Longitudinal data from 290 firefighters who had completed questionnaires 4, 11, and 29 months after exposure to a natural disaster were used to examine the role of intrusive and distressing memories in the etiology of posttraumatic stress disorder. At 42 months, all those who were at risk of having developed a psychiatric disorder (N=113) and a randomly selected comparison group (N=34) who had never developed symptoms were interviewed using the Diagnostic Interview Schedule.The intrusion subscale score of the Impact of Events Scale accounted solely for the etiological link between the disaster and posttraumatic disorders. Avoidance had no direct relationship with the onset of symptoms and appeared to be a defensive strategy to contain the distress generated by the re-experiencing of the disaster, indicating the importance of separating these phenomena from disorder mood and arousal in posttraumatic stress disorder. An information processing model was validated using three different data sets, which suggests its robustness. Using cross-lagged panel correlations, a bidirectional relationship was demonstrated between disorder and intrusive recollections, suggesting that the intensity of recurring memories of a traumatic experience is as indicative of a disturbance of mood and arousal as the exposure to the trauma.