For deaf students who use sign language, spatial-gestural information may be an integral dimension of the semantic network which serves to reconstruct meaning during reading. The purpose of this study was to investigate deaf students' comprehension of paragraphs with and without accompanying sign language pictures above English words. Subjects were 49 elementary and 32 secondary students enrolled in a full-time residential program for the deaf. An adapted version of the Silvaroli Classroom Reading Inventory was prepared for the experiment. The adaptation resulted in three forms per graded reading level—a total of 54 for grades one through six. One edition of each of the three forms appeared in standard English print. The second edition was signed with a pictorial representation of the American Sign Language equivalent above each word. The third edition appeared in Cloze format with every fifth sign deleted. Comparison of means for standard English and signed editions produced alpha levels of .0001 for the elementary and secondary students. Cloze sign format also produced gains over the standard English presentation, p = .0001 for the elementary and p = .0001 for the secondary students. For the 32 secondary students, there was a difference (p<.05) between the signed and Cloze sign presentations indicating that, although signed and Cloze sign were both more effective than the standard English, statistical evidence did not indicate equal effectiveness for both editions. The results of the study support a language difference viewpoint. This notion should be considered in teaching reading to the deaf.