We examined the effects of three different training conditions, all of which involve the motor system, on kindergarteners’ mental transformation skill. We focused on three main questions. First, we asked whether training that involves making a motor movement that is relevant to the mental transformation—either concretely through action or more abstractly through gestural movements that represent the action —resulted in greater gains than training using motor movements irrelevant to the mental transformation. We tested children prior to training, immediately after training, (...) and 1 week after training, and we found greater improvement in mental transformation skill in both the action and move-gesture training conditions than in the point-gesture condition, at both posttest and retest. Second, we asked whether the total gain made by retest differed depending on the abstractness of the movement-relevant training, and we found that it did not. Finally, we asked whether the time course of improvement differed for the two movement-relevant conditions, and we found that it did—gains in the action condition were realized immediately at posttest, with no further gains at retest; gains in the move-gesture condition were realized throughout, with comparable gains from pretest-to-posttest and from posttest-to-retest. Training that involves movement, whether concrete or abstract, can thus benefit children's mental transformation skill. However, the benefits unfold differently over time—the benefits of concrete training unfold immediately after training ; the benefits of more abstract training unfold in equal steps immediately after training and during the intervening week with no additional training. These findings have implications for the kinds of instruction that can best support spatial learning. (shrink)