The latest classifications of mood disorders, such as DSM III and ICD 10, differ from previous ones by emphasizing the bipolar-unipolar distinction. This fairly recent distinction now enjoys a wide consensus. It is supported mainly by family studies. The first works originated from the school of Kleist and Leonhard and were quickly followed by reports by Perris and Angst (1966). Studies that followed the publication of DSM III confirm that the prevalence of bipolar mood disorders is twice as high among first-degree relatives of bipolar probands when compared with those of unipolar probands. Family studies suggest that unipolar and bipolar disorders are distinct but still share some degree of genetic overlap. Other genetic approaches such as twin and adoption studies are less conclusive. Recent linkage studies with genomic markers (RFLPs) cannot directly support the validity of the bipolar-unipolar distinction. They suggest the role of a major gene locus, notably in the Xq27-Xq28 region, only in a limited number of bipolar pedigrees. Most mood disorders might involve a polygenic environment; in this context, a more limited number of factors could be associated with the polarity of the disorder.