To test the acoustic scanning hypothesis for reading, three experiments were conducted with monolingual and bilingual subjects. Ss performed a crossout task, cancelling letters in a text as they read it for comprehension. Letters remaining uncancelled were then analyzed. In Experiment 1, letters were frequently unmarked in function words, words which are highly predictable since they serve primarily to mark case relationships between content words. Changing passage difficulty did not effect the ratio of letters marked in content and function words. No significant difference was found between monolingual and highly-proficient bilinguals. In Experiments 2 and 3, Anglos marked letters in content words and especially in syllables receiving primary or secondary stress, leaving uncancelled those in unstressed function word position. Less proficient foreign students marked almost equal numbers of letters in function and content words and in stressed and unstressed positions. They also marked letters in digraphs as frequently as in single grapheme to phoneme correspondences. In the experiments, Anglos used acoustic scanning along with prediction of syntax for the task; less proficient bilinguals did not. The relationship between acoustic scanning and syntactic processing is discussed and suggestions for classroom application of the findings are included.