This study represents classroom observations of a secondary school which employs ability grouping for grades seven through twelve. The effects of ability grouping on the interaction in the classroom was measured with use of Flanders Interaction Analysis. Two history teachers with identical grades and phases were used in this study which began in the Fall of 1970 and ended that January. Each teacher's ability groups were observed twice during the semester, once at the beginning of the year and once near the end of the term. Darwin's Chi Square and graphic representations of frequency counts were used to measure differences between the two teachers and within their own ability groups. Preliminary results show that the upper phases tend to receive more empathy, praise and use of their ideas and less direction and criticism than the lower phases according to the Flanders Interaction Analysis system. The two teachers also tend to teach their lower phases in a similar manner. (Author/AG)