In the past, many assumptions have been made about warfare in prehistory, from beliefs in an almost constant pattern of aggression in early societies, to the imagination of a 'golden age' of peaceful societies. A major reason for these disagreements has been the limited range and nature of evidence used to identify warfare. The present paper attempts to develop a series of correlates intended to extend the range of evidence that could be utilised in this discussion with reference to warfare in pre-state communities. The proposed correlates are divided into four sections, comprising correlates related to possible causes of war, preparation for aggression, functional evidence of fighting and, finally, the possible after-effects of warfare. The paper suggests that war is a phenomenon that manifests itself over extended time scales, beyond the possibly short-lived episodes of actual combat; in seeking for symptoms of the 'before' and 'after' it may be possible to develop a more comprehensive understanding of the nature and effects of prehistoric warfare in general. Suggested correlates are compared against available evidence in three case studies, from differing periods in time, location, and types of society. The presence of much weaponry in the Later Bronze Age in the Middle Thames region suggests the possibility of an aggressive period, whereas in Gallia Belgica in the mid first century BC, there is historical confirmation of warfare. The evidence of the change in late prehistoric Hohokam communities of East Central Arizona has been widely regarded as indicating the probability of aggression. The results of the studies cannot confirm nor deny these assumptions, but may help to provide a more inclusive foundation for future discussion and research. The paper attempts to assess the potential usefulness of correlates as a tool to provide a methodological context for the study of complex human interactions such as war.