One of the most intractable debates in the field of world politics concerns the linkage of systemic structure to international conflict. The dialogue has focused on the relative merits of bipolar versus multipolar and, more recently, polycentric structures. Advocates of each System have their adherents and, for some time now, have agreed to disagree. Most of the debate over structure and conflict thus far has been cast in terms that do not facilitate its resolution. The objective of this study is to work toward a more compelling empirical judgment of the competing claims. Specifically, that involves revision of the central concepts. Structure cannot be assessed only in terms of distribution of power; the concept also should incorporate the notion of autonomous decision centres. With respect to conflict, most commonly referred to as instability, war is held to be a less comprehensive measurement than international crisis. Renewed testing focuses on the linkage of structure to conflict as so defined. Data from the International Crisis Behaviour Project on 280 cases from 1929 to 1979 provide the evidence to compare the phases of structure. The differences that emerge among multipolarity, bipolarity and polycentrism with respect to patterns of conflict are generally consistent with theoretical expectations.