In this study the effect of repeated reading on the acquisition of orthographic knowledge was examined. Acquisition of orthographic knowledge was assessed by the effect of word length on reading speed. We predicted that the effect of length in a set of words and pseudowords would decrease after the repeated reading of these (pseudo)words. The study involved fourth and fifth grade dyslexic children, in addition to normal readers in second and fourth grade. Words and pseudowords ranged from four to six letters and were read 16 times. A length effect was found in the dyslexic and younger normal readers, but not in the older normal readers. The length effect did not change from pre-test to post-test, although a large overall improvement in reading speed was found in all groups. These results suggest that repeated reading did not alter the predominantly sub-lexical reading procedure of the dyslexic and younger normal readers. Implications for the interpretation of the length effect and the notion of word-specific orthographic knowledge are discussed.