The effectiveness of CO2-baited Centers for Disease Control and Prevention miniature light traps elevated in the tree canopy (approximately 7.6 m) was compared with light traps placed at ground level (approximately1.5 m) and grass-sod infused gravid traps for collecting Culex pipiens, Culex restuans, and Culex salinarius and detecting West Nile virus (WNV) activity in an operational surveillance program that encompassed 12 ecologically diverse sites in Connecticut in 2004 and 2005. More than twice as many Cx. pipiens were collected on average in light traps suspended in the tree canopy than in either light or gravid traps placed at ground level. This difference was generally restricted to those collection sites where markedly greater numbers of Cx. pipiens were collected with all trapping methods but was not associated with site-specific urbanization indices. Culex restuans was not preferentially attracted to light traps suspended in the tree canopy. No differences in the overall abundance of this species were recorded with either of the 2 trapping procedures, but both light traps were more effective than the gravid traps. Culex salinarius was significantly more attracted to ground-based light traps than traps suspended in the tree canopy, while gravid traps were ineffective at all sites regardless of the level of urbanization or any other specific land-use characteristic. CO2-baited light traps placed in the tree canopy were generally superior to ground-based light traps for detecting WNV in Cx. pipiens. West Nile virus-infected females were collected more regularly, and the frequency of infected pools was significantly greater. Twofold higher minimum field infection rates (maximum likelihood estimation [MLE] = 6.7 vs. 3.0 per 1,000 mosquitoes) were also recorded from canopy collections of this species, and virus was detected in canopy-collected females several weeks before it was detected in collections from light traps at ground level. We conclude that the use of CO2-baited light traps placed in the tree canopy for targeted trapping of Cx. pipiens and subsequent detection of WNV are likely to yield better overall results than light traps placed at ground level in this region of the northeastern United States. The virus isolation data obtained from Cx. pipiens collected in gravid traps compared favorably both temporally and spatially with results from canopy trap collections. There were no significant differences in the overall frequency of WNV-infected pools or MLEs for Cx. pipiens, but fewer total WNV isolations were made from Cx. pipiens collected in the gravid traps and virus was detected more infrequently. Results reaffirmed the utility of gravid traps as effective surveillance tools for detection of WNV in Cx. pipiens in the northeastern United States. However, findings also demonstrated that CO2-baited light traps placed in the tree canopy provided more consistent results where weekly detection of virus amplification is a critical objective. The comparative effectiveness of ground- and canopy-based light traps for detection of WNV-infected Cx. restuans and Cx. salinarius was inconclusive owing to the limited number of virus isolations that were made from these species during the 2 years of study. However, WNV virus isolations were made several weeks earlier and more frequently from Cx. restuans collected in traps placed in the canopy rather than at ground level in 2004. Results support the view that ground-based light traps are more effective for detection of WNV in Cx. salinarius.