Three studies examined the terror management function of romantic commitment. In Study 1 (N = 94), making mortality salient led to higher reports of romantic commitment on the Dimensions of Commitment Inventory (J. M. Adams & W. H. Jones, 1997) than control conditions. In Study 2 (N = 60), the contextual salience of thoughts about romantic commitment reduced the effects of mortality salience on judgments of social transgressions. In Study 3 (N = 100), the induction of thoughts about problems in romantic relationships led to higher accessibility of death-related thoughts than did the induction of thoughts about either academic problems or a neutral theme. The findings expand terror management theory, emphasizing the anxiety-buffering function of close relationships.