People with central field defects resulting from age-related macular degeneration (ARMD) read very slowly. In this study, oral reading rates were determined for unrelated sequences of words in samples of normal young and old subjects with simulated central scotomata of 2 degrees, 4 degrees, and 8 degrees. Scotomata were stabilized at the fovea of the right eye by electronic feedback of eye position, monitored using a SRI dual-Purkinje Eyetracker. Reading rates were determined by jumping print after each stationary presentation through an increasing number of character spaces on different trials. This procedure mimicked the sequence of retinal images produced during the saccades and fixations of normal reading, but without requiring subjects to make accurate eye movements. In Experiment 1, the letter size that yielded the optimal reading rate was found to increase systematically with scotoma size. However, the optimal reading rate decreased more or less linearly as the scotoma size increased. Experiment 2 showed that the optimal reading rate was obtained for essentially the same duration of text presentation, regardless of scotoma size. Experiment 3 investigated the effect of spatial remapping, in which print obscured by the scotoma was stretched electronically to reappear at the scotoma margin. Compared to a nonremapped control condition, spatial remapping produced small but significant increases in reading rate for both 4 degrees and 8 degrees scotomata. Across experiments, average reading rates were faster for the young than the old subjects. Overall, the results define how reading rate is expected to decrease for central scotomata of different sizes and suggest that spatial remapping of print may improve reading rates in patients with ARMD.