A prospective study of routine school vision tests in 1809 children aged 8 and 10 y was undertaken in schools in the Cambridge Health District over two years (1988-1990) to establish whether a significant number of defects of vision were detected. Although the visual acuity of 31% of children who had an abnormal test was recorded as abnormal, most abnormalities were minimal. Only 15 (0.83%) had newly diagnosed problems requiring treatment. Almost all children with marked visual abnormalities had already been detected before school entry, at the 5 y school vision test or on another occasion. Near vision testing did not identify any previously undiagnosed child with significant defects who did not also have distant vision abnormalities. Satisfactory colour vision test results had been recorded for most children at the 5 y school entry vision test. These data do not justify the continued use of routine screening in junior schools.