Balance ability in dyslexia is an issue of considerable theoretical and applied significance, but the literature currently lacks consensus. This study applied objective measures to established balance tasks. 17 dyslexic adults and 20 controls matched for age and IQ undertook the heel-to-toe balance test for 1 minute. Further “dual task” tests were also undertaken in which the subject had to balance while undertaking secondary cognitive tasks (counting, slow choice reaction, fast choice reaction). Two factor analyses of variance revealed significant between-group balance differences in the dual task conditions. 24–82% of the dyslexic group showed balance impairment, depending on the criterion chosen. At the group level, the results are directly consistent with the Nicolson and Fawcett (1990) automatisation deficit hypothesis, but the considerable within-group heterogeneity deserves further investigation.