A seroepidemiologic study of a measles epidemic in an elementary school in which 74.5% of the students had received measles virus vaccine from their private physicians was undertaken to determine if there was a loss of immunity in previously immunized children. Of the 94 cases of clinical measles, 54 had received vaccine by various regimens. The vaccine efficacy for all immunized children was 49.3%, but for children who received live vaccine after one year of age, the current recommendation, it was 83%. There was no significant difference in the clinical illness between the immunized and unimmunized, except for atypical measles in children who had received killed vaccine. Only one of the immunized children was shown to have measles antibodies present prior to the onset of illness. However, convalescent antibody titers in the immunized children were significantly higher than in the unimmunized, suggesting a secondary antibody response. Fourteen per cent of the immunized children without clinical measles who were tested at the midpoint of the epidemic had high antibody titers, compatible with subclinical infection, and four immunized children with low titers in the initial survey were shown to develop fourfold rises in titer without clinical illness.