To guide vaccine design, we assessed whether human monoclonal antibodies (MAbs) b12 and b6 against the CD4 binding site (CD4bs) on HIV-1 gp120 and F240 against an immundominant epitope on gp41 could prevent vaginal transmission of simian HIV (SHIV)-162P4 to macaques. The two anti-gp120 MAbs have similar monomeric gp120-binding properties, measured in vitro, but b12 is strongly neutralizing and b6 is not. F240 is nonneutralizing. Applied vaginally at a high dose, the strongly neutralizing MAb b12 provided sterilizing immunity in seven of seven animals, b6 in zero of five animals, and F240 in two of five animals. Compared with control animals, the protection by b12 achieved statistical significance, whereas that caused by F240 did not. For two of three unprotected F240-treated animals there was a trend toward lowered viremia. The potential protective effect of F240 may relate to the relatively strong ability of this antibody to capture infectious virions. Additional passive transfer experiments also indicated that the ability of the administered anti-gp120 MAbs to neutralize the challenge virus was a critical influence on protection. Furthermore, when data from all of the experiments were combined, there was a significant increase in the number of founder viruses establishing infection in animals receiving MAb b6, compared with other nonprotected macaques. Thus, a gp120-binding, weakly neutralizing MAb to the CD4bs was, at best, completely ineffective at protection. A nonneutralizing antibody to gp41 may have a limited capacity to protect, but the results suggest that the central focus of HIV-1 vaccine research should be on the induction of potently neutralizing antibodies.