Background: Previous twin studies have supported a genetic contribution to the major categories of psychotic disorders, but few of these have employed operational diagnostic criteria, and no such study has been based on a sample that included the full range of functional psychotic disorders. Methods: A total of 224 twin probands (106 monozygotic, 118 dizygotic) with a same-sex co-twin and a lifetime history of psychosis was ascertained from the service-based Maudsley Twin Register in London, England. Research Diagnostic Criteria psychotic diagnoses were made on a lifetime-ever basis. Main- lifetime diagnoses of DSM-III-R and International Statistical Classification of Diseases, 10th Revision schizophrenia were also made. Probandwise concordance rates and correlations in liability were calculated, and biometrical model fitting applied. Results: A substantial genetic contribution to variance in liability was confirmed for the major diagnostic categories except Research Diagnostic Criteria depressive psychosis and unspecified functional psychosis, where familial transmission was confirmed but the relative contribution of genetic and common environmental factors was unclear. Heritability estimates for Research Diagnostic Criteria schizophrenia, schizoaffective disorder, mania, DSM-III-R schizophrenia, and International Statistical Classification of Diseases, loth Revision schizophrenia were all between 82% and 85%. None of the estimates differed significantly from any other. Conclusions: Heritability estimates for schizophrenia, schizoaffective disorder, and mania were substantial and similar. Population morbid risk estimates were inferred rather than directly measured, but the results were very similar to those from studies where morbid risks were directly estimated.