Influenza activity was studied in the Rome population from 1956 to 1976 by analysis of mortality from respiratory causes and from all causes. During cold weather months, type A influenza virus was associated, as a rule, with epidemic excess deaths at two year intervals while type B virus was prevalent twice during epidemic excess mortality periods. Excess absenteeism from work and virus isolation data were also compared with epidemic excess mortality during four consecutive years. The evidence obtained indicated that influenza virus isolation alone does not represent a reliable index of epidemic influenza activity in this population. The proportion of deaths attributed to respiratory causes consistently increased in every epidemic, the most pronounced increases occurring during large epidemics. The break-down by age of deaths from respiratory causes in the course of two epidemic periods showed that the percentage distribution of deaths was essentially the same as in non-epidemic periods. This evidence indicates that the same factors influencing the age-related distribution of mortality from respiratory causes during non-epidemic periods, probably affect the fatal outcome of influenza during epidemics.