Children who were abnormally hypermetropic at the age of 6 months were randomly allocated treatment with spectacles or no treatment. The eventual incidence of squint was the same in both groups (approximately 24%). The last known visual acuity of the two groups was not significantly different either. Therefore there is no indication to screen infants with a view to preventing squint/amblyopia by optical correction of hypermetropia. If, however, the children allocated treatment are divided into two subgroups--those who wore glasses consistently and those who probably or certainly did not do so--the incidence of squint was the same, but the last known acuities of those who consistently wore glasses may be better than those who did not do so. This suggests that it may yet prove possible to prevent severe amblyopia.