Two thousand five hundred seventy intact pairs and 724 single responders from Norwegian twins aged 18-25 years completed questionnaires with information about anxiety and depression and perceived cotwin closeness. The aim of the study was the univariate estimation of sex-specific genetic and environmental effects on an index tapping symptoms of anxiety and depression. An index of social closeness between cotwins was significantly related to the cotwin correlation for anxiety/depression scores. MZ pairs were reported to be closer than DZ pairs, and like-sexed DZ pairs were closer than unlike-sexed pairs. The symptom data were adjusted for this apparent violation of the "equal-environment" assumption in twin studies, but the adjustment did not dramatically affect the parameter estimates of genetic and environmental effects on anxiety/depression. A model specifying male (aM) and female (aF) genetic additive effects, shared environment for males (cM), and individual environmental effects (eM and eF) fitted the adjusted data very well. An alternative model, specifying aM = aF, cM = cF, and eM = eF, and no correlation between those environmental factors shared by brothers and those shared by sisters, fitted equally well. Estimated proportions of total variance from the first model were aM2 = .30, aF2 = .52, and cM2 = .21. The estimates from the second model were aM2 = aF2 = .43 and cM2 = cF2 = .11.