Using family study methodology and assessments by blind raters, this study tested hypotheses about patterns of familial association between DSM-III attention deficit disorder (ADD) and antisocial disorders (childhood conduct (CD) and oppositional disorder (OPD) and adult antisocial personality disorder) among 457 first-degree relatives of clinically referred children and adolescents with ADD (73 probands, 264 relatives), psychiatric (26 probands, 101 relatives) and normal controls (26 probands, 92 relatives). Among the 73 ADD probands, 33 (45%) met criteria for OPD, 24 (33%) met criteria for CD, and 16 (22%) had no antisocial diagnosis. After stratifying the ADD sample into those with CD (ADD + CD), those with OPD (ADD + OPD) and those with neither (ADD) familial risk analysis revealed the following: (1) relatives of each ADD proband subgroup were at significantly greater risk for ADD than relatives of both psychiatric and normal controls: (2) the morbidity risk for ADD was highest among relatives of ADD + CD probands (38%), moderate among relatives of ADD + OPD (17%) and ADD probands (24%) and lowest among relatives of psychiatric and normal controls (5% for both); (3) the risk for any antisocial disorder was highest among relatives of ADD + CD (34%) and ADD + OPD (24%) which were significantly greater than the risk to relatives of ADD probands (11%), psychiatric (7%) and normal controls (4%); and (4) both ADD and antisocial disorders occurred in the same relatives more often than expected by chance alone. Although these findings suggest that ADD with and without antisocial disorders may be aetiologically distinct disorders, they are also consistent with a multifactorial hypothesis in which ADD, ADD + OPD and ADD + CD fall along a continuum of increasing levels of familial aetiological factors and, correspondingly, severity of illness.