A total of 124 patients who had survived at least 2 years after allogeneic bone marrow transplantation (BMT) were studied. Serum was collected at least once yearly. IgG antibodies were determined by enzyme-linked immunosorbent assay for measles and mumps. Rubella antibodies were analyzed by radial hemolysis. Antibody levels were interpreted as representing immunity, seronegativity, or seropositivity, but with uncertain immunity. The median follow-up of the patients was 6.5 years (range, 2 to 13.5 years). The calculated probabilities of being immune to measles at 3, 5, and 7 years from BMT were 47%, 27%, and 20%, respectively. The corresponding probabilities for mumps were 37%, 12%, and 6%, respectively; and for rubella, 47%, 33%, and 28%, respectively. The probabilities for being seronegative for measles, mumps, and rubella at 5 years after BMT were 60%, 73%, and 52%, respectively. When compared with those patients who had experienced previous natural measles disease (P < .05), the only factor that was important for immunity to measles after BMT was whether the patient had been immunized before BMT. There was no influence of donor seropositivity on the probability of becoming seronegative to mumps during follow-up. We conclude that most allogeneic patients will become seronegative to measles, mumps, and rubella during follow-up. Therefore, long-term B-cell memory function is not maintained, regardless of the immune status of the donor.