The present study examined: (1) gender and age differences of mean gender identity disorder (GID) trait scores in Japanese twins; (2) the validity of the prenatal hormone transfer theory, which predicts that, in dizygotic (DZ) twin pairs, twins with an opposite-gender co-twin more frequently exhibit GID traits than twins with a same-gender co-twin; and (3) the magnitude of genetic and environmental influences on GID traits as a function of age and gender. Data from 1450 male twin pairs, 1882 female twin pairs, and 1022 DZ male-female pairs ranging from 3 to 26 years of age were analyzed. To quantify individual variances in GID traits, each participant completed four questionnaire items based on criteria for GID from the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR). Our most important findings were: (1) Japanese females exhibited GID traits more frequently than males and Japanese children exhibited GID traits less frequently than adolescents and adults (among females, the prevalence was 1.6 % in children, 10 % in adolescents, and 12 % in adults; among males, the prevalence was 0.5, 2, and 3 %, respectively); (2) the data did not support the prenatal hormone transfer theory for GID traits; and (3) a large part of the variance for GID traits in children was accounted for by familial factors; however, the magnitude was found to be greater in children than in adolescents or adults, particularly among females. This study suggests that although the prevalence is likely to increase, familial effects are likely to decrease as individuals age.