It is recognised that mood disorder diagnostic categories are simplifications with limited validity, and while dimensional measures may be more valid than categories, their utility is uncertain. It has been argued that the criteria for bipolar disorder (BD) are too narrow, and that a ‘bipolar spectrum’ should be recognised. The validity and utility of a dimensional measure of mania, the Mood Disorder Questionnaire (MDQ), was investigated in a cohort (n = 68) of young adults being treated for an episode of major depressive disorder (MDD). MDQ score was higher in men and correlated positively with number of depressive episodes, personality measures, and negatively with reaction time. In those on antidepressants at three month follow up (n = 36), MDQ correlated moderately with restlessness (r = .39, p = .01) and suicidal thoughts (r = .34, p = .02). A genetic study of MDD, BD and categorically defined bipolar spectrum disorder (BSD) found an association with a single nucleotide polymorphism (rs1202874) in GPR50, on Xq28. When BD and BSD groups were combined, the association strengthened (p = .0014; OR 1.97, 95% CI 1.26-3.06). The MDQ was investigated in a sample (n=2942), from the population-based Generation Scotland biobank. The MDQ showed high internal reliability, and in a subset with MDD (n=620), a three component structure. MDQ was higher in men, and in those with recurrent depression, and correlated negatively with age of onset (r = -.191, p = 2 x 10-6). A trimodal distribution of age-of-onset was observed in those with chronic or highly recurrent MDD. Controlling for age, gender and current distress, MDQ correlated negatively with general intelligence (r = -.100, p = 1 x 10-8) in controls. Overall there was reasonable evidence that the MDQ had antecedent, concurrent and predictive validity. There was less evidence to support the reliability or validity of BSD. The findings suggested that in those with MDD (particularly with risk factors such as male gender, early age of onset and recurrence) the MDQ may be useful to (1) identify those who may require more intensive monitoring and (2) inform treatment decisions. Thirdly, classifying mood disorders on the basis of prior course, and including dimensional measures, may be more clinically useful.