A slow splat. A relatively slow impact by a companion body a third the size (and 4% the mass) of the moon may have created the moon's far side highlands as well as the near side's "Man in the Moon."

Martin Jutzi and Erik Asphaug

Are Mars's moons homegrown—or snatched from the asteroid belt?

Long after astronomers found moons orbiting other planets in our solar system, Mars remained a loner. It wasn’t until the late 1800s, when astronomer Asaph Hall tried, failed, and then—at the urging of his wife—tried again, that scientists got their first peek at the Red Planet’s two tiny moons, which Hall named Phobos and Deimos.

A century later, spacecraft images revealed that the moons look like asteroids—dark, crater-pocked, and potato-shaped—suggesting Mars had snatched them from the nearby asteroid belt. Now, planetary scientists have conducted the first computer simulations that bolster a controversial alternative idea: The satellites formed as our own moon did, after a big object smashed into the planet and kicked up debris.

Although the impact hypothesis has been around for decades, it has been hotly contested. "It was embarrassing how badly people came after me," says geologist Robert Craddock of the Smithsonian Institution’s National Air and Space Museum in Washington, D.C., who presented the idea at a 1994 conference. "The general tone was this couldn't possibly work." He submitted a scientific paper to several journals, including Nature and Science, which, he says, all rejected it. So, like Hall before him, he gave up. Years later, Craddock resurrected the paper and published it in 2011 after another scientist became interested in his work—and after some of his critics had retired or died.

That once-rejected work has now inspired two teams of planetary scientists to model what might have happened if an object roughly 10 times as massive as Ceres, the largest asteroid, smashed into ancient Mars. Objects of this size were flying around the early solar system, and in Craddock's 2011 paper this one giant stone killed three martian birds. First, it spun the planet up, making it turn once every 24 hours and 37 minutes, matching the length of a martian day; without this impact, the planet would turn more slowly. Second, it produced one of the planet's largest impact basins, measuring many thousands of kilometers across. And third, it kicked up martian debris, which formed a disk around the planet that then created the moons.

Planetary scientists Robin Canup and Julien Salmon of the Southwest Research Institute in Boulder, Colorado, presented their simulations at a conference late last year. And planetary scientists Robert Citron of the University of California, Berkeley, and Hidenori Genda and Shigeru Ida of the Tokyo Institute of Technology published their own calculations in the 15 May issue of Icarus. Although the two teams worked unaware of each other, they reached the same conclusion: A large impact can indeed form both moons. In particular, the disk of debris the impact creates was compact, hugging close to the surface. That's important, because both moons are much closer to Mars than our moon is to Earth.

Both teams also find that the moon-spawning disk was much more massive than Craddock had estimated. As a result, Canup suspects the disk bred a third moon much larger than the two we see, estimating that its diameter was roughly 300 kilometers. That's only a tenth as big as our moon but dwarfs Phobos and Deimos, which are just one or two dozen kilometers across.

What happened to it? Canup suspects it crashed into Mars. Because of gravitational interactions between a moon and its planet, any moon gradually moves either farther from or closer to the planet. If a moon revolves more slowly than its planet spins—as both Deimos and our moon do—it moves outward. But if a moon revolves faster than the planet spins—as Phobos does—it moves inward and the planet's gravity tears it to pieces, which smash into the surface below. That's the unpleasant fate that awaits Phobos in roughly 40 million years. Canup thinks the same thing happened to the large moon Mars once had. "It probably only lasted a few hundred million years," she says.

"I think they're absolutely on the right track," says William McKinnon, a planetary scientist at Washington University in St. Louis. "It makes it much more believable to have a simulation that actually works." McKinnon says that if the moons are really captured asteroids, they should have elliptical orbits far from the planet, but they actually have circular paths close to the planet, as expected if the satellites formed by impact.

But other experts remain skeptical. Scott Murchie, a planetary scientist at Johns Hopkins University in Baltimore, Maryland, who has been studying the martian moons for more than a quarter-century, says the new work is valuable. But, he adds, "I think by far the interpretation most consistent with observable characteristics of the moons aside from their orbits—[their] density, spectral properties, [dark color]—would be that they're captured asteroids." The new simulations don't address the moons' composition—only their sizes and orbits.

All the scientists agree on how to resolve the controversy: Send a spacecraft to the moons. A moon-spawning collision should have vaporized water ice and hydrogen, leaving the satellites with none. "Boy, if they have substantial internal ice, I think that would be the strongest argument against impact," Canup says.

The Soviet Union sent landers to Phobos that failed shortly before their country did. Scientists in the United States have recently proposed missions of their own—for example, landing a probe to scrutinize the small moon's minerals—but NASA has not yet decided whether to support any of them.