Nicolson and Fawcett (Cognition 1990; 35: 159-182) have suggested that a deficit in the automatization of skill learning could account for the general impairments found in dyslexia. Much of the evidence for their claims has been collected via a dual task paradigm, which might allow for alternative explanations of the data. The present study examines automatic skill learning in a single task paradigm and extends previous studies by independently examining the contribution of stimulus-based and response-based learning. The task replicates Mayr's (J. Exp. Psychol.: Learning Memory Cognition 1996; 22: 350-364) methodology in the Serial Reaction Time task by exposing participants to two structured displays, simultaneously. Learning is measured by comparing RT to the learned sequence against RT to a random display. This study demonstrates learning for both dyslexic and control groups for a spatial sequence which was observed and a concurrent non-spatial sequence which was responded to via a keypress. Learning of the sequence did not seem to depend on awareness of the sequence structure. These results suggest that automatic skill learning is intact in dyslexic individuals.