A study of the information gained by students in the process of taking a cloze test in English as a second language was conducted with 34 full-time undergraduate foreign students at a state university. The subjects came from a variety of native language backgrounds. A 35-item cloze test was developed in which every eighth word was deleted, beginning with the thirty-second word. Each deletion belonged to one of these structural categories: local (clues found within two words of the blank space), clause (clues found within the simple clause but beyond two words), subordinate sentence (clues found within the subordinate or main clause attached to the clause within which the blank appears), and same paragraph (clues found within the paragraph). The test was administered in a class period, and three weeks later, the intact version of the entire passage was given to the class to read, followed immediately by a second administration of the same cloze test. Exact-word scoring was used. Analysis of the pre-and post-test results showed a significant improvement in scores, but indicated that the location of clues had no significant effect on the responses. The results suggest that the structural categories of the deletions are not crucial in the linguistic processing of written text. (MSE)