The role of the weapons trade has often been ignored in the study of conflict. This article presents empirical research that strongly supports a broad conceptualization of armed conflict in Africa and addresses a critical gap in the literature on the arms trade and war. Using multivariate binomial logistic regression, it finds that arms transfers are significant and positive predictors of increased probability of war. The models, which also include posited conflict-related domestic political, economic, military, and cultural variables, explain 90% of the cases and two-thirds of the war cases, thus making a further contribution to the understanding of conflict dynamics and correlates. In addition to arms transfers, six other variables - semi-democracy, regime transitions, military spending, cumulative previous conflict, and ethno-political groupings - correctly predict war incidence, while economic development exhibits the expected negative relationship; two other factors - repression and colonial legacy - contrary to expectations, are also negatively related. The analysis and findings, while not asserting explicit causal links between arms transfers and the incidence of political violence, support the view that weapons acquisitions are virtually necessary as ingredients in the recipe for war, and that meaningful restraint by suppliers and recipients alike is needed to break the nexus between arms and conflict in sub-Saharan Africa.