It is paradoxical that although antibodies to the outer surface protein (Osp) A of Borrelia burgdorferi protect mice against infection and that immunization of uninfected mice with Osp-A is protective, antibodies to Osp-A induced early in natural infection of mice are not curative. A region recognized by a neutralizing mAb is also recognized by sera from chronically infected or immunized mice but is not bound by sera from mice infected for 15 days. Infection in mice, despite the presence of early Osp-A antibody, may therefore be explained in part by the lack of response to this epitope. Sera from infected humans recognizes this region, although in this case the immune response to Osp-A occurs only late in infection. Nonetheless, the fact that both human sera and sera from mice immunized with Osp-A bind a conformationally dependent epitope that localizes to the same region tentatively suggests that humans are able to respond to a protective epitope and that an Osp-A-based vaccine may elicit protective immunity in humans.