This paper examines Schachter’s (1964) two-factor theory of emotion and reviews published work which either attempts directly to test the theory, or is relevant to its evaluation. Of the three propositions of the theory, the first, that unexplained arousal can be experienced as different emotions according to cognitive circumstances, is identified as the most distinctive. Any adequate test of the theory, therefore, must address itself to this proposition. Of the few attempts to provide such a test only one has produced even partial support, while two further studies have reported results inconsistent with this proposition. Limited support has been found for the second proposition of the theory, that emotional experience will be labile only in the presence ofunexplained arousal. The third proposition, that emotion will be experiencedonly to the extent that physiological arousal is experienced, has received more empirical support than the other two propositions, but is the least distinctive of the three. The dearth of experimental support for this theory, together with other considerations, suggest that it needs to be modified in some key respects. However, further, more adequate tests of the crucial first proposition are required before substantial theoretical revisions are made.