The lexical quality hypothesis assumes that skilled readers rely on high quality lexical representations that afford autonomous lexical retrieval and reduce the need to rely on top-down context. This experiment investigated this hypothesis by comparing the performance of adults classified on reading comprehension and spelling performance. "Lexical experts", defined by above average performance on both measures, were compared with individuals who are good readers/poor spellers, poor readers/good spellers, or poor on both measures. Sentences finishing with a homograph (e.g., "She danced all night at the ball") were followed by a probe word and participants had to decide whether it had occurred in the sentence. Critical probe words were related to either the sentence-congruous or the sentence-incongruous meaning of the homograph (e.g., "waltz" vs. "throw"). Lexical experts showed less interference from related probes than the other groups. When the sentences were presented at fast rates, poorer spellers showed interference for sentence-congruous but not sentence-incongruous probes. However, at slower presentation rates, all groups showed equivalent interference for both types of probes. The results support the lexical quality hypothesis by showing that high quality lexical representations, indexed by better spelling, are associated with reduced reliance on sentence context.