It's long been theorized that the moon was created when a Mars-sized proto-planet slammed into the Earth. Unfortunately, though, the chemistry of the two bodies has stubbornly failed to bear this out.
Now, though, two very different models of such a collision have been developed, both explaining satisfactorily - but incompatibly - how the Earth-moon system ended up as it did.
The big problem with current models is that they predict that the Earth and moon should have different oxygen isotope compositions, as much of ther material for the moon would have come from the impacting body, sometimes known as Theia. Observations, though, have shown that they don't.
But a new model developed by the Southwest Research Institute (SwRI) matches reality rather better, by assuming that the impactor was much bigger than previously believed.
In this scenario, both the impactor and the target are about the same size, with each about four to five times the mass of Mars. This would lead to a disk of debris with a smilar composition to the Earth, eventually coalescing to form the moon.
This theory doesn't explain everything, as it produces an Earth that is rotating at least twice as fast as is implied by the current angular momentum of the Earth-moon system.
However, in a separate paper also published today, Dr Matija Ćuk of the SETI Institute and Dr Sarah Stewart of Harvard have an explanation.
They show that a resonant interaction between the early moon and the sun - known as the evection resonance - could have decreased the angular momentum of the Earth-moon system by just the right amount soon after the impact that formed the moon.
"By allowing for a much higher initial angular momentum for the Earth-Moon system, the Ćuk and Stewart work allows for impacts that for the first time can directly produce an appropriately massive disk with a composition equal to that of the planet's mantle," says Dr Robin Canup of SwRI.
There is, though, another possibility implied by the model, involving a much smaller, high-velocity impactor hitting a target that is rotating very rapidly because of an earlier impact.
Meanwhile, though, Washington University in St Louis planetary scientist Frédéric Moynier says he's found isotopic differences between the Earth and moon after all.
His team has analyzed 20 samples of lunar rocks from four different locations, as well as one lunar meteorite, and found that they contain much lower levels of zinc than Earth, but are enriched in the heavy isotopes of zinc.
The implication is that conditions during or after the formation of the moon led to more extensive volatile loss and isotopic fractionation than was experienced by Earth; and the simplest explanation is wide-scale melting.
However, this doesn't help solve the problem of which - if either - computer model is correct.
"The ultimate likelihood of each impact scenario will need to be assessed by improved models of terrestrial planet formation, as well as by a better understanding of the conditions required for the evection resonance mechanism," says Canup.