This article takes a closer look at multiparty wars in the modern state system. Because wars are different, a classification system of interstate wars is useful for theoretical concerns and empirical testing. We label multiparty wars ''complex'' because we believe that the processes that bring them about are much more complicated and difficult to understand than dyadic wars. Complex wars are identified and classified on the basis of their size, the issue that gives rise to them, and whether they are preceded by certain behavioral processes. To see if the classification is empirically useful, a set of empirical expectations on which wars should differ is derived and employed to conduct a series of tests. It is found that dyadic wars are associated with different patterns from complex wars, and that within the set of complex wars, the larger wars differ from the smaller ones on several dimensions. It is also found that larger wars and wars over certain issues are likely to be preceded by different behavioral processes, especially in terms of the use of power politics practices. The classification breaks down each complex war into its component dyadic participants on a number of variables so that one can compare originating dyads with each joiner dyad to show how wars differ according to type.Not all interstate wars are alike. Yet quantitative scholars typically treat them as if they are. This was true of early analyses of the interstate war data (like Singer and Small 1968; Bueno de Mesquita 1981) as well as later ones. 2 It is equally true of analyses of the more recent data on militarized interstate disputes (MIDs) (see Bremer 1992; Senese and Vasquez 2008). Sometimes scholars look at certain 1 Authors' note: We would like to thank Amy Beth Schoenecker, Andy Owsiak, and Michelle Zetek for research assistance. Douglas Lemke provided comments on a previous version of this manuscript. We would also like to thank Chris Leskiw and Susan Sample for sharing their data and Frank Wayman for answering some of our questions on specific wars. For research support John Vasquez would like to thank the Helsinki Collegium for Advanced Studies, the Academy of Finland, and the Thomas B. Mackie research fund of the University of Illinois at UrbanaChampaign. The sole responsibility of the analysis remains with the authors. The data used in this article are available for purposes of replication at the ISQ Web site and at http://tigger.uic.edu/~bvaler. 2 For example, Geller and Singer (1998:156 note 1) state: ''…the explanation of war presented in this book treats all wars as the product of a set of factors drawn from multiple analytic levels.'' Ó