A rapid response to a threatening face in a crowd is important to successfully interact in social environments. Visual search tasks have been employed to determine whether there is a processing advantage for detecting an angry face in a crowd, compared to a happy face. The empirical findings supporting the ''anger superiority effect'' (ASE), however, have been criticized on the basis of possible lowlevel visual confounds and because of the limited ecological validity of the stimuli. Moreover, a ''happiness superiority effect'' is usually found with more realistic stimuli. In the present study, we tested the ASE by using dynamic (and static) images of realistic human faces, with validated emotional expressions having similar intensities, after controlling the bottom-up visual saliency and the amount of image motion. In five experiments, we found strong evidence for an ASE when using dynamic displays of facial expressions, but not when the emotions were expressed by static face images.Keywords: Anger superiority effect; Dynamic faces; Face-in-the-crowd effect; Face perception; Emotion.Evolutionary arguments may lead us to expect that threatening faces are detected more efficiently among a crowd of distractor faces than nonthreatening faces. Because facial threat provides a particularly potent social Please address all correspondence to Corrado Caudek, Department of Neuroscience, Psychology, Pharmacology, and Child's Health (NEUROFARBA), University of Florence, Via di San Salvi 12, Complesso di San Salvi, Padiglione 26, 50135 Firenze, Italy. E-mail: corrado.caudek@unifi.it These findings suggest that the ASE might be mediated by mechanisms that are summoned only in part (or not at all) by stimuli lacking the typical dynamic characteristics of facial emotion in natural settings.