Results of three experiments support hypothesized properties of the security motivation system, a special motivational system for handling potential threats, as proposed by Szechtman and Woody (2004). First, mild stimuli suggesting potential harm produced a marked state of activation (evident in both objective and subjective measures), consistent with the hypothesis that the security motivation system is finely tuned for the detection of potential threat. Second, in the absence of corrective behavior, this evoked activation is persistent, supporting the hypothesis that once stimulated, the security motivation system produces an enduring motivational state involving the urge to engage in threat-reducing behavior. Third, engagement in corrective behavior was effective in returning activation levels to baseline, whereas cognitive reappraisal was not. These findings are consistent with the hypothesis that deactivation of the security motivation system depends on performance of security-related behaviors, rather than non-behavioral events such as cognitive re-evaluation of threat.