This study analyzes the reciprocal relationship between a direct measure of gun availability and 3 types of violent crime across the 120 most populous counties in the USA. Survey data from the Centers for Disease Control and Prevention's Behavioral Risk Factor Surveillance System are used to construct a measure of household gun prevalence. Hypotheses derived from four competing perspectives concerning the role of guns in the production of violence are tested. Strong support is found for the view that easy access to guns raises the risk of serious violence by giving the perpetrator the power to inflict greater victim injury. By contrast, no support is found for the argument that widespread legal gun ownership lowers violent crime by deterring prospective offenders. (Published Abstract)