Laboratory and epidemiologic studies have established human herpesvirus 8 (HHV8) as an etiologic agent of Kaposi's sarcoma. With strong evidence linking HHV8 infection with the number of sexual partners among homosexual men, the challenge now is to determine the specific sexual acts associated with HHV8 transmission. Initial studies of specific practices, however, have differed in their conclusions; the paper by Dukers et al. in this issue of the Journal is the first to associate penile-oral intercourse with HHV8 transmission. Many sources of bias may contribute to the conflicting findings of studies reported to date: HHV8 research still lacks an adequately specific and sensitive serologic assay; identification of relevant exposure periods and measurement of sexual practices are imperfect; and sufficient adjustment for confounding is problematic. These numerous potential biases may be particularly important when trying to detect underlying associations that may be of low-order magnitude. The study by Dukers et al. (Am J Epidemiol 2000;151:213-24) is an important contribution to research on HHV8 transmission, but we do not yet know enough about the possible sexual routes of transmission to recommend avoiding any single behavior. For now, the best prevention advice is to reinforce the more general safe sex practices that have been promoted to prevent human immunodeficiency virus and other sexually transmitted diseases.