Diagnostic assays for the detection of St. Louis encephalitis (SLE) and western equine encephalomyelitis (WEE) viruses in mosquito pools and avian tissues were compared for sensitivity, accuracy and specificity. The in situ enzyme immunoassay (EIA), plaque assay on Vero cells, passage in Aedes albopictus Skuse C6/36 and C7/10 cells, antigen capture enzyme immunoassay (AC-EIA), and single and multiplex reverse transcription-polymerase chain reactions (RT-PCR) were evaluated using pools of 50 mosquitoes containing 1–2 experimentally infected individuals. RT-PCR was the most sensitive assay, with a detection limit of <0.1 plaque forming unit. AC-EIA was the fastest and most economical procedure, but was the least sensitive, detecting only 38% of positive pools. The in situ EIA included initial virus amplification on Vero cells, thereby improving assay sensitivity to detect 68% of positive pools. Passage in C6/36 and/or C7/10 cell culture revealed the presence of infectious virus in samples positive by RT-PCR, but initially negative by plaque assay on Vero cell culture, indicating that detection was related to assay sensitivity and not to the absence of intact infectious virus. Combining WEE and SLE RT-PCR assays into a multiplex assay reduced sensitivity, but still detected viral RNA at titers below plaque assay sensitivity. Plaque assay on Vero cells, mosquito cell passage, and several RT-PCR procedures were evaluated for their ability to detect WEE and SLE in white-crowned sparrow tissues during acute and chronic stages of infection. All assays detected virus during acute infection at times of high viremia; however, only RT-PCR assays were positive by day 7 when virus was not detected in sera. RT-PCR detected SLE RNA in spleen tissue from one bird 51 d after infection. Assay sensitivity also was compared using extracts of homogenized bird organs spiked with known titers of WEE and SLE. Trizol RNA extraction followed by Qiagen one-step RT-PCR was the most sensitive method, but occasionally resulted in the presence of secondary bands confounding interpretation and requiring confirmatory assays. A balanced surveillance program should combine systems that allow the detection of new agents and the sensitive monitoring of endemic agents to provide an early warning of pending health risks.