At first, Misumi explained why we planed this symposium : Most of the measurement scales of teachers' leadership which have been used these days seemed to be improper. Using such scales, the development of teachers' leadership study could not be expected. Then we proposed the teachers' leadership and the school morale measurement scales which have been made for the past year. We wanted to be criticized on them at this symposium so that we might develop a better measurement method of the teachers' leadership. Regarding this study as one step of studying teachers' leadership, we planned to discuss at the symposium, what it should be. Next, Misumi compared the present situation of the leadership studies in Industrial Organization or Government Offices with those in educational world. He explained the theoretical framework of P-M leadership theory. Two symposists made the following wing reports. Symposist : Yoshizaki At first, he reviewed the past studies of teachers' leadership mainly in Japan. Then he broughtup five problems. Next, he explained our study : The measurement of teachers' leadership was conducted by distributing a questionnaire to pupils, asking them to rate their teachers. Subjects were 3010 5 and 6 graders. First we conducted questionnaire items normal varimax method and there were five factors in teachers' leadership behavior. Secondly we conducted them on a group principal method for selecting questionnaire items of P (Performance) and M (Maintenance) scales. The school morale measurement scales were made at the same procedure as the leadership scales. Symposist : Shinohara Teachers' leadership was divided into four types, depending on the intensity of the two dimensions (P dimension, M dimension). The relationship of four leadership types to pupils' school morale was examined. As the result, PM came first, followed by M, P, and pm in that order, and the differences were statistically significant.