Two classes of randomly assigned first graders were given an experimental form of a reading-typewriting program with one typewriter for each child. The program consisted of 25 workbook lessons, each relating to a single phoneme. The computer was used as a program design tool to determine an optimal teaching order of phonemic elements. The children first typed the new letter several times in isolation, then combined it with previously learned letters and phonetic combinations. Lastly, the children moved to meaningful words. In conjunction with the workbook lessons, there was also a primer which contained phonetically consistent stories composed only of phonetic elements which the children had already mastered, accompanied by workbook exercises for each of the 20 stories. At the end of the first year the Stanford Achievement Test, the Reading Skills Diagnostic Test (Bloomer), a phonetic spelling scale, a handwriting scale, and a typewriting test were administered. The experimental group was significantly superior in paragraph meaning, word study skills, and spelling on the achievement test and in simple phonics, complex phonics, and spelling on the reading test. No difficulty in left to right progression or in writing was evidenced by the experimental group. Tables and references are included. (Author/CM)