The purposes of this study were to: (1) develop a valid and reliable instrument for direct recording of leadership behavioral characteristics of nursery school children; (2) develop a valid and reliable scale for teachers' rating of leadership behavioral characteristics of nursery school children; and (3) determine the degree of correlation between the observational measure of children's leadership behavioral characteristics and the teachers' ratings of children's leadership behavior. The Nursery School Leadership Observation Schedule (NSLOS) was so devised as to cover all the relevant behavior patterns which could be recorded within a five-minute observation period. The Nursery School Leadership Rating Scale (NSLRS) consisted of the same set of behavior units as the NSLOS in order to secure correlations of the two schedules. Each scale was a seven-point graphic series. The correlation of the observed and rated behaviors fulfilled the function of adding or subtracting further confidence in the validity of the instruments. The findings of the observation recorded on the NSLOS showed a relatively high and significant intraclass reliability among four observers. However, the ratings as graded by the raters on the NSLRS indicated that the intraclass reliability among the raters was high and significant, but the intraclass reliability of the fellowship ratings was comparatively lower. Possible explanations are offered. It can be assumed that the NSLOS and the NSLRS are in general valid and reliable for the measurement of leadership characteristics among nursery school children. Both instruments are appended. (RC)