The stability of 3 cognitive vulnerabilities - a negative cognitive style, dysfunctional attitudes, and rumination-as well as depressive symptoms as a benchmark were examined to investigate whether cognitive vulnerabilities are stable, enduring risks for depression. A sample of adolescents (6th-10th graders) completed measures of these 3 cognitive vulnerabilities and depressive symptoms every 5 weeks for 4 waves of data across 5 months. Mean-level and differential stability were examined for the sample overall and by age subgroups. A negative cognitive style exhibited mean-level stability, whereas rumination and dysfunctional attitudes showed some mean-level change. Absolute magnitudes of test-retest reliabilities were strong for depressive symptoms (mean r = .70), moderately high for a negative cognitive style (mean r = .52), and more modest for rumination (mean r = .28) and dysfunctional attitudes (mean r = .26). Structural equation modeling showed that primarily enduring processes, but not contextual forces, contributed to the patterning of these test-retest reliabilities over time for a negative cognitive style and dysfunctional attitudes, whereas both enduring and contextual dynamics appeared to underlie the stability for rumination. Theoretical and clinical implications of these findings are discussed.