In this paper, the percentage of all deaths that are related to pneumonia and influenza (P & I ratio) is evaluated as an indicator of influenza activity and as a method to quantify the impact of influenza on mortality. Analysis of data from 1962-1979 indicates that the P & I ratio is an accurate indicator of influenza activity, as confirmed by independent systems of morbidity and laboratory surveillance. A rule using the P & I ratio for detecting an outbreak of an influenza epidemic was evaluated. It was found to be much more specific than a rule based on the regression model, but did not provide more timely detection of outbreaks of influenza during the past 17 years. The use of the P & I ratio in the autoregressive integrated moving average (ARIMA) model of forecasting provided the most accurate prediction of the expected pneumonia and influenza mortality.