Understanding what possibilities are involves being able to tell the difference between things that could happen and things that can’t happen. How does the ability to distinguish the possible from the impossible emerge over ontogenesis? The current study modified the “Y-shaped tube task” (Redshaw & Suddendorf 2016) to test children’s ability to distinguish the possible from the impossible. In the Y-shaped tube task, the experimenter holds a ball above a tube shaped like an upside-down “Y” and asks a participant to catch it. A participant who sees that there are two possible places for the ball to come out should cover the openings under each branch of the Y. But children might cover both openings for other reasons as well: for example, because there are exactly two good places to put one’s hands. If children cover both openings because they understand that the ball could come out either way, then they should also find efficient, effective solutions to two new trial types presented in this paper. The “U-shaped” tube is shaped like an upside-down “U”, with an opening at the end of each leg. However, the U-shaped tube was designed with one obviously impossible opening. In order to come out from the impossible opening, a ball would need to levitate upward and pass through a block of solid wood at the top. If children can distinguish possibilities from impossibilities, they should only cover the possible opening. The “Critical tube” has three openings at the bottom, but only two are possible; again, in order to come out from the impossible opening, a ball would need to levitate upward and pass through a block of solid wood. If children can distinguish possibilities from impossibilities, they should cover both possible openings, avoiding the impossible opening. Twenty four 36-month-olds and 24 48-month-olds were tested on all three tubes. Results indicate that less than 20% of 36-month-olds and at most about half of 48-month-olds distinguish the possible from the impossible. This result converges with several other findings in the literature that suggest that the ability to choose actions that account for multiple incompatible possibilities often emerges after the fourth birthday.