In two studies, the first in a school in Peterborough and the second in a school in Norwich, more than 233 children aged 8-12 years received either an orthoptic examination, or an optometric examination, together with an examination using coloured overlays and a test of reading fluency. In both studies more than one-third of the children reported visual symptoms. More than one-third of the children chose to use an overlay, and they read more quickly with it than without. The colour of the overlay chosen was weakly related to the binocular amplitude of accommodation: overlays reflecting greater energy at long wavelengths were chosen more frequently by children with a higher amplitude of accommodation. Although the visual symptoms were strongly related to the use of an overlay, in neither study was the benefit from an overlay strongly related to the orthoptic or optometric findings. Nevertheless, children who used an overlay had slightly, but significantly, reduced mean binocular amplitude of accommodation and fusional reserves. On average, children with 'sensory' or 'motor' instability of the nonius strips of the Mallett unit read more slowly than others, as did those with poor stereopsis. However, 60% of those demonstrating sustained overlay use gave a normal response on the Mallett aligning prism test, compared with 80% of those who did not use an overlay for a sustained period. Another indicator of decompensated heterophoria, Sheard's criterion, did not differentiate subjects who used overlays from those who did not. Although binocular and accommodative anomalies do not appear to be the underlying mechanism for the benefit from coloured filters in most cases, there may be some individuals who respond to coloured filters and in whom these ocular motor factors require treatment. Children with visually precipitated symptoms and/or reading difficulties need both a careful evaluation of their accommodative and binocular status, and an investigation of the effect of coloured filters.