Is Earth locked into an extinction cycle? Are we really under threat from a meteorite or asteroid shower? David Raup discusses the work that suggests such events might happen at regular intervals

As a geology student at Harvard in the 1950s, I learned much about the history of the Earth and how to interpret it. I learned that geologic features develop gradually and slowly by processes readily observable today and that the same is true for the evolution of life. Continents certainly do not move: the old idea of continental drift is physically impossible.

And rocks surely do not fall out of the sky. Catastrophic change is anathema. All this was part of a strong paradigm that had served geology and palaeontology well since the great 19th-century geologist, Sir Charles Lyell, taught us how to look at Earth history. And the Lyellian paradigm really did work well to explain the origins of many, perhaps most, features of the Earth.

It was thus rather shocking when, in 1980, a team led by an experimental physicist and Nobel laureate, Luis Alvarez, claimed that a large comet or asteroid struck the Earth 65 million years ago and was responsible for a great mass extinction - an extinction that killed off all dinosaurs and about three-quarters of all other species on Earth.

ÖйúAƬ

The Alvarez proposal set off a firestorm of controversy. If a large body had hit the Earth, where was the crater? If the environmental shock was severe enough to kill off so many species, how did any living thing survive? Besides, the dinosaurs and many other victims of the mass extinction had been in decline for several million years before the alleged event. Did they know the impact was coming? But thanks largely to the controversy, geologists and palaeontologists around the world weighed in with research projects to learn more about the critical rock interval surrounding that extinction.

Now, 20 years later, hundreds of research reports have established what appears to be overwhelming evidence, from many different sources, of an impact almost precisely 65 million years ago. Even the probable crater has been found buried under Yucatan and the adjoining Gulf of Mexico. The causal link to mass extinction is still inference, of course, since no dinosaurs have been found impaled by meteorite fragments, but the coincidence in timing is compelling, especially since intensive field work has shown that most of the animal groups were actually thriving right up to the end.

ÖйúAƬ

We have also learned that comets and asteroids hitting Earth are much more common than originally thought. More than 100 impact craters have been confirmed. This in turn has raised the possibility that impacts may have had more influence on the history of life than we had thought.

There have been six mass extinctions such as the one that killed the dinosaurs and some of these are used routinely by geologists to mark the boundaries between geological time periods. But how many of these might also have been triggered by the impact of comets and asteroids?

This question came to the fore in 1984 when Jack Sepkoski and I published a paper in Proceedings of the National Academy of Sciences (USA) presenting a statistical analysis of the extinction record of the past 250 million years. We concluded that the principal extinctions were evenly spaced, occurring every 26 million years. This really upset our colleagues in geology and palaeontology. Criticisms of our data and methods came thick and fast. Many of the criticisms were indeed quite valid. Our data were limited and it is true that statistical claims of cycles in nature often turn out to be artefacts of poor data, inappropriate methods or even wishful thinking.

Physicists and astronomers, on the other hand, were enthusiastic about our claim of a 26-million-year periodicity and immediately launched attempts to find an acceptable mechanism. The periodic extinctions, if real, did not have to be caused by showers of comets or asteroids but this was certainly the best bet, given the apparent success of the Alvarez explanation of the dinosaur extinction.

About two months after our paper was published, five research papers on the subject appeared in a single issue of Nature. Two of these reported that the ages of impact craters are also periodic, with intervals close to our 26 million years, and one reanalysed our extinction data, coming up with 30 instead of 26 million years.

Several of the papers proposed solar system or galactic explanations. Two argued that the normal oscillation of our solar system through the plane of the galaxy causes perturbation of comet orbits every 30 million years or so, in turn increasing the likelihood of stray comets hitting Earth. Two others postulated that our Sun is actually part of a binary system, with a small companion star in an orbit that disturbs Oort Cloud comets every 26 million years, thus producing comet showers on Earth. The companion star was dubbed Nemesis after a rather nasty Greek goddess. About a year later, another idea was published in Nature, that periodic comet barrages might also be produced by an unseen tenth planet in our solar system (Planet X) if its orbit was just right.

ÖйúAƬ

Debate among astronomers raged for several years over these proposals and occasional research papers still appear from time to time. The companion star, Nemesis, has never been seen although searches are ongoing and the automated sky surveys now getting started may well find Nemesis if it is there. The stability of the Nemesis orbit over many millions of years has been challenged and other details of the several mechanisms have been found wanting by some analysts. Even the name Nemesis has been challenged.

Stephen Jay Gould, in a Natural History essay, found the name most inappropriate: Nemesis was known for targeting the vain and powerful whereas the mass extinctions appear to be nearly indiscriminate in their killing. Gould suggested the name be changed to Siva, a god who "does not attack specific targets for cause or for punishment". It is amusing that so much debate and solid research has been devoted to a star that has never been seen and may well not exist.

ÖйúAƬ

The debate over periodic extinction and Nemesis has been seen by some as science at its best: specialists in disciplines as far apart as astrophysics and palaeontology working together to come up with different ways of looking at the natural world through the medium of peer-reviewed journals. But to others, it is science gone mad.

Where are we now, almost 20 years after the rash proposal that the dinosaurs were killed by a large rock falling out of the sky? The role of large-body impact in wiping out the dinosaurs is broadly accepted in the scientific community, although there is the ever-lurking possibility that the mass extinction was actually caused by the unusual vulcanism at that time.

On the question of a 26 or 30-million-year extinction cycle, the jury is still out. The original extinction data have been reanalysed independently 13 times in published papers: five found statistically acceptable periodicity whereas eight did not. From 1984 until his death last year, Sepkoski increased the size of the extinction data base tenfold and the periodicity looks, to me at least, far better than it did before.

The Nemesis, Planet X, and galactic oscillation ideas continue to be debated in the astronomical literature, although more sporadically than before. The important thing is that the theory is still in the consciousness of the scientific community, ready for new and independent evidence to appear that can settle the argument one way or the other.

The work started by Alvarez and his ilk has given us a much broader view of environmental history and the pace of change. We now know that the Earth is not a closed system and that sudden, rare events can have great effects.

In retrospect, I wonder how we could have expected that in the few thousand years we have watched and recorded events on Earth we could have gained a complete picture of natural phenomena covering hundreds of millions of years.

ÖйúAƬ

David M. Raup is emeritus professor of geophysical sciences, evolutionary biology and conceptual foundations of science at the University of Chicago. His bookThe Nemesis Affair is published next month by WW Norton, priced Pounds 8.95.