In this investigation, carried out under a Youth Health Care postgraduate course in Nijmegen, the stereopsis of a school population of 730 children, aged 4-18 years, was recorded and clinically evaluated. Stereopsis was measured using the TNO test, a random-dot stereo test especially designed for the early detection of amblyopia. The main aims of this study were to establish the validity of the TNO test as a screening test for amblyopia, obtain information about the variability and age dependence of stereopsis, and to evaluate the efficacy of amblyopia prevention. The most important findings can be summarized as follows: 1 All amblyopes are detected by the TNO test's recommended referral criterion of 240 sec arc (binocular threshold parallax in sec arc). 2 The red-green anaglyphs used in the test do not pose problems for individuals with a colour vision deficiency. 3 The ability to discriminate depth improves by a factor of two over the age interval 4-12 years. 4 A stereoacuity of £ 120 sec arc is a good predictor of normal or correctable normal vision, and may therefore help in evaluating the often incomplete results of eye tests of young children. 5 Where there is an increased perinatal risk, there is a greater chance of disturbed binocular vision. 6 It is estimated that 75% of amblyopes remain amblyopic, possibly because of delayed detection; 60% of the amblyopes in the population examined were not identified before the age of 5 years.