Impact of a Bifenthrin-Treated Lethal Ovitrap on Aedes aegypti Oviposition and Mortality in North Queensland, Australia

Transcription

1 VECTOR CONTROL, PEST MANAGEMENT, RESISTANCE, REPELLENTS Impact of a Bifenthrin-Treated Lethal Ovitrap on Aedes aegypti Oviposition and Mortality in North Queensland, Australia CRAIG R. WILLIAMS, 1,2,3 SCOTT A. RITCHIE, 1,4 SHARRON A. LONG, 4 NIGEL DENNISON, 5 AND RICHARD C. RUSSELL 6 J. Med. Entomol. 44(2): 256Ð262 (2007) ABSTRACT Lethal ovitraps (LOs) containing an insecticide-treated ovistrip are used as a lureand-kill device for the container-breeding dengue vector, Aedes aegypti (L.). We aimed to afþrm that the pyrethroid bifenthrin could be used effectively in LOs against Ae. aegypti in north Queensland, Australia, by quantifying oviposition in and mortality caused by LOs. Small cage experiments in which individual gravid Ae. aegypti were given a choice of LOs and untreated ovitraps revealed that although LOs were less acceptable for oviposition, they provided an average 64.6% adjusted mortality. Although 92% of mosquitoes ovipositing in LOs died, 61.8% of mosquitoes that visited but did not oviposit in an LO also died, demonstrating that lethal contact occurred without egg laying. The bifenthrin content of strips ( 0.1 mg/cm 2 ; 7 mg/strip) did not decrease signiþcantly after 4 wk of Þeld exposure nor did the toxic effect of the LOs. Large cage trials with groups of 10 Ae. aegypti conþrmed that bifenthrintreated LOs provided consistent control (average adjusted mortality 79.7%). Four-week Þeld trials in north Queensland showed that although LOs were acceptable to ovipositing Ae. aegypti (mean time to Þrst egg 10.9 d; mean 47.3), insecticide-free ovitraps were oviposited in more readily (6.8 d, 199 ). The number of laid per mosquito in laboratory LOs allowed calculation of the number of Ae. aegypti killed in Þeld-deployed LOs; rapid estimates can be made by simply dividing the number of on the strip by Overall, the studies demonstrated that bifenthrin-treated LOs have potential for use as a lure-and-kill device against Ae. aegypti and that they should be effective in the Þeld for at least 4 wk. Given that untreated ovitraps were more acceptable for Ae. aegypti oviposition, the removal of alternative oviposition sites before deployment of LOs in the Þeld should maximize their effectiveness. KEY WORDS Aedes aegypti, bifenthrin, lethal ovitrap, dengue, oviposition The vector of dengue in north Queensland, Aedes aegypti (L.), uses a range of water-þlled containers around homes, including roof gutters (Montgomery and Ritchie 2002) and subterranean sites such as drain sumps, wells, and telecommunication pits (Russell et al. 1996, Kay et al. 2002) for larval habitat. Locally, vector control consists of larval source reduction and pyrethroid application inside houses within 100 m of a dengue case (Ritchie 2005). Although effective in controlling dengue outbreaks (Ritchie et al. 2002), this approach is labor-intensive and involves the exposure of the operators, the public, and nontarget organisms to insecticides. 1 School of Public Health and Tropical Medicine, James Cook University, P.O. Box 6811, Cairns, Queensland, 4870 Australia. 2 Sansom Research Institute, University of South Australia, GPO Box 2471, Adelaide, South Australia, 5000 Australia. 3 Corresponding author, craig.williams@unisa.edu.au. 4 Tropical Public Health Unit, Queensland Health, P.O. Box 1103, Cairns, Queensland, 4870 Australia. 5 Queensland Health ScientiÞc Services, Pesticide Laboratory, P.O. Box 594 ArcherÞeld, Queensland, 4108 Australia. 6 Department of Medical Entomology, University of Sydney and ICPMR, Westmead Hospital, Westmead, New South Wales, 2145 Australia. Lure-and-kill strategies involve using pheromones or kairomones to attract insects to a lethal device, such as an insecticide-treated target. This ensures a targeted delivery of a minimum amount of insecticide, enhancing control efþcacy while minimizing nontarget impact. Although lure and kill has been successfully used to control other dipterans such as tsetse ßies (Brightwell et al. 1991), its application toward mosquitoes is in its infancy. A lethal ovitrap (LO) comprising a water-þlled black plastic cup with an insecticide-treated oviposition substrate (ovistrip) was shown to be highly effective in the laboratory in killing gravid Ae. aegypti attempting to lay (Zeichner and Perich 1999), indicating a potential for operational use. Field trials in Brazil and Thailand revealed that LOs deployed en masse at a cluster of homes could reduce Ae. aegypti populations (Perich et al. 2003, Sithiprasasna et al. 2003). In Brazil, Ae. aegypti population reduction required almost 3 mo of LO deployment in one village, but 1 mo at another (Perich et al. 2003). In Thailand, signiþcant population reduction measures were observed in only four of six LO deployments (Sithiprasasna et al. 2003). The inconsistencies in /07/0256Ð0262$04.00/ Entomological Society of America

2 March 2007 WILLIAMS ET AL.: LETHAL OVITRAPS FOR Ae. aegypti 257 Fig. 1. Schematics of laboratory experimental setup (experiments 1 and 2). Gravid Ae. aegypti were given a choice of an LO or an untreated (insecticide-free) oviposition site. these promising results have been variously attributed to immigration of Ae. aegypti from adjacent areas, reduced lethality of the ovistrip after Þeld exposure and competition from insecticide-free alternative breeding sites (Perich et al. 2003, Sithiprasasna et al. 2003). LOs with a cloth strip treated with bifenthrin are used by vector control staff in north Queensland (Ritchie 2005). Bifenthrin was chosen because of its low mammalian toxicity, strong substrate binding properties (Lee et al. 2004), low irritancy against mosquitoes compared with other insecticides (Hougard et al. 2002), including deltamethrin, which was used in previous LO Þeld trials (Perich et al. 2003, Sithiprasasna et al. 2003), and the known susceptibility of local Ae. aegypti to bifenthrin-treated LOs in large cage pilot studies (S.A.R., unpublished data). In this study, we aimed to afþrm the use of the LOs as operational tools by investigating the impact of bifenthrin-treated LOs against north Queensland Ae. aegypti in the laboratory. SpeciÞcally, we aimed to quantify the acceptability of the LOs as oviposition sites for gravid Ae. aegypti compared with insecticidefree breeding sites, and the mortality caused by the LOs, and to determine whether the concentration of bifenthrin and the toxic effect of the LOs diminished during a 4-wk Þeld deployment. In practice, 4 wk is the maximum Þeld duration for LOs in north Queensland, after which time the strips can disintegrate or become covered in microbial growth. Additionally, using egglaying data from laboratory studies, we hoped to determine the mean number of laid by a gravid Ae. aegypti on a lethal ovistrip; such data would allow an estimate of the rate of mosquito visitation to LOs in the Þeld and potential number of mosquitoes killed. Overall, the results should allow an evaluation of the potential usefulness of LOs in Ae. aegypti control operations. Materials and Methods Treatment of Ovistrips for LOs. Ovistrips were made from red velour cloth ( common red ßannelette, Spotlight, South Melbourne, Victoria, Australia), because of the substrate reßectance preferences of Ae. aegypti (Muir et al. 1992), the use of red strips in previous LO studies (Zeichner and Perich 1999), the visibility of against the color, and its current use in north Queensland (Ritchie 2005). The cloth was cut into 7.5- by 2.5-cm strips for small cage experiments and by 5-cm strips for large cage experiments and Þeld trials. The strips were soaked for 10 min in a solution of 12.5 ml/liter (label rate) Bistar80 SC (FMC Chemicals, Pinkenba, Queensland, Australia), for a Þnal concentration of 0.1% active ingredient (AI) bifenthrin. The Bistar solution absorption rate of the cloth is liters/m 2, and a by 5-cm ovistrip should contain 0.1 mg/cm 2 or 7mgof bifenthrin (C.R.W., unpublished data). The cloth was dried ßat on nonabsorbent plastic sheeting, because vertical line drying leads to a loss of bifenthrin through runoff (S.A.R., unpublished data). The bifenthrintreated ovistrips were stored for 12Ð16 wk in a lightproof container at 25 C before use, unless otherwise indicated. Experiment 1: LO Acceptability and Mortality Study in Small Cages. This experiment was designed to determine the lethality of LOs toward individual Ae. aegypti when given a choice of insecticide-treated and untreated ovitraps. Gravid Ae. aegypti were obtained from a laboratory colony (F2-5) at James Cook University, derived from Þeld specimens collected in Cairns. Experiments were conducted in a laboratory with natural light, a mean minimum-maximum (minmax) temperature of 20.8Ð30.7 C, and mean min-max relative humidity of 63.8Ð86.2%. A pair of 225-ml black plastic cups were placed 30 cm apart in a plastic cage (52 cm in length by 33.5 cm in width by 25 cm in height) (Fig. 1). One cup contained a 7.5- by 2.5-cm bifenthrin-treated ovistrip fastened to the inside with a paperclip. The second cup contained an identical but untreated ovistrip. A 50% concentration hay infusion (Reiter et al. 1991) was added to both cups (100 ml). One gravid Ae. aegypti was introduced into the cage between 1000 and 1400 hours and retrieved 24 h later. Mosquito survival and laid were recorded for each mosquito, which was then dissected to count remaining, enabling calculation of total clutch size.

3 258 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 2 Bifenthrin-treated ovistrips in LOs were Þeld aged outdoors, in areas sheltered from direct sunlight and rain, for up to 4 wk. Experimental trials as described above were conducted with ovistrips aged 0, 1, 2, 3, and 4 wk. A trial with a pair of control cups with untreated ovistrips also was conducted. Field conditions for strip aging were mean minimum-maximum temperature 19.2Ð38.0 C and relative humidity of 36.0Ð 82.8%. In each trial, 16 replicate gravid Ae. aegypti were placed in individual plastic cages (as described above). To check for changes in the toxic effect of LOs after Þeld exposure, contingency table analysis and the chisquare statistic were used (Zar 1999). Mortality data from all Þve LO trials were tested for homogeneity and pooled to compare with the number dead/alive in the control trial (pairs of untreated cups) by using the 2 statistic with Yates correction (Zar 1999). The number of in untreated cups and LOs was compared for each trial using paired t-tests on log (x 1)-transformed data by using SPSS statistical software (release , SPSS Inc., Chicago, IL). Experiment 2: LO Acceptability and Mortality Study in Large Cages. To validate Þndings from small cages, large cage trials more representative of Þeld situations were conducted. We considered that in small cages, mortality may be overestimated, because mosquito ßight is limited and mosquitoes may choose to rest in LOs for lack of alternative refugia. Experiments were conducted in a house with natural light, a mean min-max temperature of 26.2Ð31.3 C and mean min-max relative humidity of 51.1Ð66.8%. A pair of 1.2 liters black golf divot buckets (as used for ovitraps in north Queensland) was placed 2 m apart in a 3- by 3-m geodesic dome screened tent (Fig. 1). One LO bucket contained a by 5-cm bifenthrintreated ovistrip, fastened to the inside with a paperclip. The second bucket contained an identical but untreated ovistrip. A 50% concentration hay infusion (Reiter et al. 1991) was added to each cup (500 ml). Ten gravid Ae. aegypti were introduced into the cage between 1000 and 1400 hours and retrieved 24 h later. Mosquito mortality and any laid were recorded. Females that did not lay were later dissected to conþrm that they were gravid. Five trials were conducted with ovistrips Þeld aged for up to 4 wk as in experiment 1. In each trial, Þve replicate experiments with 10 gravid Ae. aegypti each were conducted. A sixth trial (also with Þve replicates) made up a pair of control buckets with untreated ovistrips. To check for changes in toxic effect of LOs after Þeld exposure, mortality data from the Þve replicates in each trial were compared using the KruskalÐWallis test and a nonparametric posthoc procedure (Zar 1999). The number of in untreated cups and LOs was compared for each trial by using a paired Wilcoxon test with SPSS software. Nonparametric methods were used because of the small number of replicates in each trial. Experiment 3: LO Acceptability Field Studies. To quantify the Þeld acceptability of LOs compared with insecticide-free oviposition sites, individual LOs (1.2 liters bucket, by 5-cm bifenthrin-treated ovistrip) were deployed at houses in suburban Cairns. Three trials were conducted to assess Þeld acceptability and the effect of ovistrip storage before use. Trial 1 involved the deployment of individual LOs with ovistrips that were treated with bifenthrin the previous day (i.e., the strip had not been stored) at 10 houses in May 2004 (late wet season). Identical buckets with untreated ovistrips were deployed at the same houses, at least 10 m away from the LO. All buckets were set in shaded areas, sheltered from direct sunlight, rain, and wind. Traps were inspected daily for the presence of on ovistrips. The number of was estimated using a torch, 10 hand lens and hand counter. After 14 d, the traps were retrieved and the counted on each ovistrip in the laboratory. Untreated ovistrips were immersed in infusion for egg hatching, but from LOs were scraped from ovistrips with a needle into hay infusion for hatching to prevent contact between insecticide and recently hatched larvae. Larvae were reared to instar IV for identiþcation. Trial 2 (also in May 2004) was identical to trial 1, except that LO ovistrips were stored for 12 wk before use and were set at 12 houses. Based on the results of trials 1 and 2, trial 3 (November 2005, late dry season) involved the deployment of LOs containing ovistrips stored for 12 wk before use at 12 houses. Identical buckets with untreated ovistrips were placed at the same houses, at least 10 m away from the LO. Identical methods to trials 1 and 2 were used, except that the deployment was for 28 d, and the traps were topped up weekly with tap water because evaporation rates were very high during this period. To compare the oviposition acceptability of ovitraps with treated and untreated ovistrips, survival analysis using the KaplanÐMeier method (Pagano and Gauvreau 2000) was used to calculate the cumulative probability that an treated or untreated ovistrip would remain egg free, with death for each trap being the day it Þrst became egg positive. The log-rank test (Pagano and Gauvreau 2000) was used to compare the cumulative survival (egg-free) probabilities of treated and untreated ovistrips. The mean number of laid on treated ovistrips and untreated ovistrips was compared using a t-test on log (x 1)-transformed data. Because of ant predation on in trial 3 (Þve of 12 untreated ovistrips and four of 12 treated ovistrips), the maximum recorded number of on a particular strip was used in this analysis. Whereas such predation can make the comparison of mean egg numbers imperfect, gross differences in oviposition should still be valid and detectable. All analyses were performed in SPSS (release , SPSS Inc.). Experiment 4: Effect of Field Exposure on Lethal Ovistrip Toxicity and Bifenthrin Content. The persistence of the toxic effect of the bifenthrin ovistrips was evaluated by exposing female Ae. aegypti (James Cook University colony) to treated ovistrips aged for 0, 1, 2, 3, or 4 wk in Þeld conditions (see experiment 1). An untreated strip served as a control. Groups of 10 mosquitoes were placed in an inverted 70-ml spec-

4 March 2007 WILLIAMS ET AL.: LETHAL OVITRAPS FOR Ae. aegypti 259 Table 1. Small cage LO choice trials (experiment 1) No. on ovistrips t-test / mosquito a No. laid No. that laid in U only:lo only:both Mortality (%) Control U t /16 n.a. 3/16 U U U P (18.8) U LO U t /16 6:0:4 11/16 0-wk age LO P (68.8) U LO U t /16 7:4:1 13/16 1-wk age LO P (81.3) U LO U t /16 8:2:1 13/16 2-wk age LO P (81.3) U LO U t /16 6:4:2 11/16 3-wk age LO P (68.8) U LO U t /16 6:5:2 9/16 4-wk age LO P (56.3) U LO U t /80 33:15:10 57/80 Overall LO 1.49 b 0.3 P (71.3) Individual mosquitoes (n 16 for each treatment) were given a choice of a cup with an untreated (U) ovistrip and a bifenthrin-treated LO. s are presented with SE. n.a., not applicable. a Eggs per mosquito was sum of laid and remaining in mosquito as determined by dissection. b Value of 1.49 per mosquito on bifenthrin-treated strips is used for calculations in Table 5. imen jar atop an ovistrip for 4 min, the contact duration required for 100% lethality for bifenthrin-treated LOs (S.A.R., unpublished data). Mosquitoes were then anesthetized with CO 2 placed in a resting cup with 15% sucrose on a cotton pledglet, and mortality was assessed after 24 h for three replicates of each Þeld age treatment. Bifenthrin content in the same ovistrips was determined by extracting three times in hexane in an ultrasonic bath (20 min) before gas chromatography with electron capture detection (GC-ECD) (e.g., Jiang et al. 2004). Results were calculated from a calibration curve derived from four standards bracketing the expected concentration. Bifenthrin content in ovistrips of different Þeld ages was analyzed using a KruskalÐWallis test (Zar 1999). Results Small Cage Oviposition Choice Trials (Experiment 1). In small cage trials, the presence of a miniature LO was associated with 56.3Ð81.3% 24 h mortality for gravid Ae. aegypti (Table 1). There was no signiþcant variation in mortality as ovistrips became more Þeld aged ( , df 4, P 0.49). Thus, mortality data from all Þve LO trials were pooled and compared with the mortality in the control trial (a pair of untreated Table 2. Number of mosquitoes that laid and did not lay in LO and their subsequent mortality in small cage trials (experiment 1) Ovistrips Þeld age (wk) Overall Total Laid in LO Dead Alive Did not lay in LO Dead Alive Total mosquitoes ovistrips). Mortality was signiþcantly less in the control trial ( , df 1, P ). Although LOs were acceptable for oviposition, more were usually laid in untreated cups, particularly for ovistrips Þeld aged 2 wk or less (Table 1). Individual female mosquitoes laid in untreated cups, LOs or both within a 24-h period. Overall, 25 mosquitoes laid in LOs of which 23 (92%) died (Table 2). However, 34 of the 55 mosquitoes that did not lay in LOs also died. Oviposition in an LO was almost always lethal for a gravid Ae. aegypti, but lethal contact occurred without oviposition. Overall, Ae. aegypti mortality for all ovistrips (combined Þeld ages) was 71.3% (Table 1), but adjustment for control mortality (Abbott 1925) reduced it to 64.6%. Large Cage Oviposition Choice Trials (Experiment 2). In large cage trials, 24-h mortality for gravid Ae. aegypti ranged from 78 to 92% in tents with LOs (Table 3). Nonparametric multiple comparisons showed that mortality in the control trial was significantly less than in the LO trials (KruskalÐWallis: , P 0.019) and that there was no signiþcant variation in mortality as ovistrips became more Þeld aged (Table 3). Although LOs were acceptable for oviposition, more were usually laid in untreated cups. However, this trend was not statistically significant (Table 3). Overall, Ae. aegypti mortality for all ovistrips (combined Þeld ages) was 85% (Table 3), equating to 79.7% after adjusting for control mortality (Abbott 1925). Field Acceptability of LOs (Experiment 3). The Þeld studies relate only to Ae. aegypti, because almost all of the larvae hatched from LOs were this species (33/35, 49/51, and 97/97 in trials 1, 2, and 3, respectively). In 14-d trials, almost all LOs became eggpositive (Table 4), demonstrating their acceptability for Ae. aegypti oviposition. Furthermore, there was no signiþcant difference in cumulative probabilities of oviposition in LOs and untreated ovitraps over 14 d (Table 4), although total oviposition was signiþcantly

5 260 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 2 Table 3. Large cage lethal ovitrap choice trials (experiment 2) SE No. laid Wilcoxon Mortality (%) Control U U U Z 0.67, P (26)a U U LO 0-wk age U Z 1.47, P (88)b LO U LO 1-wk age U Z 0.37, P (88)b LO U LO 2-wk age U Z 1.75, P (78)ab LO U LO 3-wk age U Z 0.67, P (92)b LO U LO 4-wk age U Z 0.00, P (78)b LO U LO overall U Z 1.70, P (85) LO Groups of 10 mosquitoes (n 5 groups for each treatment) were given a choice of a bucket with an untreated ovistrip (U) and a bifenthrin-treated strip (LO). The same letters following mean mortality results indicate no signiþcant difference. greater in the controls. LOs with ovistrips stored for 12 wk (trial 2) seemed to have greater acceptability for Ae. aegypti oviposition than LOs with freshly treated ovistrips, with a slightly shorter mean time to Þrst oviposition (Table 4). Therefore, it was decided that future experiments would only be performed with ovistrips stored at least 12 wk before use to ensure maximum LO acceptability. In trial 3 (28 d), LOs again proved acceptable for Ae. aegypti oviposition, albeit signiþcantly less so than untreated ovitraps, because they took longer to become egg positive (Table 4). Approximately 4Ð20 times as many were laid in untreated ovitraps than in LOs (Table 4). Effect of Field Exposure on Persistence of Bifenthrin and Toxic Effect (Experiment 4). There was no detectable change in impact of bifenthrintreated ovistrips after Þeld exposure, with 100% mortality recorded in all treatments, compared with a control mortality of 7.5%. Similarly, bifenthrin content did not change signiþcantly after Þeld exposure, with mean content decreasing only slightly from mg/cm 2 at 0-wk Þeld age to mg/ cm 2 at 4-wk Þeld age (KruskalÐWallis , P 0.384). The bifenthrin concentration on ovistrips was very similar to the amount predicted based on the absorption rate of the red velour cloth (0.1 mg/cm 2 ; 7 mg/strip). Estimating the Number of Ae. aegypti Killed by Field-Deployed LOs. The mean number of laid per mosquito in an LO was calculated from laboratory experiment 1 (Table 5). The mortality attributable to the LOs was calculated from overall mortality minus the mortality in control trials (without LOs). These values (Table 5) allowed an estimation of the number of Ae. aegypti that an LO potentially killed in the Þeld from the number of in the LO. Rapid estimates without conþdence intervals also can be made by simply dividing the number of on the strip (e)by Discussion Laboratory experiments revealed that bifenthrintreated LOs were lethal to north Queensland Ae. aegypti. The level of mortality in laboratory experiments (up to 92% in 24 h) was comparable with that of deltamethrin-treated LOs in previous laboratory stud- Table 4. Oviposition of Ae. aegypti in LO and insecticide-free ovistrips in suburban Cairns (experiment 3) Trial 1: LO ovistrips freshly treated Egg positive d to Þrst egg 14-d trials 28-d trial Trial 2: LO ovistrips stored 12 wk Egg positive d to Þrst egg Trial 3: LO ovistrips stored 12 wk Egg positive d to Þrst egg LO 8/ / / Untreated ovistrips 9/ / / Statistical result Log rank 1.52 P t P Log rank 1.12 P t P Log rank 4.19, P t 3.19, P Ovistrips were either freshly treated with bifenthrin or were stored for 12 wk before use. s are presented SE.

6 March 2007 WILLIAMS ET AL.: LETHAL OVITRAPS FOR Ae. aegypti 261 Table 5. on LO ovistrips and LO-attributable mortality proportions in laboratory experiments Mortality Ae. aegypti killed by LO estimated from e a No. killed (N) SE (SD/ n) CI 95 LO-attributable (bðc) Control (b) Overall (a) Eggs (e)/ mosq. SD 1.49 b (e/1.49) i.e. e/2.84 ((e/2.97) 0.525)/ 80 (n 1.96 SE) Experiment 1 (small cage, n 80) (100/1.49) ((100/2.97) 0.25)/ ( ) 31.36; 39.11) Example for 100 (e) on an LO Þeld deployed for 1 mo These calculations enable the number of mosquitoes killed by a Þeld-deployed LO to be estimated from the no. of present on the ovistrip (e). An example for 100 on an ovistrip is presented. a Number of on ovistrip. CI, conþdence interval. b Value from Table 1. ies (up to 98% in 48 h; Zeichner and Perich 1999). Furthermore, three separate laboratory experiments revealed that bifenthrin-treated ovistrips suffered no loss in toxicity and analysis by using GC-ECD con- Þrmed no loss of bifenthrin content in LO strips over 4 wk. Analysis using GC/ECD also conþrmed no loss of bifenthrin content in LO strips over 4 wk. These results demonstrate that bifenthrin may be used effectively for ovistrips in LOs in north Queensland and that bifenthrin-treated LOs can be expected to have similar impact for up to 4 wk of Þeld use, as the deltamethrin-treated LOs previously used (Zeichner and Perich 1999, Perich et al. 2003, Sithiprasasna et al. 2003). Furthermore, monitoring of Þeld-deployed LOs for up to 4 wk has shown that larval development does not occur in LOs, even when hatched are present, indicating toxicity of bifenthrin to early instars (S.A.R., unpublished data). In laboratory experiments, the presence of in the LOs, even when an alternative untreated oviposition site was available, was evidence of mosquito visitation to the LO. However, the LOs proved lethal even in the absence of oviposition. These studies were not designed to record all visits to LOs, so we presumed that when mosquitoes died without ovipositing in an LO (34/80 mosquitoes in experiment 1), they had made some lethal contact with the bifenthrintreated ovistrip (although a small number would have died of natural causes, as indicated by mortality in the control trials: 18.8% mortality in 24 h). Overall, there was comparable mortality attributable to LOs in both small and large cage trials, indicating that the extreme proximity of Ae. aegypti and LOs in small cages (experiment 1) did not lead to increased mortality, and validating the use of small cages for the collection of data from individual gravid mosquitoes. Laboratory and Þeld studies showed less oviposition in LOs compared with untreated ovitraps, and Þeld studies demonstrated that LOs were less acceptable as oviposition sites, not only in terms of the number of laid but also in the cumulative probability of becoming egg positive over time. These Þndings suggest that LOs function better in the absence of competing untreated oviposition sites as hypothesized by Perich et al. (2003). We are now able to estimate the number of Ae. aegypti killed by an LO from the number of on the ovistrip, and such estimates could be used to assess the efþcacy of vector control operations which incorporate LOs. For example, using data from a 2004 trial with 76 LOs in Cairns, with a mean of 19.9 per strip (S.A.R., unpublished data), along with the calculations prescribed in Table 5, we estimate that a mean of 7.0 females (CI Ð7.8) visited the LO strip and died. For the entire intervention, a total of 533 (CI Ð593) female Ae aegypti were estimated to have been potentially killed. If Ae. aegypti population size estimates are made before control operations from pupal survey data, simulations, or both with life table models (e.g., Focks et al. 1993), then the direct contribution of LOs to Ae. aegypti population reduction could be assessed. However, our

7 262 JOURNAL OF MEDICAL ENTOMOLOGY Vol. 44, no. 2 estimates are based on laboratory data and the assumptions that wild and laboratory Ae. aegypti populations have a similar susceptibility to bifenthrin, that all on the ovistrip were Ae. aegypti and that the mode of toxicity in the Þeld was the same as that in the laboratory. Successful dengue control in north Queensland has been achieved using LOs as part of the vector control program, along with source reduction and interior pyrethroid application (Ritchie 2005). The Þndings reported here afþrm the use of the bifenthrin-treated LOs, but they demonstrate that competition with alternative breeding sites is likely to reduce Þeld efþcacy. Controlled Þeld studies will be required to reveal the speciþc impact of mass deployed LOs on extensive Þeld populations of Ae. aegypti in north Queensland. Acknowledgments Brian Montgomery and Paul Zborowski assisted in the development of LOs for use in Cairns. We thank the people of Cairns for access to properties. This work was funded by Australian National Health and Medical Research Council grant to S.A.R. and R.C.R. References Cited Abbott, W. S A method of computing the effectiveness of an insecticide. J. Econ. Entomol. 18: 265Ð267. Brightwell, R., R. D. Dransfield, and C. Kyorku Development of a low cost tsetse trap and odour baits for Glossina pallidipes and G. longipennis in Kenya. Med. Vet. Entomol. 5: 153Ð164. Focks, D. A., D. G. Haile, E. Daniels, and G. A. Mount Dynamic life table model for Aedes aegypti (Diptera: Culicidae): simulation results and validation. J. Med. Entomol. 30: 1018Ð1028. Hougard, J.-M., S. Duchon, M. Zaim, and P. Guillet Bifenthrin: a useful pyrethroid insecticide for treatment of mosquito nets. J. Med. Entomol. 39: 526Ð533. Jiang, W., R. T. Kon, R. A. Othoudt, R. A. Leavitt, S. Kumar, L. D. Geissel, and E. A. Gomaa Method development, validation, and analysis of bifenthrin residues in fresh and dry cilantro foliages and cilantro seeds using GC-ECD. Bull. Environ. Contam. Toxicol. 73: 9Ð16. Kay, B. H., P. A. Ryan, S. A. Lyons, P. N. Foley, N. Pandeya, and D. Purdie Winter intervention against Aedes aegypti (Diptera: Culicidae) larvae in subterranean habitats slows surface recolonization in summer. J. Med. Entomol. 39: 356Ð361. Lee, S., J. Gan, J. S. Kim, J. N. Kabashima, and D. E. Crowley Microbial transformation of pyrethroid insecticides in aqueous and sediment phases. Environ. Toxicol. Chem. 23: 1Ð6. Montgomery, B. L., and S. A. Ritchie Roof gutters: a key container for Aedes aegypti and Ochlerotatus notoscriptus (Diptera: Culicidae) in Australia. Am. J. Trop. Med. Hyg. 67: 244Ð246. Muir, L. E., B. H. Kay, and M. J. Thorne Aedes aegypti (Diptera: Culicidae) vision: response to stimuli from the optical environment. J. Med. Entomol. 29: 445Ð450. Pagano, M., and K. Gauvreau Principles of biostatistics, 2nd ed. Duxbury Press, Belmont, CA. Perich, M. J., A. Kardec, I. A. Braga, I. F. Portal, R. Burge, B. C. Zeichner, W. A. Brogdon, and R. A. Wirtz Field evaluation of a lethal ovitrap against dengue vectors in Brazil. Med. Vet. Entomol. 17: 205Ð210. Reiter, P., M. A. Amador, and N. Colon Enhancement of the CDC ovitrap with hay infusions for daily monitoring of Aedes aegypti populations. J. Am. Mosq. Control Assoc. 7: 52Ð54. Ritchie, S. A Evolution of dengue control strategies in north Queensland, Australia. Arbovirus Res. Aust. 9: 324Ð 330. Ritchie, S. A., J. Hanna, S. Hills, J. Piispanen, W. McBride, A. Pyke, and R. Spark Dengue control in north Queensland, Australia: case recognition and selective indoor residual spraying. WHO Dengue Bull. 26: 7Ð13. Russell, B. M., P. N. Foley, and B. H. Kay The importance of surface versus subterranean breeding during winter in north Queensland. Arbovirus Res. Aust. 7: 240Ð 242. Sithiprasasna, R., P. Mahapibul, C. Noigamol, M. J. Perich, B. C. Zeichner, B. Burge, S.L.W. Norris, J. W. Jones, S. S. Schleich, and R. E. Coleman Field evaluation of a lethal ovitrap for the control of Aedes aegypti (Diptera: Culicidae) in Thailand. J. Med. Entomol. 40: 455Ð462. Zar, J. H Biostatistical analysis, 4th ed. Prentice Hall, Inglewood Cliffs, NJ. Zeichner, B. C., and M. J. Perich Laboratory testing of a lethal ovitrap for Aedes aegypti. Med. Vet. Entomol. 13: 234Ð238. Received 4 May 2006; accepted 11 August 2006.