Previous studies investigating the familial nature of sexual orientation have often relied on the reports of siblings. They have generally obtained extremely high accuracy rates for proband rating of the sexual orientation of siblings. However, participants in these studies have probably been unusually open about their sexual orientation, and thus it is uncertain if more representative participants would have as accurate knowledge about their families. An unselected sample of twins from the Australian Twin Registry rated their own, their cotwins', and their siblings' sexual orientations. We examined accuracy of heterosexual and homosexual probands' assessments of their twins' sexual orientations, as well as cotwins' agreement about other siblings' orientations. Concordance between twins' ratings of their cotwins' orientations with the cotwins' self-rated orientations was considerably lower than that found in previous studies, as was the level of agreement between members of a twin pair in the assessment of other siblings' sexual orientations. Marital history as a proxy variable for determining the sexual orientation of older subjects did not support its use with women, though its use for older men received some weak support.