Part one of this analysis is an empirical test demonstrating the existence of a short-term (15-30 years) and a long-term (80-120 years) periodic fluctuation in the historical occurrence of violence, with the amount of violence used to determine periodicity and the type of war to explain fluctuations in intensity. Data further indicate that the long-term periods of intense conflict are associated with social change or turmoil evidenced by civil conflict. Part two offers explanations of the short-term fluctuation as a generation effect in decisionmaking or a fading of remembrance and experience as a permissive condition for future wars, and of the long-term fluctuation in an "action-reaction" process to changes in political philosophy, in the broad sense of the term. While these theories go beyond the empirical data, they are consistent with it and provide a framework for additional research. 32 pp. Refs.