The sample consisted of 1,140 male felons recently admitted from the community to North Carolina prisons between March and June 1983. Inmates completed the Diagnostic Interview Schedule to measure the symptoms necessary for DSM-III diagnoses. Data on criminal history, type of offense resulting in current incarceration, and current sentence terms were collected. Most of the inmates did not develop PTSD or its symptoms from a combat-related traumatic stressor event. A relationship between PTSD (and its symptoms) and arrest and incarceration for expressive violence was found when demographic, antisocial personality, and problem drinking factors were controlled. For the majority of those who experienced at least one PTSD symptom and had an arrest history for expressive violence, the PTSD symptoms preceded or occurred in the same year as the violence arrest. This temporal ordering is consistent with viewing PTSD symptoms as etiologically relevant to the occurrence of violence. 4 tables and 32 references