Based on the data of the normative study of the "Hong Kong test of specific learning difficulties in reading and writing," and the "Test of visual-perceptual skills (non-motor)-Revised," 99 children aged between 6 and 10 1/2 years were identified as children with dyslexia out of the normative sample of 690 children. By excluding 12 children known to score below average in IQ, 87 children, including 20 children not tested for IQ, could be regarded as children with dyslexia, yielding a prevalence rate of 12.6% and a boy:girl gender ratio of 1.6 to 1. The figures would become 9.7% and 2.0 to 1 if the 20 children were omitted from computation. However, gender imbalance could not be readily explained by gender differences in reading-related cognitive abilities, as there were only minor and subtle differences. Regression analyses to evaluate the relative contribution of different cognitive abilities to reading and writing suggested that orthographic knowledge and naming speed were most important among children with dyslexia. Implications of the findings and the need for early intervention are discussed. (Contains 2 tables.)