Our understanding of neuropsychiatric abnormalities in patients with deletions of the long arm of chromosome 18 (18q- syndrome) is based mainly on sporadic case reports. We characterized the neuropsychiatric phenotype in 27 patients across a wide age range (2-47 years) with breakpoints ranging from 18q22.3-18q21.2. Adaptive behavior scores (Vineland Composite) were significantly higher in females than in males (62 +/- 5 vs. 43 +/- 3). Intelligence ranged from borderline to severely deficient (IQ, 73- < 40), with academic achievement similarly impaired. Performance in specific neuropsychological functions, including attention, novel problem solving, memory, language, visuomotor integration, and fine motor dexterity, was consistently in the moderately-to-severely impaired range. Behavioral problems were common in both sexes, including aggressivity, hyperactivity, and temper tantrums. Contrary to the few previous reports, we found no evidence of psychosis in any patients. In a subset of patients selected on the basis of no prior knowledge of behavioral problems, 1 of 16 patients (6%) had autism, as defined by the Autistic Diagnostic Interview--Revised (ADI-R) [Lord et al., 1994: J Autism Dev Disord 24:659-685]. Thus, the prevalence of autism in 18q- syndrome is probably no greater than that in other developmental disabilities with a similar level of cognitive impairment. In contrast to what has been believed since 18q- was first described 30 years ago, we found no relationship between chromosome deletion size and any measure of cognition or behavior; nor were there any correlations between any of these measures with the presence or absence of abnormalities on MRI or somatosensory-evoked potentials.