Objectives: High rates of some depressive symptoms occur in both mixed and pure manic episodes. This study examined whether manic subjects identify these depressive symptoms by self-report consistently with observer ratings, whether dysphoric symptoms are self-rated differently in mixed compared to pure manic episodes, and whether discriminative self-rated dysphoric symptom sets agree with those established by observer ratings., Methods: Ninety-four inpatients meeting DSM-IV criteria for mania were classified as in pure or mixed episodes. Dysphoric symptoms were evaluated with the Hamilton Depression Rating Scale (HDRS) and the self-rated Carroll Depression Scale (CDS). Total scores and individual symptom scores on the two scales were compared, as were differences between the manic and mixed subtypes. Positive predictive values (PPV) of individual CDS statements for a diagnosis of a mixed bipolar episode were calculated. Those with a PPV of 0.5 or greater were summed across all subjects and the distributions within the bipolar manic and mixed groups inspected., Results: Self-rated depressive symptoms were highly concordant with observer-rated depressive symptoms in mania. Differences were demonstrated between mixed and pure manic subjects based on self-report, and these differences were similar to those observed with HDRS evaluations. A group of 8 dysphoric symptoms discriminated mixed from pure manic episodes on both scales. These symptoms were depressed mood, pathological guilt, suicidal tendency, anhedonia, psychomotor agitation, psychic and somatic anxiety, and general somatic symptoms (fatigue)., Conclusions: Manic patients report depressive symptoms consistently with observer ratings. Self-rated dysphoric symptoms differ significantly between mixed and pure manic episodes. Patient self-rating is another tool which may help in the diagnosis of mixed mania and the recognition of depressive symptoms during manic episodes., Limitations: The current study included patients who were evaluated during inpatient hospitalization only. The study included only subjects capable and willing to give written informed consent. Generalizability to other bipolar patients is not established.