In two experiments, adult readers' comprehension of rapid-sequential presentations of text segments of different sizes was investigated with multiple-choice comprehension questions. Subjects were separated into two groups based on their performance on the standard Nelson and Denny (1976) reading test. In Experiment 1, when the display rate was held constant, skilled subjects yielded better comprehension performance for the larger relative to the smaller segment size (i.e., two- and one-word, respectively), whereas the opposite trend was found for less-skilled subjects. In Experiment 2, skilled subjects' ability to retain paragraph meaning increased with segment size, whereas less-skilled subjects showed a reverse trend. Taken together, these results suggest that skilled and less-skilled adult readers differ in the reading unit size they employ.