A measles outbreak in early 1989 among approximately 4200 students at a high school and two intermediate schools in suburban Houston, TX, was investigated to evaluate reasons for vaccine failure and to predict the efficacy of a booster dose of measles vaccine. Seventy-seven cases occurred (71 at the high school, 6 at intermediate schools; attack rate, 3.2 and 0.3%, respectively). Vaccination in the first year of life an 13 to 14 years since last vaccination were independent risk factors for being a case. Forty-three (18%) of 239 sera collected from students just before revaccination during the outbreak were negative by enzyme immunoassay; a neutralization assay confirmed these 43 lacked antibody predicting protection against measles infection. Of 43 enzyme immunoassay-negative students 24 gave another blood sample 9 to 10 months after revaccination. Revaccination appeared to reduce the portion of all students with neutralization titers predicting susceptibility to measles illness with rash from 7.9% to 3.0% and left the portion predicted to be susceptible to illness without rash unchanged (45%).