An outbreak of 18 cases of measles in a primary school in the Australian Capital Territory in August and September 1993 provided the opportunity to study measles immunisation status and measles vaccine efficacy. Parents of 384 (78 per cent) of 491 children answered a questionnaire on recent illness consistent with measles and measles immunisation. Parents transcribed details of measles immunisation from the personal health record of the child to the questionnaire. Thirty-three per cent of cases and 3.4 per cent of the other children had not been immunised. Overall, 95 per cent of children had been immunised. The efficacy for all measles vaccines was estimated to be 90 per cent (95 per cent confidence interval (CI) 75 to 96) and for measles-mumps vaccine 87 per cent (CI 70 to 95). All of the immunised cases had received measles-mumps vaccine. There was no increased risk of measles infection in those who had been immunised at under 15 months of age compared with those immunised at 15 months or older, or in those who could not provide a date of immunisation compared with those who could. None of the children who had received two doses of vaccine caught measles. The date of immunisation was provided by 65 per cent of the respondents who said their children had been immunised. Asking parents to provide this date instead of viewing the health record is a less expensive way of assessing immunisation status but this method needs to be evaluated. Measles outbreaks still occur in highly immunised populations when vaccine efficacy appears to be acceptable.