Two experiments assessed the timecourse of meaning activation for ambiguous words. Subjects read sentences ending in homographs, and named a subsequent target. When the sentence was not biased toward either meaning of the homograph, the results showed the effects of the frequency of the meanings: The dominant meaning was activated more quickly and maintained longer than the subordinate. When the sentence was strongly biased toward one meaning, only the target related to that meaning was facilitated, regardless of the interval between the sentence and the target. In the second experiment, the homographs were replaced by words to which the targets were not associated. No priming of targets was found, indicating that the results of Experiment 1 were not due to direct activation of target information by the context. The results are discussed in light of criticisms of earlier research that did not consider the timecourse of activation and in light of recent hypotheses about the nature of context effects.