The usefulness of the Diagnostic and Statistical Manual's (4th ed.; DSM-IV; American Psychiatric Association, 1994) tolerance criterion as an indicator of dependence has been debated. The authors of this study evaluated the performance of DSM's cannabis tolerance criterion, operationally defined as a percentage increase in quantity needed to get high, in distinguishing adolescents with and without cannabis dependence. Two samples of adolescent cannabis users (ages 12-19) provided data (ns = 417 and 380). Tolerance, defined as a percentage increase (median increase = 300% and 175%, respectively, in the samples), had only moderate overall sensitivity and specificity in distinguishing those with and without cannabis dependence. Results suggest limitations of the DSM-IV's change-based operational definition of tolerance in adolescents.