The study investigated the relationship between vocabulary knowledge and novel word reading. Fourth-grade students were assessed on standardized measures of word identification, decoding, and receptive vocabulary, as well as on an experimental word identification measure using words that students in the fourth grade are unlikely to have seen before in print. In the experimental measure, pairs of words were matched on printed frequency and orthographic pattern (with a variety of spelling patterns represented), but differed in terms of the frequency of expected oral exposure for children (i.e., higher vs. lower). Results showed that students' receptive vocabulary knowledge was significantly related to performance on both the standardized and experimental measures of word identification, even after accounting for the substantial amount of variance explained by decoding ability. Students performed better reading the words with higher expected oral frequencies on the experimental task than on those items with lower expected oral frequencies. The results point to the benefits, albeit modest, of oral word familiarity for reading words when they are first encountered in print and suggest that this top-down effect is not limited to exception words, as has been suggested, but has a wider scope.