Three hundred words were programed for presentation by teaching machine or by workbook to mentally handicapped, nonreading subjects selected from the public school system (69 subjects) and from an institution (30 subjects). Both groups were matched on chronological age, mental age, programed words known, Gray Oral Reading Test (raw score) and a paragraph reading test score. One group received instruction by teaching machine, one by programed workbook, and one by conventional classroom methods. Administration required approximately 20 hours over a 5-month period. Measures of gain and retention included programed words known, the Gray Oral Reading Test, spelling words known, and a paragraph reading test administered post-instruction and at 30- and 60-day intervals. Statistical analysis revealed both the institutional and public school groups scored significantly higher than the classroom group on the measures of programed words learned (p=.05 and .01 respectively) and on the paragraph reading test (p=.01 for both). Retention remained significantly high (p=.01) for the public school group but was not demonstrated for the institutional group. The technique was not effective for teaching spelling and in most cases scores on the Gray Oral Reading Test were not sifnificantly different among the groups. (Author/RS)