This essay closely examines the chronological and geographic scope of the Salem witchcraft outbreak of 1692, the only mass witch hunt in American history. Drawing upon data about 152 accused witches in over two dozen communities, as well as scholarship in early-modern witchcraft and collective violence, it challenges the traditional image of Salem as a yearlong epidemic of hysteria. Analysis reveals, instead, that accusations progressed as a sequence of limited and brief flare-ups and that the accused were generally logical targets. Within most communities, the witch hunt passed quickly and the number of accused was small. Even in the high profile centers of the storm, such as Salem and Andover, the episode was limited in duration. Everywhere, the victims of 1692 most often resembled those who were traditionally associated with witchcraft in seventeenth-century England and New England. Despite its reputation for irrationality and excess, Salem witchcraft demonstrated the kind of constraints, limits, and coherence that scholars have found in riots, crowds, and other forms of collective violence. Such an approach helps explain how the outbreak spread to numerous communities as well as why the episode came to an end.