A study examined the use of cloze procedure as a measure of learning from content instruction. Subjects were 35 physician assistant (PA) students in a cancer prevention course and 31 education students (ED) enrolled in reading instruction courses. Cloze passages were constructed from published selections on the reading process and occupationally related cancer. Every fifth word was deleted, except for the first and last sentences. Scoring was for exact replacement. Analysis of covariance was used to compare posttest scores of the two subject groups with corresponding pretest scores as covariates. PA students surpassed reading students on total scores on the cancer passage, showing the greatest gains on content related words. Superiority of the PA students on this passage was ascribed to the effects of content instruction. However, the ED students did not achieve superior scores on the reading passage, possibly because they failed to incorporate this subject matter from their reading education courses. (Suggestions for improving the selection and use of content related words are included.) (HTH)