Effects of the different components of the Distar language program were tested on 48 trainable mentally retarded (TMR) children. Three designs were used comparing IQ, sex, continuation status, selected standardized test results, and replicates. Results indicated that the degree of previous formal language training (continuation status) did not have any effect on current functioning. The Distar program produced steady, acceptable progress of the TMR children throughout the school year. In addition, the Distar program proved more effective when compared with the Peabody Language Development program received by 40 TMR children. (The report presents final, third-year data representing part of a series of studies on the most effective language training activities for TMR children. The major portion consists of statistical data. Results of the first 2 years' studies are reviewed. (IM)