Two methods for detection of influenza virus in 451 clinical respiratory specimens were compared: (i) 24-well-plate centrifugation with Madin-Darby canine kidney (MDCK) cells and staining with monoclonal antibody pools to influenza viruses A and B (Centers for Disease Control, Atlanta, Ga.) in an indirect immunofluorescence assay after incubation for 40 h, and (ii) conventional tissue cell culture with primary monkey cells and hemadsorption. For 100 of these specimens, direct examination of smears by the direct fluorescence assay with monoclonal antibodies (Boots Cell Tech/API Analytab Products, Plainview, N.Y.) was also performed. Influenza A virus was recovered from 28 specimens by tissue cell culture after incubation for an average of 4.75 days (range, 2 to 14 days). Influenza B virus was recovered from 35 specimens by tissue culture after incubation for an average of 5.4 days (range, 3 to 14 days). By the centrifugation assay, 23 specimens were positive for influenza A virus and 30 were positive for influenza B virus. All specimens positive by the centrifugation assay were also positive by conventional tissue cell culture. The sensitivities of the centrifugation assay were 82% for detection of influenza A virus and 86% for influenza B virus (84% overall); the specificity of the assay was 100%. Of the 100 specimens studied by direct examination, 15 were positive for influenza virus by both conventional culture and centrifugation assay; however, the direct-smear results for these 15 specimens were negative in 13 cases and inconclusive in 2. The centrifugation assay is a rapid and specific method for detection of influenza A and B viruses in clinical specimens, and it can serve as a valuable and cost-efficient adjunct to conventional culture methods.