The effect of epidemics of influenza A on mortality in the United States was assessed by studying the monthly numbers of deaths during the years 1968-1976. Deaths from all causes at all ages and among persons aged 65 and over, and also deaths from acute respiratory diseases, and from cardiovascular causes were studied. Deaths from acute respiratory diseases were closely correlated with those from influenza and were taken to be an indication of the severity of influenza outbreaks. This indicator combined with a regression function expressing seasonal variation and secular trend was used to predict total mortality, cardiovascular mortality, and deaths among persons aged 65 and over. In each case the predictions proved to be reasonably close to the observed numbers of deaths. Excess mortality from all causes above that expected from seasonal variation occurred principally in three periods during the eight years of study: 1968-1969, 1972-1973, and 1975-1976, each of which coincided with an epidemic of influenza A of the H3N2 subtype. Similar excesses were seen among persons aged 65 and over and in cardiovascular deaths during the two earlier periods. It is concluded that excess mortality occurred during at least three of the major outbreaks of influenza during the period 1968-1976. This points to the need of studying the effectiveness of immunization in preventing the disease.