Although twin and adoption show bipolar disorder (BP) has a strong genetic component, few chromosomal regions have been consistently implicated by molecular genetic studies. To address this issue, we sought to determine if quantitative dimensions of bipolar disorder symptoms would be useful for detecting genes that underlie the susceptibility to bipolar disorder. Subjects were 520 individuals diagnosed with bipolar I, bipolar II or schizoaffective disorder, bipolar type who had participated in the NIMH genetics initiative for bipolar disorder. We constructed symptom scores from 29 psychiatric symptoms recorded in the Diagnostic Interview for Genetic Studies (DIGS). Principal components factor analysis followed by a varimax rotation was used to extract symptom dimensions. Factor scores were calculated for all genotyped individuals in the sample, regardless of affection status. Heritable factors were used in a variance-components linkage analysis, which utilized the exact likelihoods of allele-sharing identical-by-descent for each pair of relatives within each pedigree. The principal components factor analysis resulted in five independent dimensions: depressed state, psychosis, sleep disturbances, psychomotor acceleration, and irritability. Two factors were significantly heritable: depression (h2=0.53, p<0.001) and irritable vs. euphoric mania (h2=0.35, p=0.03). These were subsequently used in a linkage analysis that resulted in LOD scores of <2.0, which are not statistically significant. The five constructs developed through factor analysis appear to be consistent with previous factor analyses. Notably, only the dimensions associated with the type of mood disturbance showed high heritability, which suggests that careful measurements of depression, euphoria and irritability may be particularly useful in clarifying the genetic etiology of bipolar disorder in future studies.