Every pre-teen in a middle school science course learns about the solar system as a neat, orderly place. Since the formation of the solar system as we know it today, the eight planets have orbited around the sun in exact patterns, in an exact order. Who doesn’t remember the mnemonic  devices? “My Very Educated Mother Just Served Us Nachos”, “Many Very Educated Men Just Screwed Up Nature”. However, it seems that the universe isn’t as precisely organized as scientists have thought.

MIT researchers have recently concluded that the solar system once acted like giant snow globe. Planets catapulted closer to the sun and then farther away, sending asteroids swirling like the white flakes contained in those glassy spheres.

Francesca DeMeo, a Hubble Fellow at the Harvard-Smithsonian Center for Astrophysics who earned her undergraduate and Masters degrees from MIT and earned her PhD at the Paris Observatory in France, conducted a study on the composition and history of asteroids in the asteroid belt. Over the course of her research, DeMeo realized that some asteroids were obviously not in the same place where they had formed. “We’d find a few asteroids, that we think formed close to the Sun, at the outer part of the belt near Jupiter,” she told Ars Technica, a technology news and information website.

DeMeo organized the tens of thousands of asteroids and categorized them according to position and composition. Traditionally, scientists have viewed asteroids as pieces of a planet that failed to form due to the pull of Jupiter’s gravity. Now, it seems that, at some point, Jupiter bowled through the asteroid belt, sweeping away asteroids and scattering those remaining.

This discovery also has implications concerning how water arrived on Earth. Scientists believe that water was brought to Earth through long-ago collisions with asteroids. If this is true, the stirring of the planets might have been essential for the presence of water on Earth, and the existence of other Earth-like planets may be less frequent  than previously hypothesized.