In 1973, WHO proposed that "excess mortality" be used for comparative assessment of the severity of influenza epidemics between countries. In this study, analysis of the data for deaths in Japan between 1980-1994 revealed excess mortality, which could be a useful indicator in Japan. There were six influenza epidemics, and statistically significant excess mortality rates were calculated in both deaths from acute respiratory diseases and all causes. The total number of excess deaths from acute respiratory diseases during influenza seasons for the 15 years was 13,931. Statistically significant excess mortality rates were also shown for other diseases, including ischemic heart disease, cerebro-vascular disease, diabetes, nephritis, chronic liver disease, malignant neoplasm, and chronic respiratory disease. Therefore, "excess mortality" is clearly present in recent years in Japan. In analyses of age-specific excess mortality rates, the epidemic was found to be most fatal to the aged. Further, the value of % excess (= excess rate divided by expected rate x 100) suggested that the number of deaths from all causes would be a more useful indicator than that from acute respiratory diseases. It was confirmed that excess mortality was applicable to regional surveillance of epidemic influenza.