Around 250 million years ago a terrible calamity overtook life on Earth. Up to 96 per cent of all species became extinct, not overnight, but in a geologically brief span of time, maybe a few hundred thousand years. According to even the most conservative calculations, three-quarters of species disappeared at this time. Nothing like it has happened before or since. Palaeontologists have long been aware of this event, for it has left a strong imprint.
More than 150 years ago, this mass extinction was chosen as the boundary between the Permian and Triassic periods, one of the major divisions of the geological record. The changes in fossil faunas and lithologies made rocks from above and below the boundary distinct. As interest in mass extinctions has grown in recent years, the awesome scale of happenings at the Permian-Triassic boundary has become increasingly clear. Doug Erwin of the Smithsonian Institution in Washington DC has rightly called it the ‘mother of all mass extinctions’.
Advertisement
In 1980, Luis and Walter Alvarez of the University of California at Berkeley put forward the theory that the dinosaurs died off when a large meteorite collided with the Earth at the end of the Cretaceous period, around 64 million years ago. A violent end for such famous monsters is enough to catch anyone’s attention and this idea triggered increased interest in competing theories.
But this event, in which no more than a quarter of the world’s species succumbed along with the dinosaurs, was minor compared with the changes at the Permian-Triassic boundary. The Cretaceous-Tertiary debate continues, but now palaeontologists are focusing on the much bigger extinction 251 million years ago. Information is scanty, but it comes from the type of rocks deposited, the fossils they contain, and the types of carbon atoms that existed in the atmosphere 251 million years ago. The picture that is emerging is one of calamitous change that perturbed an otherwise stable biosphere – events that may have ominous parallels today.
The world was a very different place in late Permian and early Triassic times. Virtually all the continents were united in a supercontinent, Pangaea, that stretched uninterrupted almost from the North to the South Pole. On the opposite side of the globe, the vast Panthalassa Ocean stretched 220 degrees around the world. By comparison, today’s biggest ocean, the Pacific, reaches round only 130 degrees. Pangaea branched into northern and southern arms, separated around the equator by a large wedge-shaped sea called Tethys. This sea was warm and tropical. Carbonate sediments accumulated in the shallow waters around its edge, as they do now around the Bahamas and in the Caribbean. The sediments ultimately became the limestones that we now see in places such as the Italian and Austrian Alps, preserving fossils of the organisms that perished 251 million years ago.
Elsewhere, on the colder fringes of Pangaea, the Late Permian strata are mudrocks containing a lot of silica – once the skeletons of tiny radiolarian plankton. Within the supercontinent, the climate appears to have been hot and very dry; desert sands are a common rock type. Bays of the Tethys Sea around Pangaea frequently dried out, the water evaporating to leave large deposits of salt. Thick layers of salt that accumulated in Late Permian times lie beneath the floor of the North Sea today.
Life on and around the sea floor in late Permian times was very different from the same environment today. In contemporary sea floor communities, most stationary feeders live within the sediment, probably because there they can hide from predators and feed at the surface. But in Permian times a range of organisms lived and fed at a variety of different heights on and above the sea bed. Feeding space was divided into distinct but closely spaced levels that palaeontologists call tiers. The tiers probably arose through intense competition for nourishment in crowded seas. These complex tiered communities had thrived for a hundred million years or more; the mass extinction 251 million years ago eliminated them and changed the nature of the sea floor.
Before disaster struck, the sea communities contained certain characteristic organisms. Crinoids, also called sea lilies, filled the top tiers, raising their feeding cups on stems that measured as much as half a metre long and were anchored to the sediment below. The crinoids were a diverse and abundant group in Permian and older periods, but only a single genus survived into the Triassic period. Other echinoderm groups fared equally badly. Only a few species of sea urchins and starfish survived; one group, the blastoids, disappeared completely and is known only to a few palaeontologists.
The middle tiers were occupied by groups including the bryozoans and the rugose corals. The bryozoans, also called sea mats and sea mosses, form small colonies that spread like branches and fans. More than three-quarters of bryozoan families perished in Late Permian times, although their decline may have started slightly earlier. The rugose corals were even less fortunate; they were annihilated. These corals could live in isolation or in colonies, and many looked rather like modern corals although they are only distant relations. Rugose corals are common and distinctive fossils.
Organisms in the lowest tiers lived and fed at the sea floor. The principal occupants of a Permian seabed were brachiopods, which comprise the majority of specimens, and clams and snails. Brachiopods thrived right up to the end of the Permian period; 160 species belonging to 60 genera have been found in the youngest Permian rocks of south China. Nearly all these species vanished by the earliest part of the Triassic period, and only a few brachiopods survive to the present day, the little-known descendants of a group that once dominated the world.
Clams fared better, with little more than 10 per cent of their families disappearing. Snails have a curious fossil record across the boundary. Most disappeared from the fossil record in the Late Permian, but many reappear in middle Triassic rocks, having been ‘missing’ for more than 25 million years. Groups that apparently rise from the dead in this way have been named ‘Lazarus’ taxa, after the biblical character that Jesus reputedly brought back to life . Several other groups which survived this crisis contain Lazarus taxa, suggesting that the fossil record across the boundary between Permian and Triassic rocks is particularly poor – an issue to which we shall return.
Species that swam in the Permian seas and oceans were as hard hit in this mass extinction as those that lived on the bottom. At the smallest scale, it looks as if plankton were wiped out. For example, the silica shells, or tests, of the radiolarians contributed much to deep sea sediments from the latest Permian period, but they disappear abruptly when Triassic rocks appear.
The principal swimming invertebrates in the Permian seas were the nautiloids and ammonoids; the former survived virtually unscathed as a unique success amid the general annihilation. The latter were almost wiped out, although some of them may have been in decline before the major extinction. Only a few species survived to repopulate the Triassic seas, eventually diversifying into the many variations that characterise Jurassic rocks.
Taking the marine fossil record at face value, it seems that many groups were declining before receiving the coup de grace at the Permian-Triassic boundary. Indeed, many palaeontologists think that the decline was spread over the last ten million years of the Permian period. This idea has important implications: any mass extinction model that fits a slow decline over such a long time implies some gradual deterioration of the environment as opposed to a sudden change. But such a postulated gradual environmental decline may merely be an artefact of relatively poor preservation of the creatures that make up the fossil record.
Another characteristic of Late Permian times is a gradual retreat of the seas from the continental margins, perhaps to a level where shallow seas on the continental shelves reached an all-time low. These shelf seas, teeming with life, provided a rich assortment of fossils. When the shelf seas were least extensive, they left fewest traces in the rock record. Marine strata from latest Permian times are rare. With so few rocks to investigate it is not surprising that so few fossils have been found. The gradual decline in diversity may simply be reflecting this, giving a distorted view of the history of marine life at this time.
Erik Flugel of Erlangen University in Germany has evidence to support this theory. He has been collecting fossils from some of the rare sections in Greece and China where the latest Permian strata can be examined. Rather than containing a record of increasingly more impoverished communities, these areas show a marine fauna that was thriving until the end of the Permian period. Reefs built of algae, sponges and bryozoans rather than the corals of modern reefs, reached a peak of abundance and diversity in the very last Permian years. I have seen similar evidence in limestones of the same age in northern Italy. The clear implication from both these surveys is that marine life was thriving right up to the last moment of the Permian period, only to be devastated by some catastrophe.
The lie of the land
But evidence from 250 million years ago does not come solely from rocks formed under the sea. As the shelf seas began to shrink, the area of dry land grew. If the fossil record from the sea became increasingly poor because the area of shelf sea was decreasing, one might imagine that the terrestrial record should become correspondingly better. Unfortunately, the terrestrial fossil record is never anywhere near as good as the marine record because fossilisation is a much more uncertain process on land. Many organisms that die in the sea are soon buried by sediment, giving them a good chance of ultimately becoming fossils. Animals that die on land tend to lie around for a long time, exposed to the elements and scavengers. They rarely survive to be buried by the slow accumulation of sediment, and are consequently much rarer as fossils. This is the case around the Permian-Triassic boundary, but nevertheless the record is good enough to show sweeping changes among terrestrial creatures at more or less the same time as the changes in the sea.
The animals that inhabited the late Permian land lived in a range of ecological niches similar to those today. The synapsids, a group of reptiles showing some of the features we associate with mammals, were the dominant and most successful group. The pelycosaurs, primitive creatures looking like large modern lizards, often equipped with dorsal sails, were being replaced by the rapidly diversifying therapsids. This latter group contained the cow-sized dinocephalian therapsids, the gorgonopsians, large, ugly carnivores, the herbivorous dicynodonts and the therocephalians, mostly small, carnivorous creatures. Also appearing for the first time were another group of synapsids, the cynodonts, which were the most mammal-like of all these reptiles.
The mass extinction scythed through this diverse and flourishing fauna: 23 terrestrial families disappeared including 15 out of the 20 synapsid families. Only a few dicynodonts, therocephalians and the cynodonts survived. There appears to have been no sanctuary for these creatures; even the groups that lived over the widest area were affected and both small and large animals died out.
The flourishing life on land and sea in Permian times only serves to emphasize the effect of the extinction. Earliest Triassic rocks are surprisingly similar throughout the world. They usually consist of dull siltstones and dark grey shales, which have so far not inspired a great deal of research. These types of rock generally accumulate in deep water, where there is little energy supplied by currents, for example, to move heavier sediment particles.
The oceans and land of the earliest Triassic world must have been a curiously empty place. Palaeontologists know of very few species and very few individuals; the earliest Triassic fossils, in the monotonous siltstones, are exceptionally rare. What fossils there are tell a curious story. On the sea floor, the complex tiered communities of a few million years before had gone, replaced by simple communities with a single low level feeding tier occupied by clams and a few snails. This state of affairs persisted for more than 10 million years. It was not until middle Triassic times that these communities started to recover. For example, reefs are unknown in the early Triassic but they reappear in the middle Triassic, looking much the same as the late Permian reefs.
The prolonged failure of the bottom-living communities to recover from the mass extinction is as big a puzzle as the cause of the extinction itself. It is not a phenomenon seen following any other mass extinctions, and free-swimming species do not show the same hiatus. Ammonoids recovered fairly quickly and soon radiated into diverse forms. They were soon joined by several groups of terrestrial reptiles which returned to the sea in the form of plesiosaurs, ichthyosaurs and placodonts.
On land, a few genera of survivors dominated the earliest Triassic faunas. Lystrosaurus, a dicynodont, is the only significant large animal in the assemblages of this age from South Africa and Antarctica, for example. But within five million years, many groups, the cynodonts and stereospondyl amphibians in particular, had populated the world with a diverse range of animals once again filling the full range of ecological niches.
The abrupt changes that mark the transition from Permian to Triassic time are not limited to the fauna preserved as fossils. Sea level was exceptionally low at the end of the Permian period, leading to a very limited area in which shallow marine sediments were deposited. In the modern world there is a rough correlation between the number of species and the area available to them in which they can live. So many researchers have suggested that the marine mass extinction was a result of a loss of habitat area. Not only does this fail to explain the extinctions on land, but it is also inconsistent with more recent episodes of sea level fluctuation, as Dave Jablonski of Chicago University has cogently argued.
He has compiled data on the fortunes of marine fauna over the past few million years. Over this period, sea levels fell and rose dramatically as the ice sheets waxed and waned. Despite these changes, there has been no mass extinction. Jablonski believes that ocean islands can provide a refuge for marine life when sea levels fall, because of their shape. Most ocean islands are essentially volcanoes; they are shaped roughly like cones, so as the sea recedes, the surrounding area of shallow sea increases.
Steve Stanley of the Johns Hopkins University in Baltimore, Maryland, has put forward a different theory – that the mass extinction is related to a glacial period. The greatest ecological diversities come about in warm tropical climates, but when glaciation is at a peak, these areas contract and diversity falls. Unfortunately for Stanley’s proposition, however, there is only weak evidence of glaciation at the time of the boundary, and a major glaciation ended in middle Permian times.
There are other unusual events in the Late Permian period. Huge volumes of salts were precipitated from sea water in the shallow areas around the edges of the Tethys Sea. This has prompted several geologists, including Renato Posenato of the University of Ferrara, Italy, to propose that the late Permian oceans were correspondingly less salty than sea water today and may have been merely brackish. Brackish waters generally support only a small range of faunas, so such high evaporation could explain the mass extinction. But like the falling sea-level theory, this cannot account for extinctions on land. Also, evaporites (the salts deposited around the Tethys Sea) precipitated before the mass extinction: why was there a gap between precipitation and the decline of life in the seas? Another problem with the theory is that middle Triassic and Jurassic rocks also contain large evaporite sequences, without accompanying episodes of extinction.
A few years ago, a fourth theory was aired, albeit briefly, by several Chinese geologists. If a meteorite impact annihilated the dinosaurs at the end of the Cretaceous period, the argument went, a similar event could have created the Permian-Triassic boundary. To date there is no evidence to support this theory. One of the best indicators of a meteorite impact is an unusually high level of iridium, an element that is otherwise very rare in the Earth’s crust. But no extra iridium has been found in the Permian-Triassic boundary sediments. For a while, geologists ran out of ideas for the cause of the mass extinction. But now new data – measurements of the ratios of carbon isotopes in ancient rocks in particular – have revived the debate.
Carbon atoms come in two stable forms: carbon-12 atoms contain 12 nucleons (protons and neutrons) in their nuclei and carbon-13 atoms contain 13. Because these two isotopes have significantly different masses, they tend to fractionate in any chemical reaction. This means that they accumulate in varying proportions in the different carbon reservoirs at the Earth’s surface. In particular, organic carbon – living things and their remains – contains more of the lighter isotope than usual. So, when organic carbon is preserved in the geological record, in the form of coal and black shales, the carbon left on the rest of the globe will be proportionately richer in the heavier isotope. When plants and animals thrive on Earth, they lock up more of the lighter isotope, leaving atmospheric carbon, and carbon in, for example, ground water, containing a higher proportion of the heavier isotope. A record of the inorganic carbon ratios indicates how life has fared throughout geological history. Fortunately for geologists, certain types of limestones record this in the ratios of carbon isotopes when they formed.
Over the past few years William Holser, of the University of Oregon in Eugene and Mordeckai Magaritz, of the Weizmann Institute of Science in Rehovot, Israel, have been studying the variations in the ratio of carbon isotopes across the Permian-Triassic boundary. They have found some dramatic changes. In the Late Permian, carbon isotopes are very heavy, probably because coal had been forming over wide areas over the preceding hundred million years, locking away more carbon-12 than usual. Then, in the last million years of the Permian period, the proportion of carbon-12 increased rapidly. The ratio of carbon-13 to carbon-12 reached a minimum at the same time as the mass extinction.
The only feasible source of such a swing is the oxidation of a lot of coal and black shales, returning carbon-12 to the surface of the Earth and atmosphere. This is where the large fall in sea levels enters the story. A significant drop in sea levels would expose large areas of land, once underwater as the continental shelves, to erosion and oxidation. And oxidising organic matter increases the carbon dioxide in the atmosphere at the expense of oxygen. So much organic matter appears to have been oxidised in the last years of the Permian period that the oxygen levels in the atmosphere may have declined substantially. Calculations suggest that there may have been as little as half the present-day level of oxygen. This latest scenario could explain the extinction of the terrestrial vertebrates; active tetrapods need a high level of oxygen which the latest Permian atmosphere may not have been able to supply.
But then how did the marine mass extinction come about? Sea life can survive without high levels of oxygen. Most sea water contains more than enough dissolved oxygen for most metabolic requirements. Even if the oxygen content were halved this would have little effect on life. Oxygen only becomes a limiting factor when it drops below 20 per cent of normal values. At these levels, water has little oxygen available for living things. When there is no oxygen left the sea is described as anoxic, a situation threatened now for the North Sea. Such fatal conditions appear to have prevailed over the world in the earliest part of the Triassic period.
After the rapid fall in sea level at the end of Permian times, there was an even more spectacular rise – calculations suggest that this happened phenomenally quickly, as fast as several tens of centimetres per year. Tony Hallam, of the University of Birmingham, and I have focused research on the first sediments to be deposited in the Triassic seas. We found that the waters at the sea floor were anoxic or at best very poorly oxygenated. Anoxia and rising sea levels tend to occur together in the geological record, but no one is sure why. The telltale evidence in early Triassic rocks consists of myriads of tiny pyrite crystals, finely laminated sediments and a fauna consisting of a particular type of clam. This was a very specialised creature, that thrived with only low levels of oxygen. Finely laminated sediments form where mud settles gently on the sea floor; it can survive to become a rock with the layering intact only if worms and other burrowing creatures do not churn it up. If the layers persist, then this indicates that the sediment housed few creatures. Crystals of pyrite – iron sulphide – are a sign of reducing conditions at the sea floor; oxygenated water would have oxidised the sulphur to sulphate, forming other minerals.
So the Permian-Triassic mass extinction appears to be a story of death by suffocation for both terrestrial and marine life. This explanation best fits the available evidence but we are still left with many questions unanswered. For instance, why was there such a dramatic fall in sea level and then a rise across the Permian-Triassic boundary?
For those who are interested in relating past calamities to present day environmental changes, a parallel can be drawn with events around the Permian-Triassic boundary. The rapid combustion of fossil fuels is increasing carbon dioxide levels in the atmosphere. Although there may be more atmospheric oxygen around now than there was in the Late Permian, ensuring that we do not suffocate, global warming may produce a rise in sea levels similar to that in the earliest Triassic period. Let us hope not, for we do not wish to return to the stagnant, lifeless world of 251 million years ago.
Paul Wignall is a lecturer in palaeontology at the University of Leeds.
* * *
FOSSILS THAT RISE FROM THE DEAD
When palaeontologists claim to have found a mass extinction in the fossil record, what they are saying is that they have found a horizon where many species disappear. This disappearance could represent extinction, but it may not. It is fairly common for groups of fossils to vanish and then reappear at higher (younger) levels, after apparently being extinct for some time.
Such ‘born again’ species, known as Lazarus taxa, (and occasionally as ‘zombie taxa’) form an unrecorded component of the fossil record. The proportion of Lazarus taxa gives geologists a feel for the quality of the fossil record. The majority of species alive at times for which the record is particularly complete will be found as fossils; the unrecorded Lazarus taxa should constitute only a small percentage of the total. By this reckoning, the fossil record of Early Triassic times is exceptionally bad; many groups that we know existed then, because they form Late Permian and Middle Triassic fossils, are missing just above the boundary.
A related effect of the imperfection of the fossil record is ‘back-smearing’ of mass extinctions, making them appear more gradual than they are. To pick out a mass extinction properly, palaeontologists need to record all the species alive immediately beforehand. They can then compare this distribution with the impoverished faunas from after the event.
There is almost no chance of finding samples of all the species alive at any one time. What happens instead is that many of the species that are wiped out will make their final appearance in the fossil record some time before the actual extinction. Sudden extinctions are smeared out back in time and appear gradual. The mass extinction at the Permian-Triassic boundary has traditionally been considered a gradual process, but it could equally well be a sudden event, smeared out by the vagaries of the fossil record.