Preschoolers' understanding of number conservation in a significant context was studied in a series of 2 experiments. In experiment I, nonconservers in pretest(standard number conservation task)participated in a "puppet" experiment in which the conservation task was placed in the context of a puppet skit. Half of them were tested in a condition where the transformation was made by one of the puppets with an explicit reason, while for the other half transformation was made by the tester without any apparent reason. The mean success rates were 67% and 24% respectively. In experiment II, nonconservers in pretest were tested in 3 conditions in the puppet experiment: whether the transformation was made by the tester with a reason, or by one of the puppets without any apparent reason, or by the tester without any apparent reason. The mean conservation rates were 73%, 67%, and 21% respectively. Implications of these results were discussed concerning children's understanding of the significance of transformation in a conservation task.