The idea that an asteroid impact 65 million years ago killed off the dinosaurs and roughly three out of four other species then living is now firmly lodged in the scientific mind and the public imagination. The evidence for the impact is overwhelming. The huge crater it left has been discerned beneath thick sediments in Mexico, the mark of its tidal waves can be seen in the middle of America and its distinctive debris and ash has been found all over the world. The death of the dinosaurs doesn't keep resurfacing because there is a hot debate or a lively controversy; it does so because it's just such a cracking story. New wrinkles to the tale, such as last week's announcement of evidence for centuries of near-sterility in the oceans following the impact, get media attention simply because astronomy, dinosaurs and mayhem are the ideal ingredients for popular science.
But our fascination isn't quite that simple. There's something more to it: the worrying implication, raised half jokingly by the Trevor MacDonalds and Sue MacGregors when the science correspondent has done his bit, that it could happen again. This catastrophe was not a one-off. There have been many asteroid impacts in the history of the earth; left unchecked, there will be many more.
These impacts could kill billions of people, and you could be one of them. The risk of your dying in such an event is around one in 30,000, which, though small, is far from insignificant. It's certainly far greater than the chances of your winning the lottery. Like the destruction of the dinosaurs, this risk has slipped into the popular imagination, lubricated by documentaries and magazine covers. We all know of it, and we could, scientific opinion assures us, easily and cheaply do something about it. But we don't. It's a mistake that, while unlikely to prove tragic, reveals a lot about how science is used and what science means.
The truth of the matter is straightforward. The solar system is not just a neatly concentric set of planets. There are lots of smaller lumps of ice and stone and iron whirling around the sun, too. They hit the earth all the time, and they come in a range of sizes, the big ones proportionately rarer than the small. The vast majority are just pretty shooting stars - far too tiny to worry about. But watch these meteors for long enough and you will see some bigger ones. Watch for half a million years and you can expect to see one that outdoes a major nuclear war for nastiness, laying waste a continent, blacking out the sun for a year or more, blighting the world's crops.
If the risk of such an impact is 500,000 to one in a given year, then over a 70-year span, the cumulative risk to any individual of living through or dying in such an event is about one in 7,000. If such an impact leaves about 25 per cent of the earth's population dead, most of them through starvation, that gives any person a risk of about one in 30,000 during the course of an average lifetime. Some estimates make the risk smaller, and others make it larger, but that is a good average. It means that the huge unpleasantness of impacts offsets their great infrequency enough to make them roughly as dangerous as air travel, which entails a risk of about one in 20,000.
In the case of air travel, this is a level of risk people feel quite strongly about. On 13 February, President Clinton set a goal of reducing the risk of dying in an air crash by 80 per cent. Nasa, whose first A stands for aeronautics, and which thus has a thumb in the air-safety pie, will be spending about $100m a year on the project. Most people thought this investment quite wise, but a scientist called David Morrison raised an inquiring eyebrow.
Morrison, who enjoys the wonderful title "director of space" at Nasa's Ames Research Center in California, chaired a committee which produced a report in 1992 for the US Congress on the asteroid risk. Its advice was simple. There are probably a couple of thousand asteroids of the once- every-half-a-million-years, climatic-catastrophe type in earth-crossing orbits. Only a couple of hundred have been identified. Mount a thorough survey to find the rest of them, extrapolate their orbits for a few centuries in a computer, and see if one of them ever comes to occupy the same point in time and space as the earth.
If none of them is going to, that's good. And if one of them is indeed on a collision course, that's not too bad, either. Once the risk moves from the statistical to the actual, things can be done about it, especially as the survey would typically give its warning decades or centuries in advance. A nuclear explosion off to one side of the incoming rock could nudge it into an orbit that missed the earth. The technology to fly spacecraft to asteroids exists, as do the bombs. Putting them together into a successful mission over a period of years would be a tricky problem, but far less tricky than, say, waging the Gulf War.
The Spaceguard survey that Morrison and his committee suggested as a way of finding almost all the asteroids was not a huge affair. It required six specially designed telescopes of modest size operating for three decades and a data system to handle what they saw. Its costs were estimated at about $10m a year - a tenth of the price-tag for the air-safety programmes proposed two weeks ago and half a percent of Nasa's budget for space science next year.
But Nasa's big bucks, like those of its equivalents elsewhere, are fiercely fought over. They are spent on what the agencies' bosses and their beneficiaries ask for. And no-one is asking for asteroid surveys except the people already doing them, who were well represented on Morrison's committee. The focus of modern astronomy is not on the objects nearest to the earth but on those furthest away: vast black holes at equally vast distances, infant galaxies half as old as time and the fading embers of the Big Bang itself. There is, admittedly, a very small space mission heading off to a nearby asteroid at the moment, but that probably owes more to pressure brought to bear by the senior senator for Maryland, where it was built, than to a widespread scientific constituency.
When the Spaceguard report was released, Morrison pointed out to the press that there were about as many people involved in full-time searches for dangerous asteroids as there were employed in a typical McDonalds. Since then, one new asteroid surveillance system has been started in the US, largely thanks to the interest of the military, some of whom see protection against asteroids as a reasonable mission, or an excuse to try out neat weapons technology, or both. Other American searches, though, have closed down, as has the Australian programme, the only search that covered the southern skies. According to Duncan Steel, who used to work on the Australian search, there are now only about half a dozen people employed to track earth-crossing asteroids.
In short, nothing much is being done about the end of the world because it is a minority interest among scientists and no-one else feels particularly affected. For all that, asteroids are an otherworldly risk. They nicely highlight the worldiness of the relationship between science and policy. A theoretical danger can only be built into a policy-inducing risk with the help of a group of people who care about it, a constituency with a particular stake in the problem. Science simply doesn't matter much in policy debates unless there are interest groups to make use of it, lobbies with more clout than a burger-bar's worth of astronomers.
Then there's the problem of science going back on itself. People used to fear the skies, worry about Jove's thunderbolt and tremble at the sight of comets. Then the scientists took it on themselves to set the peasants right. The heavens were revealed as well-ordered clockwork, the history of the earth and life as one of slow gradual change rather than catastrophic fits and starts. By the middle of the 19th century, French astronomer Francois Arago was able to speak with pride of the fact that science had stopped people from worrying about comets, and that, as long as scaremongering journalists were assiduously slapped down, the sorry age of celestial superstition was gone for good. Science made the world seem sensible and its catastrophic demise silly. When science then comes back and says that the end of the world is, after all, a real possibility, it is not surprising that people laugh.
Sometimes, though, it's hard not to think that there is a deeper reason for "impact denial". Perhaps people do not want to see themselves connected to the universe in this sort of way. The geologists who for years resisted the impact explanation for the dinosaurs' death simply didn't want asteroids to play as big a role in the history of the earth as, say, the wanderings of one of its own tectonic plates. Tough - they do. Humans and the earth they live on are linked to the universe in all sorts of strange, indirect, unsettling ways. Worse yet, humanity now has the power to change these connections. We can empty seas and denude vast forests. We can warm an entire planet, and now, given just a little warning, we can push aside flying mountains. It's genuinely frightening to contemplate such power, especially when you realise how poorly decisions about using it are made or not made. Better to deny the risk of asteroid impact than to accept the fact the humans can redirect the stars in their courses. It's a delusion - a dangerous one, in this case - but you can understand it.