James (1890/1950) argued that self-esteem fluctuates around a typical level in response to successes and failures in domains on which self-worth is staked. In a test of James’s hypothesis, 32 students applying to graduate school completed a measure of contingencies of self-worth, and then for 2 months reported global self-esteem twice a week and any day they heard from a graduate program. As predicted, participants who were higher in basing their self-esteem on academic competence showed greater increases in self-esteem on acceptance days and greater decreases in self-esteem on rejection days than participants low in this contingency of self-worth. Daily affect was significantly related to daily self-esteem, but this association was stronger for students who based their self-esteem on academic competence, suggesting that affect and self-esteem are more strongly related when events are relevant to one’s contingencies of self-worth. Implications for research on stability of self-esteem are considered.