Duels remained an important and surprisingly common means of settling disputes in the American South until after the Civil War. We examine two historical puzzles. First, why did dueling persist as a preferred tool to resolve conflicts in the South? Second, why did duelers use relatively inaccurate weapons when deadlier weapons were available? We construct a game theoretic model and conduct simulation exercises to find the following results. One, when the public views dueling as an appropriate means of mitigating the effects of libel, then it encourages socially desirable behavior such as reduced libel and more moderate behavior. Two, a sufficiently high mortality rate may deter libel without resulting in many dueling deaths. Third, if mortality rates are too high, dueling is no longer an effective institution. We compile a novel data set of newspaper accounts of duels from digitized archives to present empirical evidence that buttresses our insights from the model.