Abstract
Most interstate wars are fought or begin between neighbors. This relationship between contiguity and war has long been known, but ignored within peace research. The major reason for this is that it has been seen as essentially a trivial relationship, reflecting the opportunity for war rather than the real underlying cause of war. Recent work on territory and the issues over which wars are fought has begun to question that interpretation. This article maintains that the clustering of war among neighbors may be theoretically significant. It presents a territorial explanation of the relationship and juxtaposes it with the proximity and interaction explanations. Each of the three explanations is appraised in terms of explicit criteria to see which is the most adequate. Ultimately, however, each explanation must specify a set of tests that would falsify it, and this article does so for the territorial explanation. The article concludes with some implications for peace in the post-Cold War era if the theoretical explanation is correct.

This publication has 15 references indexed in Scilit: