The question of whether alliance formation contributes to peace or to war is analyzed for the modern Great Power system over the last 500 years. It is found that for none of the last five centuries has alliance formation been correlated (tau-b) with war involving the Great Powers, and has often been correlated with peace. Furthermore, disproportionately few wars have been preceded by alliance formation (compared to random probability models), suggesting that alliances are not among the more important general causes of war-involving the Great Powers. Except for the nineteenth century, alliances have frequently been followed by war. This observed association is probably spurious, however, particularly in the sixteenth to eighteenth centuries: Alliances have often been formed in response to unstable conditions, rising tensions, and the anticipation of a probable war. These patterns are interpreted in terms of the nature of military technology and organization, the polarization of the alliance system and other balance of power considerations, and the motivations underlying alliance formation.