Twin studies of psychiatric illness assume that the genetic and environmental risk factors for psychiatric illness are similar in twins and non-twins and in monozygotic (MZ) and dizygotic (DZ) twins. To test this assumption, we examine whether the treated incidence of psychiatric illness in twins deviates from population expectations or differs between MZ and DZ twins. Using first admissions to the Swedish Psychiatric Registry for the years 1979–83 for all twins born 1886–1958 from the Swedish Twin Registries, we calculated Standardized Morbidity Ratios (SMRs) using national incidence rates together with individually computed person-years at risk in the twin cohort. The diagnoses examined, for which there was more than 393000 person-years of risk, were schizophrenia, other non-affective psychoses (ONAP), bipolar affective illness (BPAI), unipolar affective illness (UPAI) and neurotic depression (ND). The SMRs (and 95% CIs) for all twins were: schizophrenia 0·86 (0·69–1·06), ONAP 1·05 (0·88–1·24), BPAI 1·09 (0·90–1·32), UPAI 1·05 (0·85–1·29) and ND 0·99 (0·88–1·10). This pattern of results did not differ substantially when examined separately by gender or birth cohort. Relative risks for first admissions for MZ v. same-sex DZ twins or same v. opposite-sex DZ twins did not differ significantly from unity for any of the disorders examined. In Sweden, the treated incidence of psychotic and affective disorders in twins does not differ from that found in the general population and does not differ across zygosity groups. These results support the validity of the twin method for the study of psychotic and affective disorders.