What determines the geographic extent of territory? We microfound and extend Boulding’s “Loss
of Strength Gradient” to predict the extensive and intensive margins of conflict across space. We show
how economies of scale in the production of violence and varying costs of projecting violence at a
distance combine to affect the geographic distribution of conflict and territory. We test and probe
the boundaries of this model in an experiment varying the fixed costs of conflict entry. As predicted,
higher fixed costs increase the probability of exclusive territories; median behavior closely tracks
equilibrium predictions in all treatments.