To identify the mosquito species competent for West Nile virus (WNV) transmission, we evaluated 10 California species that are known vectors of other arboviruses or major pests: Culex tarsalis, Cx. pipiens pipiens, Cx. p. quinquefasciatus, Cx. stigmatosoma, Cx. erythrothorax, Ochlerotatus dorsalis, Oc. melanimon, Oc. sierrensis, Aedes vexans, and Culiseta inornata. All 10 became infected and were able to transmit WNV at some level. Ochlerotatus, Culiseta, and Aedes were low to moderately efficient vectors. They feed primarily on mammals and could play a secondary role in transmission. Oc. sierrensis, a major pest species, and Cx. p. quinquefasciatus from southern California were the least efficient laboratory vectors. Cx. tarsalis, Cx. stigmatosoma, Cx. erythrothorax, and other populations of Cx. pipiens complex were the most efficient laboratory vectors. Culex species are likely to play the primary role in the enzootic maintenance and transmission of WNV in California.

Three California Culex species previously identified as efficient laboratory vectors of West Nile (WN) virus were tested for their capability to vertically transmit WN virus. Wild-caught Culex pipiens pipiens L., Culex pipiens quinquefasciatus Say, and two populations of Culex tarsalis Coquillett females were inoculated intrathoracically with 10(2.7 +/- 0.1) plaque-forming units of WN virus. F1 progeny were reared at 18 degrees C and subsequently tested as adults for infectious virus on Vero cell culture. Virus was not detected in 197 pools comprising 4,884 Cx. p. pipiens. The minimum filial infection rate (MFIR) for Cx. p. quinquefasciatus was approximately 3.0/1,000 for 665 progeny tested in 28 pools. There was no virus detected in 102 pools of 2,453 progeny from Cx. tarsalis collected in Riverside County. The MFIR for Cx. tarsalis collected in Yolo County was approximately 6.9/1,000 for 2,165 progeny tested in 86 pools. Mosquito progeny infected vertically during the fall could potentially serve as a mechanism for WN virus to overwinter and initiate horizontal transmission the following spring.

A blinded laboratory evaluation compared the accuracy, sensitivity, and specificity of an in situ enzyme immunoassay (EIA), VecTest wicking assay, and reverse transcription-polymerase chain reaction (RT-PCR) to detect and distinguish West Nile (WN) and St. Louis encephalitis (SLE) viruses in pools of 50 mosquitoes. Adult female Culex tarsalis Coquillett were inoculated with either WN or SLE viruses, held for 0-11 d at 28 degrees C, killed by freezing, and then were added to 49 or 48 uninfected mosquitoes to make up 14 pools positive for WN virus, 14 positive for SLE virus, 14 positive for both WN and SLE viruses, and 14 negative for virus. Pools were number coded and tested blindly. Virus was not detected in known negative pools. VecTest and RT-PCR assays were comparably sensitive and accurate, detecting virus in pools containing females held for 3 d postinoculation; only RT-PCR detected SLE virus in pools on days 0-1. The VecTest and RT-PCR produced a single false-positive result for WN and SLE, respectively. RT-PCR detected RNA in samples positive by the VecTest, indicating that the detergent in the wicking buffer did not prevent RT-PCR from confirming VecTest results. Detector antibodies used in the in situ EIA cross-reacted between SLE and WN viruses, reducing accuracy. Both the VecTest and RT-PCR provided rapid and specific results, but they detected only those viruses known to be present. Plaque assay on Vero cells was comparably sensitive and had the added benefit of detecting newly emerging viruses, but this method required virus culture followed by identification, thereby delaying reporting.