Progress in identifying the genetic basis of bipolar affective disorder has been disappointing, most probably because of the genetic and phenotypic heterogeneity of the condition. These setbacks have led to the adoption of alternative strategies such as the use of endophenotypes or intermediate traits to identify those individuals at genetic risk for developing the disorder. Gottesman and Gould [Am J Psychiatry (2003), 160:636], in a review of the endophenotypic concept, have suggested five criteria that should be characteristic of a trait in order for it to qualify as an endophenotype. These five criteria are used in order to assess the viability of using personality traits as endophenotypes for genetic analyses of bipolar disorder. A review of the literature suggests that certain personality traits or temperaments are associated with the illness in a state independent manner, that personality is at least partly heritable, and that various temperaments aggregate in the non-affected relatives of bipolar probands. Nevertheless, it is unclear whether specific personality traits co-segregate with affectively ill individuals. We conclude that personality profiling of probands and their relatives may facilitate molecular genetic work, but given the fact that personality is itself a complex trait, its use as an endophenotype has certain limitations.