We examined teachers' verbal interactions employed to enhance the skills of students, through analysis of the records of their spoken words and actions in two groups of classes showing higher attitude scores: one group comprised classes with markedly high learning outcome (skills) (high score group; hereafter referred to as the HSC group), and the other comprised those with no marked outcome (low score group; LSC group). In October and November 1999, long-jump exercise lessons were given to fifth and sixth graders, constituting six classes. In all classes, in which the same teaching plan (problem-solving-type program) was used, all spoken words and actions of the teachers were recorded by VTR in the 2nd, 5th, 8th and 11th lessons, each of which constituted the core of the learning stage. All elements of the spoken words, which were connected with the behavior of the teachers, were classified by parts of speech, the minimum elements of sentences, and their frequencies were counted in each 45-minute session (the analyzed parts of speech are hereafter referred as to AS). The results obtained were as follows: 1) The frequencies of ordinary parts of speech in Japanese (for example: nouns or meishi; verbs or doshi; adjectives or keiyoshi; adverbs or fukushi; post-positional articles or joshi; interjections or kandoshi; and adjectival verbs or keiyodoshi) were compared between the two groups, but no significant differences were noted. This was considered to bebecause all six teachers analyzed in this study showed high attitude scores: this supports the findings of the previous report. 2) Comparison of the frequency of teachers' interactional words use in the HSC and LSC groups showed that adjectives (keiyoshi [positive]) that served to improve skills were used significantly (P<5%) more frequently in the HSC group than in the LSC group. 3) The frequencies per lesson of interactional words were compared for adjectives (keiyoshi [positive]) used generally in class, and for parts of speech concerning characteristics of long-jump skills (daimeishi [personal], meishi [parts of body, movement, timing]) used when teachers helped children to solve problems (targets). It was found that in each case, the frequencies of use were significantly (P<5%) higher in the HSC group than in the LSC group. Furthermore, when teachers promoted understanding about problems (targets), parts of speech (kandoshi [positive]) to improve class atmosphere were used significantly (P<5%) more frequently in the LSC group than in the HSC group. 4) On the basis of the results mentioned in 3) and the results from previous research (Kamihara, Umeno, 2000), the parts of speech used by teachers in promoting children's understanding of problems (targets) and in helping children solve problems (targets) were broadly classified into three categories: parts of speech relating only to attitude scores, those relating only to skill, and those relating to both attitude scores and skill. 5) The verbal interaction used to enhance skills in the USC group was shaped from the teacher's understanding of the formative information about children's problems (targets) as the foundation, arid based on "verbal interaction for understanding children's problems (targets)." Furthermore, in the process of "monitoring," as the teacher possessed the "observational ability" to immediately and definitively assess the content of the "corrective (skills) feedback" given to children, it was clear that this teacher's verbal interaction was deeply connected to the context of solving each individual child's problems.