The effect of the number of characters on the reading rate of character strings moving from right to left on a CRT was investigated (Experiment 1). Characters were added one by one to the right end of the string, and the number of characters displayed simultaneously was gradually increased from 1 to 30. Subjects (21 undergraduates) were asked to adjust the moving rate of characters until the optimal moving rate for reading was obtained. It was found that the optimal moving rate increased proportionately as the number of characters displayed simultaneously increased from 1 to 5, but it was stabilized at about 190 milliseconds/character when the number of characters exceeded five. This result was interpreted as follows: characters are processed one after another when the number of characters is five or less, while chunked characters are analyzed for morphemes when the number of characters exceeds five. A model for sentence structure analysis that includes morpheme processing was constructed according to this interpretation. The validity of the model was tested (Experiment 2, subjects were 20 undergraduates.), which showed that when the number of characters was five or less, characters were analyzed one after another for the detection of morphemes.