In this paper I would like to reexamine the result of a previous longitudinal study of the development of speaking skills among Japanese EFL junior high school students. The main purpose of this follow-up study is to prove the core finding of the original to be in agreement with the other cases. The core finding of the original is that in the former period of speaking development, fluency ("speech rate") was ameliorated, whereas in the latter period, only syntactic complexity was improved. This study is a cross-sectional study whose participants consist of 16-7th year students, 16-8th year students, 16-9th year students and 10 adult native Australian speakers. Unlike our hypothesis, the results indicated that "speech rate" had been being improved from early stage (7th year students) to native speakers'. In addition to that, in the middle of the development, the change of syntactic complexity proved to be accompanied. It seems reasonable to conclude that participants would make a change of their spontaneous utterances on the basis of largely "memory-based" into mainly "rule-based" utterances.
