The English classroom practice by the direct method has proved to be effective in extensive reading for senior high school classes. In the class, a question and answer interchange, summarization and storytelling have been tried, and summarization has been a difficult task for average students. The aim of this research was to investigate the effects, on the communication skills of the subjects, of transmitting the instructor's expectation as to the constituent information to be used for the summarization to them. Four measures were employed: i)useful communication unit score, ii)useless communication unit score, both based on the evaluation of the test material by six teachers, iii)summarization judgment, a measure based on the interaction between the useful and useless communication units, and iv)the number of words. Three matched groups, each of 29 first-year students of a senior high school, were trained and tested according to the following program. In the pre-test, all the subjects were asked to read and summarize in the written form within 40 minutes a chapter of an adventure story. In the training period, all the subjects were given the same lesson for understanding the material by questions and answers. After this, for the control group(C-group)the instructor asked two students to summarize the material orally; for the first experimental group(EI-group)he told them a model of the summarization and then let a student tell the class his own summarization; for the second experimental group(EII-group)he asked the students and let them answer the questions which covered the same information as the model given to EI-group, and then asked a student to summarize the lesson. This training procedure was repeated four times. The post-test was the same as the pre-test in the test material, the length of time, the form of answering, etc. The result was that EII-group was superior to the others in general, and C-group showed little progress in the manipulation of information. EI-group came in between, similar to EII-group in that useless communication units did not increase; similar to C-group and inferior to EII-group in the increase of useful communication unit score, though the difference between the useful communication unit score of EI-group and that of EII-group was moderate. From this it is supposed that how to organize information was better acquired by questions and answers. Contrary to the author's expectation, the number of words increased, in all the three groups. This suggests that the progress in the understanding of the material will develop the tendency to express oneself in a less succinct way, which seems to be an important cause of the difficulties that summarization tends to have. The largest increase in the control subjects, and the tendency to restrain in both the groups of experimental subjects show the necessity of the proper help on the part of the instructor.