The relationship between combat stress, DSM-III-defined post-traumatic stress disorder (PTSD), and a variety of behavioral factors was examined in a large nonclinical population. A total of 2858 randomly selected American Legion members who had served in Southeast Asia completed a questionnaire which elicited information on military service, personal health, and a variety of mental health outcomes. The data confirm the utility of the PTSD diagnosis as a distinct clinical entity. The frequency of PTSD and the extent of symptoms developed varied with the severity of criteria used for determining the extent of traumatic exposure. The PTSD rate ranged from 1.8 to 15.0% of the total sample, depending on whether "exposure" to combat was defined relatively narrowly or broadly. A distinct linear dose-response relationship between combat stress and a quantitative measure of PTSD intensity was observed. The frequency of PTSD diagnosis was not affected by the presence of either physical or mental health problems which predated military service. A strong, stable relationship was found between combat stress and PTSD intensity for cohorts with differing intervals since the experience of combat trauma, which persisted up to 20 years after discharge from the military. The data thus support a broader approach to defining traumatic events which recognizes individual differences in response to combat, as well as the existence of other behavioral outcomes as residual effects of combat. Implications of these findings and the importance of treating veterans with varying presentations of PTSD are discussed.