The instructional strategy most frequently employed for learning disabled children (Lerner, 1971) is to direct instruction through intact perceptual channels while simultaneously remediating perceptual deficits. This study investigated whether the assumption underlying tks strategy, that children learn more effectively through the stronger modality, is justified. The sample consisted of 44 7- to 9-yr.-old boys and girls (mean of 8 yr., 7 mo.), 11 in each of four groups. Scores on an auditory (Lindamood & Lindamood, 1971) and a visual (Beery, 1967) perceptual test were used to identify children who demonstrated a strength on one test but a deficit on the other. Two perceptual groups were formed by taking those high on the auditory but low on the vis~lal test, and those high on visual but low on auditory. Within each percepmal group, normal and learning disabled children were identified. IQs obtained from a group test administered by the district ranged from 95 to 123, with a mean of 109. Grade equivalents for learning disabled children were at least 6 mo. below grade placement on one or more of three subtests of the WRAT (Jastak & Jastak, 1965 ) ; for normals, grade equivalents were at or above grade placement on all subtests. The experimental task was learning paired-associates having pairs that were nonmeaningful auditory or visual stimuli. Using the number of correct responses over 10 trials as the dependent variable, results of a repeated-measures analysis of variance showed no significant differences between normal and learning disabled children. A significant interaction was obtained between perceptual groups and type of paired-associates (F1,40 = 7.43, fi < .01). The visual group scored significantly higher than the auditory group on visual paired-associates, but for auditory paired-associates, the auditory group's performance was significantly superior. However, results of a test for simple effects indicated that both perceptual groups scored significantly better on visual than on auditory paired