After the origins of humanity, the question people most like to ask about the distant past is: what killed the dinosaurs? By the end of the Cretaceous Period, 65 million years ago, they had all gone. Their disappearance has long been recognised as abrupt, at least by the leisurely standards of geological time. Nowadays, their extinction inspires and sells books and movies by the dozen. Yet for many years their disappearance was seen as something of a non-question. Textbooks concluded chapters on dinosaurs with a few desultory speculations about their demise before moving on to describe the subsequent Age of Mammals. This was partly due to lack of evidence, but it was also informed by a firm belief in evolution as a progressive force. With the inevitability of clockwork, the dinosaurs had to make way for the superior evolutionary accomplishments of mammals.
This may explain why even professionals approached their extinction as dilettantes, advancing ideas that had no scientific rigour. Dinosaurs were wiped out by epidemics of slipped discs, cataracts, toothache or parasites; they were felled by hormonal over-activity leading to bone disorders or thinning of their eggshells; they were victims of racial senescence; they were out-competed by mammals; they couldn’t digest the newly evolved flowering plants. Unconstrained by evidence, these ideas could only get sillier. Dinosaurs perished through too much sex, not enough sex, stupidity, suicide, hayfever (triggered by those flowering plants) or boredom. More serious scientists meanwhile advanced a variety of causes that had more to do with environmental circumstances than any inherent problems of dinosaurs or their evolutionary history.
Climate change was always a contender, but nobody has yet agreed on the kind of change that would have been especially uncomfortable for dinosaurs. Perhaps the climate became too hot, too cold, too dry, too wet or too changeable. Perhaps an increase in oxygen in the atmosphere led to fires, and the dinosaurs burned or choked to death. Perhaps they suffered from volcanic eruptions, mountain-building, a fall in sea-level, a shift in the Earth’s rotational poles, destruction of the ozone layer, cosmic radiation from a supernova, or the impact of an asteroid. These hypotheses had the advantage of being (in some cases) testable. They also had a more insidious effect: if dinosaurs died out because of some external change rather than any inherent flaw, the question naturally arises of whether, had this change not occurred, the dinosaurs might still be here.
Under the influence of palaeontologists such as Stephen Jay Gould, the notion of progressive evolution has given way to a more makeshift world view, in which circumstances play as great a part as natural selection in shaping the history of life. In this view, chance happenings and sudden disasters may have resounding evolutionary consequences. Such sudden disasters were meat and drink to early geologists, raised on the cramped chronology of the Bible. If the world had been created in 4004 BC, the only way to move mountains was all at once, in some divinely-ordained spasm. In the 19th century, however, geologists came to realise that the Earth was much older, allowing plenty of time for mountains to rise, and be eroded again, grain by grain. Given enough time, a rivulet could carve out the Grand Canyon.
With no need any longer to appeal to ad hoc catastrophes, 19th-century geologists such as Lyell perhaps went too far in the other direction, denying any world-shaping role for processes that could not be observed at the present time. This doctrine, ‘uniformitarianism’, came to govern geological thought: catastrophes formed no part of it. It was therefore only natural that Darwin, who read Lyell’s Principles of Geology during his five years on the Beagle, should look on evolution as a gradual process, without sudden jumps.
There is a difficulty with this view, however: the Earth’s slowly sedimenting rocks are divided into distinct Systems, each of which corresponds to a chronological Period many millions of years long. Each Period has its characteristic fossils, and is demarcated from later and earlier Periods by sharp discontinuities in the types of fossil present. These breaks are hard to equate with gradual change, except by supposing that they represent times when there was more erosion than sedimentation. The rocks that would have been laid down at what we see as the ends of a particular Period have all been eroded away, so that when sedimentation resumed, evolution had moved on to produce a new suite of fossils, giving the impression of a marked break. Some discontinuities are particularly hard to square with this view: according to Lyell, the enormous evolutionary change in fossils between the Cretaceous and overlying Tertiary Periods can only be accounted for by the erosion of sediments as deep as the whole of the Tertiary itself. In other words, by assuming that 65 million years of prehistory had just disappeared. But the continuing discussion of dinosaur extinction as a particular event suggested that sudden, unique changes could have a place in the schemes of the most die-hard uniformitarian.
Walter Alvarez was raised, like all geologists, as a good uniformitarian. In the Seventies, when astronauts were discovering a lunar landscape shattered by catastrophic impacts, he was a postdoctoral researcher at Columbia. Uniformitarian geologists didn’t ‘see’ the evidence for these impacts as their concern. Instead, the talk was all about plate tectonics – the modern equivalent of ‘continental drift’ – in which the Earth’s surface is fragmented into ‘plates’ that move in relation to one another. Although tectonic plates move with majestic slowness, given a few hundred million years they can separate to create oceans, or collide to create mountains as high as the Himalayas. Plate tectonics represent uniformitarianism on a grand scale and in the face of this triumph for Lyellian precepts, craters on the Moon (and their causes) received scant attention.
Alvarez was as taken up with tectonics as any other young geologist, and found himself in the Apennines studying a magnificent sequence of limestones, spanning the Cretaceous-Tertiary (‘K/T’) boundary. The limestones were formed on the floor of a deep ocean, since raised up by the collision between Africa and Europe: the idea was to look for clues that might reveal something of the complex tectonic history of the Mediterranean. This project failed, but yielded other dividends. Alvarez’s interest was taken by a thin layer of clay separating the underlying Cretaceous limestones from those of the overlying Tertiary. Remembering Lyell’s suspicions about the possible time-lapse between the Cretaceous and Tertiary, Alvarez wanted to find out how long it might have taken for the clay to be deposited. The answer required some ingenious physics, and Alvarez called in his father, Luis, a particle physicist, and Nobel laureate, at Berkeley. What they needed to do was measure clay deposition by some kind of steady geochemical clock. The surface of the Earth is showered with minute grains of space dust at a rate that could be calculated. The trick, Luis Alvarez suggested, would be to measure the minuscule amount of space dust found in the clay layer. Calibration with known rates would reveal how long it took for that layer to be deposited.
Space dust (from meteorites, say) may look like any old dust, but it contains a lot more of certain elements than are to be found in Earth dust. One such is the metal iridium. Any iridium found in the clay layer would act as a marker for vanishingly small amounts of space dust. The task of finding iridium was given to a pair of heroically patient laboratory geochemists, who found it at a concentration of nine parts per billion. This was almost a hundred times higher than they had expected. After ruling out other possibilities, the Alvarez team was forced to conclude that the Earth had been hit by a city-sized asteroid right at the K/T boundary. The impact would have released an energy tens of thousands times greater than the simultaneous detonation of the world’s nuclear arsenal: a bigger explosion than we can imagine. The asteroid left its mark in the iridium-enriched dust that settled out in the months following the impact. Work in places as far apart as Denmark and New Zealand revealed similar ‘spikes’ of iridium enrichment in K/T boundary sections, suggesting that the phenomenon was global.
The Alvarezes went further than the evidence strictly warranted, by supposing that the boundary and the iridium layer were more than coincidental, that the impact of the asteroid was responsible for the marked turnover in animal and plant life that many had observed at the end of the Cretaceous. The media loved it: in the public mind, the Alvarezes’ paper, published in Science in 1980, boiled down to the claim that a bomb from outer space killed the dinosaurs.
Then, however, came the longer job of filling in the gaps. In T. Rex and the Crater of Doom, Walter Alvarez describes the way the impact crater was traced to a circular geological formation buried beneath the Yucatán Peninsula of Mexico, ending his story with the great astronomical event of 1994, when fragments of a shattered comet collided with Jupiter–an epilogue which makes it plain that uniformitarian geology has now admitted catastrophes to its canon. The book tells the story simply and well. Yet the central assumption, of a link between impact and extinction, is not addressed. The evidence for an impact at the end of the Cretaceous can no longer be doubted, and the physical calculations make it plain that its results would indeed have been catastrophic. But even if dinosaurs met a catastrophic end, rather than fading out in the long Mesozoic twilight, we will never be able to prove it. The evidence can only be circumstantial – we cannot go back and see what would have happened had the asteroid missed.
Alvarez starts his book by painting a picture of the awfulness of such an impact:
In the zone where bedrock was melted or vaporised, no living thing could have survived. Even out to a few hundred kilometres from ground zero, the destruction of life must have been nearly total. Sterilised by the intense light from shock-compressed air and from the fireball of rock vapour, crushed when pores and cracks in rock were slammed shut by the passing shock wave, and bombarded by the falling debris of the ejecta blanket, little or nothing was left alive in this central area.
This is not science; it is advocacy. Alvarez hopes to convince us that the consequences of the impact were so terrible that no dinosaur could possibly have survived them. Why other creatures, from turtles to mammals, should have made it through is explained in an ad hoc fashion, in terms of body size or greater numbers.
Other scientists are still plugging away at less glamorous ideas and recent work shows that matters were more complicated than Alvarez suggests. Volcanic eruption was intense at the end of the Cretaceous, and the climate was coming to the end of a long period of warmth and stability. Recent study by palaeontologists of the extinction records of a wide variety of animal and plant groups shows that although some became extinct at or near the K/T boundary, many others persisted unchanged.
It’s easy to forget how scarce fossils are, especially of large land animals such as dinosaurs; it is also easy to forget the scarcity of good, undisturbed K/T boundary sequences in settings that might contain dinosaur fossils. Virtually the only places on Earth where the two are known to occur together are in the foothills of the Rocky Mountains, in Montana and Alberta. The record there is equivocal. The few remaining dinosaur species could have perished suddenly, but the evidence is equally consistent with a gradual (though relatively rapid) decline over a couple of million years at the end of the Cretaceous, before the impact. The impact was the insult added to an injury already sustained, as a result of long-term climate change, possibly volcanic eruptions, and a host of subtle ecological interactions. Either way, results from a handful of localities can hardly be generalised over the whole Earth.