Mosquitoes collected during the epidemic of West Nile virus (WN) in Staten Island, NY, during 2000 were identified to species, grouped into pools of up to 50 individuals, and tested for the presence of WN by using TaqMan reverse transcriptase polymerase chain reaction (RT-PCR) to detect West Nile viral RNA, Vero cell plaque assay to detect infectious virus, and VecTest WNV/SLE Antigen Panel Assay. A total of 10,866 specimens was tested in 801 pools. Analysis of results indicated that TaqMan RT-PCR detected 34 WN-positive pools, more than either of the other techniques. The plaque assay detected 74% of the pools positive by TaqMan, and VecTest detected 60% of the pools positive by TaqMan. The VecTest assay detected evidence of West Nile viral antigen in 67% of the pools that contained live virus detected by plaque assay. A WN enzyme immunoassay performed similarly to the VecTest WN assay. Differences in performance were related to relative sensitivity of the tests. Infection rates of WN in Culex pipiens and Cx. salinarius calculated by the 3 techniques varied, but each estimate indicated a high infection rate in the population. Positive and negative attributes of each procedure, which may influence how and where they are used in surveillance programs, are discussed.