Antiviral antibody production is often sustained for long periods after resolution of an acute viral infection. Despite extensive documentation of this phenomenon, the mechanisms involved in maintaining long-term antibody production remain poorly defined. As a first step towards understanding the nature of long-term humoral immunity, we examined the anatomical location of antibody-producing cells during acute viral infection. Using the lymphocytic choriomeningitis virus (LCMV) model, we found that after resolution of the acute infection, when antiviral plasma cells in the spleen decline, a population of virus-specific plasma cells appears in the bone marrow and constitutes the major source of long-term antibody production. Following infection of adult mice, LCMV-specific antibody-secreting cells (ASC) peaked in the spleen at 8 days postinfection but were undetectable in the bone marrow at that time. The infection was essentially cleared by 15 days, and the ASC numbers in the spleen rapidly declined while an increasing population of LCMV-specific ASC began to appear in the bone marrow. Compared with the peak response at 8 days postinfection, time points from 30 days to more than 1 year later demonstrated greater-than-10-fold reductions in splenic ASC. In contrast, LCMV-specific plasma cell numbers in the bone marrow remained high and correlated with the high levels of antiviral serum antibody. The presence of LCMV-specific plasma cells in the bone marrow was not due to persistent infection at this site, since the virus was cleared from both the spleen and bone marrow with similar kinetics as determined by infectivity and PCR assays. The immunoglobulin G subclass profile of antibody-secreting cells derived from bone marrow and the spleen correlated with the immunoglobulin G subclass distribution of LCMV-specific antibody in the serum. Upon rechallenge with LCMV, the spleen exhibited a substantial increase in virus-specific plasma cell numbers during the early phase of the secondary response, followed by an equally sharp decline. Bone marrow ASC populations and LCMV-specific antibody levels in the serum did not change during the early phase of the reinfection, but both increased about two-fold by 15 days postchallenge. After both primary and secondary viral infections, LCMV-specific plasma cells were maintained in the bone marrow, showing that the bone marrow is a major site of long-term antibody production after acute viral infection. These results documenting long-term persistence of plasma cells in the bone marrow suggest a reexamination of our current notions regarding the half-life of plasma cells.