STUDY OBJECTIVES WERE--(1) TO ANALYZE 5,000 ENGLISH WORDS TO FIND THE UNDERLYING PATTERNS IN THE SPELLING-TO-SOUND RELATIONSHIPS, (2) TO DEVELOP A MODEL FOR MAPPING SPELLING-TO-SOUND CORRESPONDENCES WHICH NOT ONLY DISPLAYS IN AN ECONOMICAL FASHION THE UNDERLYING SPELLING PATTERNS, BUT ALSO REFLECTS THE SPELLING-TO-SOUND HABITS OF THE MATURE READER, (3) TO FORMULATE ALL THE NECESSARY RULES, BASED UPON THE MODEL, TO PREDICT THE PRONUNCIATIONS OF ALL WORDS ANALYZED, AND (4) TO ABSTRACT RULES AND SUGGESTIONS WHICH COULD AID IN THE TEACHING OF READING. THE DATA AND MODEL WHICH WERE DEVELOPED PROVIDED IMPLICATIONS THAT SIMPLE GRAPHEME-TO-PHONEME CORRELATIONS ARE NOT ONLY UNPRODUCTIVE FOR PREDICTION OF SOUND FROM SPELLING, BUT ALSO ARE IN CONFLICT WITH THE UNDERLYING SYSTEM OF ORTHOGRAPHY. IT WAS RECOMMENDED THAT BOTH MORPHOLOGY AND SYNTAX BE CONSIDERED IN THE SELECTION AND ARRANGEMENT OF WORDS TO BE USED IN THE DESIGN OF TEACHING MATERIALS FOR THE FIRST THREE OR FOUR GRADES. (JM)