The present study manipulated fear-arousal (fear, non-fear) and anonymity (high, low) in a 2×2 factorial design. From the theory of deindividuation (Zimbardo, 1969), the two varlables were expected to induce an internal state of deindividuation, and thereby disinhibit aggressive behavior. Fifty-seven male undergraduates were randomly assigned to each of the four experimental groups. The subjects were asked either to take a pill which had side-effects (fear condition), or to take coffee (non-fear condition). In addition, the subjects in the low anonymity condition were asked their name and about their personal backgrounds, and they were always called by name during the experiment. They were also given a name tag to wear. In the high anonymity condition, subjects were not asked their name nor anything about their personal backgrounds. Instead, they were given a white robe to wear to decrease individuality. The subjects were then given an opportunity to deliver electric shocks to another subject (confederate) through a Buss-type aggression machine. Both the intensity and duration of the shocks were recorded during the administration of aggression. Deilldividuation was measured on a postsession questlonnaire that assesed the subjects' memory of their own aggressive behavior. Prior to statistical treatment, two orthogonal variates, direct aggression and indirect aggression, were identified by a principal component analysis of the aggression data. The effects of fear arousal and anonymity manipulation upon the variates were as follows: (a) fear arousal increased indirect aggresson but did not affect direct aggression; (b) anonymity manipulation affected direct aggression but did not affect indirect aggression. The questionnaire data did not confirm the mediation of the deindividuated intemal state. It may be concluded that fear arousal and anonymity manipulation affected different aspects of aggressive behavior, though it remains uncertain whether or net the effects were mediated by the internal state of deindividuation.