From January 1982 to December 1986, 113 three-year-old children (100 boys and 13 girls) visited the department of pediatrics, Asahikawa Medical College, because of delayed development of language (their expressive language age less than two-year-old). Of these children, 102 children (90%) have visited until they graduated from junior high school for the evaluation of intelligence quotient (IQ), diagnosis, the type of attended school and complications. The mean follow-up period was 10.8 years. The 113 children ware classified as 32 cases of developmental language disorder (DLD), 38 of autistic disorder (Au), 39 of mental retardation (MR), and 4 of deafness based on the results of clinical examination (DSM-III-R), ABR and WPPSI/WISC-R. The purpose of this study is to compare the assessment of language development at the age 3 with the prognosis for intelligence, academic achievement and behavioral adjustment. At the age of three, we divided them into three groups using the Enjoji shiki hattatsu kensa-hyo. Group A including 31 children (29 boys and 2 girls) means delayed development in verbal expression only. Group B including 23 children (17 boys and 6 girls) means delayed development in verbal expression and comprehension. Group C including 59 children (54 boys and 5 girls) means delayed development not only in verbal expression and comprehension but also in communication skills. RESULTS ABR: Four (2 boys and 2 girls) of 113 children did not show any significant waves on ABR at aged 3, and were also diagnosed as deafness by another audiometry. Comparison between the assessment of verbal expression at aged 3 and full scale IQ (FSIQ): FSIQs in 77% of group A were more than 70, while FSIQs in 79% of groups B and C were 70 or below. The assessment of verbal comprehension at aged 3 was significantly related with FSIQ (x2 = 23.88, p < 0.01). Classification of disorders and type of schools according to the assessment at aged 3: [Group A] Thirty one children were classified as 25 cases of DLD and 6 of MR. Before a graduation from junior high school, 20 children attended regular classes and 8 attended special classes for MR. [Group B] Twenty three children were classified as 4 cases of DLD, 10 of MR, 5 of Au and 4 of deafness. Before a graduation from junior high school, 4 children attended regular classes, 8 attended special classes for MR, 6 attended special schools for MR and 4 attended schools for deafness. [Group C] Fifty nine children were classified as 3 cases of DLD, 23 of MR and 33 of Au. Before graduating from junior high school, 10 children attended regular classes, 18 attended special classes for MR, 19 attended special schools for MR and 2 entered educational facilities. CONCLUSIONS 1. Poor mental outcome could be predicted by delayed development of both expressive and comprehensive language, particularly associated with dysfunction of communication skills at the age of three. 2. ABR is a useful method for detecting of hearing loss in non-cooperative young children with delayed development of language.