Fossils in Cretaceous rocks are quite different from those in Tertiary strata. The change is so striking that it marks a major boundary in geological time - between the Mesozoic and Cenozoic eras. Palaeontologists who follow the changes in fossil lineages up through the rocks have long been aware of the discontinuity at what is known as the Cretaceous/Tertiary (K/T) boundary. One reason why the fossil lineages could not be joined together would be that the crucial interval of time was missing, in other words that there was an unconformity, a break in the rock record, at the K/T boundary. But as geologists found more and more localities with Cretaceous and Tertiary rocks and still there was this abrupt break in the fossil record it seemed more likely that the record was correct and we were blind to what it was really portraying.
Walter Alvarez of the University of Berkeley was among the geologists impressed by the completeness of the geological record in the Apennines in central Italy. Here there are many hundreds of metres of limestones, representing deep ocean environments of the Cretaceous and Tertiary Periods. The K/T boundary was marked in the fossils by a major break between coarse limestones with large Cretaceous foraminifers and denser limestone containing the small earliest Tertiary forms. Between these two limestones was a layer of clay about a centimetre thick. Alvarez and his colleagues were curious about how much time that clay layer represented. Their method of determining how long it took for the clay to form involved measuring the concentration of the element iridium. This is rare in rocks of the Earth's crust; most of the iridium at the Earth's surface comes from dust constantly raining in on the planet from outer space.
The results surprised them all. There was far more iridium than anyone expected - more than was found in 3 metres of the underlying limestone. If the iridium had accumulated gradually, it would have taken around 340 000 years to have deposited that 1-centimetre clay layer. This sort of sediment would normally have taken around 13 000 years to be deposited. The Apennine layer would have needed a special flux of iridium - an observation which led the physicist Luis Alvarez with his son Walter, both of the University of California at Berkeley, and nuclear chemists Frank Asaro and Helen Michel at the Lawrence Berkeley Laboratory to develop the theory that the Earth had been hit by an extraterrestrial body at the time of the K/T boundary.
For many scientists the idea of an impact was not wholly unreasonable. There are plenty of asteroids in the Solar System with orbits that intersect the Earth's path, and a long record of cratering on the Moon and the Earth. Alvarez and his team calculated that bodies of around 10 kilometres diameter - big enough to supply the telltale iridium - hit the Earth every 100 million years on average; the energy released by such an impact would have been enormous - equivalent to 108 megatons of TNT. They suggested that an impact would send enough dust into the stratosphere to block light from the Sun for several months, stopping photosynthesis and breaking the food chain. The result, they said, would be just the sort of mass extinction seen in the fossil record at the K/T boundary.
Others were sceptical, for a number of good reasons. Many palaeontologists argued that several of the extinctions were gradual and began before the K/T boundary; they preferred an explanation based on gradual environmental change. The catastrophists countered that the gradual nature of the extinction record was a statistical artefact. Because the fossils were rare, the last examples of some of the key species had not yet been found.
There were also suggestions that the iridium could have come from deep-seated volcanic eruptions, which tap regions of the Earth that have kept the composition of the primitive Solar System and are as rich in iridium as a meteorite would be. For many geologists there were already plenty of satisfactory explanations for the extinctions that did not require recourse to something outside the range of present-day phenomena that are available as models.
The impact theory was never going to be accepted without an answer to the unresolved question: if an object 10 kilometres across really had hit the Earth at the end of Cretaceous time, then where had it fallen? If it fell on land where was the crater? And if the object had fallen into the sea, there was only a two-thirds chance that the impact site would be preserved, because the ocean floor is continually recycled back into the Earth's mantle.
Throughout the 1980s geologists scrutinised rocks from the K/T boundary. They found a number of interesting particles, worldwide, in rocks from land and sea. First, there were some small roundish objects termed spherules; they are generally less than 1 millimetre in diameter and now mostly made of an assortment of secondary minerals, replacing the original constituents. Most were spherical, but there were also some shaped like droplets and dumbbells. They resemble the glass droplets formed by impacts, called tektites, which condense from rock that has been vaporised or liquefied when hit by a large body.
Secondly, there were grains of quartz with a particular crystallographic structure. On a microscopic scale, the quartz is crisscrossed by several sets of narrow, parallel bands of glass, which mark the lines of shock fractures that form as a result of a sudden pressure wave. Similar features have been found in quartz around impact craters. Thirdly, unusual concentrations of iridium were found at many sites. Spherules have now been found at 71 sites, shocked quartz at 26 and iridium at 102.
It became apparent to geologists in Canada and the US that grains of shocked quartz were most abundant and biggest in the Rocky Mountains and that they became smaller farther away. This led to suggestions that the crater that produced these grains might be the poorly known structure that lies underneath the state of Iowa at a place called Manson. This still seems highly likely and the site is being examined again. But if there is a crater at Manson, it is not big enough. At only 35 kilometres diameter it would imply a bolide only around 4 kilometres across, too small to account for the extra iridium.
So the search for the main crater continued, with the focus shifting to Haiti in the Caribbean, at the approximate position of the K/T boundary. At one locality there was a thick bed of what had been called a volcanogenic turbidite - the name given to a bed of volcanic particles that had slipped down a submarine slope and tumbled into deeper water, rearranging the particles. Geologists examined particles from this bed and came to the conclusion that it was a bed of spherules associated with both shocked quartz grains and an iridium anomaly.
The Haiti bed was up to half a metre thick, around 20 times thicker than North American layer, and it had some of the largest spherules, 8 millimetres in diameter. But the spherules were particularly exciting because some of the original glass remained - enough for chemical and mineralogical analysis. The glass turned out to be almost free of volatiles, particularly water, a feature typical of impact glass. It had a wide range of chemical compositions that could be explained by a mixture of two source materials. On melting one formed a black glass with a composition similar to that of the Earth's crust, and the other was a much rarer, bubbly, yellow glass with a high component of the element calcium - a composition unknown from volcanic rock.
A further pointer towards an impact in this part of the world came with the reinterpretation of some sedimentary deposits at a K/T boundary site at Brazos River in Texas, by Joanne Bourgeois, a sedimentologist at the University of Seattle in Washington. If the bolide landed in the sea, it would have produced giant waves, tsunamis, which radiated outwards and stripped sediment from the adjacent continental shelves. This would have created a massive disturbance in the sedimentary record. At this site, the rocks were mudstones from relatively shallow water and the sequence was thought to be mostly complete, without unconformities. Fossils picked out the boundary, where there was also an unusual thick unit of what seemed to be coarse sandstone; at its base was fossil debris and pieces of sediment that had been ripped up from older layers below. The whole of the bed was made from sediment re-sorted at K/T boundary time. Bourgeois and her colleagues interpreted the unit as the much sought-after tsunami bed, the result of an impact at sea.
Other, similar units of the same age also occur in the region of the Gulf of Mexico and the Caribbean. But scientists were generally unconvinced that they were tsunami deposits, because the classical explanation for these sediments was also tenable. The beds formed at a time when sea level was falling in Europe and on the North Atlantic margins. The Brazos River sediments were thought to be part of the coarser coastal layer, which migrated offshore to lie on the muds.
The Gulf of Mexico and the Caribbean areas had become the centre of attention for the site of the crater. By now, of course, it would be covered by up to 65 million years' worth of younger sediment. The search had to employ indirect methods, such as maps of slight anomalies in the Earth's gravitational and magnetic fields. Several possible sites were later discredited. One, first suggested by Glen Penfield of the Mexican state oil company Pemex, remained as a good candidate - that on the Yucatan Peninsula in Mexico.
This showed up on both gravity and magnetic maps as a roughly circular structure of around 180 kilometres diameter, big enough to be formed by a body between 10 and 20 kilometres across. It was buried beneath several kilometres of carbonates, but three wells had been drilled into it in the search for oil in the 1950s. The drill had cut through Tertiary rocks near the surface, but company records showed that the oldest of these sediments, towards the bottom of the cores, contained some Cretaceous fossils. The three wells from the centre of the structure held a surprise - at their bases was a rock recorded as an andesite, or a volcanic, crystalline rock.
Unfortunately the core material was thought to have been destroyed in a warehouse fire, so geologists could not re-examine it. But one geologist, Alan Hildebrand, then a graduate student at the University of Arizona, was persistent. He hunted round for bits of material that had been given to specialists for additional identification and still kept at other institutions. Eventually he found some of the 'andesite' and found that it was a most peculiar rock containing glass and beautiful shocked quartz grains. It was, he thought, the melt rock from the bottom of the crater. With it there was a breccia, made of broken fragments of igneous and sedimentary rocks. The breccia was also found in wells outside the structure as a layer between Cretaceous and Tertiary sediments.
The sediments immediately above the andesites and breccias were examined again. They did indeed contain the small early Tertiary microfossils that had passed unnoticed amongst the large Cretaceous ones, which could have been reworked from material around the crater. There were also indirect signs, such as fossils in the Tertiary sediments from the inside of the structure coming from deeper water than those outside: in early Tertiary times there had been a deep hole in the Yucatan peninsula.
The theory was now mature. An extraterrestrial object around 10 kilometres across had hit the Earth at K/T boundary time leading to great environmental devastation and the death of the dinosaurs - and there had been a major impact which caused the Chicxulub crater on the Yucatan Peninsula of Mexico. At the time, carbonate sediments were being deposited under only a few metres of water on the edge of the continental shelf. The rock that the asteroid hit would have contained plenty of calcium carbonate, the source of the calcium-rich glass; the vaporised carbon dioxide could have led to global warming and acid rain. The impact of such a body would have caused massive earthquakes, perhaps hundreds of times bigger than the largest measured earthquake. Massive tsunami waves, probably caused by soft sediments collapsing down the adjacent continental slope, would have radiated outward from the Yucatan Peninsula. They would have scoured sand from the sea floor and dumped in its place a chaotic jumble of blocks and fine sand.
Unable to see the rocks of the crater itself, geologists homed in on the nearest outcrops where the K/T boundary was exposed at the Earth's surface. Walter Alvarez, Berkeley colleagues Alessandro Montanari and I, together with Jan Smit of the Free University of Amsterdam (who had arrived at the idea of the K/T impact at around the same time as the Alvarezes) decided to go to Mexico and look for outcrops.
By the end of Cretaceous time, the continental edge was inland from where it is now and northeastern Mexico lay under water perhaps 1 or 2 kilometres deep. The sequences from deep water were a mixture of mud brought down by rivers from the North American continent and plankton from surface waters; they were mostly complete, without unconformities. In fact the Cretaceous sediments of the Mendez formation are so remarkably similar to those of the overlying Tertiary Velasco formation in that part of Mexico that, apart from a very slight change of colour, it is almost impossible to tell them apart without looking at their fossils. Several complete K/T boundary sites have been described from that part of Mexico but nothing remarkable had been noted.
After we had visited several of these sites and found nothing unusual, we were ready to return. But on the last day we decided to visit a little-known locality which had been described in the 1930s by J. M. Muir, a geologist looking for oil. The site was in one of the river valleys, the Arroyo el Mimbral, where Muir had described an unusual sandstone bed. After several hours of driving on the highway and a couple of hours on rough dirt roads we approached our goal and immediately began to get excited. A thick, massive bed quite unlike anything we had seen in the rest of the Mexican K/T sequences was jutting out of the river bank. When we reached the outcrop, we found a thick bed of spherules underneath the sandstone bed - spherules that later proved to contain original glass with the same distinctive compositions as that from Haiti. And at the base of the sandstone bed there were large fragments of fossil plants.
Here, graphically displayed, was confirmation of the entire story, reading from bottom to top. With the impact of the bolide bits of the rock that it hit and droplets of molten material flew through the air from Yucatan and were the first to be deposited on the sea floor. The impact generated tsunami waves which washed away material from the shallow shores of the Gulf of Mexico and dragged it back into the deep sea. These waves took many an hour to travel, so this sediment was deposited on top of the spherules. The waves rocked back and forth across the Gulf of Mexico as they gradually died down, leaving gentle ripples in fine sediment on the top of the tsunami bed. These rocks contain the highest concentrations of iridium, which was in the finest dust, so it was the last to settle after the impact.
At the start of 1992, Alvarez, Smit and the team went back to northeastern Mexico with a clearer idea of what to look for. Searching in the field with geologists from Pemex, we found that this prominent sandstone bed - which can be traced on published geological maps - outcrops in patches over a large area. It is situated exactly at the K/T boundary as indicated by the fossils, and beneath it is a layer of spherules that ranges between centimetres and metres in thickness. Pieces of the Yucatan melt rock have now been dated using an extremely precise method based on the radioactive decay of potassium-40 to argon-40. They give the same answer - 65 million years - as glass from Haiti and Mexico. The Yucatan crater is exactly the same age as the K/T boundary.
In fact, there may be several craters of the right age. The two North American examples have received the most attention but recently the giant Popigay crater in Siberia, 20 kilometres across, was redated to have an age near to that accepted for the K/T boundary. There is evidence for several impacts, closely spaced in time. The large shocked quartz grains from the Rocky Mountains are more likely to have come from the Manson impact, but the Mexican-Caribbean glass probably came from Yucatan. In the Rockies, the boundary layer has two distinct parts, a lower sublayer with altered spherules and an upper one with iridium and shocked quartz. Between them there are casts of plant roots which must have taken at least one growing season to form.
A likely scenario would be a sequence of impacts, which could arise from several processes. A comet or a large meteorite breaking up as it rounded the Sun would bring a rain of objects over several years; separate impacts of objects from a comet shower could last for up to 1.5 million years, according to Piet Hut, an astronomer at Prince ton and his colleagues.
There are still dissenters from the whole idea of an impact. One camp, led by Charles Officer and Charles Drake at Dartmouth College, doggedly persist in interpreting the geochemical anomalies as a result of volcanism. There certainly was unusual volcanic activity around this time; one of the largest outpourings of lava onto the Earth's surface that has ever taken place formed the Deccan Traps of India, at roughly the time of the K/T boundary, but is thought to have started just beforehand. Specialists argue that this type of volcanism could not have produced the shocked quartz, the high iridium concentrations or the kind of glass found in the spherules.
Harder to convince have been the palaeontologists. The fossil record has been fiercely scrutinised, more data has been collected and much has been reinterpreted. This is particularly true for the dinosaurs, the group of fossils which most captures the imagination when thinking of K/T extinctions.
The record of latest Cretaceous dinosaurs comes essentially from North America. It is apparent that their diversity declined drastically from about 100 to 70 million years ago, although there now seems no evidence that the dinosaurs were in decline just before the K/T boundary. The death of those dinosaurs that remained is still therefore catastrophic.
The record from the deep sea is perhaps more significant; there are far more specimens and species of plankton than land animals, and these organisms were a vital part of the marine ecosystem. There have been numerous studies of the boundary record; some investigators still record a decline starting around 20 000 years before the boundary, but in recent years almost everyone has recognised that there was a mass extinction of overwhelming significance at this time and that most of the extinctions came at the boundary.
Nothing about the K/T boundary is simple, however. Two cases have emerged where more and better data has pushed the case the other way and shown that extinctions were gradual phenomena happening before the boundary. Both are from the shallow water fauna, specifically the inoceramid and rudist bivalves. For the inoceramids most of the extinctions were concentrated in a pulse of extinctions around 2 million years before the boundary. But the rudist bivalves show a steady, gradual decline in the final 10 million years of Cretaceous rocks. The decline coincides with the loss of their shallow-water habitat worldwide as the sea level fell and water drained back into the ocean basins.
Was it just bad luck that the Earth was bombarded by these extraterrestrial objects at a time when things were going so badly anyway? Why did it happen just when a deep-seated plume of lava finally reached the surface, to pour lava across the Indian continent, releasing large quantities of dust and volatiles? What was happening in the major marine regression with significant consequences for seawater chemistry and circulation patterns?
But the argument has heightened concentration on the world at the end of Cretaceous time, and the debate has changed focus. The question is now: what is the background and what the main extinction mechanism? Maybe impacts happen all the time, and the biosphere must already be in crisis for them to push it over the edge.
Nicola Swinburne is a postdoctoral researcher at the University of California at Berkeley.
All comments should respect the New Scientist House Rules. If you think a particular comment breaks these rules then please use the "Report" link in that comment to report it to us.
If you are having a technical problem posting a comment, please contact technical support.
18:53 05 September 2008
