Based on the author's work and that of collaborators, as well as other contemporaneous research, this article reaffirms the existence of a broad bipolar spectrum between the extremes of psychotic manic-depressive illness and strictly defined unipolar depression. The alternation of mania and melancholia beginning in the juvenile years is one of the most classic descriptions in clinical medicine that has come to us from Greco-Roman times. French alienists in the middle of the nineteenth century and Kraepelin at the turn of that century formalized it into manic-depressive psychosis. In the pre-DSM-III era during the 1960s and 1970s, North American psychiatrists rarely diagnosed the psychotic forms of the disease; now, there is greater recognition that most excited psychoses with a biphasic course, including many with schizo-affective features, belong to the bipolar spectrum. Current data also support Kraepelin's delineation of mixed states, which frequently take on psychotic proportions. However, full syndromal intertwining of depressive and manic states into dysphoric or mixed mania-as emphasized in DSM-IV-is relatively uncommon; depressive symptoms in the midst of mania are more representative of mixed states. DSM-IV also does not formally recognize hypomanic symptomatology that intrudes into major depressive episodes and gives rise to agitated depressive and/or anxious, dysphoric, restless depressions with flight of ideas. Many of these mixed depressive states arise within the setting of an attenuated bipolar spectrum characterized by major depressive episodes and soft signs of bipolarity. DSM-IV conventions are most explicit for the bipolar II subtype with major depressive and clear-cut spontaneous hypomanic episodes; temperamental cyclothymia and hyperthymia receive insufficient recognition as potential factors that could lead to switching from depression to bipolar I disorder and, in vulnerable subjects, to predominantly depressive cycling. In the main, rapid-cycling and mixed states are distinct. Nonetheless, there exist ultrarapid-cycling forms where morose, labile moods with irritable, mixed features constitute patients' habitual self and, for that reason, are often mistaken for "borderline" personality disorder. Clearly, more formal research needs to be conducted in this temperamental interface between more classic bipolar and unipolar disorders. The clinical stakes, however, are such that a narrow concept of bipolar disorder would deprive many patients with lifelong temperamental dysregulation and depressive episodes of the benefits of mood-regulating agents.