into the Darling Downsover the majority of this year. The lack of differenti-ation between the Darling Downs and the Namoi wasa result of signiÞcant immigration from the DarlingDowns into the Namoi (Scott et al. 2005). In this year,the local AMOVA analysis, the high migration rates,and assignment tests all support the suggestion therewas quite a signiÞcant and probably prolonged adultmoth movement between growing regions during theyear. Accordingly, it might be expected that the man-agement of H. armigera on the Darling Downs mayhave been inßuenced by the occurrences and insec-ticide resistances fromthepest in severalother regions(DawsonÐCallide valleys, Macintyre Valley, and Mur-rumbidgee Valley).Growing Season (April 2002–March 2003). In the

Þnal year of study, very signiÞcant differentiation wasevident between occurrences of H. armigera withinthe Darling Downs. Toowoomba, Kingaroy, Murgon,and Dalby each formed distinct groupings with sig-niÞcant AMOVA supports (Table 2). Nm values at thistime were comparatively very low (0.25Ð0.97), withproportions of immigrants low from other regions. The

Table 5. AMOVA results between Darling Downs and other major crop growing regions in eastern Australia (refer to Fig. 2) fromNovember 1999 to March 2003 (subset of data from Scott et al., 2005)

Regions compared% between regions

(P value)% between collections

(P value)% within collections

(P value)

1999Ð2000Darling Downs to Central Highlands (QLD) 5 (0.01) 12 (0.01) 83 (0.01)

2000Ð2001Darling Downs to Dawson-Callide Valleys (QLD) 4 (0.01) 14 (0.01) 82 (0.01)Darling Downs to Lower Namoi Valley (NSW) 2 (0.01) 13 (0.01) 85 (0.01)

2001Ð2002Darling Downs to Dawson-Callide Valleys (QLD) 0 (0.19) 14 (0.01) 86 (0.01)Darling Downs to Central Highlands (QLD) 4 (0.01) 13 (0.01) 83 (0.01)Darling Downs to Macintyre Valley (QLD) 1 (0.01) 10 (0.01) 89 (0.01)Darling Downs to Lower Namoi Valley (NSW) 0 (0.06) 13 (0.01) 87 (0.01)Darling Downs to Murrumbidgee (Southern NSW) 2 (0.01) 14 (0.01) 84 (0.01)

2002Ð2003Darling Downs to Dawson-Callide Valleys (QLD) 12 (0.01) 19 (0.01) 69 (0.01)Darling Downs to Macintyre Valley (QLD) 1 (0.03) 25 (0.01) 74 (0.01)Darling Downs to Murrumbidgee (Southern NSW) 9 (0.01) 20 (0.01) 71 (0.01)Darling Downs to Katherine (NT) 13 (0.01) 19 (0.01) 68 (0.01)

AMOVA analysis was annual for geographic regions that had substantial H. armigera pest pressure (P values � 0.05 are signiÞcant).

Table 4. Assignment of migrants (subset of data from Scott et al. 2005) identified using Monte-Carlo resampling of gametes (Paetkauet al. 2004) in GeneClass 2.3 (Piry et al. 2004)

Collection

% assigned

n n unassignedCentral

HighlandsDawsonCallide

DarlingDowns

MacintyreValley

NamoiValley

MurrumbidgeeSource

unknown

Darling Downs, Jan.ÐMar. 2000 225 20 3.9 96.1Darling Downs, Oct.ÐDec. 2000 63 0 1.6 87.3 11.1 0.0Darling Downs, Jan.ÐMar. 2001 30 7 4.3 87.0 8.7 0.0Darling Downs, JulyÐSept. 2001 15 6 0.0 11.1 77.8 11.1 0.0Darling Downs, Oct.ÐDec. 2001 30 13 0.0 17.6 70.6 5.9 0.0 0.0 5.9Darling Downs, Jan.ÐMar. 2002 15 4 0.0 9.1 72.7 0.0 18.2 0.0Darling Downs, Oct.ÐDec. 2002 108 21 1.1 94.3 2.3 2.3 0.0Darling Downs, Jan.ÐMar. 2003 95 20 89.3 10.7

Shaded areas indicate unavailability of samples.

December 2005 SCOTT ET AL.: REGIONAL GENE FLOW OF H. armigera in AUSTRALIA 2217

low migration rates and high levels of local differen-tiation are consistent with a minimal H. armigeramovement year. Perhaps an exception to the minimalmovement was the migration event during JanuaryÐMarch 2002 from the Murrumbidgee (18.2%) into theDarling Downs, which also was seen as a shift ingenotype between February 2002 and October 2002by PCA analysis. This Murrumbidgee to DarlingDowns moth immigration event may have contributedin part to the large genotype shift, but it is more likelythat this shift resulted from other immigration eventsbetween April and September 2002 that we were un-able to describe because of sample unavailability.Pheromone trap data from Jimbour and Brooksteadindicated that a large immigration event may haveoccurred between 29 September and 12 October 2002,because during that period there was a very high ßightdensity recorded, and it preceded the predicted emer-gence of local overwintering H. armigera.

The Darling Downs in this year was geneticallydifferent from all other regions studied at the time(DawsonÐCallide valleys, Macintyre Valley, Murrum-bidgee, and Katherine). Genotypes in the DarlingDowns were most different from those in Katherine(13%, P � 0.01; AMOVA) and the DawsonÐCallidevalleys (12%, P� 0.01; AMOVA) (Table 5). AMOVAdifferentiation was smallest between the DarlingDowns and the Macintyre Valley, and this is supportedby the assignment data that show the largest of theimmigration events into the Darling Downs havingcome for the Macintyre Valley (2.3%, OctoberÐDe-cember 2002; 10.7%, JanuaryÐMarch 2003), withsmaller contributions into the Darling Downs fromboth the DawsonÐCallide valleys (1.1%, OctoberÐDe-cember 2002), and the Murrumbidgee Valley (2.3%,OctoberÐDecember 2002).

Conclusions

Using microsatellites, we have detected and de-scribed population structuring and the level of H.armigeramovement on the local (intra-regional) scalewithin the Darling Downs, Australia. From April 2001to March 2002, genetic measures were consistent withsigniÞcant genetic mixing between the Darling Downsand other distant growing regions across (i.e., mixingbetween Darling Downs, Lower Namoi Valley, andDawsonÐCallide valleys). This is in concordance withthe high levels of migration observed in other Aus-tralian agricultural and horticultural regions in thisyear (Scott et al. 2005). In contrast, in other years suchas April 2002-March 2003, there was signiÞcant localstructuring of populations within the Darling Downsand very low levels of adult moth migration into theregion. The low migration in the Darling Downs isconsistent with observations in other Australian grow-ing regions where migration accounted for �10% ofindividuals sampled in most regions through this sea-son (Scott et al. 2005). These data provide furtherevidence moth movement differs from season to sea-son, highlighting the importance of studies in groupssuch as the Lepidoptera extending over consecutive

years, because short-term sampling may be misleadingwhen population dynamics and migration change sosigniÞcantly. It is outside the scope of the currentdiscussion to analyze the biological basis of such vari-ation in migration patterns, but it may be a result ofhost availability, meteorological conditions such asrainfall, wind direction, and speed of severe storms orto other limiting factors (Rochester 1999). During thecurrent study, most cropping regions suffereddrought, with rainfall varying from average to verymuch below average (Australian Bureau of Meteorol-ogy data), and these conditions may have effected themigratory behavior. This example also illustrates thatmicrosatellite analysis can detail pest populationstructuring and pest movement at a geographic scaleof 30Ð50 km in years where there is low inter-regionalgene ßow.

This research also demonstrates the importance ofmaintaining coordinated insecticide resistance man-agement strategy, because in some years H. armigerapopulations may be independent within a region andthus signiÞcantly inßuenced by local managementpractices. However, periods with high migration willoccur and resistance may rapidly spread. Current re-search is incorporating the resistance versus suscep-tible status and migratory origin of H. armigera todirectly track the movement of resistance across grow-ing regions.

Acknowledgments

We thank the growers and collectors from the DarlingDowns and in particular Jamie Hopkinson, Ian Crosthwaite,Geoff Cornwell, Hugh Brier, Rebecca Creagh, and John Gill-ingham. The map of the Darling Downs was generouslyprovided by Greenmount Press, and the Australian map pre-pared by Trevor Wardill. Finally, we thank Mark Schutze, JoKent, and Joanne McDiarmid for technical support. Financialsupport for this project was provided by the Grains Researchand Development Corporation, the Cotton Research andDevelopment Corporation, and The University of Queens-land.

References Cited

Cornuet, J.M., S.Piry,G.Luikart,A.Estoup, andM.Solignac.1999. New methods employing multilocus genotypes toselect or exclude populations as origins of individuals.Genetics 153: 1989Ð2000.

Daly, J.C., andP.Gregg. 1985. Genetic variation inHeliothisin Australia: species identiÞcation and gene ßow in twopest species H. armigera (Hubner) and H. punctigeraWallengren (Lepidoptera: Noctuidae). Bull. Entomol.Res. 75: 169Ð184.

Eanes, W. F., and R. K. Koehn. 1978. An analysis of geneticstructure in the monarch butterßy, Danaus plexippus L.Evolution 32: 784Ð797.

Excoffier,L., P.R. Smouse, and J.M.Quattro. 1992. Analysisof molecular variance inferred from metric distancesamong DNA haplotypes: application to human mitochon-drial DNA restriction sites. Genetics 131: 479Ð491.

Farrow, R. A., and J. C. Daly. 1987. Long-range movementsas anadaptive strategy in the genusHeliothis(Lepidoptera:Noctuidae): a review of its occurrence and detection infour pest species. Aust. J. Zool. 35: 1Ð24.

2218 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 98, no. 6

Fitt,G.P., and J.C.Daly. 1990. Abundanceofoverwinteringpupae and the spring generation of Helicoverpa spp.(Lepidoptera; Noctuidae) in northern New South Wales,Australia: implications for pest management. J. Econ. En-tomol. 83: 1827Ð1836.

Fitt, G. P., and M. L. Dillon. 1993. Spatial population mod-elling of Helicoverpa spp.: studies of adult behaviour andmovement, pp. 449Ð451. In S. A. Corey, D. J. Dall, andW. M. Milne [eds.], Pest control and sustainable agricul-ture. CSIRO Publications, Melbourne, Australia.

Fitt, G. P., M. L. Dillon, and J. G. Hamilton. 1995. Spatialdynamics of Helicoverpa populations in Australia: simu-lation modelling and empirical studies of adult move-ment. Comput. Electron. Agric. 13: 177Ð192.

Hartl, D. L., and A. G. Clark. 1997. Principles of populationgenetics, 3rd ed. Sinauer, Sunderland, MA.

Ji, Y.-J.., D.-X. Zhang, G. M. Hewitt, L. Kang, and D. M. Li.2003. Polymorphic microsatellite loci for the cotton boll-worm Helicoverpa armigera (Lepidoptera: Noctuidae)and some remarks on their isolation. Mol. Ecol. Notes 3:102Ð104.

Korman, A. K., J. Mallet, J. L. Goodenough, J. B. Graves, J. L.Hayes, D. E. Hendricks, R. Luttrell, S. D. Pair, and M.Wall. 1993. Population structure in Heliothis virescens(Lepidoptera: Noctuidae): an estimate of gene ßow. Ann.Entomol. Soc. Am. 86: 182Ð188.

McKechnie, S. W., M. E. Spackman, N. E. Naughton, I. V.Kovacs, M. Ghosn, and A. A. Hoffmann. 1993. Assessingbudworm population structure in Australia using the A-Trich region of mitochondrial DNA. In Proceedings of theBeltwide Cotton Conference, New Orleans 2: 838Ð840.

Michalakis, Y., and L. Excoffier. 1996. A generic estimationof population subdivision using distances between alleleswith special reference for microsatellite loci. Genetics142: 1061Ð1064.

Nibouche, S., R. Bues, J.-F. Toubon, and S. Poitout. 1998.Allozyme polymorphism in the cotton bollwormHelicov-erpa armigera (Lepidoptera: Noctuidae): comparison ofAfrican and European populations. Heredity 80: 438Ð445.

Orloci,L. 1978. Multivariate analysis in vegetation research.Dr W. Junk B.V., The Hague, The Netherlands.

Paetkau, D., R. Slade, M. Burden, and A. Estoup. 2004. Ge-netic assignment methods for the direct, real-time esti-mation of migration rate: a simulation-based explorationof accuracy and power. Mol. Ecol. 13: 55Ð65.

Peakall, R., and P. E. Smouse. 2001. GenA1Ex genetic anal-ysis in Excel. Population genetic software for teachingand research, version 5. Australian National University,Canberra, Australia. (http://www.anu.edu.au/BoZo/GenA1Ex/).

Peakall,R., P.E. Smouse, andD.R.Huff. 1995. Evolutionaryimplications of allozyme and RAPD variation in diploidpopulations of buffalograss (Buchloe dactyloides (Nutt.Engelm.). Mol. Ecol. 4: 135Ð147.

Piry, S., A. Alapetite, J.-M.Cornuet,D. Paetkau,L.Baudouin,and A. Estoup. 2004. GeneClass2: a software for geneticassignment and Þrst-generation migrant detection.J. Hered. 95: 536Ð539.

Rannala, B., and J. L. Mountain. 1997. Detecting immigra-tion by using multilocus genotypes. Proc. Natl. Acad. Sci.U.S.A. 94: 9197Ð9201.

Rochester, W. A. 1999. The migration systems of Helicov-erpa punctigera (Wallengren) and Helicoverpa armigera(Hubner) (Lepidoptera: Noctuidae) in Australia. Ph.D.dissertation, The University of Queensland, Brisbane,Australia.

Rochester, W. A., M. L. Dillon, G. P. Fitt, and M. P. Zalucki.1996. A simulation model of the long-distance migrationof Helicoverpa spp. moths. Ecol. Model. 86: 151Ð156.

Scott, K. D., K. S. Wilkinson, M. A. Merritt, L. J. Scott, C. L.Lange, M. K. Schutze, J. K. Kent, D. J. Merritt, P. R.Grundy, and G. C. Graham. 2003. Genetic shifts in He-licoverpa armigera (Hubner) (Lepidoptera: Noctuidae)over a year in the Dawson/Callide Valleys. Aust. J. Agric.Res. 54: 739Ð744.

Scott, K. D., C. L. Lange, L. J. Scott, andG. C. Graham. 2004.Isolation and characterisation of microsatellite loci fromHelicoverpa armigeraHubner (Lepidoptera: Noctuidae).Mol. Ecol. Notes 4: 204Ð206.

Scott, K. D., K. S. Wilkinson, N. Lawrence, C. L. Lange, L. J.Scott, M. A.Merritt, A. J. Lowe, andG. C. Graham. 2005.Gene-ßow between populations of cotton bollworm He-licoverpa armigera Hubner (Lepidoptera: Noctuidae) ishighly variable between years. Bull. Entomol. Res. 95:381Ð392.

Slatkin, M. 1985. Rare alleles as indicators of gene ßow.Evolution 39: 53Ð65.

Slatkin,M. 1987. Gene ßow and the geographic structure ofnatural populations. Science (Wash. DC) 236: 787Ð792.

Slatkin, M., and N. H. Barton. 1989. A comparison of threeindirect methods for estimating average levels of geneßow. Evolution 43: 1349Ð1368.

Stokes, N. H., S. W. McKechnie, and N. W. Forrester. 1997.Multiple allelic variation in a sodium channel gene frompopulations of AustralianHelicoverpa armigera (Hubner)(Lepidoptera; Noctuidae) detected via temperature gra-dient gel electrophoresis. Aust. J. Entomol. 36: 191Ð196.

Tan, S., X. Chen, A. Zhang, and D. Li. 2001. Isolation andcharacterization of DNA microsatellites from cotton boll-worm (Helicoverpa armigera,Hubner). Mol. Ecol. Notes1: 243Ð244.

Weir,B. S. 1990. Genetic data analysis. Sinauer, Sunderland,MA.

Wilson, G. A., and B. Rannala. 2003. Bayesian inference ofrecent migration using multilocus genotypes. Genetics163: 1177Ð1191.

Wright, S. 1931. Evolution in Mendelian populations. Ge-netics 16: 97Ð159.

Zhou, X., O. Faktor, S. W. Applebaum, and M. Coll. 2000.Population structure of the pestiferous mothHelicoverpaarmigera in the eastern Mediterranean using RAPD anal-ysis. Heredity 85: 251Ð256.

Received 22 March 2005; accepted 12 August 2005.

December 2005 SCOTT ET AL.: REGIONAL GENE FLOW OF H. armigera in AUSTRALIA 2219