WHENEVER AN URBAN disorder breaks out, numerous explanations are offered for the cause of the trouble. Some theories are extremely unitary and parsimonious: the rioter is basically "criminal" and loves to destroy property, or he riots to cover basically larcenous motives, or he is a pawn in the hands of outside agitators sent by Peking, Moscow, or (in the minds of some midwesterners), New York. Other explanations are more sophisticated: the trouble lies in poor housing, job discrimination and inferior schools for black children; or, the black, seeing an affluent America all around him, feels "relatively deprived" even though his own lot may be slightly better than it formerly was (Merton and Kitt, 1950). Another explanation points out that the progress made over the past ten years in removing inequalities has led to a sense of rising expectations in the ghetto, so that failure to make continuing rapid progress toward equality is perceived as an intolerable frustration (Davirs, 1969).Each of these latter commonly held points of view implies that deprivations and injustice produce a sense of frustration which ultimately boils over into riot, and all of these ideas are related to a principle long known to psychology: that frustrations produce a build-up of tensions within individuals which may eventuate in aggression and violence. The idea, which is essentially "hydraulic" in its approach to human motivation in that it views motives as building up like water pressure, is stated in its simplest form in the works of certain neo-Frcudians (Hartmann et al., 1949) and in the frustration-aggression hypothesis of Dollard and his colleagues (Dollard et al., 1939). While having the advantage of parsimony, this principle has not been a generally good predictor of human behavior in controlled experimental settings (Kaufman, 1970) and has, in its simplest form, been largely abandoned by researchers. However, some investigators ( e.g., Berkowitz, 1969) have been reluctant to give up the frustration-aggression hypothesis entirely and have revised it and specified the intervening variables in the relationship.Although the frustration-aggression hypothesis may still have much to say on the origins of human aggression, the approach taken in this paper will be that it is not a satisfactory single explanation. As we shall see, research has uncovered some subtle and complex variables in the process, one of the more important being the presence within the situation of stimuli having some relation to aggression. The hypotheses which have guided the research to be reviewed and those which have developed as a result of it are, for the most part, untested in the "real world." They are the product of the psychological laboratory, where evi-