Seediscussions,stats,andauthorprofilesforthispublicationat:http://www.researchgate.net/publication/40034761

Alethalovitrap-basedmasstrappingschemefordenguecontrolinAustralia:II.ImpactonpopulationsofthemosquitoAedesaegypti

ARTICLEinMEDICALANDVETERINARYENTOMOLOGY·DECEMBER2009

ImpactFactor:2.86·DOI:10.1111/j.1365-2915.2009.00834.x·Source:PubMed

CITATIONS

29

READS

69

8AUTHORS,INCLUDING:

CraigRWilliams

UniversityofSouthAustralia

76PUBLICATIONS1,159CITATIONS

SEEPROFILE

RebeccaMayAdamczyk

QueenslandHealth

6PUBLICATIONS66CITATIONS

SEEPROFILE

SharronLong

JamesCookUniversity

17PUBLICATIONS454CITATIONS

SEEPROFILE

ScottARitchie

JamesCookUniversity

223PUBLICATIONS4,531CITATIONS

SEEPROFILE

Availablefrom:RebeccaMayAdamczyk

Retrievedon:02October2015

Medical and Veterinary Entomology (2009) 23, 303–316

A lethal ovitrap-based mass trapping schemefor dengue control in Australia: II. Impacton populations of the mosquito Aedes aegypti

L. P. R A P L E Y1, P. H. J O H N S O N1, C. R. W I L L I A M S2, R. M. S I L C O C K1,M. L A R K M A N1, S. A. L O N G3, R. C. R U S S E L L4 and S. A. R I T C H I E1,3

1School of Public Health, Tropical Medicine and Rehabilitation Sciences, James Cook University, Cairns, Australia, 2Sansom

Research Institute, University of South Australia, Adelaide, Australia, 3Tropical Population Health Services, Queensland Health,

Cairns, Australia and 4Department of Medical Entomology, ICPMR, Westmead Hospital, University of Sydney, Westmead,

Australia

Abstract. In Cairns, Australia, the impacts on Aedes aegypti L. (Diptera: Culicidae)populations of two types of ‘lure & kill’ (L&K) lethal ovitraps (LOs), the standardlethal ovitrap (SLO) and the biodegradable lethal ovitrap (BLO) were measured duringthree mass-trapping interventions. To assess the efficacy of the SLO, two interventions(one dry season and one wet season) were conducted in three discrete areas, eachlasting 4 weeks, with the following treatments: (i) SLOs (>200 traps, ∼4/premise),BG-sentinel traps (BGSs; ∼15, 1/premise) and larval control (container reduction andmethoprene treatment) and (ii) larval control alone, and (iii) untreated control. FemaleAe. aegypti populations were monitored for 4 weeks pre- and post-treatment in allthree areas using BGSs and sticky ovitraps (SOs) or non-lethal regular ovitraps (ROs).In the dry season, 206 SLOs and 15 BGSs set at 54 and 15 houses, respectively, caughtand killed an estimated 419 and 73 female Ae. aegypti, respectively. No significantdecrease in collection size of female Ae. aegypti could be attributed to the treatments.In the wet season, 243 SLOs and 15 BGSs killed ∼993 and 119 female Ae. aegypti,respectively. The mean number of female Ae. aegypti collected after 4 weeks withSOs and BGSs was significantly less than the control (LSD post-hoc test). The thirdmass-trapping intervention was conducted using the BLO during the wet season inCairns. For this trial, three treatment areas were each provided with BLOs (>500,∼4/premise) plus larval control, and an untreated control area was designated. Adultfemale Ae. aegypti were collected for 4 weeks pre- and post-treatment using 15 BGSsand 20 SOs. During this period, 53.2% of BLOs contained a total of 6654 Ae. aegyptieggs. Over the intervention period, collections of Ae. aegypti in the treatment areaswere significantly less than in the control area for BGSs but not SOs. An influx ofrelatively large numbers of young females may have confounded the measurementof changes in populations of older females in these studies. This is an importantissue, with implications for assessing delayed action control measures, such as LOsand parasites/pathogens that aim to change mosquito age structure. Finally, the highpublic acceptability of SLOs and BLOs, coupled with significant impacts on femaleAe. aegypti populations in two of the three interventions reported here, suggest that

Correspondence: Scott Ritchie, School of Public Health, Tropical Medicine and Rehabilitation Sciences, James Cook University, PO Box 6811,Cairns, QLD 4870, Australia. Tel.: +61 740503619; fax: +61 740311440; e-mail: scott.ritchie@jcu.edu.au

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society 303

304 L. P. Rapley et al.

mass trapping with SLOs and BLOs can be an effective component of a dengue controlstrategy.

Key words. Aedes aegypti, BG-sentinel trap, biodegradable lethal ovitrap, dengue,lethal ovitrap, ‘lure and kill’, mass trapping, vector control, Australia.

Introduction

Novel measures to control dengue have been championedrecently (Kroeger & Nathan, 2006; Reiter, 2007; Farrar et al.,2007; Morrison et al., 2008). This is partly in response to fail-ures to reduce epidemic dengue transmission, and the relianceon pesticide use in and around homes that increases the risk ofpublic exposure to insecticides. Reiter (2007) also emphasizedthat novel methods should be developed to exploit character-istic behaviours of Aedes aegypti L, the recognized principalvector of dengue, such as its highly domestic niche and thepropensity to oviposit in several water-holding containers insuccession (skip oviposition; Colton et al., 2003). In pursuit ofthis goal, several workers have developed and tested a lethalovitrap (LO). The LO is an example of a ‘lure and kill’ (L&K)trap, whereby the target insect (e.g. Ae. aegypti ) is lured tothe ovitrap by an infusion known to be attractive to oviposit-ing females and is killed when it contacts a lethal ovistripin the trap. When employed in large numbers, kairomone-baited insect traps can significantly reduce populations of targethematophagous insects (Kline, 2006), and mass trapping withLOs would specifically target gravid, older mosquitoes that aremore at risk of being infected with dengue virus.

Zeichner & Perich (1999) tested a small (473 mL capacity)black plastic cup, featuring a velour paper ovistrip treatedwith a synthetic pyrethroid, and filled with hay infusion toattract and kill gravid Ae. aegypti. Ultimately, they patenteda deltamethrin-treated LO that has been used in publishedtrials in Thailand (Sithiprasasna et al., 2003) and Brazil (Perichet al., 2003). Typically, several LOs are placed inside andoutside houses or buildings, and refilled with water and anew lethal strip as needed. Kittayapong et al. (2008), using apermethrin-treated LO in conjunction with source reduction,larval control and insecticide-treated jar covers, reportedsuccessful reduction of Ae. aegypti populations and denguetransmission in Thailand.

Our research unit has developed a larger LO using a 1.2-L black plastic bucket and a bifenthrin-treated lethal ovistrip,referred to here as a standard lethal ovitrap (SLO), that canbe deployed outdoors (Williams et al., 2007a). In Australia,Queensland Health personnel have begun using this SLO indengue interventions. An integrated programme of larval con-trol and reduced interior residual spraying undertaken by theDengue Action Response Team (DART) of Queensland Health,under the auspices of the state Dengue Fever Management Planfor North Queensland (http://www.health.qld.gov.au/dengue)has led to the suppression of dengue transmission and elim-ination of the virus in four outbreaks from 2004 to 2008

(Montgomery et al., 2005; Ritchie, 2005; S. Ritchie, unpub-lished data).

An important drawback to plastic SLOs is that they canbecome significant mosquito breeding sites themselves oncethe insecticidal strip breaks down, thus necessitating laboriousand costly retrieval operations. To overcome this limitation,Ritchie et al. (2008) developed a LO composed of a biodegrad-able thermoplastic starch The first biodegradable lethal ovitrap(BLO), composed of 30–35% plastic/amylose maize polymersblend, was shown to be as attractive to ovipositing Ae. aegyptias a regular plastic ovitrap (RO) (Ritchie, 2001; Ritchie et al.,2008). In a companion paper, we demonstrate that mass trap-ping with the LO or the BLO is acceptable to the public, andthe BLOs exhibited a good biodegradation rate (Ritchie et al.,2009).

Field studies measuring the impact of LOs on populations ofAe. aegypti, while generally favourable, have produced mixedresults. Perich et al. (2003) placed 10 LOs, 5 inside and 5outside, each of 30 contiguous premises in two communitiesin Brazil, and 30 premises immediately outside this treatmentcluster to provide a treatment buffer. While three populationmeasures were used (containers positive for larvae and/orpupae and adult females collected inside houses), the meannumber of adult females collected inside houses was a criticalmeasure. The mean number of female Ae. aegypti collected perhouse was significantly reduced 1 month after LO deploymentin one municipality, but it took almost 3 months in a differentmunicipality (Perich et al., 2003). A significant reduction inadult females per house was obtained in only one village. Non-serviceable containers within the treatment area were suspectedto have reduced the impact of the LOs by providing alternativeovipositional sites and by producing adult Ae aegypti. Usingthe Zeichner & Perich (1999) LO in Thailand, Sithiprasasnaet al. (2003) reported a negligible impact on a number ofmeasures of Ae. aegypti abundance in a 1999 trial, possibly as aresult of fungal growth on the lethal ovistrip. While control wasbetter in their 2000 trial, it was not absolute and competitionfrom alternative oviposition sites was again proposed as thereason for suboptimal results.

We describe the impact of mass-deployment of SLOs andBLOs on field populations of Ae. aegypti in Cairns. We alsoexamine the impact of a different type of L&K technology,the BGS, which attracts host and harbourage-seeking females(Krockel et al., 2006; Williams et al., 2006a) for removal trap-ping of Ae. aegypti when used in conjunction with SLOs. Wediscuss why control technologies that target older mosquitoes,such as the SLO and BLO, may not significantly reducemosquito populations, but may, nonetheless, be effective inreducing dengue transmission.

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 305

Materials and methods

Lure and kill interventions using lethal ovitraps and BGS traps

Study site. Two field trials of L&K traps against Ae. aegyptiwere conducted during the dry (June to July 2006) and wet(March to April 2007) seasons in the Cairns suburb of MachansBeach (population 893). Aedes aegypti is common and the areahas had multiple outbreaks of dengue (Ritchie et al., 2001;Ritchie et al., 2002; Hanna et al., 2006). The area is geograph-ically isolated within sugar cane fields and beach scrub, whichhelps reduce the immigration of mosquitoes from non-treatedurban areas that could confound treatment effects (Koenraadtet al., 2007). Meteorological data (rainfall, temperature andevaporation) were obtained from an Australian Bureau ofMeteorology station located ∼1 km from the study site.

Treatments. Using a geographic information system (GIS)(MapInfo; http://www.mapinfo.com.au), Machans Beach wassplit into two comparably-sized (200-m radius) treatmentenclaves and a third reference enclave formed from housesoutside the treatment enclaves served as the untreated con-trol (Fig. 1). Populations of Ae. aegypti were monitored (seepopulation assessment) within 50 m of a hypothetical denguecase house in all enclaves for 3–4 weeks prior to the appli-cation of treatments to estimate baseline population densitiesand their degree of fluctuation. Treatments were carried outwithin 200 m of a hypothetical dengue case house, with bothtreatment enclaves containing >70 houses. Two teams of threepeople implemented the treatments over 3 days (dry season3–5 July 2006 and wet season 26–28 March 2007). The twotreatments, (i) adult L&K trapping (SLOs and BSGs) plus lar-val control and (ii) larval control alone, are described below.

Adult lure and kill trapping plus larval control. This treat-ment was undertaken to replicate the activities carried outby the DART when they respond to a reported dengue case.SLOs (1.2-L bucket with a 13.5 cm by 5 cm bifenthrin-treatedovistrip, 1 L of water and a 0.5-g alfalfa pellet according toWilliams et al., 2007a) were placed outside houses, usually at adensity of 4/premise, for a total of >200 traps within the 200-mradius of the house. They were set in secluded and shelteredsites to reduce degradation of the pesticide strip as a resultof rainfall and direct exposure to sunlight (Williams et al.,2006b). BGSs were run continuously (1/premise, for a total of∼15 traps) in premises within 100 m of the proxy case house insheltered areas frequented by the house occupants such as thelaundry, verandah or work shed. Both SLOs and BGSs weredeployed for a period of 4 weeks in the dry and wet seasons.The duration of the intervention was limited to 4 weeks becausethat is, on average, how long LOs are effective without needingto be revisited to top-up the water and replace the insecticidalstrip. Larval control was applied as described below.

After 4-weeks deployment, the SLOs were collected. Thestructural condition of the SLOs and extent of ovipositionwithin them are reported in the companion paper (Ritchieet al., 2009). Lethal ovistrips were removed and mosquito eggs

counted and removed for rearing to fourth instar for identifica-tion (Williams et al., 2007a). The mean number of Ae. aegyptieggs/SLO was estimated by multiplying the number ofeggs/ovistrip by the percent that were Ae. aegypti, from whichthe number of Ae. aegypti females killed by SLOs and the asso-ciated standard error were estimated (Williams et al., 2007a).

Larval control. Larval control was undertaken as a once-offstand-alone treatment at the start of the intervention in bothtreatment enclaves. Larval control involved reducing actual andpotential Ae. aegypti immature habitats by turning over bothwet and dry containers, and treating permanent wet containersand roof gutters with s-methoprene pellets. No insecticidesthat killed adult mosquitoes were employed. Larval controlwas conducted in every yard, with permission, that was within200 m of the case house, which included yards whose propertyboundary was bisected by the 200-m boundary (Fig. 1).

Population assessment: dry season intervention. Mosquitoescollected by the BGSs set for L&K were also used as a popu-lation measure. Although the BGSs were run continuously, themosquitoes captured during 2 consecutive days each week wereused to measure populations of Ae. aegypti. In the untreatedcontrol area, 15 BGS traps were only run during the 2-dayperiod. Mosquito collections were returned to the laboratoryfor identification. We also used ROs to estimate female Aeaegypti populations. Single ROs consisting of a 1.2-L blackplastic bucket with an untreated flannel cloth ovistrip, contain-ing 1 L of tap water and a 0.5-g lucerne pellet (Ritchie, 2001)were placed outdoors at 10 premises in each enclave and mon-itored weekly. Eggs were hatched to determine the percentageAe. aegypti.

Population assessment: wet season intervention. During thewet season intervention, female Ae. aegypti were monitoredusing BGSs as above. All female Ae. aegypti caught in BGSswere frozen for later dissection to determine their physiologicalstatus (nulliparous, parous and gravid) both pre- and post-intervention for each enclave. Sticky ovitraps (SOs) (Ritchieet al., 2003) were used to monitor gravid females. Everyweek, 20 SOs each containing 1 L of tap water and 0.5 g oflucerne (Ritchie, 2001) were placed outdoors, two per premise,at select premises, in secluded and sheltered sites (Williamset al., 2006b). At 5 weeks post-treatment, an additional SOcollection was analysed as above to determine if the gravidfemale population rebounded after removal of the SLOs.

To compare dry and wet season Ae. aegypti populations,Breteau Indices (BIs) (Service, 1993) were calculated fromcontainer survey data that were collected from the untreatedcontrol enclave at the start of the intervention.

Data analysis. Counts of female Ae. aegypti collected bymonitoring traps (BGSs, ROs and SOs) in both interventionswere log (X + 1) transformed to obtain a normal distribu-tion and tested for treatment effects using a general linearmodel repeated measures (spss V14) with a type III error

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

306 L. P. Rapley et al.

Fig. 1. Map of standard lethal ovitrap (SLO) mass-trapping trial at Machans Beach showing the two treatment areas including the hypotheticaldengue case houses for the dry season. The wet season lethal ovitrap intervention was similar with treatment enclaves shifted to the northeast.

term. For both the wet and dry seasons this test was con-ducted for two time periods: pre- and post-intervention. Thepost-intervention period was analysed at 3 and 4 weeks afterthe intervention started. The analysis at 3 weeks was conductedbecause a high number of SLOs were expected to be dry after 4weeks and hence would no longer be contributing to mosquitocontrol. Finally, to identify significantly different treatments,the least significant difference (LSD) post-hoc test wasapplied.

The frequency of parous and nulliparous mosquitoes wascompared between treatment enclaves and the untreated controlenclave. A 2 × 2 contingency table with Yates correctionand the chi-square statistic (Zar, 1999) was used to testfor independence between physiological state and treatmentarea. Separate tests were performed on pooled data for bothpre-intervention and post-intervention periods. Weekly datawere pooled after heterogeneity chi-square testing (Zar, 1999)revealed no significant difference between frequencies.

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 307

Biodegradable lethal ovitrap intervention

Study site. The Cairns suburb of Parramatta Park (popula-tion 2900) was chosen as the experimental site because of itshigh Ae. aegypti populations and past occurrence of dengue(Ritchie et al., 2004; Hanna et al., 2006). Daily temperatureand evaporation was obtained from an Australian Bureau of

Meteorology station located ∼5 km from the study site, withrainfall collected ∼1 km away.

Treatment and sampling. The southern end of ParramattaPark was split into three distinct geographical areas of compa-rable size for use as treatment areas (Grimshaw, Victoria andPalm) using GIS (MapInfo) (Fig. 2). A fourth similar sized

Fig. 2. Map of biodegradable lethal ovitrap (BLO) mass-trapping trial in Parramatta Park showing the three treatment areas. The untreated controlarea was located ∼1 km west and is not shown. ‘K’ denote key premise containing 14 Aedes aegypti breeding sites found at the end of theintervention.

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

308 L. P. Rapley et al.

area was located in the neighbouring suburb and used as anuntreated control area. Aedes aegypti populations were mon-itored for 4 weeks pre- and post-treatment (30 January to 20March 2008) in each area using BGSs and SOs as describedfor the wet season SLO intervention. All female Ae. aegypticaught in BGSs were frozen for later dissection to determinetheir physiological status both pre- and post-intervention foreach enclave.

Over 3 days (wet season, 18–20 February 2008) three teamsof three people each implemented larval control, as describedbelow and placed BLOs [1.2-L bucket with a 13.5 cm by5 cm bifenthrin-treated ovistrip, 1 L of tap water (Ritchieet al., 2008)] outside houses (usually 4/house; Williams et al.,2006b) in secluded and sheltered sites similar to those used byWilliams et al. (2006b) for SLOs, and left for 4 weeks. At thebeginning and end of the intervention, yards were surveyedfor containers, and larvae from positive containers were iden-tified and used to calculate the BI for each enclave. Larvalcontrol was conducted in yards and consisted of source reduc-tion (turning over small containers) and treatment of immove-able artificial containers with Bacillus thuringiensis israeliensis(Bti) (Vectobac WG) at a rate of 80 mg/L (240 000 ITU/L).Roof gutters were treated with s-methoprene pellets (Ritchieet al., 2001).

At the end of this period the BLOs were collected andinformation relating to their performance in the field recorded(Ritchie et al., 2009). The number of mosquito eggs present oneach lethal ovistrip was counted under a stereo-microscope.The number of Ae. aegypti eggs was estimated using thepercentage of sticky ovitraps collections that were Ae. aegypti,and the number of female Ae. aegypti potentially killed wasestimated using the formula of Williams et al. (2007a).

Data analysis. Mean number adult female Ae. aegypti col-lected by BGSs and SOs was log (X + 1) transformed to obtaina normal distribution and tested for treatment effects usinga general linear model repeated measures (spss V14; SPSSInc, http://www.spss.com/au/) with a type III error term, asdescribed for the LO interventions. The frequency of parousand nulliparous mosquitoes was compared between combineddata from treatment enclaves and the untreated control enclaveas described above. Pooled data were tested in 2 × 2 contin-gency tables after heterogeneity chi-square testing (Zar, 1999).

Results

Impact of lure and kill trapping using lethal ovitraps andBGS traps

The deployment of SLOs in the L&K enclave resulted inalmost twice the mortality of female Ae. aegypti in the wetcompared with the dry season (Table 1). Continuously runBGSs collected a total of 73 and 119 female Ae. aegyptiin the dry and wet season, respectively. All larvae hatchedfrom a sample of eggs (dry season n = 98, wet season n =157) collected on bifenthrin-treated ovistrips were Ae. aegypti,therefore the number of adults killed could be determined

Table 1. Impact of lure and kill trapping using standard lethalovitraps (SLO) and BG-sentinel traps (BGSs) plus larval control fora 4-week intervention at Machans Beach for both a dry and a wetseason.

Dry season Wet season

Lethal ovitrapsNumber of houses with traps/total

number of houses54/72 64/90

Number of traps 206 243Number of eggs on ovistrips 1189 2817% Ae. aegypti eggs (n) 100% (98) 100% (157)Mean ± SE no. adult female Ae.

aegypti killed∗419 ± 23 993 ± 56

Mean ± SE no. adult female Ae.aegypti killed/trap

2.03 4.08

BGS trapsNumber of houses with traps

(1 trap/house)15 16

Adult female Ae. aegypti caughtand killed

73 119

Mean no. adult female Ae. aegyptikilled/trap

4.87 7.44

∗Based on formula of Williams et al. (2007a) where 2.84 eggs perfemale Ae. aegypti killed by a LO.

(Williams et al., 2007a) without having to adjust egg counts forthe presence of other species (Table 1). Although more SLOswere used during the wet season, the higher mortality wasprincipally as a result of the increase in mosquito populations,as demonstrated by the doubling of the number of Ae. aegyptikilled in SLOs in the wet compared with the dry season.Furthermore, in the wet season, a greater percentage of SLOswere positive for mosquito eggs and more adult female Ae.aegypti were caught through the continuous use of BGSsduring the intervention period (Table 1).

Population assessment–comparison between seasons. Pre-treatment collections of female Ae. aegypti in BGSs werehigher in all enclaves in the wet compared with the dry season(Figs 3a and 4a). This was reflected in the higher BI in theuntreated control enclave in the wet (75) compared with thedry (33) season, and was probably because of higher rainfall inMarch and April 2007 (wet season–282 mm) compared withJune and July 2006 (dry season–135 mm).

Population assessment–dry season. Collections of femaleAe. aegypti using BGSs 3 and 4 weeks after the interventionwere significantly different between enclaves, with collectionsin the L&K enclave being significantly smaller than in both thelarval control enclave and untreated control enclave (Table 2).However, despite the declining number of mosquitoes caughtin BGSs in the L&K enclave after the intervention (Fig. 3a),the overall difference between the L&K enclave and theuntreated control enclave could not be attributed to the inter-vention, as collections in the L&K enclave were already sig-nificantly smaller prior to the intervention (Table 2). There

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 309

Fig. 3. Population samples of Aedes aegyptiin the dry season lethal ovitrap interven-tion showing (a) mean ± SE number adultfemales per BG-sentinel (BGS) over 48 h(n = 15 for each enclave) and (b) mean ±SE number Ae. aegypti eggs per regular ovi-trap set for weekly intervals (n = 10 foreach enclave). Black arrow indicates the dateexperimental treatments were applied (inter-vention date).

was evidence that L&K trapping improved mosquito controlbeyond that of larval control on its own, as significantly fewermosquitoes were collected by BGSs in the L&K enclave com-pared with the larval control enclave after the intervention(Table 2). The low mosquito numbers in all enclaves, particu-larly the L&K enclave prior to the intervention, was probablybecause of the prevailing dry weather.

The use of ROs produced highly variable data, and despitemean egg numbers decreasing in the L&K enclave afterthe intervention (Fig. 3b), collections were not significantlylower compared with the other two enclaves post-intervention(Table 2).

Population assessment–wet season. Collections of femaleAe. aegypti 3 weeks after the intervention were significantlydifferent between enclaves, with fewer mosquitoes caught inBGSs in the L&K enclave compared with either the larvalcontrol enclave or the untreated control enclave (Fig. 4).However, unlike the dry season, the differences could beattributed to the intervention, as collections in the L&Kenclave only became significantly smaller after the intervention(Table 3). Four weeks after the intervention, BGSs collectionsin the L&K enclave were still lower compared with the larvalcontrol enclave or the untreated control enclave (Fig. 4a).

The deployment of SLOs in the L&K enclave also resultedin a dramatic decrease in the number of female Ae. aegypti

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

310 L. P. Rapley et al.

Fig. 4. Population samples of Aedes aegyptiin the wet season lethal ovitrap interven-tion showing (a) mean number ± SE adultfemales per BG-sentinel (BGS) over 48 h(n = 15 for each enclave) and (b) mean num-ber ± SE adult female A. aegypti per stickyovitrap set for weekly intervals (n = 20 foreach enclave). Black arrow indicates the dateexperimental treatments were applied (inter-vention date).

caught in SOs (Fig. 4b). In the L&K enclave, the weekly

mean number of female Ae. aegypti per SO declined from

0.55 pre-intervention to 0.07 post-intervention, a decrease

of 87%. Three and four weeks after the intervention, this

decrease was significantly (P < 0.05) different than the larval

control enclave and both larval control and untreated control

enclaves, respectively (Table 3). Five weeks post-intervention,

a significant increase (Paired t-test; t = 2.36; d.f. = 9; P =0.04) in the number of female Ae. aegypti caught in SOs

in the L&K enclave was observed (Fig. 4b), and the catch

of gravid Ae. aegypti in the L&K enclave and the untreatedcontrol enclave were not significantly different.

Pre-intervention, the frequencies of parous and nulliparousfemales were not significantly difference between the L&Kenclave and the untreated control area (χ2 = 0.92; P =0.34), and there was no difference between the larval controlonly and untreated control area (χ2 = 0.43; P = 0.33). Post-intervention, there were no significant changes in the frequencyof parous mosquitoes in the L&K enclave relative to theuntreated control area (χ2 = 0.17; P = 0.68), or in the larvalcontrol only area (χ2 = 0.00; P = 1.00) (Table 4).

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 311

Table 2. Statistical analysis (repeated measures anova and post-hoc LSD) of data from monitoring BG-sentinel (BGS) traps (female Aedesaegypti ) and regular ovitraps (Ae. aegypti eggs) collected from the standard lethal ovitrap (SLO) dry season intervention, in the two treatmentenclaves (lure and kill traps + larval control and larval control) and the untreated control enclave.

Trap Test Pre-intervention (−4 weeks) Post-intervention (+3 weeks) Post-intervention (+4 weeks)

BG-Sentinel anova F = 3.38, P = 0.04 F = 6.12, P = 0.01 F = 5.67, P = 0.01LSD L&K + larval controla L&K + larval controla L&K + larval controla

Larval controlab Larval controlb Larval controlb

Untreated controlb Untreated controlb Untreated controlb

Regular ovitrap anova F = 0.04, P = 0.96 F = 1.12, P = 0.34 F = 0.90, P = 0.42LSD L&K + larval controla L&K + larval controla L&K + larval controla

Larval controla Larval controla Larval controla

Untreated controla Untreated controla Untreated controla

Represents post-hoc LSD grouping. Enclaves with the same letter did not differ significantly.

Table 3. Statistical analysis (repeated measures anova and post-hoc LSD) of data from monitoring BG-sentinel (BGS) traps (female Aedesaegypti ) and regular ovitraps (Ae. aegypti eggs) collected from the standard lethal ovitrap (SLO) wet season intervention, in the two treatmentenclaves (lure and kill traps + larval control and larval control) and the untreated control enclave.

Trap Test Pre-intervention (−3 weeks) Post-intervention (+3 weeks) Post-intervention (+4 weeks)

BGS anova F = 0.28, P = 0.76 F = 3.74, P = 0.03 F = 3.01, P = 0.06LSD L&K + larval controla L&K + larval controla L&K + larval controla

Larval controla Larval controlb Larval controlb

Untreated controla Untreated controlb Untreated controlb

Sticky ovitrap anova F = 0.18, P = 0.84 F = 5.06, P = 0.01 F = 5.47, P < 0.01LSD L&K + larval controla L&K + larval controla L&K + larval controla

Larval controla Larval controlb Larval controlb

Untreated controla Untreated controlab Untreated controlb

Represents post-hoc LSD grouping. Enclaves with the same letter did not differ significantly.

Impact of lure and kill trapping using biodegradable lethalovitraps

Most (93%) premises in each of the treatment areas wereaccessed in order to conduct larval control and set BLOs(Table 5). Overall a total of 553 BLOs was placed in 166premises, with a mean of 3.3 per premise. Over half of therecovered BLOs had eggs on the ovistrips (Table 5), and atotal of 6654 eggs was laid mostly (53%) in the Grimshaw area.The total number of female Ae. aegypti killed was estimated tobe 2345 + 132. Larval productivity was high in the treatmentareas, with BIs of 58, 57 and 156 for Grimshaw, Victoria andPalm, respectively. The untreated control area had a BI of 100.The post-treatment BI decreased to 19 and 24 at Victoria andPalm areas, respectively, but rose to 62 in Grimshaw. Thislatter result was largely because of a single premise, whichwe could not access initially, and was later found to have 13flooded containers producing Ae. aegypti (Fig. 2). Excludingdata from this key premise, the BI at Grimshaw dropped from58 to 47. We were unable to conduct a post-treatment containersurvey in the untreated control area.

Population assessment. The mean number of female Ae.aegypti collected by BGSs and SOs in the untreated controland three intervention areas during the pre-treatment periodwas comparable (P > 0.05) (Table 6). However, all threeintervention areas showed a significant (P < 0.05) decrease in

BGS collections for the 3- and 4-week post-treatment periods(Fig. 5a). SO collections also showed a significant difference 3and 4 weeks post-treatment, but in this case mean collectionsfrom the Grimshaw treatment area were significantly higherthan the other two treatment areas and the untreated controlarea (Fig. 5b), perhaps as a result of mosquitoes from the keypremise (Fig. 2).

Table 4. Summary of physiological status of female Aedes aegypticollected in BG-sentinel (BGS) traps in the lure and kill enclave(SLO + larval control), larval control only enclave and untreatedcontrol enclave, pre- and post-intervention at Machans Beach in thewet season.

Physiological status (%)∗

Nulli- Not TotalEnclave parous Gravid Parous known dissected

Pre-interventionLure and kill 46 (44%) 10 (10%) 48 (46%) 12 116Larval control 31 (30%) 13 (13%) 59 (57%) 19 122Untreated control 32 (32%) 22 (22%) 47 (46%) 21 122Post-interventionLure and kill 13 (36%) 2 (6%) 21 (58%) 3 39Larval control 40 (38%) 12 (12%) 52 (50%) 13 117Untreated control 33 (40%) 8 (10%) 41 (50%) 12 94

∗Per cent based on total of nulliparous + gravid + parous.

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

312 L. P. Rapley et al.

Table 5. The impact of mass trapping using biodegradable lethalovitraps (BLO) for a 4-week deployment in three treatment areas(Grimshaw, Victoria and Palm).

Grimshaw Victoria Palm Total

Number of houseswith traps/totalhouses

66/69 54/61 46/49 166/179

Number of traps 213 183 156 552Number of traps

with eggs (%)130 (61) 97 (53) 67 (43) 294 (53)

Number of eggson ovistrips

3534 2009 1111 6654

Mean ± SE∗ no.adult femaleAe. aegyptikilled

1245 ± 70 708 ± 40 391 ± 22 2345 ± 132

Mean no. femaleAe. aegyptikilled/trap

5.8 3.8 2.5 4.2

∗Based on formula of Williams et al. (2007a) where 2.84 eggs perfemale Ae. aegypti killed by a LO.

Table 6. Statistical analysis (repeated measures anova and post-hocLSD) of data for female Aedes aegypti collections from monitoringBG-sentinel (BGS) traps and sticky ovitraps for the three treatmentareas (lure and kill + larval control, larval control alone, and untreatedcontrol area) in the biodegradable lethal ovitrap (BLO) intervention.

Pre Post PostTrap Test (−4 weeks) (+3 weeks) (+4 weeks)

BG-Sentinel anova F = 1.78 F = 4.39 F = 4.96P = 0.16 P = 0.01 P = 0.00

LSD Palma Palma Palma

Grimshawa Grimshawa Grimshawa

Victoriaa Victoriaa Victoriaa

Untreatedcontrola

Untreatedcontrolb

Untreatedcontrolb

Sticky ovitrap anova F = 0.97 F = 3.60 F = 4.89P = 0.41 P = 0.02 P < 0.01

LSD Grimshawa Grimshawa Grimshawa

Victoriaa Palmb Palmb

Palma Victoriab Victoriab

Untreatedcontrola

Untreatedcontrolb

Untreatedcontrolb

Enclaves with the same letter did not differ significantly.

Pre-intervention, the frequency of parous compared withnulliparous females revealed no significant difference betweenthe L&K enclave and the untreated control area (χ2 = 3.06;P = 0.08). Post-intervention, again the frequency of parousmosquitoes in the L&K enclave relative to the untreated controlarea were not significantly different (χ2 = 2.26; P = 0.13)(Table 7).

Discussion

Similar to results reported in Brazil (Perich et al., 2003) andThailand (Sithiprasasna et al., 2003), mass trapping (Kline,

2006) using SLOs resulted in inconsistent impacts on femaleAe. aegypti populations. The dry season intervention usingSLOs, BGSs and larval control reduced populations, but notsignificantly so. The wet season intervention caused significantreductions in both BGS and SO collections. The BLO + larvalcontrol wet season intervention significantly reduced the meannumber of female Ae. aegypti collected in BGSs, but not inSOs. Thus, over the three interventions, significant treatmenteffects were found in two-third of BGS collections and one-third of the ovitrap collections, i.e. 50% of the interventions intotal. Nonetheless, with the exception of SO collections in theBLO intervention in the Grimshaw area (Fig. 5b), L&K masstrapping was not associated with an increase in female Ae.aegypti. Proximate explanations for the insignificant populationchanges between treatment and control areas include (i) lowpopulations in the control area (BGS collections dry seasonL&K enclave, Fig. 3a), (ii) high variability and low replicationof sampling method (RO collection, dry season L&K enclave,Fig. 3b), (iii) the natural collapse of populations within thecontrol area (SO collections, BLO intervention, Fig. 5b) and(iv) seeding of the intervention area from a key premise thatwas not treated (SO collections, BLO intervention Grimshawenclave, Figs 2 and 5b). Indeed, the low collections duringthe pre-treatment phase of the dry season L&K intervention(Fig. 3a) indicate that the intervention should have beenpostponed. Other factors, such as untreated, cryptic breedingsites, new mosquito breeding sites (including pet bowls),failure of the Bti treatment, immigration of adults frombeyond the treatment zone (Koenraadt et al., 2007) and fungalcontamination of SLOs (Sithiprasasna et al., 2003) could havereduced the direct impact of LOs. The post-treatment BIsindicate the continued presence of larvae in the treatmentareas, especially Grimshaw. No consistent impact on agestructure was evident from the lack of significant change inthe percentage of parous mosquitoes.

What can we conclude about the impact of the SLO/BLOmass trapping approach? In the companion paper, we foundthat the public accepted the use of SLOs and BLOs, withonly a small percentage of traps being removed. GravidAe. aegypti were attracted to and readily oviposited in theSLOs and BLOs. Laboratory assays of the lethal ovistrips after4-weeks field exposure (Williams et al., 2007a; Ritchie et al.,2009) demonstrated that they still killed over 70% of femaleAe aegypti in small cage trials. Estimates using the formulaof Williams et al. (2007a) suggest that hundreds of femaleAe. aegypti were killed. A significant decline in female Ae.aegypti that can be attributed to the treatment was found in halfof the population measures, and in all but one case (Grimshawtreatment enclave, Fig. 5b) female Ae. aegypti populationsdeclined in the intervention area. However, we do not knowthe impacts on older mosquitoes (those that are potentiallyinfective), as we do not have a method for directly measuringtheir numbers.

Insignificant declines in populations are to be expected,as outlined below, and serve to highlight the problems ofmeasuring the impact of a delayed action control methodon insects that are present in low densities. Lethal ovitrapsare a delayed action control method, as are life-shorteningstrains of the endosymbiont Wolbachia (Brownstein et al.,

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 313

Fig. 5. Population samples of Aedes aegyptiin the biodegradeable lethal ovitrap (BLO)intervention showing (a) mean ± SE numberof Ae. aegypti adult females per BG-sentinel(BGS) set for 48 h (n = 15 for each treatmentarea) and (b) mean number ± SE of Ae. aegyptiadult females per sticky ovitrap set for weeklyintervals (ca. n = 15 for each treatment area).Black arrow indicates the date experimentaltreatments were applied (intervention date).

2003) and various pathogenic fungi (Blanford et al., 2005).Lethal ovitraps specifically target ovipositing mosquitoes that,in the case of Ae. aegypti, are likely to be >5 days old. FemaleAe. aegypti must mate, obtain a bloodmeal, and develop eggsbefore visiting an ovitrap. Using CIMSiM (Focks et al., 1993),we estimate that Ae. aegypti begins ovipositing 4–8 days old at28◦C, the average temperature in Cairns during the wet season.The extrinsic incubation period for dengue in Ae. aegypti is∼10 days at 28◦C (Focks et al., 1995). Allowing for time toobtain a viraemic bloodmeal, these mosquitoes are probably11–14 days old before they begin transmitting virus. However,the bulk of the population consists of mosquitoes <6 days old,which are not targeted by these methods. Assuming a dailysurvival probability of 0.85 for female Ae. aegypti (based ona range of 0.80–0.91; Focks et al., 1993; Muir & Kay, 1998;

Harrington et al., 2001), 63% of the population is <6 days old.Thus, relatively large populations of young mosquitoes woulddampen the impact of control that targets older mosquitoes, andhence, the total population may not be significantly reduced.

Measuring the impact of delayed action control methods that‘change the age structure of vector populations’ is difficult.In the case of the LO interventions (Perich et al., 2003;Sithiprasasna et al., 2003; this study), adult females werecollected using aspirators, BGSs and ovitraps. The first twomethods collect across all age groups, with both aspiratorand BGS collections in Cairns consisting of 17 and 42%nulliparous Ae. aegypti, respectively (Williams et al., 2006a).In this study, 35% of the female Ae. aegypti collected byBGSs in the SLO wet season intervention untreated controlarea were nulliparous. Use of sampling methods that target

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

314 L. P. Rapley et al.

Table 7. Summary of physiological status of female Aedes aegypticollected in monitoring BG-sentinel (BGS) traps in the lure and killenclaves (biodegradable lethal ovitrap + larval control) and untreatedcontrol enclave pre- and post-intervention at Parramatta Park.

Physiological status (%)∗

Nulli- Not TotalEnclave parous Gravid Parous known dissected

Pre-interventionLure and kill 22 (9%) 192 (75%) 42 (16%) 101 357Untreated control 22 (20%) 71 (63%) 19 (17%) 33 145Post-interventionLure and kill 33 (14%) 170 (74%) 28 (12%) 45 276Untreated control 13 (14%) 57 (61%) 23 (25%) 23 116

∗Per cent based on total of nulliparous + gravid + parous.

older mosquitoes, such as gravid traps and sticky ovitraps,should help, but even then a majority of the population will beyounger than the minimal age of infected mosquitoes.

Alternative methods for assessing population interventionsare required. Ultimately, methods to accurately age mosquitoeswould enable us to disregard young mosquitoes in populationestimates. Assessment of parity allows us to measure thenumber of mosquitoes that have oviposited and are likely tobe >6 days old. Novel gene expression methods offer promiseto categorize mosquitoes into even older age groups (Cooket al., 2006, 2007), but the method is highly technical andnot widely available. Additionally, sampling methods withhigh variability greatly reduce the ability to detect changes inpopulations, especially when populations are low. In extremecases (e.g. Fig. 3b), the 95% confidence interval (CI) canextend below zero before the intervention, making detectionof change impossible. Scheduling of intervention experimentswhen Ae. aegypti populations are low (e.g. dry season in ourcase) should be avoided, as population measures are morevariable and less accurate (Morrison et al., 2004). Use ofcontained field trials (Benedict et al., 2008) involving releaseof mosquitoes of known number and age would eliminate erroras a result of immigration and multiple age cohorts.

Lure and kill traps such as SLOs do not increase themortality rate in mosquito populations nearly as effectivelyas conventional pesticides can. Rather, some proportion of thepopulation is attracted to the traps each day, and an incrementalreduction in the older individuals in a population can beachieved. This is an important distinction. Large interventionsusing fogging and indoor applications of residual insecticidesmay rapidly reduce mosquito populations across all ages (e.g.see Perich et al., 2001; Ritchie et al., 2004). This is mosquitocontrol in the broad sense, and as biting by female mosquitoesof all ages suddenly decreases, is interpreted by the public as asuccess. Incremental, delayed action control strategies such asSLOs do not kill young females, but do reduce the probabilitythat a mosquito will live long enough to transmit a pathogen.Younger mosquitoes will continue to bite, arousing skepticismin the public. Thus, while such delayed action control canbe effective disease control, it is not necessarily effectivemosquito control. It is also important to emphasize thorough,

residual larval control to prevent continuing reinfestation ofthe treatment area.

Sithiprasasna et al. (2003) attributed their inconsistentresults in 1999 to fungal growth recorded on the insecticideimpregnated ovistrips. In the current study, we observed somefungal growth on our bifenthrin-treated ovistrips after the wetseason LO and BLO interventions. This may have contributedto the slightly reduced lethality of field-exposed ovistrips(83%) compared with unexposed bifenthrin-treated ovistrips(100%). It should be noted that the reduced efficacy of thefield-exposed ovistrips may have also led to an over estimationin the number of mosquitoes killed by the SLOs, as calculationsof adult mortality were based on Williams et al. (2007a) whorecorded no reduction in lethality of bifenthrin-treated ovistripsafter 4 weeks of field exposure. The effectiveness of SLOs wasalso reduced because of traps drying out towards the end ofthe intervention period. This was particularly a problem inthe wet season and was probably as a result of the highermean daily evaporation rate during the 4-week period SLOswere deployed in the field in the wet (5.8 mm) compared withthe dry season (4.3 mm). Trap structural failure (drying outof SLOs) may have also contributed to the increase in adultcatches in SOs 5 weeks after the wet season LO intervention,although we suspect that both mosquito immigration into thetreated area (Koenraadt et al., 2007; Williams et al., 2007b),undetected breeding sites and the creation of new mosquito-producing containers may have also played a role.

Although BGSs were used in both the dry and wet seasoninterventions, our principal strategy to achieve mosquitocontrol was through the mass deployment of LOs. Williamset al. (2007b) found no significant decrease in Ae. aegyptipopulations within a cluster of houses where 16 BGSs wererun continuously for 15 days. Nevertheless, it is interestingto compare the success of the two trapping methods. Inboth SLO + BGS interventions, the mean number of femaleAe. aegypti caught per BGS was nearly twice as many asthe estimated number killed by each SLO. However, theBGS collects female Ae. aegypti in varying physiologicalstates, including a high percentage of nullipars (Williamset al., 2006a), whereas the SLO and BLO target gravid, olderfemales. In the current study 35% of mosquitoes caught by theBGSs in the reference enclave were nulliparous. Thus, for asingle premise during the wet season intervention, the BGSscaptured a mean of 4.8 gravid or parous Ae aegypti whereasfour SLOs killed an estimated 16.3 of these groups. We havedemonstrated that the use of L&K traps has the potentialto improve Ae. aegypti control in north Queensland. Indeed,the DART have used a SLO mass trapping scheme since2004, which helped to eliminate dengue transmission in fouroutbreaks in north Queensland (Montgomery et al., 2005; S.Ritchie, unpublished data), although these were not controlledstudies and must be interpreted cautiously. The success ofthe SLOs is partly because of the fact that ovipositing Ae.aegypti visit a number of sites to lay their entire egg batch(Colton et al., 2003; Reiter, 2007). This ‘skip oviposition’behaviour increases the likelihood that gravid females visita SLO, especially when the traps are deployed immediatelyafter the elimination or treatment of alternative breedingsites. Also, when used in conjunction with the elimination

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

Dengue control in Australia 315

of alternative breeding sites, they may reduce dispersal ofpotentially infective mosquitoes by providing a suitable, albeitlethal, oviposition site. However, before LO mass trapping isadopted for dengue control, assessment of its efficacy duringdengue outbreaks must be conducted.

Acknowledgements

We thank the Dengue Action Response Team and AvrilUnderwood who assisted with the interventions, and AnnaKoetz who assisted with the mosquito dissections. Karel vanHorck of Queensland Health generously produced GIS maps.We thank Michelle Rapley and Amanda Markovitz for aidingin mosquito monitoring, and the community of MachansBeach and Parramatta Park for access to their propertiesand cooperation. We acknowledge the support of BrianMontgomery, Ross Spark and Brad McCulloch of QueenslandHealth. This work was funded by an Australian National Healthand Medical Research Council grant 379615 to SAR and RCR.

References

Blanford, S., Chan, B.H.K., Jenkins, N.E., Sim, D., Turner, R.J.,Read, A.F. & Thomas, M.B. (2005) Fungal pathogen reduces poten-tial for malaria transmission. Science, 308, 1638–1641.

Benedict, M., D’Abbs, P., Dobson, S. et al. (2008) Guidance forcontained field trials of vector mosquitoes engineered to contain agene drive system: recommendations of a scientific working group.Vector-Borne and Zoonotic Diseases , 8, 127–166.

Brownstein, J.S., Hett, E. & O’Neill, S.L. (2003) The potential ofvirulent Wolbachia to modulate disease transmission by insects.Journal of Invertebrate Pathology , 84, 24–29.

Colton, Y.M., Chadee, D.D. & Severson, D.W. (2003) Natural skipoviposition of the mosquito Aedes aegypti indicated by co-dominantgenetic markers. Medical and Veterinary Entomology , 17, 195–204.

Cook, P.E., Hugo, L.E., Iturbe-Ormaetxe, I. et al. (2006) The use oftranscriptional profiles to predict adult mosquito age under fieldconditions. Proceedings of the National Academy of Science, 103,18060–18065.

Cook, P.E., Hugo, L.E., Iturbe-Ormaetxe, I. et al. (2007) Predictingthe age of mosquitoes using transcriptional profiles. Nature Proto-cols , 2, 2796–2806.

Farrar, J., Focks, D., Gubler, D. et al. (2007) Towards a global dengueresearch agenda. Tropical Medicine and International Health, 12,695–699.

Focks, D.A., Haile, D.G., Daniels, E. & Mount, G.A. (1993) Dynamiclife table model for Aedes aegypti (Diptera: Culicidae): analysisof the literature and model development. Journal of MedicalEntomology , 30, 1003–1017.

Focks, D.A., Daniels, E., Haile, D.G., Daniels, E. & Keesling, J.E.(1995) A simulation model of the epidemiology of urban denguefever: literature analysis, model validation and samples of simula-tions results. American Journal of Tropical Medicine and Hygiene,53, 489–506.

Hanna, J.N., Ritchie, S.A., Richards, A.R. et al. (2006) Multipleoutbreaks of dengue serotype 2 in north Queensland, 2003/04.Australia and New Zealand Journal of Public Health, 30, 220–225.

Harrington, L.C., Edman, J.D., Costero, A.C., Clark, G.G., Kit-tayapong, P. & Scott, T.W. (2001) Analysis of survival of youngand old Aedes aegypti (Diptera: Culicidae) from Puerto Rico andThailand. Journal of Medical Entomology , 38, 537–547.

Kittayapong, P., Yoksan, S., Chansang, U., Chansang, C. & Bhumi-ratana, A. (2008) Suppression of dengue transmission by applica-tion of integrated vector control strategies at sero-positive GIS-based foci. American Journal of Tropical Medicine and Hygiene,78, 70–76.

Kline, D.L. (2006) Semiochemicals, traps/targets and mass trappingtechnology for mosquito management. Journal of the AmericanMosquito Control Association, 23 (Suppl.), 241–251.

Koenraadt, C.J., Aldstadt, J., Kijchalao, U., Kengluecha, A., Jones,J.W. & Scott, T.W. (2007) Spatial and temporal patterns in therecovery of Aedes aegypti (Diptera: Culicidae) populations afterinsecticide treatment. Journal of Medical Entomology , 44, 65–71.

Krockel, U., Rose, A., Eiras, A.E. & Geier, M. (2006) New tools forsurveillance of adult yellow fever mosquitoes: comparison of trapcatches with human landing rates in an urban environment. Journalof the American Mosquito Control Association, 22, 229–238.

Kroeger, A. & Nathan, M.B. (2006) Dengue: setting the globalresearch agenda. The Lancet , 368, 2193–2195.

Montgomery, B.L., Ritchie, S.A., Hart, A.J., Long, S.A. & Walsh, I.D.(2005) Dengue intervention on Thursday Island (Torres Strait) 2004:a blueprint for the future? Arbovirus Research in Australia, 9,268–273.

Morrison, A.C., Astete, H., Chapilliquen, F. et al. (2004) Evaluationof a sampling methodology for rapid assessment of Aedes aegyptinfestation level in Iquitos, Peru. Journal of Medical Entomology ,41, 502–510.

Morrison, A.C., Zielinski-Gutierrez, E., Scott, T.W. & Rosenberg, R.(2008) Defining challenges and proposing solutions for con-trol of the virus vector Aedes aegypti. PLoS Medicine, 5,doi:10.1371/journal.pmed.0050068.

Muir, L.E. & Kay, B.H. (1998) Aedes aegypti survival and dispersalestimated by mark-release-recapture in northern Australia. AmericanJournal of Tropical Medicine and Hygiene, 58, 277–282.

Perich, M.J., Sherman, C., Burge, R., Gill, E., Quintana, M. & Wirtz,R.A. (2001) Evaluation of the efficacy of lambda-cyhalothrin appliedas ultra-low volume and thermal fog for emergency control of Aedesaegypti in Honduras. Journal of the American Mosquito ControlAssociation, 17, 221–224.

Perich, M.J., Kardec, A., Braga, I.A. et al. (2003) Field evaluationof a lethal ovitrap against dengue vectors in Brazil. Medical andVeterinary Entomology , 17, 205–210.

Reiter, P. (2007) Oviposition, dispersal, and survival in Aedes aegypti :implications for the efficacy of control strategies. Vector-Borne andZoonotic Diseases , 7, 261–273.

Ritchie, S.A. (2001) Effect of some animal feeds and oviposition sub-strates on Aedes oviposition in ovitraps in Cairns, Australia. Journalof the American Mosquito Control Association, 17, 206–208.

Ritchie, S.A. (2005) Evolution of dengue control strategies innorth Queensland, Australia. Arbovirus Research in Australia, 9,324–330.

Ritchie, S.A., Hart, A., Long, S.A., Montgomery, B.I., Walsh, I. &Foley, P. (2001) Update on dengue in north Queensland. 2001.Arbovirus Research in Australia, 8, 294–299.

Ritchie, S.A., Hanna, J., Hills, S., Piispanen, J., McBride, W., Pyke, A.& Spark, R. (2002) Dengue control in north Queensland, Australia:case recognition and selective indoor residual spraying. Dengue Bul-letin, 26, 7–13.

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316

316 L. P. Rapley et al.

Ritchie, S.A., Long, S., Hart, A., Webb, C.E. & Russell, R.C. (2003)An adulticidal sticky ovitrap for sampling container-breedingmosquitoes. Journal of the American Mosquito Control Association,19, 235–242.

Ritchie, S.A., Long, S., Smith, G., Pyke, A., & Knox, T. (2004)Entomological investigations in a focus of dengue transmission inCairns, Queensland, Australia using the sticky ovitrap. Journal ofMedical Entomology , 41, 1–4.

Ritchie, S.A., Long, S.A., McCaffrey, N., Key, C., Lonergan, G. &Williams, C.R. (2008) A biodegradable lethal ovitrap for control ofcontainer-breeding Aedes. Journal of the American Mosquito ControlAssociation, 24, 47–53.

Ritchie, S.A., Rapley, L.P., Williams, C.W. et al. (2009) A lethalovitrap-based mass trapping scheme for dengue control in Australia:I. Public acceptability and performance of lethal ovitraps. Medicaland Veterinary Entomology , 23, 295–302.

Service, M.W. (1993) Mosquito Ecology: Field Sampling Methods ,2nd edn. Elsevier Science Publishers Ltd, Essex, U.K.

Sithiprasasna, R., Mahapibul, P., Noigamol, C. et al. (2003) Fieldevaluation of a lethal ovitrap for the control of Aedes aegypti(Diptera : Culicidae) in Thailand. Journal of Medical Entomology ,40, 455–462.

Williams, C.R., Ritchie, S.A., Russell, R.C., Geier, M. & Long, S.A.(2006a) Field efficacy of the BG-Sentinel compared with the CDC

backpack aspirator and CO2- baited EVS trap for collection ofadult Aedes aegypti in Cairns, Queensland, Australia. Journal ofthe American Mosquito Control Association, 22, 296–300.

Williams, C.R., Long, S.A., Russell, R.C. & Ritchie, S.A. (2006b)Optimizing ovitrap use for Aedes aegypti in Cairns, Queens-land, Australia: effects of some abiotic factors on field effi-cacy. Journal of the American Mosquito Control Association, 22,635–640.

Williams, C.R., Ritchie, S.A., Long, S.A., Dennison, N. & Rus-sell, R.C. (2007a) Impact of a bifenthrin-treated lethal ovitrap onAedes aegypti oviposition and mortality in north Queensland, Aus-tralia. Journal of Medical Entomology , 44, 256–262.

Williams, C.R., Long, S.A., Webb, C.E., Bitzhenner, M., Geier, M.,Russell, R.C. & Ritchie, S.A. (2007b) Aedes aegypti populationsampling using BG-Sentinel traps in north Queensland Australia:statistical considerations for trap deployment and sampling strategy.Journal of Medical Entomology , 44, 345–350.

Zar, J.H. (1999) Biostatistical analysis . Prentice Hall, Upper SaddleRiver, NJ.

Zeichner, B.C. & Perich, M.J. (1999) Laboratory testing of a lethalovitrap for Aedes aegypti. Medical and Veterinary Entomology , 13,234–238.

Accepted 17 August 2009

© 2009 Queensland HealthJournal compilation © 2009 The Royal Entomological Society, Medical and Veterinary Entomology, 23, 303–316