One of the great mysteries of our Solar System is why Uranus is tilted on its side. Surely, if the solar system formed from the same rotating cloud of dust and gas, then all the bodies within it should rotate in the same way. And yet Uranus' axis of rotation lies at 97 degrees to the plane of the solar system.

The standard explanation is that Uranus must have been involved in some kind of interplanetary collision with and earth-sized protoplanet in the early days of the solar system. That's a tempting idea but it has some shortcomings. For example, it doesn't explain why the orbits of the moons of Uranus are similarly tilted, not that of its rings.

Today, Gwenael Boue and Jacques Laskar at the Observatoire de Paris in France put forward another idea. They say that Uranus may have become tilted during the period soon after formation when the planets were migrating to the orbits we see now. They point out that the presence of satellites around a planet can increase its rate of precession, if it has a high initial inclination of more than say 17 degrees. This increase can be by as much as a factor of 1000 if the mass of the moon and the radius of its orbit have certain values. For Uranus, this is for a moon of 0.01 Uranian mass and at 50 Uranian radii.

The problem, of course, is that Uranus does not have such a moon. Its most distant companion is Oberon with a mass of just 10^-5 Uranian masses and an orbit of 23 Uranian radii.

Boue and Laskar's idea is that Uranus once had a moon of the required size and orbit, which caused the planet to tilt during the planetary migration, but that this moon was ejected during a close encounter towards the end of the migration.

To study whether this idea is feasible, they simulated the process of giant planet migration in the early solar system some 10,000 times. They then discarded all scenarios in which the planets collided or did not end up in the correct final order. They then selected only those outcomes in which Uranus had an inclination of more than 17 degrees and also rejected any simulation in which Uranus came within 50 Uranian radii of another planet, since that would be likely to eject Oberon as well as the additional hypothesised moon. That left 17 simulations.

Boue and Laskar then added the additional moon to see how it would effect the tilt of Uranus and repeated each of these 17 scenarios a further 100 times. In 37 cases, the new moon helped Uranus onto its side and then ended up being ejected after a close encounter with another gas giant.

That's an interesting result and not just because of the tilt: some models of planet formation predict that Uranus ought to have had another moon (albeit somewhat smaller than the one Boue and Laskar introduce). Consequently, this idea has the elegant property of explaining two mysteries for the price of one, never a bad thing in science.

Ref: http://arxiv.org/abs/0912.0181: A Collissionless Scenario For Uranus Tilting