The purpose of this study was to investigate the relations between labeling effects and eye movements on perceptual matching tasks. Two topics of special interest were studied:(1)the effect of training by labeling meaningful words, and(2)the process of perceptual transfer. In the experiment a 2 × 2 factorial design was used, which incorporated labeling training(yes or no)and matching tasks(pre-test or post test). Ss were 48 young children(5 years old)divided into two groups: an experimental and a control group. Stimuli consisted of 3 similar drawings for the labeling session, and 6 pairs of drawings made from the previous drawings, for the test sessions. Labels of the drawings were "inu"(dog)and "risu"(squirrel). One label was assigned to 2 drawings. Eye movements were measured by the ophthalmography which could take a photograph of a corneal reflected light. Movements were only measured while Ss were looking at a pair of drawings on the rear screen in both test sessions. Main results were as follows. 1.In the experimental group Ss were influenced by the labeling training: 78.8% of them answered "same" to mismatched pairs. In the control group all of Ss answered correctly. 2.Eye movements of Ss in the experimental group were also influenced by labeling training. When they answered incorrectly, fixations and its duration were decreased significantly. However, the number of shift movements between drawings did not show any change. The above results showed that meaningful labeling by children have effects on the perceptual matching tasks; it also showed a change in their eye movements. Through these findings we may conclude that a)the transfer of answers involves a change of eye movements and b)eye movements may be influenced by a symbolic factor.