This paper reviews published literature on the long-term persistence of immunologic memory for HBsAg after a course of hepatitis B vaccine and the functional significance this has for policy on booster vaccination. Several studies have shown that vaccine induced antibody (anti-HBs) specific for the surface antigen (HBsAg) of hepatitis B virus (HBV) is protective at a serum concentration of 10 milli-International Units per milliliter (mIU ml-1). When acquired passively (e.g. from hepatitis B immune globulin), susceptibility to infection returns as antibody declines. However, vaccine induces active synthesis of anti-HBs accompanied by immunologic memory for HBsAg that affords ongoing protection independent of antibody. Persistent memory over periods of 5 years or more is evident from large, rapid increases in antibody following booster vaccination, even in subjects who have lost antibody. Complementary studies, using an in vitro enzyme linked immunosorbent assay (spot-ELISA), show that the number of memory B lymphocytes able to produce anti-HBs does not diminish as the level of antibody declines. That immunologic memory provides effective immunity is suggested by serologic studies over periods of 5 years or more of vaccinees frequently exposed to HBV. Although many failed to maintain at least 10 mIU ml-1 of antibody, there have been very few clinically significant breakthrough infections. Thus, it appears unnecessary to give healthy vaccinees a booster vaccination when the level of anti-HBs falls below 10 mIU ml-1. Current studies suggest good retention of immunologic memory in healthy vaccinees over periods of 5-12 years. While additional studies will better define the limits of this phenomenon, routine booster vaccination should not be needed to sustain immunologic memory and protection within 5 years and perhaps longer after the primary vaccination series.