The cataclysm that made the man in the Moon began in the far reaches of the Solar System, says Ivan Semeniuk

   THIS MUCH we know: the Solar System is not a safe place. Sixty-five million years ago, the dinosaurs were clobbered by a wayward asteroid or comet. A hundred and eighty million years earlier, a similar event appears to have swept the trilobites and most of their contemporaries into the dustbin of prehistory. And a far greater assault occurred just as life was gaining a foothold.

   There is new evidence that a sudden barrage of deadly debris crashed against the Earth and Moon 3.9 billion years ago. Thousands of giant impacts pummelled the Earth--most as big as the event that wiped out the dinosaurs, and some much larger. Those vast impacts left behind continent-sized craters and liberated enough heat to vaporise oceans.

   What triggered this onslaught? "Something in the structure of the Solar System must have changed," says Harold Levison of the Southwest Research Institute in Boulder, Colorado. Levison has his own personal Solar System, a virtual model, which he is using to simulate those cataclysmic events. And he is pointing the finger at two unlikely instigators: Uranus and Neptune. If he's right, these two distant giants caused the worst assault on Earth's surface since life began.

   The evidence of an ancient barrage has been staring at us for aeons. Even a casual glance at the full Moon reveals the image of a hollow-eyed face, frozen in a Munch-like scream. The tortured expression of the man in the Moon is actually an arrangement of giant circular impact basins. These "seas" are filled with dark lava that stands out against the rougher and brighter highlands. The basins are the most obvious sign today of giant impacts in the Moon's distant past.

   The bodies that created these seas were probably at least 50 or 100 kilometres across, travelling at tens of kilometres a second. They would have excavated, melted and scattered vast amounts of debris across the lunar surface, leaving a gigantic crater. Later, volcanic activity below the surface flooded many of these low-lying basins with darker material.
2
   The fact that there were giant impacts isn't too surprising. After all, rocky moons and planets are all made of smaller rocks that collided and coalesced. It is reasonable to assume that these collisions continued for some time after the Solar System formed, gradually tapering off as available material was absorbed or parked in stable orbits. On Earth, the combination of oceans, atmosphere and tectonic activity wiped out all visible traces of these formative collisions, but on the unaltered crust of the Moon the scars accumulated.

   Formative collisions

   But this picture changed after Apollo. The Apollo astronauts bobbed across the lunar terrain in search of geological souvenirs, and returned hundreds of kilograms of rock to Earth. Geochemists then dated these rocks by extracting argon-40. This element is produced by the radioactive decay of potassium, building up slowly and steadily inside the lunar rocks. But if a rock is heated to its melting point, any argon-40 in it is released as a gas. So the amount of argon-40 tells you how much time has passed since a rock last melted.

   Lunar scientists had believed in a gradual decline of collisions, so they expected the ages of lunar rocks to cluster towards the time of maximum bombardment, namely right after the Moon's formation 4.5 billion years ago. But most of the Apollo rocks proved to be about 3.9 billion years old--more than half a billion years younger than the Moon. Only a few, from the Apollo 16 site, were closer to 4.5 billion.

   As most of the Apollo rocks came from the giant impact basins, it appeared that the basins did not accumulate gradually, but instead were created long after the Moon formed, in a period now called the "late heavy bombardment". It's hard to see how this could have been the tail end of the process that formed the Moon, says Graham Ryder of the Lunar and Planetary Institute in Houston, Texas. "That would mean you'd have to have at least that amount of impacting--probably an increasing amount--prior to 3.9 billion years ago," he says. To Ryder, that's just too much bombardment; a sudden cataclysm around 3.9 billion years ago seems more likely.

   But other researchers were less sure, pointing out that the Apollo samples are biased towards the equator of the Moon--so they might only be giving us the history of a few impact basins, rather than the full story.

   Then last December, Barbara Cohen of the University of Tennessee, Knoxville, published her investigation of a group of lunar meteorites--rocks from the Moon that have landed on Earth. Cohen chose meteorites that differed in composition from the Apollo samples, so that they should have a wider range of origins. "We went in expecting to ref-ute the cataclysmic theory," says Cohen. But they couldn't. None of the lunar meteorites could be dated before 3.9 billion years ago.

   So it seems that the Moon was subjected to an intense bombardment around 3.9 billion years ago, lasting perhaps 100 million years. Planetary scientists are hard pressed to explain what caused it. "This is one reason why many people are uneasy with the idea of a cataclysm," says Ryder.

   As the only concrete evidence comes from lunar samples, the fateful event might have been confined to the Earth-Moon system. Perhaps Earth started off with a second moon, or a loose cluster of moonlets. As moon number one--the Moon we know today--moved outwards in its orbit, it could have destabilised the orbits of the other moons, so that they came close to Earth and got torn into fragments by our gravity.

   But there are doubts that this can deliver enough hits quickly enough to account for the cataclysm. Moreover, says Levison, there are indications that the late heavy bombardment was more widespread. Huge craters similar to the Moon's impact basins can be found on Mars and Mercury. And meteorites from Mars and the asteroid Vesta show signs of heavy impacting around 3.9 billion years ago.

   What could have unleashed such widespread devastation? One theory is that two huge asteroids collided and scattered fragments throughout the inner Solar System. But to explain the number and size of the impacts, these asteroids would have had to be an unlikely 10,000 times as massive as the whole asteroid belt today. Levison thinks we must look further afield for the culprit.

   The two outermost giants in our Solar System have a reputation for unusual goings on. Uranus spins on its side, presumably because of a collision that knocked it over. Miranda, a satellite of Uranus, has a crater so huge that the collision that formed it must have nearly blasted the moon to bits. Neptune's largest moon, Triton, orbits backwards in relation to the rest of the Solar System. "It all suggests there were large bodies roaming around out there in the past," says Joseph Hahn of the Lunar and Planetary Institute. Large bodies with a violent streak, what's more. But what could all this have to do with our Moon?

   In 1975, George Wetherill of the Carnegie Institution of Washington proposed that leftovers from the formation of Uranus and Neptune caused the late heavy bombardment. He reasoned that Uranus and Neptune might have formed significantly later than the other planets. These planets are thought to have grown from a swirl of icy bodies called planetesimals. Out beyond the orbit of Saturn this raw material was thinly spread, and might have taken hundreds of millions of years to gather into planets. "People who work with simulations tend to fail to produce Uranus and Neptune in what might be regarded as a reasonable amount of time," says Hahn. "People are still arguing about how much time you really need to form these planets."

   Whenever they finally formed, the gravity of these new planets would have catapulted leftover ice and rocks in all directions. Some of this material would have been kicked out to the limits of the Sun's gravitational influence, joining the Oort Cloud, which occasionally sends us comets today. Other chunks of ice would have been diverted inwards, to wreak havoc on the inner Solar System.

   When Wetherill tested the idea with a computer simulation he found he could reproduce a kind of late bombardment, but the timing was wrong. Instead of a narrow spike 3.9 billion years ago, the bombardment was staggered over a much longer interval. That made the Uranus-Neptune theory a poor match for the measured ages of the lunar samples.

   A generation later, Wetherill's idea has been reborn. In 1995, Levison was using a simulation of the early Solar System to study the formation of the Oort Cloud. Unlike Wetherill's original model, which relied on statistical approximations to predict where debris ends up, Levison's is a true simulation. At any given moment, it calculates the gravitational pull on each individual object and how that object moves, then it steps forward in time by a short interval and does the entire calculation again.

   Levison's simulation followed comets ejected from the neighbourhood of Uranus and Neptune. To his surprise, a sudden surge of these icy bodies was sent spiralling down into the inner Solar System, lasting about as long as the late heavy bombardment. "I realised I could get the narrow spike," says Levison.

   Excited by this result, Levison set about creating a new series of simulations, this time with the late heavy bombardment in mind. The key question was no longer the duration of the bombardment, but its intensity.

   The odds of any one comet hitting the Moon are low, so to plaster the Moon with impact basins requires a vast number of comets. But planetary scientists think the total mass of planetesimals beyond Saturn was less than 50 times the mass of the Earth, or else Neptune and Uranus would have grown larger than they actually are.

   Fortunately, Levison doesn't need too much raw material. In a paper to be published in the June issue of Icarus, he says that his simulation would produce a cataclysm on the Moon as long there were at least 32 Earth masses of planetesimals available. "If I needed a thousand Earth masses, or ten thousand, I knew no one would believe it," says Levison. "When the number 32 jumped out of my computer I said to myself, 'This is it! This has got to be right!'"

   The strange picture that Levison's model paints is of an early Solar System that ends at Saturn. Uranus and Neptune struggle into existence more that half a billion years late, emerging out of the icy flotsam and jetsam beyond Saturn's orbit. And once they have grown to a critical size, after perhaps 700 million years, their gravity becomes strong enough to cast icy planetesimals inwards. Twenty or so carve out the shadowy face on the Moon, and many more hit the Earth.

   For Levison, the best test of this model will be whether we find evidence for cataclysmic impacts beyond the inner Solar System. "On its way in, this stuff should have knocked the hell out of Jupiter's satellites," he says. The Jovian moon Callisto has a surface mostly made of ice, and Levison's version of the late heavy bombardment should have melted it. "If we find that Callisto never melted, that could rule the model out."

   Levison notes a curious side effect of his idea. All these icy comets would have dumped enough carbon dioxide on Mars to produce a thick, insulating atmosphere that would have allowed liquid water to exist on the Red Planet's surface. Photographs taken in the past year by the Mars Global Surveyor spacecraft do show apparently water-carved landforms, raising scientists' hopes that life once existed there. So while the missiles flung by Uranus and Neptune battered early organisms on Earth, they might have allowed a brief flourishing of life on Mars.

   Life in hell

   The Moon may still bear the scars of the late heavy bombardment, but it was our home planet that got the worst of it. With its larger size and mass, Earth would have attracted at least ten times as many impacts as the Moon, no matter what the cause of the bombardment. "Earth got beaten up, there's no doubt about it," says Hal Levison.

   Yet some traces of microbial life date back to the time of the bombardment, suggesting that the very first life forms were around earlier still. So could life have survived 200 impacts, each big enough to boil an ocean?

   "These are some really catastrophic scenarios," says Barbara Cohen of the University of Tennessee, Knoxville. But even after an impact of the most destructive kind, she thinks, conditions on Earth could have got back to normal within 10,000 years--roughly the average time between hits. "Maybe life could go dormant for that long and survive, perhaps in a spore form," she says.

   Or maybe it found places to shelter. The ocean floor may have provided the best haven from the hellish conditions closer to the surface. If so, we may have evolved from organisms that once thrived around deep-sea vents.

   Ivan Semeniuk is a science writer and broadcaster based in Toronto