The relationship between dyslexia and visual stress (sometimes known as MearesIrlen syndrome) is uncertain. While some theorists have hypothesised an aetiological link between the two conditions, mediated by the magnocellular visual system, at the present time the predominant theories of dyslexia and visual stress see them as distinct, unrelated conditions, a view that has received some support from studies with children. Studies of visual stress in adults are rare, yet recent reports of a high incidence of this phenomenon amongst university students with diagnosed dyslexia call for further investigation of the issue. This study sought to clarify the relationship between visual stress and dyslexia by comparing the reading performance of dyslexic and non-dyslexic adults with, and without, colour. Degree of susceptibility to visual stress was determined by means of a symptom rating scale. Optimal colour was determined using an Intuitive Colorimeter, which was also employed to assess reading speed under the two experimental conditions. Only the dyslexic students with high visual stress showed significant gains in reading speed when using optimal colour. The use of response to treatment (rather than symptomatology) as a diagnostic criterion for visual stress is questioned, especially when applied to adults, as this may give misleading findings. On the basis of reported symptomatology, students who experience high levels of visual stress are more likely to show improvements in reading rate with optimal colour if they also have dyslexia than if they do not have dyslexia. Although not establishing an aetiological link, these findings imply an interaction between the two conditions with major implications for theory, diagnosis and treatment.The view that visual factors are involved in dyslexia has a long and controversial history. The current predominant theory -which is so well supported that it could be said to have acquired the status of orthodoxy -is that dyslexia is caused by a genetically-based anomaly in neurological systems sub-serving phonological processing (see