In an effort to improve oral reading, beginning and remedial reading programs in English focus on phonological awareness skills and recoding with grapheme--phoneme correspondences. The meanings of the words children practice reading aloud are given little emphasis. Some studies now suggest semantic knowledge may have a direct effect on children's oral reading, but it is unclear whether it is due to knowledge of a given word, general semantic knowledge (vocabulary size), or morphological awareness. We asked third and fourth graders with reading difficulty and their typically achieving peers (N = 95) to read polysyllabic words (N = 48) in isolation. We tested children's semantic knowledge for those specific words, general semantic knowledge (vocabulary size), morphological awareness, and orthographic and phonological knowledge. Using generalized linear mixed-effects models, we found a word-specific semantic effect--along with word-specific orthographic and phonological effects--and general effects of semantic knowledge, morphological awareness, and phonological awareness. The results add to the studies showing the importance of semantic information but is unique in clarifying that a general semantic effect may be at least partly morphological. The findings support a distributed processing account of reading acquisition in which readers use all reliable information to pronounce words, not only letter--sound consistencies. There may be implications for curriculum design. The word-specific semantic effect may suggest that beginning readers should practice reading words in their phonological lexicons. The general morphological effect suggests that children might benefit from learning morphological units early in their reading development.