The original version of the "Helen Keller Story" and a linguistically more complex version of it were used to test the hypothesis that reading comprehension is affected by the complexity of linguistic structures. Complexity was measured by four readability measures, the mean number of words per T-unit, and the Schmidt-Kittrell Linguistic Complexity Scale, but none of these measures accurately reflected the relative complexity of the two versions. When 87 university students were tested on the two versions, using the cloze technique, however, significant differences were found, indicating that the complex version of the story was more difficult to comprehend. Samples of the structures that make the version more difficult are illustrated in the study. (Author/AA)