Although family, twin, and adoption studies indicate that attention deficit hyperactivity disorder (ADHD) is a familial condition with a robust genetic component, molecular genetic studies of candidate genes have produced inconsistent findings. One of the challenges to elucidating the genetic architecture of ADHD is its potential genetic heterogeneity. Therefore, efforts are needed to identify etiologically homogeneous subgroups of subjects with ADHD for use in genetic studies. The current article reviews evidence suggesting that parsing ADHD subjects based on comorbidity with conduct and bipolar disorders may yield familial subtypes that are suitable for genetic analyses.