We quantified the genetic influences affecting the liability to nightmares, and the association between nightmares and psychiatric disorders in a community-based sample. In 1990, 1,298 monozygotic (MZ) and 2,419 dizygotic (DZ) twin pairs aged 33-60 years responded to a questionnaire study in the Finnish Twin Cohort. Structural equation modeling was used to estimate genetic and environmental components of variance in the liability to nightmares. Records on hospitalization and long-term antipsychotic medication were used to estimate the period prevalence of serious psychiatric disorders. Nightmares were reported more frequently in females both in childhood and as adults. The correlation between occurrence in childhood and as adults was 0.69 in males and 0.71 in females. Polychoric correlations of occurrence within the twin pairs were 0.45 in MZ and 0.21 in DZ pairs in childhood, and as adults 0.39 and 0.18, respectively. The best fitting genetic model was that specifying additive genetic and unshared environmental effects. The estimated proportion of genetic effects in childhood was in males 44% (95% confidence interval [CI] 35-52%) and in females 45% (95% CI 38-52%) of the phenotypic variance. As adults the values were in males 36% (95% CI 27-44%) and in females 38% (95% CI 31-45%). Nightmare frequency and psychiatric disorders were linearly associated. Among those with the most frequent nightmares odds ratios (95% CI) were 3.67 (2.48-5.42) for childhood and 5.87 (4.08-8.45) for adults compared with those never having nightmares. Nightmares are quite a stable trait from childhood to middle age. There are persistent genetic effects on the disposition to nightmares both in childhood and adulthood. Nightmares are significantly associated with psychiatric disorders. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 88:329-336, 1999.