Several studies of the heterosexual transmission of human immunodeficiency virus have reported no association between transmission and number of exposures. In contrast, this study showed that for a susceptible sex partner, the number of exposures to an infected index case is indeed associated with transmission, but in a nonlinear fashion. Factors that can dilute an association between transmission and number of exposures include measurement error in calculating number of exposures, use of inappropriate statistical models, and failure to account for variations in transmission rates. For example, the practice of anal intercourse and the experience of bleeding during intercourse increase the likelihood of transmission. We also observed that transmission occurred with fewer exposures among couples who did not use condoms compared with couples who did. The number of exposures also affects the independent association between other risk factors and transmission and thus should be considered when analyzing other sources of risk.