An epidemic of influenza B occurring between Jan-Mar 1986 allowed analysis of this illness in 325 participants in an efficacy trial of influenza A vaccines. Eighty-six individuals with respiratory illness were cultured. Ninety-one randomly selected pairs of sera bracketing the epidemic were assayed for seroresponse to influenza B. To determine the optimal serologic assay, 32 paired sera from people with culture proven influenza B were run using 3 assays (Figure).An ELISA using purified hemagglutinin-neuraminidase was the most sensitive although 31% of individuals with proven influenza B did not have an antibody rise. Thirty percent of the illnesses were influenza B by culture. Of those ill, 26% had a serologic rise; however, 48% of those with no recorded illness also had a serologic rise. The overall seroresponse rate was 42% by ELISA, which was equal in adults and children. A higher percent of culture proven influenza B illnesses were febrile (64% v. 32%, p=.004). Children (1-16 yrs) had more frequent influenza B culture positive illness than adults (17% v. 6%, p=.007). Efforts to identify influenza B illness during a winter season depends heavily on culture documentation. Neither clinical illness nor serologic assessment would fully characterize the epidemic. However, the latter determination emphasizes the high rate of apparently asymptomatic or minor illness with influenza B, particularly in adults.