The purpose of the present study is to provide data on serial learning from the developmental point of view under the conditions of serial tasks which can be learned by different learning modes. For this purpose, several used serial tasks were varied as to the degrees of perceptual and spatial cues. Therefore, the tasks were different from each other as to the organizational possibilities. The organizational possibilities of the serial tasks were controlld by two factors, presentational position and visibility. The former consisted of both C and D conditions. In Condition C, five colored circles were projected on the same position on a day-light screen, and in Condition D, the five circles were projected in five different positions on the screen. The latter consisted of three conditions ; V-O, V-CA, and NV-CA conditions. In the V-O Condition, the positions where the circles were projected were visible during the presentation and the recall period. In V-CA Condition, the positions were visible during the presentation, but the screen was covered with a curtain during the recall period. In NV-CA Condition, the positions were not visible during the presentation period, and the screen was covered with a curtain durins the recall period. So, Condition D has more perceptual and spatial cues than Condition C, and Condition V-O has more cues than Condition V-CA, and Condition V-CA has more cues than Condition NV-CA. Three age groups of 36 children(mean age 5 :08, 7 : 06, and 11 : 07 years)each were asked to recall the order of five colored circles presented by their color names. The followings were the main results. 1. The analysis of the number of trials and errors to criterion on presentational position showed that 5 year old children learned the tasks more easily under Condition D than under Condition C, and that 7 and 11 year old children learned the tasks under Condition D as easily as under Condition C. 2. From the data of the analysis of the number of trials and errors to criterion on visibility, no. statistically significant differences were found among visibility groups in each age group. 3. The data on the serial position effect revealed that 5 year old children showed the salient reversed U curve under Condition C. Seven and eleven year old children showed that the serial curves of Condition C and D were similar and linear. 4. The data on ancillary behaviors showed that 5 year old children tended to name color-names aloud during the presentation while the 7 and 11 year old children tended to move their lips silently. Therefore it may be stated that the 5 years old children learned the serial tasks more efficiently when perceptual and spatial cues were available for organizing the order of items than when their cues were not available. For the younger ones learned the tasks with the perceptual operation depending upon their cues, while the olders did efficiently whether their cues were or not because of the learning of the tasks with the verbal-conceptual operation which is free from their cues; also they can organize the order by only temporal cues.