Zadra, Desjardins, and Marcotte (2006) have made a valuable contribution to the empirical testing of the Threat Simulation Theory (TST) (Revonsuo, 2000a) in recurrent dreams. For the most part, their results are in accordance with the theory, while some findings seem to conflict with the predictions of TST. In our commentary, we consider some alternative ways to interpret the results, and we conclude that many prominent features of most recurrent dreams seem to be manifestations of a threat simulation function, leading to repeated rehearsal of threat perception and avoidance, but a minority of recurrent dreams seem to have origins unrelated to threat simulation. © 2005 Elsevier Inc. All rights reserved.