Abstract: This study recounts the successful experiment with 23 classes in 7provinces using 3 years as one cycle for the guided teaching of self-study in mathematics since 1980. In the four indexes, i. e. schoolachievement, growth of ability in self-study, transfer of ability in self-study,and the overall development of all the subjects, the experimentalclasses were practically all more superior to the comparative classes.Results obtained from the 77 experimental classes in 16 provincesstarting from 1981 and the 70 classes in 22 provinces starting from 1982also support the experimental results obtained in 1980. This type ofteaching method is not only of benefit to the guidance of poor students,it also helps to develop the intelligence and wisdom of the superiorstudents, truly enabling the manoeuvering of the student's learninginitiativeness and enthusiasm, and also to bring into play the effect ofthe teacher's instructional guidance, lighten the teacher's low effectlabour, impelling intellectualization. This method not only can developthe student's convergent thinking, it also develops the divergent thinking.

Key words: NULL