How well antibodies can protect against disease due to HIV-1 infection remains a pivotal but unresolved issue with important implications for vaccine design and the use of prophylactic antibody to prevent infection after accidental exposure to the virus and to interrupt transmission of virus from mother to child. Strong doubts about the possible utility of antibodies in vivo have been raised because of the relative resistance of primary viruses to antibody neutralization in vitro. Primary viruses are likely to be close to the viruses transmitted during natural infection in humans. Vaccine studies have been of little value in assessing antibody efficacy in vivo because none of the strategies described to date have elicited significant neutralizing antibody responses to primary viruses (reviewed in ref. 1). Passive immunization studies are similarly hindered by the paucity of reagents able to neutralize primary viruses effectively² and a single study has suggested some benefit³. Here we describe experiments to explore the ability of passive antibody to protect against primary virus challenge in hu-PBL-SCID mice. In this model, severe combined immunodeficient (SCID) mice are populated with human peripheral blood mononuclear celI (PBMCs) and infected with HIV-1 (ref. 4). We find that the potent neutralizing human monoclonal antibody lgG1b12 at high dose is able to completely protect even when given several hours after viral challenge. The results are encouraging for antibody-based postexposure prophylaxis and support the notion that antibody induction could contribute to an effective vaccine.