Some recent studies on dyslexia have suggested a selective abnormality in the magnocellular visual pathway. To verify this hypothesis, we investigated motion-onset visual evoked potentials (VEPs) (predominantly testing the magnocellular system) as well as pattern-reversal VEPs (presumably testing the parvocellular system) in 20 dyslexics and 16 controls (both groups with a mean age of 10.0 years). Although the latencies and amplitudes of the main positive peak of pattern-reversal VEPs did not differ between the dyslexic and control group, the motion specific negative peak of motion-onset VEPs was significantly delayed (p<0.001) in dyslexics. Our results confirm a selective magnocellular pathway disorder in dyslexics and indicate that the motion-onset VEPs might serve as an objective method for early diagnosis of dyslexia.