Arms races between predators and prey may be driven by two related processes—escalation and coevolution. Escalation is enemy-driven evolution. In this top-down view of an arms race, the role of prey (with the exception of dangerous prey) is downplayed. In coevolution, two or more species change reciprocally in response to one another; prey are thought to drive the evolution of their predator, and vice versa. In the fossil record, the two processes are most reliably distinguished when the predator-prey system is viewed within the context of the other species that may influence the interaction, thus allowing for a relative ranking of the importance of selective agents. Detailed documentation of the natural history of living predator-prey systems is recommended in order to distinguish the processes in some fossil systems. A geographic view of species interactions and the processes driving their evolution may lead to a more diverse array of testable hypotheses on how predator-prey systems evolve and what constraints interactions impose on the evolution of organisms. Scale is important in evaluating the role of escalation and coevolution in the evolution of species interactions. If short-term reciprocal adaptation (via phenotypic plasticity or selection mosaics among populations) between predator and prey is a common process, then prey are likely to exert some selective pressure over their predators over the short term (on ecological time scales), but in the long run predators may still exert primary “top-down” control in directing evolution. On the scale of evolutionary time, predators of large effect likely control the overall directionality of evolution due to the inequalities of predator and prey in control of resources.