This undated artist rendering shows the twin Grail spacecraft mapping the lunar gravity field. The two probes are scheduled to enter orbit around the moon over New Year's weekend. (AP/NASA)

She hangs there nightly, a yellow or white or spookily orange disk, the bringer of tides, the caster of romantic shadows. She waxes and wanes and sometimes she turns ruddy as the shadow of the Earth crosses her face. For all her beauty, though, our moon hides a lumpy, unflattering secret: She’s lopsided. Her backside is much thicker than her front. And no one knows why.

It’s unseemly, really. After more than 100 robotic and human missions to the moon, scientists still can’t account for why one half — the half we can’t see — is taller than the other.

Twin NASA probes that arrived at our satellite this weekend may finally reveal a shocking truth: that early on, a smaller twin moon smushed into her. As this intruder splatted into its big sister, it shattered “like a mega-avalanche,” said Erik Asphaug, the planetary scientist at the University of California at Santa Cruz who published the twin-moon idea in the journal Nature in August. His co-author was Martin Jutzi of the University of Bern in Switzerland.

This collision would have spread a wide hump of rock onto the back of the moon. There, the material cooled and hardened into a thick crust: the far-side lunar highlands.

“This is one of those ideas that all sorts of people will try to prove wrong,” said Maria Zuber, the MIT scientist heading up the new NASA moon mission. “But it’s extremely testable.”

And so GRAIL will test it. Designed to probe the moon’s interior, the two washing-machine-size spacecraft will reveal the thickness of the moon’s crust, its topmost layer.

If the two-moon theory is correct, the backside crust will be much thicker than that of the front side. The hump should taper toward the equator.

GRAIL could also spot another hidden feature predicted by the theory. If a second moon did crash into the first, the collision would have occurred when the big moon was young and hot. A thin layer of molten heavy elements including uranium and potassium still burbled just under the crust.

The backside impact would have squeezed this liquid, pushing it around to the front side. There it would have cooled and hardened, leaving a telltale layer.

The existence of both features — a thick backside crust and a thin, dense layer under the front’s crust — would offer strong support for the twin-moon theory, Asphaug said.

When Zuber first heard the notion, she scoffed. “This is going to be nonsense,” she recalled thinking. But computer simulations run by Asphaug and Jutzi were compelling leading Zuber to reverse course. “It’s a plausible scenario,” she said.

The idea is also simple, another stroke in its favor. By contrast, other explanations for the moon’s front-back discrepancy tend toward the complicated and unsatisfying.

“There are all these theories out there,” Asphaug said, “that have big warts on them.”

Such as: Maybe the front side of the moon was terribly unlucky, flattened by seven or eight big space rocks. The problem: Asteroids and comets arrive from all directions; there’s no reason impacts should cluster. “It’s like flipping a coin and getting heads eight times,” Asphaug said.

Another theory suggests that the backside hump is a tidal bulge. Planets and moons sport such bulges when they get tugged at — and the Earth tugs on the moon a lot. The problem: Tidal bulges tend to be symmetrical, so there should be a bulge on both sides of the moon.

Asphaug’s theory requires a very specific sequence of events some 4.5 billion years ago, when the infant Earth was a molten ball.

Long before life appeared, rocky debris ricocheted around the early solar system. Something the size of Mars plowed into the Earth, sending huge globs of molten material hurtling into space. The largest glob coalesced into the moon. This catastrophic-impact theory of moon formation is widely accepted by scientists.

To that, Asphaug and Jutzi threw in a twist: What if a second, smaller glob of Earth-stuff also got blasted free? If it launched at a particular angle, the glob would have coalesced into a second body and drifted behind the moon in roughly the same orbit.

After a few million years, the pull of the sun would have drawn the smaller moon closer to the bigger moon. Eventually, the two bodies collided — in slow motion. A fast collision would have excavated a giant crater. But a slow collision — just the type predicted by the computer simulations — would have pancaked the small moon onto the surface, leaving evidence for GRAIL to spot.

It’s a quirk of happenstance that GRAIL will be able to test the theory at all. Zuber proposed the $400 million mission five years ago, long before Asphaug and Jutzi published their idea. Zuber wanted to probe other, more general questions: Does the moon have a solid core? How long did the moon take to cool after it formed? And did the moon once have a magnetic field?

“You might think we already know all there is to know about the moon,” said Zuber. ”Of course, that’s not the case.”

The twin GRAIL probes arrived in a high lunar orbit this weekend, but they won’t begin collecting data until March.

By then, thrusters will have dropped the pair to just 35 miles above the surface. Flying in formation — one ahead of the other — the probes will map minute fluctuations of the moon’s gravity over its entire surface. This new gravity map will be 100 to 1,000 times as accurate as current maps. From it, scientists will infer the internal structure of the moon “from crust to core,” Zuber said.

Asphaug said there’s an even better way to test the long-shot idea, though it’s one that GRAIL can’t carry out: Study rocks from the far side of the moon. The Apollo astronauts collected hundreds of pounds of moon rocks — but all of them came from the Earth-facing side.