Family studies may elucidate etiological relationships between two psychopathological conditions. This study explored the prevalences of personality disorders (DSM-III-R) and the variation of personality traits measured by the Munich Personality Test (MPT) in first-degree relatives of patients with bipolar disorder in comparison to control families recruited in the general population. Although the overall prevalence of having any personality disorder did not distinguish both groups of relatives we found significantly more compulsive personality disorders among relatives of probands with bipolar disorder. Relatives of patients with bipolar disorder also revealed significantly higher mean scores of "rigidity' (MPT); other personality traits, including neuroticism and extraversion, did not distinguish both groups. The observed differences in personality features between both groups of relatives are not mediated by current or previous axis I disorders. Therefore, they may reflect overlap of etiological factors of familial origin.