READING ABILITIES OF PRIMARY CHILDREN WITH VISUAL PERCEPTION PROBLEMS WHO WERE TAUGHT BY CONVENTIONAL METHODS WERE COMPARED TO THE ABILITIES OF CHILDREN GIVEN SPECIAL TRAINING DESIGNED TO IMPROVE FAULTY OR IMMATURE VISUAL PERCEPTION. ONE CONTROL GROUP PARTICIPATED IN SPECIAL ACTIVITIES SUCH AS PICTURE COLORING AND THE OTHER CONTROL GROUP RECEIVED NO INSTRUCTION DURING THE EXPERIMENTAL PERIOD. THE STUDY WAS DESIGNED TO DETERMINE THE BEST METHODS FOR TEACHING CHILDREN OF AVERAGE INTELLIGENCE WHO DEMONSTRATED READING FAILURE, REVERSAL TENDENCIES, CONFUSED VERTICAL ORIENTATION, AND FAULTY VISUAL PERCEPTION AND REPRODUCTION. THE SUBJECTS CONSISTED OF FIRST- AND SECOND-GRADERS FROM LOW SOCIOECONOMIC GROUPS. SCREENING WAS BASED ON INTELLIGENCE, GROSS NEUROLOGICAL NORMALCY, ADEQUATE SOCIAL AND EMOTIONAL ADJUSTMENT, ADEQUATE VISUAL AND AUDITORY ACUITY, READING DEFICIENCY, AND VISUAL PERCEPTUAL IMPAIRMENT. PRE- AND POST-TEST INSTRUMENTS USED INCLUDED THE KEYSTONE VISUAL TEST, WIDE RANGE ACHIEVEMENT TEST, AND THE BENDER-GESTALT. FINDINGS SHOWED NO EVIDENCE THAT CHILDREN RECEIVING THE SPECIAL VISUAL PERCEPTION TRAINING MADE BETTER GAINS IN READING OR DID NOT MAKE BETTER GAINS ALTHOUGH THEY MADE SUBSTANTIALLY IMPROVED BENDER-GESTALT DRAWINGS. EXTENSIVE USE OF CLASSROOM READING MATERIALS RESULTED IN IMPROVED READING SCORES. (AL)