Vol. 32, no. 2 Journal of Vector Ecology 285 Comparison of light,, and resting boxes for West Nile virus surveillance Gregory M. Williams 1 and Jack B. Gingrich 2 1 Hudson Regional Health Commission, 595 County Ave., Secaucus, NJ 07094, U.S.A. 2 Department of Entomology and Wildlife Ecology, University of Delaware, Newark, DE 19716, U.S.A. Received 27 March 2007; Accepted 13 June 2007 ABSTRACT: West Nile virus surveillance was conducted at five sites in New Castle County, DE, and one site in Salem County, NJ, from June through September, 2004, using dry ice-baited Centers for Disease Control miniature light, infusion-baited, and resting boxes. All trap types were simultaneously placed at each site every two weeks and run overnight. Collected mosquitoes were identified to species, pooled, and analyzed for virus using a real-time reverse transcriptase polymerase chain reaction test. In total, 47,972 mosquitoes in 29 species or species groups were analyzed. Light collected 60,201 mosquitoes in 28 species or species groups. Gravid collected 3,195 mosquitoes in 19 species or species groups. Resting boxes collected 99 mosquitoes in nine species or species groups. In total, 1,0 mosquito pools were tested for WNV resulting in ten positive pools. All positive pools consisted of Culex pipiens, Culex restuans, or Culex salinarius. Seven positive pools were from and three were from light despite testing almost 14 times as many pools from light. The overall infection rate from was nearly 33 times greater than the infection rate from light, 2.29 and 0.07 infected mosquitoes per 1,000, respectively. The results demonstrate the advantage of using for West Nile virus surveillance over light or resting boxes. Journal of Vector Ecology 32 (2): 285-291. 2007. Keyword Index: Trap comparison, West Nile virus, mosquito infection rate, Delaware. INTRODUCTION Several types of have been employed to monitor West Nile virus (WNV) infections in mosquitoes. While all of these differ somewhat in design and function, they can be grouped into three main categories: that collect questing mosquitoes, that collect ovipositing mosquitoes, and that collect resting mosquitoes. Many WNV surveillance programs use carbon dioxide-baited light and infusion-baited (Andreadis et al. 2001, Hribar et al. 2003, Mans et al. 2004, Nasci et al. 2001b). Each category of trap has distinct advantages and disadvantages for arbovirus surveillance. Baited light collect a greater number and variety of mosquitoes than other trap types (Gingrich and Casillas 2004, Hribar et al. 2003). Unfortunately, the majority of these mosquitoes are unfed nullipars (Barr et al. 1986, Reisen and Pfuntner 1987) and therefore could only be infected with WNV through transovarial transmission. Gravid may collect relatively few individuals from a few species, but they effectively sample Cx. pipiens and Cx. restuans (Gingrich and Casillas 2004, Reiter 1983), both implicated as potential vectors of WNV (Nasci et al. 2001a, Savage et al. 1999, Kilpatrick et al. 2005, 2006). An added advantage is that most of the individuals collected in are females (Gingrich and Casillas 2004, Reiter et al. 1986), making them more suitable for arbovirus surveillance. Resting boxes preferentially sample Culiseta melanura and are relatively poor at sampling other species (Crans 1995). Many of the mosquitoes collected are males, but a large proportion of females are often engorged (Edman et al. 1968). While the overall utility of resting boxes is limited for WNV surveillance, blood-fed mosquitoes are a valuable resource for analysis, and resting boxes may be the only efficient way of sampling Cs. melanura. This study compared the variety, numbers, and infection rates of mosquitoes collected with carbon dioxide-baited light,, and resting boxes to determine which trap type is the most effective for WNV surveillance at our research sites. MATERIALS AND METHODS Mosquitoes were collected from five sites located in New Castle County, DE, and one in Salem County, NJ, every two weeks from June through September, 2004, for a total of 48 trap nights. Collection sites were at the Augustine Wildlife Area, Landfill, Fort Delaware State Park, Killcohook National Wildlife Refuge, the City of New Castle, and the Ommelanden Hunter Safety Training Center. The 2,630-acre Augustine Wildlife Area is located along the Delaware River, just north of Port Penn, DE. The research site was in an area of marsh covered mostly with Phragmites (Poales: Poaceae) and dense brush. Cherry Island Landfill is a 513-acre landfill positioned along the
286 Journal of Vector Ecology December 2007 Delaware River in Wilmington, DE. This surveillance site was located in a row of trees along the river. Fort Delaware State Park is part of Pea Patch Island in the center of the Delaware River, just to the northeast of Delaware City, DE. The surveillance site was located along the edge of a heavily wooded section toward the northern end of the island. The Killcohook National Wildlife Refuge was located along the eastern shore of the Delaware River in Pennsville, NJ. The research site was along the edge of a Phragmites marsh next to Finn s Point National Cemetery. The New Castle site was situated among a Phagmites marsh along the Delaware River in New Castle, DE, within an industrial complex. The Ommelanden Hunter Safety Training Center was along the Delaware River in New Castle, DE, within a stand of mixed hardwood trees. A Phragmites marsh was located just north of the site. Collections were made with dry ice-baited Centers for Disease Control (CDC) miniature light (Model 512, John W. Hock Co., Gainesville, FL), (Model 2800, BioQuip Products, Inc., Rancho Dominguez, CA), and resting boxes. Gravid were baited with a grass and hay infusion (Scott et al. 2001). Resting boxes were five-sided plywood cubes, roughly 0.03 m 3, and painted flat black. Mosquitoes were removed from the resting boxes with a battery powered aspirator (Hausherr s Machine Works, Toms River, NJ). The different trap types were placed approximately 10 m apart at each site in the early evening and retrieved the following morning. Light were hung from the trees approximately 1.5 m off the ground. Gravid were placed on the ground in partially shaded areas. Resting boxes were placed on the ground in vegetation and left on-site all season to allow mosquitoes to acclimate to the box. Mosquitoes were able to enter and leave the resting box at will and only mosquitoes present on a collection day were aspirated. All sites were trapped within two consecutive evenings to minimize the effects of variable weather conditions on the collections. Mosquitoes were identified through the use of morphological keys (Darsie and Ward 1981). Culex pipiens and Cx. restuans are morphologically similar, and even slight damage to the specimens made species identification difficult. Therefore, the Cx. pipiens/restuans group consisted of individuals that were either Cx. pipiens or Cx. restuans but could not be distinguished. The Culex spp. group consisted of individuals that were known to be in the Culex genus but could not be identified to species. The individuals in this group were most likely Cx. pipiens, Cx. restuans, or Cx. salinarius. The Anopheles crucians complex consisted of An. crucians Wiedemann and Anopheles bradleyi King, which are also morphologically similar. Because of their limited role, if any, in WNV transmission, the two species were not differentiated. For large collections, subsets of approximately 2,000 mosquitoes were randomly selected, identified, and pooled in groups of or less for testing. The subset and the remaining mosquitoes were weighed in order to estimate the total number of each species in a collection. Pools were placed into 2.0 ml polypropylene microcentrifuge tubes (USA Scientific Inc., Ocala, FL) along with a steel BB (Daisy Outdoor Products., Rogers, AR) for automated tissue maceration immediately after identification and stored at 80 C until extraction. For WNV detection in mosquito pools, real-time reverse transcriptase polymerase chain reaction (RT-PCR) tests were performed at the U.S. Army Center for Health Promotion and Preventive Medicine, Fort Meade, MD following the exact protocol described in Williams and Gingrich (2007). Maximum likelihood estimations (MLE) of infection rates were calculated using PooledInfRate 2.0 software (Biggerstaff 2003). Average collections from each trap type were compared with the Tukey multiple comparison test using the GLM procedure in SAS 8.02 software (SAS 1999). For all analyses, a P value 0.05 was considered significant. For comparing means, standard error was used as the measure of variation. RESULTS In total, 63,495 mosquitoes were collected in 29 species or species groups (Table 1). Of those, 47,972 (75.6% of the total) were analyzed for the presence of WNV RNA. All untested mosquitoes were from large light trap collections. Culex salinarius was the most abundant species collected, followed by Aedes vexans and mixed Cx. pipiens/restuans. Together, these three species or groups comprised 82.2% of the total collection. Light collected 60,201 mosquitoes (94.8% of the total collected) in 28 species or species groups. Light collected more total mosquitoes in every species except Aedes j. japonicus, Cx. pipiens, and Cx. restuans, where collected more total mosquitoes. Gravid collected 3,195 mosquitoes (5.0% of the total collected) in 19 species or species groups. Although collected more individual Cx. pipiens and Cx. restuans, light collected more individuals classed as Cx. pipiens/restuans mixed species. Overall, light collected more total Cx. pipiens and Cx. restuans combined, but the specimens were of poorer quality than those from making identification to species-level difficult. Resting boxes collected 99 mosquitoes (0.2% of the total collected) in nine species or species groups. Resting boxes were originally employed to collect Culiseta melanura, a species not usually caught by other trap types (Edman et al. 1968). However, the areas where the study was conducted were not preferred Cs. melanura habitats (Lake et al. 1962), and none were collected. Gravid collected significantly more Ae. j. japonicus, Cx. pipiens, and Cx. restuans per night than light or resting boxes (Table 2). For Aedes albopictus, light and collected significantly more mosquitoes per night than resting boxes. Light collected more Aedes cantator, Ae. vexans, Aedes spp., An. crucians complex, Anopheles punctipennis, Anopheles quadrimaculatus, Anopheles walkeri, Coquillettidia perturbans, Culex erraticus, Cx. salinarius, Culex spp., Psorophora columbiae, Ps. howardii, and Uranotaenia sapphirina per night than either or resting boxes. Light also collected
Vol. 32, no. 2 Journal of Vector Ecology 287 Table 1. Mosquitoes collected and tested from dry ice-baited CDC miniature light,, and resting boxes, 2004. Species Total no. collected a Total no. tested No. collected from light No. tested from light No. collected and tested from No. collected and tested from resting boxes Aedes albopictus (Skuse) 44 40 25 21 19 0 Ae. canadensis (Theobald) 1 1 1 1 0 0 Ae. cantator (Coquillett) 765 749 763 747 2 0 Ae. j. japonicus (Theobald) 12 12 0 0 12 0 Ae. sollicitans (Walker) 11 8 8 5 3 0 Ae. taeniorhynchus (Wiedemann) 2 2 2 2 0 0 Ae. triseriatus (Say) 25 21 18 14 7 0 Ae. trivittatus (Coquillett) 19 19 19 19 0 0 Ae. vexans (Meigen) 15,459 11,083 15,446 11,070 12 1 Aedes spp. 194 166 194 166 0 0 Anopheles crucians complex 984 854 969 839 15 0 An. punctipennis (Say) 291 238 285 232 4 2 An. quadrimaculatus Say 2,129 1,766 1,932 1,569 121 76 An. walkeri Theobald 181 160 177 156 4 0 Anopheles spp. 221 221 221 221 0 0 Coquillettidia perturbans (Walker) 3,238 3,016 3,229 3,007 7 2 Culex erraticus (Dyar&Knab) 38 38 36 36 1 1 Cx. pipiens Linnaeus 887 887 170 170 712 5 Cx. pipiens/restuans 6,223 4,355 4,790 2,922 1,428 5 Cx. restuans Theobald 851 851 106 106 743 2 Cx. salinarius Coquillett 30,514 22,273 30,414 22,173 95 5 Cx. territans Walker 2 2 1 1 1 0 Culex spp. 1,065 915 1,065 915 0 0 Psorophora ciliata (Fabricius) 26 25 26 25 0 0 Ps. columbiae (Dyar & Knab) 105 73 100 68 5 0 Ps. ferox (Von Humbolt) 42 35 42 35 0 0 Ps. howardii Coquillett 2 2 2 2 0 0 Uranotaenia sapphirina (Osten Saken) 110 106 106 102 4 0 Damaged 54 54 54 54 0 0 Totals 63,495 47,972 60,201 44,678 3,195 99 a Total no. collected is an estimate based on the weight of the collection compared to the weight of the total no. tested.
288 Journal of Vector Ecology December 2007 Table 2. Mean catch of mosquitoes per trap night for three trap types ± standard error, 2004. a Species Light trap Gravid trap Resting box F-value P-value Aedes albopictus 0.5 ± 0.17 a 0.4 ± 0.13 a 0 b 5.31 0.006 Ae. canadensis 0.02 ± 0.02 a 0 a 0 a 0.98 3785 Ae. cantator 16.2 ± 5.27 a 0.04 ± 0.03 b 0 b 9.24 0.0002 Ae. j. japonicus 0 a 0.3±0.11 b 0 a 6.45 0.0021 Ae. sollicitans 0.1 ± 0.07 a 0.1 ± 0.07 a 0 a 1.73 0.1803 Ae. taeniorhynchus 0.04 ± 0.03 a 0 a 0 a 2 0.1393 Ae. triseriatus 0.4 ± 0.18 a 0.2 ± 0.05 ab 0 b 3 0.0531 Ae. trivittatus 0.4 ± 0.26 a 0 a 0 a 2.41 0.0935 Ae. vexans 328.6 ± 102.11 a 0.3 ± 0.13 b 0.02 ± 0.02 b 10.13 < 0.0001 Aedes spp. 4.1 ± 1.68 a 0 b 0 b 5.94 0.0034 Anopheles crucians complex 20.6 ± 5.35 a 0.3 ± 0.31 b 0 b 14.26 < 0.0001 An. punctipennis 6.1 ± 1.38 a 0.1 ± 0.05 b 0.04 ± 0.03 b 18.38 < 0.0001 An. quadrimaculatus 41.1 ± 11.20 a 2.7 ± 1.21 b 1.6 ± 0.38 b 11.71 < 0.0001 An. walkeri 3.8 ± 0.86 a 0.1 ± 0.06 b o b 18.44 < 0.0001 Anopheles spp. 4.7 ± 2.98 a 0 a 0 a 2.43 0.0917 Coquillettidia perturbans 68.7 ± 22.16 a 0.1 ± 0.06 b 0.04 ± 0.04 b 9.38 0.0002 Culex erraticus 0.8 ± 0.25 a 0.02 ± 0.02 b 0.02 ± 0.02 b 8.47 0.0003 Cx. pipiens 3.6 ± 1.72 a 15.8 ± 2.9 b 0.1 ± 0.09 a 18.36 < 0.0001 Cx. pipiens/restuans 101.9 ± 73.51 a 31.7 ± 9.59 a 0.1 ± 0.09 a 1.46 <0.2357 Cx. restuans 2.3 ± 0.97 a 16.5 ± 2.61 b 0.04 ± 0.03 a 32.01 <0.0001 Cx. salinarius 647.1 ± 171.08 a 2.1 ± 0.75 b 0.1 ± 0.05 b 13.95 < 0.0001 Cx. territans 0.02 ± 0.02 a 0.02 ± 0.02 a 0 a 0.51 0.6006 Culex spp. 22.7 ± 6.57 a 0 b 0 b 11.64 <0.0001 Psorophora ciliata 0.6 ± 0.36 a 0 a 0 a 2.28 0.1059 Ps. columbiae 2.1 ± 0.85 a 0.1 ± 0.06 b 0 b 5.83 0.0037 Ps. ferox 0.9 ± 0.27 a 0 a 0 a 10.94 <0.0001 Ps. howardii 0.04 ± 0.04 a 0 b 0 b 0.98 0.3785 Uranotaenia sapphirina 2.3 ± 1.03 a 0.1 ± 0.07 b 0 b 4.54 0.0124 Damaged 1.1 ± 1.10 a 0 a 0 a 0.98 0.3785 a Means followed by the same letter across a row are not significantly different. Significance was determined by P 0.05. For all comparisons, df = 2,136. more Aedes triseriatus per night than resting boxes. There were no statistically significant differences among in the mean number of mosquitoes collected per night for all other species because the number of each species collected was generally low or that species was collected infrequently. There was also no difference among in the mean number of Cx. pipiens/restuans collected per night because of the large variation in the light trap collections. The number of individuals in this group was more of an indication of the condition of the collection rather than the relative abundance of Cx. pipiens and Cx. restuans. Resting boxes were relatively ineffective for sampling any species at these study sites. Even An. quadrimaculatus, the most common species in resting box collections, was collected in greater numbers in light. Ten positive pools were found among the 1,0 pools analyzed for WNV (Table 3), three from light and seven from. Positive pools consisted of Cx. pipiens (1), Cx. pipiens/restuans group (7), Cx. restuans (1), and Cx. salinarius (1). There were no positive pools from resting boxes. The details of each positive pool are provided in Table 4. The Cx. pipiens/restuans group had the highest infection rate of 1.65 infected mosquitoes per 1,000 individuals (Table 5). Culex pipiens and Cx. restuans individually had very similar infection rates of just over one infected mosquito per 1,000 individuals. The overall MLE of the infection rate for was over 32 times greater than that of light
Vol. 32, no. 2 Journal of Vector Ecology 289 Table 3. Number of mosquito pools tested for West Nile virus and number of positive pools, by trap type and species, 2004. Numbers in parenthesis refer to the number of positive pools. Species Total no. pools No. from light No. from No. from resting boxes Aedes albopictus 22 11 11 0 Ae. canadensis 1 1 0 0 Ae. cantator 32 30 2 0 Ae. j. japonicus 7 0 7 0 Ae. sollicitans 6 4 2 0 Ae. taeniorhynchus 2 2 0 0 Ae. triseriatus 16 9 7 0 Ae. trivittatus 4 4 0 0 Ae. vexans 254 247 6 1 Aedes spp. 6 6 0 0 Anopheles crucians complex 43 41 2 0 An. punctipennis 40 35 3 2 An. quadrimaculatus 92 59 15 18 An. walkeri 33 31 2 0 Anopheles spp. 9 9 0 0 Coquillettidia perturbans 98 91 6 1 Culex erraticus 12 10 1 1 Cx. pipiens 55 (1) 14 39 (1) 2 Cx. pipiens/restuans 128 (7) 82 (2) 44 (5) 2 Cx. restuans 55 (1) 14 39 (1) 2 Cx. salinarius 491 (1) 466 (1) 21 4 Cx. territans 2 1 1 0 Culex spp. 27 27 0 0 Psorophora ciliata 6 6 0 0 Ps. columbiae 25 21 4 0 Ps. ferox 13 13 0 0 Ps. howardii 1 1 0 0 Uranotaenia sapphirina 18 16 2 0 Damaged 2 2 0 0 Totals 1,0 (10) 1,253 (3) 214 (7) 33, 2.29 and 0.07 per 1,000 individuals, respectively. For all species, the had higher infection rates than the light, with the exception of Cx. salinarius. This is most likely because 99.6% of the Cx. salinarius mosquitoes were collected in light. DISCUSSION Light clearly collected more mosquitoes from more species than the other trapping methods. The primary difficulty with light trap collections is that the majority of mosquitoes collected are unfed nullipars. Parity rates were not determined for this study, but the percentage of or blood-fed mosquitoes collected from light has been estimated to be between 21 and 35% (Barr et al. 1986, Reisen and Pfuntner 1987) while in, up to 95% of the mosquitoes may be parous (Reiter et al. 1986). As a result, light need to collect 2.7 to 4.5 times as many mosquitoes as to attain a similar total number of parous mosquitoes. Looking at the most commonly infected mosquitoes from this study, Cx. pipiens and Cx. restuans, collected 2,883 of these species while light collected 5,066 individuals. This number falls short of the approximately 8,000 to 13,000 mosquitoes required from the light to achieve a similar number of parous mosquitoes as the. This is most likely the reason for the lower infection rates from the light for these species. While it is possible that using simultaneously with light at each site lured mosquitoes away from the light and into the, it is unclear if the light trap infection rates would have been higher in
290 Journal of Vector Ecology December 2007 Table 4. Collection date, location, size, and trap type for each mosquito pool testing positive for West Nile virus, 2004. Species Date collected Site No. in pool Trap type Cx. pipiens 8/10/2004 New Castle 40 Cx. pipiens/restuans 7/13/04 7/27/04 8/10/04 8/24/04 8/24/04 8/24/04 9/9/04 New Castle Augustine 22 1 light light Cx. restuans 9/7/2004 New Castle Cx. salinarius 8/11/2004 Augustine light Table 5. Maximum likelihood estimation of the mosquito infection rates per 1,000 mosquitoes for West Nile virus by species and trap type, 2004. Species MLE (95% CI) Total Light Gravid Cx. pipiens 1.13 (0.07-5.55) 0 1.41 (0.08-6.93) Cx. pipiens/restuans 1.65 (0.73-3.26) 0.68 (0.12-2.24) 3.74 (1.40-8.35) Cx. restuans 1.19 (0.07-5.86) 0 1.37 (0.08-6.73) Cx. salinarius 0.04 (0.00-0.22) 0.04 (0.00-0.22) 0 Overall na 0.07 (0.02-0.08) 2.29 (1.01-4.55) the absence of the. In fact, the low infection rates found in light trap collections in 2003 (Williams and Gingrich 2007), without the use of, suggest that this was not the case. Despite their apparent lack of sensitivity, light offer the advantage of sampling large numbers of species not collected effectively with. Gravid collected fewer individuals and fewer species than light. The infection rate for each infected species was higher in the, with the exception of Cx. salinarius. In addition, the specimens collected from the were in much better condition for morphological identification than the light trap specimens. Fewer specimens also resulted in fewer analyses and reduced costs. Based on the overall infection rates, the offered a distinct advantage over the light. However, do not efficiently collect several species of mosquitoes that may be important WNV vectors, such as Cx. salinarius. In the present study, the Cx. salinarius infection would have been completely missed without the use of light. For this reason, both trap types may be required. Gravid are better suited to endemic surveillance purposes where timely results are required. Light may be better suited for epidemic research purposes when it is unclear which mosquito species are serving as potential vectors. Resting boxes were ineffective for WNV surveillance at our research sites because they collected too few specimens to be useful for surveillance purposes. In more suitable habitats, resting boxes can collect significant numbers of Cs. melanura. However, Cs. melanura is reported to have limited potential to serve as an enzootic vector of WNV and no potential to serve as a bridge vector (Turell et al. 2005). Resting boxes are better suited for eastern equine encephalitis surveillance, for which Cs. melanura is the principal epiornitic vector (Howard et al. 1988). Acknowledgments We thank the Delaware Department of Natural Resources and Environmental Control, Mosquito Control
Vol. 32, no. 2 Journal of Vector Ecology 291 Division for providing funding for this project and access to research sites. We are indebted to LTC Charles Cannon, Alexandra Spring, Georgette Sala, and Melissa Miller from the U.S. Army Centers for Health Promotion and Preventive Medicine, Fort Meade, MD, for the use of their laboratory facilities and assistance with the analysis of mosquito pools. Finally, we thank the staffs at the Landfill, Finns Point National Cemetery, Fort Delaware State Park, Killcohook National Wildlife Refuge, and the Ommelanden Hunter Safety Training Center for access to research sites and cooperation with this project. REFERENCES CITED Andreadis, T.G., J.F. Anderson, and C.R. Vossbrink. 2001. Mosquito surveillance for West Nile virus in Connecticut, 2000: Isolation from Culex pipiens, Cx. restuans, Cx. salinarius, and Culiseta melanura. Emerg. Infect. Dis. 7: 670-674. Barr, A.R., A.C. Morrison, and P. Guptavanij. 1986. Parity rates of mosquitoes collected in the San Joaquin marsh. Proc. Calif. Mosq. Vector Contr. Assoc. 54: 117-118. Biggerstaff, B.J. 2003. PooledInfRate: a Microsoft Excel Add-In to compute prevalence estimates from pooled samples. Centers for Disease Control and Prevention, Fort Collins, CO. Crans, W.J. 1995. Resting boxes as mosquito surveillance tools. Proc. N. J. Mosq. Contr. Assoc. 82: 53-57. Darsie, R.F. and R.A. Ward. 1981. Identification and Geographical Distribution of the Mosquitoes of North America, North of Mexico. Mosquito Systematics Supplement Volume 1, 313 pp. Fresno, CA: American Mosquito Control Association. Edman, J.D., F.D. Evans, and J.A. Williams. 1968. Development of a diurnal resting box to collect Culiseta melanura (Coq.). Am. J. Trop. Med. Hyg. 17: 451-456. Gingrich, J.B. and L. Casillas. 2004. Selected mosquito vectors of West Nile virus: comparison of their ecological dynamics in four woodland and marsh habitats in Delaware. J. Am. Mosq. Contr. Assoc. 20: 138-145. Howard, J.J., C.D. Morris, D.E. Emord, and M.A. Grayson. 1988. Epizootiology of eastern equine encephalitis virus in updstate New York, USA. VII. Virus surveillance 1978-1985, description of 1983 outbreak, and series conclusions. J. Med. Entomol. 25: 1-514. Hribar, L.J., J.J. Vlach, D.J. Demay, L.M. Stark, R.L. Stoner, M.S. Godsey, K.L. Burkhalter, M.C. Spoto, S.S. James, J.M. Smith, and E.M. Fussell. 2003. Mosquitoes infected with West Nile virus in the Florida Keys, Monroe County, Florida, USA. J. Med. Entomol. 40: 361-363. Kilpatrick, A.M., L.D. Kramer, S.R. Campbell, E.O. Alleyne, A.P. Dobson, and P. Daszak. 2005. West Nile virus risk assessment and the bridge vector paradigm. Emerg. Infect. Dis. 11: 425-429. Kilpatrick, A.M., L.D. Kramer, M.J. Jones, P.P. Marra, and P. Daszak. 2006. West Nile virus epidemics in North America are driven by shifts in mosquito feeding behavior. PLoS Biol. 4(4): e82 doi:10.1371/journal. pbio.0040082. Lake, R.W., J.M. Doll, and D. MacCreary. 1962. Preliminary investigations of the distribution of Culiseta melanura (Coquillett) in Delaware. Proc. N. J. Mosq. Exterm. Assoc. 49: 169-174. Mans, N.Z., S.E. Yurgionas, M.C. Garvin, R.E. Gary, J.D. Bresky, A.C. Galaitsis, and O.A. Ohajuruka. 2004. West Nile virus in mosquitoes of northern Ohio, 2001-2002. Am. J. Trop. Med. Hyg. 70: 562-565. Nasci, R.S., H.M. Savage, D.J. White, J.R. Miller, B.C. Cropp, M.S. Godsey, A.J. Kerst, P. Bennett, K. Gottfried, and R.S. Lanciotti. 2001a. West Nile virus in overwintering Culex mosquitoes, New York City, 2000. Emerg. Infect. Dis. 7: 1-3. Nasci, R.S., D.J. White, H. Stirling, J. Oliver, T.J. Daniels, R.C. Falco, S. Campbell, W.J. Crans, H.M. Savage, R.S. Lanciotti, C.G. Moore, M.S. Godsey, K.L. Gottfried, and C.J. Mitchell. 2001b. West Nile virus isolates from mosquitoes in New York and New Jersey, 1999. Emerg. Infect. Dis. 7: 626-630. Reisen, W.K. and A.R. Pfuntner. 1987. Effectiveness of five methods for sampling adult Culex mosquitoes in rural and urban habitats in San Bernardino County, California. J. Am. Mosq. Contr. Assoc. 3: 601-606. Reiter, P. 1983. A portable, battery-powered trap for collecting Culex mosquitoes. Mosq. News 43: 496-498. Reiter, P., W.L. Jakob, D.B. Francy, and J.B. Mullenix. 1986. Evaluation of the CDC trap for the surveillance of St. Louis Encephalitis vectors in Memphis, Tennessee. J. Am. Mosq. Contr. Assoc. 2: 209-211. SAS Institute Inc. 1999. SAS/STAT User s Guide, Version 8, SAS Institute Inc., Cary, NC. 3848 pp. Savage, H.M., C. Ceianu, G. Nicolescu, N. Karabatsos, R. Lanciotti, A. Vladimirescu, L. Laiv, A. Ungureanu, C. Romanca, and T.F. Tsai. 1999. Entomologic and avian investigations of an epidemic of West Nile fever in Romania in 1996, with serologic and molecular characterization of a virus isolate from mosquitoes. Am. J. Trop. Med. Hyg. 61: 600-611. Scott, J.J., S.C. Crans, and W.J. Crans. 2001. Use of an infusion-baited trap to collect adult Ochlerotatus japonicus. J. Am. Mosq. Contr. Assoc. 17: 142-143. Turell, M.J., D.J. Dohm, M.R. Sardelis, M.L. O Guinn, T.G. Andreadis, and J.A. Blow. 2005. An update on the potential of North American mosquitoes (Diptera: Culicidae) to transmit West Nile virus. J. Med. Entomol. 42: 57-62. Williams, G.M. and J.B. Gingrich. 2007. Mosquitoes infected with West Nile virus in New Castle County, Delaware, 2003. J. Am. Mosq. Contr. Assoc. (In press).