The relationship between 18 preparatory teachers' verbal behavior and their students' attainment of number concepts in investigated. The study is based on research of teacher observational systems used to identify "low inference" teacher-variables; this literature is reviewed in this report. The six variables measured by analyzing ten 20-minute recordings over six monthes were: warmth, performance (demanded of students), verbal (number of interchanges), cognitive stimulation, positive feedback, and explicit feedback. Tests on number attainment, number extension, number concepts, form consistency and semantic differential were used to assess student improvement. Only the number concepts test was appropriate for a pre- and post-measure. Thus, results on this test were used to predict scores on the others, and the difference between predicted and attained scores was used to measure improvement attributed to instruction. Each of the six teacher-variables was scaled on a high-medium-low basis. Analysis of covariance was used to investigate whether or not a teacher's score on any variable was related to her children's improvement. The performance variable was significant in favor of low or high teacher scores. Trends are discussed on a post-hoc analysis. Statistical data and the number concepts test are included. (JP)