Journal of the American Mosquito Control Association, 22(1): 15-21, 2006Copyright (C) 2006 by the American Mosquito Control Association, Inc.

DOES 1-OCTEN-3-OL ENHANCE TRAP COLLECTIONS OF JAPANESEENCEPHALITIS VIRUS MOSQUITO VECTORS IN

NORTHERN AUSTRALIA?

ANDREW E VAN DEN HURK, BRIAN L. MONTGOMERY, PAUL ZBOROWSKI, NIGEL W. BEEBE, R. D.COOPER, AND SCOTT A. RITCHIE

ABSTRACT. The responses of Japanese encephalitis virus (JEV) mosquito vectors to 1-octen-3-ol (octenol)and CO2 were evaluated using Centers for Disease Control (CDC) light traps at 3 sites in northern Australia.There was no significant difference between the number of Culex sitiens subgroup mosquitoes or Cx. geliduscollected in CDC light traps baited with either CO2 alone or CO2 + octenol on Badu Island. At both mainlandlocations, using octenol in combination with CO2 significantly increased collections of Cx. sitiens subgroupmosquitoes. Collections of nontarget species, such as Ochlerotatus spp., Anopheles spp., and Verrallina spp.were also significantly increased with the addition of octenol. At all 3 locations, reducing collections of nontargetmosquitoes by not using octenol increased the proportion of Culex spp. collected, thus potentially reducing thetime and resources required to sort and process collections for JEV detection. Our results also indicate that trialsinto the efficacy of using octenol as an attractant should be carried out in each area prior to the implementationof a mosquito-based arbovirus surveillance system.

KEY WORDS Arbovirus, mosquito, surveillance, octenol, Australia

INTRODUCTION

Since 1995, incursions of Japanese encephalitisvirus (JEV) have occurred almost annually into theTorres Strait of northern Queensland (Mackenzie etal. 2002). In 1998, the virus was reported for thefirst time on the Australian mainland (Hanna et al.1999). Virus isolation studies and vector compe-tence experiments have demonstrated that membersof the Culex sitiens subgroup, particularly Culexannulirostris Skuse, are the major vectors of JEVin northern Australia (Ritchie et al. 1997, van denHurk et al. 2003). This species is also a major vec-

tor of endemic arboviruses responsible for humandisease in Australia, including Murray Valley en-cephalitis (MVE), Kunjin (KUN), and Ross River(RR) viruses (Mackenzie et al. 1994, Russell 1995).

Despite the reappearance of JEV almost everywet season in northern Australia, there is no evi-dence to suggest that it has become enzootic. Con-sequently, to detect incursions of JEV into northernAustralia and to determine if it had become enzo-otic on mainland Australia, a sentinel pig programwas established in the Torres Strait and on CapeYork Peninsula (Shield et al. 1996, Hanna et al.1999). However, the deployment and maintenance

Department of Microbiology and Parasitology, Schoolof Molecular and Microbial Sciences, The University ofQueensland, St. Lucia, Queensland 4072, Australia.

Tropical Public Health Unit, Queensland Health, 19Aplin Street, Cairns 4870, Australia.

Institute for Biotechnology of Infectious Disease, Uni-versity of Technology, Sydney, Westbourne Street, GoreHill, New South Wales 2065, Australia.

Australian Army Malaria Institute, Gallipoli Barracks,Enoggera, Queensland 4052, Australia.

School of Public Health and Tropical Medicine, JamesCook University, PO Box 6811, Cairns, Queensland 4870,Australia.

of sentinel pigs in remote areas is expensive, and,as pigs are the primary amplifying hosts of JEV,they may potentially serve as a source of the virus.An alternative to the sentinel pig program is a

system based on the detection of viral RNA in mos-quitoes collected by new-generation propane or so-

lar powered traps. In these traps, CO2 is providedby the combustion of propane, such as in the Mos-quito Magnet (MM; American Biophysics, NorthKingston, RI) or via compressed gas cylinders inthe newly developed Banks Updraft Trap (BUT; D.Banks and S. A. Ritchie, unpublished data). Asthese traps produce their own electricity, they areideal for remote locations, where accessing powermains and/or recharging or replacement of batteriesis not possible. Detection of viral RNA in deadmosquitoes is undertaken using either a seminestedreverse transcriptase-polymerase chain reaction(RT-PCR) (Johansen et al. 2002) or real-timeTaqMan RT-PCR (Pyke et al. 2004). In a field trialof this system, Ritchie et al. (2003) demonstratedthat JEV RNA could be detected in a single in-fected mosquito within pools of up to 5,000 mos-

quitoes, stored for 14 days under wet-season con-

ditions of high temperature and humidity. It was

concluded that pool sizeg of 500 mosquitoes wouldrepresent the ideal compromise between detectabil-ity and labor (Ritchie et al. 2003).

In 2002, 2 MM Pro*'s were deployed to detectincursions of JEV into the Torres Strait. To date,JEV RNA has been detected in 7 pools collectedfrom Badu Island and 3 pools from St. Pauls com-

munity on Moa Island (A.T. Pyke and S.A. Ritchie,Queensland Health, unpublished data). However,the processing of these samples for virus detectionwas very labor intensive, as there were over100,000 mosquitoes in each collection, with Culexspp. comprising less than 1% of the catch. The col-lections were dominated by Ochlerotatus spp., pri-

15

16 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 22, No.

marily Oc. kochi (D6nitz) and Oc. culiciformis(Theobald), which are relatively inefficient vectorsof JEV (van den Hurk et al. 2003).The efficacy of a mosquito-based sentinel system

could be improved if the number of Ochlerotatusspp. within the trap collections were reduced, with-out compromising collections of Culex spp. Oneway this could be achieved would be to remove 1-octen-3-ol (octenol), which has been used in com-bination with CO2 to increase MM and Centers forDisease Control (CDC) light-trap collections ofmosquitoes during studies of arbovirus ecology innorthern Australia since 1995 (Ritchie et al. 1997,Harley et al. 2000, "van den Hurk et al. 2001). Oc-tenol acts synergistically with CO2 to increase col-lections of some mosquito species (Kline 1994). Instudies undertaken in southeast Queensland, Aus-tralia, the addition of octenol to CO_-baited lighttraps significantly increased collections of Oc. vi-gilax (Skuse) and Verrallina funerea (Theobald),but did not affect collections of Cx. annulirostrisand Cx. sitiens Wied. (Kemme et al. 1993, Van Es-sen et al. 1994). We report on field trials conductedon Badu Island and on mainland Australia to ex-amine the responses of individual mosquito taxa tothe presence of octenol as an additional attractantwith CO2-baited CDC light traps. Importantly, wedetermined whether removing octenol would affectthe proportion of Culex spp. in trap collections byreducing the numbers of nontarget Ochlerotatusspp.

MATERIALS AND METHODS

Study sites: Field studies were undertaken at 3locations: Badu Island in the Torres Strait (1007'S,14209'E) between February 13 and February 19,2003; Mount Bundey Army Training Area in thenorthwest Northern Territory (1252'S, 13135'E)between March 13 and March 23, 2003; and theNorthern Peninsula Area of Cape York Peninsula(NPA) (1053'S, 14223'E) between March 22 andMarch 25, 2004. Badu Island is a 101 km graniticisland with vegetation characterized by open Cor-ymbia tessellaris (F. Muell.) K.D. Hill and L.A.S.Johnson and Corymbia polycarpa (F. Muell.) K.D.Hill and L.A.S. Johnson woodland, Melaleuca spp.,Pandanus spp., sedgeplains, and patches of man-

groves near the coast (Fox et al. 2001). The studysite at Mount Bundey, located ca. 120 km east ofDarwin, consists of low woodland dominated by C.polycarpa, Eucalyptus tetrodonta F. Muell., andMelaleuca viridiflora Sol. ex Gaertn. (Kinhill En-gineers Pty Ltd. 1991). Dominant shrubs includePandanus spiralis R.Br., Grevillea spp., and Ter-minalia spp., while the main grasses are Eriachnespp., Sorghum spp., and Themeda triandra Forssk.The NPA has vegetation dominated by E. tetrodon-ta and Corymbia spp. woodland, a conspicuous Liv-istonia spp. subcanopy, and Schizachyrium spp. tus-

sock grasses; tracts of semideciduous vine forestalso occur (Fox et al. 2001).

Study design: The standard CDC miniature lighttrap (Model 512; John W. Hock, Gainesville, FL)was used for all collections. Paired comparisonswere undertaken between CO2-baited CDC lighttraps with and without octenol as an added attrac-tant. The CO2 bait consisted of kg of dry ice,which was either wrapped in 4 sheets of newspaperon Badu Island and the NPA or housed in an in-sulated metal bucket at Mount Bundey. Althoughwe did not measure the amount of dry ice at thestart and finish of each trap night, previous studieshave calculated CO2 release rates of approximately400 ml/min for kg of dry ice (Van Essen et al.1994, Johansen et al. 2003). The light source was

a 6-V incandescent bulb, which had been blackenedwith a permanent marker pen to reduce the intensityof light emitted. Octenol was released at a rate of4.5 mg/h from 5-ml vials attached adjacent to theopening of the trap cylinder (Van Essen et al.1994). For the 3 study sites, traps were set over atotal of 44 paired trap nights, with 20 paired trap-night collections obtained from 4 locations on BaduIsland, and 12 paired collections from 3 locationsat Mount Bundey and on the NPA, respectively. OnBadu Island, results for 2 trap pairings were notobtained due to trap theft, while paired-trap col-lection was not obtained on the NPA due to trapmalfunction. Individual locations were separated by>-500 m on Badu Island and by >2 km at MountBundey and on the NPA. At each location, pairedtraps were set ca. 50 m apart, run for 14 h (1700-0700 h), with attractant treatments at each locationbeing rotated every night. Mosquitoes were killedon dry ice and either immediately identified to spe-cies or stored at -70C to await identification.

Identification of cryptic species in the Anophelesfarauti complex and Cx. sitiens subgroup: All iso-morphic members of the Anopeles farauti complexcollected from Mount Bundey and the NPA wereidentified using a polymerase chain reaction-restric-tion fragment length polymorphism (PCR-RFLP)procedure (Beebe and Saul 1995). As a previousstudy had identified only An. farauti sensu strictoLaveran occurring on Badu Island (Foley et al.1991), specimens collected from that location were

not identified using the PCR-RFLRIn the Cx. sitiens subgroup, Cx. annulirostris, Cx.

palpalis Taylor, and Cx. sitiens share overlappingmorphology, which can make morphological iden-tification of these species difficult. Consequently,morphological identification of a subsample of Cx.sitiens subgroup mosquitoes from Badu Island andMount Bundey was verified using a PCR-RFLP as-

say developed by Beebe et al. (2002). No attemptwas made to morphologically separate the speciesin the Cx. sitiens subgroup collected from the NPA.Instead, a subsample of specimens from each of theNPA trapping sites was identified using PCR-RFLPanalysis.

MARCH 2006 NORTHERN AUSTRALIAN MOSQUITOES AND OCTENOL 17

Table 1. Total mosquitoes collected from Badu Island, Mount Bundey, and the Northern Peninsula Area (NPA).

Badu Island Mount Bundey NPA

CO_ + light CO2 + light CO2 + lightSpecies CO2 + light + octenol CO2 + light + octenol CO2 + light + octenol

Aedes lineatopennis 0 0 3 34 0 0Ae. scutellaris 2 0 0 287 221Aedeomyia catasticta 0 0 0 0 8 6Anopheles amictus 0 0 3 26 0 0An. annulipes 0 0 16 74 10 0An. bancroftii 0 0 109 807 53 142An. farauti s.s. 13 17 0 5 7An. hilli 0 0 3 8 0An. hinesorum 0 0 0 51 29An. meraukensis 0 0 95 158 0 0An. novaguinensis 0 0 41 278 350 1,088An. powelli 0 0 0 0 70 110An. torresiensis 0 0 24 198 75 95Coquilletidia xanthogaster 0 0 364 362 0 0Cq. crassipes 0 0 0 0 8 7Culex bitaeniorhynchus 23 4 5 2 0 0Cx. cubiculi 0 0 0 0 23 6Cx. gelidus 133 218 5 14 3Cx. hilli 23 6 0 0 0 0Cx. pullus 832 224 197 32Cx. quinquefasciatus 19 5 0 0 0 16Cx. sitiens subgroup 3,283 2,902 3,504 14,914 4,894 7,386Cx. starckeae 0 0 0 0 0Ficalbia spp. 0 0 0 0 0Mansonia uniformis 0 0 8 105 0 0Ochlerotatus culiciformis 2,280 15,889 0 0 101 448Oc. elchoensis 0 0 3 183 0 0Oc. imprimens 0 0 0 0 3 0Oc. kochi 6,315 29,761 0 4 1,381 9,302Oc. littlechildi 103 229 0 0 17 42Oc. normanensis 0 0 307 5,593 0 0Oc. notoscriptus 24 55 2 115 230Oc. palmarum 0 0 0 0 0Oc. purpureus 0 12 12 30 0 0Oc. rupestris 0 0 0 0 42 52Oc. stoneorum 162 638 5 28 0 0Oc. tremulus 2 0 6 88 208Oc. vigilax 314 695 2 115 18 43Tripteroides atripes 0 0 0 0 38 78Tp. magnesianus 0 0 0 0 14 32Tripteroides spp. 7 0 0 0 0 0Uranotaenia spp. 0 0 0 0 11 15Verrallina carmenti 167 458 0 0 31 110Ve. funerea 20 0 0 0 26 268Ve. reesi 0 0 13 66 0 0Total 12,892 50,891 5,355 23,228 7,926 19,976

Specimens only identified to genus level.

Analysis: To determine the effects of octenol onindividual taxa, mean trap catches were log (n +1) transformed and compared using a paired t-test.Only species for which total numbers exceeded 100mosquitoes per location were statistically analyzed.For each trap collection, totals for the individualmembers of the Cx. sitiens subgroup were com-bined before analysis. The proportion of Culex spp.collected in CO2 + octenol-baited CDC light trapsand in traps where CO2 was used alone was arcsinetransformed and compared using a paired t-test. All

statistical tests were undertaken using SPSS forWindows Version 12 (SPSS 2003).

RESULTS

Mosquito collections: A total of 120,268 mos-quitoes, comprising 47 species in 11 genera, were

collected from the 3 study areas, with 63,783,28,583, and 27,902 obtained from Badu Island,Mount Bundey, and the NPA, respectively (Table1). At all 3 locations, members of the Cx. sitiens

18 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION VOL. 22, No.

Table 2. Mean SD number of mosquitoes collected from Badu Island, February 13-19, 2003 (n 18 trapnights); Mount Bundey, March 13-23, 2003 (n 12 trap nights); and the Northern Peninsula Area (NPA), March

22-25, 2004 (n 11 trap nights) using CDC light traps baited with either CO2 or CO2 + octenol'2.

Location Species CO2 + light CO2 + light + octenol t-value P

Badu Island

Mount Bundey

NPA

Culex gelidus 7.4 13.4 12.1 30.2 0.40 0.698Cx. sitiens subgroup 182.4 101.4 161.2 132.9 1.24 0.231Ochlerotatus culiciformis 126.7 +_ 208.0 882.7 867.4 5.97 <0.001Oc. kochi 350.8 301.0 1,653.4 1,248.7 5.54 <0.001Oc. littlechildi 5.7 12.7 12.7 20.8 1.53 0.144Oc. stoneorurn 9.0 _+ 18.4 35.4 57.4 2.83 0.012Oc. vigilax 17.4 16.0 38.6 34.0 1.95 0.067Ve. carmenti 9.3 8.5 25.4 +_ 24.4 2.11 0.050Anopheles bancroftii 9.1 _+ 8.4 68.1 38.8 5.33 <0.001An. merahkensis 7.9 7.6 13.2 6.0 2.57 0.026An. novaguinensis 3.4 _+ 8.3 23.2 36.8 3.98 0.002An. torresiensis 2.0 2.4 16.5 +_ 14.0 5.35 <0.001Coquilletida xanthogaster 30.3 27.5 30.2 _+ 28.6 0.03 0.979Cx. sitiens subgroup 292.0 +_ 242.1 1,242.2 792.8 5.65 <0.001Cx. pullus 69.3 _+ 112.7 18.7 35.9 2.69 0.021Oc. elchoensis 0.3 0.5 15.3 +_ 27.3 3.91 0.002Oc. normanensis 25.6 30.7 466.1 +_ 413.9 7.11 <0.001Aedes scutellaris 26.1 57.8 20.1 _+ 19.2 0.95 0.366An. bancroftii 4.8 _+ 6.5 12.9 19.7 2.37 0.039An. novaguinensis 37.3 59.6 111.4 216.6 0.98 0.352An. torresiensis 6.8 15.2. 8.6 15.1 0.98 0.352An. powelli 6.4 6.7 10.0 _+ 9.9 0.79 0.451Cx. pullus 17.9 23.9 2.9 4.5 4.42 0.001Cx. sitiens subgroup 444.9 431.6 671.5 638.8 2.33 0.042Oc. culiciformis 9.2 10.7 40.7 36.3 3.81 0.003Oc. kochi 125.5 _+ 137.7 845.6 +_ 445.2 6.32 <0.001Oc. notoscriptus 10.5 7.2 20.9 _+ 18.8 1.90 0.087Oc. trernulus 8.0 8.3 18.9 +_ 21.0 3.88 0.003Tripteroides atripes 3.5 4.9 7.1 5.3 2.33 0.042Verrallina carmenti 2.8 _ 3.2 10.0 8.6 4.17 0.002Ve. funerea 2.4 __+ 5.4 24.4 47.2 4.56 0.001

Log (x + 1) trap numbers compared using paired t-test.Degrees of freedom 17, 11, and 10 for Badu Island, Mount Bundey, and the NPA, respectively.

subgroup accounted for >80.0% of Culex spp. col-lected using either trap bait. Molecular verificationof a subsample of mosquitoes morphologicallyidentified as Cx. annulirostris (n 135) collectedfrom Badu Island revealed that 94.1% of the sam-ple was Cx. annulirostris, with the remainder beingCx. palpalis (4.4%) and Cx. sitiens (1.5%). Simi-larly, molecular verification of morphologicallyidentified Cx. annulirostris (n 175) collectedfrom Mount Bundey revealed that 92.6% of thesample was Cx. annulirostris and 7.4% Cx. palpal-is. Out of 41 morphologically identified Cx. pal-palis tested by PCR-RFLR 87.8% were identifiedas Cx. palpalis, with the remaining 12.2% beingCx. annulirostris. Finally, on the NPA, 100% of Cx.sitiens subgroup mosquitoes were identified as be-ing Cx. annulirostris (n 109).On Badu Island, trap collections were dominated

by Oc. kochi and Oc. culiciformis, with these 2 spe-cies accounting for 66.7% and 89.7% of mosqui-toes collected in CO2 and CO2 + octenol-baitedCDC light traps, respectively. Of the non-Culexspp. collected from the NPA, Oc. kochi was thedominant species collected, comprising 49.1% and

74.2% of mosquitoes collected in CO and CO2 +octenol-baited CDC light traps, respectively. AtMount Bundey, Cx. sitiens subgroup mosquitoesdominated collections, accounting for ca. 65% ofmosquitoes collected, irrespective of trap bait. InCOz + octenol-baited traps, Oc. normanensis (Tay-lor) was the next dominant species collected, com-prising 24.1% of trap collections. In traps baitedwith CO2 alone, Cx. sitiens subgroup mosquitoesaccounted for 65.4% of trap collections. Of the oth-er species, Cx. pullus Theobald, Coquillettidia xan-thogaster (Edwards), and Oc. normanensis com-prised 15.5%, 6.7%, and 5.7% of the trapcollections, respectively.

Treatment effects on individual taxa: On BaduIsland, there was no significant difference (P >0.05) between CO2 or CO + octenol-baited CDClight trap collections of Cx. sitiens subgroup mos-quitoes or Cx. gelidus Theobald (Table 2). Collec-tions of Cx. sitiens subgroup mosquitoes wereslightly reduced in CDC light traps supplementedwith a combination of CO2 and octenol. In contrast,at Mount Bundey and the NPA, significantly more(P < 0.05) Cx. sitiens subgroup mosquitoes were

MARCH 2006 NORTHERN AUSTRALIAN MOSQUITOES AND OCTENOL 19

2O30O

1OOO0

Oclenol N ocenol Octenol N octenol

Badu Is NPA

Colletion location

Oclnol Nooclno

Mtl.lndey

Fig. 1. Total number of Culex spp. (D) and nontarget species (.) collected in CDC light traps baited with COalone or CO2 + octenol at 3 locations in northern Australia.

collected in CDC light traps baited with CO + oc-tenol than those with CO2 alone. Indeed; there was

a 325.6% and 50.9% increase in collections of Cx.sitiens subgroup mosquitoes when octenol was usedat Mount Bundey and on the NPA, respectively.With the exception of Oc. littlechildi Taylor and

Oc. vigilax on Badu Island and Oc. notoscriptus(Skuse) on the NPA, collections of Ochlerotatusspp. and Verrallina spp. from all locations weresignificantly increased (P < 0.05) when octenolwas used with CO2-baited CDC light traps. AtMount Bundey, collections of all Anopheles spp.were also significantly increased (P < 0.05) whenoctenol was used. Similarly, on the NPA, all col-lections of Anopheles spp. were increased with theaddition of octenol, although the effect was signif-icant only with Anopheles bancroftii Giles. In con-

trast, collections of Cx. pullus were significantly in-creased (P < 0.05) when octenol was not used as

an additional attractant.The reduction in the numbers of nontarget spe-

cies resulted in a significant increase in the propor-tion of Culex spp. collected in CDC light traps bait-ed with CO2 alone compared with CO2 + octenolon both Badu Island (t 5.50; df 17" P < 0.001)and the NPA (t 7.74; df 10" P < 0.001) (Fig.1). At Mount Bundey, not using octenol with CO2-baited CDC light traps did increase the proportionof Culex spp. collected (t 1.91" df 11" P0.082), although this increase was not significant.

DISCUSSION

The primary vectors of JEV are Culex mosqui-toes, of which members of the Cx. sitiens subgroupare the most important vectors in the Australasianregion (Ritchie et al. 1997, van den Hurk et al.2003). The addition of octenol to CO2-baited CDC

light traps elicited different responses in the 3 pop-ulations of Cx. sitiens subgroup mosquitoes thatwere included in the study. The results of the PCR-RFLP analysis suggest that Cx. annulirostris com-prised the majority of mosquitoes belonging to theCx. sitiens subgroup collected from all of the studylocations and that the observed responses to CO2and octenol were not due to interspecific variationbetween the members of this complex. On BaduIsland, collections of Cx. sitiens subgroup mosqui-toes were not significantly affected by the additionof octenoi, reflecting the results obtained in south-east Queensland by Kemme et al. (1993) and VanEssen et al. (1994). Indeed, slightly fewer mosqui-toes of this group were collected in the traps baitedwith a combination of CO2 and octenol comparedwith CO alone. However, at Mount Bundey andthe NPA, collections of Cx. sitiens subgroup mos-quitoes were significantly increased (P < 0.05)when octenol was used in combination with COz.As octenol is a volatile compound present in thebreath of ruminants (Kline 1994), the increasedcol-lections of the mainland Cx. sitiens subgroup mos-quitoes may reflect the response of these mosqui-toes to the presence of cattle and buffaloes at thesecollection locations. These large vertebrates are ab-sent from Badu Island.

Surveillance for JEV in northern Australia gen-erally consists of detecting specific antibodies insentinel pigs (Shield et al. 1996) or viral antigen inmosquito pools by either virus isolation (Ritchie etal. 1997) or PCR-based assays (Ritchie et al. 2003).When planning a remote mosquito-based surveil-lance system, it is critical to establish if the num-bers of nontarget species in the collections shouldbe reduced to minimize processing time, but at thepotential expense of collecting fewer Culex spp.vectors. For instance, on Badu Island, not using oc-

20 JOURNAL OF THE AMERICAN MOSQUITO CONTROL ASSOCIATION Vow. 22, No.

tenol significantly increased the proportion of Culexspp., without significantly affecting the numbers ofCx. sitiens subgroup mosquitoes collected. Basedon the results of the current study, octenol was re-moved from the MM Pro deployed on Badu Islandin 2003, resulting in an increase in the proportionof Culex spp. from -< 0.01 to 0.2 (S. A. Ritchie, A.T. Pyke, R Zborowski and A. E van den Hurk, un-published data). From a total of 53 pools, compris-ing 3,867 Culex spp., JEV was detected in 2 poolsof Cx. gelidus and pool of Cx. sitiens subgroupmosquitoes at a combined maximum likelihood es-timation of infection rate (MLE-IR) of 0.79 per1,000 mosquitoes. Infportantly, the time taken tosort and pool the Culex spp. from these collectionswas reduced from ca. 15 h to 3 h. Decreasing theturnaround time between sample receipt and outputof results would increase the efficacy of the mos-quito-based sentinel system for detecting JEV onBadu Island.The situation on the NPA and at Mount Bundey

appears to be different. Although the proportion ofCulex spp. may have increased when octenol was

not used, the number of actual Cx. sitiens subgroupmosquitoes collected was significantly reduced. Inlocations with a high field-infection rate, such asBadu Island, this may not be an issue, but some-where such as the NPA, significantly more mos-quitoes would need to be processed to detect thevirus. On Badu Island, during recognized periodsof JEV activity, MIRs for Cx. annulirostris havebeen as high as 2.97 per 1,000 mosquitoes (Ritchieet al. 1997). However, on the NPA, only isolateof JEV has been obtained (MLE-IR of 0.04 per1,000 mosquitoes), despite processing 28,806 Cx.sitiens subgroup mosquitoes during 2 recognizedincursions of JEV, in 1998 and 2004 (van den Hurket al. 2001, 2005). Consequently, on mainland Aus-tralia, we recommend that octenol be used in con-junction with CO2 to maximize collections of Cx.sitiens subgroup mosquitoes, which, in turn, shouldincrease the chances of detecting JEV. Our field tri-als have also demonstrated that, before any remotemosquito-based sentinel system is deployed, it isimportant to undertake preliminary studies to de-termine if there is intraspecific variation in the re-

sponse of the targeted species to the combinationof CO2 and octenol.

ACKNOWLEDGMENTS

The authors wish to thank Badu Island and allNorthern Peninsula councils for providing us withthe opportunity to conduct our field studies in theircommunities. We especially wish to thank HarrySeriat, Ronald Williams, Sean Lawler, and DarrenWaterson for their assistance with trapping and Pe-ter Whelan for identification of some of the MountBundey mosquito specimens. We are grateful toPublic Health Virology, Queensland Health Scien-tific Services, for access to their facilities and re-

sources. This study was funded by the AustralianHealth Minister's Advisory Council (AHMAC) Pri-ority Driven Research Program.

REFERENCES

Beebe NW, Saul A. 1995. Discrimination of all membersof the Anopheles punctulatus complex by polymerasechain reaction-restriction fragment length polymor-phism analysis. Am J Trop Med Hyg 53:478-481.

Beebe NW, van den Hurk AE Chapman H, Frances SP,Williams CR, Cooper RD. 2002. Development andevaluation of a species diagnostic polymerase chain re-action-restriction fragment-length polymorphism pro-cedure for cryptic members of the Culex sitiens (Dip-tera: Culicidae) subgroup in Australia and the southwestPacific. J Med Entomol 39:362-369.

Foley DH, Bryan JH, Booth D, Freebairn CG. 1991.Anopheles farauti Laveran sensu lato and An. hilliWoodhill and Lee (Diptera: Culicidae) in Torres Strait,Australia. J Aust Entomol Soc 30:307-312.

Fox ID, Neldner VJ, Wilson GW, Bannink PJ. 2001. Thevegetation of the Australian tropical savannas. Bris-bane, Australia: Environmental Protection Agency.

Hanna JN, Ritchie SA, Phillips DA, Lee JM, Hills SL,van den Hurk AE Pyke AT, Johansen CA, MackenzieJS. 1999. Japanese encephalitis in north Queensland,Australia, 1998. Med J Aust 170:533-536.

Harley D, Ritchie S, Phillips D, van den Hurk A. 2000.Mosquito isolates of Ross River virus from Cairns,Queensland, Australia. Am J Trop Med Hyg 62:561-565.

Johansen CA, Hall RA, van den Hurk AE Ritchie SA,Mackenzie JS. 2002. Detection and stability of Japaneseencephalitis virus RNA and virus viability in dead in-fected mosquitoes under different storage conditions.Am J Trop Med Hyg 67:656-661.

Johansen CA, Montgomery BL, Mackenzie JS, RitchieSA. 2003. Efficacies of the MosquitoMagnet and coun-terflo geometry traps in North Queensland, Australia.J Am Mosq Control Assoc 19:265-270.

Kemme JA, Van Essen PH, Ritchie SA, Kay BH. 1993.Response of mosquitoes to carbon dioxide and 1-octen-3-ol in southeast Queensland, Australia. J Am MosqControl Assoc 9:431-435.

Kinhill Engineers Pty Ltd. 1991. Mount Bundey TrainingArea--draft environmental impact statement. Adelaide,Australia: Department of Defence. 212 p.

Kline DL. 1994. Olfactory attractants for mosquito sur-veillance and control: 1-octen-3-ol. J Am Mosq ControlAssoc 10:280-287.

Mackenzie JS, Johansen CA, Ritchie SA, van den HurkAE Hall RA. 2002. The emergence and spread of Jap-anese encephalitis virus in Australasia. Curr Top Mi-crobiol lmmunol 267:49-73.

Mackenzie JS, Lindsay MD, Coelen RJ, Broom AK, HallRA, Smith DW. 1994. Arboviruses causing human dis-ease in the Australasian zoogeographic region. Arch Vi-rol 136:447-467.

Pyke AT, Smith IL, van den Hurk AE Northill JA, ChuanTE Westacott AJ, Smith GA. 2004. Detection of Aus-tralasian flavivirus encephalitic viruses using rapid fluo-rogenic TaqMan RT-PCR assays. J Virol Methods 117:161-167.

Ritchie SA, Phillips D, Broom A, Mackenzie J, PoidingerM, van den Hurk A. 1997. Isolation of Japanese en-

MARCH 2006 NORTHERN AUSTRALIAN MOSQUITOES AND OCTENOL 21

cephalitis virus from Culex annuiirostris in Australia.Am J Trop Med Hyg 56:80-84.

Ritchie SA, Pyke AT, Smith GA, Northill JA, Hall RA,van den Hurk AE Johansen CA, Montgomery BL, Mac-kenzie JS. 2003. Field evaluation of a sentinel mosquito(Diptera: Culicidae) trap system to detect Japanese en-

cephalitis in remote Australia. J Med Entomol 40:249-252.

Russell RC. 1995. Arboviruses and their vectors in Aus-tralia: an update on the ecology and epidemiology ofsome mosquito-borne arboviruses. Rev Med Vet Ento-mol 83:141-158.

Shield J, Hanna J, Phillips D. 1996. Reappearance of theJapanese encephalitis virus in the Torres Strait, 1996.Comm Dis Intell 20:191.

SPSS. 2003. SPSS 12 for Windows. Chicago, IL: SPSSInc.

van den Hurk AE Johansen CA, Zborowski P, PhillipsDA, Pyke AT, Mackenzie JS, Ritchie SA. 2001. Flavi-viruses isolated from mosquitoes collected during thefirst outbreak of Japanese encephalitis virus on CapeYork Peninsula, Australia. Am J Trop Med Hyg 64:125-130.

van den Hurk AE Nisbet DJ, Hall RA, Kay BH, Macken-zie JS, Ritchie SA. 2003. Vector competence of Aus-tralian mosquitoes (Diptera: Culicidae) for Japanese en-cephalitis virus. J Med Entornol 40:82-90.

van den Hurk AE Ritchie SA, Smith GA, Montgomery,BL, Mackenzie JS. 2005. A Japanese encephalitis od-yssey: entomological studies in northern Australia,2002-2004. Arbovirus Res Aust 9:370-377.

Van Essen PHA, Kemme JA, Ritchie SA, Kay BH. 1994.Differential responses of Aedes and Culex mosquitoesto octenol or light in combination with carbon dioxidein Queensland, Australia. Med Vet Entomol 8:63-67.