Although the distinction between bipolar and unipolar disorders served our field well in the early days of psychopharmacology, in clinical practice it is apparent that their phenotypes are only partially described by current diagnostic classification systems. A substantial body of evidence has accrued suggesting that clinical variability needs to be viewed in terms of a broad conceptualization of mood disorders and their common threshold or subthreshold comorbidity. The spectrum model provides a useful dimensional approach to psychopathology and is based on the assumption that early-onset and enduring symptoms shape the adult personality and establish a vulnerability to the subsequent development of Axis-I disorders. To obtain a clearer understanding of the depressive phenotype, it is pivotal that we increase our detection of hypomanic symptoms so that clinicians can better distinguish bipolar II disorder from unipolar depression. Diagnostic criteria sensitive to hypomanic symptoms have been identified that suggest bipolar II disorder is at least as prevalent as major depression. Moreover, the comorbidities of these illnesses are very different and alcoholism in particular appears to be a greater problem in bipolar II disorder than in unipolar depression. Structured clinical interviews and patient self-report questionnaires have also successfully identified the presence of hypomanic symptoms in patients with unipolar disorder and support the concept of a spectrum of bipolar illness. In conclusion, the importance of subthreshold syndromes should not be underestimated as failure to recognize bipolar spectrum disorder could delay treatment and worsen prognosis.