A study investigated the hypothesis that high frequency words can be effectively specified when word shape data are combined with available contextual clues. To test this hypothesis, an augmented contextual restriction task similar to one used by R. N. and L. R. Haber was employed. The task differed in terms of increased passage difficulty and the provision of length alternatives. Subjects, 20 college students, rated the appropriateness of each alternative word choice for each of 46 items. They were told to circle all of the words from the total list of alternatives that could be used to construct a reasonable sentence in English, then to underline the single word that they felt had been actually used in the reading passage. Results showed that while word shape exerted significantly more constraint than word length, word length still contributed markedly to high frequency word specification. (Excerpts from the restriction task are appended. (Author/FL)