To develop and test a scheme for analyzing adult-child verbal interaction, tape recordings of such interactions were made in 12 homes representing high verbal ability and average verbal ability. Dialogues obtained were coded in nine categories of utterances, and their frequency and patterning were related to the verbal ability of the 12 7 1/2- to 8 1/2-year-old male subjects. These children were part of a longitudinal study of the pre and post natal period as it relates to learning abilities and disabilities, and had been tested on verbal ability at age 4 and at age 7. A complete medical, social, and psychological history was also available for each child. Data from the present research included seven hours of tape from each family, representing 21 20-minute sessions; the family's commentary on each of the sessions; the researcher's notes from home visits; and the mothers' attitude responses to a questionnaire. Apparent within-group differences produced very large standard deviations and hence findings of non-significance. It was concluded that further naturalistic study would be productive, but that tighter experimental controls should be established. (NH)