Data are presented from a national probability sample of males interviewed by telephone and asked their sexual orientation. Of these males 3.7 percent reported that they were homosexual or bisexual. Homosexual/bisexual men were compared with heterosexual ones on the demographic variables. This sample produced larger numbers in those groups which appear to be underrepresented in the usual samples drawn from the gay world. These groups include those with little education, married men, older men, minorities, and those living in small towns. It is suggested that probability samples which do not draw directly or heavily from the gay world for homosexual respondents obtain a broader sampling of those having homosexual feelings or behaviors.