Information about menstrual cycle variables was obtained by questionnaire using 462 female twin pairs. The twins were either members of the Institute of Psychiatry Volunteer Twin Register, or of the Birmingham Population-based Register. The two samples were analysed separately using univariate and multivariate methods so that an independent replication was obtained. Maximum likelihood estimation was used to fit simple models of genetic and environmental variation to these data. The results suggest that age of menarche, menstrual cycle regularity and premenstrual symptom reporting may be heritable, whereas menstrual cycle length is not. The results should be interpreted with caution as not all variables were replicated in the smaller sample, and the method of retrospective menstrual cycle data collection has been questioned.
