The eye movements of 13 dyslexics, 10 normal, and 16 retarded readers were recorded while they sequentially fixated 7 equidistantly spaced digits from left to right (L----R) and right to left (R----L). This nonreading task simulated the sequential motor components of the reading process. Dyslexics and their controls were selected on the basis of comprehensive and quantifiable criteria. All three groups were of the same chronological age. The dyslexics and retarded readers had also the same reading difficulties, but the causes of their problems were different. Dyslexics made significantly more regressions than the normal or the retarded readers in both directions (L----R and R----L). These results were in line with the ones reported in similar studies and suggest that the dyslexics' erratic eye movements and their excessive number of regressions are present even in nonreading, nonverbal sequential tasks. It may, therefore, be possible to use the eye movement records from nonreading sequential tasks for the objective diagnosis of dyslexia. As such tasks do not depend on reading, they may even be used before reading age.