In this study, we examine whether an actual (rather than hypothetical) man being labeled ''gay'' either by himself or by another influences American (US) undergraduates' attributions of the man's masculinity, femininity, and likeability, replicating (with refinements) a similar study from the 1970s. One hundred ninetytwo male and 591 female undergraduates, almost exclusively white, in Kentucky observed two gender-typical white men (one very masculine and the other of average masculinity, both low in femininity, both gay) play a word game on videotape; prior to playing, each man labeled either himself or the other man as either gay or adopted. Male participants rated the men as less masculine and more feminine than female participants, but the label used did not differentially influence male and female participants. Both male and female participants rated each man less masculine and more feminine when labeled gay than when the other man was labeled gay, and rated the more masculine man less masculine and more feminine when labeled gay than when labeled adopted. Whether either man was labeled by himself or by the other man, or whether either man was a labeler or in the presence of a self-labeler, had no effect on participants' ratings of the men's masculinity or femininity. Both men were rated as likeable across all conditions. While the stereotype of gay men as more feminine and less masculine than other men appears robust since the 1970 study, the dislike of gay men appears to have abated.