Asteroid ApophisFAJARDO, Puerto Rico—A fascinating idea came up in an informal chat I had yesterday with asteroid expert Erik Asphaug of the University of California, Santa Cruz. The early solar system was a veritable shooting gallery. Our moon is thought to have formed when a Mars-size body hit Earth and threw out a cloud of debris that coalesced in orbit around our young planet. In his talk Monday at the annual Division for Planetary Sciences meeting here, Asphaug reported that the incoming body had to hit at a fairly low velocity. Any faster, and the debris would have scattered into interplanetary space. In that case, whatever body coalesced would not have been a moon, but a planet in its own right. Riffing on his talk, Asphaug has a provocative answer for one of my favorite questions in planetary science: Why doesn't Venus have a moon? How did it manage to dodge all the bullets flying around the early solar system? Asphaug suggests that maybe it didn't. Maybe Venus got hit worse than we did, so that a planet rather than a moon was the outcome.

And where is that planet now? Maybe Mercury is it. Others, too, have proposed that Mercury formed from a collision on the early Venus or even a second giant impact on Earth.

Speaking of collisions, scientists here have offered both bad news and good news about how likely we are to get pummeled by an asteroid. Yesterday morning, Renu Malhotra of the University of Arizona looked to lunar craters for clues to the impact risk. In 2003, researchers using data obtained by the Clementine spacecraft in 1994 discovered a peculiar fact about craters on the moon: the youngest, whose bright debris deposits give them away, are not scattered uniformly over the surface. Rather, they cluster on the leading hemisphere—that is, the side that points in the direction the moon is moving in its orbit around Earth. If you go out and look at the moon, it orbits toward your east, and there are 70 percent more young craters on the east side than on the west. At first blush, this sounds unsurprising: moving through the solar system is like driving in a hailstorm, and you'd expect more hail to hit your windshield than your rear window.

But it's not so simple. Malhotra and her colleague Takashi Ito of the National Astronomical Observatory of Japan ran computer simulations using the known population of near-Earth asteroids and estimated there should be only a 30 percent difference in the collision rate. Their conclusion: other asteroids are lurking out there. "We have to consider the intriguing possibility there's an unseen population of low-velocity impactors," Malhotra says. And if those bodies pelt the moon, Earth is in trouble, too.

The following talk, by Malhotra's graduate student David Minton, offered a happier thought. He focused on asteroids of the size that took down the dinosaurs and would do the same to us. In his simulations, the planets' gravitational influences remove these bodies from the asteroid belt early on, leaving fairly few to pose a risk to us today. Dinosaur-killers should strike only once every billion years, much less often than the usual estimate of every 100 million years. During the comments period after Minton's talk, Bill Hartmann of the Planetary Science Institute in Tucson, Ariz., said that Martian cratering indeed shows signs of having slowed down over time.

Impact expert Alan Harris of the Space Science Institute in Boulder, Colo., was skeptical about both of these talks. He says that the 100-million-year rate is estimated directly from the observed number of such bodies. "What you see is what you get," he remarks. As for the lunar-cratering asymmetry, Harris says that fresh craters are hard to identify and that the discrepancy Malhotra sees is well within the noise.

More good news/bad news about asteroids came today. Steve Chesley of the NASA Jet Propulsion Laboratory in Pasadena, Calif., and David Tholen of the University of Hawaii presented updated estimates about the most significant specific threat to our planet: that the asteroid 99942 Apophis will strike in the year 2036. This 200-meter-wide asteroid will come distressingly close on April 13, 2029—within just 39,000 kilometers—but the orbital uncertainties for that date are very low and scientists have ruled out a collision then. They are less sure about 2036. The asteroid can hit if its trajectory is just right—specifically, if its 2029 near miss brings it within 600 meters of a certain location in space that astronomers call the "keyhole." Data released yesterday reduce the chance of that from 1 in 45,000 to 1 in 250,000.

The bad news is that there's now a chance of a collision in 2068. That collision, too, would require Apophis to pass through a keyhole in 2029—and Chesley reports that the asteroid is aimed straight for it. The saving grace is that the keyhole is just two meters wide, so the probability of a collision is less than one in a million. Clearly, though, we'd better keep an eye on this beast. Scientists plan to ping it with the Goldstone planetary radar in 2012 and the Arecibo Observatory radar in 2013.

Moreover, Harris told the meeting that asteroid surveys have so far found only 40 percent of the estimated population of Apophis-size bodies. Eventually scientists should have a complete census of civilization-killing asteroids. Unfortunately, that leaves smaller ones that could still take out a country or city. Harris says it's much more probable that a 25-meter asteroid will hit next week than some unknown Apophis-size asteroid will hit later this century. "It's 10 times more likely that an unknown asteroid will slam into us from behind while you're looking at Apophis," he reports.


From carbon planets to the lakes of Titan: Dispatch from the annual planets meeting

What caused Saturn to lurch? Second dispatch from the annual planets meeting

Image of asteroid Apophis (circled): UH/IA